FRTR & BH & Rem2

FROM REALISM TO RAPPROCHEMENT:
THE AUTOBIOGRAPHICAL INTERPRETATION
OF COLLINGWOOD'S PHILOSOPHY
Glenn C. Shipley
Loyola University of Chicago
PhD Dissertation
1983
ACKNOWLEDGMENTS
In writing this dissertation I have received support and

encouragement from many people. Fr. Edward Maziarz, my Director,
has been unflagging in his enthusiasm for the project and in his
belief in my ability to bring it to a successful conclusion. Dr.
James Blachowicz and Dr. Peter Maxwell have done me the great
honor of reading every word of it with critical intelligence and
many suggestions for improving both its style and content. And
if it were not for the patience, confidence and support of Dr.
Francis Catania, I would never have completed my doctoral
studies. I am also grateful to my typist, Jane Strom, whose
skillful and tactful work has made concluding the final stages
of this project immeasurably easier.
I owe a greater debt than words can express to my wife,

who has suffered the fate of a "dissertation widow" with love
and good humor; to my children, who accepted their father's
frequent disappearances to work on his "dissertation book;" and
to my mother, who knows what it means to go it alone to finish a
lifetime project, and who taught me all those things which I
rely upon without thinking, so that I can now be free to rely
upon my thinking.
ii
LIFE
The author, Glenn C. Shipley, is the son of Herman

Shipley and Frances (Brachle) Shipley. He was born November 15,
1938, in Villa Park, Illinois. He is married to Marilyn
(Sheridan) Shipley, and is the father of Gregory, William, and
Mark Shipley.
He graduated from St. Alphonsus Grammar School in 1952,

and from St. Michael High School in 1956, both of Chicago,
Illinois. In September of 1956 he was enrolled at Loyola
University of Chicago under a Loyola University Competitive
Scholarship. In 1962, as a member of the United States Air
Force, he attended the Armed Forces Language Institute at the
University of Indiana in Bloomington. In 1965 he returned to
Loyola, from which he received, in February of 1967, the degree
of Bachelor of Science with a major in philosophy.
He began graduate study in philosophy at Loyola

University under a Graduate Assistantship in February of 1967,
and in March of 1967 was awarded a three year National Defense
Graduate Fellowship (NDEA Title IV). In June of 1971 he received
the degree of Master of Arts. He was awarded an Arthur J.
Schmitt Dissertation Fellowship for the academic year 1972-1973.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . ii
LIFE . . . . . . . . . . . . . . . . . . . . . . . . . iii
LIST OF TABLES . . . . . . . . . . . . . . . . . . . .. viii
LIST OF IN-TEXT REFERENCES . . . . . . . . . . . . . .. ix
Chapter
PART I
I. INTRODUCTION: COLLINGWOOD AND HIS INTERPRETERS
1. Collingwood: His Life, His Writings, and His Era. . . 1

2. T. M. Knox and the "Radical Conversion Hypothesis". . 9
3. Collingwood's Interpreters: An Overview . . . . . . . 22
4. On Interpreting Collingwood . . . . . . . . . . . . . 42
II. THE AUTOBIOGRAPHICAL INTERPRETATION OF COLLINGWOOD'S

PHILOSOPHY
1. The Autobiography as Literary Evidence . . . . . . . 53

2. The Autobiography as Historical Interpretation . . . 60
3. The Autobiographical Interpretation: Four Themes . . 64
4. The Critique of Realism . . . . . . . . . . . . . . . 67
5. The Logic of Question and Answer . . . . . . . . . . 79
6. History and Philosophy. . . . . . . . . . . . . . . . 89
7. Rapprochement Philosophy . . . . . . . . . . . . . . 98
iv
PART II
III. REALISM AND IDEALISM
1. Introduction . . . . . . . . . . . . . . . . . . . . 106
2. Realism and Idealism in Religion and Philosophy . . 109
3. Idealism, the Absolute, and the Metaphysic
of Knowledge . . . . . . . . . . . . . . . . . . 124
4. Absolute Idealism and the Forms of Experience. . . . 128
5. Speculum Mentis and the Emergence of
Explicit Anti-Realism . . . . . . . . . . . . . 139
6. Conclusion . . . . . . . . . . . . . . . . . . . . . 148
IV. QUESTIONS AND ANSWERS, LOGIC AND DIALECTIC
1. Introduction . . . . . . . . . . . . . . . . . . . . 154
2. Abstract and Concrete Universals . . . . . . . . . . 158
3. Science and Supposal . . . . . . . . . . . . . . . . 167
4. Conclusion: Three Logics . . . . . . . . . . . . . . 181
V. HISTORY AND PHILOSOPHY
1. Introduction . . . . . . . . . . . . . . . . . . . . 191
2. The Identity of History and Philosophy . . . . . . 194
3. The Concrete Universal as Absolute Object . . . . 198
4. The Ideality of History a Scale of Forms . . . . .. 206
5. Conclusion . . . . . . . . . . . . . . . . . . . . 233
VI. RAPPROCHEMENT, RELIGION, AND ABSOLUTE KNOWLEDGE
1. Introduction . . . . . . . . . . . . . . . . . . . 243
2. Religion, Philosophy, and Incomplete Rapprochement . 247
3. Speculum Mentis: Rapprochement and Developing Series 265
4. Speculum Mentis: Retrogressive Identity . . . . . . 272
5. Speculum Mentis: Progressive Identity . . . . . . . 286
6. Disputed Questions . . . . . . . . . . . . . . . . . 293
v
PART III
VII. ANTI-REALISM AND THE PHILOSOPHY OF MIND
1. Introduction . . . . . . . . . . . . . . . . . . . . 320
2. Empirical Thinking and the Essay on
Philosophical Method . . . . . . . . . . . . . . 328
3. Sensation, Imagination, and Empirical Thought . . . 335
4. Attention, Freedom and Corrupt Consciousness . . . . 355
5. Idealism and the Limitations of Phenomenology . . . 366
6. The New Leviathan: Attention as a Linguistic Act . . 390
7. Conclusion . . . . . . . . . . . . . . . . . . . . . 415
VIII. LOGIC, LANGUAGE, AND MENTAL ACTS
1. Introduction . . . . . . . . . . . . . . . . . . . . 423
2. Questions, Answers, and Presuppositions . . . . . 427
3. Dialectical Logic and Philosophical Methodology . 458
4. Language and Logic in The Principles of Art . . . 518
5. Language and Mind: The New Leviathan. . . . . . . . 538
6. Conclusion . . . . . . . . . . . . . . . . . . . . . 592
IX. THE PHILOSOPHY OF HISTORY
1. Introduction . . . . . . . . . . . . . . . . . . . . 607
2. Anti-Realism and History
a) The Definition of History . . . . . . . . . . . 617
(b) The Outside and Inside of Historical Events . . 625
(c) Individuality, Universality, and the
Subject Matter of History . . . . . . . . . 636
(d) Historical Re-enactment . . . . . . . . . . . . 645
(e) The A-Priori Imagination . . . . . . . . . . . 655
(f) Re-enactment: Beyond Realism and Idealism . . . 667
vi
3. Evidence, Inference, and Necessity
(a) Historical Evidence . . . . . . . . . . . . . 694
(b) Historical Inference . . . . . . . . . . . . 704
(c) Historical Necessity . . . . . . . . . . . . 712
4. Conclusion . . . . . . . . . . . . . . . . . . . . 719
X. METAPHYSICS AND RAPPROCHEMENT
1. Introduction . . . . . . . . . . . . . . . . . . . 721
2. The Autobiographical Themes in the Essay on
Metaphysics . 736
3. Obstacles to Understanding Collingwood's
Reformed Metaphysics . . . . . . . . . . . . . 752
4. The Unity of the Autobiographical Themes . . . . . 767
5. The Rehabilitation of Reformed Metaphysics . . . . 795
(a) From Anti-Metaphysics to Reformed Metaphysics .797
(b) Metaphysics as an Historical Science . . . . 805
(c) The Absoluteness of Presupposing . . . . . . 815
(d) Ontology and Reformed Metaphysics . . . . . . 825
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . 836
LIST OF TABLES
1. The Interpretations of Collingwood's Philosophy . . . . 41
2. The Critique of Realism . . . . . . . . . . . . . . . . 77
3. Question and Answer Logic . . . . . . . . . . . . . . . 84
4. The Principles of History . . . . . . . . . . . . . . . 91
5. Rapprochement Philosophy . . . . . . . . . . . . . . . 103
6. The Meanings of History . . . . . . . . . . . . . . . . 240
7. The Knower-Known Relationship in Perception . . . . . . 388
8. The Philosophy of Mind . . . . . . . . . . . . . . . . 419
9. Question-and-Answer logic in The Essay on Metaphysics . 430
10. Philosophical Method . . . . . . . . . . . . . . . . . 463
11. Language and Expression in The Principles of Art . . . 520
12. Language, Logic, and Mental Acts in The New Leviathan . 540
13. The Presuppositions of the Analysis of Mental Functions 557
14. Metaphysics and Anti-Metaphysics . . . . . . . . . . . 747
viii
LIST OF IN-TEXT REFERENCES
In this dissertation references to Collingwood's pub-

lished philosophical works are made in the fashion that has
become accepted by Collingwood's interpreters. The references
are made in-text, by abbreviation of the title of the work
followed by the page or pages on which the reference is found.
These abbreviations are as follows:
A An Autobiography. London: Oxford University Press, 1939.
EM An Essay on Metaphysics. Oxford. The Clarendon Press, )97~.
EPA Essays in the Philosophy of Art. Edited by Alan Donagan.

Bloomington: Indiana University Press, 1964.
EPH Essays in the Philosophy of History. Edited by William

Debbins. New York: University of Texas Press, 1965.
EPM An Essay on Philosohical Method. London: Oxford University

Press, 1933.
FR Faith and Reason: Essays in the Philosophy of Religion.

Lionel Rubinoff, editor. Chicago Quadrangle Books, 1968.
IH The Idea of History. T. M. Knox, editor. London: Oxford

University Press, 1946.
IN The Idea of Nature. T. M. Knox, editor. London: Oxford

NL The New Leviathan. London: The Clarendon Press, 1942.
PA The Principles of Art. Oxford: The Clarendon Press, 1938.
RP Religion and Philosophy. London: Macmillian and Company,

1916.
SM Speculum Mentis. Oxford: The Clarendon Press, 1924.
ix
In addition the following standard works on Collingwood are also
given in-text references:
CEPC Krausz, Michael, editor. Critical Essays in the Philosophy

of R. G. Collingwood. Oxford: Clarendon Press, 1972.
CRM Rubinoff, Lionel. Collingwood and the Reform of

Metaphysics. Toronto: University of Toronto Press, 1970.
FYC Johnston, William M. The Formative Years of R. G.

Collingwood. The Hague: Martinus Nijhoff, 1967.
LPC Donagan, Alan. The Later Philosophy of R. G. Collingwood.

Oxford: The Clarendon Press, 1962.
MHD Mink, Louis O. Mind, History, and Dialectic: The Philosophy

of R. G. Collingwood. Bloomington: Indiana University
Press, 1969.
FROM REALISM TO RAPPROCHEMENT: THE
AUTOBIOGRAPHICAL
INTERPRETATION OF COLLINGWOOD'S PHILOSOPHY
Glenn C. Shipley
Loyola University of Chicago
During his lifetime at Oxford, the English philosopher,

R. G. Collingwood (1889-1943), published a number of books both
in history and in philosophy, including An Autobiography (1939),
in which he offered a self-interpretation of his intellectual
development. This dissertation is an attempt to measure the
degree to which the Autobiography can successfully serve as an
interpretation of the whole of Collingwood's published
philosophy.
In Part I the problem of interpreting Collingwood's

philosophy is surveyed, and it is shown that the main problems
to which Collingwood's interpreters address themselves are those
which Collingwood also discusses in the Autobiography. These
problems cluster around four major themes: the realism-idealism
controversy, question-and-answer (Q-A) logic, the philosophy of
history, and philosophical rapprochement.
In Part II the early writings (1916-1932) are examined

on each of these issues. After exhibiting an early tolerance of
realism, Collingwood's anti-realism surfaced in Speculum Mentis,
along with a commitment to absolute idealism. While there is no
evidence of a systematic Q-A logic in his early writings, the
epistemic functions of questioning and asserting are clearly
present and indicate an incomplete resolution of problems on
logic and methodology. Similarly the philosophy of history and
the notion of rapprochement identity show a steady development,
but leave serious ambiguities unresolved.
In Part III the later writings (1933-1946) are surveyed

on each of these same issues, and it is shown that Collingwood's
anti-realism not only remained constant, but hardened--which led
to certain difficulties in his analysis of perception. Q-A
logic, rather than being an alternative to formal logic,
actually functions as an interpretative methodology, while the
dialectical logic of the Essay on Philosophical Method becomes
the criteriology of rapprochement completion in all his later
works. In the final chapter of Part III the unity of the
autobiographical themes is reconstructed and used to overcome
the obstacles to a reformed metaphysics as an historical science.
FROM REALISM TO RAPPROCHEMENT: THE AUTOBIOGRAPHICAL
INTERPRETATION OF COLLINGWOOD’S PHILOSOPHY
Copyright, 1983: Glenn Shipley, PhD

PART I
THE PROBLEM OF INTERPRETING

COLLINGWOOD'S PHILOSOPHY
Chapter I
INTRODUCTION: COLLINGWOOD AND HIS INTERPRETERS
1. Collingwood: His Life, His Writings, and His Era.
Robin George Collingwood (1889-1943) was the youngest of
four children and the only son of W. G. Collingwood (FYC, 6).
His parents were artists, amateur archeologists, and friends of
John Ruskin. The elder Collingwood was first Ruskin's student,
later his personal secretary and confidant. and after Ruskin's
death in 1900 he became his biographer. The young Robin grew up
in an atmosphere heavily influenced by the Ruskinean ideal of
the universal man (FYC, 1, 17-36, 143-46).1 Like Ruskin himself
and also the young J. S. Mill, Collingwood was educated at home
until he was thirteen (partially due to the poverty of his
parents); and also like Mill, he was started on classical
languages at an early age--Latin at four and Greek at six (A,
1). He was able to read the English proofs for his father's
books by the age of five (FYC, 7), and by the time he started at
____________________
1 An appreciation of the degree to which Collingwood
valued Ruskin's thought can be gained by a reading of
Collingwood's own assessment of it in his early essay, "Ruskin's
Philosophy," delivered as an address at the Ruskin Centenary
Conference, 1919, but published in 1922 and reprinted in EPA, 1-
41.
2
Rugby at thirteen he could read and speak German and French
almost as easily as English (A, 6). At Rugby he taught himself
enough Italian to read Dante in his spare time (A, 7). As a
child he also accompanied his parents on archeological
expeditions: he claims to have attended his first "dig" as a
three-week old infant --in the toolbag of his parents (A, 80).
From Rugby Collingwood won a scholarship to University
College, Oxford, which he attended from 1908 until he graduated
in 1912 with "Firsts" in both Classical Moderation and Literae
Humaniores. He was hired as a tutor at Pembroke College while
still wearing his scholar's robes from final examinations.2 By
1913 he was given an independent hand at excavations on Roman
ruins in England, and was already being regarded by Haverfield
(his Oxford mentor in the subject) as his successor. After the
outbreak of war in 1914 he entered the British Admiralty
Intelligence where he remained until the end of the war in 1918.
During this time he wrote the manuscript for Religion and
Philosophy, which was published in 1916.3 In 1918 he married
Ethel Graham, moved his quarters to a country house, and began
an extremely active academic life as lecturer and tutor in both
philosophy and Roman and British History. In 1934 he was
relieved of some of this burden of tutoring and teaching when he
was appointed to
__________________
2 R. B. McCallum, "Robin George Collingwood,"
Proceedings of the British Academy, XXIX (1944), p. 463.
3 Ibid., p. 464.
3
the chair of Waynflete Professor of Metaphysical Philosophy.
From 1918 to 1934 he published four books--Roman Britain (1923,
revised edition in 1934), Speculum Mentis (1924), The Archeology
of Roman Britain (1930), and An Essay on Philosophical Method
(1933)--and numerous short articles and monographs.4
The last-mentioned book was produced during a leave of
absence from his teaching responsibilities in 1932, due to a
prolonged illness (A, 117-18)--the beginning of the ill-health
against which the remainder of his life was to be an heroic
struggle (FYC, 12). Between 1935 and 1941, when he resigned his
professorship, Collingwood managed to find time to do the
writing on archeology, history and philosophy that form the bulk
of his later published writings--Roman Britain and the English
Settlements (1936, with J. N. L. Myres), The Principles of Art
(1938), An Autobiography (1939), and An Essay on Metaphysics
(1940).5
Most of the material posthumously published as The Idea
of Nature (1945) and The Idea of History (1946) was also written
during this period, in the form of lectures (IN, v; IN, v-vi).
The New Leviathan (1942), the final work published during his
lifetime, was written under extreme
________________
4 T. M. Knox, "Notes on Collingwood's Philosophical
Work, with a Bibliography," Proceedings of the British Academy,
XXIX (1944), pp. 469-75.
4
stress, Collingwood’s state of health rapidly deteriorating, and
England being torn by war; its final chapters, he notes, were
written during the bombardment of London (NL, v; LPC, 316). In
1939 Collingwood felt well enough to sign on as First Mate on a
sailing yacht, the Fleur de Lys, for a trip from the coast of
France to Greece and the Greek islands--the other members of the
crew being mostly Oxford students. His account of that journey--
the First Mate's Log--was published in 1940 by Oxford University
Press in a limited edition. In 1942 his marriage with Ethel
Graham was dissolved at his wife's request (they had two
children, a son and a daughter), and he married Kathleen Frances
Edwardes, who bore him a daughter (EPH, x-xi). But his health
never fully regained its vigor, and he was forced to retire to
Coniston, to the house he inherited from his father, where he
died in 1943 of pneumonia.6
The years through which Collingwood lived and worked at
Oxford were among the most violent and revolutionary that Europe
has seen, and certainly the most profoundly threatening that
England has endured. His work spans the first half of the
twentieth century, bracketed at one end by the First World War
and the Russian Revolution, and at the other by World War II and
the Fascist holocaust. At Oxford the intellectual climate was no
less subject to violent upheavals. When Collingwood began there
in
_______________
6 McCallum, p. 468; cf. IN, xxi.
5
1908, the 19th century British idealist movement begun by T. H.
Green, with F. H. Bradley and B. Bosanquet its most respected
spokesmen, had just about spent itself, and a realist reaction
headed by Cook Wilson was already in full swing (A, 15-21). By
the time he retired his professorship in 1941 realism had given
way to logical positivism (A, 52). H. J. Paton writes that at
Oxford, in the period between the wars, "Collingwood and I were
the only representatives of our generation--a slender bridge
between predecessors at least ten years older and successors at
least ten years younger."7
In his Autobiography Collingwood corroborates this
remark when he writes that in the area of archeology he was the
only remaining Oxford resident trained by Haverfield as a
Romano-British specialist, and therefore felt an obligation to
keep alive that branch of studies which was left vacant when
Haverfield died in 1919, because most of his students had died
during World War I (A, 120).
In such highly troubled waters Collingwood felt himself
to be the vessel not only of Oxford Romano-British archeology
but also of philosophy. In philosophy the burden was even
greater, and in carrying it (alone, he felt) Collingwood was
buffeted by all the prevailing winds of
__________________
7 H. J. Paton, "Fifty Years of Philosophy," in
Contemporary British Philosophy, Third Series, ed. by:H. D.
Lewis (New York, 1956/1961), p. 345.
6
his era. Consequently even though his early works (Religion and
Philosophy, Speculum Mentis, and even the Essay on Philosophical
Method) have led to his rejection by contemporaries as a latter
day idealist, the body of his later writings continues to arouse
the interest of people of very different philosophical
persuasions--perhaps because many of the sources of these
persuasions were also influential on Collingwood himself.
Thus in the last of his books published during his
lifetime, The New Leviathan, one finds evidence relating him to
most of the major contemporary schools of thought. For example:
(1) pragmatism: "Reason is always essentially practical; because
to be reasonable means to be interested in questions beginning
with 'why'; and this happens because people crave for
reassurance against the fallibility of their knowledge" (NL,
14.31); (2) phenomenology and existentialism: "Man as mind is
whatever he is conscious of being" (NL, 1.84; emphasis his); (3)
linguistic analysis: "Language is not a device whereby knowledge
already existing in one man's mind is communicated to another's,
but an activity prior to knowledge itself, without which
knowledge could never come into existence" (NL, 6.41); and (4)
even Marxism: "Is there nowhere such a thing as 'purely
theoretical thinking'? There is; but it is not real thinking,
and it does not lead to real knowledge . . . . Real thinking . .
. always starts from practice and returns to practice; for it is
based on 'interest' in the thing thought about" (NL, 18.13).

7
Perhaps because such diverse inclinations are reflected
in his philosophy, he has been claimed for, and damned by, most
of these same schools of thought--and this is reflected in the
diversity of interpretations concerning his philosophy in the
growing body of secondary literature about him. Here one finds
him claimed not only by representatives of the traditions just
mentioned, but also (incredibly enough) logical positivism,
radical empiricism, idealism, cultural anthropology, and systems
theory. His roots have been located in Ruskinean moralism,
German Hegelianism, English and/or Italian idealism, and Cook
Wilsonean realism. Affinities have been found between his
philosophy and that of Ryle, Strawson, Wittgenstein, Dewey,
Husserl, Kierkegaard, Barth, and Sartre. And his best work has
been said to be in history, the philosophy of history,
esthetics, the philosophy of mind, metaphysics, ethics,
epistemology, and the history of ideas.
In spite of all these affinities, and in spite of the
remarkable breadth of his interests in this age of
specialization, he fits neatly into none of the contemporary
schools of thought, and succeeded in developing no appreciable
following of his own (cf. MHD, vii-viii, and FYC, vii, 137-46).
This is at least partially a matter of choice on Collingwood's
part: he sought no following, refused to engage in public
debate, and preferred taking his case in writing directly to the
public
8
(A, 56, n. 1, and A, 118). It is also partly due to the
circumstances
of his life: with positivism and analytic philosophy coming into
prominence in his own university, and Fascism and Communism
vying for dominance in the political arena around him,
Collingwood's attempts to steer an independent course led to his
estrangement from nearly all of his contemporaries. Even those
close to him felt a little irritated at him for being "rather
too quick in claiming all knowledge as his portion"8 and often
those who disagreed with him to his face were told, as it were,
to "bathe in Jordan."9 And finally, his isolation is very much
the result of the diversity and incompleteness of his output,
which makes it difficult to find a single insight that unifies
all of his multi-faceted output, or even to find a capstone to
complete the arch.
But whatever the reason for the difficulty in achieving
a clear focus in the surviving portrait of Collingwood's
thought, a portion of the blame for this difficulty must be
shared by the first interpretative authority to reflect on the
whole of Collingwood's output, and to this problem we must now
turn.
_________________
8 Paton, p. 345.
9 McCallum, p. 466.
9
2. T. M. Knox and the "Radical Conversion Hypothesis."
Before Collingwood died in 1943 he named T. M. Knox, a
friend and former student, as his literary executor.9a
Collingwood's will authorized his executor to publish only as
much of his unpublished writings as met high standards--thus
leaving to Knox's judgment what the public should see of the
unfinished works (IH, v). This material included Collingwood's
lecture notes on the philosophy of history, the philosophy of
nature, and ethics, and essays on philosophical theology and
cosmology with which he closed lectures on ethics and on the
philosophy of nature respectively (IN, v; IH, v-vi; CRM, 397).
In addition to this material there was an incomplete sketch of
The Principles of History which Collingwood hoped to be his
magnum opus, but which is now forever lost--and a large number
of other unspecified materials, not available for public
inspection.l0
9a According to David Boucher, Knox was chosen to edit

several works only after Collingwood’s death – see “The
Principles of History and the Cosmology Conclusion to the Idea
of Nature,” Collingwood Studies Vol 2, 1995, pp 142-46.
10 In 1969 I wrote to the Delegates of the Clarendon

Press, Oxford, for information about the status of these
manuscripts. They forwarded my letter to Collingwood's widow
(his second wife), who replied to me as follows: "Dear Mr.
Shipley, The Clarendon Press, Oxford, has passed your letter on
to me. There are a considerable number of unpublished papers of
R. G. Collingwood in my possession. Within the next year or so I
hope to deposit these in the Bodleian Library, Oxford. A number
of manuscripts will be reserved for some years. The incomplete,
unpublished Principles of History is lost. Yours Sincerely, Kate
Collingwood." In May of 1972 I received word from the Bodleian
Library that they still had not received any of the manuscripts
promised to them by Mrs. Collingwood, and stating that they had
been unable to contact her by mail. In June of 1979 I received
word from the Bodleian Library that they had recently received
the manuscripts from Mrs. Collingwood, and that
10
Under the titles of The Idea of Nature and The Idea of History
Knox published only the nearly completed lectures on cosmology
and history. He appended a part of The Principles of History
(actually only a portion of the first third of what Collingwood
had planned to write (IH, vi) and several completed essays
(delivered originally as lectures)as "epilegomena" in The Idea
of History, but omitted doing the same for The Idea of Nature on
the grounds that Collingwood seems to have become dissatisfied
with it: for the sketch of his own cosmology which had closed
his original lectures, Collingwood had substituted a shorter
concluding passage when he set about revising these lectures for
publication sometime after 1939 (IN, v.).ll
___________________
they filled the equivalent of five boxes measuring 14 x 11 x 4
inches "and are packed tight." A partial listing of their
contents included translations, letters, lectures, and notes on
everything from idealism and realism to the epistemology of
logic and English folklore. Most of the material is open to
inspection by scholars but, according to the terms set by Mrs.
Collingwood, is not to be photocopied.
1l There is something odd about Knox's editorial

judgment in this matter. Knox argues (as we shall see in a
moment) that Collingwood's best work was done between 1928 and
1936, and that after his radical conversion to historicism
between 1936 and 1938 his judgments were unsound and not to be
trusted in matters philosophical. On these grounds Knox
published the lectures on the philosophy of history and the
philosophy of nature, and included the terminal essays on
history that form the "epilegomena" to The Idea of History. But
on these grounds the essay on cosmology which closed
Collingwood's lectures on the philosophy of nature in 1934 and
1937 is a product of his mature, middle period. But instead of
publishing it as an "epilegomenon" to The Idea of Nature Knox
accepts Collingwood's later, possible judgment (which is
supposedly unsound) to omit this terminal essay. Instead Knox
published the short concluding piece which argues that
11
When Knox published The Idea of History in 1946 he added
an "Editor's Preface" in which he not only explained significant
editorial details about the manuscripts, but also proposed an
interpretation of the whole of Collingwood's philosophy, on the
basis of which he evaluated the posthumously published works,
and placed them in their setting within the context (as he saw
it) of Collingwood's entire published output. Therefore the
significance of the totality of Collingwood's published works to
date has rested upon Knox's judgment in both constitutive and
retrospective senses: constitutive because two important works,
including their present form, were directly due to his editorial
labors; and retrospective insofar as his account of
Collingwood's development is both (a) the only evidence that the
public has for the "high standards" that Knox used in deciding
which works should be suppressed and which deserved publication,
and (b) the only justification of these standards, based on the
interpretation of the group of writings that Collingwood did
publish in his own lifetime, that Knox offered for extending
that total output.
Knox's interpretation of Collingwood's development can
be outlined as follows:
________________
science "depends on historical thought for its existence" (IH,
177). But this is inconsistent with the editorial policy of The
Idea of History.
12
(1) Dividing Collingwood's writings into three groups,
Knox finds in the "juvenilia" (Religion and Philosophy (1916)
and Speculum Mentis (1924) evidence of skepticism and
dogmatism:12 religious skepticism insofar as religion is
described (in Speculum Mentis) as an erroneous mistaking of
imagining for thinking, and philosophical dogmatism insofar as
only philosophy is asserted as providing the full truth for
which religious assertion is only the symbol (IH, xiv-xv). Knox
says that the shift from the earlier work (in which religion,
theology and philosophy are identified) to the position in
Speculum Mentis marks the ascendancy of a dogmatic strain in
Collingwood's thought which "affected its content and . . . was
linked with a change in his attitude to religion, always one of
his strongest interests" (IH, xv).
(2) By 1932, during the "middle period" (as in
Collingwood's masterpiece, the Essay on Philosophical Method of
1933) the mature and
____________________
12 Knox does not specify what he means by dogmatism and
skepticism. But it seems that for Knox (1) one may understand by
"dogmatism" the imposition of an external source or standard of
truth on the internal doctrines of a body of knowledge; and (2)
"skepticism" to mean the failure or refusal to provide a
criterion for truth or falsity within a body of knowledge. That
is why historicism is a skepticism for Knox: history cannot
provide a criterion of truth or falsity for philosophy.
Therefore any criterion of truth which proceeds from, or is
grounded in, presuppositions which are themselves unquestioned
and/or unquestionable, is dogmatic; and any body of knowledge
resting on presuppositions which cannot themselves be judged to
be true or false is skeptical. Cf. IH, xvii.
13
undamaged mind of Collingwood embraced metaphysics as a separate
and distinct study of the one, the true, and the good (IH, xi).
This would have allowed for a truth-criterion (hence escaping
skepticism) not itself based on the unquestioning acceptance of
religious doctrine (hence not dogmatic). Knox thinks this much
is indicated in Collingwood's essay, "Faith and Reason," which
in assigning independent functions to each faculty escaped the
ascription of a monopoly of truth to any discipline (dogmatism)
as well as the denial of a truth universal and valid for all
thought (skepticism) (IH, xvi).
(3) But in the 1940 Essay on Metaphysics (and the
remainder of his later philosophy, his third period) Collingwood
lapsed back into the latent skepticism and dogmatism of his
youth by denying the independent status of metaphysics (which is
reduced to history). His "reform of metaphysics" is based upon
reducing the metaphysician's task to the historical work of
discovering the "absolute presuppositions" of science in a given
era. Since these absolute presuppositions were characterized by
Collingwood as (a) themselves neither true nor false, and (b)
religious in nature, they indicate a radical change in
Collingwood's mature position. This time Collingwood proposed a
philosophical skepticism and a religious dogmatism. Absolute
presuppositions are dogmatic as unquestionable and religious
(that is, held by an act of "unquestioning acceptance" or

14
"natural piety"--i.e. faith); and they indicate a kind of
skepticism insofar as they rest on no higher criterion of truth
and are themselves neither true or false (IH, xv-xvi).
Knox groups The Idea of History and The Idea of Nature
in the second, mature period, and The Principles of Art
overlapping the second and the third. He claims to have
documentary evidence that in Collingwood's second period he
still held that metaphysics as an autonomous branch of knowledge
was possible, and that in the third period not only is
metaphysics declared to be an historical science, but
"philosophy as a separate discipline is liquidated by being
converted into history" (IH, x; cf. xi-xii).
Unfortunately the work from which these quotations are
taken is The Principles of History which was never completed,
never published, and is now lost and so incapable of being
publicly examined; so there is at present no documentary
evidence to verify these statements. Even assuming that Knox is
a reliable authority, it is impossible to evaluate these
fragments without the full textual context--to see in what ways
Collingwood meant them to be taken, or qualified them, or
posited them as provisional assertions to be later corrected and
modified, etc. In future chapters we shall find direct evidence
in Collingwood's writings that he

15
often employed a dialectical strategy of positing a remark as a
starting point, then modifying the remark, and finally
contradicting it altogether--all in the span of one essay.
Unless we are to remain "scissors and paste" historians, and
therefore show that we have not learned even the first lesson
that Collingwood wished to teach us, we cannot take Knox's
fragments from Collingwood's "nachlass" manuscripts uncritically.
We are therefore left with Knox's arguments, based on
available evidence, concerning Collingwood's radical change of
mind--which Knox says occurred somewhere between 1936 and 1938
(IH, xi). The failure to acknowledge this change of mind is one
of the reasons Knox rejects Collingwood's Autobiography as a
reliable account of his development--the other being that in the
Autobiography Collingwood seems to wish his readers to believe
that he had worked out his theory of absolute presuppositions
and the purely historical character of metaphysics prior to 1932
and the Essay on Philosophical Method. Regarding the latter
point (1) Collingwood made no such claim in the Autobiography--
he says only that "these ideas . . . became clear to me soon
afterward," i.e. after returning to Oxford in 1918, and he says
nothing about any reduction of philosophy to history (Knox even
hedges by calling the claim about the dating of the discovery of
absolute presuppositions an "inference" that Collingwood wished
his readers to make from the text); and (2) even if

16
the claim were made, it is arguable that it has its roots in
doctrines already present not only in his early period, but in
the works of his mature middle period as well.
Supposing that one were to have unimpeachable evidence
that there was a radical change of mind as Knox says there was:
what reasons does Knox assign for it? In his "Editor's Preface"
he gives no less than three: (1) Collingwood's mind changed:
beginning in 1932, "tiny blood-vessels began to burst in the
brain, with the result that the small parts of the brain
affected were put out of action" (IH, xxi); (2) he changed his
mind himself: he came to think, like Croce, that "philosophy as
a separate discipline is liquidated by being converted into
history" because "the skeptical and dogmatic trends, present in
Collingwood's earlier thought, triumphed over the temporary
defeat they had sustained between 1932 and 1936" (IH, xi)); and
(3) Collingwood was an inconsistent, even ficklel3 thinker:
"Collingwood believed in the coincidentia oppositorum, as many
passages in his writings testify. I am suggesting that his own
later philosophy provides a striking illustration of this
phenomenon" (IH, xvii). "He brought a powerful mind to bear on
whatever happened to be engrossing his energies . . . and he
seems to have been inclined
________________
13 Knox does not use the term, "fickle." The term is
mine; the accusation is Knox's.
17
to draw the conclusion that philosophy was simply identical with
whatever he happened to be studying most intensively at the
time" (IH, xv)).
None of these reasons are acceptable. (1) The ad hominem
"brain pathology" explanation, while certainly verified by
documentary evidence, is not detailed enough to make any
accurate assessment of what portions of Collingwood's brain (to
say nothing of his mind or judgment) were affected. The evidence
in fact seems to point to brain damage in the motor areas:
writing to Croce in January of 1939, Collingwood says that "just
a year ago . . . I was partly paralyzed by a stroke which
deprived me of the power of speech . . . . I am making good
recovery: I can use my hand and foot moderately well, and can
speak now well enough for the purposes of my profession" (LPC,
316). Even if we were to assume that Knox's acquaintance with
Collingwood was so intimate that he could detect hemorrhagic
capillaries in Collingwood's brain as early as 1933, why should
we accept such an account as philosophically relevant? Why
indeed, when Collingwood was healthy enough to write the "second
book in his series," The Principles of Art (the first being the
1933 Essay on Philosophical Method) prior to his first
debilitating stroke in 1938, and healthy enough after his stroke
to act as First Mate on a sailing schooner which voyaged for
some months in 1939 in the Mediterranean--and to write a lively
account of
18
the experience? And why again when he was clear-headed enough to
write not only the Essay on Metaphysics, which so scandalizes
Knox, but also The New Leviathan, a work which many critics felt
would alone earn him a respected place in philosophical
literature? John Passmore's two-sentence estimate of the
situation is worth quoting:
It is sometimes suggested by Idealist admirers of

Collingwood that the brain disease from which he began to
suffer in 1933 is reflected in his ultimate heterodoxies.
When one contemplates the speculative freedom of these later
works, one can only wish that his contemporaries could have
been similarly afflicted.14
(2) Asserting that Collingwood's youthful skepticism and
dogmatism overwhelmed his better judgment is less a change of
mind than a relapse; interestingly enough, Knox suggests that
both Collingwood's earlier and later philosophy represent lapses
into a youthful realism, while it is only his middle period
which steered clear of the "shoals of skepticism and the billows
of dogmatism" (IH, xviii, xiv). But the clear evidence of the
Autobiography is that as of 1939 Collingwood interpreted the
whole of his philosophy as a response to the threat of
"realism." It is incredible to think, therefore, that his
literary executor could seriously entertain the hypothesis that
in the end he merely surrendered to its doctrines. Since
Collingwood was an historian of unimpeachable ability, who knew
well how to use evidence, and was fully aware that autobiography
________________
14 John Passmore, A Hundred Years of Philosophy (London,
1957), p. 306 n. 1; quoted by Rubinoff, CRM, 376 n. 17.
19
is history in which the subject happens to be oneself, if no
radical change of mind is recorded in the Autobiography it is
unlikely to be because of a lapse of memory, but rather because
Collingwood meant to deceive his readers. For what reason? Knox
fails to provide us with any.
(3) Knox's accusations of deliberate inconsistency and
vacillation by preoccupation also cannot go without challenge.
While it is easy to find passages which discuss the coincidentia
oppositorum (e.g. SM, 197-98, 249), there are others which
confine it to the scientific level of thought, where it is
contrasted with the synthesis of opposites (e.g. SM, 310). If by
a "coincidence of opposites" Knox means the simultaneous
affirmation of a pair of contradictory statements, then if
Collingwood asserted this he did indeed, as Knox charges, turn
traitor to his profession as philosopher. But the burden of
proof for this is on Knox: charges of "dogmatism and skepticism"
do not constitute sufficient evidence for the simultaneous
acceptance of a pair of contradictory statements. But if by
"opposites" Knox does not mean contradictories but contraries,
then what sense does the charge make that his later philosophy
is a prime instance of it? For it is just as true that
Collingwood accepted a coincidence of opposites as contraries in
his mature middle period--e.g. in the Essay on Philosophical
Method, where the

20
"overlap of classes" is described in terms of relations of
opposition and relations of distinction (EPM, 74-75).
Furthermore the charge of inconsistency by
preoccupational vacillation (what I have called "fickleness") is
self-refuting. Knox prefaced this remark with the observation
that it was the power of Collingwood's mind that caused him to
become so engrossed in his subject matter that he simply
identified philosophy with whatever he happened to be working on
at the time. But the charge does not bespeak a powerful mind but
a weak one--drifting this way and that according to what
"happens" to occupy it. This suggests an erratic and drifting
route for Collingwood's rudderless vessel. Opposed to this
charge (perhaps the most insulting that Knox levels at him) we
have Collingwood's autobiographical account of the logic of his
philosophical program, which leaves little room for topics to
merely "happen" to occupy his interest. He writes there that he
planned a series of books, beginning with the Essay on
Philosophical Method and continuing with The Principles of Art,
and that he planned (as of 1939) to devote all his remaining
time and energy to completing the series (A, 117-19). The
interpreter faced with a choice between Knox's version and
Collingwood's own account of the development of his thought
might prefer Knox's, but then he must supply convincing reasons
for rejecting Collingwood's. But Collingwood's version promises
to be systematic, and
21
Knox's to be haphazard and disjointed. Therefore an interpreter
cannot be blamed for choosing Collingwood's on these grounds
alone.
But when stripped of its pathophysiological banalities
and uncritical appeal to unexaminable evidence, there is a
positive service that Knox's Preface performs. Knox shows that
one of the crucial problems in Collingwood's mature philosophy,
one that may lie deeper that the more apparent problem of the
relationship of philosophy to history, is the issue of the
functions and autonomy of metaphysics and religion; Knox's
"scepticism-dogmatism" argument rests precisely on the relative
priority or independence of reason and faith. We shall see in
the next section how Collingwood's views on religion and
metaphysics set his interpreters at odds with each other. We
have Knox to thank for calling attention to this dimension of
the problem.
Since Knox's Preface, three principal interpreters of
Collingwood's mature philosophy have grappled with its central
paradox. Of these three, one--Alan Donagan--accepts as decisive
Knox's conclusion that Collingwood's thought suffered a drastic
reversal sometime between 1936 and 1938 (LPC, 1). But Donagan
rejects Knox's "brain-damage" reason for this reversal, and
argues rather that philosophically acceptable reasons must be
found for it (which he claims to provide) (LPC, 12-18). Of the
other two, Louis 0. Mink comes close to acknowledging a radical

22
change of mind, insofar as he accepts a greatly reduced version
of the reversal and locates it much earlier--between 1916 and
1924, when Collingwood discovered "dialectic," first applied in
Speculum Mentis (MHD, 20). Lionel Rubinoff, on the other hand,
roundly attacks what he aptly calls "The Radical Conversion
Hypothesis" initially woven by Knox and later embroidered by
Donagan (CRM, 21). Rubinoff dismisses Knox's reference to
Collingwood's illness as being "of no philosophical relevance"
(CRM, 18), and appears to be the only major interpreter to take
the Autobiography seriously.
In the next section we shall take up the interpretations
offered by these and several other notable commentators on
Collingwood's philosophy, and in the final section of this
chapter we shall try to sketch how the interpretation that will
be offered in succeeding chapters differs from theirs.
3. Collingwood's Interpreters: An Overview.
In 1972 there appeared a collection of fourteen essays
(all previously unpublished) entitled Critical Essays on the
Philosophy of R. G. Collingwood, edited by Michael Krausz and
published by Oxford University Press. Aside from its con-tents
(which represent fairly well the current state of the question
concerning his philosophy), the mere appearance of this book
could not help but both please and displease the late Waynflete
Professor of Metaphysical Philosophy.

23
On the one hand it defies his express wishes on the
matter. Towards the end of his life Collingwood's ill health
forced him to recognize the possibility that he might not be
able to finish all the projects that he had set for himself. He
therefore wrote his Autobiography "to put on record some brief
account of the work I have not yet been able to publish, in case
I am not able to publish it in full" (A, 118). In it he wrote
this request not to be the subject of scholarly inquiry:
I am nearly fifty, and cannot in any case hope for more

than a few years in which I can do my best work. I take
this opportunity, therefore, of saying that I will not
be drawn into discussion of what I write . . . . Some
readers may wish to convince me that it is all
nonsense . . . . Some may wish to show me that on this
or that detail I am wrong. Perhaps I am; if they are in
a position to prove it, let them write not about me but
about the subject . . . . And if there are any who think
my work good, let them show their approval of it by
attention to their own. So, perhaps, I may escape
otherwise than by death the last humiliation of an aged
scholar, when his juniors conspire to print a volume of
essays and offer it to him as a sign that they now
consider him senile. (A, 118-19).
The appearance of this volume of essays indicates that it was
only by death that Collingwood escaped that "last humiliation."
But on the other hand the book is scarcely a humiliation
to the memory of the late Collingwood. On the contrary it
illustrates to a surprising degree the extent to which
Collingwood's thought is still very much alive--a liveliness
that could not have but pleased the philosopher-historian who
argued so eloquently for the notion of history as a process

24
of re-thinking past acts of thought still living in the present
(IH, 218). The contributors to this volume realized that they
were defying his wishes, but as students of his philosophy they
found his works, as one of them so succinctly put it, "too
incisive to dismiss and too unclear to adopt" (Mink, in MHD,
vii) and therefore demanding interpretation to an unusual degree.
To a reader familiar with Collingwood primarily as the
author of The Idea of History, and who accepts the ac-count of
Collingwood's development as given by T. M. Knox in the Preface,
this collection of critical essays would come as something of a
surprise. He would be startled at the wide range of topics on
which Collingwood wrote systematic treatises of some brilliance
and originality: besides philosophy of history, the topics
discussed in essays in this volume include esthetics, philosophy
of mind, philosophical method, philosophy of religion,
metaphysics, philosophy of nature, ethics, social and political
philosophy, and even philosophy of education. He would also be
surprised to find no less than half of the essays dealing with
Collingwood's views on metaphysics: as we have just seen, Knox
had found these views dogmatic, sceptical, and in general
scandalously inferior to Collingwood's best efforts in
philosophy and history (IH, xv-xvii). He would also be startled
at the evidence presented by some of the authors for
Collingwood's
25
anticipation of issues the importance of which have only
recently begun to be appreciated: a case in point being Stephen
Toulmin's essay comparing Collingwood to Thomas Kuhn, whose
essay, The Structure of Scientific Revolution, is still being
seriously debated by philosophers and historians of science
(CEPC, 201-21). The central thesis of the latter was discussed
by Collingwood some twenty years earlier in An Essay on
Metaphysics (cf. EM, 48, 74-76).
But it is disconcerting to encounter the evidence, cited
by many of the contributors to this volume, supporting Knox's
argument for the ultimate inconsistency of Collingwood's
philosophy. What is disconcerting about it is that it still
appears to be impossible for the reader to assume any consistent
or even comfortable posture toward this prickly and ill-
assimilated man. As represented by his three principal
interpreters (Donagan, Rubinoff, and Mink) he remains something
of a puzzling figure. Did his attempts to work out a
reconciliation between philosophy and history fail insofar as at
various times he subordinated the one to the other--especially
in his final, allegedly historicist phase? Are Collingwood's
earlier "idealist" reflections on the nature of philosophical
thinking truly "repudiated" by his final analytic philosophy of
mind? Was his revolutionary logic of presupposition, question
and answer really at variance with contemporary logic? Or is
there some
26
comprehensive framework detectable in Collingwood's writings--a
context which he sketched out in his youth and into which all
his later, more detailed writings fit as parts of a systematic
whole?
A case in point is Collingwood's views on religion--the
subject of his earliest publications and by most accounts one of
Collingwood's deepest and most enduring interests. We have seen
that in Knox's view the place of faith and religion is of prime
importance in evaluating Collingwood's alleged dogmatism and
scepticism. In the lead essay of Krausz's collection, however,
Collingwood's views on the religious doctrine of the fall and
redemption of man are singled out by Alan Donagan as "less
blasphemous than laughable" when used to interpret what "any
ordinary Christian believes that Christian redemption is
redemption from" (CEPC, 19). The passage Donagan cites is from
Speculum Mentis, and in it Collingwood is using the fall as a
metaphor symbolizing man's lapse into forbidden knowledge (the
error of abstraction--the separation of subject and object), and
redemption as God's acceptance of this burden of human error as
His own--presumably in the person of Jesus (SM, 302-03). For the
view of the "ordinary Christian" Donagan chooses to compare this
passage to one from John Bunyan's Grace Abounding and The
Pilgrim's Progress, which speaks of the burden of human
suffering that is the lot of every man. Donagan's objection is
that no attempt at a literal paraphrase of religious metaphor
can absorb the truth of passages like the latter

27
without remainder; but this is what Collingwood thought
philosophy could do for religious truth (CEPC, 18).
Now it is the measure of the difference between Donagan
and Lionel Rubinoff that the same passage that Donagan holds up
for ridicule is later cited by Rubinoff as "one of the most
important passages in all of his writings" (CEPC, 101).
Rubinoff's reading of Collingwood is almost a literal rendering
of passages such as this one from The Idea of History:
The task of religious thought and religious practice

(for in religion the theoretical and practical
activities are fused into one) is to find the relation
between these two supposed conceptions of myself as
finite and God as infinite . . . . (I)n religion the
life of reflection is concentrated in its intensest
form, and . . . the special problems of theoretical and
practical life all take their special forms by
segregation out of the body of the religious
consciousness, and retain their vitality only so far as
they preserve their connection with it and with each
other in it (IH, 314-15).
Rubinoff argues that Collingwood's use of the religious metaphor
of the fall and redemption are apt precisely because
Collingwood's entire philosophy is a sustained attempt at
reconciliation of all the divergent tendencies within man--
subject vs. object, thought vs. action, faith vs. reason,
history vs. philosophy, etc.--and the first level on which that
reconciliation takes place is that of religion, with
Christianity as its highest manifestation (CEPC, 106).

28
The two articles by Donagan and Rubinoff therefore take
radically opposing views on the same issue, and their divergent
interpretations extend to the whole of Collingwood's philosophy.
Donagan argues that the collapse of Collingwood's program for an
idealistic metaphysics of the Absolute (in his early philosophy
as expressed in Speculum Mentis and An Essay on Philosophical
Method) left him with a crippled historicist substitute and no
viable philosophy of religion (CEPC, 18). Rubinoff holds that
for Collingwood not only is absolute idealism the only
philosophy adequate to the Christian solution to the twin
problems of alienation and irrationalism that plague the modern
world, but religion itself (at least as Christianity) is a
necessary condition of the possibility of all other forms of
experience--presumably (using the scheme of Speculum Mentis)
art, science, history, and philosophy (CEPC, 86-88).
The reader's suspicion that Donagan and Rubinoff have
their own, divergent meanings for the expression, "absolute
idealism," is partly confirmed by the fact that Donagan
formulates the position in terms of an anti-realist or anti-
abstraction principle (viz. that all abstractions are partial
truths and to that extent erroneous) which is explicit in
Collingwood's early writings, while for Rubinoff the term refers
to the "unified life of the mind," the divisions of which mark
the various subject-object alienations within contemporary
consciousness. If these
29
positions appear as obverse and reverse of the same coin, the
way to distinguish heads from tails would be that Rubinoff
accepts, and Donagan denies, that something describable as "the
absolute standpoint" is possible.
In his book, The Later Philosophy of R. G. Collingwood,
Donagan forcefully (if not always persuasively) argues that the
reason that Collingwood abandoned the idealistic position of his
youth is that he came to realize, from his analysis of mental
functions in The Principles of Art (1938) that all thinking is
conceptual and hence abstract.15 But this position renders
anything like "absolute knowledge" (an absolute identity between
subject and object) impossible, and therefore represents a
"repudiation" of his earlier idealism, in which abstraction (the
cardinal doctrine of realism--on Donagan's reading of
Collingwood) is regarded as the root of all error, and itself a
falsification (LPC, 14, 47-50, 285-89; cf. CEPC, 18). The
philosophy of mind that survives self-destruction by
contradiction forms Collingwood's "later philosophy" which
parallels conclusions of Ludwig Wittgenstein and anticipates
Gilbert Ryle's concept of mind (cf. LPC, 37, 42-43).
____________________
15 The references that Donagan cites at LPC, 14 and 47-
54 to support this assertion--namely, PA, 254 and NL, 7.22, 7.3-
7.31, and 7.38--do not make the claim that "all thinking is
conceptual and hence abstract." See below, pp. 562-76.
30
Donagan reconstructs this philosophy of mind around four
principles: (1) the Principle of Intentionality ("if a man is
conscious he must be conscious of something"); (2) the Principle
of Order ("if a man is conscious of one of his own acts of
consciousness, then it is not by that act itself, but by another
act of consciousness which may be said to be of a higher
order");16 (3) the Law of Primitive Survivals ("when a function
of consciousness (B) is brought into existence having a lower-
order function (A) as its object, unless the lower-order
function (A) continues to exist in its primitive state, the
higher-order function (B) cannot exist at all"); and (4) the Law
of Contingency ("the earlier terms in a series of mental
functions do not determine the later") (LPC, 27-29). In the
resulting hierarchy of levels of consciousness, Donagan argues,
there is no upper limit (LPC, 29,91-92); therefore there is no
such thing as 'absolute knowledge" as an upper limit to
knowledge (cf. LPC, 258).
Lionel Rubinoff, on the other hand, argues persuasively
(if not always forcefully) in his own book, Collingwood and the
___________________
16 None of Donagan's references to Collingwood's texts
for evidence for the "principle of order" support Donagan's
formulation and subsequent employment of this principle: at LPC,
28 and more directly at LPC, 105 and 168, Donagan cites NL,
4.31, 5.91 and 5.92 as evidence that Collingwood's philosophy of
mind "was fundamentally anti-Cartesian; . . . he repudiated
Descartes' doctrine that acts of consciousness are, as it were,
self-illuminating" (LPC, 25). In Donagan's view, Collingwood
came to hold that no act of consciousness can have itself as an
object (LPC, 108, 167-68). For Collingwood's views on self-
consciousness and Cartesianism, see NL, 1.84-1.85, 5.34-5.39;
IH, 141, 291-94, 297, 306; PA, 206, 222-23, 247-52.
31
Reform of Metaphysics, that such a position could never satisfy
Collingwood, whose whole philosophy is an extended argument for
rapprochement--including the rapprochement of subject and object
in absolute knowledge. There was no radical conversion" in
Collingwood's development, Rubinoff argues, because Collingwood
remained true to the idealistic program laid out in Speculum
Mentis (CRM, 23). Taking his clue from the description in
Speculum Mentis of the three ways that the "prize of truth" can
be awarded (to one, to two or more, or to none of the competing
forms of experience), Rubinoff constructs a framework of "three
ontological levels of experience" on which all of Collingwood's
writings can be located.
At the first level, consciousness assumes an absolute

distinction between subject and object, and views the whole
of reality as an expression of whatever experience it is
presently identified with. . . . At the second level, the
distinction between subject and object remains but each
experience now regards itself as only one among a variety
of equally valid standpoints. At the third level the
subject-object distinction has been finally overcome and
some recognition is given to the fact that the forms of
experience, rather than being coordinate species of a
genus, are on the contrary a scale of overlapping forms. On
the basis of these distinctions the implicit rationale of
Collingwood's published works may now be reconstructed.
According to this reconstruction each work may be seen as
exemplifying one or another type of philosophy operating on
one or more of the three levels of experience (CRM, 29-30).
Rubinoff's strategy in answering the charges of Knox and
Donagan, therefore, is to locate the source of one of a pair of
conflicting
32
assertions on one level of this ontological schema, and the
other on a different level. Thus when Collingwood asserts in The
New Leviathan that there is no upper limit to the levels of
consciousness he is speaking at the "second level" in which new
forms of experience are always possible because they are
regarded merely as coordinate species of a genus. But when
Collingwood asserts in Speculum Mentis that absolute knowledge
forms the upper limit of forms of experience, he is speaking
from the "third level" at which subject and object are
identified, the "absolute standpoint" (cf. CRM, 69-73, 369-72).
Where Collingwood's project seems to falter, Rubinoff calls up a
reserve battery of idealistic arguments, from Hegel to Husserl,
and from Bradley to Bosanquet and Blanshard.17 And the summit of
Collingwood's idealistic efforts is a description of mind as
"pure act''l8—which stands in stark contrast to Donagan's static
hierarchy of levels of abstract concepts, related by the logic
of the Principia Mathematica.
___________________
17 The most noteworthy use of idealistic arguments in
Rubinoff's book are the sections dealing with the logic of the
"concrete universal" (interestingly enough in the chapter on
"Philosophy as Absolute Knowledge") and with the theory of mind
as pure act. CRM, 150-83, 315-22.
18 Rubinoff uses in this portion of his argument

Collingwood's translation of G. de Ruggiero's Modern Philosophy.
CRM, 315-22.
33
To the extent that he, too, finds a temporarily
schizophrenic Collingwood unacceptable, Louis O. Mink sides with
Rubinoff--but not by calling up the arguments of Hegelian
idealism. In his perceptive and refreshingly tactful book, Mind,
History, and Dialectic: The Philosophy of R. G. Collingwood,
Mink asserts with Rubinoff, and against Donagan, that
Collingwood is a dialectical philosopher rather than an
analytical one. But he holds, in direct opposition to Rubinoff,
that Collingwood "retained dialectic and abandoned the absolute"
(MHD, 78). And like Donagan, Mink finds Collingwood's religious
philosophy the least interesting of all his thought. Mink goes
so far as to say that Collingwood had no sympathy for the
philosophy of religion, and that Religion and Philosophy was
Collingwood's only non-dialectical book (MHD, 16, 20, 260 n. 7).
But Mink is no subscriber to Donagan's thesis that Collingwood's
later philosophy of mind represents a complete break with his
earlier philosophical program as exemplified in Speculum Mentis
and made explicit in the Essay on Philosophical Method (MHD, ix,
16, 20).
Mink employs his considerable interpretative skills in
bringing out the "recessive" themes in Collingwood's philosophy.
In doing so he tries to display the "figure in the carpet" (a
tripartite dialectic of experience, of concepts, and of mind)
that makes it possible to soften many of the apparent
absurdities and contradictions in Collingwood's phil-

34
osophy (MHD, 80; cf. MHD, 52, 118, 237). In his essay for the
Krausz collection, for example, he takes up Collingwood's famous
description of history as the "rethinking of past acts of
thought" and shows that each term of the expression requires
qualifications in the light of Collingwood's "dialectic of
process." In Mink's reconstruction "history" in this expression
must be taken as a philosophical rather than an empirical
concept, and is thus subject to the characterization of
philosophical concepts that Collingwood gave in his Essay
on_Philosophical Method. As a philosophical concept it has a
dialectical structure, which means that the elements designated
by the term "history" are related in a developing series or a
"scale of forms" (CEPC, 157-72).
Mink describes such a dialectical system as having four
properties: it is (1) connective (the terms in the series are
related generically to a single essence or general description),
(2) cumulative (members of the series are preserved and modified
in successive forms), (3) asymmetrical (no member is the mere
duplicate of another, but rather differs both in degree and in
kind from the others), and (4) non-deterministic (in the series
a prior term is necessary but not a sufficient condition for the
generation of its successor). Mink argues that since
intellectual history (e.g. the history of science, art, or
religion) deals with a subject which involves acts that are
purposively
35
connected, cumulative in effect, non-cyclic and non-
deterministic in their growth, it is clear that it answers to a
dialectical system. All history (in the philosophical sense) is
therefore the re-thinking (i.e. the dialectical analysis) of
past acts of thought (CEPC, 172-76).
But while these efforts by sympathetic admirers of
Collingwood's philosophy may go a long way towards giving the
reader an understanding of the intricacies and articulations of
Collingwood's philosophy, their solutions to its central paradox
would fail to satisfy many of the other contributors to the
Krauts volume--and especially as that paradox is stated in the
Essay on Metaphysics. Thus W. H. Walsh points out that when
Collingwood wrote in that work that "absolute presuppositions"
(the true object of the metaphysician's search rather than the
"pure being" of the ontologists, which Collingwood rejects as an
empty concept) are neither true nor false, he qualified himself
as a "metaphysical neutralist"--Walsh's term for a philosopher
who limits himself to description only, refusing to apply
criteria which would allow one to make a judgment on the truth
or falsity, reality or unreality, etc. of the object described.
As merely descriptive and factually encountered factors
operative in the thought of those who are engaged in any piece
of scientific thinking, such presuppositions may escape the
positivistic condemnation of metaphysical assertions as neither
factually verifiable nor analytically tautologies, but they are

36
also rendered immune to any sort of justification--they can
merely be reported. Metaphysics is thus an historical science,
as Collingwood's central thesis of the Essay on Metaphysics
maintains; but how then account for the refutability, adequacy,
or success of one set of "consupponible" presuppositions to
another (CEPC, 134-53; especially, 142-46, 149)?
Similarly, Stephen Toulmin argues that although
Collingwood was one of the few pioneer thinkers to come to grips
with the central and still unanswered metaphysical question
about conceptual changes in the history of science (or
scientific revolutions, as Kuhn was later to call them), his
"relativism" (roughly the equivalent to Walsh's "neutralism")
prevented him from giving a rational account of why they occur.
Instead Collingwood resorted to a quasi-causal, psychologistic
explanation in terms of unconscious mental "strains" occurring
in a constellation or set of presuppositions (Kuhn's "crisis in
normal science") which are "taken up" or resolved when a new
conceptual framework replaces an old one (in Kuhn's terms, when
a new "paradigm science" appears, completing a conceptual
revolution). Toulmin's dissatisfaction with both Kuhn and
Collingwood adds fuel to Walsh's charge: if two sets of
presuppositions differ, must there not be some mutual
presuppositions with respect to which, or by reference to which,
they are in agreement (CEPC, 201-21, especially 209-13)?

37
Finally, Nathan Rotenstreich adds his eloquent voice to
this dissenting chorus with a deft discussion of Collingwood's
proposed reform of metaphysics vis-a-vis the tradition he
proposed to reform--Aristotle, Spinoza, Kant, and Hegel.
Rotenstreich shows that as often as not when Collingwood used
historical examples of a metaphysician (in Collingwood's
reformed sense) who pointed out the presuppositions of science
in one period or another, he (Collingwood) ignored the
evaluative activities of these same philosophers, who were not
concerned with reportage only, but with justification and
critique as well. What remains of Collingwood's truncated
version of metaphysical history is what Rotenstreich calls a
"cultural anthropology of metaphysics," which "does not
distinguish between the intention and the intentionality of a
metaphysical system which attempts to be categorical" (CEPC,
179-221; esp. 179-80, 197-99).
From this seeming dead end two escapes are possible:
expansion and revision. Both are represented by essays in the
Krausz volume. The first is taken by Errol Harris. Like
Rubinoff, Harris is a sympathetic student of Collingwood and of
the great idealists of this and the last centuries, and he
supplements Collingwood's thought by evoking a frankly
idealistic context for it. But unlike Rubinoff, Harris does not
hesitate to criticize some of Collingwood's positions as
untenable.19 In his essay for the

38
Krausz volume, Harris renews his argument with Gilbert Ryle over
Collingwood's defense of the Ontological argument, first carried
out in the pages of Mind in 1935 and 1936, just after
Collingwood's Essay on Philosophical Method was published.20
Ryle had argued that Collingwood's use of Anselm's argument made
the common idealistic mistake of thinking that concrete matters
of fact (concerning the existence of anything whatever) could be
established by the use of a priori arguments that can only
establish their conclusions hypothetically. Harris' response is
to defend the idealist's use of such arguments by showing the
legitimacy of "categorical universal" judgments, "the concrete
universal" of Bosanquet and Bradley, and the Absolute of Hegel
(CEPC, 113-33).
Michael Krausz, on the other hand, takes the alternative
route. Collingwood had said in the Essay on Metaphysics (1) that
all questions
_________________
19 Cf. Errol E. Harris, "Collingwood on Eternal
Problems," Philosophical Quarterly, I, no. 3 (April, 1951), pp.
228-41. Reprinted in his Nature, Mind and Modern Science
(London, 1954), pp. 3-42.
20 G. Ryle, "Mr. Collingwood and the Ontological

Argument," Mind, XLIV (April, 1935), pp. 137-51; E. E. Harris,
"Mr. Ryle and the Ontological Argument," Mind, XLV (October,
1937), pp. 474-80; G. Ryle, "Back to the Ontological Argument,"
Mind, XLVI (January, 1937), pp. 53-57. These essays are printed
in John H. Hick and Arthur C. McGill, eds., The Many-Faced
Argument (New York, 1967), pp. 246-74.
39
have presuppositions from which they logically arise; (2) that
to be true or false a proposition has to be an answer to a
question; and (3) that absolute pre-suppositions are neither
true nor false because they are not answers to questions, but
stand relative to all questions, in a body of inquiry, as their
presuppositions. Krausz argues that Collingwood's formulation of
the relation of absolute presuppositions may be interpreted to
mean either that it cannot be the answer to a question in any
given systematic inquiry, or in any systematic inquiry whatever.
In the second case it is impossible to explain how what is taken
as an absolute presupposition at one time can become a relative
presupposition (i.e. one that is itself an answer to a question,
and therefore verifiably true or false) at another. Krausz's
strategy is to argue in favor of the first interpretation, which
involves altering Collingwood's theory of meaning to conform
with more contemporarily acceptable accounts, notably that of P.
F. Strawson, which allow for senses of truth and falsity not
specified by Collingwood (CEPC, 222-40).
It is unnecessary to delve any further into Krausz's
collection of critical essays,21 since we have at this point a
fairly synoptic
_______________
21 There are included in this volume illuminating
articles by W. van Leyden, Peter Jones, and Richard Wollheim on
Collingwood's esthetics and philosophy of mind; by Leon J.
Goldstein on the constitution of the historical past (which
deserves special marks for its care in comparing Collingwood's
actual historical praxis, in his writings on Roman Britain,
40
overview of the range of alternatives available to the reader
interested in finding out the extent of Collingwood's
coherence--or incoherence--as a philosopher. Many of these ideas
and issues will arise again in chapters to come, but before
stating our own reasons for rejecting the approaches of the
principal interpreters of Collingwood some sort of summary of
conclusions is in order. Table I sets forth observations which
seem to follow from our brief survey of Collingwood's
interpreters.
We shall see in Chapter II that the list of issues
which we have found to be the central core of concern to
Collingwood's interpreters turns out, interestingly enough, to
be the very set of issues that Collingwood himself presents in
his Autobiography. And yet not one of the contributors to this
volume of essays seriously considered the Autobiography as an
interpretation valid for the whole of his philosophy. This is
all the more the pity, because it seems that Collingwood is
worthy of at least that degree of attention that he lavished on
the unfortunate
_________________
with the theory of historical imagination in The Idea of
History--perhaps the first time an author has approached
Collingwood's work on the philosophy of history on his own
terms); by Sherman M. Stanage on "Collingwood's Phenomenology of
Education: Person and the Self-Recognition of the Mind" (based
on a few scant remarks by Collingwood on the speaker-hearer
situation and the learning of language, in The Principles of
Art); and by A. J. M. Milne on Collingwood's ethics and
political theory (which might serve as an antidote to Walsh's
re-marks in an earlier essay about Collingwood's lack of
appreciation for the social sciences.)
41
TABLE 1
THE INTERPRETATIONS OF COLLINGWOOD'S PHILOSOPHY

_______________________________________________________________
1. The major issues with which Collingwood's interpreters are

concerned are:
a. his attitudes towards, and arguments about, realism and

idealism;
b. his formulation of the logic of questions, answers, and

presuppositions, and in general his position on
philosophical methodology;
c. his remarks about the nature of history and the

relationship of history to philosophy, and especially of
history to metaphysics; and
d. the way or ways in which he worked out a reconciliation

of all the disparate forms of knowledge (religion, art,
science, history, philosophy) within an overall
philosophy of mind.
2. The major alternatives which interpreters have presented for

dealing with the central paradoxes of Collingwood's
philosophy are as follows:
a. since Collingwood's philosophy is not coherent as it

stands, it is necessary to divide his published works into
two or more groups, based on the contradictory premises on
which they are based, and then argue the relative merits
of one group over the other or others;
b. since Collingwood's philosophy is not coherent as it

stands, it is necessary to propose a revision of a portion
of it in order to render the remainder coherent;
c. since Collingwood's philosophy is coherent as it stands,

its apparent inconsistency can be resolved by assimilating
it to a larger and more complete schema--e.g. historical
idealism;
d. since Collingwood's philosophy is coherent as it stands,

its apparent inconsistency can be resolved by showing the
essential core of truth or coherence that unifies its
diverse aspects.
______________________________________________________________
42
Albert Memorial during his military service in World War I. For
in so much of what he writes he seems, like the lines he quotes
from Wordsworth's "Leech-Gatherer" to describe this monument,
Like one whom I had met with in a dream; Or like a man from
some far region sent, To give me human strength, by apt
admonishment (A, 29).
It would be apt, therefore, to admonish oneself as an
interpreter of Collingwood to ask what relationship there is
between what he did and what he had tried to do--the very
question he wished to put to Scott, the architect of the Albert
Memorial--and to start by taking seriously what he said he had
tried to do, before rejecting or revising what he did.
4. On Interpreting Collingwood.
The very issue that is at the center of controversy
concerning Collingwood's philosophy recoils upon the method an
interpreter chooses to employ in dealing with that philosophy.
For (1) if history and philosophy are not identical, then in
dealing with Collingwood's philosophy in an historical manner,
the question of its truth or falsity cannot arise: the historian
would merely point to the "facts" of the matter, record any
lapses in coherence, and let the matter stand. And in dealing
with it in a philosophical manner, all the works of Collingwood
suddenly assume equal standing, and it becomes impossible to see
how later positions develop out of earlier ones, or how one work
has priority over another. One is there-

43
fore stuck again with apparent contradictions in his fundamental
assertions. But (2) if history and philosophy are identical,
then one cannot deal with Collingwood's philosophy in a non-
evaluative manner: understanding what he said at various points
in his career would therefore be only a prelude to measuring its
implicit promise against its explicit performance, and
evaluating the outcome. But this alternative necessitates the
evaluation of some texts as central and others as peripheral,
and therefore choosing the "facts" to which one is to attend.
But if in getting caught up in interpretative
controversies we were to sidestep the issue of truth in
Collingwood's philosophy, we would be showing that we had not
learned the main lessons he wished to teach us about history: to
think historically is not to merely record facts and refrain
from judgment. History is essentially a judgmental affair,
because it involves selective attention to a chosen set of
facts. This is supremely the case when it comes to the history
of philosophy, where the "facts" to which the historian
selectively attends are meanings and meaning-complexes. One must
therefore take a philosopher for what he said (because this
constitutes all the evidence we have for what he was thinking)
but only as a symbol for for what he meant. Conflicts in
evidence (e.g. contradiction in the texts) may make it difficult
or
44
impossible to get beyond the symbol to the meaning, but this
then is the point at which the borders between history and
philosophy become precarious. But since so much of Collingwood's
thought resides here, we must not fear to explore these disputed
territories.
To think philosophically about the past, according to
Collingwood, is also to take it as intentional--that is, as
something already deliberated upon and thought about with us in
mind as the intended heirs to a mental estate. The question one
ought to ask with respect to Collingwood's philosophy should
then be not what chronological series of literary events
occurred in his lifetime, but rather what order or sequence of
thoughts he intended for his readers to follow if they are to
understand his thinking. Now it is in his Autobiography that
Collingwood publicly specified for his philosophical heirs what
sort of program he wished them to inherit, and in it he tried to
make clear how they should proceed to lay claim to this
inheritance. We therefore respectfully decline to accept the
limited inheritance offered to us by Mr. T. M. Knox, and declare
our intentions to carry our suit to a higher court. In doing so
we propose to accept his Autobiography as Collingwood's only
public, legal will.
In arguing our case we shall seriously attempt to live
up to the highest standards of historical scholarship--these
standards being those governing the philosophical interpretation

45
of another philosopher's literary remains. Most of the canons
for such an endeavor are usually given in negative terms. Two of
them were cited by Collingwood himself in the Autobiography: (1)
never accept criticism of any author before satisfying yourself
of its relevance; and (2) reconstruct the problem, or never
think you understand any statement made by a philosopher until
you have decided, with the utmost possible accuracy, what the
question is to which he means it for an answer (A, 74). To these
we add the five interpreter's fallacies stated by Richard
Robinson22 (which we shall number consecutively to the two
Collingwood canons just cited). One should avoid committing any
or all of the following atrocities to a philosophical text: (3)
mosaic interpretation (the habit of laying any amount of weight
on an isolated text or single sentence, without determining
whether it is a passing remark or a settled part of your
author's thinking); (4) misinterpretation by abstraction
(assuming that because an author mentions X and X appears to the
interpreter to be a case of Y. that the author also meant,
asserted, or was aware of Y); (6) the fallacy of insinuating the
________________
22 Richard Robinson, Plato's Earlier Dialectic (Oxford,
1953), pp. 1-5. It is interesting that Robinson seems to
envision the task of an interpreter in terms which sound like a
direct quotation from Collingwood: "The purpose of an
interpreter . . . is to make himself and others rethink the very
thoughts that were thought by someone long ago. Interpretation
is not just any sort of commentary, including the revelation of
the historical causes and consequences of a given thought. It is
the re-creation of that thought" (Ibid., pp. 5-6).
46
future (assuming that the author held doctrines that did not
become explicit until later); and (7) going beyond a thinker's
last word (ascribing to him not merely all the steps he took in
a certain direction but the next step too).23
There may be more rules than this, but these are the
most helpful ones this author has ever encountered, and suffice
for the task at hand. They serve to eliminate every
interpretation of Collingwood that has yet been offered. For (1)
would it not be an error to accept Knox's criticism of
Collingwood's radical conversion to historicism as a lapse into
dogmatism and scepticism due to a cerebrovascular accident,
without satisfying ourselves first that it is relevant? And (2)
would it not be a mistake to reconstruct Collingwood's mature
philosophy of mind, as Donagan does, without understanding what
the question was to which it was meant for an answer? And was
this question not "How can a thinking person understand his own
mind without resorting to the errors of realism?" And (3) is it
not a mosaic interpretation when Rubinoff takes the metaphorical
remark in Speculum Mentis about the three ways the "prize of
truth" may be awarded, and then erects on this frail motif the
"three ontological levels of consciousness" on which are mapped
all of Collingwood's
_______________
23 We also declare ourselves bound by the full set of
grammatical and logical rules necessary for any discourse to
make sense and be coherent. These we omit stating because they
are assumed in any piece of rational inquiry.
47
writings, from books to essays to letters and even translations?
24 And (4) is it not a case of misinterpretation by abstraction
when Mink rules out Collingwood's views on religion and its
philosophy (in the book by that name) on the grounds that they
are not dialectical? For how does Mink know that because
Collingwood argues that philosophy is at least dialectical that
he would accept the further statement that that is all it is?
Does he not say it is also analytical, and is that not what he
is engaged in doing in Religion and Philosophy? And (5) is it
not a case of misinterpretation by inference to assert, as once
again Donagan does, that because in Speculum Mentis Collingwood
connected the doctrine of realism to the mental function of
abstraction, and argued that all abstraction is falsification,
and then later in The Principles of Art argues (if he in fact
did so) that all concepts are abstract, that he therefore
"repudiated" his earlier rejection of realism? And (6) is it not
an insinuation of the future to argue, as Donagan does, that
Collingwood's philosophy of mind is a specimen of linguistic
analysis of the sort carried out in Ryle's Concept of Mind, and
then to assert (on the deception of this analogy with Ryle) that
it is fundamentally anti-Cartesian? But on the other hand (if we
may be permitted
_________________
24 I do not wish to imply that Rubinoff is the only
interpreter to commit this error: Donagan and Mink are equally
guilty of mosaic interpretation--Donagan's four principles of
the philosophy of mind and Mink's tripartite dialectic of
experience, concepts, and mind are also examples of it.
48
to invent a new version of (6), is it not an insinuation of the
past to justify Collingwood's philosophy of mind, as Rubinoff
does, by appeal to the idealism of Hegel? And (7) is it not
going beyond an author's last word to argue, as Rubinoff does,
that because Collingwood took several steps in the direction of
a descriptive phenomenology of consciousness in Speculum Mentis,
that he therefore would take the next step too, and endorse a
"transcendental phenomenology" of mind (CRM, 54, 152-53, 311-15)?
And yet historical scholarship, like history itself, is
a developmental process, as Collingwood says, in which
successive terms sum up and go beyond previous terms without
being necessitated by them. Our own interpretation of
Collingwood is itself subject to this description. Therefore in
what follows we shall see how Collingwood's philosophy, in
senses yet to be specified, displays most of the characteristics
that his interpreters have attributed to it--perhaps even some
that they did not:
(1) As Knox points out, Collingwood's most apparent
problem is his "historicism"--his tendency to identify
philosophy with history; and beyond this is a deeper, less
apparent problem of establishing a workable relationship between
metaphysics (which Collingwood recognized to be at the center of
philosophy) and religious faith. And as Knox (and later such
authors as Walsh, Toulmin, and Rotenstreich) argue, his final
49
position as stated in the Essay on Metaphysics does display
aspects both sceptical and dogmatic. But unlike Knox, we find no
need to posit a radical reversal in his philosophy, and by
viewing the paradoxical statements of the Essay on Metaphysics
in the light of the overall philosophical orientation provided
in his Autobiography, and the development of his thought in his
published writings, this dogmatism and scepticism are
transformed into something more akin to conviction and necessary
self-criticism--philosophical virtues rather than vices. For
there is no need to take the Essay on Metaphysics as a work all
by itself, abstracted from his other writings. When restored to
its rightful context, its central paradox is illuminated and
refined, and the limits of its applicability are re-established;
metaphysics may be more than an historical science, but
Collingwood's point is that it is at least that.
(2) And as Donagan argues, Collingwood's unique
achievement in his later writings is a philosophy of mind that
is carried out by arguments which stress the importance of
expressive, linguistic structures in the life of thought. But
unlike Donagan we find no grounds for arguing that his final
philosophy of mind "repudiates" his earlier anti-realistic
stance, or even that it was "anti-Cartesian" or anti-intuitional
in denying the thesis that self-consciousness is possible at
all. The linguistic basis for his conclusions was present even
in his
50
early writings, and he never varied in his assault on what he
took to be the main tenet of realistic philosophy. For
Collingwood, refuting this central tenet does not mean
maintaining that "all abstraction is falsification," as Donagan
thinks, but rather it involves showing that the proposition
"knowing makes no difference to the object known" is false.
(3) And as Mink argues, the central "figure in the
carpet" of Collingwood's philosophy is recognizable in the
methodology first explicitly spelled out in the Essay on
Philosophical Method--a work that is, as Knox first said it was,
a philosophical classic. We also will agree that the philosophy
of mind that Collingwood worked out in his later writings
exhibits the structure of a scale of forms, a structure first
exhibited in Speculum Mentis, which Mink calls Collingwood's
first dialectical book. However we find no need to argue, as
Mink does, that Collingwood "discovered" dialectic after writing
Religion and Philosophy, nor that Collingwood had no sympathy or
interest in the philosophy of religion. On the contrary we will
find that Collingwood's philosophy is profoundly religious, and
although he discontinued talking about "absolute knowledge" in
his later writings, he modified his view of a philosophical
absolute rather than dropping it altogether (as his discussion
of the ontological argument in both the Essay on Philosophical
Method
51
and the Essay on Metaphysics shows). And although some of the
functions of a philosophical absolute were taken over by
"absolute presuppositions," it is clear that Collingwood took as
one of the absolute presuppositions of contemporary science the
doctrine of Christianity that God exists.
(4) And as Rubinoff argues, Collingwood's philosophy as
a whole cannot be understood apart from his overall orientation
towards a rapprochement of the alienating forces typical of the
contemporary human situation. The paradigm for this rapprochement
was indeed the relation of philosophy and history, which the
position Collingwood called "realism" was committed to ignoring.
And the means that Collingwood used to re-establish continuity
between forms of knowledge was by arguments which are, in some
sense of the term, "idealistic," relying as they do on premises
that are incompatible with the realistic thesis that "knowing
makes no difference to the object known." However there is no
need to call in Hegel's aufheben--or even Bradley's experiential
Absolute--to save Collingwood's rapprochement project. If
because of irreconcilable contradictions on fundamental issues
Collingwood's philosophy cannot be approached on its own terms,
it must be declared to be to that extent inconsistent and in
need of revision, rather than declared to be incomplete and in
need of assimilation to, or absorption by, the philosophy of
Hegel. If we are in debt to Rubinoff for exposing the

52
fallacy of the "radical conversion hypothesis," we also declare
ourselves free of his own fallacy, the "radical consistency
hypothesis." The plain fact is that Collingwood did change his
mind on several issues, and stated publicly that he had done
so--but not on the fundamental issues discussed in the
Autobiography, as we shall see.
In short, the thesis of our interpretation is that it
is possible to make sense of Collingwood on his own terms, if we
are careful to avoid making the errors of our interpretative
predecessors. Perhaps by so doing we shall avoid the wrath of
the shade of Collingwood, who warned us of a haunting should we
fail to take heed to the requirements of philosophical
interpretation:
The reader . . . must approach his philosophical author

precisely as if he were a poet, in the sense that he must
seek in his work the expression of an individual experience,
something which the writer has actually lived through, and
something which the reader must live through in his turn by
entering into the writer's mind with his own. To this basic
and ultimate task of following or understanding his author,
coming to see what he means by sharing his experience, the
task of criticizing his doctrine, or determining how far it
is true and how far false, is altogether secondary. A good
reader, like a good listener, must be quiet in order to be
attentive; able to refrain from obtruding his own thoughts,
the better to apprehend those of the writer; not passive,
but using his activity to follow where he is led, not to
find a path of his own. A writer who does not deserve this
silent, uninterrupting attention does not deserve to be read
at all (EPM, 215).
CHAPTER II
THE AUTOBIOGRAPHICAL INTERPRETATION

OF COLLINGWOOD'S PHILOSOPHY
1. The Autobiography as Literary Evidence.
In an essay written just two years prior to the
publication of his own Autobiography, Collingwood wrote the
following account of the requirements for adequate
autobiographical composition:
If anyone of us were setting out to compose such an account

((viz. an autobiography - "a strictly historical account of
my own past")), he would be confronted with two kinds of
tasks . . . . The first task is that of recollecting: he
must search his memory for a vision of past experiences,
and use various means of stimulating it, for example by
reading letters and books that he once wrote, revisiting
places associated in his mind with certain events, and so
forth. When this is done, he has before his mind a
spectacle of the relevant parts of his own past life; he
sees a young man undergoing such and such experiences, and
knows that this young man was himself. But now begins the
second task. He must not merely know that this young man
was himself, he must try to rediscover that young man's
thoughts. And here recollection is a treacherous
guide . . . because thought is not wholly entangled in the
flow of experience, so that we constantly reinterpret our
past thoughts and assimilate them to those we are thinking
now. There is only one way in which this tendency is to be
checked. If I want to be sure that twenty years ago a
certain thought was really in my mind, I must have evidence
of it. That evidence must be a book or letter or the like
that I then wrote . . . . Only by having some such evidence
before me, and interpreting it fairly and squarely, can I
prove to myself that I did think thus. Having done so, I
rediscover my past self, and re-enact these thoughts as my
53
54
thoughts; judging now better than I could then, it is to be

hoped, their merits and defects. (IH, 295-96, emphasis
mine.)
It would be hard to believe that a trained archeologist
and historian, having just written such a clear account of the
criteria for autobiography, an account which expresses such a
hard-headed view of what counts as evidence for such a literary
project, could have forgotten about them completely when it came
time two years later to write his own autobiography. Yet such is
the charge of more than one of Collingwood's interpreters--
including the man Collingwood named as his literary executor
(cf. IH, x-xi).
Since the charge has been made it must be confronted in
the same spirit of historical objectivity to which Collingwood
himself subscribed. In the second part of his task, Collingwood
wrote, "there is nothing which the autobiographer does . . .
that the historian could not do for another" (IH, 296). If the
autobiographer, in short, performs his task with the same rigor
that is expected of historians, he functions as an historian of
a subject matter which merely happens to be the events of his
own life. If in fact Collingwood was in good faith with this
principle when he wrote his Autobiography (and he at least
claimed that he was--see A, 107 and "Preface"), his own
interpretation of his intellectual development (and this is the
main concern of the Autobiography) stands as one among other

55
such interpretations, each of which must be judged in accordance
with the same criteria: each must be an interpretation of
Collingwood's thought, based on the evidence provided by
Collingwood's published writings, critically evaluated.
One serious objection to taking the Autobiography in
this way is that at the time Collingwood wrote it (1938) two of
his major works (An Essay on Metaphysics and The New Leviathan)
and several articles had not yet been written. Consequently a
good part of his interpretation of what his philosophy actually
achieves remains speculation about what he intended it to
achieve, and not what, on the basis of documentary evidence, it
already had achieved. It is because his work remained incomplete
at the time of the writing of his Autobiography that several of
his interpreters have felt justified in rejecting the latter and
arguing that his later works break entirely with the positions
he had maintained prior to his writing of the Autobiography.
But while it is certainly quite appropriate to raise the
question of the de facto adherence of these later works to the
philosophical doctrines of Collingwood's earlier writings, or of
the adherence of both of these to the interpretation offered in
the Autobiography, it is capriciously arbitrary at best, and
maliciously prejudicial at worst, to exclude the latter as a
possible interpretation valid for the whole of Collingwood's
philosophy, including the later works. That the

56
Autobiography should not be allowed a privileged position among
the interpretations offered of Collingwood's philosophy is a
defensible corollary of the principle cited by Collingwood above
(viz. that correct autobiography is an application of correct
historiography); but that it should be given no consideration is
just as clearly ruled out by the same principle.
A second objection is that in addition to proposing an
interpretation or an account of the development of his thought,
Collingwood's Autobiography also cites evidence to support this
interpretation, and since some of this evidence is not publicly
available, the interpretation based on this evidence is also
open to question.
Now the evidence in the Autobiography is of three sorts:
(1) direct statements by Collingwood of positions he holds, at
least at the time of the writing of the Autobiography; (2)
references to published works that he had completed some years
before, some with and some without qualifying remarks to
indicate the extent to which he still agreed or disagreed with
what he had written; and (3) references to unpublished
manuscripts. Certainly there is no difficulty with taking
statements from the first group as evidence that as of 1938
Collingwood held the positions that he says he does. And just as
certainly, statements from the second can be checked for their
accuracy by consulting the published works

57
to which they refer. Where such reference is factually mistaken,
one can only register the lapse and credit Collingwood only as
holding to the position he is discussing (if he gives textual
indication of it) as his own as of 1938. The third group
requires special comment.
In the Autobiography Collingwood refers to four pieces
of documentary evidence that he gives every indication of having
consulted during his composition of the text, but which are not
accessible to other historians. Two of these documents--an
unpublished book called Truth and Contradiction which he wrote
in 1917, and a book-length essay written in 1920 and jokingly
entitled Libellus de Generatione (as if written by one of the
Italian idealists)--were destroyed by Collingwood after he wrote
the Autobiography (A, 42, 99, and 99 n.l). The other two may
still exist: one is a paper he wrote around 1918 and read to his
colleagues at Oxford, but apparently never published (A, 44);
the other is a 1928 paper which he calls his "Die manuscript"
(because it was written at a country-house at Le Martouret near
Die in France), and which he may have published under a more
descriptive title (A, 107). Of these four documents the first
two were seen by at least one other person apiece--the first by
"a publisher" to whom it was sent (and by whom it was refused)
(A, 42); and the second by Guido de Ruggiero, "for whom I typed
a copy, thinking that it might amuse him as an historian of
philosophy" (A, 99). Whether the latter

58
copy still exists is, to my knowledge, unknown; but in any event
no one has questioned the fact that the original 1917 and 1920
manuscripts did exist until Collingwood destroyed them. The 1918
Oxford paper may be among those unpublished papers still in the
possession of Mrs. Kathleen F. Collingwood. If in fact he did
publish the "Die manuscript" under a different title, it is
still identifiable only in terms of the content he assigns to
it, and is hence not unimpeachable evidence.
Since the remainder of the pieces of literary evidence
that Collingwood cites in the Autobiography refers to books and
articles still publicly available, such evidence is not in
question, and one can still test his interpretation by comparing
it with the relevant texts. The case is not so clear with
respect to the four items mentioned above, and hence they must
be treated as evidence of nothing more than what Collingwood's
views were at the time of his writing of the Autobiography. The
doctrines that Collingwood claims to have espoused in these
documents, insofar as he mentions what these doctrines are, and
insofar as he does not directly re-pudiate the position stated,
must be treated just like the other direct and contentful
statements he makes in the Autobiography--that is, they must be
taken as stating doctrines that he does not assign only to one

59
period or to one document, but puts forward as positions that he
currently held.1
____________________
1 Since the issue is crucial for the interpretation of
Collingwood's philosophy as contained in Autobiography, and
since the issue is made so sensitive by later writers on
Collingwood's philosophy, at the risk of tedium I propose the
following propositional version of what I take to be the limits
of autobiographical interpretation:
Supposing in manuscript M1 at time T1 an interpreter finds

author A1 making these statements:
S1: "I hold P1''
S2: "In M2 at T2 I held P2"
S3: "I still hold P2"
Then an interpreter is justified in making at least the

following assertions:
Il: Ml is evidence that A1 held S1, S2, and S3 at T1.
I2: M1 is evidence that A1 held P1 and P2 at T1.
He is clearly not justified if he were to say:
I3: M1 is evidence that A1 held P2 at T2.
He is not justified because I3 is not constructable on the basis

of S1- S3 of M1. It is valid on the evidence of M2 only if M2 at
T2 contains the assertion, "I hold P2." In short, S2 is itself
an interpretation requiring M2 for its justifying evidence. The
situation is not changed if the interpreter is the author
himself, and the manuscripts are his own writings.
60
2. The Autobiography as Historical Interpretation.
Prior to the Autobiography Collingwood had published
only four philosophical works: Religion and Philosophy (1916),
Speculum Mentis (1924), An Essay on Philosophical Method (1933),
and The Principles of Art (1938). The latter two were intended
as part of a projected series of philosophical works, about
which we will have more to say presently. Concerning the former
two Collingwood has several remarks in the Autobiography.
Religion and Philosophy had been written "some years earlier"
than 1916 "to tidy up and put behind me a number of thoughts
arising out of my juvenile studies in theology" (A, 43). The
main effect of these studies, at least for the development of
his later thought, was his recognition of the falsity of the
claim that empirical psychology had "already exploded the
pretensions and inherited the possessions of the old pseudo-
sciences of logic, ethics, political theory, and so forth," and
was hence the science of human affairs the world was seeking (A,
92).
If this claim never for a moment deceived me, that is a

benefit I owed to my early studies in theology. Like every
one else who studied that subject in those days, I read
William James' Varieties of Religious Experience and a lot
of other books in which religion was treated from a
psychological point of view . . . . I was profoundly
shocked by the Varieties . . . because the whole thing was
a fraud. The book professed to throw light on a certain
subject, and threw no light on it whatever. And that
because of the method used. It was not because the book was
a bad example of psychology, but because it was a good
example of psychology, that it left its subject completely
unilluminated. And in Religion and Philosophy I attacked,
not William
61
James, but any and every psychological treatment of

religion, in a passage of which the crucial words are "the
mind, regarded in this way, ceases to be mind at all." (A,
93).2
According to the above account, then, Collingwood's first
published philosophical work attacked the psychologistic
reductions of religion and mind to neuropathology and physiology
respectively, defending them as on the contrary functions of
consciousness, reason, and will (cf. A, 94-95).
His second philosophical work attempted to establish a
position independently of idealism, but repudiating realism -- a
position which belonged to no recognizable or ready-made class.
I became used to it . . . when ((for example)) one of the

"realists" (not an Oxford man), reviewing the first book in
which I tried to indicate my position, dismissed it in a
few lines as "the usual idealistic nonsense." The book was
Speculum Mentis, published in 1924. It was a bad book in
many ways. The position laid down in it was incompletely
thought out and unskillfully expressed . . . . But any one
who had been intelligent enough to see what I was trying to
say would have realized . . . that it was neither "usual"
nor "idealistic." (A, 56-57).
In a footnote to this passage Collingwood adds that, having just
reread Speculum Mentis for the first time since it was
published, he found it better than he had remembered.
______________________
2 The exact words are: "The mind, regarded in this
external way, really ceases to be a mind at all. To study a
man's consciousness without studying the thing of which he is
conscious is not knowledge of anything, but barren and trifling
abstraction." (RP, 42; FR, 77).
62
It is a record, not so very obscure in expression, of a good

deal of genuine thinking. If much of it now fails to satisfy
me, that is because I have gone on thinking since I wrote
it, and therefore much of it needs to be supplemented and
qualified. There is not a great deal that needs to be
retracted. (A, 56, n. 1).
Surprisingly enough, this is as much as Collingwood has
to say in his Autobiography about his first two published
philosophical books. For the remaining evidence that his ideas
developed as he said they did in the period from 1912 to 1932
Collingwood refers the reader to the short articles he published
in philosophical periodicals, "where they were rendered useless
by the fixed determination of the persons who read such
periodicals not to think about history" (A, 116, n. 1). Two
volumes of these essays have appeared since Collingwood's death
(EPA in 1964 and EPH in 1965). His first two books and these
essays (several of which are included by Knox in the concluding
portions of The Idea of History are the only sources presently
available for critically reconstructing the development of
Collingwood's philosophy during the period prior to the
appearance of the Autobiography.
What sort of conclusion then can one draw from a careful
reading of the Autobiography concerning Collingwood's
development as a philosopher? Every indication in the
Autobiography leads the reader to conclude that if there was any
"development" of his ideas, during which time a possible change
of mind may have occurred, it was in the period

63
from 1912 to 1932 (A, 23, 28, 116-17), and that after 1932 he
was engaged in preparing his conclusions for publication.
The ideas very briefly summarized in this chapter ((viz. ch.

X--"History as the Self-Knowledge of Mind")) and the two
preceding it were being worked out for nearly twenty years
after I became a teacher of philosophy ((in 1912)). They
were repeatedly written down, corrected, and rewritten . . .
. None of these writings has ever been intended for
publication, although much of their substance has been
repeatedly given in lecture form; but I am publishing this
short summary because the main problems are now ((i.e. in
1938)) solved, and publishing them in full is only a
question of time and health. (A, 116-17).
Both of these conditions, however, were to prove problematic:
By about 1930 my health was beginning to suffer from long-

continued overwork . . . . By this time I had in my head a
great deal which I believed the public would value; and the
only way of giving it to the public was by writing books.
On this, therefore, I decided to spend my leisure; and
planned a series, to begin with an Essay on Philosophical
Method. This I wrote during a long illness in 1932. It is
my best book in matter; in style, I may call it my only
book, for it is the only one I ever had the time to finish
as well as I knew how, instead of leaving it in a more or
less rough state. After settling accounts with my
archeological studies . . . I wrote in 1937 the second book
of my series, The Principles of Art. Before it had gone
through the press I was overtaken by the more serious
illness which gave me both the leisure and the motive to
write this autobiography; whose purpose is to put on re-
cord some brief account of the work I have not yet been
able to publish, in case I am not able to publish it in
full. Henceforth I shall spend all my available time in
going on with the series. (A, 117-18).
This passage is crucial for any attempt to reconcile
Collingwood's later philosophy to the interpretation of it
offered in the Autobiography. It will be the purpose of the
later chapters of this dissertation to

64
examine the extent to which this program--his "series"--was
carried out. Right now I merely wish to make two observations.
First, Collingwood leaves no doubt that he considers the Essay
on Philosophical Method as the key to understanding his mature
philosophy. Even the Autobiography itself, the purpose of which
is clearly stated in the text just quoted, is regarded as an
interim report of work in progress. No other work is singled out
for such high marks, and consequently any account of his mature
philosophy which ignores it must do so in defiance of
Collingwood's own clearly stated intentions.
Secondly, Collingwood gives no indication in the
Autobiography of any radical change of mind either before or
after he had begun his "series." He therefore clearly intends
his readers to approach the body of his later writings (i.e.
after 1932) as the fulfillment of a single-minded project, and
the earlier writings as a development leading up to it. An
interpretation of Collingwood's philosophy would be consistent
with the plan of the Autobiography only if it follows this
pattern.
3. The Autobiographical Interpretation: Four Themes.
My life's work hitherto, as seen from my fiftieth year,"
wrote Collingwood approximately five years before his death,
"has been in the main an attempt to bring about a rapprochement
between philosophy and history" (A, 77). The entire problem
concerning Collingwood's mature

65
philosophy is contained in this sentence, and especially in the
meaning of the term, rapprochement. According to Knox and
Donagan the reconciliation which Collingwood outlines in the
Autobiography, and continues in the Essay on Metaphysics,
amounts to nothing short of radical historicism: philosophy as a
separate discipline is liquidated by being absorbed into history
(IH, x). According to Mink and Rubinoff no such reduction occurs
in Collingwood's philosophy, since the rapprochement is
dialectical in nature, and in a dialectical relation the relata
are not separate or mutually exclusive, but rather "overlap" (a
technical term the meaning of which will be examined in Chapter
9).
In the Autobiography neither of these positions is
directly supported. There is no mention of any serious change of
mind, or of any radical reduction of philosophy to history; and
there is no discussion of dialectic or dialectical relations.
After two introductory chapters ("Bent of a Twig" and "Spring
Frost") recounting his early educational experiences at home and
at Rugby (1902-08), there follows two chapters ("Minute
Philosophers" and "Inclination of a Sapling") on Collingwood's
encounters with the Oxford Realists, first as a loyal, but
somewhat sceptical student initiate, and later as a rebelliously
independent tutor. Chapter V ("Question and Answer") encompasses
the years (1915-18) of his work in the Admiralty Intelligence

66
Division during World War I, during which time his daily
communings with the grotesque Albert Memorial led to the
development of his "question and answer logic"3 which was to be
the foundation of all his later philosophical and historical
work. Chapter VI ("The Decay of Realism") records his return to
Oxford in 1918 as a complete opponent of Oxford realism. The
next five chapters ("The History of Philosophy," "The Need for a
Philosophy of History," "The Foundations of the Future,"
"History as the Self-Knowledge of Mind," and "Roman Britain")
deal with the gradual development of his views on historical and
philosophical thinking, listing the principles which became part
of his mature philosophy of history, and giving some indication
of the progress of his work on the archeology of Roman Britain.
A final chapter ("Theory and Practice") records his political
views, especially concerning fascism and socialism, and his
assessment of the rapidly degenerating situation in pre-World
War II Europe.
Even from this brief topical survey one can see that
four themes dominate the Autobiography's interpretation of
Collingwood's philosophy. (1) Out of a total of twelve chapters,
no less than three (III, IV, and VI) deal with Collingwood's
reaction to Oxford realism, and the theme recurs throughout
Collingwood's discussion of his own positive
____________________________
3 Hereafter "Question and Answer" will be referred to by

the abbreviation, Q-A: e.g., "question and answer logic" appears
as "Q-A logic."
67
contributions to philosophy and history. In fact, as we shall
see, the Autobiography exhibits a rather surprising consistency
when seen from the point of view of this rejection of realist
doctrines. (2) Five chapters (VII through XI) are devoted, as
one might expect, to Collingwood's overthrow, based on his
rapprochement between philosophy and history, of the skeptical
conclusions of the Oxford realists. (3) The key to this
reconciliation is discussed in a crucial chapter (V) on his
"revolutionary" Q-A logic. (4) Finally, the theme of
rapprochement is extended, in the final chapter, but also in
remarks scattered throughout the other chapters, to other,
philosophically opposing doctrines: theory and practice, freedom
and obligation, etc. These four themes--the critique of realism,
Q-A logic, philosophy and history, and rapprochement
philosophy--are central concepts in the autobiographical
interpretation.
4. The Critique of Realism.
Collingwood writes that his tutors at Oxford were
members of the "realist" school of philosophers, a school whose
primary function was the destructive criticism of idealism, and
which converged towards the "zero line of complete scepticism"
(A, 18-19).
When I began to read philosophy there in 1910, Oxford was

still obsessed by what I will call the school of ((T. H.))
Green . . . . The philosophical tendencies common to this
school were described by its contemporary opponents as
Hegelianism. This title was
68
repudiated by the school itself, and rightly . . . . This

movement never in any sense dominated philosophical thought
and teaching at Oxford . . . . When I say that Green's
school at this time obsessed Oxford philosophy, what I mean
is that the work of that school presented itself to most
Oxford philosophers as something which had to be destroyed,
and in destroying which they would be discharging their
first duty to their subject. The question what positive
views they themselves held was of secondary importance (A,
15, 16, 19).
At the time of his graduation Collingwood felt that he
was "logically bound to remain a 'realist"' until he had
satisfied himself "either that the positive doctrines of the
school were false, or that its critical methods were unsound"
(A, 23). In connection with the relation between methods and
doctrine there appeared to be three alternatives, between which,
he says, he did not decide until after he had begun to teach
philosophy at Oxford: (a) there was no connection between them
(i.e., both were false); (b) the positive teachings were
mistaken but the critical methods sound; or (c) the positive
doctrines were correct but the critical methods were invalid.
The fourth alternative--viz. that both the positive doctrines
and the critical methods were valid--was apparently ruled out by
Collingwood on the basis of the negativity of the latter: their
positive teachings were incapable of resisting attack by their
own critical methods (A, 23).
Collingwood's description of this method is given in

acid terms:
69
(T)he 'realists'' chief, and in the last resort, it seemed

to me, only method was to analyze the position criticized
into various propositions, and detect contradictions between
these . . . (f)ollowing as they did the rules of
propositional logic . . . . (A, 42).
On any given issue a realist would "fish the problem P out of
the hyperuranian lucky-bag, hold it up, and say 'what did So-
and-so think about this?'" and only after this would they ask,
"Is he right?" (A, 68-9). In short, they separated the
historical question, "What did X think about P?" from the
philosophical question, "Was X right in thinking A about P?" (A,
27, 59). The presupposition of this procedure was that there
were a set of "eternal problems" in philosophy, to which
philosophers gave various answers at various times (A, 60, 69).
The answers given by different philosophers were to a presumably
identical set of questions--where "the sameness was the sameness
of a 'universal', and the difference the difference between two
instances of that universal" (A, 62). Since truth and falsity
were regarded as properties of propositions (A, 34), the
"answers" could be compared to one another to see if they were
contradictory or not (A, 40-42).
As for the positive content, Collingwood writes that
this consisted of a single assertion, dogmatically maintained:
"except for ((the)) one nonsensical phrase ((that)) knowledge
making no difference to what is known, 'realism' had no positive
doctrines of its own at all but had stolen all that it had from
the school of thought which it was

70
primarily concerned to discredit" (A, 44-45). The dogma was
propagated in a number of ways, all of which embody the central
assertion in one way or another. In its metaphysical form it is
the doc-trine that "the known is independent of, and unaffected
by, the knowing of it" (A, 45). In epistemology the "Oxford
'realists' talked as if knowing were a simple 'intuiting' or a
simple 'apprehending' of some 'reality'" (A, 25).
What all these "realists" were saying, I thought, was that

the condition of a knowing mind is not indeed a passive
condition, for it is actively engaged in knowing; but a
"simple" condition, one in which there are no complexities
or diversities, nothing except just the knowing. They
granted that a man who wanted to know something might have
to work, in ways that might be very complicated, in order
to "put himself in a position" from which it could be
"apprehended"; but once the position had been attained
there was nothing for him to do but "apprehend" it, or
perhaps fail to "apprehend" it. (A, 25-26).
As Collingwood paraphrased the way one member of the
movement stated it, knowing is "the simple 'compresence' of two
things, one of which ((is)) a mind" (A, 25). In ethics "the
great principle of realism, that nothing is affected by being
known" becomes the principle that "(m)oral philosophy is only
the theory of moral action: It can't therefore make any
difference to the practice of moral action" (A, 48). In
political theory the realists denied "the conception of 'common
good', the fun-damental idea of all social life," by "insisting
that all 'goods' were private" (A, 49).

71
In Collingwood's estimation the net result of all this
was nothing short of disastrous--for philosophy, for
civilization, and for the realist movement itself. In a gradual
and piecemeal fashion, a process of self-stultification occurred
within the ranks of the Oxford realists.
In this process, by which anything that could be recognized

as a philosophical doctrine was stuck up and shot to pieces
by the "realist" criticism, the "realists" little by little
destroyed everything in the way of positive doctrine that
they had ever possessed. (A, 49).
But although "the fox was tailless, and knew it," he did not
count it a misfortune: the realists
were glad to have eradicated from the philosophical schools

that confusion of philosophy with pulpit oratory which was
involved in the bad old theory that moral philosophy is
taught with a view to making the pupils better men. They
were proud to have excogitated a philosophy so pure from the
sordid taint of utility that they could lay their hands on
their hearts and say it was no use at all; a philosophy so
scientific that no one whose life was not a life of pure
research could appreciate it, and so abstruse that only a
whole-time student, and a very clever man at that, could
understand it. They were quite resigned to the contempt of
fools and amateurs. If anybody differed from them on these
points, it could only be because his intellect was weak or
his motives bad. (A, 51).
Collingwood writes that at the time of the outbreak of
World War I he had not satisfactorily decided which of his three
alternatives concerning Oxford realism was correct (A, 27-28).
As far as he had advanced was to work out the first of his two
rules for sound scholarship which he tried to instill in his
students.
72
I . . . taught my pupils, more by example than by precept,

that they must never accept any criticism of anybody's
philosophy which they might hear or read without satisfying
themselves by first-hand study that this was the philosophy
he actually expounded; that they must always defer any
criticism of their own until they were absolutely sure they
understood the text they were criticizing; and that if the
postponement was sine die it did not greatly matter. (A,
27).
"This did not as yet involve any attack" writes Collingwood,
"upon the realists' critical methods" (A, 27). Using this rule
himself he came to realize that with respect to what the realist
movement primarily was--viz. an attack on "the school of
Green"--they misspent their shot. The position they assaulted
was not Hegelianism, nor was it even idealism in the proper
sense (A, 15-16, 19).
But when Collingwood returned to Oxford after the war he
was already convinced that both the critical methods and the
positive doctrines of the realists were in error (his first
alternative) (A, 42, 44), and his "logic of question and
answer," worked out during his wartime reflections on the A1bert
Memorial, had led to a second pedagogic maxim: "reconstruct the
problem" or "never think you understand any statement made by a
philosopher until you have decided, with the utmost possible
accuracy, what the question is to which he means it for an
answer" (A, 74).
In an (unpublished) paper read at Oxford in 1918,
Collingwood writes, he assailed the cardinal principle of the
realists:
73
I read a paper to my colleagues, trying to convince them

that ((the realists')) central positive doctrine, "knowing
makes no difference to what is known", was meaningless. I
argued that anyone who claimed . . . to be sure of this, was
in effect claiming to know what he was simultaneously
defining as unknown. For if you know that no difference is
made to a thing by the presence or absence of a certain
condition c, you know what θ is like with c, and also what
θ is like without c, and-on comparing the two find no
difference. This involves knowing what θ is like without c;
in the present case, knowing what you defined as the
unknown. (A, 44).
In addition to this "refutation of realism" Collingwood proposed
an alternative theory of knowledge based on the centrality of
the questioning, rather than the merely asserting, activity:
The questioning activity, as I called it, was not an

activity of achieving compresence with, or apprehension of,
something; it was not preliminary to the act of knowing; it
was one-half (the other half being answering the question)
of an act which in its totality was knowing. (A, 26).
We will presently consider the "logic of question and answer" in
more detail, but here it is worth noting that Collingwood
proposed his "revolutionary" logic as an alternative to
propositional logic:
For a logic of propositions I wanted to substitute what I

called a logic of question and answer. It seemed to me that
truth, if that meant the kind of thing which I was
accustomed to pursue in my ordinary work as a philosopher or
historian--truth in the sense in which a philosophical
theory or an historical narrative is called true, which
seemed to me the proper sense of the word--was something
that belonged not to any single proposition, nor even, as
the coherence-theorists maintained, to a complex of
propositions taken together; but to a complex consisting of
questions and answers. (A, 36-37).
74
As this passage makes clear, Collingwood's alternative logic was
intended as an instrument for the discovery of philosophical and
historical truths--a task for which he felt propositional logic
ill-suited.
As a corollary to his Q-A logic, Collingwood denied the
realist's assumption that there are "eternal problems" in
philosophy. This occurred in two phases (A, 68). In the first,
Collingwood discovered through his historical research and in
his teaching experience that there is in fact no set of
permanent, eternal questions in philosophy (A, 60-68):
I found (and it required a good deal of hard detailed work

in the history of thought) that most of the conceptions
around which revolve the controversies of modern philosophy,
conceptions designated by words like "state", "ought",
"matter", "cause", had appeared on the horizon of human
thought at ascertainable times in the past . . . and that
the philosophical controversies of other ages had revolved
around other conceptions, not indeed unrelated to ours, but
not . . . indistinguishable from them. (A, 68).
Secondly he attacked the problem in principle. There can
be no absolute distinction between historical and philosophical
questions both because the distinction presupposes the
permanence of philosophical problems (which was false on
historical grounds), and because in any case of a philosophical
question one and the same passage is used as historical evidence
that it was a problem and as philosophical evidence of what that
problem was (A, 69-70).

75
Having disposed of the central positive doctrine of the
Oxford realists, as well as their epistemology and logic,
Collingwood went on to reject their moral and political
theories. He writes that since 1919 he lectured almost every
year on moral philosophy, and although his reconciliation-
philosophy was still incomplete, the rudiments were present even
then of a solution to the realist separation or distinction
between "facts" and "theories" (A, 148-49).
My first efforts in this direction were attempts to obey

what I felt as my call to resist the moral corruption
propagated by the "realist" dogma that moral philosophy does
no more than study in a purely theoretical spirit a subject
matter which it leaves wholly unaffected by that
investigation. The opposite of this dogma seemed to me not
only a truth, but a truth which, for the sake of his
integrity and efficacy as a moral agent in the widest sense
of that term, ought to be familiar to every human being:
namely, that in his capacity as a moral, political, or
economic agent he lives not in a world of "hard facts" to
which "thoughts" make no difference, but in a world of
"thoughts"; that if you change the moral, political, and
economic "theories" generally accepted by the society in
which he lives, you change the character of his world; and
that if you change his own "theories" you change his
relation to that world; so that in either case you change
the way in which he acts . . . . There were, I held, no
merely moral actions, and no merely political actions, and
no merely economic actions. Every action was moral,
political and economic. (A, 147, 149).
Collingwood regarded this as only a "theoretical" rapprochement,
and the conclusion of the Autobiography describes Collingwood's
bitter and painful discovery that a "practical" rapprochement
was also necessary. But this meant a unification of what he
calls the "three R. G. C.'s"--the "gloves-on" university
professor; the family man of practical affairs; and the

76
"man of action", a "gloves-off" philosopher for whom "the
difference between thinker and man of action disappeared," and
who used to "stand up and cheer, in a sleepy voice," whenever he
began reading Marx (even though, he says, he was "never at all
convinced either by Marx's metaphysics or by his economics") (A,
150-53). The closing lines of the final chapter give the reader
a sense of just what a threat Collingwood regarded the
"realists" to be, not only to philosophy but to civilization as
well. Recalling his remarks about the realists' reduction of
ethics to pure theory, Collingwood links them with the recent
rise in England of what he regarded as a fascist movement:
I am not writing an account of recent political events in

England: I am writing a description of the way in which
those events impinged upon myself and broke up my pose of a
detached professional thinker. I know now that the minute
philosophers of my youth ((viz. the realists)) for all their
profession of a purely scientific detachment from practical
affairs, were the propagandists of a coming Fascism. I know
that Fascism means the end of clear thinking and the triumph
of irrationalism. I know that all my life I have been
engaged unawares in a political struggle, fighting against
these things in the dark. Henceforth I shall fight in the
daylight. (A, 167).
The results of our survey of Collingwood's critique of Oxford
realism can be summarized as follows:

77
TABLE 2
THE CRITIQUE OF REALISM
1. Any doctrine which asserts as its basic principle that the

known is independent of, and unaffected by, the knowing of
it, is realism (A, 44-45).
2. Realism's ultimate method is destructive criticism; that is:
a. the analysis of a position into various propositions; and
b. the use of the rules of propositional logic to detect

contradictions between these propositions (A, 42).
3. The consequences of realism are:
a. the separation of the historical question of fact from the

philosophical question of truth, and the metaphysical
assumption that the latter are eternal (A, 59);
b. an epistemology which defines knowledge as the simple

apprehension of an object (A, 25-26);
c. a metaphysics which deals with a body of eternal truths

concerning the world's general nature (A, 65-67), and
which denies the reality of becoming (A, 99);
d. an ethics which regards itself as merely moral theory, and

hence makes no difference to the practice of moral action
(A, 47-48, 147);
e. a political theory which denies the conception of a

"common good" and insists that all "goods" are private (A,
49).
4. The basic principle of realism cannot withstand destructive

criticism: it involves the meaningless assertion (on
propositional grounds) that one can know what is
simultaneously defined as the unknown. Realism hence cannot
live up to its own claims, and fails as a philosophy (A, 23,
44).
78
5. The method of realism is false as a philosophical method,

because
a. there is not, and cannot be, a one-one correspondence

between indicative sentences in a language and logical
propositions (A, 35); and
b. meaning, contradiction and agreement, truth and falsity do

not belong to propositions by themselves, but to
propositions as answers to questions (A, 33).
6. The consequences of realism, as well as being disastrous for

civilization, are philosophically erroneous:
a. philosophical and historical questions are inseparable,

and there are no eternal questions and concepts (A, 68-69);
b. knowledge is a complex process consisting of questions and

answers, and questioning activity being one half (the
other half being answering the question) of an act which
in its totality is knowing (A, 26);
c. questions concerning the world's general nature are based

on beliefs or presuppositions made by the phys cists of an
era, these presuppositions being subject to change but not
to the distinction between truth and falsity (A, 66);
d. knowledge of the situation in which one is called upon to

act affects the action of the agent in that situation (A,
147-48);
e. actions (which are moral, political and economic at the

same time) based on false knowledge of the situation in
which one is called upon to act, do not serve the good of
the nation as a whole but the good of a class, section, or
only oneself (A, 147-49, 155).
79
5. The Logic of Question and Answer.
Collingwood states that the first steps that he took in
his youthful revolt against the doctrine of realism were
methodological: in place of the "propositional logic" accepted
not only by realists but by idealists as well (A, 52),
Collingwood formulated a "logic of question and answer",
philosophically more appropriate and historically more sound (A,
26, 28, 30). He went so far as to write it up in book-length
form ("during my spare time in 1917"), and offered it, under the
title Truth and Contradiction, to a publisher, but was refused
on the grounds that "the times were hopelessly bad for a book of
that kind" (A, 42). The book was never published, and
Collingwood later destroyed the only draft of it (A, 99, n.l).
The roots of Collingwood's "revolutionary" logic, as
stated in the Autobiography, are complex, and the rules of this
logic shade off imperceptably into his views on history and
metaphysics. With respect to the former, three areas of his
experience seem to have contributed to the formulation of his
views. The first was his field experience in archeology--
initially under the tutelage of his father, then, after 1913,
directing his own excavations (A, 23-24, 30). This experience.
he writes, impressed upon him the importance of the "questioning
activity" in knowledge: in archeological field work "one found
out nothing at all except in answer to a question; and not a
vague question either, but

80
a definite one" (A, 24). In addition to teaching him that the
questioning activity was not preliminary, but rather integral to
the activity of knowing, it also taught him that the
intuitionistic epistemology (knowledge reduced to direct
acquaintance with an object) and propositional logic (truth as a
property of indicative assertions) espoused by the realists were
inadequate (A, 26-27, 30-31).
The second was his pre-war experience as a teacher at
Oxford. This yielded the first of his two pedagogical rules for
philosophical interpretation: "never accept criticism of any
author before satisfying yourself of its relevance"--that is,
one should satisfy oneself by first-hand study that this was the
philosophy the author actually expounded (A, 27, 74). But (as we
have already seen) while this "did not as yet involve any attack
upon the 'realists'' critical methods", when coupled with his
archeological experience it converged as a "flank attack on
'realism' as a philosophy which erred through neglecting
history" (A, 28).
The third source and by far the most important,
according to the Autobiography, was Collingwood's daily
communings with the Albert Memorial:
A year or two after the outbreak of ((the first world))

war, I was living in London and working with a section of
the Admiralty Intelligence Division in the rooms of the
Royal Geographical Society. Every day I walked across
Kensington Gardens and past
81
the Albert Memorial. The Albert Memorial began by degrees to

obsess me . . . . Everything about it was visibly misshapen,
corrupt, crawling, verminous; for a time I could not bear to
look at it, and passed with averted eyes; recovering from
this weakness, I forced myself to look, to face day by day
the question: a thing so obviously, so incontrovertibly, so
indefensibly bad, why had Scott done it? . . . . What
relation was there, I began to ask myself, between what he
had done and what he had tried to do? (A, 29).
His reflections on the unfortunate Albert Memorial led
Collingwood to formulate the second of his two pedagogical
rules: "reconstruct the problem", or "never think you understand
any statement made by a philosopher until you have decided, with
the utmost possible accuracy, what the question is to which he
means it for an answer" (A, 74). This was a direct
generalization arising from the analysis of his aesthetic
experience of the Albert Memorial: Collingwood forced himself to
reconstruct the problem that Scott had set for himself in
designing such an artistic monstrosity:
Had he tried to produce a beautiful thing; a thing, I meant,

which we should have thought beautiful? If so, he had of
course failed. But had he perhaps been trying to produce
something different? If so, he might possibly have
succeeded. If I found the monument merely loathsome, was
that perhaps my fault? Was I looking in it for qualities it
did not possess, and either ignoring or despising those it
did? (A, 29-30).
In addition to affording Collingwood yet another occasion for
examining the role of questioning in knowledge, his reflections
on the Albert Memorial provided him with a clue for solving the
problem about
82
"eternal questions" in philosophy, and especially in metaphysics
(A, 60). Applying his second pedagogical maxim to political
theory he discovered that "the history of political theory is
not the history of different answers given to one and the same
question, but the history of a problem more or less constantly
changing, whose solution was changing with it" (A, 62). Hence
the realist assumption that different philosophies were
different attempts to answer the same question was a "vulgar
error, consequent on a kind of historical myopia which, deceived
by superficial resemblances, failed to detect profound
differences" (A, 60-61). Just as the ideal nature of the state
exhibits essential differences for philosophers living at
different times and in different societies, so the ideals of
personal conduct are subject to essential changes (A, 61-65).
The question "What is the ideal state?" and "What sort of
behavior is moral?" are not the same questions when asked by
different philosophers, because they have different essential
meanings and different presuppositions. The clearest application
of the principle was to metaphysics, where Collingwood finally
laid to rest "the philosophers convictions about the eternity of
problems or conceptions" (A, 65):
It became clear to me that metaphysics (as its very name

might show, though people still use the word as if it had
been "paraphysics") is no futile attempt at knowing what
lies beyond the limits of experience, but is primarily at
any given time an attempt to discover what the people of
that time believe about the world's general nature; such
beliefs being the presuppositions of all their "physics",
that is, their inquiries into its detail. Secondarily,
83
it is the attempt to discover the corresponding

presuppositions of other peoples and other times, and to
follow the historical process by which one set of
presuppositions has turned into another. (A, 65-66).
Hence after discovering the clue during his reflections on the
Albert Memorial and after generalizing the maxim and applying it
to the fields of political theory, ethics, and finally
metaphysics, Collingwood concluded that "there was no recognized
branch of philosophy to which the principle did not apply that
its problems, as well as the solutions proposed for them, had
their own history" (A, 67). Except in the sense used to
designate collectively a series of problems connected by a
process of historical change, such that their continuity, but
not their differences, are discernible--except in this
inaccurate sense "(t)he conception of 'eternal problems'
disappeared entirely" (A, 67-68, n.l).
Based on the generalizations from these three regions of
his experience, Collingwood formulated his revolutionary Q-A
logic, the rules of which (in the autobiographical version) may
be tabulated as follows:
84
TABLE 3
QUESTION AND ANSWER LOGIC - AUTOBIOGRAPHY
1. A body of knowledge consists not of propositions but of these

together with the questions they are meant to answer (A, 30).
a. "Proposition" denotes an assertive act of thought or

what in those acts of thought is asserted (A, 30).
b. A proposition is always a logical and not merely a

linguistic entity ((i.e. it states what ought to be the
case rather than what merely is the case concerning
assertive acts of thought)) (A, 31).
2. In order to find out what a proposition means the question to

which the proposition was meant as an answer must be known
(A, 31).
3. No two propositions can agree with or be contradictory to one

another unless they are answers to the same question (A, 33).
a. The sameness of two questions is the sameness of an

historical process, and the difference between two
questions is the difference between one thing ((the
first question)) which in the course of that process has
turned into something else ((the second question)) (A,
62).
b. An historical process is a process of becoming such that

if a process P1 turns into a process P2, there remains
in P2 a trace or survival of P1 (A, 98-99).
4. Truth and falsity belong to a complex consisting of questions

and answers such that:
a. each answer and its question must be relevant to a

complex of questions and answers ((i.e., to a systematic
inquiry));
85
b. each question within that complex must "arise"

((Collingwood leaves the meaning of "arise" unresolved
in the Autobiography));
c. each answer must be the "right" answer to its question

(where "right" means "enabling the inquiry to proceed",
and not "true"--the right answer could be false);
d. each answer is to a certain specific question in the

questions and answer complex (A, 37). (cf. A, 31-32--
correlativity of Q & A).
5. Questions not only have answers, they also have

presuppositions which are not subject to the distinction
between what is true and what is false (A, 66).
a. Some presuppositions ((relative presuppositions)) may be

the answer to another question (A, 66).
b. Some presuppositions are "absolute"--that is, are not

answers to any questions at all (A, 67).
6. The question "To what question did So-and-so intend this

proposition for an answer?" is an historical question, and
cannot be settled except by historical methods (A, 38-39).
a. The settlement of an historical question results from

arguing back from the propositional answer to its
question (A, 70).
b. In arguing back from a propositional answer to its

question, one and the same piece of evidence states the
answer and allows the historian to identify the question
(A, 70).
86
Collingwood's informal presentation of this Q-A logic is
interspersed with commentary on it, some of which is worth
mentioning here as a sort of concluding appendix to this section.
It is worth noting, to begin with, that Collingwood's Q-
A logic is formulated in the context of a theory of knowledge,
and a theory of knowledge that looks to history as a paradigm of
knowing rather than to mathematics (A, 36-37). This is the
significance of the first statement in the above table. The
"body of knowledge" or systematic inquiry which he has in mind
is an inquiry in which discoveries are still being made (at
least for the inquirer), and not a closed system with fixed
relationships (cf. A, 75).
Secondly, Q-A logic is at once a theory of meaning, of
logical validity (agreement and contradiction), and of truth:
If the meaning of a proposition is relative to the question

it answers, its truth must be relative to the same thing.
Meaning, agreement and contradiction, truth and falsehood,
none of these belonged to propositions in their own right,
propositions by themselves; they belonged only to
propositions as the answers to questions: each proposition
answering a question strictly correlative to itself. (A, 33;
cf. A, 37).
Another way of putting the matter would be to say that meaning,
validity, and truth are functions of a Q-A complex which, for
Collingwood, is taken as the primary logical unit. Items 2, 3,
and 4 of the above table take up each of these successive
functions.
87
Thirdly, it is over these three functions of the Q-A
complex that Collingwood felt he departed from the doctrines of
propositional logic--the mathematical logic accepted in part by
the idealists, and in total by the realists (A, 33-36, 42). As
in his refutation of the cardinal principle of the realists,
Collingwood was careful to formulate "the central doctrine of
propositional logic" which he was concerned to reject:
that there is, or ought to be, or in a well-constructed and

well-used language would be, a one-one correspondence
between propositions and indicative sentences, every
indicative sentence expressing a proposition, and a
proposition being defined as the unit of thought, or that
which is true or false. (A, 35-36).
This central doctrine would clearly be ruled out of Q-A logic on
the grounds that both meaning and truth are functions of a Q-A
complex, and not of answers or of assertive acts of thought
alone. Yet it is presupposed by all the various well known
theories of truth":
One school of thought holds that a proposition is either

true or false simply in itself, trueness or falseness being
qualities of propositions. Another school holds that to call
it true or false is to assert a relation of "correspondence"
or "non-correspondence" between it and something not a
proposition, some "state of things" or "fact". A third holds
that to call it true or false is to assert a relation
between it and other propositions with which it "coheres" or
fails to "cohere". And, since in those days there were
pragmatists, a fourth school should be mentioned,
holding . . . that to call a proposition true or false is to
assert the utility or inutility of believing it. All these
theories of truth I denied. (A, 36).
88
It is no wonder that, as Collingwood remarks in a later chapter,
"(s)o far as my philosophical ideas were concerned, I was now
cut off not only from the 'realist' school . . . but from every
other school of thought in England, I might almost say in the
world" (A, 53).
Fourthly, it is in connection with his application of Q-
A logic to the supposedly "eternal problems" of metaphysics that
the relationship between Q-A complexes and presuppositions makes
its appearance (A, 66-67). So also the example which Collingwood
uses to illustrate that contradiction or agreement (what we have
called "validity") is a function of the Q-A complex, is the
classical metaphysical problem of "the one and the many":
For example, metaphysicians have been heard to say "the

world is both one and many"; and critics have not been
wanting who were stupid enough to accuse them of
contradicting themselves, on the abstractly logical grounds
that "the world is one" and "the world is many" are mutually
contradictory propositions . . . . There is no contradiction
between saying that something . . . is one, and saying that
it is many. Contradiction would set in only if that
something were said to be both one x and many x's . . . .
Thus, if a given doctrine D is criticized as self-
contradictory because it is divisible into two parts E and
F. where E contradicts F. the criticism is valid only if the
critic has correctly reconstructed the questions to which E
and F were given as answers (A, 40-41).
Fifthly, and finally, the way Collingwood formulates his
Q-A logic presupposes an understanding of "history" as he uses
the term. This is clear from items 3a and 3b of the above table,
which specify
89
"sameness of question" in terms of historical process, and item
6, which makes historical methodology integral to the discovery
of "sameness of meaning" in a Q-A complex.
6. History and Philosophy.
By the time of the outbreak of World War I, but before
his reflections on the Albert Memorial, Collingwood had not
successfully resolved his "threefold question" concerning the
critical methods and positive content of Oxford realism, but his
archeological research and early philosophical teaching
experience aided him in mounting what he calls a "flank attack"
on the same problem:
Working simultaneously along these two lines, I could see

them tending to converge in an attack on "realism" as a
philosophy which erred through neglecting history. If I had
thought it possible to forewarn the "realists" of this
attack, I should have said, "You must pay more attention to
history. Your positive doctrines about knowledge are
incompatible with what happens, according to my own
experience, in historical research; and your critical
methods are misused on doctrines which in historical fact
were never held by those to whom you ascribe them." (A, 28,
emphasis mine).
By 1920, Collingwood writes, he had completely worked
out the idea of a "living past", and was prepared for a frontal
assault on the realists' view of the past as consisting of
corpse-like "events". The realists' neglect of history was a
result of their refusal to admit the reality of becoming. The
overcoming of the error involved in the

90
recognition that historical processes survive in the present,
and this is the key concept in deciding how history can be
reconciled with the "wisdom" sought by philosophers. The past
that the historian studies is part of the situation within which
he is called upon to act (A, 114); and the "events" of the past
are processes of becoming which survive as active features of
the present (A, 98-100).
By 1930, according to the Autobiography, Collingwood
had worked out the principles on the basis of which history as
"a science of human affairs" could be constructed (A, 115).
These principles Collingwood connects with his Q-A logic and
with the maxims drawn from his experience as an archeologist (A,
106-109, 122, 130). Again, for brevity's sake, we shall tabulate
these principles here in slightly altered order from their
appearance in the Autobiography, but with a notation that
facilitates reconstruction of that order:

91
TABLE 4
THE PRINCIPLES OF HISTORY
1. (LG) History is concerned not with events but with processes

(A, 99).
a. Processes are things which do not begin or end ((as

events do)) but turn into one another; if process Pi
turns into P2, Pi goes on in the changed form P2, and
P2 has previously been going on in earlier ((implicit))
form, P. (A, 98).
b. ((In an historical process)) Pi leaves traces (evidence)

of itself in P2, so that an historian living in P2 ((a
situation)) can discover that what is now P2 was once PI
by the interpretation of evidence (A, 98; cf. A, 96).
_________________
NOTE: In Table 4, the following abbreviations are used:
LG = Libellus de Generatione (see page 94, below), mentioned in
Chapter IX of the Autobiography; ARCH-1,-2,-3 = Archeological
principles, mentioned in Chapter XI; HIST -1, -2,-3,-4 =
Historical principles mentioned in Chapter X.
Collingwood clearly indicates that the principles employed in

archeology are applicable to all of history (A, 121, 130, 133).
He also states that the idea of a living past, expressed in
terms of historical processes, became his "first principle of a
philosophy of history"; it is therefore listed as such in the
table. "ARCH-2" follows it, because it amplifies "LG", and is a
natural bridge to "HIST-1". "HIST-1,-2,-3 are listed in that
order in Chapter X of the Autobiography. "HIST-4" is not so
numbered by Collingwood, but he indicates that it forms the
conclusion of a train of thought that "was not complete until
about 1930" (A, 115). Finally, it is worth noting that
Collingwood states in a footnote that the principles we are
calling "HIST-1,-4" were discussed in a paper delivered before
the British Academy after his election to that body in 1934.
They appear almost verbatim in The Idea of History in the
"epilegomenon" called, after the paper, "Human Nature and
History" (IH, 205-231; see especially IN, 215).
92
2. (ARCH-2) There are no mere "events" in history; what is

miscalled an event is really an action, and expresses some
thought (intention, purpose) of its agent; the historian's
purpose is to identify this thought (A, 127-28; cf. A, 130).
3. (HIST-1) All history is the history of thought (A, 110).
a. The thought must be expressed: either in language or in

one of the many other forms of expressive activity (A,
111).
b. The historian must be able to think over again for himself

the very same thought whose expression he is trying to
interpret (A, 111).
4. (HIST-2) Historical knowledge is the re-enactment in the

historian's mind of the thought whose history he is studying
(A, 112).
a. The sameness of the thought is not the sameness of a

universal but the sameness of an historical process (A,
62).
b. The difference between the thought of the agent and the

reenacted thought of the historian is a difference of
context: to the historian it is a past thought living in
the present ("encapsulated", not "free"), while to the
agent it is a present thought (A, 113).
5. (HIST-3) Historical knowledge is the re-enactment of a past

thought encapsulated in a context of present thoughts which,
by contradicting it, confine it to a plane different from
theirs (A, 114).
a. An Encapsulated thought is a thought which, though

perfectly alive ((not a mere "event", which ends and
begins)), forms no part of the Q-A complex which
constitutes the "real" life, the superficial or obvious
present, of the mind in question (A, 113; cf. A, 140-41).
93
b. Present and past planes of thought are distinguished by

observing the way in which ((their respective)) problems
arise (A, 114).
c. Every historical problem ultimately arises out of "real"

life--i.e., out of practical problems (A, 114).
6. (HIST-4) The science of human affairs (i.e. moral and

political wisdom) is history (A, 99, 115).
a. Knowledge achieved by historical inquiry is not knowledge

of his (the historian's) situation as opposed to knowledge
of himself, it is knowledge of his situation which is at
the same time knowledge of himself (A, 114).
b. There must be a kind of action which is not deter-mined

according to rule, and where the process is directly from
knowledge of the situation to an action appropriate to
that situation without passing through the stage of
formulating a rule appropriate to the situation (A, 103).
b. History offers insight into the situation in which one is

called upon to act, rather than ready-made rules for
acting in all situations of a given kind (A, 100-102; cf.
A, 114).
7. (ARCH-1) Success in historical studies depends upon clear

application of Q-A logic to historical problems (A, 121-122,
124).
8. (ARCH-3) No historical problem should be studied without

studying its second-order history, that is, the history of
historical thought about it (A, 132).
94
Again, several comments about these principles are in
order. The first is that what is stated by Collingwood to be the
"first principle" of a philosophy of history, as of 1920, is
expressed by him in two ways, which are not transparently
identical in meaning. The first is "that the past which an
historian studies is not a dead past, but a past which in some
sense is still living in the present"; the second is in the form
in which it appears above. The idea of a living past is
described later in terms of the principle of Encapsulation (5.
of Table 4), so that its later reiteration eliminates the need
to argue for an identity of meaning in these two expressions, as
well as the necessity to speculate why Collingwood took them to
mean the same thing.
Secondly, Collingwood writes that this first principle
of history initially appeared in an essay of short book-length
(Libellus de Generatione) which "was primarily a study of the
nature and implications of process or becoming."
Secondarily, it was an attack on "realism", showing how the

non possumus of "realists" towards a theory of history
arose from their refusal to admit the reality of becoming,
and from their analysis of the true proposition ''P1
becomes P2" into the complex of propositions ''P1 is P1",
"P1 is not P2" "P1 ends where P2 begins", "P2 is P2", and
"P2 is not Pi'', all of them either tautologous or false.
(A, 11).
Hence Collingwood's first principle of history was formulated,
according to the Autobiography, in direct opposition to what he
took to be a realist's position.

95
Thirdly, Q-A logic appears as an integral part of his
listing of historical principles: it appears in his principles
of re-enactment (4), of Encapsulation (5), and of successful
historiography (7).
Fourthly, Collingwood states that in connection with the
principle of re-enactment (4), the question of what the
difference is between the thought of an historical agent and the
re-enacted thought in the historian's mind, was the most
difficult of all the questions he encountered in his study of
historical method (A, 112). The answer that he gives to this
question in the Autobiography and the example which accompanies
it are given in terms of his Q-A logic. According to the
principle of Encapsulation (see 5 in Table 4), present thoughts
and past, Encapsulated thought are distinguished by the way in
which questions arise in each. For Admiral Nelson at the naval
battle at Trafalgar the question, "Shall I take off my
decorations?" and its answer, "In honour I won them, in honour I
will die with them," occur in a primary Q-A series that involve
the battle and his participation in it on the decks of the
Victory. But this question does not arise in a primary Q-A
series involving the Encapsulated thought (the historian does
not contemplate removing his own decorations in fear of losing
his life).
But a question arising in ((the historian's)) primary

series may act as a switch into another dimension. I plunge
beneath the surface of my mind, and there live a life in
which I not merely
96
think about Nelson but am Nelson, and thus in thinking

about Nelson think about myself. But this secondary life is
prevented from over-flowing into my primary life by being
what I call incapsulated, that is, existing in a context of
primary or surface knowledge which keeps it in its place
and prevents it from thus overflowing. Such knowledge, I
mean, as that Trafalgar happened ninety years ago: I am a
little boy in a jersey: this is my father's study carpet,
not the Atlantic, and that the study fender, not the coast
of Spain. (A, 113-14).
That primary or surface knowledge, some examples of which
Collingwood uses from his juvenile re-enactments of the naval
engagement at Trafalgar, serve to "contradict" the imaginative
experience in which one takes on the role of being Nelson (and
in the process forgets oneself or loses oneself in the imagined
object). It is this "contradiction" which confines Encapsulated
thought to a plane different from the context of present
thoughts.
Finally, history as the science of human affairs
(6)corrects the false claim that psychology, a natural science
of mind, is the source of wisdom, especially in matters moral
and political (A, 92, 94, 116, 126).
The nineteenth century, likewise in search of a science of

human affairs, tried to realize it in the shape of a
"psychology" in which the mental was reduced to the
psychical, the distinction between truth and falsehood
thrown overboard, and the very idea of a science negated,
psychology itself being involved in the resulting
bankruptcy. But the revolution in historical method . . .
swept away these sham sciences and . . . brought into
existence a genuine, actual, visibly and rapidly progressing
form of knowledge which now for the first time was putting
man in a position to obey the oracular precept "know
thyself" . . . . (A, 116).
97
But if a "science of human nature" cannot be achieved except by
historical methods, part of the reason is that actions--
historical processes--are not always performed in accordance
with rules. Collingwood points out two occasions in which agents
necessarily act without knowledge of any rule appropriate to the
situation: (1) when a situation requires one to act and yet does
not recognizably belong to any rule-governed types; and, (2)
when the situation is recognized as of a rule-governed type, but
the required act "involves a certain misfit between yourself and
your situation"--presumably because the agent requires more of
himself than action according to type or to rule (A, 103-104).
Of these two cases in which it is necessary to act other-

wise than according to rule, the first arises out of the
agent's inexperience and ignorance of life . . . . The
second arises only for people of experience and
intelligence, and even then occurs only when they take a
situation very seriously; so seriously as to reject not only
the claims of . . . desire, and . . . self-interest, but
((also)) . . . right conduct, or action according to the
recognized rules (A, 105).
Such rule-free occasions call for improvised actions appropriate
to the recognized realities of the situation; and the function
of historical thinking was to provide "insight" into such
situations, the reality of which included the incapsulated past
as part of itself (A, 106, 101)

98
7. Rapprochement Philosophy.
We have already seen that Collingwood regarded his
philosophy as "in the main an attempt to bring about a
rapprochement between philosophy and history"; that he rejected
realism as a "philosophy which erred through neglecting
history"; that philosophy and history shared a common
methodology by employing Q-A logic; and that historical
questions are not separate from philosophical questions because
"all history is the history of thought". The final step in the
autobiographical account is therefore to elucidate the meaning
of Collingwood's "rapprochement" philosophy.
One aspect of this rapprochement occurred to Collingwood
in the course of his philosophical teaching at Oxford: if
philosophers were to deal with the history of their own subject
they ought to do so in a manner that met the contemporary
standards of historical thinking (A, 77)--some of which
Collingwood states in the Autobiography, and which we have
summarized in Table 4.
But, in addition to making philosophy more historically
respectable, it was necessary to make history more
philosophical: this Collingwood did in his own historical work.
As an example of it he cites the chapter on "Art" which he wrote
for the first volume of the Oxford History of England--a chapter
in which he showed how a revival

99
of Celtic art forms was possible after two centuries, during
which time only Romanized art was produced. In solving the
problem Collingwood made use of a modified form of his principle
of incapsulation:
Incapsulation is not an "occult entity". It was my name for

such facts as this--familiar enough to everybody--that a
man who changes his habits, thoughts,etc., retains in the
second phase some residue of the first. (A, 141).
The principle also operates in habits transmitted from one
generation to the next, and without the need of positing any
occult entities "like racial temperament or an inheritance of
acquired psychical characteristics" (A, 142):
(T)he transmission by educational means of any moral ideal

which involves the outlawry of an institution or custom, and
the repression of a desire for it, entails the simultaneous
transmission of that desire itself. The children of each
generation are taught to want what they are taught they must
not have. (A, 143).
This was the means by which a suppressed art-form was preserved
over two centuries of time in Romanized Britain:
(T)he less successful the Britons were in Romanizing

art, . . . the more they were likely to cherish the memory
of their own fashions and ensure that these fashions were
never wholly lost to sight by the rising generation. (A,
144).
Collingwood calls on this example as an illustration of the use
of his rapprochement philosophy:
I found it possible to assert a connection between two

facts, both of them notorious, which had not previously been
thought of as
100
connected. One was the Celtic revival; the other was the
badness of Romanizing British art . . . . (T)he idea which
I expressed in the chapter on "Art" in the Oxford History
of England . . . I would gladly leave as the sole memorial
of my Romano-British studies, and the best example I can
give to posterity of how to solve a much-debated problem in
history, not by discovering fresh evidence, but by
reconsidering questions of principle. It may thus serve to
illustrate what I have called the rapprochement between
philosophy and history, as seen from the point of view of
history (A, 144-145).
In addition to a reconciliation from the historical
direction, there is presumably a reconciliation from the point
of view of philosophy:
This meant, in the first instance, a special branch of

philosophical inquiry devoted to the special problems raised
by historical thinking. Epistemological problems, such as
one might group together under the question "how is
historical knowledge possible?" Metaphysical problems,
concerned with the nature of the historian's subject-matter:
the elucidation of terms like event, process, progress,
civilization, and so forth. (A, 77).
We have already observed how Collingwood resolved the first
question (viz. that one can know the past if it is not a "dead"
past, but is rather "living", incapsulated in the present, and
known by critical evaluation of evidence). In the second group,
Collingwood deals only with the terms "event" and "process"--the
subject matter of the first two principles in Table 3; the
remainder are left unelucidated in the Autobiography.
"But this demand for a new branch of philosophy," adds
Collingwood, "soon developed into the demand for a new kind of

101
philosophy" (A, 77), a "Copernican revolution" (A, 79, n.l) in
which historical knowledge is shown to be an element in all
thinking (A, 86-88, 67-68). Collingwood cites the example of a
scientist who, in framing a theory, makes use of historical
experiments as to what experiments had been tried and with what
results, and insists that this historical knowledge is an
essential element in all scientific thinking (A, 87).
Metaphysics itself is an historical science, insofar as "the
question what presuppositions underlie the 'physics' or natural
science of a certain people at a certain time" is a purely
historical question, and the origins of these beliefs about the
world's general nature have come into existence by certain
changes out of other such beliefs (A, 66-67).
Finally, a similar analysis occurred in connection with
moral philosophy, which also involved an historical element:
If knowledge as to the facts of one's situation is called

historical knowledge, historical knowledge is necessary to
action . . . . Immediately after the War, therefore, I began
to reconsider in detail all the familiar topics and problems
of moral philosophy . . . . In the first place, I subjected
these topics and problems to what I called an historical
treatment, insisting that every one of them had its history
and was unintelligible without some knowledge of that
history. Secondly, I attempted to treat them in another way,
which I called analytic. My notion was that one and the same
action, which as action pure and simple was a "moral"
action, was also a "political" action relative to a rule,
and at the same time an "economic" action as means to an
end. (A, 149).
102
The reconciliation between theory and practice, like that
between history and philosophy, proceeded from two sides. From
the theoretical side thought and action were shown to be
mutually dependent, "thought depending upon what the thinker
learned by experience in action, action depending upon how he
thought of himself and the world" (A, 150). We have already seen
how the need for a practical rapprochement impressed itself on
the "three R. G. C.'s".
Unfortunately this is as much as Collingwood has to say
on the subject of rapprochement philosophy in the Autobiography,
and the reader is directed to complete the task of working out
the details for himself (A, 149). Unlike the discussion of Q-A
logic and the principles of history, there are no explicit rules
for rapprochement. However, from the above discussion several
generalizations about rapprochement philosophy are possible:

103
TABLE 5
RAPPROCHEMENT PHILOSOPHY
1. The subject matter of rapprochement philosophy are viewpoints

presumed to be distinct and opposite (e.g. philosophy and
history, theory and practice, etc.). Such an accepted state
of unreconciled opposition between viewpoints is
characteristic of the "realist" philosophy (A, 148)
2. A reconciliation of opposing viewpoints proceeds from each

viewpoint towards its presumed opposite (e.g. from philosophy
toward history, and from history toward philosophy) (A, 77,
144-45).
3. What must be shown for minimal reconciliation is that there

is a relation of mutual dependence of each view-point on its
presumed opposite (A, 150).
4. A relationship of mutual dependence between opposing

viewpoints is established by subjecting them to both:
a. an historical treatment, in which they are shown to

satisfy the conditions (or principles) of historical
inquiry (to be related as two phases of a process, capable
of re-enactment, etc.); and
b. an analytic treatment, in which they are shown to satisfy

the conditions (or principles) of Q-A logic (they must be
answers to the same question, be part of a Q-A complex,
etc.). (A, 148; cf. A, 31, 42).
104
It is questionable, given Collingwood's remarks about
rule-free situations and improvisation, whether he would agree
that any such semi-formalization of the methodology of
reconciliation is possible. Improvisation apparently extends
even to the level of methodology:
Obscure provinces, like Roman Britain, always rather appeal

to me. Their obscurity is a challenge; you have to invent
new methods for studying them, and then you will probably
find that the cause of their obscurity is some defect in the
methods hitherto used. When these defects have been removed,
it will be possible to revise the generally accepted
opinions about other, more familiar, subjects, and to
correct the errors with which those opinions are perhaps
infected. In this sense, knowledge advances by proceeding to
"from the known to the unknown", but from the "unknown" to
the "known". (A, 86).
Collingwood extends this remark to include "obscure subjects"
including (at that time in England) historical methodology, the
systematic study of which he hoped would reveal epistemological
truths "concealed from the 'realists' by their obviously
conventional and second-hand ideas about the methods of natural
science" (A, 86).
It is also possible that Collingwood's failure in the
Autobiography to specify the sorts of conditions under which
questions are said to "arise" is related to his reversal of the
Aristotelian maxim about the direction in which the mind works
in coming to know a subject (for Aristotle it proceeded from the
known to the unknown). If obscurity is a stimulus for the
invention of new methods which, in turn, reveal errors

105
in previous treatments of the subject, then it is hardly
possible to predict in advance the questions that will arise: it
presumes prior knowledge of the errors that have been made.
But on these issues Collingwood is enigmatically silent.

PART II
THE EARLY WRITINGS (1916-1932)

CHAPTER III
REALISM AND IDEALISM
1. Introduction.
In the Autobiography Collingwood described his weaning
from the brand of realism that he had been taught at Oxford, the
turning point occurring during his reflections on the Albert
Memorial sometime around 1917. At one end of this development is
the teaching of Cook Wilson and the other Oxford realists, and
at the other is the publication of Speculum Mentis in 1924,
which the Autobiography acknowledges was perceived by at least
one reviewer as the "usual idealistic nonsense." In between
these two points there is the publication of his first book,
Religion and Philosophy, in 1916, but actually "written some
years earlier, in order to tidy up and put behind me a number of
thoughts arising out of my juvenile studies in theology (A, 43).
Prior to the appearance of Speculum Mentis Collingwood published
a lecture entitled "Can the New Idealism Dispense with
Mysticism?" and several other articles dealing with the
distinction between science and history and Croce's philosophy.
From these points we should be able to sketch in the curvature
of his thought and compare it to that described in the
Autobiography.
106
107
Along the way we should be able to reconstruct what
Collingwood's understanding was of the realism-idealism issue,
before it hardened into the form it takes in his later writ-
ings--viz. the anti-realism of the Autobiography. This is a
crucial topic, since the starting point for virtually every one
of Collingwood's publications is a critique of the realistic
position of the subject under investigation. Furthermore his
interpreters, as we saw in Chapter I, have used this issue to
discredit the accuracy of the autobiographical account. Indeed
it is difficult to understand how Collingwood could deny that in
the years following the war he thought and wrote as an advocate
of a school of thought widely recognized under the title
"idealism"--as the 1924 article on the "New Idealism"
illustrates. Yet in the Autobiography Collingwood writes that
Speculum Mentis was neither "usual" nor "idealistic" (A, 57). We
shall have to decide on the basis of his use of the term whether
this is a sheer piece of effrontery or if it would be like Hegel
denying that he was an idealist --where the term means
"subjective idealism" of the sort he attributed to the Kantian
philosophy.
The idealism-realism issue is also a strategic boundary
in the interpretation of Collingwood, since (as we also saw in
Chapter I) Rubinoff uses Speculum Mentis as a basic program for
interpreting the entire remainder of Collingwood's philosophy,
and views Collingwood as an

108
unregenerated idealist of an Hegelian pedigree, whereas Donagan
reconstructs what he regards as the steps by which Collingwood
worked his way free from his youthful idealism and forged a
philosophy of mind on linguistic and analytic principles. And
where Rubinoff has to explain why such terms as "absolute
knowledge" and "dialectic" tend to disappear in the later
writings, Donagan has to account for Collingwood's continued and
uncompromising anti-realism in these same works. And finally,
where Rubinoff has to account for the reversal from the
condemnation of abstraction in Speculum Mentis to the apparent
endorsement of it for all higher thought in The New Leviathan,
Donagan, who uses this as an index of Collingwood's conversion
from idealism, must account for the striking difference between
the descriptions of this process of abstraction in the same two
works-a point that we will examine more carefully in later
chapters.
In all of this discussion we must therefore try to be as
clear as the texts will allow us to be on the senses of the
terms "idealism" and "realism" as Collingwood formulates them.
This will necessitate trying to be clear about certain other
related terms, since what we are trying to do is to flesh out
the bones of the abstract formula for realism from the
Autobiography, "Knowing makes no difference to what is known."
What does Collingwood mean by each of the terms, "object,"
"knowing," and "makes a difference"? As we shall see in this and
subsequent chapters, the sense of

109
the formula shifts with the meaning of the terms: where
"knowing" means "perception" the difference knowing makes to the
object of perception is not completely equivalent to the
situation in which "knowing" means "history" and the object is
the thought of an historical agent. And when Collingwood rejects
realism with his autobiographical anti-realistic argument, one
must pay careful attention to the situation in which it is
applied in order not to discard the baby with the bath water.
When he rejects the abstractions of realism in Speculum Mentis,
for example, we must ask if there is any acceptable sense of
"abstraction" that is salvagable from the condemnation as
distinct from the realism that it is aimed at overcoming. And
similarly when we find the project of "absolute knowledge"
collapsing for lack of any concrete subject matter at the end of
Speculum Mentis, we have to ask if from the wreckage we can find
any principles on which a more solid structure of thought can be
erected. (Some of this work will occupy us in the next three
chapters.)
2. Realism and Idealism in Religion and Philosophy.
If we were to begin where a beginning should really be
made, we would probably never achieve the limited goals we have
set for ourselves. A proper assessment of Collingwood's
background would require sketching out not only the realistic
doctrines of Cook Wilson and the Oxford realists but also of the
shadowy figure of Bradley, whose thought was

110
still very much alive—even if only as the target of realistic
critics.1 From this point one could trace the issues backwards
or forwards: backward into the roots of Collingwood’s idealism
in Bradley, Green, Hegel, Kant, and Berkeley, and the sources of
the realism he opposed in Mill, Hume, and Locke; and forward
into the fruits of idealism contemporaneous to Collingwood in
the Italian idealists, Croce, Gentile, and de Ruggiero, and the
development of realism with Alexander, Moore, Ayer and Russell.
All of this would make fascinating reading and would be a
welcome study and an invaluable background work for the
understanding of Collingwood’s philosophy. Having said that we
shall say little more about these matters, except by way of
occasional footnotes suggesting interesting parallels or
contrasts. Our task is to understand Collingwood’s philosophy
as interpreted by the Autobiography, and limitations in both
space and our own background necessitate leaving these matters
for another time.
The place for us to start is with Collingwood’s early
publication, Religion and Philosophy, which we anticipate will
give evidence of Collingwood’s early tolerance of realism. As
expected, this work exhibits an ambivalent attitude toward
realism. On the one hand Collingwood writes that what he is
saying “contains little if anything which
____________________
1 C.J. Warnock, English Philosophy Since 1900 (Oxford,
1964), pp 1-12. Cf. John Passmore, A Hundred Years of Philosophy
(Baltimore, 1957/1968), pp 48-71, 240-57.
111
contradicts the principles of either Realism or Idealism in
their more satisfactory forms,"2 and adds that "There is an
idealism with which I feel little sympathy, and there is a
so-called realism which seems to be only distinguishable from
that idealism by its attempt to evade its own conclusions" (RP,
101, n. 1; FR, n. 1).3 On the other hand (and even this passage
shows a leaning in this direction), in the same work Collingwood
admits that in at least one controversy the position he is de-
fending "would claim the title of idealism" (RP, 94-95). The
only way for us to understand where Collingwood stands in this
early work is therefore to look a little more closely at the
positions he is analyzing and the arguments he puts forward to
confirm or reject them.
Religion and Philosophy is laid out in three parts. Part
I examines the general nature of religion, and attempts to
examine the
____________________
2 In this same footnote he cites with apparent approval
the work of Joachim on The Nature of Truth on the one hand and
Prichard's Kant's Theory of Knowledge and Carritt's Theory of
Beauty on the other as representing the "more satisfactory forms
of the theories he hopes not to be contradicting. In the
Autobiography Collingwood identified Carritt as his realist
tutor, and H. A. Prichard as following Cook Wilson, and H. H.
Joachim as a close personal friend of Bradley and later of
Collingwood (A, 18, 20-22).
3 In-text citations from Religion and Philosophy are

followed by citations from the publication, Faith and Reason,
ed. by L. Rubinoff, because the former is out of print, and
portions of it appear in the latter work. Where no second
citation follows the reference to RP, it does not also appear in
FR.
112
distinction of religion from ritual, conduct, and feeling, and
to identify it as "creed," i.e. its intellectual element.
Furthermore it examines the distinction between religion and
conduct (thought and action), history, science, and philosophy.
To anticipate later conclusions, we might add that it is in this
part that Collingwood not only engages in wholesale
identification of apparently distinct subject-matters (religion
as creed is identified with history, philosophy, and even, by
implication, science; thought and action are identified; etc.),
but also sets up many of the problems against which he would
later struggle in Speculum Mentis. Part II takes up the
metaphysical issues of proving the existence of God, the dualism
of matter and mind, personal identity, and evil--these issues
all approached in a manner which first analyzes a claim, then
its counter-claim, and then attempts to state on which side the
truth appears to reside. Part III is more properly theological
in tone, dealing in successive chapters with the self-expression
of God in Man, in the person of the Christ; God's redemption of
Man; and the problem of Miracles.
As can be seen from this glance at the table of con-
tents, there is a good deal of interesting material in this very
early work, and we must resist the temptation to deal with all
of it (some of it will appear in later chapters--the issue of
the "identities" will appear, for example, in

113
Chapter VI on rapprochement). What we seek to clarify at this
point is how Collingwood stood on the issues of idealism and
realism, and this requires us to limit ourselves primarily to
the material presented in Part II, "Religion and Metaphysics."
In particular we want to find out the way in which Collingwood
understood the independent status of objects of knowledge.
As in most of his subsequent writings, Collingwood
prefaces his positive treatment of the subject with a critique
of false views of the subject matter. In Part I Collingwood
singles out psychology as characteristic of the way the phe-
nomenon of mind in general, and religious thinking in particu-
lar, is improperly approached. When composing the Autobiography
Collingwood recalled this passage with evident approval (A, 93),
and in re-reading it he may have prepared himself for writing
the anti-psychological chapters in the Essay on Metaphysics,
which it strikingly anticipates. Psychology, Collingwood writes,
is distinguished from the philosophical sciences of logic and
ethics (which also study the mind) not by its subject-matter but
by its method.
The method peculiar to psychology may perhaps be described

as follows. The psychology of knowing differs from logic or
the philosophical theory of knowledge in that it treats a
judgement--the act of knowing something--as an event in the
mind, a historical fact. It does not go on to determine the
relation of this mental event to the "something" known, the
reality beyond the act which the mind, in that act,
apprehends. Such a further investigation would be
metaphysical in character and is therefore avoided by psy-
chology. Now this formula can be universalized, and thus
gives us the definition of the
114
psychological method. Take the mental activity as a

self-contained fact; refuse, so far as that is possible, to
treat of its metaphysical aspect, its relation with real
things other than itself; and you have psychology. Thus in
scientific thought as studied by logic we have a judgement
in which the mind knows reality: psychology, treating the
judgement as a mere event, omits its reference to reality,
that is to say, does not raise the question whether it is
true. (RP, 40; FR, 75-76).
In a footnote Collingwood adds that the same omission or ab-
straction is made by formal logic, which he takes to be a psy-
chological rather than a philosophical science (RP, 40, n. 2;
FR, 76, n. 2).
Here we have in germ a strategy which comes to fruition
in Collingwood's later philosophy of history as a re-
interpretation of the act-object distinction, i.e. the dis-
tinction between an act of consciousness and the object of such
an act, where the former is regarded by realists as an event in
the subjective or psychological life of a conscious agent (cf.
IN, 282-301). Our present interest is in the "metaphysical"
aspect of the "reality beyond the act" which the mind
apprehends. Throughout this chapter and those which follow,
where Collingwood's anti-realism is being evaluated we will use
this independent reality of the object as an index to measure
the degree of his anti-realism, bearing in mind that in each
case we must try to assess the meaning of the terms involved. It
is, after all, the object's independence from the act of
knowledge which is at the center not only of Collingwood's

115
autobiographical realist formula, but also of the historical
issue as Collingwood inherited it.4
The passage about psychology which we have just quoted
is taken from a chapter on religion and history, in which
Collingwood is trying to demonstrate that religious creed is not
devoid of reference to historical factuality. The attack on
psychology is carried out because the psychology of religion,
while pretending to deal with the phenomenon of religious
consciousness, fails to do so precisely because it ignores this
factual reference: a mind regarded in an external way, i.e.
without reference to its object, "really ceases to be a mind at
all," and the knowledge gained by studying it in this way "is
not knowledge of anything, but barren and trifling abstraction"
(RP, 42; FR, 77). In a later passage he comments that empirical
psychology treats mind exactly as if it were matter (RP,
__________________
4 Cf. Bertrand Russell, "Logical Atomism," in Contemporary
British Philosophy, ed. J. H. Muirhead (London, 1924), I, p.
360: "For some years I was a disciple of Mr. Bradley, but about
1898 I changed my views, largely as a result of arguments with
G. E. Moore. I could no longer believe that knowing makes any
difference to what is known." A more technical version of the
formula was given by Russell some years earlier in "Meinong's
Theory of Complexes and Assumptions," Mind, n.s. XIII (1904), p.
204: "every presentation and every belief must have an object
other than itself, and, except in certain cases where mental
existents happen to be concerned, extra-mental; . . . and . . .
the object of a thought, even when this object does not exist,
has a Being which is in no way dependent upon its being an
object of thought." R. M. Chisholm in his "Editor's
Introduction" to Realism and the Background of Phenomenology
(New York, 1960), p. 3, n. 1, quotes Russell as saying that he
had been led to accept these theses by Mr. G. E. Moore, and that
"Except Frege, I know of no writer on the theory of knowledge
who comes as near to this position as Meinong."
116
76). But when it comes to describing the nature of history in
contrast to this Collingwood endorses what appears to be a
blatantly realistic definition:
History must be regarded not as a mechanical process, nor

yet as a gradual accumulation of truths, but simply as ob-
jectivity; as the real fact of which we are conscious.
History is that which actually exists; fact, as something
independent of my own or your knowledge of it. In this
sense there would be no philosophy without it; for no form
of consciousness can exist without an object. (RP, 49; FR,
83).
At this point in his development Collingwood had not yet come to
a full realization of what he would later call the ideality of
history, and as the subsequent discussion illustrates, history
is not yet confined to deeds of men. In Chapter V we shall see
how this primitive idea of history became refined in the essays
written between 1920 and 1930.
But what of this concept of a fact as "something inde-
pendent of my own or your knowledge of it"? Where we might be
tempted to soften its realistic impact by qualifying it as dis-
tinct from an object (the factuality of anything is its given-
ness, its aspect of independence, which says nothing about the
facticity of objects, or about their independent existence or
reality), it is better to let it serve as a statement forming
the limit of Collingwood's early attitude toward realism. We
shall never again encounter him affirming anything like it--
unless it is as a premise which he was setting out to demolish.
The problem of the reality of objects makes its appearance in a
later chapter in the discussion of what

117
he calls the "plain man's metaphysic:"
Popular metaphysic distinguishes two categories of reality,

mind and matter. Mind is a reality whose qualities are
thought, will, and so forth; it is not extended over space
or divisible into parts. Matter, on the other hand, occupies
space, and is homogeneously subdivisible into smaller parts;
it has no consciousness of itself as mind does . . . . Mind
is active . . . matter is passive . . . . We have thus three
hypotheses before us. Either the world is entirely material,
or it is entirely spiritual, or it is a compound of the
two . . . . (M)aterialism will admit the existence of
thought, but will try to explain it as a kind of mechanism;
the opposite theory (which for the sake of convenience I
shall call idealism) will admit the existence of mechanism,
but will try to describe it in such a way that its operation
is seen to be a form of spiritual activity. (RP, 72-73).
In a footnote Collingwood adds that the sense of the term
"idealism" which is opposed to materialism "must be carefully
distinguished from Idealism as a theory of knowledge. The
former, concerned with the antithesis between mind and matter,
has no connexion whatever with the latter, which concerns the
quite different antithesis of subject and object, and is opposed
not to Materialism but to Realism" (RP, 73, n. 1).
Here a preliminary distinction has been drawn between
metaphysical and epistemological idealism, and although one
might question whether they truly have "no connexion whatever"
with one another,5 such an equivocation on the term might allow
Collingwood to deny being an idealist (in the
__________________
5 We shall try to avoid being drawn into a protracted
historical discussion of what the various meanings of the term
"idealism" have been--since the title encompasses everything
from Platonic archetypalism to Husserlian essentialism, and
along the way from one end of this spectrum to the other, cuts
across both Kantian and Hegelian territory. Nevertheless to say
that the metaphysical and epistemological
118
metaphysical sense) without being committed thereby to a denial
that he was an idealist in the epistemological sense. More
importantly, Collingwood does not view the realism-idealism con-
troversy in a metaphysical sense--as it might be viewed by
someone who takes realism to be a statement about independent
existence of objects, in which case it is saying something about
reality rather than something about our knowledge of it. For
Collingwood realism is not a metaphysical issue but an
epistemological one, i.e. one which says something not about
reality but about our knowledge of it.
After some argumentation criticizing rigid adherence to
the materialistic hypothesis (partly based on the impossibility
of importing mind-characteristics into a purely mechanistic
description of the world, and partly based on the failure of
mechanists to defend the principle of causality), Collingwood
pays the mechanistic devil his due. Materialism satisfies the
scientist's demand for uniformity, regularity, and
generality--all of which are satisfied by a materialistic
philosophy (RP, 91). But that is not all.
Another merit of materialism is its insistence on fact, on

reality as something beyond the power of the individual
__________________
forms of idealism have nothing whatsoever to do with one another
ignores the fact that Berkeley's idealism arises from a critique
of the concept of matter from a practitioner of epistemological
empiricism.
119
mind to create or alter. Matter is supremely objective. And

when it is said that mind is the only reality, the
suggestion at once arises that the world is less solid, less
satisfying, less "real" than we believed . . . . Materialism
. . . is right as against those theories which make the
world an illusion or a dream of my own individual mind; but
while it is right to insist on objectivity, it goes too far
in describing the objective world not only as something
different from, and incapable of being created or destroyed
by, my own mind, but as something different and aloof from
mind in general. (RP, 92-93).
We shall find Collingwood making this point in various forms
repeatedly throughout his philosophical writings, but usually
not in defense of materialism. Whatever mind may make of the
world, it is not simply the creation of one's own imagination
--a view that Collingwood called "subjective idealism" (RP,
120)6 and one which he tended to associate with solipsism.
____________________
6 This is the title Hegel applied to the Kantian cri-
tical philosophy. Hegel faulted Kant for stopping short with the
analysis of experience at that point at which objects of
experience are shown to be mere appearances (phenomena), whereas
the view that Hegel advocates is absolute idealism, which holds
that objects of experience are not only mere phenomena for us
but in their own nature, since their existence is founded not in
themselves but in the universal divine Idea. Cf. The Logic of
Hegel, tr. from The Encyclopedia of the Philosophical Sciences
by William Wallace (Oxford, 1873), pp. 93-94. Kant himself
refuted a form of idealism which he called "material
idealism"--the theory which declares the existence of objects in
space outside us to be merely doubtful and indemonstrable (the
problematic idealism of Descartes) or to be false and impossible
(the dogmatic idealism of Berkeley)--Immanuel Kant's Critique of
Pure Reason, tr. by N. Kemp Smith (New York, 1961), B 274, p.
244. Collingwood's use of the term is closer to Kant's "material
idealism" than to Hegel's "subjective idealism."
120
The outcome of the discussion of the matter vs. Mind
controversy is that "we cannot conceive matter without ascribing
to it some qualities of mind, nor mind without ascribing to it
some qualities of matter" (RP, 93-94)--in short, the "mixed"
hypothesis. But even here there is a definite tilt towards
idealism.
To ask whether mind is a form of matter or matter a form of

mind is largely a question of words. The important thing is
that we should be able to bring the two into relation at
all; that we should hold such a conception of matter as does
not prevent us from admitting truth, morality, and life as a
whole to be real facts, and that we should hold such a
conception of mind as does not reduce the world to an
illusion and experience to a dream. The first of these
errors is that of crude materialism, and the second that of
an equally crude idealism. The view for which we are
contending would claim the title of idealism rather than
materialism, but only because the current conception of mind
seems a more adequate description of the world than the
current conception of matter. We are laying stress on the
fact that the world is the place of freedom and
consciousness, not of blind determinism; and at present this
can best be conveyed by saying that mind is the one reality.
(RP, 94-95).
A similar fate awaits the immanence-transcendence question in a
later chapter. God must be regarded as both immanent and
transcendent: immanent because all human knowledge and goodness
are the very indwelling of his
_______________________
7 Collingwood closes the chapter with the suggestion
that he does not wish to exclude a "higher materialism" which
would "regard matter as nothing else than mind itself in its
concrete existence, and mind as the life and operation of mat-
ter"--but this must wait for physics to develop to the point
where it can adopt a principle which would regard all matter as
in its degree a form of life (RP, 95).
121
spirit in the mind of man; and transcendent because God has
attained these things whether or not man attains them, his being
not depending upon the success of human endeavor (RP, 119). But
when he asks if God exists only as a "spirit in our hearts" or
if he is also a real person "with a life of his own, whether we
know him or not," he confesses that this is not an easy phil-
osophical problem to solve.
The difficulty of answering this question is bound up with a

well-known philosophical puzzle, the puzzle of how to prove
the existence of anything except as present to the mind. If
it is true that things cease to exist when we are not
thinking of them, . . . then it follows by the same argument
that God is immanent only, and exists nowhere but in the
mind of men. But we cannot really believe that these things
are so . . . . The arguments for pure immanence are at
bottom identical with the philosophical creed of subjective
idealism, and with that creed they stand or fall. (RP,
119-20; FR, 188-89).
If God is not entirely an immanent idea, what objective
status does the concept have? Can the existence of God be
proven? Collingwood argues that the traditional proofs for the
existence of God are not so much impossible as premature: what
is required is to first define what one means by "God," which
involves finding some definite content to the concept-the task
of theology. "No one can prove that God exists, if no definite
significance is attached to the words" (RP, 64).8
______________________
8 Collingwood does not deny that the existence of God
can be proven--he merely postpones it until after an adequate
concept of God has been developed. But after he spends the
remainder of Religion and Philosophy expending considerable
effort to develop just such a concept, he does not conclude his
study with any such proof. Nor does one appear in
122
The only course open to the sceptic who doubts the existence of
God is to discover what he actually thinks, and then to find out
if that idea was justified or not (RP, 69).
This is an extension of a principle which Collingwood
lays down as fundamental to all thought, and which is an early
statement of the principle of intentionality, the modification
of which we will be watching closely in succeeding chapters. In
its earliest formulation it is expressed as follows:
The mind is specifically that which knows the

object . . . . The mind seems to be not so much that which
thinks as the thinking itself; it is not an active thing so
much as an activity. Its esse is cogitare. . . . All con-
sciousness is the consciousness of something definite, the
thought of this thing or of that thing; there is no thought
in general but only particular thoughts about particular
things. The esse of mind is not cogitare simply, but de hac
re cogitare. (RP, 100; FR, 172).
When two minds think the same thought it is never exactly the
same in terms of emphasis or applications peculiar to the
individual; but that does not mean that such a difference
destroys the identity of the truth, or its ability to be the same
truth thought by two different minds (RP, 106; FR, 177).
___________________
any of his writings. What one finds instead is a favorable
discussion of the ontological argument in the Essay on
Philosophical Method (EPM, 124-26), and a very idiomatic use of
the argument in the Essay on Metaphysics (EM, 185-90), which
declares that the statement-God exists" is not a verifiably true
or false proposition, but an absolute presupposition, and
therefore that the ontological proof is only a way for
Christians like St. Anselm to say what in fact they believe. If
Religion and Philosophy is Collingwood's theology, his
philosophy of religion remains unpublished. (Cf. RP, 16; FR,
53-54).
123
In Chapter VI we shall take up the issue of Collingwood's use of
the term "identity" as he uses it in Religion and Philosophy and
Speculum Mentis.9
On the strength of the passages we have been examining
we can say that in this early publication Collingwood already
shows a leaning toward an epistemological form of idealism, but
at the same time he rejects both a "metaphysical" or subjective
idealism which would assert that objects of knowledge are simple
creations or imaginings of subjective consciousness. With
respect to realism he displays a certain tolerance--as in the
notion of "fact" as something independent of any knowledge of
it, and in the distinction between an act of consciousness and
its object (although he says nothing about the esse of the
object to correspond to the esse of
____________________
9 Our purpose in this chapter is to examine Collingwood's
early views on realism and idealism, and it is only secondary to
this that we have become involved in the issues of the
philosophy of religion. But we do not mean to leave the reader
dangling on the issues we have just raised--some of which we
will meet again repeatedly in different chapters. The conception
of God that Collingwood develops in Religion and Philosophy is
basically that of the Christian God--centering on the
Incarnation as the means by which the creator God overcomes his
transcendence. The unity of man and God occurs through a
"concrete union . . . attained in and by the identification of
the self in all its aspects with the perfect mind of God" (RP,
150; FR, 254). This is achieved in the person of the Christ, who
"has absolute experience of the nature of God and lives in
absolute free obedience to his will" (RP, 166; FR, 267). Insofar
as any man achieves a similar concrete identity of will with the
mind of God he achieves union with God (RP, 160, 167; FR, 262,
268). Whether such a person as the Christ actually existed is an
historical issue that Collingwood declines to consider (RP, 151;
FR, 254).
124
consciousness). Similarly he shows a tolerance for the positive
aspects of materialism as a philosophy which, with proper
modifications, is able to satisfy the scientific demand for
uniformity and generality. We have not yet found any form of the
anti-realism argument as outlined in the Autobiography, nor any
linkage of realism and abstraction. The situation, in short, is
pretty much what we would expect it to be on the grounds of the
autobiographical interpretation.
3. Idealism, the Absolute, and the Metaphysic of Knowledge.
In a short piece published after a symposium in 1923
(the year before Speculum Mentis), Collingwood responded to a
paper by Evelyn Underhill on the "new idealism"--by which she
meant the philosophy of Croce and Gentile (FYC, 85). Under hill
charged the Italian idealists with failing to provide room for
mysticism in their account of the forms of experience, and
Collingwood (who had just finished translating Guido de
Ruggiero's Modern Philosophy 1O) undertook a defense of Gentile's
___________________
10 Collingwood translated three of Croce's works and two
books by Croce's disciple, Guido de Ruggiero--see Johnston, FYC,
66. Johnston's book also has a sketchy chapter on Collingwood's
relation to Croce and the Italian idealists, whose influence on
Collingwood Johnston aptly summarizes under the rubric
formulated by Collingwood himself, that "to borrow is to
interpret" (FYC, 89). "Collingwood asks us to focus not on what
was borrowed, but on what led the borrower to select what he did
. . . . It was the multiplicity of his interests and his command
of many fields of learning which made Collingwood 'capable of
borrowing' from Croce, Gentile, and Vico. It was his almost
unique intellectual versatility which 'laid (Collingwood) open
to their influence'" (FYC, 88-89).
125
religious philosophy against her attack. The interest for us of
this brief and tactful article is in the distinctions that
Collingwood draws concerning idealism, which if they do not
directly reflect his commitment to the school of thought he is
discussing, at least indicates to us his understanding of it.
Collingwood corrects Miss Underhill's misconception that
the new idealism "dispenses with mysticism" in the sense that it
ignores an intuitive or immediate consciousness of the supreme
reality as one, eternal, and spiritual. On the contrary, both
Croce and Gentile identify mysticism with religion (FR, 270).11
Gentile in particular does not deny the existence of an absolute
object of thought such as that which mystics
contemplate--something that is one, eternal, and unchanging; nor
is Gentile's philosophy exhausted by calling it a philosophy of
change like that of Bergson for whom absolute reality is an
absolute flux (FR, 273). This misconception is taken over from
Bosanquet, whose view Collingwood corrects as follows:
_____________________
1l Collingwood recognizes that Croce did not represent
religion as one of the "necessary forms of the spirit" in his
systematic philosophy (FR, 270), but adds that it provides the
hint of a new attitude towards religion that "in Gentile blos-
soms into a complete new attitude to religion" (FR, 271-72).
Johnston adds that Croce "accords religion scant place," and
that in this "he differs significantly from Collingwood, who all
his life regarded religion as a necessary, even indispensable
component of culture" (FYC, 70).
126
reality, for Gentile, is history. Now history is not, as

Miss Underhill assumes, a synonym for change. Change
is . . . a realistic concept, history an idealistic. That
which changes is a mere object, which need not know that it
is changing, and indeed which no one need know to be
changing. The philosophy of change is a "metaphysics of
being," that is, a philosophy which tries to describe the
world as a thing in itself without raising the question how
it comes to be known. And there can be little doubt that the
philosophy of change makes the world unknowable. That which
has a history, on the other hand, is a mind, for matter may
change but it cannot be said to have a history . . . . Hence
Gentile's philosophy is a "metaphysic of knowledge," that is
to say, a philosophy which never loses sight of the question
"how do we come to know what we know?" (FR, 274).
In the light of what we shall find Collingwood saying in
Speculum Mentis in this and the next three chapters, it can
hardly be doubted that Collingwood would describe his own
philosophy as a "metaphysic of knowledge." But is that
idealism? By 1923 it is clear that Collingwood had recognized
the essential ideality of history, and had already moved away
from a conception of history as simply "factuality." In Chapter
V we shall observe the staging of this development; at this
point we are interested in the way in which Collingwood
describes idealism. In the essay under consideration Collingwood
finds in Gentile an expression of the common ground of all
idealism (FR, 277), the double aspect of mind as both active and
passive, expressed by Gentile as the identity of act and fact:
(C)hange in a mind must be change for that mind, a change

of which That mind is conscious; and to be conscious of it,
the mind must somehow be raised above it. How is this ap-
parent contradiction to be realized? How is the mind to be
at once in change and out of change?
127
Only if the mind originates change in itself. For then, as

the source and ground of change, it will not be subject to
change; while on the other hand, as undergoing change
through its own free act, it will exhibit change. This
double aspect of the mind as active and passive is the very
heart of Gentile's philosophy. It is his favorite
distinction of act and fact. (FR, 275).
But if the identity of act and fact is the equivalent of Croce's
principle of immanence, what room is there for a principle of
transcendence? Gentile assigns the name of religion or mysticism
to the losing of the mind in its object--the transcendent
element of all human life. The synthesis of the immanent element
of life (which Gentile calls art) with the transcendent is
philosophy, which seeks the absolute, defined (citing Hegel for
support) as "that which has reconciled its own opposite to
itself, and therefore no longer stands in opposition to it" (FR,
276). Collingwood therefore states that in pursuing the
absolute, Gentile's philosophy "is as convinced of the necessity
of transcendence as Miss Underhill herself . . . . That
reconciliation of the opposing principles of immanence and
transcendence which both regard as possible, necessary, and
indeed actual, she calls mysticism, and ((Gentile)) calls it
philosophy" (FR, 276-77). It is, he adds, in basic agreement
with Hegel's Absolute Spirit and the post-Kantian tradition in
philosophy (FR, 277). For this tradition, mysticism is a thing
which cannot be dispensed with -- not as something intuitional
or wholly immediate, but as something assimilated by the labor
of the life of the mind (FR, 278-79).

128
The necessity of mystical experience lies in the principle

that we discover new truths neither by the inference of the
logic-books nor by the intuition of Aristotle, but by an act
of mind which reaches out beyond the given, grasps the new
thought as it were in the dark, and only after that
consolidates its new conquest by building up to it a bridge
of reasoned proof. (FR, 281).
In this discussion of Gentile we find Collingwood de-
fending a form of absolute idealism, and defending it against
attack not from the point of view of someone who finds it
problematic in the sense of failing to show how the object is
unaffected by the knowing of it (epistemological realism), but
from the point of view of someone who charges it with failing to
provide sufficient grounds for mystical religious experience. In
the next section we shall find Collingwood widening this
discussion to include all forms of experience, and deepening his
commitment to absolute idealism at the same time that he begins
to lay the foundation for moving beyond it.
4. Absolute Idealism and the Forms of Experience.
In 1916, when Religion and Philosophy was published, we
found Collingwood stating that the argument he was advocating
did not conflict with either realism or idealism in their more
satisfactory forms. By 1924 not only had his point of view
shifted in the direction of absolute idealism, but his earlier
tolerance had completely vanished. In Speculum Mentis
Collingwood was prepared to take a stronger stand against re-
alism and all its ramifications, and it was this work that
earned him the

129
reputation of being an "idealist." It is a label that he
explicitly rejected, but in much the same sense that he had
rejected it in Religion and Philosophy:
Idealism . . . is the doctrine that the world is made, so to

speak, of mind; and is regarded as the opposite of ma-
terialism or the doctrine that the world is made of matter.
Both of these theories begin by abstracting the object of
knowledge from the subject, and both go on by inquiring into
the nature of the object in this abstraction, regarded as a
thing in itself. Both agree in committing the fundamental
error of separating the metaphysical inquiry as to what the
world is in itself from the psychological inquiry as to how
we come to know it. Idealism in this sense leaves
unreconciled the opposition between subject and object, and
therefore sets the object outside the subject; . . . it
tries to bridge the gap by ascribing to the object some kind
of consubstantiality with the subject, turning it into
another mind, a society of minds (spiritual pluralism) or an
infinite mind (theism). With anything which deserves the
name of idealism in this sense we have nothing to do except
reject it. (SM, 266-67).
What is being rejected is the "metaphysical idealism" that he
had contrasted with materialism in Religion and Philosophy.
But what of the "epistemological idealism" that is
contrasted with realism over the issue of whether knowledge
makes a difference to the object known? Speculum Mentis widens
our horizon beyond the object of religious knowledge; it is an
analysis of successive "forms of experience," each of which
competes with all the others for the "prize of truth" --that is,
the successful fulfillment of its claim to be true knowledge
(SM, 42). Each form of experience is a conscious attitude with
respect to the known, an activity which has both cognitive and

130
practical aspects (SM, 39, 42, 44). In short, each is knowledge
claiming to be wisdom.
Five such forms of experience are examined--art, reli-
gion, science, history, and philosophy--although the list is not
exhaustive of all possible forms (SM, 41, 57, 280). As concrete
activities engaging the whole self, each form of experience
regards the others as illegitimate contenders for the prize of
truth, "rival ways of conceiving the whole"--the "whole" being
that conception of reality which will allow the mind to live the
unified life that it sees and needs in order to be totally
satisfied (SM, 36-37, 47-48). Consequently the forms of
experience cannot be regarded as mere species of a genus, each
taking a portion of the prize: "each denies the others; and
because they are not species they have not that indifference
with respect to one another which characterizes abstract logical
classifications" (SM, 55). Each iS to be examined on its own
merits or in accordance with its own claims, in order to
discover whether its claim is consistent with its actual
performance.
Our map ((of knowledge)) . . . is to be a statement of the

essential nature or structure of each successive form of
experience, based on actual knowledge of that form from
within, and concentrated upon the search for inconsisten-
cies, rifts which when we come to put a strain on the fabric
will widen and deepen and ultimately destroy it. (SM, 45).
When such inconsistencies appear it is the task of a higher order
form of
131
experience to repair the damage by constructing "from without" a
self-conscious justification of the primary, inconsistent form
(SM, 250, 252-55).
Since we will be engaging in an explicit discussion of
the form of experience called science in Chapter IV, of history
in Chapter V, and of the overall argument of Speculum Mentis in
Chapter VI, rather than attempting a detailed analysis of the
contents at this point we shall first present Collingwood's own
summary of the five forms of experience in the Outlines of a
Philosophy of Art which appeared the following year (reprinted
in EPA, 45-154), and then focus on those sections of Speculum
Mentis which particularly reveal Collingwood's views on idealism
and realism.
After stating at some length a general theory of art,
Collingwood locates art within the context of the life of the
spirit, which he characterizes as follows:
The life of the spirit is an indivisible whole within which

are necessary and permanent distinctions: permanent in the
sense that the spirit in its own activity perpetually
affirms them, and necessary in the sense that the attempt
not to affirm them would merely result in affirming them
over again. Fundamentally, the spirit is awareness or
consciousness, which implies a prima facie distinction
between the conscious spirit and the world of which it is
conscious; but since this awareness is itself an act, a
self-modification on the part of the spirit, the passivity
of pure awareness rests upon the creativity of action, and
the life of the spirit is a whole within which
consciousness and action, awareness of the world and modi-
fication of the world are correlative elements. The unity
of these two elements is feeling, where that of which we
are aware is our own states, and
132
these states are identical with the feeling of them: they

are at once states of consciousness and objects of
consciousness.(EPA, 137-38).
Coming upon this passage after reading Collingwood's defense of
"the new idealism" it is hardly possible to ignore Collingwood's
commitment to an idealism at least similar to that of the
Italian philosophers, Croce and Gentile. For certainly the view
expressed here represents the unity of "act and fact" or the
duality of spirit or consciousness as both active and passive--a
view that Collingwood had called the common ground of all
idealism. Furthermore it appears that this distinction is at
least post-Kantian, and probably Hegelian: the con-
sciousness-object distinction is described as occurring within
spirit ("a self-modification on the part of spirit") rather than
forming the limit of spirit beyond which lies an unintelligible
thing-in-itself.
Furthermore Collingwood imbues this structure with an
internal dynamism which propels it through the stages of the
"forms of experience":
Hence a rhythm in which awareness and activity concentrate

themselves into the unity of feeling, and feeling again
articulates itself into awareness and activity, is funda-
mental in all aspects of spiritual life. But life is not a
mere rotation of three psychological categories in a
rhythmical monotony. This triple rhythm is present in all
life, but it is never twice alike; its whole character is
altered by the specific difference of the experience in
which it is embodied. These differences emerge in the course
of a process which on its theoretical side may be called the
spirit's attempt to know itself, on its practical side the
spirit's attempt to create itself. To know itself means also
knowing its world, and to create itself means creating its
world; its world in the former case means the world
133
of which it is aware, in the latter case the world in which

it can live. (EPA, 138).
It is clear that when Collingwood refers to "objects" or "the
world" he means primarily objects of consciousness--what he had
called, in Religion and Philosophy, de hac re cogitare. It
refers to the objective correlate of acts of consciousness, and
when the term "object" is used henceforth without further
qualification this is the sense we shall have in mind. If we
mean to refer to a "thing" we shall call it a "physical object"
or an object having extra-mental reference.
The first stage in the life of spirit is the life of art
or the pure act of imagination--i.e. "the act of consciousness
which presents to itself an object of whose relation to other
objects it takes no cognizance" (EPA, 139). It is important to
bear in mind that by imagination Collingwood does not mean a
faculty which creates objects which appear to be presented in
perception; in an earlier section of this essay Collingwood
makes this point more clearly than he ever did in his subsequent
writings:
In art there are always a subject and an object, a con-

templator and something contemplated. But the subject's
activity, the object's nature, and the character of the
relation between them have certain peculiarities which
distinguish the case of art from other cases. What the
subject does is to imagine: the object is an imaginary
object, and the relation between them is that the individual
or empirical act of imagining creates the object. In
knowledge, on the other hand, the object is real; and the
relation between them is that the empirical act of knowing
presupposes the object and does not create it.
134
This may be said without prejudice to the idealistic view

that there is an absolute or transcendent sense in which
knowing creates its object; for no idealist is so innocent
as to confuse knowledge with imagination and to suppose
that what we generally call knowing is simply imagining.
(EPA, 52).
The caveat against the "idealist" here is clearly against the
"subjective idealist." Furthermore by "the empirical act of
knowing" Collingwood clearly means perception. We shall find
that throughout Collingwood's writings he never intimates that
in perception consciousness creates an object ex nihilo: and
this is the abhorrent sense of "idealist" that he rejected
throughout his published works.
If the essence of artistic consciousness or aesthetic
experience is its monadism, i.e. its contemplation of an object
without relating it to other objects, its practical aspect
appears in play--the immersing of ourselves in an activity
without any question as to the relation of this activity and
anything else. "Just as art does not explain itself by stating
reasons, so play does not explain itself by stating reasons; and
immediacy means the absence of reasons" (EPA, 139-40). But
neither simple imagination nor simple play can remain in this
immediacy as a complete and self-contained form of consciousness:
It is only within a consciousness which distinguishes truth

from falsehood that we can find in actual existence that
consciousness which does not distinguish them . . . . The
question "what am I?" can therefore no longer be answered in
terms of imagination; I am not
135
merely an imaginer but a thinker. The question "what is my

world?" must be answered by saying that it is a world not
merely of fancies but of realities. But if I who think am
also the I who imagine, it would seem natural to superimpose
the act of thinking on the act of imagining in such a way
that the real is merely one division of the imaginary. The
only world whose existence we have learned to recognize is
the world of our own imaginations; and when the distinction
between reality and unreality forces itself upon us . . . we
impose this distinction upon the world of imaginations, and
regard certain imaginations as true and others as false. To
do this is to break with the life of art; for . . . now we
are asserting one imaginary object as real, and denying
another as unreal; and to do this is to embark upon the life
of religion. (EPA, 140-41).
I have quoted this passage at some length because it is
our first contact with a transition between forms of con-
sciousness, and is expressed without much of the complicating
circumlocutions of Speculum Mentis. In fact it is exemplary in
its simplicity, and gives us a clear sense of how Collingwood
envisioned distinctions occurring within the unified life of the
spirit: when a distinction like that between reality and
unreality "forces" itself upon us, it is imposed on the unity of
the form of consciousness which thereafter regards its objects
under this oppositional distinction. In religion this takes the
form of mythological or metaphorical expression, insofar as
religion "says one thing and means another" by using imagery to
convey a truth. Nevertheless religion marks an advance in the
life of spirit:
(I)n religion that indifference to the distinction between

real and unreal, which is the essence of art, is abolished.
Religion is essentially a quest after truth and explicitly
conscious of itself as such a quest. But the truth which it
can and does discover is a truth
136
which is always hidden from view in a reliquary of

symbolism: we see the imagery, but we do not see the truth;
we are only conscious that the truth is there, and its
presence converts the beauty of the imagery into holiness.
But inasmuch as this holiness is a property of a mere
symbol, religion always contains an element of idolatry and
superstition. (EPA, 141).12
Religion, like art, cannot survive the disruptive force
of the tension between metaphorical and literal language. When
it attempts to overcome its own superstitious tendencies it is
forced to distinguish thought from its own imagery, and in the
process "the symbol loses its holiness and becomes merely sig-
nificant" (EPA, 142). With the distinction between metaphorical
and literal expression the life of explicit or self conscious
thought is reached. While art forgets the presence of thought
and concentrates on the pure imagery of language, and where
religious thought was immediately identified with the language
expressing it, scientific consciousness separates thought from
language and intellect from imagination.
Here thought is regarded as an activity self-contained and

self-sufficient, and its object as a self-contained and
self-sufficient intelligible world, reached through, but
lying behind, the sensible world. The aim of science is to
apprehend this purely intelligible world as a thing in
itself, an object which is what it is independently of all
thinking . . . . The world of thought is the universal, the
timeless and spaceless, the absolutely necessary, whereas
the world of sense is the contingent, the changing
__________________
12 Remarks like this make it difficult to assess Col-
lingwood's true estimate of religion, which here seems to have
slipped somewhat under the high position he accords it in
Religion and Philosophy. In Chapter VI we shall review
Collingwood's remarks about religious consciousness and try to
judge the extent to which his view of religion is reductive.
137
and moving appearance which somehow indicates or symbolizes

it. (EPA, 142).
The distinction is the separation of reality from appearance,
the necessary from the contingent, which ultimately poses prob-
lems for scientific consciousness. Science cannot bridge the gap
between the abstract universal and the particular, the necessary
and the contingent, reality and appearance.l3
The overcoming of this opposition is the achievement of the

form of experience known as history. Appearance and reality,
imagination and thought, have been merely distinguished and
not related: they must somehow be brought together again and
shown to be equally necessary, each to the other. This need
is satisfied by the historical consciousness, whose object
is the individual; no longer an abstract universal divorced
from its own equally abstract particulars, but a universal
that particularizes itself, a particular constituted by its
own universality. For history, the truth is no longer an
abstract necessity which nowhere actually exists; it is
concrete and actual, it is real in every sense of the word,
while the truth of science is a reality which is in one
sense utterly unreal, and ideal never realized, a law which
has no instances. (EPA, 143).
___________________
13 Collingwood's view of science in these early writings
is primarily based on the classical Greek notions of science,
perhaps modified by Renaissance advances, but hardly based on
first-hand knowledge of work in the physical and biological
sciences of his day. It is therefore somewhat understandable why
this form of experience is least articulated and not altogether
satisfactory. Having stated the requirement for other forms of
experience that the philosopher of it must be one who not only
observes the experience of others but has engaged in the
activity for himself (EPA, 153), Collingwood was in a weak
position to describe the nature of scientific experience.
Nevertheless we shall try always to evaluate what he says about
it as a reflection of his own understanding of it rather than
what natural science is in itself or in the contemporary
understanding of it. What he does understand by it is sometimes
quite remarkable, as we shall see in Chapter IV.
138
But history too has its fatal flaw. The historian presumes a
world of fact which is already there for him to discover, and
for which the historian is a mere spectator. The fact is a thing
in itself, "a thing whose existence and nature are supposed to
be wholly independent of the thinker" (EPA, 143). This
constitutes the separation of subject and object--a relic of the
abstractness of science remaining in history.14
This abstractness is only overcome in philosophy. The object

of philosophy is nothing short of reality, a reality which
includes both the fact of which the historian is aware and
his awareness of that fact. The philosopher . . . is not,
like the historian, outside his own picture; he sees himself
as part of the historical process which he studies, and
therefore part of his problem is to understand how that
historical process has thrown up in its development an
organ--namely himself--which is at once a part of it and the
spectator of it. With this clue in his hand . . . he is able
to reinterpret that process itself, and to see in every
phase of it a nisus towards self-consciousness. And in
realizing that history is the emergence of the spirit's
consciousness of itself he is actually achieving that
consciousness . . . . His knowledge is therefore explicitly
action; he is creating himself by knowing himself, and so
creating for himself an intelligible world, the world of the
spirit in general. (EPA, 143-44).
Since we are treading on the margins of territories that we
prefer to leave for later chapters, we shall leave off this
discussion of the forms of experience with the conception of
philosophy just given. In Chapter VI
_______________
14 As Collingwood's idea of history developed he struggled
against this very criticism, and his development of the concept
of "re-enactment" is an attempt to overcome the difficulty
completely. The extent to which this constitutes an adequate
rapprochement between history and philosophy will be evaluated
in Chapter IX.
139
we shall be in a better position to evaluate Collingwood's
project for a philosophy of "absolute knowledge."
But we have several particular ideas from Speculum
Mentis to attend to, and to these we must now turn.
5. Speculum Mentis and the Emergence of Explicit Anti-Realism.
In the last section when we discussed the way in which
the five forms of experience competed, as rival ways of con-
ceiving the whole, for the prize of truth, we noted that when a
particular form of experience exhibits inconsistencies it be-
comes the task of a higher form of experience to resolve the
disparity. That form of experience is philosophy, whose main
task is to attain self-consciousness and therefore to overcome
all the contradictions that arise due to the subject-object
dichotomy. In Speculum Mentis this process gives rise to the
construction of what Collingwood calls "dogmatic philosophy"--
i.e. a form of consciousness which is conscious of itself, but
imposes on itself the limits of a particular form of experi-
ence--art, religion, etc.--which it undertakes to justify as the
only true form of experience. But these are errors that the mind
makes about itself, and since the esse of mind is de hac
re_cogitare, it follows that when the mind takes itself as its
object, an error about that object is an error that it makes
about itself--an error that is immediately reflected in the
mind's activities under the

140
erroneous conception of itself: "the mind, having formed a false
conception of itself, tries to live up to that conception. But
the falseness of the conception just means that it cannot be
'lived up to"' (SM, 250).
Art, religion, science, and history are thus philosophical

errors, and owe their characteristics, and the char-
acteristics attributed by them to their ostensible objects,
to the initial error on which each is based . . . . Each
grasps "one aspect of the truth," as we say, forgetting
that truth is a whole whose aspects cannot be thus
separated: each is true, even while it is false . . . .
Error is always present in truth, but negatively present,
that is to say, it is present as that which is
denied. . . .This interdependence of truth and error, error
containing truth positively and truth containing error
negatively, is not only a fact easily verified in empirical
observation . . . but is a corollary of the fact that all
knowledge is self-knowledge, and every error an error about
the knowing mind. Hence an abstraction which separates
subject and object also separates truth and error (good and
evil, and so on), and . . . is a logical consequence of
realism, for if there were a world of real objects
completely other than the mind, absolute errors could no
doubt be made concerning them. (SM, 250-52).
Here for the first time we encounter Collingwood at-
tributing to realism all the errors that the mind makes about
itself. But if we are to become aware of the genesis of this
generalization, we must return to the conception of "fact" with
which we began our investigation of realism in the section on
Religion and Philosophy. Where in that earlier work Collingwood
appeared to be content with leaving uncriticized the concept of
historical "fact" as "something independent of my own or your
knowledge of it," this immunity is no longer respected. The
philosopher knows what

141
the historian does not know, "that his own knowledge of facts is
organic to the facts themselves, that his mind is these facts
knowing themselves and these facts are his mind knowing itself"
(SM, 295; cf. SM, 287). Modern realism arises from the discovery
of the concept of fact, which soon develops into the "historical
form of dogmatism" (SM, 281):
Historical dogmatism is the assertion of fact as ultimately

real, and fact means not only the facts of "history" but the
facts of perception. Such a dogmatism may take a
considerable number of forms ((several of which Collingwood
lists: monism and pluralism, intuitionism and intel-
lectualism, materialism and spiritualism, etc.)) . . . .
That which unites all these divergent views is their common
assertion of the positivity of the object, that is, their
denial that the object is conditioned or affected by
becoming known to any thinking mind or to what, with a
question-begging epithet, is sometimes called finite mind:
its finiteness being just this indifference to it on the
part of its object. Such a realistic account of the object,
as positive fact indifferent to its being known, is at first
sight compatible with any theory as to what the ultimate
nature of this object may be; and so we get all manner of
realisms . . . all equally capable of being held in
combination with the fundamental thesis of realism, which is
distinguished from them as "theory of knowledge" from
"metaphysics." (SM, 282-83).
Here we find a clear statement of what the Autobiography called
the central doctrine of realism. This formulation never changed
substantially in Collingwood's writings.15
______________________
15 Cf. EPA, 182-83 (1925); EPH, 99-100 (1933); EPM,
161-62, 169-70 (1933); IH, v, 142 (1936). The term "realism"
does not appear in some of the later writings because the move-
ment ceased to maintain its positions under that title. But
Collingwood's arguments were directed against the same positions
as they were now maintained under the titles of positivism and
empiricism: cf. PA, 130-31, 149-51 (1938); EM, 34-35, 37-38,
337-38 (1940); NL, 5.2, 5.31-5.32, 5.39 (1942).
142
Just as in the Autobiography it is presented primarily as a
negative thesis: the negative formula that knowledge can make no
difference to its object. But the realist is caught in a
positive antithesis to which he is equally committed--a
coincidentia oppositorum which could not help but embarrass a
subscriber to formal logic. If reality consists of a collection
of objects independent of the mind which knows them, then a
pluralistic realism is affirmed--a universe in which objects and
minds both occur. But on the other hand the historian presumes
that all facts fall into place in a single all-embracing system
of fact, and this system is the absolute, the ultimate
reality--a monistic realism which conflicts with its latent
pluralism precisely over the status of the thinker. Collingwood
phrases the dilemma as follows:
For either the thinker himself falls inside the absolute

whole or he does not. If he does then differences in his
thought about it make a difference to it, and the more
concretely real--that is, organized and interconnected in
all its parts--it is, the more fundamental these differences
will be and the more completely the positivity of fact is
lost. If he does not, then the monistic doctrine is
surrendered and we return to pluralism. (SM, 283-84).
While this is an issue that appears to arise only from within
the perspective of the form of experience called philosophy,
Collingwood emphasizes that it is a problem that arises from
within the historical standpoint itself. The fundamental
principle of history, the concreteness of the object, makes it
impossible for it to define its object in such a

143
way that it leaves out the subject, i.e. the historian, and
"compels us to recognize an object to which the subject is
organic, in the sense that the subject's consciousness of it
makes a real difference to it as a whole and to all its parts,"
so that "(b)eing known, whether truly known or erroneously
known, must make a difference to the object" (SM, 244).
Just as the formulation of the central doctrine of
realism corresponds to the statement of it in the Autobiography,
so one also finds in Speculum Mentis a rejection of the
principle in an argument employing the same strategy. However
the argument introduces a complication in the form of the
principle of abstraction, an issue to which we shall return
after having a look at the argument.
(A)ny object considered in abstraction from a mind which

knows it is neither material nor mental, but an illusion, a
false abstraction. Thus we do not say that the objective
world in itself is mental. If we are asked what it is apart
from a mind that knows it, we shall answer that it is not
"apart from" such a mind; it is "with" it in the sense of
being known by it. If we are asked what it would be apart
from such a mind, we shall answer that the very question
implies the suggestio falsi that we can describe that which
by definition is unknown. (SM, 267-68; cf. SM, 241).
Notice that the nerve of this argument is the absurdity that
results from suggesting that we can describe (and hence know)
what has been defined as something that is unknown, i.e. the
world as it is "apart from" the mind which knows it. This is
slightly different wording from the Autobiography,

144
but essentially the same argument as the one that he claims to
have given at Oxford in a paper which refuted the central
negative tenet of realism (A, 44). It is also noteworthy that
although the phrase "in abstraction from" is employed in
connection with this refutation of realism, its use in the
argument is confined to the meaning, "apart from," as is
indicated by the fact that Collingwood replaces the phrase with
this expression without any change in the resulting sense of the
argument.
But since Alan Donagan has made the issue of abstraction
central to his interpretation of Collingwood's development (see
Chapter I), we should take careful note of this meaning of the
term in Speculum Mentis, so that when we encounter it again in
its altered meaning in Chapter VIII we shall have some basis for
comparison. Furthermore it plays such a prominent role in
Speculum Mentis and its connection with realism is so strong
that it would be hard to ignore it. Collingwood himself ties the
concepts of realism and abstraction together:
Such was . . . and to some extent still is, the belief of

eminent realists, who sum up their own position in the
negative formula that knowledge can make no difference to
its object. On the other hand, it is not possible to assert
so much as this without asserting more, namely the principle
of abstract thought; for what is explicitly asserted is the
complete separateness of subject and object, their
independence of one another: and this implies that there are
facts in existence which are thus completely independent. It
is therefore correct to maintain that realism commits its
author to the principle of pluralism; and pluralism only
means the scientific abstraction of the universal from its
particulars. This path, therefore, leads from historical
dogmatism back to scientific dogmatism. (SM, 283).
145
It is this double implication that causes the basic inconsis-
tency within realism itself, and is responsible for the dual
role that realism is assigned in Speculum Mentis. On the one
hand realism is identified with the particular form of dogmatism
which he calls "historical"--history being taken as "the
assertion of concrete fact," and realism as the assertion of
this fact as ultimately real (SM, 281-87, 201-11). On the other
hand realism is identified with all forms of dogmatism, insofar
as realism is the willful resistance to any doubt that subject
and object are in all cases separate from one another, such an
act of will being the essence of the dogmatic attitude (SM,
259). Realism thus appears to be identified with both the
principle of abstraction (where abstraction means separation of
subject and object), and the principle of concrete factuality.
Collingwood himself recognizes this dual role, and charges the
realists themselves with it:
In spite of the simplicity of these difficulties, they have

not as yet been fairly faced by a single realist with whose
work the present writer is acquainted . . . . The fact is
that modern realism is essentially inconsistent. It is a
halt, or rather a confused running to and fro, between two
principles, the abstract concept and the concrete fact (SM,
284-85).
Speculum Mentis thus emerges as a philosophical work
which not only rejects realism but does so in large part on the
grounds of its principle of abstraction. It is the latter
principle which is responsible for all the errors that con-
sciousness makes about itself. "There is not

146
only one dogmatism; there are as many types of dogmatism as
there are types of abstraction" (SM, 268).16 "Every error is a
lapse from concreteness into abstraction, and all abstraction is
dogmatism" (SM, 288). "(T)he abstraction which separates subject
and object also separates truth and error (good and evil, and so
on) . . . . Such a melodramatic view of life is a logical
consequence of realism (SM, 252; cf. SM3259). "Abstract know-
ledge is the same as error . . . . But all error contains an
element of truth and the error appears as the externality of the
object, its otherness with respect to the mind" (SM, 313).
Furthermore realism and abstractionism are also held
responsible for most of the pernicious consequences of realism
that Collingwood listed in the Autobiography: an ethics which
separates knowledge from conduct (SM, 169-72); an epistemology
which defines knowledge in terms of intuition (SM, 188-94; cf.
SM, 255, 262, 283, and 293); a metaphysics which deals with
hypostatized or abstract universals concerning "nature or the
objective world" (SM, 271-81, especially 273 and 277; cf. SM,
158-63); and political theory which, in the guise of utili-
tarianism, pits the subjective will of the individual (as
desire) against the objective
________________________
16 How many types is that? Clearly more than Collingwood
enumerates in Speculum Mentis, since the abstraction that he
embraced in later years is not one of those considered among the
abstractive processes of the four sub-philosophical forms of
experience. See Chapter VIII, below. We leave it to the reader
to judge if that form of abstraction also qualifies as a
dogmatism.
147
will of society (as law), so that "in such a society one regards
every one else as a means to his own ends"--a society in which,
in other words, all goods are regarded as private (SM, 169-76,
221-31).
Nor is this anti-realism peculiar to Speculum Mentis
alone; many of these themes appear in different guises in ar-
ticles which he published during this period, and some of them
are brought out even more forcefully in later publications. It
is also instructive to pay attention to the synonyms that
Collingwood uses for realism. Thus what is called the "plain
man's metaphysic" in Religion and Philosophy, and "realism" in
Speculum Mentis, is called "the plain man's realism" defined as
"to think of the object as a 'thing in itself,' a thing existing
in and by itself" in a paper of 1928 on "The Limits of
Historical Knowledge" (EPH, 99). In a 1921 paper on "Croce's
Philosophy of History," Collingwood criticizes Croce for his
"vacillation between naturalism, for which some statements are
just true and other just false, and idealism, for which truth
and falsehood are inextricably united in every judgement" (EPH,
12)--the former representing the Croce who is "the realist,
dualist, empiricist, or naturalist, who delights in formal
distinctions and habitually works in dualistic or transcendent
terms"(EPH, 8). In passages such as these it is clear that
Collingwood tends to use interchangeably terms like realism,
naturalism, dualism, and empiricism.

148
This synonymous use of terms is further exemplified in a
1923 paper on "Sensation and Thought" in which Collingwood
attacked what he called "the empiricist fallacy" of divorcing
sensuous appearance from objective reality, or assuming "that a
distinction could be made between what a thing looks like and
what it is"l7--a distinction which is taken as "the infallible
mark of dogmatism (and consequently of realism) in all its
varieties" in Speculum Mentis (SM, 255; cf. SM, 77). Here we
have empiricism and naturalism described in terms identical with
the epistemological doctrine of realism. In a 1925 paper on
"Economics as a Philosophical Science" Collingwood reinforces
this identification of realism and empiricism by defining
"empirical thought" as "that which conceives its object as
substance or thing," as opposed to philosophical thought, which
conceives its object as activity.18 The distinction is virtually
repeated in "Political Action," a paper delivered before the
Aristotelian Society in 1928.19
6. Conclusion.
To say that "realism" as described in Speculum Mentis is
a protean monstrosity would hardly be an exaggeration, and one
cannot help but
____________________
17 Proceedings of the Aristotelian Society, XXIV (1923-
1924), p. 66.
18 International Journal of Ethics, XXXVI (1925), p. 162.
19 Proceedings of the Aristotelian Society, XXIX (1928-

1929), p. 155.
149
wonder why any person of sound mind and civilized demeanor would
ever willingly describe himself as a realist. Are we dealing
with something which has any historically evidenced counterpart,
or is Collingwood presenting us with caricature as a foil for
his protagonist, absolute idealism? While we can certainly say
that at this point that the autobiographical interpretation on
the issue of realism is well supported by the evidence we have
been examining,20 we are left with a number of troubling
problems which we shall have to monitor closely in the chapters
which follow.
_________________
20 Two themes from his discussion of realism in the
Autobiography have not been discussed at this point: the non--
eternity of philosophical problems and the falsity of realistic
logic with its claim to one-one correspondence of propositions
with indicative sentences. We shall be examining Collingwood's
views on logic in the next chapter. Concerning the eternity of
problems, the closest one comes to a statement of his position
at this time is in a 1927 paper on "The Theory of Historical
Cycles" (EPH, 76-89), which argues that in one sense it is true
to say the the problem of politics is always the same, but in an
equally important sense it is always different. The abstract
goal of providing for the needs and betterment of a society
remains the same, but in each case this involves solving
different concrete problems (EPH, 85-87). The reader of the
quotations of the last few pages would not have much difficulty
in constructing an argument to overcome this difficulty: the
problems that philosophers are concerned with are eternal--the
same questions from generation to generation--only insofar as
they are initially taken as abstracted from the historical
situations in which they arise. The refusal to see these
questions and their historical context as related is a willful
dogmatism, and a species of the more fundamental error of
separating subject (the philosopher) and object (the problems he
considers). From an abstract point of view, then, there are
eternal questions and concepts--and such is the point of view of
the realist; from a concrete point of view, there are not--and
such is the viewpoint of the philosopher as historian.
150
The first of these is the issue we just raised. If there
are distinctions which Collingwood is willing to make within the
ranks of those who have been called idealists, so that when
Collingwood denies that Speculum Mentis is idealistic he can
remain in good faith with his readers by assuming that they
understand him to mean that he was not a subjective idealist,
are there no such equivalent distinctions to be found among the
ranks of the realists? While Collingwood tells his readers that
there are as many forms of error as there are forms of realism,
he fails to tell the reader what those forms might be. We shall
find in succeeding chapters that this issue remains unresolved,
and that the "realists" remain not only a shadowy group of
figures warming themselves by the bare fire of their negative
thesis, but also that the position being rejected becomes
progressively more indistinct as his polemic against it
increases in its fury.
What is clear from this polemic is that in discussing
realism Collingwood himself always employs an abstract sense of
"knowing" and "object" because he was dealing with the problem
from what he considered to be the realist's own perspective,
i.e. one which takes the perception of a physical, extra-mental
object as paradigmatic for all forms of knowing.21 This means
that by "realism" Collingwood had in mind
_____________________
21 Passmore writes that "British philosophy, preoccupied
with the theory of perception, tends to classify philosophical
theories by their attitude to the perception of material things:
'realism,' for it, is the view that material things exist even
when they are not being perceived,
151
the epistemological viewpoint which looks to empiricism as its
natural point of departure. That there might be non-empiricist
realists Collingwood never appears to have considered as is
clear from the fact that even in his criticism of a realist he
admired--Samuel Alexander--his main objection is to the fact
that he adopts an acquaintance theory of knowledge (EM, 17678).
His strategy is therefore to attack any sensation-bound theory
of knowledge which neglects the active role of thought in
perception, and to do so by emphasizing the contextual and
interpretative aspects of the perceptual process. In Chapter VII
we shall find Collingwood arguing against a different op-
ponent--G. E. Moore--and adopting a similar strategy, but one
which argues that the expression "sense datum" is intrinsically
absurd. But in both earlier and later writings, his point of
departure is the abstract statement of the realists that
"knowing makes no difference to the object known," where
"knowing" means perception, and "object known" means object of
perception.
Secondly, what about the connection between realism and
abstraction? Obviously an empiricist epistemology must account
for the existence of universal concepts and many do so by
providing a theory
_________________________
and 'idealism' is, most commonly, the view that they exist only
as objects of perception." Op. cit., p. 49, note.
152
which describes universals as arising by a process of
abstraction from particular sensible instances. In Speculum
Mentis Collingwood does not enter into a criticism of any such
abstractive process, nor is it the basis for attack on
"abstraction" as the root error of realism and all forms of
dogmatism. Speculum Mentis, adopting the stance of absolute
idealism, assumes that absolute truth resides only in the whole,
and "abstraction" is whatever divides this whole into atomic
parts. Just as Collingwood did not seem to consider the
possibility of a non-empiricist realism, in Speculum Mentis he
did not appear to consider a non-abstractive realism. But we
found him in at least one passage arguing against realism in a
way which does not entail any commitment to the principle of
abstraction (SM, 267-68), at least not in all senses of the
term. In Chapter VIII we shall find Collingwood proposing a
peculiar description of abstraction that does not involve
"separating" what is abstracted from its abstracting context,
and at that point we shall have to assess whether such a process
can be maintained within an anti-realistic framework. The
statements "all abstraction is falsification" and "abstraction
is necessary for all true judgements" may not be contradictory
when the appropriate interpreting qualifiers are added--"simple
abstraction" in the first assertion being semantically
discernible from concrete or "real abstraction" in the second.
But more of this in Chapter VIII.

153
Thirdly, we wish to point to a hidden presupposition in
these early discussions of realism and idealism, but one which
Collingwood stated early on in his career, and never abandoned.
In Religion and Philosophy Collingwood had already decided that
there was no such thing as a fixed human nature. This appears to
be a corollary of his principle that the esse of mind is de hac
re cogitare. "The question to be asked about mind," writes
Collingwood in 1916, "is not what it is, but what it does; a
question which the logic of things and qualities does not deal"
(RP, 165; FR, 266)--and herein lies an enormous part of the
program for the remainder of his philosophy. Not only does it
entail placing Collingwood on a collision course with empirical
psychology, which until the end of his life he criticized for
treating mind as if it were a thing, and acts of thought as if
they were events not significantly different from those of the
physical world; but it required him to formulate a logic
alternative to that of "things and qualities" in order to have
an instrument for dealing with what mind does without doing
violence to it in the process. In the next chapter we shall look
at Collingwood's early attempts to satisfy this requirement. We
can anticipate that, just as he found in the philosophy of
absolute idealism a kindred spirit for his anti-realistic
leanings, so he would also find in their logic an alternative to
the formal logic he presumed to be the tool of the realist
philosophers.
CHAPTER IV
QUESTIONS AND ANSWERS, LOGIC AND DIALECTIC
1. Introduction.
In the Autobiography Collingwood states that his Q-A
logic was not explicitly formulated until he was obliged to
confront the Albert Memorial on a daily basis during World War
I; but by 1917 it was, since he states that it was at that time
that he wrote a book called Truth and Contradiction, which was
refused by a publisher. After writing the Autobiography
Collingwood deliberately destroyed it (A, 29-30, 42, 99 n. 1).
In Chapter I we noted that Knox found it incredible that
Collingwood could have worked out his theory of absolute
presuppositions (which the Autobiography describes as part of
Q-A logic) prior to writing the Essay on Philosophical Method in
1932 (IH, x-xi). Since the Autobiography is not clear on what
exactly is included in the version of Q-A logic presented in
Truth and Contradiction, we have questioned Knox's judgment on
this issue. In this chapter we must try to find whatever
evidence we can in the writings prior to 1932 of the role of
questions and their presuppositions in the logical functions of
mental acts, and on the basis of this evidence to decide if
Collingwood's autobiographical interpretation can be
154
155
upheld on this point, or if he was deliberately trying to de-
ceive his readers (as one might suspect from his destruction of
the manuscript of Truth and Contradiction, which could be
interpreted as an attempt to cover his own tracks).
At this level of inquiry our task is plain: we are to
take up the early publications, including books, essays, and
published lectures, and examine them for statements concerning
questions, answers, and presuppositions. Unfortunately this task
is already complicated by the overlay of issues from the
previous chapter--especially the problems which arise due to his
early commitment to absolute idealism. For while Collingwood
disclaimed originality in rejecting propositional logic and the
propositional, correspondence, coherence, and pragmatic theories
of truth (A, 36), he appears to claim credit for recognizing
that truth is a property of the Q-A complex and not of
propositions as such (A, 38). Furthermore he clearly wants his
readers to believe that the alternative to formal logic was his
own Q-A logic, and that Q-A logic allowed him to answer his 1914
question about whether the realists' methods were sound: the
answer was that they were not, because "the 'realists'' chief
and only method was to analyse the position criticized into
various propositions, and detect contradictions between these,"
following as they did the rules of propositional logic (A, 42).

156
But if Collingwood could have been so historically
careless as to have forgotten, as other students of Oxford
realism have not, that it was Cook Wilson himself who was highly
critical of propositional logic, and who insisted that what the
actual subject or predicate of a statement was depended on what
question the statement is answering, 1 could he also have been
mistaken about other details concerning his "discovery" of Q-A
logic? What is particularly puzzling is Collingwood's failure in
the Autobiography to mention anything at all about dialectical
logic, and it is this oversight--or was it deliberate
neglect?--that is brought into focus by the issues raised in the
previous chapter. While it is true that in Religion and
Philosophy Collingwood employs a method of argument that relies
more on what he claimed in the Autobiography to be that of the
realists (i.e. analysis into contradictory propositions),
Speculum Mentis is beyond any reasonable shadow of a doubt built
upon the dialectical
___________________
1 "A point of particular importance . . . is Cook -
Wilson's criticism of the subject-predicate logic. First of all,
he sharply distinguishes between the Grammatical subject and the
logical subject, which the traditional logic is content . . . to
identify . . . . Everything depends upon what question ((a))
statement is answering . . . . (S)tress and context are ignored
by the traditional logic; thus there arises what Cook Wilson
regards as the absurd presumption that the noun which is
nominative to the principal verb in a statement is bound to
indicate the logical subject" (John Passmore, A Hundred Years of
Philosophy (Baltimore, 1957/1968), p. 244). Passmore also notes
that Cook Wilson criticized Bradley's dialectical method for
asking "unreal" questions, i.e. questions which cannot
intelligibly arise (op. cit., p. 246).
157
logic familiar to post-Kantian idealist philosophers. If
Collingwood was an anti-realist primarily because he became an
absolute idealist, then it is understandable that he would also
be opposed to a logic of atomic propositions, each externally
related to an equally atomic and distinct state of affairs in
the physical world, and that he would offer in its place a
dialectical logic of developmental processes more in keeping
with the active role of thought in reconciling oppositions. But
this is not the genesis of his thought as he outlined it in the
Autobiography, where the turn away from realism is described
without any mention of absolute idealism or dialectical logic.
In fact it is astonishing to find that there is no mention of
any of the Italian idealists in the Autobiography other than one
brief reference to Guido de Ruggiero as the recipient of a copy
of his manuscript, Libellus de Generatione, which outlined the
logic of historical process (A, 99). His description of this
process is also the closest that Collingwood comes in the
Autobiography to discussing dialectical logic (the term is not
used). In fact in reading the Autobiography one is inclined to
believe that while Collingwood read widely in his youth, in his
maturity he was philosophically influenced only by indigenous
British philosophers.
We already know this not to be the case. But if we are
to understand the Autobiography as an act of self-inter-
pretation, our concern

158
is less with such oversights as these--whatever their reason
might be 2 --than with the extent to which the account he gives
can make sense of his published writings, and it is to these
that we now turn without further prologue.
2. Abstract and Concrete Universals.
As one might expect, there is no mention of the ques-
tioning activity in Religion and Philosophy, nor of the Q-A
complex as the unity of knowledge. But there are scattered
remarks about logic, the universal, abstraction, and scientific
thinking which provide us a few clues about the way in which he
conceived these subjects prior to his turn to absolute idealism.
In the first chapter Collingwood tries to analyze what would be
meant by a philosophy of religion, and in the process outlines
what a "philosophy of" anything means:
The philosophy of any subject means careful reflexion upon

that subject; thus we have the philosophy of art, of
conduct, of science
_______________________
2 One such reason could very well be the political and
military polarizations which were occurring at this time in
Europe. In May of 1936 Italian forces entered the Abyssinian
capital of Addis Ababa, and in July of 1936 the Spanish Civil
War began. The Preface to Collingwood's Autobiography is dated 2
October 1938. In May of 1939 Germany and Italy signed the "Pact
of Steel." From June through August of 1939, Collingwood sailed
to Greece and Italy as First Mate of the schooner yacht, Fleur
de Lys. In The First Mate's Log he writes with outrage and
contempt of an incident with an Italian fascist harbor patrol in
Messina (FML, 170-74), and at the end of his account of the
voyage he describes the discovery by the crew, mostly from
Oxford, of the German-Russian alliance and the Nazi invasion of
Poland. In such circumstances it is understandable why
Collingwood might not wish to make an issue of expressing his
indebtedness to German and Italian idealism.
159
and so on. To do a thing, and to understand what one is

doing and how one does it, seem to be different
things; . . . to conduct an argument is science, to reflect
upon it is logic . . . . But the theory of knowledge or
logic does not consider differences of the object, but only
processes of the subject; and therefore there is no
distinction between the philosophy of religion (as theory of
religious knowledge) and the theory of knowledge in general.
If there is a general philosophy of knowing, it includes
religious knowledge as well as all other kinds; no separate
philosophy is required. (RP, 15; FR, 53).
As Collingwood's thought developed, nearly every statement in
this passage is modified to the point of contradiction, with the
exception of the definition of logic as a "theory of knowedge."3
By the time he came to write Speculum Mentis he had already
abandoned the idea of a "general philosophy of knowing"--as if
knowledge were a genus and cases of it were particular species;
and in the same work he denies that a theory of knowledge does
not consider differences of the object, but only "processes of
the subject." In fact even in Religion and Philosophy
Collingwood could
________________
3 In his chapter on Cook Wilson and Oxford philosophy,
Passmore writes: "Cook Wilson's main theme is logic, but logic
conceived in the Oxford manner, as a philosophical investigation
into thought rather than as the construction of a calculus. The
Boole-Schroder logic, indeed, Cook Wilson condemned as 'merely
trivial,' in comparison with 'the serious business of logic
proper '--inquiry into 'the forms of thought"' (op. cit., p.
240). It must be recalled that Collingwood learned his logic
from Cook Wilson, and never seems to have accepted any other
view than that the true task of logic is to understand the
"forms of thought"--cf. Chapter VIII, below. As Donagan rightly
notes, Collingwood completely failed to appreciate the
revolution in logic occurring during his lifetime.
160
not consistently maintain that logic considers only processes of
the subject, since in this work we have already found him
attacking psychology for treating a judgment (the act of knowing
something) as an event in the mind without going on to determine
the relation of this mental event to the reality beyond the act
which is being apprehended (RP, 40; FR, 75-76).
A more promising and enduring starting point is with his
remarks about the universal-particular distinction. In analyzing
the relationship of philosophy and history (as part of his
efforts to relate religion and history), Collingwood examines
the attempt to distinguish history and philosophy on the grounds
that they deal respectively with the particular and the
universal:
History, it is sometimes said, is knowledge of the par-

ticular, philosophy knowledge of the universal. But the
particular is no mere particular; it is a particular of this
or that universal; and the universal never can exist at all
except in the form of this or that particular. "The
universal" and "the particular" considered as separate
concrete things are fictions; and to equate the distinction
of philosophy and history with such a fictitious distinction
is to admit at once that it is untenable. (RP, 49-50; FR,
83).
Later Collingwood settles in much the same way the suggestion
that the Incarnation can be interpreted by means of the same
distinction--God as the universal and man as the particular. To
regard the universal as if it were something separate and
concrete is the result of a logic gone awry.

161
(T)he universal itself, which as a matter of fact exists

only in various particulars, is sometimes falsely conceived
as if it were itself another particular; and thus arises the
notion of an archetype or ideal specimen of a class to which
every less perfect member is an approximation. These ((are))
two tendencies of false logic, the tendency to elevate one
particular into the standard and only real instance of a
universal, and the tendency to hypostasize the universal
into a perfect and ideal particular . . . . (RP, 163).
What Collingwood suggests as an alternative to the
separate universal of false logic is what he calls a concrete
identity in Religion and Philosophy and the concrete universal
in Speculum Mentis. We shall deal with the latter in the next
section, but it will be helpful to have a provisional idea of
what he means by these terms. Evidently a universal refers to a
unity of some sort, and especially a unity which is capable of
being shared by two minds thinking about the same thing--such is
the minimal sense of "universal" at least since the time of
Socrates and Plato. When Collingwood asks how it is possible for
two minds to think the same thought, he begins by assuming the
factual existence of communication and knowledge (RP, 98, 109;
FR, 170, 180) and, like Kant, asks how this is possible. Since
all consciousness is the consciousness of something definite, it
follows that if one is thinking of anything at all it must be a
thought of something concrete; Collingwood goes so far as to say
that "One simply cannot make general statements without any
thought of their instances" (RP, 46; FR, 81). Two minds share
the same thought when that thought is a concrete identity, i.e.
one which has the characteristics of

162
a whole in each of whose parts the whole is entirely present
(RP, 88-89; cf. RP, 108, 112; FR, 179, 182). Collingwood
contrasts this concrete unity (manifested in personal identity)
with the abstract unity--i.e. what a thing is in itself as
opposed to what it is in relation to its context or the whole of
which it is a part; he argues that the character or self of a
thing, what it is, cannot be distinguished from its relations,
which consist in a quality of the thing itself (RP, 110-12; FR,
181-82).
Without going further into this discussion (we shall
take it up again in more detail in Chapter VI on rapprochement
identity), we can see already a drift in the direction of
idealism of the sort we discussed in Chapter III. It is also
transparent that Collingwood is making use of distinctions that
were known to anyone familiar with the logic of Bradley.4 The
contrast is between the abstract and concrete universals, and
while it is somewhat vague and imbedded in discussion of other
topics in Religion and Philosophy, it is explicit and prominent
in Speculum Mentis. But so is, we must add, the role of
questioning in knowledge.
In Religion and Philosophy Collingwood seems anxious to
take seriously the realists' principle "that the mind is one
thing and the
_____________________
4 F.H. Bradley, The Principles of Logic (Oxford, 1883),
vol. I, p. 188; cf. Bosanquet, The Principles of Individuality
and Value (London, 1927), pp. 35-39.
163
object another," although he immediately adds that "we cannot
rest content with the statement" (RP, 99; FR, 172). The theory
of knowledge contained in this early work seems to be an attempt
at a compromise between correspondence and coherence theories of
truth with the hope expressed that it would offend neither
realists nor idealists. On the one hand mental acts (judgments)
are defined and determined by their reference to objects (RP,
99-102; FR, 17173); on the other hand the "object" toward which
mental acts are directed turns out to be an
identity-in-difference (a concrete universal) whose "inner
structure" is entirely constituted by the necessary relations it
has as part of a whole (RP, 108-14; FR, 179-84).
But by 1924 Collingwood was less willing to grant any
ground to the realist at all. In Speculum Mentis knowledge in
its irreducible and simplest state is an activity of questioning
and answering, and the attempt to identify knowledge as anything
less than this is sharply dismissed:
A crude empiricism imagines that knowledge is composed

wholly of assertion: that to know and to assert are iden-
tical . . . . Knowledge as a past fact, as something dead
and done with--knowledge by the time it gets into encyclo-
pedias and textbooks--does consist of assertion . . . . But
those who look upon it as an affair of discovery and
exploration have never fallen into that error. People who
are acquainted with knowledge at first hand have always
known that assertions are only answers to questions. (SM,
77).
164
Here we have a declaration that sounds like it could be a direct
quote from the Autobiography. But as we have already seen, there
is a great deal else in Speculum Mentis that does not. One
noteworthy difference is that there seems to be a strong
inclination to accept coherence rather than P-Q-A complexes as a
touchstone of truth: Collingwood makes it clear from the start
that the various claimants to the title of truth--his five
"forms of experience"--are to be tested on their ultimate
self-consistency, that is, the coherence they exhibit in
attempting to live up to their own claims (SM, 4445). The
"self-consistency" that Collingwood has in mind, however, is not
merely freedom from propositional contradiction:
Now the characteristic mark by which a form of experience is

shown to be satisfactory is simply that it is possible . . .
. Any scheme ((i.e. form of experience)) which is in itself
contradictory or nonsensical cannot redeem ((its)) promises,
because it cannot be put into execution; but if there is any
scheme of life which is inherently consistent and therefore,
ideally speaking, practicable, we may safely assume that
this is the scheme to adopt. Self-consistency, then, is our
test.5 (SM, 44; cf. SM, 250).
Collingwood immediately adds that any criterion of truth resting
on a presumed correspondence either with human nature or with
the facts
__________________________
5 Cf. F. H. Bradley, Appearance and Reality (Oxford,
1893), p. 120: "Is there an absolute criterion? . . . Ultimate
reality is such that it does not contradict itself; here is an
absolute criterion. And it is proved absolute by the fact that,
either in endeavoring to deny it, or even in attempting to doubt
it, we tacitly assume its validity."
165
about the world is a mistake, because both of these are
inherently doubtful: what human nature is and what the facts
about the world are, are both open to legitimate questioning
(SM, 45). So it is apparent that if the usual form of
"coherence" is not exactly what Collingwood has in mind as a
criterion of truth, "correspondence" of consciousness and object
is no longer acceptable. Nor did he appear to entertain it
seriously again for the remainder of his published career.
Speculum Mentis also in part corrects, in part develops
his earlier views on logic. As in Religion and Philosophy, what
Collingwood is seeking is a philosophical logic, by which he
apparently meant something like Kant's transcendental logic,
i.e. a logic which does not regard mental acts considered in
abstraction from their reference to objects. It must therefore
be not the formal logic of the "abstract universal" (the unity,
identity, or sameness of a concept which is indifferent to the
variation or inter-relation of its own instances), but the
dialectical logic of the "concrete universal" (an
identity-in-difference, or a unity to which difference is
essential) (SM, 162-63). Collingwood's strategy in both of these
early works is to argue that the very attitude of consciousness
that regards all concepts or universals as abstract also
prevents the latter from being identified in any meaningful way
with their objects or instances--in a word, from referring. In
Religion and
166
Philosophy Collingwood limited himself to the case of religious
consciousness, and calls such an abstract attitude with respect
to religious objects "psychology" (RP, 40; FR, 75-76) (although
he does also speak of "abstract history"--this being the "mere
verbal description of events without any attempt at
understanding them," and abstract philosophy as "the dry
criticism of formal rules of thinking without any attempt at
grasping their application"-RP, 51; FR, 85). In Speculum Mentis
such an attitude towards objects of knowledge is taken to be
characteristic of science as such (SM, 158-63). In both cases
the reification of the abstract universal, that is, making it an
object of thought "separate" from its instantiations, is taken
as characteristic of the "realistic" point of view (SM, 189,
252, 282-85). But in Speculum Mentis objects of knowledge are
not described as "real things" or "facts" independent of
anyone's knowing of them, but rather are taken as the objective
correlates of acts of knowledge (SM, 11, 159, 287, 293-95, 310).
And as a consequence the tendency to regard logic merely as a
psychological science (a tendency opposed by such diverse
thinkers as Bradley and Frege) is overcome to some extent, and
logic is allowed to have its own innings--that is, to present
its case as a justification of science as an autonomous
discipline.
But since it is in Speculum Mentis that both Q-A logic
and formal logic (as the justification of science) make their
appearance, it is
167
necessary here to assess their relationship in order to see if
Collingwood's view in the Autobiography, which presents the
former as an alternative to the latter, is reflected in his
second major work. As we shall see, the linking concept between
the two is bound up with the activity of "supposal" or
hypothesis formation, which remains somewhat ambiguous--as one
might expect, since at this stage his reconciliation philosophy
was still incompletely worked out.
3. Science and Supposal.
The form of experience which regards its objects only as
particulars of a universal, as members of a class, or as
instances of a law, Collingwood calls "science" (SM, 158-63).
Whether the abstract concept makes its appearance as a Platonic
form, a medieval universal, or a Renaissance law of nature, the
characteristic viewpoint of scientific consciousness is that it
distinguishes between universals and particulars, and assumes
that the former can be abstracted by thought --i.e., separated
or isolated and studied apart from its instances (SM, 159-60,
180). Such an attitude is variously described by Collingwood as
the abstractness of the scientific concept (SM, 162), the
principle of the transcendence of the universal to the
particular (SM, 179), or the relation of difference without
identity (SM, 243). In any event it asserts the reality of the
abstract concept as indifferent to its exemplifications or to
the mutual
168
relations its instances have other than as members of a class
(SM, 162-63). Furthermore the abstract universal of science is
contrasted with the concrete universal of history:
Classification is the key-note of the scientific spirit;

but classification is nothing but the abstractness of the
scientific concept. For a class as such is a collection of
individuals without any mutual cohesion or organization
except their common membership of the class. They have no
reference to each other, but only to the universal; and
each one refers to the universal in precisely the same way
as every other. As soon as they refer to the universal in
different ways, or, what is the same thing, as soon as they
develop a system of mutual relations between themselves,
they cease to be a mere class and become an organized and
articulated system; and the universal ceases to be an
abstract universal (class-concept) and becomes a concrete
universal, or one to which the differences between its
particulars are relevant. (SM, 162-63).
The point at which science cannot maintain its object as an
abstract universal, and is forced to assert the reality of the
concrete universal is, therefore, the point at which science
ceases to survive as an autonomous form of experience and be-
comes dependent on another form of consciousness, viz. history
(SM, 180, 186-87; cf. 193).
That point occurs when scientific consciousness attempts
to deal with "facts," or to refer to objects and events given in
sensuous appearance. In its most elementary and primitive form
science is constructed in accordance with an a priori or
deductive ideal: it is the attempt to work out the implications
of the concept of a class as such, known independently of all
experience (SM, 164). Since the relations

169
of an indeterminate plurality of abstract units (the members of
a class) to a determining unity (the class itself) is precisely
what constitutes the numerical series, mathematics is the one
and only a priori science. Mathematics "is simply the theory of
order, where order means classificatory order, structure in its
most abstract possible form" (SM, 165). Pure mathematics deals
with classes (e.g. numerical sets) whose members are themselves
classes (numbers) (SM, 169). But if it claims to be objectively
true, mathematics turns out to be an illusion: it is "the truth
about nothing," since it is "the description of the structure of
a null class"(SM, 185).6
The attempt to import some sensuous content into these
empty class-concepts gives rise to the second phase of science:
science on the empirical or inductive model--an ordered
knowledge of "facts" (SM, 177; cf. 168). But the inductive
method (observation and experiment) does not supersede a priori
deduction; on the contrary, induction itself presumes a
principle variously described as the uniformity of nature, the
law of universal causation, etc. which induction is unable to
establish by its own methods, and which rests on the principle
of uniformity itself (SM,
______________________
6 This seems to be a variant of Russell's paradox--the
class of all classes both is and is not a member of itself. For
Collingwood's further use of this paradox, see SM 169, 189, and
192. Collingwood does not, however, seem to apply it to
mathematics itself, but only to mathematics insofar as it claims
to be objectively true.
170
178-79). If it is not to presume the very principle that it
seeks to prove, empirical science is forced to alter its
conception of its own object, and this leads to the third phase
of science: science as supposal.7
Science asserts, not the actual truth, but what would be

true if something ((else)) were true which is laid down as
an hypothesis. It asserts, never that S is P. but that if
there were an S it would be P. Its procedure therefore
consists, first, in making an assumption, secondly in de-
ducing consequences of that assumption. Throughout this
process it never makes an assertion, in the sense of a
categorical judgement, at all. Its judgements are hypo-
thetical from the beginning to end. (SM, 183).8
Such a process is utterly gratuitous: without a basis in pre-
vious assertions, Collingwood insists, no hypothesis can be
framed at all, much less relevant or illuminating hypotheses
(SM, 79). Supposal without a framework of assertion is arbitrary
and meaningless; but supposal within a framework of assertion is
something more than merely hypothetical. Its
____________________
7 Collingwood, true to a long standing tradition in
British philosophy, does not provide his reader with anything
but scant clues about the historical representatives of these
movements. The third phase has taken place "almost within living
memory" in the latter part of the 19th century, and is
represented by the "critical movement" and "scientific prag-
matism" (SM, 180-82). There is no hint in Speculum Mentis that
any movement in science follows this one.
8 Collingwood casts the hypothetical assertion in

subject-predicate form but it is clear that it applies equally
to propositional functions: the procedure of "making assump-
tions" and then "deducing the consequences of these assumptions"
is appropriate for propositions, since what is assumed is a
proposition, not merely the subject of an assertion.
171
object is not only something possible (viz. the alternatives en-
visaged in the hypothesis) but something actual, and this is
beyond the reach of mere supposal (SM, 187-88, 191, 199). Unless
science is to remain a tissue of hypotheses, and fact is
therefore permanently to elude the grasp of science, it must
have a categorical basis in actual assertions. (SM, 185).
On this objective ambiguity founders the autonomy of
science. If the objects of science are pure concepts (mathe-
maticals, universals), they are true of nothing actual at all;
but if science intends its laws to apply to real objects it
cannot deal with the latter by means of purely abstract con-
cepts; and finally, if science says that its object is neither a
null class nor something strictly speaking actual, but only
something possible, it renders its conclusions arbitrary. The
issue cannot be settled from within scientific consciousness
alone, and the attempt to settle it from without gives rise to
"scientific philosophy"--that form of dogmatism which, in the
guise of formal logic and metaphysics, presents itself as a
justification of the scientific attitude (SM, 271).
Since the deduction of consequences from hypothetical
assertions is governed by the rules of formal logic, the latter
is taken (by the critical and pragmatic defenders of the third,
and presumably still current phase of science) as the
justification of the methodology of science. But mathematical or
formal logic constructs its deductions in a

172
categorical fashion: its statements are demonstrably true of
every possible object--true, that is, categorically. Insofar as
formal logic is the categorical basis for the (hypothetical)
mathematical formulation of scientific assertions, mathematical
truths themselves are taken as "true without qualification of
the entire world, actual and possible" (SM, 184-85). Unfor-
tunately this conflicts with the conclusion that mathematics had
reached about itself as the "theory of order"--viz. that it is
the description of the structure of a null class, the truth
about nothing. Formal logic of itself is not able, therefore, to
extricate third-phase science from its difficulties: laws that
are categorically true of every possible object are true of
everything in general, but of nothing in particular. Formal
logic may be able to distinguish valid from invalid inferences,
but without metaphysics as its necessary correlate, such a
vindication of the principles on which scientific thinking is
founded may only be true for thought alone, and not for thought
that is directed to an object (SM, 272-73).
(W)e must demonstrate that what we have hitherto called

logic or the theory of ((scientific)) thought is really
metaphysics or the theory of reality, and that what we have
called the laws of thought are the laws of being. But this
is precisely what we cannot do. Metaphysics is impossible
((on scientific grounds)); for its task is to vindicate the
objective validity of the ways in which we think, and if
there are any flaws in our methods of thought, these will
affect our metaphysical theory of reality and introduce into
it the very mistakes which by its help we had hoped to
eradicate. Hence the theory of being as distinct from
thinking (metaphysics) will only be the theory of thinking
as
173
distinct from being (logic) expressed in different

terminology, but subject to the same fatal weakness, namely
that just as logic can never analyze real thinking--thinking
that, going on in the logician's mind, always lies behind
((i.e. beyond)) his analysis--so metaphysics can never
analyze real being, being as it is in itself untainted by
thought. (SM, 273-74).
In short, scientific consciousness is left permanently and ir-
reparably without justification because its principle of ab-
straction renders its objects (individuals, particular matters
of fact) utterly beyond the reach of its principles (laws, ne-
cessary principles of order) (SM, 185-86, 277).
What science is left with as a justification of its own
viewpoint is a psychology of abstract consciousness which fails
for the very reasons that Collingwood had advanced in Religion
and Philosophy--viz. since it fails to take into account the
truth or falsity of the thought it examines ab extra, as an
event, it cannot justify itself as anything more than another
mental event alongside the first (rather than an explanation of
the thought it claims to be observing) (SM, 274-77). Psychology
(or, more properly, psychologism) marks the point at which
scientific consciousness fails to achieve wisdom or
self-conscious self-justification, just as history marks the
point at which science fails to achieve concrete knowledge of
fact--the difference being that history provides the
justification of the scientific attitude that psychology cannot
(SM, 186, 193). But of this we shall have more to say in the
immediately following chapter.

174
It is important to realize (both for an understanding of
the argument in Speculum Mentis and for an assessment of the
relationship of Q-A and formal logic) that when a form of
experience is show to fail in its claim to be the whole truth
about its object, it is exposed as an error, but it is not
thereby utterly discredited and completely rejected. Since it is
an error that consciousness makes about itself, that error is
overcome only when it is criticized and shown to rely on
something other than itself for its validity (SM, 244-45, 255,
288-91, 296-97). Consequently when scientific consciousness
"collapses" it does not collapse into nothing, but leaves behind
"solid assets" in the form of a pure science of mathematics and
an empirical-hypothetical science of nature, with mathematical
logic and metaphysics as their partial justification, and the
psychology of concrete mind as their absolute justification (SM,
271-72, 277-78, 280, 317; cf. "ST", 73-75). One of the "solid
assets" for Collingwood's philosophy is the status of questions
and answers in this context.
And since the purpose of this excursion through one
portion of Collingwood's "phenomenology of error"9 was to assess
the relationship
_____________________
9 Cf. SM, 289; Collingwood rejects this title as a de-
scription of what he is doing, but in at least one sense of the
term "phenomenology" --as a description of how states of
consciousness appear or manifest themselves--it is still a
correct description of what he is doing in Speculum Mentis.
Collingwood rejects the term because he claims not only to be
merely describing, but also stating the truth about the forms
of experience he is describing.
175
between Q-A logic and formal logic, it is time to make good our
previous assertion that the connecting link is an ambiguous
notion of hypothesis or supposal. The entertainment of an
hypothesis is, according to Speculum Mentis, equivalent to the
asking of a question:
Science is explicitly supposal. But supposal itself . . . is

identical with questioning, which is the cutting edge of the
mind, an activity not self-contained or independent, but
implying behind it a body of information or
assertion . . . . But it is the facts that are true; the
scientific simplification of them into instances of laws,
abstract particulars of abstract concepts, is not true but
arbitrary, useful no doubt, but useful precisely because it
is not asserted as true but merely entertained in the form
of a question. (SM, 186).
But the entertainment of an hypothesis in the form of a question
turns out to be a complex mental function, since hypothesis
formation is identified not only with questioning but also with
"intuition":
The paradox of science may be expressed by calling it in-

tuitive thought. Intuition is the questioning, immediate
side of experience: thought is the asserting, explanatory
side. Science is explicit to itself as thought, but it turns
out on inquiry to be identical with the questioning
activity; that is, it realizes the contradiction of a type
of thought which is not thought precisely because it is
thought's opposite, intuition. (SM, 188).
Collingwood is quick to add that the division of experience into
intuition and thought is an abstract fiction, and represents
only a distinction between two sides of the indivisible whole of
experience--"an immediate, intuitive, or questioning side" and a
"mediating, reflective, logical or

176
assertive side," the former being called sensation and the
latter, thought (SM, 95, 188; cf. "ST", 57-58). But the willful
enforcement of this distinction by scientific consciousness is
another manifestation of its self-contradictory or paradoxical
nature:
It is this falling-back upon intuition that constitutes the

irrationality, the arbitrariness of all science. The
assumptions made by science cannot be justified under
criticism; their only justification is the frankly irra-
tional fiat of the scientist's will. The concept is for him
an abstraction, that is to say hypostasized into a thing,
reified; hence it cannot be explained by thought, it can
only be intuited, and this intuitive attitude towards a
concept is what is meant by assumption in science. (SM,
189).
It is at this point that the ambiguity of hypothesis or
assumption becomes apparent--a ghost that not only remains
unexercised in the body of science, but which survives to haunt
Collingwood's Q-A logic. If assertion is the minimum activity
claiming truth or falsity (SM, 59-60), then (with respect to
assertion) hypothesizing, questioning, and intuition have one
and the function: each is a suspension of the activity of
asserting (SM, 78-79, 186, 188-89). But there are important
differences which such an identification overlooks, and which
are crucial not only for formal logic but for a Q-A logic which
seeks to replace it.
(1) Questions are not merely non-assertions, they are
proto-assertions -- non-assertions about to become assertions.

177
It may be true that "supposal and questioning are at bottom the
same thing" (insofar as asking a question, like framing an
hypothesis, means contemplating the non-existent in the form of
several alternatives, only one of which may be existent), but
"true questioning is a suspension ((of assertion)) which looks
forward to a renewal of this asserting activity, in the shape of
an answer" (SM, 78-79). "Hypothesis" as Collingwood uses the
term in Speculum Mentis vacillates between these two
senses--i.e. proto-assertion and non-assertion-and it is not
until the Essay on Metaphysics that the distinction is clearly
made between questions, assertions, suppositions, and
presuppositions (EM, 21-33).
(2) In Speculum Mentis meaningful questions do have a
bi-directionality, looking both forward to an answer and back-
ward to other assertions (SM, 79), but there is no clear re-
cognition that the assertions which ground questions may be
non-factual but yet meaningful--i.e. function as what he was
later to call "presuppositions." Instead Collingwood leaves the
reader to decide whether the categorical assertions, which are
logically prior to hypotheticals (SM, 183: cf. "ST", 64), are
necessarily or only factually true. Thus, for example, within
two pages he speaks about the mind "categorically asserting a
concrete fact"
178
(when it sets about framing hypotheses),l0 and about
mathematical logic being "categorically true of every possible
object" (SM, 184-85). These are the factual and the
inferentially necessary senses, respectively, of the term. It is
not until the Essay on Philosophical Method that Collingwood
would try again to say what he means by a "categorical
assertion" (EPM, 117-36).
(3) Underlying both of these ambiguities is a basic
equivocation on the use of the phrase, "mental activity," which
leaves mental dispositions (conscious attitudes involved in
raising questions, framing hypotheses, asserting a proposition,
etc.) undistinguished from logical functions (formal properties
of interrogative sentences, hypothetical assertions, categorical
propositions, etc.)ll This leads to certain
_______________________
10 Cf. EPH, 45-46 (1924): "(S)cientific thinking is an
abstract thinking, historical thinking a concrete thinking. In
other words, because the object of science is not a fact but an
abstract type or form, the judgement of science is always
hypothetical: 'if A, then B,' where it is not asserted that A
exists in the world of fact . . . . Whereas the object of
history is the fact in all its actuality, and therefore the
historical judgement is categorical . . . . The ideal of his-
tory, then, is to be a single categorical judgement, articulated
into an infinity of coherent categorical judgements, asserting
the reality and expounding the nature of an infinite individual
world of fact articulated into an infinity of individual facts.
11 The closest Collingwood comes in Speculum Mentis to

recognizing the difference is at SM, 79, where he speaks about
the "empty form of questioning," questions which ask nothing,
mere "marks of interrogation." But he is not here explicitly
distinguishing between mental dispositions and logical func-
tions, but rather two sorts of mental activities, one with and
one without a background of factual assertion.
179
puzzling and unresolved contradictions which mask rather than
illustrate Collingwood's thought in Speculum Mentis, and provide
obstacles to the success of his Q-A logic. (a) On the one hand
questioning is identified with explanation and intellection. The
characteristic mark of the former processes is their "immediacy"
(i.e. their spontaneous actuality, their positivity, or in
general their lack of dependence on something other than
themselves) (SM, 95, 188). Questions are thus immediate, and
answers mediate. (b) On the other hand questions are identified
with the logical function of assertion cast in the hypothetical
mode, while answers are identified with the corresponding
categorical function (SM, 183, 186). In this sense questions are
mediate (their truth depends on something further being
asserted), and answers are immediate (they actually assert
something positive--something true or false). It is not until
Collingwood wrote the chapter on language in the Principles of
Art that logical and psychological functions are distinguished
(PA, 225-69).
Collingwood does not seem to be alarmed at these con-
clusions in Speculum Mentis, since he presumes that such argu-
ments demonstrate that questions and answers are not independent
abstractions, but are both mental activities, and therefore have
the characteristic marks of all

180
mental activity--viz. mediacy and immediacy (cf. SM, 80).12 But
when Collingwood calls the distinction between question and
answer an "ideal distinction," and adds that "the process of
knowledge is . . . not so much an alternation of question and
answer as a perpetual restatement of the question, which is
identical with a perpetual revision of the answer," the
_____________________
l2 Cf. SM, 188-89, where parallel arguments concerning
mental and logical functions are put forward to illustrate the
irrationality of the abstract distinctions made by science: (a)
"Intuition and thought are not two separate activities which are
somehow united in the body of human experience. Experience is an
individual whole in which two sides can always be distinguished:
an immediate, intuitive or questioning side, which is
hypostatized in abstract psychology into the faculty of
sensation, and a mediating, reflective, logical side, which is
called thought. Thought is the one, sensation the many. What
characterizes the intuitive or sensuous side of experience is
just its manyness or perpetual difference from itself, flux,
novelty, or creation. What characterizes the logical or re-
flective side is its self-identity, permanence, unity. Now we
have already seen that science consists in the separation of
these two distinct elements, and the attribution of reality to
thought while denying it to sensation. But division as such is
the characteristic of sensation as opposed to thought: thought
unifies what sensation divides. Therefore any given thing which
is made the field of an unreconciled division is thereby placed
under the head of sensation, for the characteristic unity of
thought has been denied to it. If experience as a whole is now
divided into two separate parts, thought and sense, it becomes
by this very definition wholly sensuous, and each part of it is
a sensuous, not an intelligible, object." (SM, 188-89). (b)
"This argument is more familiar. though more superficial, when
stated in terms of logic. The universal has its very life and
being in its particulars, of whose multiplicity it is the unity.
If now it is disentangled from those particulars and set apart
by itself, it becomes not their universal but another particular
object, thus losing precisely its intelligibility (universality)
and becoming an object of mere intuition, a thing that we no
longer think but only imagine. It is this falling-back upon
intuition that constitutes the irrationality, the arbitrariness
of all science." (SM, 189).
181
ensuing “identity" which "contains all diversity within itself"
(SM, 80) appears dangerously similar to that absolute night in
which all cows are black. The result, in short, is as
unfortunate for his Q-A logic as it is for formal logic: if
truth and falsity are functions of the Q-A complex and there is
no adequate way to distinguish within that complex what counts
for a question and what counts for an answer, then it seems
there may be no adequate criteria for distinguishing between
truth and falsehood either. It is not until the early chapters
of the Essay on Metaphysics that he attempted to disentangle Q-A
complexes and truth criteria (EM, 21-48).
4. Conclusion: Three Logics.
From the evidence that we have just examined it does not
appear that there is any systematic "question-and-answer logic"
as such mentioned in any of Collingwood's writings through 1924,
and a glance at the articles from this date until 1932 does not
yield any significant indication that would change this
judgement. Thus we find Collingwood writing in "The Nature and
Aims of a Philosophy of History" (1925) that the way in which a
problem arises for an historian must convey some hint of the
direction in which evidence for its solution is to be sought,
and that in doing so he argues to and from this evidence, so
that there is in the last analysis no distinction between his
sources and his conclusions (EPH, 5253). While this gives the
observant reader of the Autobiography

182
some confirmation of the use of Collingwood's Q-A principle that
in arguing back from a propositional answer (the conclusion) to
its question the same piece of evidence states the answer and
allows the historian to identify the question, it does not
explicitly suggest that this was a conscious application of a
systematic Q-A logic. And again Collingwood writes in "The
Philosophy of History" (1930) that
a question must be asked with some reasonable expectation

of being able to answer it, and to answer it by ((genuine))
thinking; otherwise it leads nowhere, it is at best idle
"wondering" . . . . We express this by saying that a ques-
tion does or does not "arise." To say that a question
arises is to say that it has a logical connexion with our
previous thoughts, that we have a reason for asking it and
are not moved by mere capricious curiosity. (EPH, 137).
We see at work another of the principles of Q-A logic cited in
the Autobiography, but without any explicit mention of presup-
positions and their logical efficacy in causing such questions
to arise. In fact the omission of any explicit discussion of
presuppositions or presupposing in these early writings renders
suspicious the autobiographical suggestion that he had
formulated Q-A logic as such (including the theory of presup-
position) in the years between 1916 and 1918 when he returned to
Oxford. And once again we must add that this suggestion is
rendered even more suspicious by his failure to mention anything
about dialectical logic and his early absolute idealism.
But perhaps we are allowing ourselves to be misled by
Collingwood's use of that vague and sometimes all-inclusive

183
word, "logic," as well as by his tendency in the Autobiography
to offer Q-A logic as an actual replacement or substitute for
formal logic. In Chapter VIII we shall take a hard look at this
claim and try to settle the issue of whether and to what extent
Collingwood's Q-A logic is "logic" in this sense at all. For now
we can only make several tentative statements based on the
evidence provided in the writings we have been examining. We
note, first of all, that we have three labels to attend to, if
not three logics: formal logic (F-logic), dialectical logic
(D-logic), and Q-A logic. Collingwood's conception of logic
tends to view it in epistemological terms, but he does allow
that logic in general is concerned with the justification of
some form of knowledge. If it is to be a truly philosophical
logic, it must justify not only the conclusions reached by a
body of knowledge, but the way in which it reaches these
conclusions. The way in which conclusions are reached is by a
systematic question-and-answer process, the success or failure
of which is not measured by any external criteria, although in
empirical science, at least, the conclusions arrived at can be
shown to be free of contradiction by formal logic. But a body of
knowledge must not only be free from contradiction; it must also
be able to demonstrate that in its process of development the
methods that it employs for the discovery and verification of
these assertions do not conflict with the ideals which this
knowledge sets for itself. This is the role of dialectical
logic. It would therefore

184
appear that Q-A, F-, and D-logics are ordered in a developing
series such that Q-A and F-logics are opposed and complementary
phases by which knowledge comes to be (Q-A logic), and retains
itself in being by resisting criticism (F-logic), but are both
preliminary to the dialectical measurement of explicit
performance against implicit promise (D-logic). We hasten to add
that this suggestion of the relationship between the three
logics that appear in Collingwood's writings was never made
explicit by him, and is our own reconstruction of their apparent
relationship at this point in our investigation.
Before passing on to the next step of our survey of the
early works, we must make some final comments about D-logic. At
least one of Collingwood's interpreters, whose views this author
greatly respects, has stated that Religion and Philosophy is
Collingwood's only non-dialectical book, and that his moment of
kairos came between his first and second books, because Speculum
Mentis is a dialectical essay where its predecessor is not
(Mink, MHD, 20, 242). To this we can only comment that if
Collingwood discovered dialectic, it could not have come as much
of a surprise. For on his own principles philosophy does not
bring us to know things of which we are simply ignorant, but
brings us to know in a different way things which we already
knew in some way (EPM, 161) Dialectic could therefore only be
something of which he was already aware, so that when it became
explicit it was only a refinement of a reflection on an

185
experience with which he was already familiar. We therefore
cannot accept any interpretation of Collingwood which presumes
that there was any such "event" in the intellectual development
of his philosophy, and especially when no such event is
acknowledged in the Autobiography.
But the evidence of Collingwood's writings forces us to
admit that not only does the explicit use of dialectical
concepts, judgments, and inferences show an order of develop-
ment, but also the use of the term dialectic has a peek-a-boo
career in Collingwood's published writings. So far as this
development is concerned, Religion and Philosophy presents us
with the spectacle of an analytic of concepts and principles,
but does display a propensity for expressing first one side of
an issue and then the opposing viewpoint, and only does so in
order to resolve the issue by showing how the two opposing
viewpoints can be reconciled. Furthermore the discussion of
"concrete identity" (which evolves quite naturally into that of
the concrete universal of Speculum Mentis) is prominent in
Religion and Philosophy, and is opposed to the abstract univer-
sal of "false logic." In Speculum Mentis the concrete universal
emerges as an "identity-in-difference" whose characteristics are
those of a scale of forms, but one which lacks the unity of both
differences of kind and differences of degree that Collingwood
later specified for such a scale

186
in the Essay on_Philosophical Method (SM, 208; EPM, 54-57).
Therefore from Religion And Philosophy to Speculum Mentis to the
Essay on Philosophical Method there is a continuous development
of the idea of defining something not by subsumption under a
universal and abstract class, but by showing how its meaning in-
cludes others which are related by opposition and distinction,
embodying differences of both degree and kind, and where the
variable is identified with the generic essence.
But on the other hand the term dialectic appears and
disappears in Collingwood's writings, sometimes in a most
confusing manner. While the term is not mentioned in Religion
and Philosophy, it is over-used in Speculum Mentis, mentioned
with modesty in the Essay on Philosophical Method (e.g. EPM, 12,
210), disappears altogether in the Autobiography and the Essay
on Metaphysics but re-appears in more classical costume
______________________
13 The reader must be patient with the use of such jargon
at this stage of our investigation--the terms will be explained
in future chapters. But it is well to bear in mind that (1)
where Religion and Philosophy argues that where two terms are
"not different" they are therefore identical, and (2) Speculum
Mentis argues that two terms in a dialectical series are related
to one another in kind only, and not in degree, (3) it is only
after the Essay on Philosophical Method that terms in a scale of
forms are declared to differ both in degree and in kind from one
another. Therefore if Religion and Philosophy is a
non-dialectical book, so is Speculum Mentis--i.e. on the
standards of dialectic set in the Essay on Philosophical Method.
14 There is only one reference to dialectic in the Essay

on Metaphysics, and that is in connection with Hegel's use of
it, which is not treated in a flattering manner (EM, 318).
187
in The New Leviathan.15 Nor can one simply relate the sudden
prominent usage of the term in Speculum Mentis either with his
recognition of its systematic implications (since such a
description is present in Religion and Philosophy in the
analysis of "concrete identity") or with the properly
philosophical use of the implicit-explicit distinction (since
this is not a distinction reserved for philosophical dialectic:
at SM, 93 it is applied to psycho-analysis in the Freudian
sense--the interpretation of dreams is not the bringing into
consciousness of that which was unconscious, but the bringing
into explicitness that which was implicit).
A more promising direction for uncovering Collingwood's
mature understanding of the term is in the posthumous publica-
tion, The Idea of History. In a chapter on Hegel in Part I,
Collingwood takes up Croce's criticism of Hegel's philosophy of
history which appeared to Croce
____________________
15 At NL, 24.63-24.68 Collingwood writes of dialectical
thinking, defined as "the readiness to give up something which
at a certain time you settled on as true." As an example Col-
lingwood cites Plato's discovery that the way to find one's way
about in a Heraclitean world is to think dialectically--a
Heraclitean world being one in which change from X to not-X or
vice versa is constantly occurring. At NL, 27.82 Collingwood
writes of the dialectical spirit as the spirit of agreement and
compromise in the ensuing discussion--see NL, 24.61, 27.9). And
at NL, 24.57-24.61 Collingwood writes of dialectical discussion,
where one's aim is to show that both disagreeing parties in an
argument are correct. NL, 24.57 contrasts dialectical and
eristic discussion, and declares that all logic is concerned
with discussions. In all these usages Collingwood seems to have
in mind a sense of the term derived from classical Greece, i.e.
the sense which emphasizes the manner of conducting a debate
involving disagreeing parties.
188
to be a "gigantic blunder produced by confusing two quite
different things: namely opposition and distinction." Then he
continues:
Concepts, Croce says, are related by opposition: good and

bad, true and false, freedom and necessity, and so forth;
and the theory of their relation, he admits, has been well
expounded by Hegel in his theory of dialectic, which de-
scribes the way in which any concept stands in a necessary
relation to its own opposite, generating it at first and
then negating it, so that the way in which the concept lives
is by creating and overcoming oppositions. But the
individual things which are the instances of concepts are
never related to each other by way of opposition, only by
way of distinction: consequently the relations between them
are not dialectical, and in history, which is the history of
individual actions and persons and civilizations, there is
consequently no dialectic . . . . (IH, 118-19).
It is interesting to note, before turning to Collingwood's
comments on this Crocean criticism, that in The New Leviathan
Collingwood remarks that logic applies to propositions only
because it applies in the first instance to concepts (NL, 7.33,
7.39, 11.35). Following this line of thought one might conclude
that D-logic applies primarily to concepts, and F-logic to
propositions. But extrapolating this analogy is not very
promising, since Q-A logic is not obviously inferential, where
F-logic is, and so is D-logic. Nonetheless there is some point
in the observation that the center of gravity, so to speak, of
D-logic is concepts, which emphasizes the semantic unity of
terms, and has as its central concern the meaning of terms,
whereas F-logic takes as its minimal unit the proposition that
can be the bearer not only of meaning but of truth or falsity.

189
However in the passage that we are presently considering
Collingwood is pursuing a different line of thought:
Plausible though Croce's view is, it does not really get to

the heart of the problem. It implies that in talking of
history we should never use words like opposition or
antagonism, and synthesis or reconciliation: . . . we ought
only to say that they are different: we ought not to speak
of an opposition, but only a difference, between Whigs and
Tories, or Catholics and Protestants. Now it is true that
we do not need to use terms like opposition (let me call
them dialectical terms), when we are talking only of the
outward events of history; but when we are talking only of
the inward thoughts which underlie these events it seems to
me that we cannot avoid them. (IH, 119).
To pursue his argument from the Idea of History any further
would get us into the subject of the next chapter, where we
shall have a chance to observe how his concept of history de-
veloped in the early writings. For now we wish to call attention
to the usage of opposition as a "dialectical" term, and
synthesis as a term of reconciliation.
As we proceed with our investigation of Collingwood's
logical views it will be well to bear in mind that "dialectic"
may represent a number of different meanings in his writings, so
that we might recognize, for example, that as the methodical
program we have called D-logic it stands in a relation with
F-logic and Q-A logic so as to appear to form a scale of forms,
from another point of view any such scale can itself be defined
as "dialectical." And again, while the analysis of a concept
into opposing principles may be called a dialectical
relationship and the

190
synthesis of this opposition may be called a rapprochement (so
that analysis, dialectic, and reconciliation may be interpreted
as three phases of complete philosophical thought), from a more
inclusive sense of the term (e.g. the one so frequently used in
Speculum Mentis) all rapprochement is dialectical. It may well
be that Collingwood omitted explicit reference to dialectic and
dialectical logic in the Autobiography because the term is
capable of such a multitude of meanings that it was virtually
useless for painting the kinds of sharp contrasts that he wanted
to elicit for the purpose of highlighting what he regarded as
central to his philosophical development.
But while we may charge Collingwood's Autobiography with
an error of omission, and even with being somewhat misleading in
that it suggests to the reader that his early Q-A logic already
made explicit the role of presuppositions, it is at least
accurate to the extent that the early writings do show evidence
of a stress on the role of questioning in the process of
knowledge. We can also say that nothing we have found would give
any indication that had he had such a complete Q-A logic he
would have done anything but embrace it wholeheartedly at this
point in his development, since its spirit is present from
Speculum Mentis onwards, and many of its principles are
explicitly stated in the early writings.

CHAPTER V
HISTORY AND PHILOSOPHY
1. Introduction.
There can be little doubt that Collingwood felt that his
greatest contribution to philosophy would be his reconciliation
of philosophy and history. The Autobiography describes in some
detail how the two activities between which he divided the
majority of his time, his actual work in Roman-British history
and his philosophical lecturing and writing, gradually
converged, and how this synthesis not only was in large measure
responsible for breaking with the epistemology of his realist
tutors at Oxford, but also was the foundation for the
development of his rapprochement philosophy. If we are to
understand how and why he recognized this relationship to be
axial not only for his development but for interpreting his
entire philosophical outlook, we must take a careful look at how
this philosophy of history took shape in his early writings. And
since the Autobiography specifies that this development can be
found in his essays of the twenties (A, 107, 116-17), it is this
group of articles that we must spend most of our time on in this
chapter.
191
192
But what is it that needs reconciling? If "the phil-
osophy of" something merely means, as Religion and Philosophy
asserts, the careful reflection upon that subject (RP, 15; FR,
53), what is problematic about a philosophy of history? In the
Autobiography the opposition is approached from the direction of
the realism-idealism controversy: realists assumed the existence
of permanent philosophical problems which could be analyzed
independently of their historical setting, so that the
historical question of what someone thought about one of these
"eternal problems" was distinct from the philosophical question
of whether or not he was right (A, 59-68). In addition to this
mistake, the realists assumed that all knowledge was a simple
apprehending or intuiting of some unaffected reality, which
failed to recognize the essential role of interpretation of
evidence in the case of historical knowledge (A, 25-26, 39-40).
But from the other side, in the practice of his historical
studies Collingwood found that those historians he most
respected knew and cared little about philosophy (A, 83), so
that the need for a reconciliation from this direction was
simply the absence of any serious philosophical reflection on
the subject-matter of history. Not only did realist philosophers
neglect history as a form of knowledge not assimilable to their
sense-bound epistemology, but historians neglected philosophy
and therefore failed to pay sufficient attention to the
foundations of their own science (A, 85-90).

193
When we turn to the early writings we find out that it
was not merely neglect that is involved in the distantiating of
philosophy and history, but a basic difference in orientation of
each that must be overcome. As we shall see in the next two
sections, Collingwood's earliest attempts at the philosophy of
history tend to assume that philosophy is the realm of the
universal and history the realm of the particular, and the
resolution of the conflict that arises when it is assumed that
truth resides either with the universal or the particular is by
rooting both in a "concrete identity" or "concrete
universal"--an individual which is both universal and particular
at once. While the idea of an individual historical event as
something which is both universal in meaning and concrete in
objective reference is an idea which as a long and continuous
development in Collingwood's philosophy of history, the focus of
it shifted gradually away from the object of history and towards
the thinking role of the historian, and it is this development
that Collingwood is anxious to indicate in his Autobiography.
It is also quite a natural course of thought from the
point of absolute idealism, as we saw in Chapter III; for if the
common ground of all idealism is the role of the mind as both
active and passive (FR, 275, 277), then to view the historian as
a merely passive receptacle for the transmission of fact as
"something independent of my own or your knowledge

194
of it" is on a collision course with the philosophical view of
knowledge as active, and ultimately with the philosophical de-
mand for self-knowledge. Just as Gentile distinguished between
the realistic concept of change in material objects and the
idealistic concept of the history of mind, and argued that
reality is history (FR, 274), so Collingwood would distinguish
between physical events and re-enacted historical acts, and
argue that mind is what it does, and what it does is to make
history. The path to self-knowledge then is through history, for
it is history which narrates what it is that man has done.
But the reconciliation of philosophy and history devel-
ops as Collingwood's concept of rapprochement became more com-
plete.
2. The Identity of History and Philosophy.
In the earliest form that Collingwood's reconciliation
between history and philosophy takes, the two are simply equat-
ed: philosophy and history are the same thing. In Religion and
Philosophy the identity between philosophy (like that between
religion and philosophy, and between religion and morality) is
established in two ways: first by criticizing the views which
hold them to be separate, and then by showing that they are
mutually dependent on one another. In the case of history its
abstract separation from philosophy has been held by those who
maintain either that historical facts are independent of the
philosophical constructions
195
or interpretations that are placed on them (historical
positivism), or by arguing that the past as such is unknowable
because it depends on inferences based on fallible data
(historical scepticism). Neither of these anti-historical
arguments can withstand criticism. Historical positivism fails
because the distinction between historical interpretation and
historical fact cannot be maintained: historical interpretation
is just historical fact further specified. If one "construction"
that is put on fact differs from another, it is not merely two
"ideas" superimposed on one fact: one was an historical fact and
the other a historical error (RP, 46; FR, 80). Similarly,
anti-historical scepticism fails because it is not just
historical data used in inference that is fallible; the same can
be said of all data. "If inference as such is to be distrusted,
the evidence that leads us to distrust it is discredited with
the rest" as another misreporting of a well-attested fact. (RP,
44; FR, 79).
But not only do arguments for positivism and scepticism
fail to dislodge historical thought, but we cannot do without
either philosophical or historical thought, since they co-imply
one another:
In the first place, it appears that history cannot exist

without philosophy. There is no such thing as an entirely
non-philosophical history. History cannot proceed without
philosophical presuppositions of a highly complex character.
It deals with evidence, and therefore makes epistemological
assumptions as the value of evidence; it
196
describes the actions of historical characters in terms

whose meaning is fixed by ethical thought; it has con-
tinually to determine what events are possible and what are
not possible, and this can only be done in virtue of some
general metaphysical conclusions. (RP, 46-47; FR, 81).
We shall see in Chapter VII that at least one of these "meta-
physical conclusions" (actually a presupposition) mentioned here
is taken by Collingwood from Bradley's The Presuppositions of
Critical History, which Collingwood would later acutely
criticize (IH, 238-39). It is interesting to note that this is
the first reference to "presuppositions" which we have
encountered in Collingwood's writings, and it occurs in the
context of an argument that identifies philosophy and history.
The argument continues by showing the philosophy needs or
presupposes history:
It is equally certain that philosophy is impossible without

history; for any theory must be a theory of facts, and if
there were no facts there would be no occasion for theory .
. . . History must be regarded not as a mechanical process,
nor yet as a gradual accumulation of truths, but simply as
objectivity; as the real fact of which we are conscious.
History is that which actually exists; fact, as something
independent of my own or your knowledge of it. In this
sense there would be no philosophy without it; for no form
of consciousness can exist without an object. (RP, 47-49;
FR, 81-83).
Collingwood draws the conclusion that the relation of history to
philosophy is that neither can exist without the other, or as he
says, "each presupposes the other" (RP, 49; FR, 83).

197
Each is knowledge; and if they are different, they must be

the knowledge of different objects. How can we distinguish
these objects? History, it is sometimes said, is knowledge
of the particular, philosophy knowledge of the universal.
But the particular is no mere particular; it is a
particular of this or that universal; and the universal
never can exist at all except in the form of this or that
particular. . . . History, like philosophy, is the
knowledge of the one real world; it is historical, that is,
subject to the limitation of time . . . . It is
philosophical, that is, all embracing, universal, for the
same reason; because historical fact is the only thing that
exists and includes the whole universe. History a parte
objecti--the reality which historical research seeks to
know--is nothing else than the totality of existence; and
this is also the object of philosophy. History a parte
subjecti--the activity of the historian--is investigation
of all that has happened and is happening; and this is
philosophy too . . . . (T)he philosophical presuppositions
of history are not something different from the history
itself; they are philosophical truths which the historian
finds historically exemplified. History and philosophy are
therefore the same things. (RP, 49-51; FR, 83-85).
As can be seen from this argument, Collingwood had not at this
point penetrated very deeply into the nature of historical
thought. As he refined the concept he would narrow the object of
history from "the knowledge of the one real world" or "the
totality of existence" to res gestae--deeds of men, done in the
past. And even in Speculum Mentis he showed dissatisfaction with
such wholesale identities as he proposed in Religion and
Philosophy. Since it is not a settled part of his mature
outlook, we shall not engage in criticism of the argument here
(although we shall do so in examining the concept of
rapprochement identity in Chapter VI). But while one may quarrel
with the argument here, the spirit of reconciliation is
certainly present here, its only flaw being that the

198
resulting rapprochement between history and philosophy is less a
marriage than a fusion. While he continued to use the strategy
of "co-implication and therefore identity" in arguing for the
reconciliation of any two forms of consciousness, he became more
careful (a) to specify the differences between them as well, (b)
to distinguish what any form of consciousness is implicitly from
what it is explicitly,] and (c) to locate them not as
co-ordinate species of the genus, "knowledge," but rather as
successors on a scale of forms.
3. The Concrete Universal -_Absolute Object.
This is evident in Speculum Mentis, where philosophy and
history are described as having aims insofar as they both
essentially assert concrete reality and deny simple abstraction.
Therefore "the identification of philosophy with history is far
less violent and misleading than its identification with
science, religion, or art." But it is immediately added that
"all such identifications are barren abstractions. since to
"assert the identity without the difference or the difference
without the identity is to turn one's back on reality and amuse
oneself with paradoxes" (SM, 246). The differentiating feature
of philosophy as opposed to history, as presented in Speculum
Mentis,
_______________________
1 Cf. SM, 108, n. 1, where Collingwood acknowledges that
in Religion and Philosophy he had overlooked the distinction
between implicit and explicit--roughly, the distinction between
its "promise" or what it indirectly implies (implicit) and its
"performance" or what a form of consciousness directly asserts
(explicit).
199
is that in philosophy subject and object are identified, whereas
for historical thinking concrete fact is always assumed to be
something independent of the knowing activity of the historian
(SM, 241-43, 249). That is why Collingwood calls modern realism,
in which the object is presumed to be unaffected by the knowing
of it, the historical form of dogmatism.
Now these appear to be paradoxical assertions if taken
in conjunction with what Collingwood says about the reconcili-
ation of philosophy and history in the Autobiography. (a) If it
is philosophy, and not history, that is the form of con-
sciousness in which subject and object are identified, then how
can history be the "self-knowledge of mind" that Collingwood
intended it to be? And (b) if the form of dogmatism peculiar to
the historian is realism, then how can realism be "a philosophy
which erred through neglecting history"? We already know part of
the answer to the second question from Chapter III: what makes
history dogmatic is its unquestioning acceptance of the concept
of fact as something independent of anyone's knowledge of it,
which is the fundamental concept underlying realism. Since there
are as many forms of realism as there are of dogmatism,
historical realism follows scientific, religious, and artistic
realism. What makes historical realism an error is its
assumption that this is the whole story, and what is neglected
is the higher concept of

200
history in which fact and the knowledge of fact are reconciled
in self-consciousness--i.e. the philosophical concept of
history. But this leads us back to the first question, and
ultimately to the development of the concept of history in
Collingwood's early writings. The first stage in this
development we have just reviewed in Religion and Philosophy.
The second is in Speculum Mentis, which we will consider in this
section. The third is in the essays on history that appear in
the decade of the twenties, which we will examine in the
immediately following section.
The recognition that the form of experience known as
history is an improvement on abstract science is an achievement
of the Renaissance, and is solidified in the experimental method
of Renaissance scientists. "Experiment means the recognition of
fact, and experimental science means the assertion of fact, even
if only mutilated fact, as the true presupposition of scientific
thought" (SM, 201-202).2 The first to recognize this explicitly,
says Collingwood, was Descartes:
_____________________
2 The reader would be correct to assume that by the
"assertion of fact" at this point Collingwood means something
more than the utterance of a statement that something or other
is a fact or is the case. In Speculum Mentis to assert a fact
means to assert it as true, which is more than merely observing
it. To be asserted as scientifically true it must be capable of
withstanding Baconian cross-examination, which means subjecting
a fact to the sort of treatment an hypothesis gets in the
laboratory. Cf. SM, 53.
201
All science, said Descartes, rests upon the one indubitable

certainty that I think and that therefore I exist. Now the
thought and existence of which Descartes spoke were not
abstractions--anything thinking anything, or anything
somehow getting itself thought about . . . . Descartes meant
what he said; and what he said was that the concrete
historical fact, the fact of my actual present awareness,
was the root of science. He was only going one step beyond
Bacon, for whom the root of science was natural fact:
Descartes, more profoundly, saw that before natural fact can
be of any use to the scientist he must observe it, and that
the fact of his observing it is the fact that really
matters. Science presupposes history and can never go behind
history: that is the discovery of which Descartes' formula
is the deepest and most fruitful expression. (SM, 202).
If it is surprising to find Descartes' cogito translated as "the
fact of my actual present awareness," and given credit for
expressing the discovery that "science presupposes history,"3 it
is nonetheless revealing of Collingwood's early view of the
matter: the recognition of the historical element of science is
the first stage of both a revolution in science (the development
of the experimental method) and the beginning of a revised
awareness of history (SM, 202-203). It also brings into focus an
early association of history as a form of knowledge and
perception as a level of mental activity. The early annalists or
historical compilers were not aware of their reliance on
perception, but as the concept
_____________________
3 As we shall see, this is not the last of Collingwood's
creative interpretations of Descartes and the history of
philosophy: cf. EPM, 10-25, 124-26, 155-60; EM, 185-90. In The
Idea of History Descartes is treated as an historical sceptic,
and Vico's anti-Cartesianism is hailed as the real beginning of
scientific historiography (IH, 59-70).
202
of history developed the role of perception, as opposed to sen-
sation, became more explicit.
Sensation is the false or abstract account of perception. In

perception we are immediately aware of our object, which is
a concrete and therefore historical fact: perception and
history are thus identical. But the immediacy of perception
does not exclude mediation, it is not abstract immediacy
(sensation) but implicitly contains an element of mediation
(thought) . . . . Perception is explicitly immediate, but it
always contains within itself mediation (thought,
"interpretation of sense-data," "inference from the
immediately given," or whatever one likes to call it) and is
therefore never abstract immediacy . . . . History is thus,
as a specific form of experience, identical with perception.
(SM, 204-205).
Just as perception requires an element of memory insofar
as an object of perception is grasped not all at once or
immediately but serially, or as a "synthesis" or reconstruction,
so also the annalist or writer or memoirs is someone whose
reminiscences rely on memory--one's own or that of a
"source"--for its narrative sequence. But such a reliance on
memory must, to retain even the rudiments of reliability,
distinguish between memory and pure imaginative fantasy, and
this requires the exercise of a selective criterion (SM,
213-14). The earliest forms of historiography were dominated by
various mixtures of the true criterion of history with ideals
drawn from other forms of experience, so that in historians like
Herodotus and Thucydides factual coherence was mixed with
aesthetic or dramatic effect (SM, 214-16). It is only in the
eighteenth century that the concept of fact becomes explicit and
such errors were overcome to reach "historians' history" (SM,
216).
203
But it is at this point that fact is elevated to the
level of the absolute:
There is thus no feature of experience, no attitude of mind

towards its object, which is alien to history. Art rests on
the ignoring of reality: religion, on the ignoring of
thought: science, on the ignoring of fact; but with the
recognition of fact everything is recognized that is in any
sense real. The fact, as historically determined, is the
absolute object. The mark of the absolute object is
individuality, for individuality is concreteness. The object
as individual is the whole of what exists, and this is
concretely articulated into parts each of which is again
individual, and so to infinity . . . . The object, as a
system of fact so organized, is objective throughout, for
every part is a true microcosm, and is truly infinite. (SM,
218-19).
It is to this extent that history achieves what art, religion,
and science could not. For where esthetic consciousness ex-
presses itself in a monadic work of art, but while each such
work is its own cosmos, it is only so by ignoring not only other
works of art but the everyday world in which the artist lives
and works (SM, 84, 219). Religious consciousness locates its
individuality in God, "the monad of monads, a cosmos whose
structure is that of the absolute object"--but whose individu-
ality stands over against that of the world "whose very nature
is to be outside him," and therefore leaves the absolute in-
dividuality of God unattained (SM, 219). Scientific conscious-
ness replaces the concept of God with the concept of law, but
with the consequence that "(w)hat is individual and organized as
a system of individuals is not the world but only the concept"
which keeps law separated from true individuals to the

204
extent that any such individual is only a particular instance of
the law. The failure of each of these forms of experience is
redeemed by the success of history, which reaches the idea of an
object "beyond which there is nothing and within which every
part truly represents the whole."
This absolute whole is the concrete universal; for concrete

universality is individuality, the individual being simply
the unity of the universal and the particular. The absolute
individual is universal in that it is what it is throughout,
and every part of it is as individual as itself. On the
other hand it is no mere abstraction, the abstract quality
of individualness, but an individual which includes all
others . . . . The principle of its structure is not
classification, the abstract concept, but the concrete
concept, which is relevance, or implication. . . . (T)he
concrete universal is the daily bread of every historian,
and the logic of history is the logic of the concrete
universal. (SM, 221).
It is doubtful whether every historian would recognize
that his "daily bread" consisted of "the concrete universal" in
the sense of being the "system of systems, the world of
worlds"--including the later Collingwood himself, as we shall
see in Chapter IX: while the historian brings to his study an a
priori concept of the past, as well as presuppositions about the
coherence of the past, he feels himself under no obligation to
consider the whole of it as an absolute object consisting of an
infinity of facts (IH, 240-45, 303). In fact it is under the
burden of such a goal that history collapses by the "inner
dialectic" of its own version of the monism-pluralism dilemma.
If history exists, its object is

205
an unknowable infinite whole; and if its parts are atoms, then
history disappears and science takes its place, with its own
unresolved problems of universality and particularity (SM, 234).
In either case, the absolute object remains unachieved.
Thus history is the crown and the reductio ad absurdam of

all knowledge considered as knowledge of an objective
reality independent of the knowing mind. Here for the first
time we place before ourselves an object which satisfies
the mind; an object individual, concrete, infinite, no
arbitrary abstraction or unreal fiction, but reality itself
in its completeness . . . . The progressive alienation of
the mind from its object is in history complete. The world
is triumphantly unified as object, only to find itself
separated from the mind by a gulf which no thought can
traverse. (SM, 238).
We shall resume this discussion in the next chapter,
where we shall be concerned with the rapprochement identity of
absolute knowledge as it is presented in Speculum Mentis, but at
this point we must make several observations about the concept
of history herein presented. The reader has no doubt wondered
why it is necessary for history to postulate itself as an
absolute object at all, for if it were not for this the final
"reductio ad absurdam" would not occur at all. It appears that
the chapter on history is only a stage along the way to
establishing the demands of absolute knowledge--i.e. the demand
for an object that will fully satisfy the mind. What mind? The
mind of an absolute idealist, one concludes-for there is only
token effort to take into account the actual praxis of
historians in Speculum Mentis, other than the occasional remarks
about Herodotus and Thucydides, Mommsen and Gibbon. What we are
confronted
206
with instead is what amounts to a very abstract view of the
subject matter, one which tends to assimilate historical
thinking to perception and memory, and the historical object to
the world of fact--all of it. Little consciousness is exhibited
of the uniqueness of human acts in the range of facts, of the
need for cross-examination of facts by the historian's critical
intelligence, or the legitimate role of the historian's
presuppositions in this process.
We are therefore a long way from the view of history
presented in the Autobiography--but this itself is consistent
with the autobiographical interpretation, since Collingwood
states that in these years the rapprochement between history and
philosophy was incomplete, and was subject to a long and
painstaking development. The evidence for this development is
contained in the essays written between 1920 and 1930, and to
these we must now turn.
4. The Ideality of History as a Scale of Forms.
There is no clearer index to Collingwood's views on the
philosophy of history than that which can be obtained by careful
attention to these essays from the decade of the twenties. In
them one finds a gradual shift away from a definition of history
solely in terms of the realism-idealism controversy and toward a
clear recognition of the multi-layered

207
senses of the term. Between these two ends of the scale we shall
find not only most of the "principles of history" mentioned in
the Autobiography making their appearance, but we shall also see
parallel discussions of both the static logic of historical
assertions (for example concerning "categorical judgements" and
"the concrete universal") and the dynamics of history in terms
of processes and dialectical development.
A. In "Croce's Philosophy of History" (1921) and "Can
the New Idealism Dispense with Mysticism?" (1923) Collingwood
writes, as we have already seen, as a sympathetic critic of
Italian and German idealism, and in the process of his discus-
sion of this philosophy he formulates (and explicitly approves)
the notions of a "living past," history as dealing with thought,
and the basic process of "re-thinking" past thoughts (EPH, 6-10;
FR, 274-75).
History goes on in the mind of the historian: he thinks it,

he enacts it within himself: he identifies himself with the
history he is studying and actually lives it as he thinks
it, whence Croce's paradox that "all history is contemporary
history." . . . History is thought, annals the corpse of
thought. But has thought a corpse? and if so, what is it
like? . . . Croce's general "philosophy of the spirit"
supplies him with a ready-made answer. Nothing exists but
the spirit; but the spirit has two sides or parts, thought
and will . . . . Thought is the synthesis of subject and
object, and its characteristic is truth; will is the
creation of an object by the subject, and its characteristic
is utility . . . . Annals are not thought but willed; they
are constructed--"drawn up"--by the historian for his own
ends. (EPH, 6-7).
208
In this 1921 essay Collingwood goes on to criticize Croce for
mixing idealism with naturalism (which, we recall, is a synonym
for realism), and he calls on Croce to purge his philosophy of
its naturalistic elements in order to "reach the point of
absolute idealism" which the essay appears to assume to be a
step forward (EPH, 22). While Collingwood is critical of Croce
for reducing philosophy to the methodological moment of history,
and for absorbing philosophy into history (EPH, 20-21), the idea
of a living history of thought remains untouched. It comes
fairly close to the "first principle" of history as Collingwood
formulated it in the Autobiography (A, 110).
In the 1923 essay Collingwood presents in even stronger
terms the contrast between the idealistic and the naturalistic
view of history. We already quoted the passage in which
Collingwood contrasts Croce and Gentile on the issue of the
"metaphysic of being" which presents a philosophy of the real-
istic concept of change, and a "metaphysic of knowledge" which
presents a philosophy of idealistic concept of history (FR,
274). Although Collingwood does not explicitly state the extent
to which he would subscribe to either Croce's position or
Gentile's, these passages do show an awareness of a distinction
between process in general and historical processes. When taken
in conjunction with his charges against Croce, it is not
unreasonable to infer that Collingwood leaned toward a view on
the ideality of history not significantly

209
different (at this point) from that of Gentile. By 1923,
Collingwood appeared to have accepted the basic ideality of
history, i.e. that thought is essential to it. Later Collingwood
will drop the idealism-realism emphasis, while retaining the
distinction between natural change and historical processes.
B. In "Are History and Science Different Kinds of
Knowledge?" (1922) and "The Nature and Aims of a Philosophy of
History" (1924), Collingwood examines the epistemological claims
of history vis-a-vis other forms of knowledge. These two essays
represent a position intermediate between that of Religion and
Philosophy (in which various forms of knowledge are identified
insofar as they are directed toward the same intentional object)
and that of Speculum Mentis (in which forms of knowledge are
distinguished on the basis of the manner in which they grasp
their intentional objects--as successive approximations to the
truth).
In the 1922 essay Collingwood denies that there are any
epistemological grounds for a distinction between history and
science: "when both are regarded as actual inquiries the
difference of method and logic wholly disappears" (EPH, 33). It
is not the case that the scientist deals with the universal and
the historian with the particular exclusively, nor that it is
the function of the scientist to generalize and of the historian
to particularize; both deal with the world of individual fact,
and both activities involve the interpreta-

210
tion of individual facts in terms of general concepts (EPH,

26-27).
Interpretation is not the employment of a previously con-

structed tool (concept) upon a separately-given material
(fact): neither the concept nor the fact is "possessed"
(thought and observed respectively) except in the presence
of the other. To possess or think a concept is to interpret
a fact in terms of it: to possess or observe a fact is to
interpret it in terms of a concept. Science is this
interpretation . . . . The object which the scientist
cognizes is not "a universal," but always a particular fact,
a fact which but for the existence of his generalizing
activity would be blank meaningless sense data. His activity
as a scientist may be described alternatively as the
understanding of sense-data by concepts, or the realizing of
concepts in sensation, "intuiting" his thoughts or "thinking
out" his intuitions . . . . (T)here is no such thing as
knowledge either of the particular or of the universal, but
only of the individual: and . . . the sense-datum (pure
particular) and concept (pure universal) are false
abstractions when taken separately ((and)) yet, as elements
in the one concrete object of knowledge, the individual
interpreted fact, are capable of being analytically
distinguished. (EPH, 28-29).
In the 1924 essay Collingwood seems to reverse his
earlier position: he declares that "history and science are not
identical a parte subjecti)" because "scientific thinking is an
abstract thinking, historical thinking a concrete thinking"
(EPH, 45). We recall from Religion and Philosophy that by "a
parte subjecti" Collingwood understands precisely the activity
of the knower qua historian, scientist, etc.--in contrast to "a
parte objecti)" which refers to the reality or object studied by
a given form of knowledge (RP, 51; FR, 84-85).
In short, Collingwood seems to be saying here that,
contrary to his earlier judgment, there is an epistemological
ground for
211
distinguishing science from history--the distinction being that
between abstract and concrete thinking respectively.
The apparent contradiction is reconciled somewhat when
one pays careful attention to the shift of levels between the
two essays--the first dealing with the praxis of the two kinds
of inquiry (" . . . regarded as actual inquiries . . . "), the
second with the ideals of the two forms of knowledge. The dis-
tinction is important for an understanding of Collingwood's
emerging philosophy of history, and the point of view which he
adopts in Speculum Mentis. Both science and history, he wishes
to tell us, deal with individual facts, and both make use of
generalizations and the application of generalizations (par-
ticularization) to concrete cases. But each proposes a different
ideal for itself.
Ideally, historical thought is the apprehension of a world

of fact. Actually, it is the presentation by thought to
itself of a world of half-ascertained fact: a world in which
truth and error are at any given moment inextricably
confused together. Thus the actual object of actual his-
torical thinking is an object which is not "given" but
perpetually in process of being given . . . . The philosophy
of history, therefore, is a study of historical thinking:
not only psychological analysis of its actual procedure, but
the analysis of the ideal which it sets before itself.
Historical thought is one among a number of attitudes taken
up by the mind towards the objective world; it is an
attitude which assumes that there exists a world of
facts--not general laws, but individual facts--independent
of the being known, and that it is possible, if not wholly
to discover these facts, at any rate to discover them in
part and approximately. The philosophy of history must be a
critical discussion of this attitude. (EPH, 44).
212
But this distinction (based on mental dispositions, be
it noted) still does not extract Collingwood from his dif-
ficulties, since (a) to distinguish between what a form of ex-
perience or thought is ideally from what it is in practice, or
actually, is, on Collingwood's own grounds, to commit the error
of realism (SM, 313; cf. RP, 8, 29-33; FR, 47, 66-68); and more
seriously, (b) the historian does not always deal with the same
set of "facts" as does the scientist--a point which Collingwood
had already acknowledged in his 1923 essay on the New Idealism,
wherein he recognized that the facts with which the historian is
concerned are consciously performed processes in contrast to the
events of nature: "that which has a history . . . is a mind, for
matter may change but it cannot be said to have a history" (FR,
274). To be subject matter for science an event must merely be
capable of being thought. With events that are already thoughts
the scientist has no direct concern, since the facts which he is
proposing to explain are not artifacts but rather natural events.
Collingwood makes no note of these objective distinc-
tions in his 1924 essay since he is limiting himself to a dis-
cussion of the differences in ideals between history and other
forms of knowledge. Consequently he stresses the manner in which
the historian, as opposed to the scientist, seeks to express his
judgments about the "facts" he is investigating. Although both
hypothetical and categorical forms are used in

213
science and history, the scientist seeks to express his judg-
ments primarily in the hypothetical mode, in which the antece-
dent is the ground of the consequent ("if equals be added to
equals, the sums are equal"), whereas the historian aims at
categorical judgments. "The ideal of history, then, is to be a
single categorical judgment, articulated into an infinity of
coherent categorical judgments, asserting the reality and ex-
pounding the nature of an infinite individual world of fact
articulated into an infinity of individual facts" (EPH, 46). The
historian assumes the objective independence of these facts from
the knowing mind, so that "these actual happenings are always
the object of his thought, and never his thought itself"--this
being the realistic bias of the historian (EPH, 46-47).
As in Speculum Mentis, which was published in the same
year as the essay under consideration, Collingwood is unclear
about what he understands by "categorical judgement," and his
example does not clarify his usage. Collingwood uses as an
example of a categorical judgment used by scientists the state-
ment: "all whales are mammals"--which he then criticizes as not
truly categorical since it "does not imply an enumeration of all
actual whales but rather tells us that whatever we can identify
as a whale, if and when we do so identify anything, we can
further identify as a mammal." The implication seems to be that
(a) categorical judgments do imply an actual enumeration of all
the entities
214
involved--that is, some sort of quantification; and (b) that
categorical judgments as used by scientists refer to their
objects only hypothetically, and hence fail to be truly
categorical. In the 1924 essays the doctrine of categorical
judgments is vague, and requires further clarification.
But Collingwood does further specify what sort of
"single categorical judgement" the historian's ideal would be,
and what the limitations of such an intentional object would be.
The infinite whole of fact which it is the historian's

business to determine is . . . a world whose centre is the
historian's "immediate" perception, and whose radius is
measured by the depth to which he can see into the signi-
ficance of that perception . . . . The world of every his-
torian is limited by the limits of his knowledge . . . .
Each historian sees history from his own centre, at an angle
of his own: . . . so the various "perspectives" of
historians are arranged in a "space of perspectives;" each
historian is a monad which mirrors the universe from a point
of view that is irrevocably not any other's point of
view . . . . But a monad has no windows, and the historian
as such cannot do the work of co-ordinating the infinity of
possible perspectives. He can only travel from one per-
spective to another . . . . But in reflecting, that is
philosophizing, about his own thought he recognizes that he
is a monad, and to realize that one is in the "egocentric
predicament" is to transcend it. When thought returns upon
itself and faces the question of its own relation to its
object, by criticizing the point of view from which it has
regarded that object it transcends this point of view. (EPH,
53-55)4
___________________
4 Although Collingwood twice uses the expression, "space
of perspectives" in this essay, it never again occurs in his
writings. The term strikes the reader familiar with Husserl and
Lonergan as similar to the notion of "horizon."
215
It is here that we have the rudiments of an answer to
the first of our two paradoxical questions at the beginning of
Section 3. Although the historian's ideal is to articulate a
single categorical judgment, in practice he is always partially
separated from the object of his thought, so that his thought
and its relation to its object is at best one of similarity and
not one of total identity. As an historian, then, "he is always
the spectator of a life in which he does not participate: he
sees the world of fact as it were across a gulf which, as an
historian, he cannot bridge" (EPH, 47). But as a philosopher the
historian can also reflect on the fact that he himself "is part
of the world of fact, and that his own historical thought is a
product of the historical process which he is studying" (EPH,
47). In other words the historian qua historian is limited by
his "realist" assumption about the independence of the facts he
is observing, but the historian qua philosopher is capable of
transcending (by sympathetic interpretation and by critical
reflection) this separation of subject and object. History is
therefore the "self-knowledge of mind" when it becomes
self-consciously or critically reflective about itself, or when
it becomes philosophy. (Cf. EPH, 85-86).
Furthermore the way in which history becomes critically
reflective is by attending to the way in which historical
problems arise within the historian's own experience (EPH, 51-
53). Before something can become an

216
historical issue "the problem must arise within historical
thought; it must, that is to say, arise somehow out of the
attempt to perceive more adequately the world that exists here
and now for our perception" (EPH, 53).
Hence although Collingwood does not use the terms "en-
capsulation" and "re-enactment" in this essay, the concepts and
their relationships are the same as those which he expresses in
the historical principle which we have called "HIST-3" (number 5
of Table 4) in our summary of the Autobiography. The experience
in question, says Collingwood,
is . . . subject to the distinction between truth and

falsehood: hence we have not only to read, but to criticize.
The recognition of this truth is what differentiates history
in the higher sense of the word from the mere absorption and
repetition of stories . . . . This critical work is
sufficiently difficult to require somewhat elaborate
training, which involves the incidental construction
of . . . historical methods. As the word method suggests,
these sciences consist of empirical generalizations or rules
of procedure, instructing the student how to proceed in
typical cases . . . . Their business is to solve the problem
"how can the historian check his sources"? to which the
general answer is, "the historian who knows his business can
always invent methods of checking any source." (EPH, 51-52).
The only element lacking in this account, as compared to that of
the Autobiography, is the fully developed notion of the "real
life" or "practical problems" which constitute the "superficial
or obvious present" of the historian: these elements are added
in his 1930 essay, "The Philosophy of History," which we will
consider presently.
217
But we must observe in passing that in the above passage
Collingwood is beginning to relate several concepts which are of
very high importance for his later philosophy: (1) he is
beginning to acknowledge explicitly that there are levels of
meaning to the concept of history; (2) he recognizes that in the
higher sense of the term, truth and falsehood are distinguished
by means of critical thinking; (3) that this critical thinking
gives rise to historical methods; and (4) that these methods
consist of rules of procedure on how to proceed in typical
cases. Furthermore (5) it is noteworthy that in this 1924 essay
(as well as in Speculum Mentis, published the same year) he
first recognizes that "the various forms of thought (art,
science, history, philosophy) are not species of a genus," but
rather form what he was later to call a "scale of overlapping
forms" such that "art and science are contained in history, not
excluded from it: yet contained in a form transmuted by their
subordination to the historical end," while history "is not
contained in this manner in art or science" (EPH, 48).
Similarly, the philosopher must in a sense be an historian

and the historian in a sense a philosopher; but the phil-
osopher is suppressed in the historian, and the historian is
preserved but subordinated in the philosopher; history is
included in philosophy while philosophy is excluded from
history. (EPH, 49)
Here we catch a partial glimpse of an answer to the
second of our two paradoxical questions at the beginning of the
previous section of
218
this chapter: the realist is a philosopher who attempts to
ignore the history of his own subject matter, which, on
Collingwood's view, is a self-stultifying process. Not only is
realism the root of all forms of formalistic and abstract
dogmatism (which would mean, in terms of the account given in
Speculum Mentis, that it erred as much through neglecting
history as it did through neglecting art, religion, and
science--that is, it neglected them as concrete but partial
modes of thought), but it is a philosophy which neglects the
historical appreciation of concrete fact, which is its own
basis. The full flowering and ultimate consequence of the
realistic attitude towards the objects of knowledge is history;
but to deny that subject and object can ever be reconciled to
one another (which is the same thing as to affirm that objects
of knowledge are unaffected by the knowing of them) is to
prevent the historical viewpoint from ever reaching its highest
point of development--viz. a self-consciously critical process,
the self-knowledge of mind.
C. But the complete answer to our two questions is
possible only by an explication of two further lines of thought:
(1) the idea of history as a developmental process, and (2) the
nature of historical evidence. The latter is developed in two
essays published in 1928 and 1930, which we shall consider
shortly; the former is worked out in three essays in which
Collingwood deals with the cyclic view of history and its
companion notion
219
(in the 19th century at least) the idea of historical progress.
In "Oswald Spengler and the Theory of Historical Cycles"
(1927) Collingwood criticizes Spengler for writing history as
though he were describing a natural phenomenon, which reduces it
to so many episodes in the natural lifecycle of an atomic
"culture," which is elevated to the status of a thing with
definite characteristics. Collingwood's objection to Spengler is
similar in strategy both to his previous criticism of Croce and
to his later criticism of Toynbee: an historian cannot treat his
subject matter as if it were a natural entity with physical
properties subject to natural laws, because it commits the
positivistic error of regarding as given what in fact is
constructed by the active process of interpretation on the part
of the historian.
These are not superficial flaws . . . . They are sacrifices

of truth to method; they are symptoms of a logical fallacy
which underlies the whole book and has actually been
erected into a principle. The fallacy lies in the attempt
to characterize a culture by means of a single idea or
tendency or feature, to deduce everything from this one
central idea without recognizing that a single idea,
asserted in this way, calls up its own opposite in order to
have something to assert itself against, and henceforth
proceeds, not by merely repeating itself, but by playing a
game of statement and counter-statement with this
opposite . . . . (W)here ((Spengler)) fails is in thinking
out what he means by "characteristic." He thinks that the
characteristic is a fundamental something whose logical
consequences flow smoothly and unopposedly into all its
manifestations; whereas it is really the dominant partner
in a pair of opposites, asserting itself only so far as it
can keep its opposite in check and therefore always colored
by the hidden presence and under ground activity of this
opposite. (EPH, 63, 65).
220
Now while Collingwood does not explicitly use the
expressions "event" and "process" in this passage, it is clear
that his critique of Spengler is based on the historical
principle we have called "LG" in our summary of the
Autobiography. Any single idea, tendency, or feature of a
culture in history "calls up its own opposite" with which it
plays a game of statement and counter-statement--becomes, in the
terminology of the Autobiography, a process the components of
which are not static characteristics that do not change, but
rather "turn into one another":
(W)hile recognizing that a given culture has a certain

self-consistent character, a fundamental idea which is
working itself out into a complete social life, we must
assert that this idea or character is not static but dy-
namic; it is not a single unchanged thing . . . but a pro-
cess of spiritual development, an idea which grows out of
other ideas, in an environment of other ideas, which asserts
itself against these other ideas through a process of
give-and-take in which it modifies them and is modified by
them in turn. (EPH, 73).
In "The Theory of Historical Cycles" (1927) and "A Philosophy of
Progress" (1929) Collingwood continues to develop the concept of
history as a living thought-process rather than a dead set of
events--first by showing that the cyclical view of history is a
function of the limitations of the historian's own knowledge,
and secondly by locating the theory of historical cycles within
the context of range of possible viewpoints that the historian
can take toward his subject matter.

221
In the 1927 essay Collingwood stresses the need for a
unifying principle in history in contrast to its pluralizing
tendencies, which culminate in taking its data as so many atomic
events. In addition to being held together subjectively by the
historian's own thinking, it is held together objectively by a
continuity of problems the successive resolutions to which form
the fabric of history. So far as the historian can see history
as a whole, he sees it "as a continuous development in which
every phase consists of the solution of human problems set by
the preceding phase" (EPH, 87).
But that is only an ideal for the historian; that is what he

knows history would look like if he could see it as a whole,
which he can never do. In point of fact, he can only see it
in bits; he can only be acquainted with certain
periods . . . . At any given moment, therefore, the
historian can only present an interim report on the progress
of historical studies, and there will be gaps in it. These
gaps will appear as breaches in continuity, periods in which
the historian loses track of the development . . . . In this
condition, we see history split up into disconnected
episodes, each episode forming a relatively intelligible
whole, separated from its neighbors by dark ages. That is
the point of view from which we see history in cycles. . . .
The cyclical view of history is thus a function of the
limitations of historical knowledge. (EPH, 87-89).
In "A Philosophy of Progress" (1929) Collingwood sets
out to answer the question he had posed in his 1927 essay on
historical cycles: do the moral categories of good and bad have
a place in the evaluation of the course of history (EPH, 77)?
His answer, in brief, is that "whether

222
you think the course of events is an upward or a downward course
depends not on it but on you" (EPH, 109)--that is, on the
presuppositions of the historian. Three views are possible: the
Greco-Roman view in which history is a process of increasing
decadence from a presumed golden age; the 18th and 19th century
European view of history as giving evidence of progressive
improvement of human conditions and man's ability to solve his
perennial problems; and the more recent theory of historical
cycles, in which there is an alternation between periods of
decadence and of progress (EPH, 104-105). Since the third is
nothing more than a combination of the first two, the real
question is one which gets at the basis on which one measures
progress or decadence, and Collingwood suggests that in the end
the question is not so much factual or theoretical as much as it
is practical. The historian who works to preserve and improve
what he finds of value in the world and who therefore continues
to find the world as it is a better place on the whole than it
was, will view history as on the whole a progressive development
leading to the present. One who resigns himself to inactivity
finds it on the whole a worse place than it was in a previous
age; he feels that a past age is not again achievable, and
therefore writes as a pessimistic historian who views history as
decadence. And finally, one who finds that some things at the
present time are degenerate and require improvement (e.g. a
present form of government) while others are already notably
better (e.g. architecture and

223
engineering) will find both elements of progress and decadence
in history, and opt for some version of a theory of historical
cycles. But in any case the question is decided by the
limitations in the present practical life of the historian:
The question whether, on the whole, history shows a progress

can be answered, as we now see, by asking another question.
Have you the courage of your convictions? If you have, if
you regard the things which you are doing as things worth
doing, then the course of history which has led to the doing
of them is justified by its results, and its movement is a
movement forward. (EPH, 120).
Here we not only have an illustration of Collingwood's
use of "HIST-4" to solve an historical problem, we also have the
second phase of the answer to the first of our two paradoxical
questions. The process of history is taken by the historian as
consisting of events that occur independently of his knowing
them, but upon critical examination the nature or meaning of
these events is determined by the historian's judgment. This
judgment is limited by the limitations of the historian's own
mind. But these limitations are not merely theoretical, they are
also practical: at the level at which the historian is both
agent and patient, or at the level in which subject and object
are identified, the historical processes are additionally
determined by the historian (e.g. whether or not they shall be
progressive, retrogressive, or cyclical).
D. If such a conclusion as this seems to destroy the
last shred of historical objectivity, the balance is restored in
two further essays,

224
"The Limits of Historical Knowledge" (1928) and "The Philosophy
of History" (1930), which complete the sketch of Collingwood's
development of the concept of history prior to 1932, and before
he had written the series of lectures which were later to be
published as The Idea of History.
The 1928 essay argues that although the historian is
limited by the quantity and quality of the evidence he has at
his disposal, this does not mean that the historical sceptic is
correct in his claim that history is "the doubtful story of
successive events." On the contrary, historical scepticism is
only the negative aspect of a full definition of history which,
when confronted by conflicting evidence, cannot stop short at
the critical confrontation of one statement by its contradic-
tory, but must proceed to the dialectical task of showing why
one statement must be revised on the basis of another--why, in
short, criticism is a phase in the complete process of histor-
ical thinking, which must present its arguments on the basis of
all relevant evidence (EPH, 96). What the historian is seeking
is not "what really happened," since this has about the same
status as the Kantian unknowable "thing in itself." The past
referred to by historical thinking is not, that is, what is
demanded by that permanent tendency in all thought which is
"sometimes called the plain man's realism--to think of the
object as a 'thing in itself,' a

225
thing out of all relation to the knowledge of it, a thing
existing in itself and by itself" (EPH, 99). Such thinking leads
to the notion of a specious past--a "limbo where events which
have finished happening still go on," where the past is
something still existing in a ν ο η τ ο s τ ο π ο s of its own; a

world where Galileo's weight is still falling, where the smoke
of Nero's Rome still fills the intelligible air, and where
interglacial man is still laboriously learning to chip flints"
(EPH, 101). On the contrary, what the historian seeks is a
present filled with those symbols of the past which Collingwood
calls "evidence."
An event that has finished happening is just nothing at

all. It has no existence of any kind whatever. The past is
simply non-existent . . . . What the historian wants is a
real present . . . . He wants to reconstruct in his mind
the process by which his world--the world in those of its
aspects which at this particular moment impress themselves
on him--has come to be what it is. This process is not now
going on . . . . He is trying to know the past . . . as it
appears from its traces in the present . . . . (A)ll
historical thought is the historical interpretation of the
present. By leading to the present, it has left its traces
upon the present; and by doing that, it has supplied the
historian with evidence concerning itself, a starting point
for his investigations. (EPH, 101-102).
The process of historical criticism, therefore, is a process of
confronting one piece of evidence with another, related piece,
from the point of view of establishing how the historian's own
present came to be what it is (EPH, 98-99).
Thus far Collingwood has only determined that the his-
torian's knowledge of the past is limited by evidence, and that
this evidence must be something that constitutes part of the

226
historian's own present. In "The Philosophy of History" (1930)
he not only refines the notion of evidence, but also adds a
second set of limitations to historical thought, connects
historical inquiry with Q-A logic, again attacks the realists'
conception of a specious past, and outlines several of the
remaining "principles of history" with which we are already
familiar from the Autobiography--all this within the very
compact space of the four concluding pages of the essay.
First, with respect to evidence and the limits of his-
torical knowledge:
History is knowledge of the past, and the past consists of

events that have finished happening. The past does not exist
and cannot be perceived; our knowledge of it is not derived
from observation and cannot be verified by experi-
ment . . . . We come to know the past, not immediately, but
by interpreting evidence. This evidence (or data) is
something that exists in the present and is perceived by the
historian . . . . But data are not enough. They must be
interpreted. This requires principles, and the body of
principles constitutes historical method or technique. Some
of these principles are scientific in character, that is,
they concern particular groups of evidence . . . . Some are
philosophical, that is, they apply universally to all
evidence whatever, and compose the logic of historical
method. It is to this that we must refer such problems as,
the nature and limits of negative evidence, the possibility
of analogical argument, and so forth. Data, on the one hand,
and principles of interpretation on the other, are the two
elements of all historical thought. (EPH, 13637).
This is one of those passages where the reader is confronted
with a direct contradiction. Did not Collingwood, after all,
state in the Autobiography that history is not concerned with
past events (which begin and end) but

227
with processes (which do not)? Was this not the "first
principle" of history ("LG")? Here the interpreter must tread
with extreme caution, but in the process he will make a
discovery of the first importance in understanding how to treat
the many direct contradictions one finds in Collingwood's
writings. Making use of his "dialectical principle," Collingwood
often begins an expository section of his writing by making a
statement that he thinks corresponds to the ordinary beliefs of
his reading audience, and then in successive sentences and
paragraphs, leads him beyond this point to the position he
wishes to establish.5 In this case Collingwood is starting from
the point of view of the "plain man's realism"--that history is
knowledge of an object, the past, which exists independently of
his knowing it-and leads him to a modified notion of history as
knowledge of a significant present. This is manifestly his
intention, and it is made clear by filling in the first ellipsis
in the above quote, and adding the concluding paragraph of the
essay:
___________________
5 Cf. CRC, 9 May 1935, Collingwood to Ryle, p. 15: "(I)n
the work of any competent philosopher I find that the part
played by systematic fallacies is partial only; repeatedly, when
real difficulties arise, his insight into the subject, sharpened
by the sense of the difficulty, leads him to reject the fallacy
even at the cost of inconsistency and to adopt a better
procedure than that which he had followed . . . . Now this being
so, a philosopher named as the victim of a fallacy might . . .
say to me: 'You pillory me unfairly; on page X, it is true, I do
fall into your fallacy; but on page Y I correct it; you ought to
take my work as a whole, and interpret X in the light of Y; if
you did so you would see that the error was only a temporary
slip at worst; and, at best, you might wonder whether it was not
merely the exploration of a provisional point of view.'"
228
A "realistic" theory, according to which knowledge is the

"apprehension of a really existing object," is ruled out as
absolutely inapplicable to history. . . . (T)he historian
does not select, because no past facts are "there" before
him, to select from, until he has put them there by sheer
historical thinking. . . . Finally, since the past in itself
is nothing, the knowledge of the past in itself is not, and
cannot be, the historian's goal. His goal, as the goal of a
thinking being, is knowledge of the present . . . . But, as
historian, he is concerned with one special aspect of the
present--how it came to be what it is. (EPH, 136-39).
We shall have occasion in the sequel to examine many
other instances where Collingwood seems to take back with one
hand what he had given with the other, but in the meantime we
should note that a successful gloss on these passages would be
something like the following: "Although from the point of view
of the plain man's realism history is taken to mean knowledge of
the past, where the past consists of events that have finished
happening, since the past does not exist (as the realistic
theory might lead one to suspect) one must revise the notion of
the past to mean that aspect of the present which gives
indication (evidence) of how it has come to be what it is--and
this is the past the historian is concerned with." Such a
reading does not violate Collingwood's "first principle of
history," and in fact makes sense of the complete passage
--especially its final paragraph. But be it noted, it still does
not get Collingwood out of the woods, because the reader is
still puzzled by what it would mean for historical objectivity
for the historian to "put" facts before himself "by sheer
historical thinking." For a clarification of what this means

229
we must next pay note to the role of Q-A logic in historical

thinking:
The beginning of historical research is therefore not the

collection or contemplation of crude facts as yet unin-
terpreted, but the asking of a question which sets one off
looking for facts which may help one to answer it . . . .
And the question must be asked with some reasonable ex-
pectation of being able to answer it, and to answer it by
genuinely historical thinking . . . . We express this by
saying that a question does or does not "arise." To say that
a question arises, is to say that it has a logical connexion
with our previous thoughts, that we have a reason for asking
it and pre not moved by mere capricious curiosity. (EPH,
137)6
Now although there is no mention in this passage of Q-A logic as
such, it is clear (e.g. from the discussion of a question's
"arising" by logical connection with the historian's own pre-
vious thoughts) that he has Q-A logic in mind, just as he had
described it in the Autobiography--short, that is, of an ex-
plicit discussion of the relation of questions and presupposi-
tions.7 Hence this passage confirms that as of 1930 Colling-
_________________
6 Cf. CRC, loc. cit., p. 17: "Thus it seems to me that
the individual 'proposition' assented to on any given occasion
is assented to only in a context, never by itself; and this
context is not a fortuitous context but a necessary one; I mean,
'It is not yet noon and the sun is shining' won't do as a
substitute for 'It is not yet noon and it is half-past eleven.'
The context is not (may I say?) a merely psychological context,
consisting of anything else that we may happen to be thinking at
the time; it is a logical context, consisting of other things
which if we didn't think we couldn't think what ex hypothesi we
are thinking."
7 This distinction would not be in the province of the

working historian anyway; and not even, according to the Essay
on Metaphysics, the task of the metaphysician. It is the task 6f
the logician. See EM, 54.
230
wood was making use of the principle of history which we have
labeled "ARCH-1," and two of the corollaries of "HIST-3" in our
summary of the Autobiography.8
Finally one finds an explicit statement of "ARCH-3" in the
1930 essay:
All history must be the history of something particular, and

the most we can ever do is to express the present state of
knowledge concerning this particular subject. As no history
can be universal, so no history can be final . . . . All
history is thus an interim report on the progress made in
the study of its subject matter down to the present; and
hence all history is at the same time the history of history
. . . . (EPH, 138).
Here again we have Collingwood apparently contradicting
himself, and then revising his previous statement. In the first
part of the paper he states that second-order reflection on
history, viz. the philosophy of history, must show how history
"is somehow a universal and necessary characteristic of things,
not merely a particular and contingent characteristic of a
certain group of things" (EPH, 122).
"The philosophy of something" is a legitimate phrase only

when the "something" in question is no mere fragment of the
world, but is an aspect of the world as a whole--a universal
and necessary characteristic of things. . . . If there is to
be a philosophy of history, history must be . . . a
universal and necessary human interest, the interest in a
universal and necessary aspect of the world. (EPH, 122-23).
______________________
8 It also, incidentally, states what Collingwood means by a
question "arising"--i.e., that the question has a "logical
connection with our previous thoughts, that we have a reason for
asking it . . . ." (SM, 137).
231
If history is concerned with a "universal and necessary aspect
of the world," then how is it that "all history must be the
history of something particular"? And how can all history be at
the same time the history of history without infinite regress?
The answer, insofar as one is possible on Collingwood's grounds,
is the same for both problems: "history" is an equivocal term,
with layers of meaning which it is the work of the philosopher
to distinguish--but to distinguish in a manner which cannot but
violate the expectations of someone committed to the logic of
the abstract universal. If one considers history in terms of its
limitations (both by its given evidence and by the historian's
own principles of interpretation) history always deals with
something particular; but if one regards history as that aspect
of the present which explains how that present came into being,
history deals with a universal and necessary aspect of the
world. This sort of distinction is only possible from the
philosophical point of view--the historian is not explicitly
aware of it. But Collingwood views these levels of meaning of
the concept of history not as several static definitions
separate from one another but as themselves part of a
developmental process in which the idea of history is
progressively realized.
In the intermediate part of the 1930 paper (clearly a
prototype of his later lectures on The Idea of History--cf. EPH,
xxxiii) Collingwood
232
shows how the step which elevates history to the rank of a
philosophical science was taken:
The essence of this development is the doctrine of the

individual judgement. Ordinary logic distinguishes the
individual judgement "This S is P." from the universal
judgement "All S is P." Now, says Croce, "This S is P" is
history, "All S is P" is science. But whenever we say "All S
is P" we have before our minds a "this S." . . . . "All S is
P" means "This S. in its character as S. is P." When the
element of individuality is taken away we have, not a
universal judgement "All S is P." but nothing at all. This
conception can be expressed by saying that all knowledge is
historical knowledge (individual judgement) and that science
is history with its individual reference neglected. (EPH,
135-36).
We shall have occasion in the sequel to return to this
theme in connection with some of Collingwood's later remarks
about historical and scientific judgments, about the concrete
universal, and about the structure of philosophical concepts.
But for now it is sufficient to notice that the formal opposi-
tion between universal and particular judgments is taken (by
Croce and, on approval, by Collingwood) to be overcome by the
individual judgment of history, which is not merely a particu-
larized universal judgment, but a judgment distinguishable from
a particular judgment as the base-line of its intelligibility
(when individuality is taken away we do not have a universal
judgment but "nothing at all").9 But such distinctions as these
are made by the philosophy of history which, as Collingwood was
so often to point out, is not something different or separate
from history itself, but is history with its presuppositions
made explicit (cf. EPH, 125).
___________________
9 This view is later criticized by Collingwood (IH, 303).
233
5. Conclusion.
In our survey of Collingwood's essays in the philosophy
of history between 1920 and 1930 we managed to find direct or
indirect evidence which substantiates his claims in the
Autobiography concerning the development of his ideas on the
relation of philosophy and history--with one notable oversight.
One principle that is totally missing in these essays is the one
we have called "HIST-4"--history as the science of human
affairs. It is the absence of this principle that mars his
discussion of the object of history, so that it seems that the
object of history is the whole of changing reality, and hence
"science is history with its individual reference neglected."
Now it so happens that not only is this principle the
central issue of his 1936 British Academy lecture, "Human Nature
and Human History" (IH, 205-31; cf. A, 116-17, n. 1), but the
principles of history which we have called "HIST-1, -2, -3, and
-4" appear almost verbatim in this essay. Even assuming that
Collingwood read the substance of his 1936 lecture back into his
earlier essays on the philosophy of history, and that therefore
the development of his ideas on the relationship of philosophy
and history was not as complete as he said it was as of 1930,
the real question that arises in understanding Collingwood's
early reconciliation between history and philosophy concerns the
manner in which
234
he finally distinguishes them within some sort of unity. For
what kind of unity would that be? History, science, art,
etc.--his "forms of experience" in Speculum Mentis--are not, he
insists, species of a common genus, "knowledge," as the realists
had asserted. That meant that each of the forms of experience
has an epistemology proper to it. At the same time Collingwood
feels obliged to show that all forms of experience are mani-
festations of a single process, mental activity, and are
therefore variations on a single theme.
We have already had occasion to notice that Collingwood
was not entirely successful in his several attempts to
distinguish history from other forms of knowledge solely on the
grounds of the manner in which they express their judgments. In
his 1922 essay on science and history he admitted that both of
these forms of experience make use of particular and universal
judgments and therefore cannot be separated on the basis of the
sorts of judgments they use. But in his 1930 paper he asserts
that the judgment peculiar to history is individual judgment
("This S is P") which he distinguishes both from the particular
judgment ("Some S is P") and from the universal judgment ("All S
is P"). But since judgments of science are merely the judgments
of history with their individuality suppressed, it follows (as
Croce had said) that "all knowledge is historical knowledge" and
that "science is history with its individual reference
neglected" (EPH, 136).

235
But then it would seem that science is a species of history, or
that both of these are (as Speculum Mentis seems to leave one
believing) unsuccessful forms of philosophy. In any event we
have the sort of genus-species situation within knowledge that
Collingwood had hoped to overcome. On the other hand the attempt
to distinguish different sorts of knowledge on the grounds of
their subject matter is equally unsuccessful. If history deals
with a "universal and necessary aspect of things"--i.e. their
"becoming"--then it is clear that it is indistinguishable from
the science of nature, since everything in nature is also
subject to a process of becoming (EPH, 12224). Starting from
either direction then, one arrives at the conclusion that the
forms of knowledge are indistinguishable from one another.
"Everything," Collingwood writes in 1930, "has a past;
everything has somehow come to be what it is; and therefore the
historical aspect of things is a universal and necessary aspect
of them" (EPH, 124). On this account there would not only be a
history of Greece and Rome, but of the San Andreas Fault and the
moon.
Perhaps Collingwood assumed that the reader would recall
what he had written in his 1923 essay on the New Idealism, where
he says that "that which has a history . . . . is a mind, for
matter may change, but it cannot be said to have a history" (FR,
274), and would supply this reservation as the distinguishing
feature of history. But he does not

236
tackle the issue explicitly in his published writings until his
British Academy lecture of 1936, where it is assaulted head on
and solved in the manner described in the Autobiography.
The thesis which I shall maintain ((writes Collingwood in

"Human Nature and Human History")) is that the science of
human nature was a false attempt--falsified by the analogy
of natural science--to understand the mind itself, and
that, whereas the right way of investigating nature is by
the methods called scientific, the right way of investi-
gating mind is by the methods of history . . . . Since the
time of Heraclitus and Plato, it has been a commonplace
that the entire world of nature is a world of "process" or
"becoming." But this is not what is meant by the histor-
icity of things; for change and history are not the
same . . . . (H)uman history shows change not only in the
individual cases in which . . . ideals are realized or par-
tially realized, but in the ideals themselves . . . . The
processes of nature can therefore be properly described as
sequences of mere events, but those of history cannot. They
are not processes of mere events but processes of actions,
which have an inner side, consisting of processes of
thoughts; and what the historian is looking for is these
processes of thought. All history is the history of
thought. (IH, 209-11, 215--emphasis mine).
Collingwood's solution to the problem then is to distinguish
between two classes of intentional objects—mere events and
"actions" of historical agents, the latter having deliberate
thought as part of its essential structure. Natural science
deals with the former, history with the latter.
But Collingwood is also saying something further in
this essay, and that is that the processes with which historical
inquiry deals are developments of a self-making mind (IH, 226).
What exactly "mind" is, and what part "thought" plays in the
life of men with minds, is something that Collingwood works out
in some detail in The Principles of Art and

237
The New Leviathan. But the notion of mind as a self-making
activity is explicitly present in Speculum Mentis, and goes as
far back as Religion and Philosophy.
In the latter Collingwood wrote, it will be recalled,
that "there is no such thing as human nature in the sense of a
definite body of characteristics common to every man," and that
therefore "the question to be asked about mind is not what it
is, but what it does; a question with which the logic of things
and qualities does not deal" (RP, 164-65; FR, 266). In Speculum
Mentis the doctrine of mind as a self-making activity is put
forward as the final realization of all knowledge:
Knowledge polarizes itself into abstract or erroneous and

concrete or true. Abstract knowledge is the same as error,
because, separating what is thought to be from what is, it
erects that which it thinks into a false object over against
itself, an external world . . . . But in concrete knowledge
the mind sees itself face to face, and knows even as it is
known. Here the object is the subject . . . in the sense
that the object finds its very life in being known by the
subject, the subject in knowing the object . . . . In an
immediate and direct way, the mind can never know itself: it
can only know itself through the mediation of an external
world, know that what it sees in the external world is its
own reflection. Hence the construction of external
worlds--works of art, religions, sciences, structures of
historical fact, codes of law, systems of philosophy, and so
forth ad infinitum-is the only way by which the mind can
possibly come to that self-knowledge which is its end. (SM,
313-15).
In Collingwood's view the process of creation of these "external
worlds"--of art, religion, science, history (as historiography),
and philosophy--is the very process of history, and the
historian in the very

238
best sense must be the one who can recreate all these processes
over again by "re-thinking" them in his own historical
imagination. In the process of doing so, and in confronting the
"evidence" of these worlds in a critical fashion, he not only
discovers the past to which he is heir but re-creates that past,
and in the process builds the structure of his own consciousness:
Man has been defined as an animal capable of profiting by

the experience of others. . . . The body of human thought or
mental activity is a corporate possession, and almost all
the operations which our minds perform are operations which
we learned to perform from others who have performed them
already. Since mind is what it does, and human nature, if it
is a name for anything real, is only a name for human
activities, this acquisition of ability to perform
determinate operations is the acquisition of a determinate
human nature. Thus the historical process is a process in
which man creates for himself this or that kind of human
nature by re-creating in his own thought the past to which
he is heir. (IH, 226).
That such a view, so eloquently expressed and so close to the
standpoint of European existentialism, would bring Collingwood
to loggerheads with behavioral psychologists requires no special
demonstration; we shall see in a future chapter how this aspect
of Collingwood's philosophy developed into the anti-positivistic
polemic in the Essay on Metaphysics. His antipathy to behavioral
psychology is neither something entirely new in his philosophy
nor is it inconsistent with its central tenets. We have already
seen that as early as Religion and Philosophy Collingwood

239
assaulted "psychology" as the pseudo-science of mind, since it
treats of mental activities as mere events (RP, 42; FR, 77). And
we have had occasion to notice that Collingwood's fundamental
view of human mentality makes deliberate activity (in a manner
yet to be explained) its central and irreducible aspect. But the
issue of immediate interest here is the support that one finds
throughout Collingwood's philosophy, and not merely in the
period following the appearance of the Autobiography, for the
reconciliation of history and philosophy through the concept of
mind as a self-making activity. Although "HIST-4" does not
appear explicitly prior to the 1934 British Academy Lecture, it
is a principle latent but operative in his early philosophical
writings, and takes no great feat of interpretive skill to
discern in his discussions concerning history.
It remains to pull all of these strands together into
some sort of summary of our survey of Collingwood's earlier
reconciliation of philosophy and history. What we have seen
emerging from Collingwood's essays on the philosophy of history
is a discussion which not only confirms most, if not all, of
what he had said about his discovery of the "principles of
history" as he described them in the Autobiography, but in ad-
dition a fuller discussion of several of these topics which
parallels and complements the autobiographical sketch. The main
ideas of this complimentary account can be summarized in tabular
form as follows:
240
TABLE 6
THE MEANINGS OF HISTORY
1. (EPH-1) "History" is an ambiguous term, the meanings of

which display a progressive and orderly development with
both a temporal and logical structure, the highest point of
which relates the diversity of these meanings in the unity
of a philosophical concept. (EPH, 39-40, 124-25).
2. (EPH-2) History understood as the ideal of the historian is

the knowledge of a unified and meaningful whole of acts of
historical agents, expressed in a single categorical
judgment. (EPH, 46).
a. The objective aspect of this ideal is the drama of

history as a single developmental process, which con
sists of the successive resolution of problems relat-
ing to human self-consciousness (in the form of art,
religion, science, etc.).
b. The subjective aspect of this ideal is the continuity

of this drama provided by the historian's own con-
sciousness, i.e. his expectation that the object of
history will be an organized and coherent whole. Min-
imally this is the view that history has a plot. (EPH,
37, 111, 137-38).
3. (EPH-3) History understood as the practice of the historian

consists of the two fundamental elements: historical data,
or evidence, and principles of interpretation. Each of these
limits the ideal anticipations of the historian. (EPH, 137).
a. Objectively, historical thinking is limited by the

evidence which provides the data for historical analy-
sis. Such data (i.e., traces of the past existing in the
present) requires critical analysis before serving as
evidence of (for or against) any historical thinking at
all. -TEPH, 136).
b. Subjectively, historical thinking is limited by the

historian's own "point of view"--the principles of
241
interpretation he brings to the data he is analyzing.

Such principles include the nature and limits of
negative evidence, the possibility of analogical ar-
gument, views of history as cyclic or linearly pro-
gressive, etc. (EPH, 89, 137).
4. (EPH-4) At the level at which theory and practice (and

subject and object) are identified, the historian is also an
historical agent, not only observing history but making it.
This is the philosophical concept of history, in which
subject and object are identified--where, in the case of
history, the historian is both agent and patient. (EPH, 46,
49, 120).
5. (EPH-5) The philosophical concept of history is concerned

with a fully articulated answer to the question, "How has
the world (of human affairs) as it now exists come to be
what it is"? The answer to this question is conceived as the
first stage of an activity to either keep it that way or to
initiate an alteration of it. (EPH, 89, 92, 102, 109, 120;
cf. EPA, 144).
242
One must pay careful attention therefore to which level
of significance the particular use of the term "history" refers
in any given passage in Collingwood's writings. To accuse
Collingwood of "radical historicism" in his later years (as
critics do who read the Essay on Metaphysics without seeing it
in the context of his other writings on history) may be either
utterly true or totally false, depending on what one understands
by "history" and "historicism." As a philosophical concept,
history is identified with philosophy, and therefore the
reduction of philosophy to history is the identification of
philosophy with itself--a harmless tautology. On the other hand
as an empirical concept (the collection of events which
Collingwood calls raw data, or the uninterpreted evidence, of
history), history is not only not identical with philosophy, but
can hardly be said to be identical with itself.
Nevertheless in our examination of Collingwood's later
philosophy of history we shall have to attend to the manner in
which he tries to retain a criteria of historical objectivity
while yet rejecting the realistic historian's emphasis on
extra-mental "facts." But meanwhile we must turn now to what
Collingwood understood by his "identities" in conjunction with
the issue of what he expected of a "rapprochement" philosophy,
and therefore what method he used to achieve this end.

CHAPTER VI
RAPPROCHEMENT, RELIGION, AND ABSOLUTE KNOWLEDGE
1. Introduction.
Just as in his discussions of Q-A logic and the
philosophy of history, Collingwood made his approach in the
Autobiography to the discussion of a need for a new branch of
philosophy, devoted to the work of reconciliation, from a
critique of realism.
It was during World War I and his daily reflections on
the Albert Memorial that he discovered the "vulgar error" of the
realists--their false belief that problems of philosophy were
eternal, and that different philosophies were different attempts
to answer the same questions (A, 60). The first instance he
cites of his own clear recognition of the truth about eternal
problems came in political theory, where it became obvious to
him that the political theories of Plato and Hobbes are not two
theories of the same thing--the "nature of the state"--because
there were genuine differences between the ideal states (and not
just the empirical ones) being discussed in each case, ideals
that are ignored in treating them as if each intended the same
thing (A, 61). There is a relation between them, but the
relation is not the sameness of an abstract
243
244
universal, but the sameness of an historical process; and the
difference between one thing which in the course of that process
has turned into something else (A, 62). Collingwood went on, he
says, to apply this "clue" to the problems of ethics and
metaphysics, and in the case of the latter it bore fruit in his
ultimate discovery that metaphysics is an historical science (A,
63-65). And finally, he found that his discovery could be
generalized to the whole field of philosophy: "By degrees I
found that there was no recognized branch of philosophy to which
it did not apply" (A, 67).
Later in the Autobiography (in Chapter XII, "Theory and
Practice") Collingwood picks up the discussion of rapprochement
again and spells out the nature of his reconciliation in ethics.
In addition to his reconciliation of history and philosophy (as
instanced in his reflections on the political notion of the
state) he was also working on a rapprochement between theory and
practice in order to counteract the "moral corruption propagated
by the 'realist' dogma that moral philosophy does no more than
study in a purely theoretical spirit a subject matter which it
leaves wholly unaffected by that investigation" (A, 147). In so
doing he subjected the familiar topics and problems of moral
philosophy first to an historical treatment (to show,
presumably, how the problem developed as an historical process
wherein the differences were truly integral to it and changed
the nature of the problem as it developed), and secondly to an
analytic treatment (which, by critical argumentation,

245
both differentiated the problem into its correlative parts, and
then showed them to be aspects of one and the same identical
entity). Naturally enough in a chapter on theory and practice,
the example Collingwood used was one of action--first
distinguished as moral, political, and economic, and then
unified insofar as every action is moral, political, and
economic--not separated into three classes, but seen as three
characteristics, distinguished and yet united in any actual act
(A, 148-49).
We have recalled these passages here to remind the
reader not only of the anti-realist context of Collingwood's
remarks about rapprochement, but also to recall the two phases
(historical and analytic) of his earlier reconciliational
philosophy. We have said enough in the previous section about
the historical phase of his rapprochement methodology to give
the reader a sense both of how he felt it should be carried out
and what its intrinsic and unresolved problems were. In this
section, therefore, we shall concentrate on Collingwood's
"analytic" phase. In doing so we must pay heed to one further
proviso:
These were the lines on which I treated the subject in my

lectures of 1919. I continued to lecture upon it yearly
during the whole remainder of my life at Pembroke College,
with constant revision. The scheme I have just described
obviously represents a stage in my thought at which the
rapprochement between history and philosophy was very in-
complete . . . . The rapprochement between theory and prac-
tice was equally incomplete. I no longer thought of them as
mutually
246
independent; I saw that the relation between them was one of

intimate and mutual dependence, thought depending upon what
the thinker learned by experience in action, action
depending upon how he thought of himself and the world . . .
. But this was only a theoretical rapprochement of theory
and practice, not a practical one. (A, 149 50).
If we are to approach this topic in Collingwood's early writ-
ings, therefore, in the spirit of the Autobiography, we must not
only take into account the sense of the term "analytical" as
used by Collingwood, but also the fact that he regarded this
"analytic" to be incomplete. How it was completed (and the sense
in which his early analytic was incomplete) Collingwood leaves
to the reader to work out for himself (A, 149).1
The evidence from Collingwood's writings is insufficient
to establish what the actual topic was to which he first applied
his rapprochement method; although he lectured on ethics and
political philosophy since 1919, his first publications
specifically on these subjects were not until six years later
(cf. "EPS," 1925), and by then both Religion and Philosophy and
Speculum Mentis had appeared. While it is true that he treated
the subjects in ethics and political science in these later
publications much as he says he did in the Autobiography, for
our
_____________________
1 The only further clue the reader has from the
Autobiography is Collingwood's unreserved praise for his Essay
on Philosophical Method, which he recommends as his "best book
in matter; in style, I may call it my only book" (A, 117-18).
The reader must conclude that as of 1933 rapprochement metho-
dology was fairly complete.
247
purposes the earliest instance, in his published writings, of
rapprochement is in the area of religious philosophy. Not only
is religion as a "unified life of all the faculties" the first
published topic on which Collingwood exercised his early
reconciliational technique, but it provides us with an index as
to how that technique was modified between 1916 and 1924, when
Collingwood once again dealt with the subject in Speculum Mentis.
2. Religion, Philosophy, and the Incomplete Rapprochement.
At the time of his writing of Religion and Philosophy
(published in 1916, but written some time earlier--see RP, v)
the analytic phase seems to be only recourse Collingwood re-
sorted to in his treatment of philosophical problems, the his-
torical phase being only implicit in what he was saying, not in
what he actually did.2 We have already seen an example of this
analytic phase in Collingwood's treatment of the identity of
philosophy and history. Consequently we expect him to begin with
a pair of concepts presumed to be isolated or "separated" from
one another. In
__________________
2 Actually Speculum Mentis, for all its discussion of
history and its relation to philosophy, does not actually dis-
play in its compositional format any more concern for the his-
torical development of its subject matter than does Religion and
Philosophy, except that the former has that same vague reference
to historical development of forms of experience as does Hegel's
Phenomenology of Mind. But cf. SM, 50-55 (an "ages of man"
analogy), and 21-38.
248
this case the reconciliata (the term we shall henceforth use for
viewpoints to be held together in the relationship of
rapprochement) are religion and philosophy. But from the first
words of the book we are aware that the orientation of the
discussion to follow is primarily philosophical:
This book is the result of an attempt to treat the Christian

creed not as a dogma but as a critical solution of a
philosophical problem. Christianity, in other words, is
approached as a philosophy, and its various doctrines are
regarded as varying aspects of a single idea which, ac-
cording to the language in which it is expressed, may be
called a metaphysics, an ethic, or theology. (RP, xiii).
Collingwood also specifies the sense in which he understands the
terms "religion" and "philosophy," at least for the purposes of
discussion in this book:
Just as every man has some working theory of the world which
is his philosophy, some system of ideals which rule his
conduct, so every one has to some degree that unified life
of all the faculties which is a religion . . . . We apply
the term religion to certain types of consciousness, and not
to others, because we see in the one type certain
characteristics which in the others we consider to be ab-
sent. Further investigation shows that the characteristic
marks of religion, the marks in virtue of which we applied
the term, are really present in the others also, though in a
form which at first evaded recognition. (RP, xvii-xviii).
Collingwood did not follow up these clues about the
presence of
249
one form of consciousness in other forms until Speculum Mentis,3
and while it is interesting to note that the germ of his later
doctrine of a "scale of forms" is present in this early
discussion of the nature of religion, in Religion and Philosophy
the strategy he follows is based not on the construction of a
phenomenology of forms of experience, but rather on the analytic
method of his earlier (and still incomplete) rapprochement.
Therefore he begins with what already amount to theories of
religion, but theories which treat of religion as if it had no
direct relationship to philosophy. Furthermore it is to be noted
that Collingwood does not start out with a definition or
description of religion that omits its philosophical aspect
(e.g. "primitive religions") and then try to derive philosophy
from this description. He starts out rather with
"anti-intellectual" theories of religion--religion viewed as
(mere) ritual, as conduct, or as feeling--that is, theories that
deny that religion has an intellectual element and therefore
deny that it is
______________________
3 In addition to examining the relations of religion to
philosophy, history, and conduct (morality), Collingwood
recognized a "fourth question" concerning the relation of re-
ligion to art, but declined to deal with the issue in Religion
and Philosophy, promising to discuss the nature of metaphor,
prose, and the philosophy of language (all being in the province
of art) in a future volume (RP, xvi). In Speculum Mentis he
recalls this promise, and indicates to his readers the extent to
which he felt the earlier work to be deficient (SM, 108, n. 1).
250
identical with philosophy. Each of these theories is then
subjected to a critique, and in the process of denying the
premise (that religion is mere ritual, mere conduct, or mere
feeling, each without an intellectual element) he affirms the
consequence that creed is not nonessential to religion.
(a) The claim of religion to be mere ritual is based on
anthropology's examination of religions of lower culture, which
purport to show that ritual is prior to creed. But Collingwood
argues that aside from the issue of what relation there is
between primitive and modern religions (which may be one of
analogy only), the theory still lacks an account of how ritual
practices arise in the first place. The necessary nature of
ritual implies a grounding in fears, and the ritual act is
performed because the primitive people assume the universe to be
governed by certain powers, and that their acts will somehow
please or influence these powers in ways beneficial to the
tribe. But such a judgment about the nature of the powers that
govern the universe is a primitive theory, and the belief that
acts will influence that power is creed. (RP, 6-7; FR, 45-46).
(b) Similarly the view that holds that doctrine has
little or no bearing on conduct, and therefore holds that re-
ligion is primarily a system of rules guiding conduct, ignores
the fact that action relies on knowledge of the situation which
calls for application of the rule. There

251
is no such thing as conduct divorced from knowledge or knowledge
divorced from conduct. Furthermore the conduct that is being
recommended presumes a judgment that certain sorts of conduct
are good and others bad, which is a judgment based on creed: it
is the presumption of the truth of the moral creed that results
in the good act. (RP, 7-9; FR, 46-49).
(c) Finally, religion as pure feeling trades on the
ambiguities of the word "feeling," which in some cases (as when
it refers to very indefinite and indistinct states of mind)
seems to rule out any truly intellectual element, but in others
(as when it implies absolute and positive conviction coupled
with an inability to offer proof or explanation of the
conviction) it does not rule out knowledge per se but actually
means the same thing as knowledge, albeit unreasoned. But in the
first case one could hardly call such feeling religious, because
it does not hold to any truth at all, being too indistinct and
indefinite (not all indefinite states of mind are therefore
religious, nor are religious states of mind indefinite). And if
feeling just means emotion, it must at least admit to a kind of
emotion that is appetition for the desired thing--in this case,
God--and consequently presumes some knowledge of God as
desirable. In any event religious feeling requires some
intellectual element. (RP, 11; FR, 49).

252
It can be seen from these arguments that if there is any
"reduction" of religion that is to occur, at least Collingwood
will not be content to let this be the humiliating reduction of
religion to a lesser form of experience. Whether he allows for a
kind of exalting reduction by assimilating it to a "higher"
realm remains to be seen. But these passages (and especially
when taken with the perceptive analysis of Christian beliefs
concerning the fall and redemption of man, in the concluding
chapters of the book) show a refined sense of awareness of the
subject matter of religion, and an unwillingness to have it
sullied by inappropriate and reductive comparisons.
Now at this point the reader would expect (if the "an-
alytic" follows the pattern of the discussion of the relation of
philosophy and history) a criticism of the view that philosophy
has no religious element, the implication being that if it is
not the case that philosophy does not have a religious element,
then philosophy must have an essential religious dimension just
as religion must have an element of philosophy. In other words,
insofar as philosophy is essential to religion and religion to
philosophy, religion and philosophy mutually imply one another
and are hence identical (that is, nonseparate). But what one
finds instead in Religion and Philosophy is an identification of
the two by their object, coupled with a denial that there is any
such thing as a separate "philosophy of religion"--separate,
that it, from theology

253
or creed.
(T)he theory of knowledge or logic does not consider dif-

ferences of the object, but only processes of the subject;
and therefore there is no distinction between the philosophy
of religion (as theory of religious knowledge) and the
theory of knowledge in general. If there is a general
philosophy of knowing, it includes religious knowledge as
well as all other kinds; no separate philosophy is
required . . . . If the philosophy of religion is
indistinguishable from philosophy as a whole, what is the
relation of philosophy as a whole to religion or theology?
Philosophy is the theory of existence; not of existence in
the abstract, but of existence in the concrete; the theory
of all that exists; the theory of the universe . . . . Now
if philosophy is the theory of the universe, what is
religion? We have said that it was the theory of God and of
God's relations to the world and man. But the latter is
surely nothing more or less than a theory of the
universe . . . . Religion and philosophy alike are views of
the whole universe . . . . If religion and philosophy are
views of the same thing--the ultimate nature of the
universe--then the true religion and the true philosophy
must coincide . . . (RP, 15-18; FR, 53-55).
We have already seen that Collingwood identified history and
philosophy in much the same way--as both dealing with "knowledge
of the one, real world." This occurs, in fact, in a chapter in
which the relation between history and religion is being
discussed, and the implication towards which the reader seems to
be compelled is that since religion and philosophy are
identical, and since history and philosophy are identical,
religion is identical with history. The conclusion of the
chapter even suggests that science is also not anti-religious,
and therefore, by implication, identical with religion, phil-
osophy, and history.

254
It goes without saying that such a Mulligan stew of
subjects is more than a reader can be expected to digest. To
simply identify religion and philosophy seems not only to lead
to the confused state wherein what can be said as proper to each
is now predicable of the other (as if one could resolve an
honest dispute about some aspect of reality by engaging in
prayer or sacrifice, or that one might celebrate ritual syllo-
gism on a Sabbath morning and thereby appease a reasoning god),
but renders suspect the autonomy of all other forms of thought
as well. If two kinds of knowledge are identified by reference
to their object, and if there is no distinction in the processes
of the subject whereby a "philosophy of" something and a theory
of knowledge in general may be distinguished, then there seems
to be no remaining way of distinguishing between two ways of
knowing the same thing, and consequently between science,
religion, history, and philosophy. But then on what grounds can
one even distinguish between subjective processes of knowledge
and objective correlates of these processes?
These consequences of his position are not altogether
evaded by Collingwood, as we see in a later chapter of Religion
and Philosophy where he argues as follows:
My imagination of a table is certainly a different thing

from the table itself, and to identify the two would be to
mistake fancy for fact; but my knowledge of the table, my
thought of it in that sense, is simply the table as known to
me, as much of the table's nature as
255
I have discovered. In this sense, my "thought about" the

table-what I think the table to be--only differs from the
table itself if and in so far as I am ignorant of the
table's real nature. My thought of the table is certainly
not something "like" the table; it is the table as I know
it. Similarly, your thought of the table is what you know of
the table, the table as known to you; and if we both have
real knowledge of the table, it seems to follow that our
thoughts are the same, not merely similar; and further, if
the mind is its thoughts, we seem to have, for this moment
at least, actually one mind; we share between us that unity
of consciousness which was said to be the mark of the
individual. (RP, 100-01; FR, 172-73).4
Even with the proviso that in life "real knowledge" or the
knowledge of the real nature of something (even artifacts,
judging from the example) is seldom if ever perfect, so that "in
a sense, no two people ever do, or ever could, think or will
exactly the same thing" (RP, 106; FR, 177), such an assertion
puts an intolerable strain on the notion of identity. For even
if my knowledge of the table were "real knowledge," it would
seem strange to assert that my concept of the table is subject
of predication in the same sense that the table is--and
therefore capable of supporting
_________________________
4 In a footnote to this passage Collingwood denied that
his argument placed him in an idealistic stance: "I believe that
the argument I have tried to express contains little if anything
which contradicts the principles of either realism or idealism
in their more satisfactory forms." (RP, 101; FR, 173). It is
interesting to note that in his later philosophy when he puts
forward his theory of historical re-enactment, which makes a
similar claim to the identity of mind as in the present
paragraph, he also juxtaposes it against both realists and
idealists--see IN, 282-302.
256
articles placed on it, capable of physical destruction with
sufficient mechanical force, etc. And to assert that we never
have "real knowledge" of anything seems to grant so much ground
to scepticism that the argument in point seems futile.
Since Collingwood's rapprochement program is precisely
one of showing that two forms of thought are not "separate" but
rather are "the same"--whether that sameness be the sameness of
an historical process or the sameness of question to which they
are the responses--it is clear that at the root of
reconciliation philosophy is a notion of identity that on the
face of it (from our above example) is under disruptive
pressures which threaten to split it assunder. For (a) if "the
same" means merely "not different," then when two forms of
thought are said to be "the same" they are being said to be "not
different," and in their indifference they are not only no
longer two forms of thought (and so the comparison is fatuous),
but they are also indistinguishable from nothing at all;5 but
(b) if "the same" means "similar in some, but not all
characteristics" then when two forms of thought are said to be
"the same" they are being
________________________
5 The situation is aggravated rather than ameliorated in
the case of the forms of knowledge, since Collingwood specifies
their objects, in each case, to be the "whole universe" or "all
of existence," which leaves no room a parte objecti for a
distinction between them. And since the forms of knowledge have
been identified with each other and with philosophy, a
distinction a parte subjecti) is also impossible.
257
said to be "similar and dissimilar" in specifiable ways--which
is precisely what Collingwood seems to be denying.
But lest we be charged with a case of unfair pillory, we
hasten to add that Collingwood shows even in Religion and
Philosophy that he is aware of the difficulty, and goes so far
as to try to meet it head on. In the same chapter from which the
table example is drawn (a chapter, incidentally, which deals
with inter-subjective identity as the grounds of possibility for
religious unity of God, man, and universe) Collingwood
distinguishes between an abstract and a concrete unity, and says
that "the unity whose possibility we are concerned to prove is
the fully concrete identification, by their own free activity,
of two or more personalities" not as a universal condition but
as an ideal (RP, 106-07; FR, 178).
A person is undoubtedly himself, and can never help being

himself, whatever he does; but this merely abstract unity,
this bare minimum of self-identity, is much less than what
we usually call his character or personality. That is rather
constituted by the definite and concrete system of his
various activities or habits . . . . The same distinction
applies to the unity of a society. In one sense, any kind of
relation between two people produced a kind of social union
and identification; in another sense, only the right kind of
relation unifies them, and a different relation would
destroy the unity. In the first case, their union is what I
call the purely abstract unity; in the latter, it is the
concrete unity that has to be maintained by positive and
harmonious activity. (RP, 107-08; FR, 178-79).
Aside from the question of what the "right kind of relation"
might be, or by what criteria one decides rightness and non
rightness of relations,
258
there is still the puzzling and as yet unresolved question as to
what the distinction between abstract and concrete unity refers
to. Collingwood's answer seems to rely on a modification of the
relation of whole and part:
But is unity the same as identity? There seems at first

sight to be a very decided difference between saying that
two things are part of the same whole, and saying that they
are the same thing; the parts of one thing seem to be
themselves quite separate and self-existent things, possibly
depending on each other, but each being what it is itself,
and not the others; while the whole is simply their sum. We
have already expressed doubts as to the strict truth of this
conception . . . . (I)f a whole was to be knowable, it must
be of such a kind that the parts are not simply added in
series to one another, but interconnected in such a way that
we can somehow say that each part is the whole. In that case
each part would also be in a sense the others . . . . Each
part has its own nature, its own individuality, which is in
the strictest sense unique; and apart from the contribution
made by each several element the whole would not exist.
Change one part, and the whole becomes a different whole.
Not only does the whole change, but the apparently unchanged
parts change too. (RP, 108-10; FR, 179-80).
Collingwood uses the instance of any whole consisting of three
parts, x, y, z--whether that whole be a machine with three
working parts, a society of three members, a stanza of three
lines, or a syllogism containing three propositions. In such a
system the definition of the part x can only take the form of a
definition of the whole xyz, since the "thing" itself is "only a
relation, an interchange, a balance between the elements which
at first we mistook for its parts" (RP, 112-13; FR, 182-83).

259
The attempt to evade this analysis of relations and
relate by pointing to the difference between what a thing is in
itself and what it is in relation to its context or the whole of
which it is a part is also, Collingwood says, unsuccessful,
since "the character or self of a thing, what it is, cannot be
distinguished from its relations" (RP, 112; FR, 182) Even though
its "internal relations" seem not to change when a change of
context or of its "external relations" occurs, it is impossible
to deprive a thing of every context (as one would presumably be
forced to do to prove that what it is in itself is not affected
by its context), so that one can do no more than to replace one
context with another.6 Whether the "context" be spatial or
temporal does not greatly affect the argument, since "the
history of a thing in the past and its capabilities for the
future are as real as its present situation" (RP, 111; FR, 182).
But even this analysis of the relationship of parts to
whole, although presumed to be correct and even essential for
his argument, is regarded by Collingwood as a variety of
abstract unity. It is merely the lowest possible sort of whole,
the necessary, abstract unity
____________________
6 Collingwood seems unaware that his argument at this
point begs the question: to "change a context" does not ordin-
arily mean to deprive a thing of all contexts, so to assert the
impossibility of the latter does not affect the argument that
there is a difference between what a thing is in itself and what
it is in relation to its context.
260
of its elements. It is not yet the "contingent unity of
co-operation" between consenting minds that is required for a
community of persons. Since Collingwood wishes to show that all
personalities are components of a whole (the universe) and
therefore "necessarily identified with each other and the whole,
that is, with the universe considered as homogeneous with them,
an absolute mind, God" (RP, 114; FR, 184), this minimum,
abstract unity, although essential, falls short of the mark; it
leaves us with an abstract God.
(T)he error lies in mistaking this fundamental assumption

for the final conclusion; in assuming that this elementary,
abstract unity is the only one which concerns us . . . . To
call this formless and empty abstraction "the Absolute" is
merely to abuse language; and to suppose that this is all
philosophy has to offer in place of the concrete God of
religion is completely to misunderstand the nature and aim
of philosophy . . . . The Absolute . . . is not a label for
the bare residuum, blank existence, which is left when all
discrepancies have been ignored and all irregularities
planed away . . . . A real philosophy builds its Absolute
(for every philosophy has an Absolute) out of the
differences of the world as it finds them, dealing indi-
vidually with all contradictions and preserving every detail
that can lend character to the whole . . . . The formless
and empty Absolute of this abstract metaphysic perished long
ago in the fire of Hegel's sarcasm . . . ((as)) the
pseudo-Absolute, the "night in which all cows are black" . .
. . (RP, 114-16; FR, 184-86).
We shall return in a moment to what Collingwood understands by
"the Absolute," but for now we wish to show that in contrast to
the abstract Absolute Collingwood proposes a "concrete identity
of activity":
261
A mind is self-identical in this sense if it thinks and

wills the same things constantly; it is identical with
another, if it thinks and wills the same things as that
other . . . . Now these two cases are typical first of the
self-identity of God, and secondly of his identity with the
human mind . . . . Further, this divine mind will become one
with all other minds so far as they share its thought and
volition; so far, that is, as they know any truth or will
any good. And this unity between the two is not the merely
abstract identity of co-existence, but the concrete identity
of co-operation. (RP, 116-19; FR, 186-88).
In Religion and Philosophy, however, Collingwood at-
tempts little more than to show how such a unity is possible;
since he has not built his Absolute "out of the differences of
the world" as he found them, but rather built them out of ab-
stract concepts by dialectical analysis, his philosophy of re-
ligion at this stage remains an unfulfilled promise. Furthermore
the reader is left unsatisfied how there can be "other minds" at
all, or how one distinguishes between the mind of God and the
minds of men when they are thinking and willing the same thing.
We have covered a lot of ground since initiating our
discussion about Collingwood's rapprochement philosophy, and it
may be helpful here to pause and survey the territory. (1) We
have seen that Collingwood claimed in the Autobiography that
prior to 1932 he had only partially worked out his rapprochement
philosophy, to the point where he saw that the relation between
reconciliata was one of "intimate and mutual dependence." (2) As
the first published example of this

262
partial reconciliation we took up the analysis of religion and
philosophy in the book of that name, his first publication. (3)
In contrast to the reconciliation of philosophy and history
which we examined in the previous section (and found to involve
an argument that showed history and philosophy to mutually
implicate one another, and therefore to be "identical"), the
reconciliation of religion and philosophy took a more complex
and circuitous route. (4) The strategy of the latter argument
turned out to rely on the epistemological assumption that forms
of knowledge, and even minds, which intend the same object are
identical--thereby identifying religion and philosophy as forms
of knowledge about "all of existence" or "the universe," and
incidentally identifying all other forms of knowledge as well.
(5) Collingwood recognized that this was a kind of abstract
identity, and attempted to correct this abstract notion of
identity with a discussion of "concrete identity"--which turns
out to mean a contingent unity of the activity of co-operation
between minds. (6) Such an identity is the only true Absolute
sought by philosophy and religion alike; therefore they are
reconciled by sharing a common ideal object.
On route from reconciliation to Absolute we discovered a
number of subsidiary issues, which we shall find of importance
in considering Collingwood's further development of the notion
of rapprochement. (1) In Religion and Philosophy, at least,
Collingwood displayed both a remarkable

263
sensitivity for the issues confronting persons who attempt to
reflectively understand their religious consciousness, and an
equal obteuseness about the "division of knowledge" into
provinces, to the point where the forms of knowledge--science,
history, art-appear as species of the genus, "knowledge"
(although he never explicitly says this). (2) But in trying to
sort out the confusion that results from identifying forms of
knowledge with each other, we discovered that for Collingwood
there are levels of meaning for the concept of identity, just as
there were levels of meaning to the concept of history. So when
two reconciliata are said to be "identical" or "the same" one
must carefully attend to which sense of "identity" is being
used. We have observed at least three uses of the term so far:
(a) a bare, abstract identity in which terms are related only as
members of a class or genus; (b) a (dialectical) identity of
whole and part, in which relations and relate are so intimately
connected that any change in a part necessarily implies a change
in the relations and in the whole, and vice versa; and (c) a
concrete identity of mental activity, in which persons may co-
operate in thinking truth or in doing good. The first two in-
volve relations of necessity, the third a relation of contin-
gency. The first two are also abstract, the third is fully
concrete. (d) While the self-identity of the Absolute mind, or
of God, is not specified as a fourth kind of identity, it is at
least put forward as the limiting case

264
of the concept of identity--the self-identity of truth, in all
its diversity, with itself (RP, 117-20; FR, 186-190). (3) For
each sense of the term identity there corresponds a sense of the
term "separate"--so that there will be (a) separate members of a
class (abstract separation--entities treated as mere instances
of a class), (b) separate elements of an organic whole
(dialectical separation), and (c) personal separateness
(concrete separation--if the unity of cooperation breaks down,
for example). We would have to presume that a separateness or
disunity in the self-identity of the Absolute mind or God would
be impossible; the disunity would occur between God and man, and
would fall into the category of concrete separation.7 (4)
Finally the reader is left to his own devices to decide what
would constitute an instance of the first sense of the term,
"identity," as well as its corresponding sense of separation,
insofar as Collingwood's examples from the second meaning (what
we have called dialectical identity and separation) --namely a
three-part machine, society, poem, or syllogism (he even refers
later to the stones forming the arch of a house)--seem to leave
little room for an example of something that is not an "organic
whole."
______________________
7 This issue is taken up by Collingwood in the conclud-
ing chapters of Religion and Philosophy in the discussions on
evil, the self-expression- God in man (in the person of the
Christ), and God's redemption of man (RP, 122-93; FR, 192-211,
251-69). These are interesting and occasionally profound
chapters and utterly repudiate the thesis that Collingwood had
neither interest nor insight into the nature of religious
consciousness.
265
3. Speculum Mentis: Rapprochement and Developing Series.
Some of these themes and the issues they raise appear
again in Speculum Mentis, and others in essays published between
1916 and 1926. In a footnote in Speculum Mentis Collingwood
proposes that his chapter on religion be taken as a correction
of the views he put forward in Religion and Philosophy, and says
his "mea culpas" for the oversights of that earlier work:8
With much of what that book contains I am still in agree-

ment; but there are certain principles which I then over-
looked or denied, in the light of which many of its faults
can be corrected. The chief of these principles is the
distinction between implicit and explicit. I contended
throughout that religion, theology, and philosophy were
identical, and this I should not so much withdraw as qual-
ify by pointing out that the "empirical" (i.e. real but
unexplained) difference between them is that theology makes
explicit what in religion as such is always implicit, and
so with philosophy and theology. This error led me into a
too intellectualistic or abstract attitude towards re-
ligion, of which many critics rightly accused me; for in-
stance . . . I failed to discover any real ground for the
distinction not only between man and God, but between man
and man . . . . (SM, 108, no. 1; emphasis mine).
The way in which Collingwood revised the conception of his forms
of knowledge and their relations is based on a more careful
analysis of the language of art, religion, science, and history,
as well as on a
____________________
8 The fact that Collingwood here and elsewhere expli-
citly acknowledges changes in his position on topics discussed
in previous writings should alert the reader to beware of any
claims by overzealous interpreters to find a radical consistency
in all of his writings. It should also, however, be a warning to
all who claim a "radical conversion" where no such change of
heart is acknowledged explicitly by Collingwood. For another
explicit acknowledgement, see PA, 288, n. 1).
266
careful delineation of what the claims of any one form are
(explicitly) as distinct from what the limitations of these
claims are (implicitly) when seen from a "higher" viewpoint.9 He
withdraws his previous assertion that epistemology (in the guise
of logic) "is a masterscience having jurisdiction over the whole
field of knowledge"--which he now regards as "pure
intellectualism" (SM, 49); and he condemns all identifications
of philosophy with religion, science, art, or history as "barren
____________________
9 The "implicit-explicit" distinction, although singled
out by Collingwood as very important and the essential differ-
ence between the argument in Speculum Mentis and that of
Religion and Philosophy, is difficult to define in terms general
enough to be acceptable to all levels of the scale of forms of
knowledge as they are presented in the later work. Collingwood's
characterization of the distinction tends to be in terms of
mental dispositions. "In any given experience," he writes,
"there are certain principles, distinctions, and so forth of
which the person whose experience it is cannot but be aware:
these I call explicit features of the experience in question . .
. . On the other hand, an observer studying a certain form of
experience often finds it impossible to give an account of it
without stating certain principles and distinctions which are
not actually recognized by the persons whose experience he is
studying . . . ((for example)) theology makes explicit certain
principles which are implicitly, but never explicitly, present
in religious consciousness; and in general what we call
philosophy reveals explicitly the principles which are implicit
in what we call everyday experience" (SM, 85, n. 1). However in
the course of the argument it becomes clear that what may be
implicitly assumed in one form of experience (e.g. religion) may
become explicit either in the next form (e.g. science) which
appears as its successor in the developing series, or by the
"dogmatic philosophy" of that form itself--the philosophy which
is engendered by a form of knowledge when it attempts to justify
itself as a total outlook on reality. Roughly, then, the
"implicit-explicit" distinction is the distinction between
something presumed and something asserted by the expression of a
form of experience.
267
abstractions" (SM, 246; cf. FR, 91). Having served his time on
the Procrustean couch of logic, Collingwood is bent upon
correcting his previous mistakes, and rises now to the occasion
by roundly condemning the view that the five forms of experience
are species of a genus, which may be substituted for one another
or taken up in any order. They now form a series, and in this
series there is an element of denial and distinction:
(O)ur forms of experience are not mere species of a genus,

because each denies the others; and because they are not
species they have not that indifference with regard one
another which characterizes abstract logical classifica-
tions. They must form an order of some kind . . . . But
what is even more important than the actual order is the
suggestion of serial arrangement as such. For a series of
terms implies that each term is as it were built upon or
derived from its predecessor and therefore does not start in
vacuo, is not wholly fresh embodiment of the universal, but
is essentially a modification of the term before. Hence
even if we only recognized three terms, and made a series
by alternating them, abcabcabc . . ., there would be no
repetition, for the second a would be not the mere first a
again, but a modified by having been developed through b
out of c; the third a would be a modified by the same
process in the second degree; and so on. (SM, 55; cf. SM,
206-07).
We now see a glimmering of an answer not only to the
question of what the way was in which Collingwood modified his
original rapprochement philosophy, but also the way in which the
"sameness of question" turns out to be not the sameness of a
universal but the sameness of an historical process. The
modification consists in the addition of a notion of what Col-
lingwood would later call a "scale of forms" to the analytic

268
phase of his di-phasic (now tri-phasic) method of reconcili-
ation. We shall see in a moment the radical way in which this
modifies the notion of identity; but just on the analogy of his
three-term system, it is clear from the comparison of this
system with that offered in Religion and Philosophy that there
is a shift of emphasis here from the mere assertion that the
change of relations in a three-term system implies a change in
the relate (in a system of dialectical identity) to the
specification of how that change affects the relationship. The
terms are related not merely "internally" (Collingwood avoids
the "internal-external" metaphor in Speculum Mentis) but in an
ascending series, in an order of development, such that when a
term changes it is changed through the agency or mediation of
another term in the series into something further which
incorporates the others (its predecessors) into itself. But this
is the way that Collingwood had defined historical change in his
1927 and 1929 essays on the philosophy of history: history is a
process of spiritual development, a dynamic rather than static
concept, in which ideas grow out of other ideas and modify these
previous ideas while being modified by them (EPH, 73). The
manner of identification of the forms of knowledge and the
sameness of historical problems converge on the concept of a
developmental series.
Collingwood is quick to reassure the reader that al-
though the forms of experience and the stages of history display
a developmental
269
structure, the life of mind is "not the rotation of a machine
through a cycle of fixed phases," and consequently one should
not expect to have his nose held to a "dialectical grindstone"
(SM, 56). It will be recalled from Religion and Philosophy that
the "identity of cooperation" (which we now recognize as the
identity of historical processes) was regarded as a contingent
identity, and not a necessary one. Consequently it is not
surprising that at the very beginning of a work which is to
spell out the stages of development of a mind through successive
forms of experience, stages which parallel the development of
consciousness in mankind as a whole (cf. SM, 50-54), Collingwood
should feel bound to deny that the relation between phases is
one of compulsion or necessity--and this importation of
contingency is a measure of Collingwood's distance from Hegel.
On the other hand (as will be recalled from our previous
discussion of Speculum Mentis)--the test of each of these forms
of experience (each of which claims not only to give the truth
but "to give the absolute or ultimate truth concerning the
nature of the universe, to reveal the secret of existence, and
to tell us what the world really and fundamentally is" (SM, 41))
is to be its inherent self-consistency. The "prize of truth" for
which they all strive is to be gained by the form of knowledge
which is self-consistent, or which proves its claim by
"demonstrating the necessary inconsistency of the other forms"
(SM, 45).
270
But then it would seem that the grindstone of dialectical
necessity that Collingwood seems anxious to keep away from the
noses of his readers is kept whirling a few inches away (perhaps
to sharpen an analytical blade). How does a form of knowledge,
which retains its claim to truth by demonstrating both its
inherent self-consistency and the inherent and essential
inconsistency of rival forms of knowledge--how does it do so
without displaying itself as the necessarily true form of
knowledge and the necessary successor to its predecessors? For
if the other forms are inherently inconsistent, then it is clear
that they cannot of themselves achieve ultimate truth, and
therefore of necessity fail.
By now the reader should expect that the solution to the
problem will more likely be to revise the notion of necessity
than to withdraw one or the other of the pair of contradictory
claims.l° But to get an idea of how Collingwood actually viewed
the manner in which such a series of terms are related we must
(once again begging the reader's indulgence) take another look
at the argument of Speculum Mentis. In so doing we shall
deliberately take a different route from the one Collingwood
laid out for his readers. Instead of beginning with artistic
experience and proceeding
_____________________
1O This is Rubinoff's strategy: see CRM, 61, 176-83.
These passages are also the most flagrantly Hegelian in his
whole book.
271
through theothers until we reach philosophy, we shall begin
where we left off in our discussion of Q-A logic, i.e., with
science, and work first backwards to religion and then art, and
then forward to history and philosophy.
Our reason for this strategy is primarily to test this
new requirement for rapprochement: for if (a) the forms of
knowledge constitute a series wherein each term is a modifica-
tion of the one before, then in an analysis of a successorform
one should find traces of its predecessor--traces that will,
when analyzed show how the successor-form is "built upon or
derived from" its predecessor. Thus in the analysis of science
which constituted much of our discussion of Q-A logic, we should
be able to detect a religious element, and in the latter an
artistic element. And if (b) there is no dialectical necessity
involved in the series then when we progress from science to
history the latter will resolve the inconsistencies of the
former without being the necessary outcome of it, or necessarily
implied by it. In short, when viewed retrogressively there will
be necessary relations exhibited between terms in the series;
when viewed progressively these relations will appear to be
contingent.
One more final note before beginning our survey. From
our discussion of the analysis of science in Speculum Mentis we
notice that scientific consciousness is characterized in terms
of (1) the sort of

272
ideal object it intends to grasp, (2) the faculty of mind by
which it habitually operates, (3) its characteristic mode of
expression, and (4) its fatal weakness, or inner contradiction,
that renders it to be an unstable mode of knowledge. Besides
these four, Collingwood further characterizes it (5) by the
consequences it has for human action, insofar as it generates,
when taken as a guide for conduct, a particular form of ethics.
While important for Collingwood's rapprochement between theory
and practice, and therefore another partial confirmation of his
contention in the Autobiography that he sought to carry his pro-
gram of reconciliation even into the realm of ethics and con-
duct, we propose to ignore this fifth aspect here in order not
to further tangle the knotted threads of our argument.
4. Speculum Mentis: Retrogressive Identity.
Science, as described in Speculum Mentis, is the con-
scious attitude of regarding its object always as an instance of
a universal law (SM, 158-63). It operates by a faculty of
understanding, which distinguishes between universals and par-
ticulars, and assumes that universals can be separated or iso-
lated (abstracted) by thought and studied apart from its in-
stances (SM, 166-67). Instances of such universal laws are not
regarded as having mutual relations among themselves other than
those specified by the law (SM, 162, 166-167). But scientific
laws are not merely statements of what would be the case if
certain uniform conditions obtain (the hypothetical deductive
aspect of science),
273
science also seeks to assert its laws as holding in the real
world, and therefore referring to individuals or to the world of
facts perceived by observers (the categorical-inductive aspect
of science) (SM, 177, 183). In epigrammatic form, science is the
form of consciousness which intends an object which can be
expressed in universal, referentially true assertions.
Now it is not difficult to see that this sort of con-
sciousness presupposes and depends for its existence on sub-
ordinate acts of consciousness. Scientific consciousness pre-
supposes a distinction between universals (concepts, laws) and
the particulars (perceptions, instances) to which they apply--a
distinction, Collingwood insists, which is essential and
irreducible for scientific thinking, given the sort of object
(ideal) it intends.ll But a distinction between a universal and
something particular is a distinction between something merely
entertained as a meaning (in imagination) and something to which
this meaning refers as a real instance (in perception); it is
not only an assertion, but an assertion with a referent, i.e., a
real object (EPH, 135-36; cf. SM,
______________________
1l Collingwood states that "mathematics, mechanism, and
materialism are the three marks of all science," but he reduces
all three of these essential characteristics of science to the
"assertion of the abstract concept"--they are all "products of
the classificatory frame of mind" (SM, 167). For Collingwood's
later view of what the characteristic marks of modern science
are, see IN, 13-27.
274
238-39). It is, as we said, both hypothetical (a possible
meaning) and categorical (referential). As hypothetical sci-
entific consciousness is the exercise of making supposals, of
entertaining questions and formulating hypotheses--all of which
imply the consideration of an object without reference to its
truth or falsity, or to its existence or non-existence. This is
the defining characteristic of imaginary objects, and
Collingwood's claim is that imagined objects are what the forms
of consciousness known as religion and art intend. For art the
entertainment and expression of imagined objects is necessary
and sufficient to it as a mental activity; for religious
consciousness there is the additional requirement that its
objects be taken as truly existing--it is not indifferent to the
existence of its object as is artistic consciousness (cf. EPA,
137-41).
Therefore it is clear that (1) if scientific conscious-
ness requires that its object be expressed in scientific asser-
tions, and (2) if scientific assertions essentially imply a
distinction between universals and particulars, and (3) if this
distinction demands expression in hypothetical assertions which
(4) are not only entertained as possible meanings but also (5)
are taken as categorically true, or as truly referring to
particulars, and finally (6) if the latter two activities (that
is, (4) and (5)) are characteristics essential to artistic and
religious consciousness respectively, then (7) we have
demonstrated that scientific

275
consciousness essentially and necessarily contains elements (or
structures) that are characteristically religious and artistic.
If this is an acceptable summary of the argument in the first
three portions of Collingwood's "dialectic of experience" it
seems to be saying too much and too little at the same time. Is
it not asking too much to expect us to believethat before a law
of nature is taken to be scientific it must first be grasped
artistically and then religiously--first painted by an artist
and then worshipped by a priest?l2
________________________
12 Collingwood does point to an explicit inheritance of
religion and art in science, but he presents it in the form of a
mental disposition--a "bias toward abstraction." He writes:
"This bias is allowed unconsciously to control its development .
. . . Because the abstractness of science is a perpetuation of
the abstractness of religion, science most naturally arises out
of a religion which has not overcome this abstractness, that is
to say, out of a non-Christian religion. Hence European science
has its roots in the religion of pagan antiquity . . . . Science
in the modern world is science Christianized, science fed by a
religious consciousness in which the primary abstractness of
religion has been cancelled by the notions of incarnation and
atonement. This gives the distinction between the a priori
science of the Greeks and the empirical science of the modern or
Christian world. But religion, even in the form of Christianity,
never really transcends its abstractness . . . . The aim of
science is to avoid this fault; Greek science aims at avoiding
the specific fault of Greek religion, modern science at avoiding
that of Christianity, namely, its liability to misinterpretation
in a sense which makes God an arbitrary tyrant, whose very gifts
are an insult to a free man. The history of European science
begins with the breakdown of a religious view of the world in
the mind of ancient Greece, and the concepts of Greek science
appear as a kind of depersonalized gods" (SM, 160-61). It
appears that Collingwood would have agreed with Cornford against
Burnet: "Principium sapientiae (quae scientiae) timor dei."
276
This is something of the impression one gets by a forward
reading of Speculum Mentis. But in reversing directions we see
that it obviously is not what he has in mind. Collingwood is not
speaking generically--he is not saying that one and the same
object of consciousness is first taken as an art-object and then
as a religious object and finally as a science-object. He is
making what appears to be a logical point--that forms of
consciousness and their characteristic modes of expression are
bound to one another in a logically necessary fashion, such that
in thinking scientifically one cannot help but also think (in an
implicit way) religiously and artistically (but not necessarily
vice-versa).
But then is this not saying too little? Our previous
objection concerning the reduction of the forms of consciousness
to one another as species of a genus seems to apply here as
well. Surely in scientific consciousness there is an element of
creed, if one takes creed as the form ofexpression of a
consciousness which assumes its object to be real and asserts
itself in statements about that object that are assumed to be
true and distinct from other statements about it which are
false. But to take the object of religious consciousness to be
creed in this sense is to reduce religion to its mode of
expression, and this would be an error as grievous as merely
taking it as an event in the brain, insofar as both ignore what
is properly and peculiarly being asserted by

277
religious statements. Religion is not particularly concerned
with objects sliding down inclined planes or the displacements
of fluids by solid objects of a given weight; and science is not
particularly concerned with the effect that natural laws will
have on the moral behavior of those who believe in them. Fur-
thermore religion expresses itself not only in credal asser-
tions, but in exclamations, questions, petitions, demands,
apologies, etc. By focusing on the mode of expression, and on
only one mode at that, Collingwood seems to be ignoring what it
is that is being expressed by that mode, and this is the sort of
formalism Collingwood had criticized psychology for employing.
And nothing is added by relating a mode of expression to a
faculty of consciousness, since the same objection can be raised
about the mode of consciousness which takes its object a certain
way, but does not specify what is peculiar to the object it is
so taking.
These objections would be more biting if it were not for
the fact that Collingwood spends an entire chapter of Speculum
Mentis trying to say what religion is, not only generically (as
a form of consciousness sharing characteristics with other
forms) but specifically. Religion is a form of consciousness
which intends an object which is a unified whole (like the
object of art), which is ultimately real (like the object of
science and unlike the object of art), and which is taken as
sacred or holy--that is,

278
deserving of adoration (unlike both science and art):
God, we are told by theologians, is the ultimate reality,

conceived as spirit; spirit omnipotent, omniscient, cre-
ative, transcending all sense of immediacy, yet immanent in
his church. But this language, well enough in theology, is
very far from natural to religion . . . . From the simple
and unsophisticated point of view of the religious
consciousness, it is not the spirituality nor the immanence
of God that is important, nor even his power or goodness,
but his holiness, the necessity of falling down before him
in adoration. This sense of the holiness of God is the
explicit differentia of the religious experience. (SM,
118-19).
As in Religion and Philosophy, Collingwood is anxious not to
characterize religion in a reductive fashion, so he is at pains
not to be content, as some writers on religion had been, to
describe "the holy" as a feeling of uncanniness for the divine.
If it is a feeling, writes Collingwood, and one which is a uni-
versal characteristic of religion, "it must be bound up with its
essential nature, and capable of being deduced from it"(SM, 119).
Holiness is to religion what beauty is to art. It is the

specific form in which truth appears to that type of con-
sciousness. As religion, therefore, is a dialectical de-
velopment of art, so holiness is a dialectical development
of beauty. Now religion is art asserting its object. The
object of art is the beautiful, and therefore the holy is
the beautiful asserted as real. . . . Further, holiness,
like beauty, polarizes itself into the positively holy
(God) and the negatively holy, that which we are forbidden
to find holy or worship, the devil and all his works. But
specifically, holiness is asserted as real, and therefore
God is regarded as not our own invention, not a fancy work
of art, but a reality, indeed the only and ultimate real-
ity. Hence that rapture and admiration which we enjoy in
the contemplation of a work of art is in the case of God
fused with the conviction that we here come face to face
with something other than ourselves and our imaginings,
something infinitely real, the ground
279
and source of our own being. It is this fusion which

constitutes the sense of holiness, and forms the basis and
motive of worship. Neither the real nor the beautiful is as
such the proper object of adoration: it is only the
aesthetic attitude towards ultimate reality, or conversely
the elevation of beauty into a metaphysical principle, that
constitutes worship. (SM, 119-20).
once again it is necessary to enter a caveat here against the
possible misinterpretation of Collingwood's intent. In analyzing
religion as a form of consciousness which intends an object that
is defined by holiness, and then by defining holiness in terms
of the object of artistic consciousness (the holy as the
"beautiful asserted as real"), Collingwood is not simply saying
that religion elevates an art object, asserts its reality, and
then falls down in worship of it. That he does not mean this is
clear from his comments about idolatry:
(T)he enemy of religion is idolatry, or the attempt to

worship an object which, however exquisite to the artist's
eye, cannot claim to be the ultimate reality. The sin of the
idolater is to worship his own works of art known to be
such. This is not true religion, because true religion
worships the real God, no mere figment of the
imagination . . . . (I)n religion the mind becomes aware
that it is in danger of illusion. God and religion are
correlative; and to doubt the reality of God is to deny the
validity and legitimacy of religion. There are no religions
without a god or gods: what have passed by that name have
been either philosophies, or religions whose gods have
escaped the eye of the observer, or a kind of mechanical
contrivance put on the market by a deluded or fraudulent
inventor. (SM, 119-20).
280
In defining religious consciousness as the artistic
object asserted as real, therefore, Collingwood is attempting to
characterize what is actually a unified whole by means of terms
drawn both (a) from a lower form of consciousness (art), which
forms one of religion's distinguishable (but not truly separate)
elements, and (b) from within that form of consciousness itself
(the assertion of God as ultimate reality). But the lower term,
in this case the artistic object, "the beautiful," is
transformed into something different--into "the holy" or
something deserving of worship. If "the beautiful" is taken
generically, then "the holy" is a specification of that genus in
such a way that the generic essence is modified by the specific
difference of religious consciousness to the extent that it
becomes identified with it: the genus, in short, is identified
with the variable--the beautiful is asserted as real.
Now we have just seen that a very similar state of af-
fairs appears to be the result of Collingwood's analysis of
scientific consciousness. The genus provided by religious
consciousness (creed: the assertion of the imagined object as
real) is modified by what is specific to scientific conscious-
ness (the principle of abstraction) to become not an object of
faith and worship but an object of scientific inquiry--the
conception of reality as particular instantiations of universal
laws. And we see again

281
from this analysis that the object of religion (God as ultimate
reality) is not taken over by scientific consciousness without
modification: laws of nature are not merely the worshipped God
reduced to the status of universals. The object of scientific
consciousness takes the generic essence of religious
consciousness as already abstract it takes up only the notion of
an imaginary object asserted as real and leaves aside the sacred
or holy aspect of the object of religious consciousness. It
becomes, in terms Collingwood does not employ, wholly profane
science.
We see therefore a very important exemplification of
what Collingwood's modification of the notion of identity was in
Speculum Mentis. The relationship between forms of knowledge is
as good an illustration as one can expect to find of what
Collingwood meant by a "developing series" in which each term is
a modification of the one before. It is also the basis upon
which he will construct his analysis of a "scale of forms" in
the Essay on Philosophical Method: "In a philosophical scale of
forms the variable is identical with the generic essence itself"
(EPM, 60). We do not have to proceed much further in our
retrogressive survey before discovering another anticipation of
a doctrine in the Essay on Philosophical Method: in a scale of
forms there is no zero end of the scale, the minimum realization
of the generic essence lying not at zero but at unity (EPM, 81).
In this case the minimum specification of

282
the genus, "object of knowledge," is the object of artistic
consciousness.
It is no accident that art occupies the primary position
in Collingwood's survey of the forms of experience because to be
an object of consciousness at all is to be entertained as an
object in imagination, without which even perception is
impossible (SM, 204-05; EPA, 57).
The first stage in this process ((viz. the life of spirit or

awareness or consciousness)) is the life of art, which is
the pure act of imagination. This is not only empirically
the first stage observable in children and primitive
peoples, it is necessarily the first stage. Awareness in
itself . . . is an act of consciousness which presents to
itself an object of whose relation to other objects it takes
no cognizance . . . . (I)n religion that indifference to the
distinction between real and unreal, which is the essence of
art, is abolished. Religion is essentially a quest after
truth and explicitly conscious of itself as such a quest.
(EPA, 141).
Once again it is necessary to enter a word of caution to
avoid a misunderstanding. Collingwood is not saying that art
first asserts, then withdraws the aspect of truth or falsehood,
or the reference to reality. Art is not asserting at all: its
"apparent assertions are not real assertions but the very
suspension of assertion"--and the non-assertive, nonlogical
attitude towards an object of consciousness, the indifference to
its reality or unreality, is imagination (SM, 60). That is why
art is prior to the other forms of consciousness, and even the
discussion of its essential nature as the activity of
imagination is not taking place

283
from within artistic consciousness itself, but is thought's way
of describing its most primitive function.
To imagine is to refrain from making a distinction which we

make whenever we think: the distinction between reality and
unreality, truth and falsehood. Therefore imagining is not a
kind of thinking, nor is thinking a kind of imagining, for
each negates the specific nature of the other . . . . Hence
the relation between imagination and thought is that thought
presupposes imagination, but imagination does not presuppose
thought . . . . As thinking presupposes imagining, all those
activities whose theoretical aspect takes the form of
thought presuppose art; and art is the basis of science,
history, "common sense," and so forth. Art is the primary
and fundamental activity of the mind . . . . It is not a
primitive form of religion or science of philosophy, it is
something more primitive than these, something that
underlies them and makes them possible. (EPA, 54-55).
But although artistic consciousness is the suspension of
assertion (and it will be recalled from Chapter V that the
suspension of assertion is one of the ways Collingwood defines a
question and an hypothesis), it is not simply the "blooming,
buzzing confusion" that William James ascribed to the world of
pure sensation. Artistic consciousness has at least the co-
herence necessary to hold its object together as a single en-
tity, and in so doing it makes a rudimentary distinction between
beauty and ugliness. It is this minimal activity that permits us
to call it consciousness or mental activity at all. Insofar as
imagination is a constructive activity (which ultimately issues
in the creation of works of art) it is a representative of
mental consciousness striving to see its world as a whole, the
whole in this case being the imagined object, the work of art
(SM, 63-65).
284
And even though art proclaims its refusal to be bound either by
any necessary relation to the world of reality, or by the
restrictive, rule-governed world of thought, it nonetheless
utilizes, at least implicitly the minimum realization of thought
in its application of the principle of beauty:
Now this process of imagining a whole, or creating a work of

art is . . . no mere rudderless drifting of images across
the mind; it is a process of unification in which the mind
strives to see its world as a whole, the "world" being just
the work of art which for the time being absorbs the whole
gaze of the mind . . . (T)he whole comes into imaginary
existence only in the critical process of experimenting with
its parts . . . . The law of this process, its guiding
principle, is beauty . . . . Now art as such has nothing to
do with principles or laws . . . . ((But)) beauty is not a
concept. It is the guise under which concepts in general
appear to the aesthetic consciousness. Beauty means
structure, organization, seen from the aesthetic point of
view, that is, imagined and not conceived. (SM, 65-66).13
The work of aesthetic consciousness, then, is the creation of
coherence in its minimal form. "When one imagines," writes
Collingwood, "one must imagine something; it must be a definite
and not a self-contradictory imagination, and hence the
necessary unity of the work of art" (SM, 70).
_____________________
13 Although the general sense of what Collingwood is
saying about beauty and art is clear, it is difficult to un-
derstand how beauty can be the "law of this process" of uni-
fication, "its guiding principle--and still not be a concept.
Perhaps Collingwood is trying to present it after the fashion of
Kant's a priori forms of intuition, space and time, which were
also not concepts. Collingwood's later esthetic drops the notion
of beauty altogether, while retaining the theory of imagination:
see PA, 37-41, 137, and 149.
285
We see in the experience of art what Collingwood takes
to be the minimum sense of unity--the identity of a coherent,
imagined object in terms of structure or organization, its
ability (from the point of view of consciousness) to be held
before the gaze of consciousness as an object to it. To para-
phrase Collingwood, to be one is minimally to be a coherent
object of imagination, without consideration of truth or fal-
sity, reality or unreality. But concern only for internal co-
herence is what is ultimately responsible for what Collingwood
calls the "monadism of art:"
Every aesthetic act is an individual internally organized by

the harmonious fitting-together of subordinate aesthetic
acts . . . . Works of art always ignore one another and
begin each from the beginning: they are windowless monads;
and this is because they are acts of imagination, from which
it necessarily follows that they are careless of mutual
consistency and interested only in their internal
coherence . . . . The work of art is a monad, and monadology
is the philosophy of art. (SM, 71-72).
Since art as imagination is necessarily the fundamental
form of consciousness, it is clear that we have reached the end
of the line in our retrogressive survey of the forms of
consciousness (at least those starting with science and its
subordinate forms) necessarily presupposes its predecessors;
each of the subordinate forms, that is, is a necessary, but not
a sufficient condition for its superordinate form, without which
it could not be what it is. And although Collingwood is not
altogether rigorous in exhibiting this dependence at all the
levels of description we found him giving of the forms of
consciousness(i.e. as intentional
286
structures, as objects of consciousness, as modes of expression,
and as faculties of mind),14 the main outlines of the
rapprochement identity, when approached retrogressively, are
clear--and its ideal limit is the minimal form of unity as the
coherent object of imagination.
5. Speculum Mentis: Progressive Identity.
When Collingwood begins the chapter on science in
Speculum Mentis he summarizes the failure of both art and re-
ligion to fulfill the promise that they hold out as being com-
plete and independent forms of consciousness. In so doing he
takes up not what they are in themselves
____________________
14 While it is relatively easy to find passages in
Speculum Mentis and elsewhere that illustrate the retrogres-
sively necessary relationship of forms of consciousness, the
basis for this relationship is given in terms of intentional
structures (the way that consciousness intends its objects),
faculties (imagination, understanding, faith, reason), and ty-
pical modes of expression (hypothesis, assertion, etc.), but not
always in terms of their objects. When religious consciousness
transforms "the beautiful" into "the holy" the transition
between objects, while highly abstract, is fairly explicit. But
when science takes up its object one might expect that "the
holy" would now become "the true" or "the abstract universal" or
some such entity. But Collingwood does not carry forward the
analogy, perhaps because "the truth" is what is intended by all
forms of thought as such, from science on up; or perhaps, again,
because thought self-conscious of itself as such is best dealt
with in its expressive mode in linguistic forms. But from the
point of view of Speculum Mentis, Collingwood clearly felt that
the different descriptive terms we have distinguished were just
different ways of characterizing the same entities--art,
science, religion, etc.
287
(forms of consciousness or intentional structures), nor what
they intend as ideal objects, nor as faculties of mind, but the
way they manifest themselves--art as imaginative expression, and
religion as credal assertion. It is in this form that they most
clearly show their failure to achieve fully satisfying results.
Art and religion, to the superficial observer, are forms not

of thought but of language. Art . . . is simply language
itself, language in its pure form apart from any meaning . .
. . Art is not pure language, but thought failing to
recognize that it is thought, mistaking itself for
imagination. Religion is . . . a dialectical development of
art, art realizing that it is not bare imagination but
assertion, and then proceeding to misinterpret its own as-
sertions and to suppose itself to be asserting the image or
word when it is really asserting the meaning of the word. In
a special sense both art and religion are thus linguistic
functions, forms of expression rather than forms of thought.
(SM, 154).
But if art and religion are both "phases in the history of a
mind, preceding its attainment of complete mastery over the
means of expression," science represents thought's completion of
this development in the recognition that language is the servant
of thought rather than either the whole of thought (as art
assumes) or thought's master (as religion assumes) (SM, 155).
Scientific consciousness explicitly distinguishes between
metaphorical and literal meaning, and hence between thought and
language (SM, 157). It does so by the assertion of the abstract
concept, as we have already seen, and the historical locus of
this event was in the world of post-Homeric Greece, where
scientific
288
concepts appeared first in the guise of depersonalized gods (SM,
158, 160-61). Whereas (a) art fails completely to recognize the
claim of thought on its conscious activity, being totally
absorbed in the technique of expression, and (b) religion only
minimally recognizes the claim of thought but misinterprets it
by identifying thought with its own expression, language with
reality, (c) science represents thought fully conscious of
itself as thought, thought expressed rather than concealed (SM,
154-55).
From the point of view of the modes of expression of the
forms of consciousness, then, each is an incomplete fulfillment
of what thought is trying to achieve: each only partially
conceives of an object which will fully satisfy its (the mind's)
requirements. But when taken as ultimate and independent
statements about what consciousness is, and what the
relationship of consciousness is to its object, they are not
only incomplete they are errors--mistakes that consciousness
makes about itself and about its object. When made explicit by
the labors of thought, these errors appear not as inadequacies
of modes of expression but as actual contradictions within that
mode. Thus art appears to non-assertively assert, religion to
not mean what it says it means, and science appears to
non-referentially refer (SM, 242-43; cf. SM, 311).

289
Now if one were of a formalistic cast of mind, which
Collingwood clearly was not, one might easily demonstrate that
from contradictory statements everything follows of necessity
(because anything, including further contradictory statements,
can be derived from any rule which allows contradictories to
stand as simultaneously true). Therefore when any form of
knowledge is shown to rest on a presupposition (when what it
implicitly presumes is made explicit) that is inherently false,
or self-contradictory, it would automatically rule out the
possibility that any other form of self-consistent knowledge
could proceed from it--and certainly not of necessity. Therefore
from the collapse of one form of knowledge by self-contradiction
no other form of knowledge can follow of necessity, and the
progressive dialectical process, as we anticipated, would show
no necessity: the relation of succession is not one of necessity
but of contingency.
But this is not Collingwood's route. We saw in a pre-
vious chapter (Chapter IV) that Collingwood's commitment to an
anti-realist position concerning the relation of knower and
known necessitated the conclusion that any such error that
consciousness makes about its object recoils on itself, altering
the nature of conscious activity itself (SM, 241). When a form
of consciousness is shown to express itself in an inconsistent
manner, like the rift in the sail of a schooner, it need not
abandon ship,
290
because it has a greater resiliency than strict logic might
allow. It can repair its own damages by transforming itself into
something that at first sight it was not--spontaneous and
reflective thought. Therefore imaginative consciousness (art)
can recognize that it is not utterly free of the imposition of
thought, since it presumes a guiding principle of selectivity
(the idea of relevance) in the construction of a coherent work
of art (SM, 97). The consciousness of this controlling element
in artistic creation is the birthplace of art criticism, and of
esthetic philosophy (SM, 98-100). The distinction that is
introduced is one of form and content, of the manner in which
meaning in art is expressed as opposed to what is being
expressed by that form (SM, 96). This distinction is the
beginning of scientific consciousness,15 and recognizing the
danger that artistic consciousness may at this point cease to be
itself and become science, esthetic philosophy--thought
conscious of itself as art--resists the absorption of its
primary form (artistic experience) by attacking science, perhaps
in the form of an intuitionism that is read back into science as
its (science's) essential nature (SM, 262). Esthetic philosophy
"reduces all philosophical problems to terms of imagination or
intuition," which describe the world as one of pure change, a
monadic world in which every event is new in the sense of
irrelevant to what went before (ibid.).
_______________________
15 Notice that Collingwood here at least implicitly ac-
knowledges that it is possible to pass directly from art to
science--but only through the intermediary of some sort of ex-
plicit thought process, such as art criticism.
291
The same dogmatic service of thought can be performed
for the other forms of experience, and it is this polymorphous
perversity (Collingwood of course does not use this phrase, apt
as he may have found it) that prevents the phases of the
dialectical series of forms of knowledge from becoming neces-
sarily successive. Once a form of consciousness has at its
disposal a battery of modes of expression (questions, asser-
tions, supposals, etc.) provided by the capacities of thought,
it can respond to the threat of destruction of itself by its own
inner dialectic (self contradiction) either by becoming
explicitly what it was only implicitly (rational thought, for
example, instead of pure imagination; assertion instead of mere
questioning) or by transforming itself into another form
altogether (transforming the object of art into the object of
religion, for example). Art can defend itself by becoming
art-criticism or esthetic philosophy or it can allow itself to
be absorbed by religion, science, or history as one of its
essential constituents. Religion can develop a dogmatic defense
as theology, science can develop a dogmatic defense as
metaphysics and logic, and history (peculiarly enough, as we
noted earlier) can develop a dogmatic philosophy in the form of

292
realism; or each could be transformed in its successor (SM,
260-87).16
We have at this point shown, therefore, how Collingwood,
true to his epistemological assumptions,has maintained the
contingency of the forward motion of his process of re-
conciliation between the forms of knowledge. Progressive
rapprochement identity (according to the position proclaimed in
Speculum Mentis, at least) consists in the unifying activity of
a form of consciousness, deliberately attempting to preserve its
integrity by overcoming errors which become manifest when its
implicit assumptions are explicitly expressed. Since it can
overcome error in a variety of ways, and with differing degrees
of success, the possible modes of succession of thought forms
are not fixed, and contingency in the progressive direction is
preserved. The ideal limit to the process of overcoming
expressed errors is total self-consistency, or absolute
knowledge. Such a knowledge would not only grasp its intentional
object in a non-misleading way, but would also account for the
errors that it has made in achieving this totally adequate
knowledge. Therefore it cannot intend an object that is utterly
one, a bare blank identity (the abstract Absolute), but a
unity-in-diversity, one for which differences are essential, a
"concrete universal." But a concrete universal as an
__________________
16 Although strictly speaking, tertium non datur, I see
no reason why Collingwood would object to the suggestion that a
form of experience could simply remain what it is, ignoring the
contradiction within itself.
293
absolute object would have to be an infinite whole of fact, and
a whole in which subject and object are not conceived as
separate, but identified. Ideal maximal progressive rapprochement
identity, therefore, is one of complete identity of subject and
object.17
Since there is a whole battery of problems that arise in
connection with Collingwood's concluding chapters of Speculum
Mentis, especially concerning this progressive rapprochement
identity in the form of "absolute knowledge," we propose to deal
with these problems as a series of disputed questions with which
we will close this chapter.
6. Disputed Questions.
(1) Why do forms of consciousness succeed one another?
Collingwood's descriptive phenomenology does not set out to
answer this question, but it certainly does "arise"--even on his
own sense of that term. If there is no necessity (meaning
logical compulsion) in the forward direction in a scale of forms
of knowledge, then it would seem not merely contingent that they
succeed one another but utterly accidental, even gratuitous or
miraculous. Is the reader to assume a "nisus" or innate striving
toward greater adequacy of thought? Does thought have
_______________________
17 Sources for the statements in this paragraph (the
argument is nowhere, of course, stated as such by Collingwood)
include: SM, 238-41, 288-97. More detailed, explicit references
will be given in section 6, question 2.
294
an inborn natural tendency to supersede or overreach or
transcend itself and thus create different forms of itself? Is
the logical requirement of self-consistency a motive force at
all levels of consciousness?
In his "Outlines of a Philosophy of Art" (published one
year after Speculum Mentis) Collingwood actually does posit a
"nisus towards self-consciousness" at all levels of con-
sciousness (EPA, 144), but in Speculum Mentis he does not give a
consistent answer to this question. On the one hand, in the
introductory and less precise passages of the book, Collingwood
seems to lean towards the "nisus" thesis, if we may so call it,
insofar as he assumes that the disease of modern man is
self-alienation, the separation of the forms of experience one
from another, and the cure for this disease to be their "reunion
in a complete and undivided life" (SM, 36). But on the other
hand, Collingwood denies that there is anything like a fixed
"human nature" (SM, 296)--a denial he defended throughout his
lifetime, and against which he threw all the weight of his
reflections on the nature of history as a "self-making ac-
tivity." But if there is no such thing as a fixed human nature,
not even in the minimal sense of the term, then it is hard to
account for any tendency for consciousness to become altered at
all. In fact it seems utterly groundless to assert that
consciousness would seek to grasp a fully satisfactory object
unless there are fundamental and irreducible

295
characteristics of consciousness as an activity, and what could
this be but a minimally essential human nature?
Collingwood does not acknowledge the problem at all in
Speculum Mentis, the assumptions upon which it builds its argu-
ment being (1) that knowledge and consciousness exist (but in a
divided state), and (2) that these mental activities claim fully
satisfactory ideal objects (cf. SM, 39). It is not until he
attempted a more complete analysis of mental activity in The
Principles of Art and The New Leviathan that the
emotive-expressive aspects of mental activity was fully inte-
grated into his philosophy of mind.
(2) Even assuming that, for whatever reason, the forms
of consciousness do succeed one another in a scale of developing
forms, and in much the way Collingwood describes them, is there
any end-point to the series? We have seen in our retrogressive
survey of the forms of consciousness that there is a terminal
point at the "lower" end inasmuch as the consideration by
consciousness of a whole object, its mere entertainment by
imagination, is the minimal sense in which intentionality can
grasp its object at all. But we have also seen that the
imaginary object is intended without consideration of its
reality or unreality (i.e. without reference), and that this is
the root sense of abstraction, insofar as the act which grasps
an imaginary object does so by ignoring (Collingwood will later
say by "suppressing") its relations to any and all

296
other objects. And abstraction, as the negative side of in-
tentionality, is also the act by which consciousness cuts itself
loose from its object, or sets the object apart as "separate"
from itself. But this is the primitive act of separation of
subject and object, the root error of realism.
From this analysis it is not surprising that the repair
to the torn fabric of consciousness will be the reunification of
subject and object by an act that is the very opposite of
abstraction--viz. reconciliation. This in fact is Collingwood's
strategy in Speculum Mentis: the termination of the scale of
forms is philosophy (which succeeds history), and whereas
history achieves concrete knowledge (and therefore rectifies
implicitly the abstractness of science), philosophy achieves
absolute knowledge (and therefore overcomes the last vestige of
abstraction in the form of history's separation of subject and
object--the historian contemplating a world of facts, and not
fully aware that he is more than merely an observer of those
facts) (SM, 238-39; cf. SM, 242-43, 311, and EPA, 143-44).
We shall see in a moment what the subject matter of this
absolute knowledge is, and how it leads to further difficulties
unresolved in Speculum Mentis. For now it is sufficient to
notice that the argument that

297
Collingwood provides for us as a justification for the
transition from history to philosophy does not appear to leave
us with any alternative, and thus threatens the contingency of
the forward dialectic. He not only argues (a) that if the
distinction or separation between subject and object is invalid,
then the last veil hung between the mind and its object falls,
revealing (not Salome but) the mind itself in mirror
reflection--the "speculum mentis" of his title; but also (b)
that on either of two mutually exclusive alternatives, the
result (the identity of subject and object) follows:
If subject and object are opposite, then they can only

exist in synthesis: well and good. But if they are distinct
concrete facts, they both fall within the world of fact,
and of this world it remains true that everything in it
determines the whole and everything else, it follows that
subject and object are just as inseparable on this
hypothesis as on the other. For the concept of the world of
fact as the concrete universal has destroyed any dis-
tinction between a logic of opposition and a logic of dif-
ference.18 The fundamental principle of history itself,
namely, the concreteness of the object, thus makes it im-
possible for the object to ignore the subject, and compels
us to recognize an object to which the subject is organic,
in the sense that the subject's consciousness of it makes a
real difference to it as a whole and to all its parts . . .
. Being known, whether truly known or erroneously known,
must make a difference to the object: to deny this . . . is
to turn one's back on concrete thought and revert to the
fallacies of abstraction. (SM, 244; emphasis mine).
With this passage it appears that the veil is not so much

dropped as torn
___________________
18 Collingwood here anticipates another doctrine of the
Essay on Philosophical Method: "the kind of opposition which is
found among philosophical terms is at once opposition and
distinction" (EPM, 75).
298
down, revealing not only the mirror of the mind but
Collingwood's own presuppositions, and it becomes clear that the
position he is advocating is naked epistemological idealism
(albeit of an unusual variety). The suspicion (if one still
needs convincing) is further confirmed in a later passage in a
section dealing with the historical form of dogmatism, in which
Collingwood praises German idealism for killing "scientific
realism--the popular philosophy of today-as dead as a door nail"
(SM, 287).
But the admission also threatens Collingwood's entire
philosophical enterprise in Speculum Mentis. For if the forward
movement of the dialectic of experience is not propelled by
necessity, then the final transition, even more so than all the
intermediate ones, seems hypothetical at best and arbitrary at
worst. But the above passage, as the underlined words show,
seems to contradict the thesis that the dialectical progression
is not one of logical necessity.l9 The reader is shot as from
_________________________
19 Cf. SM, 292-93: "Not that such creation of an ex-
ternal world is capricious. The mind cannot simply think what-
ever it pleases, or even imagine whatever it pleases. It is
bound by the laws of its own nature to this extent, that even
though it can deform its nature by misconceiving it, it can
never deform it out of recognition, because misconceiving is
after all a kind of conceiving. Its scientific concepts, its
religious imagery, its aesthetic imaginings must grow out of the
soil of fact, and that fact is just its own nature as that
stands for the time being. This necessity of all its actions,
ignored in the life of imagination, is though ignored not done
away. It is transformed, by being ignored, from a rational ne-
cessity to the blind necessity of instinct . . . . The discovery
of necessity . . . is the achievement of the religious con-
sciousness; but this necessity is there from the first."
299
an historical cannon into the realm of idealistic philosophy.
We shall see in the next question that Collingwood's
attempts to escape from the consequences of his own argument by
denying that his position commits him to a metaphysical form of
idealism are not convincing enough, as they stand, to exonerate
his "absolute knowledge" from self-contradiction, and
consequently from the necessity of positing even a further form
of consciousness to repair the damage.
(3) Does philosophy have an object? The reader of
Speculum Mentis is well aware that Collingwood wants philosophy
to be a form of knowledge that (a) is self-consistent, (b) is
self-consciously reflective, and (c) achieves the object of
self-knowledge in a manner that escapes, or overcomes, the
errors of subordinate states of consciousness (SM, 45-46,
247-49). We have just seen that Collingwood hopes to fulfill
these conditions in the guise of absolute knowledge. By as-
serting that the differentia of absolute knowledge is the
identity of subject and object (SM, 249), Collingwood argues
that he has found a kind of knowledge that fulfills all the
requisite conditions. (a) If it is identical with its object,
there is no "externality of the object," and therefore no place
for necessary inconsistency to conceal itself. And (b) so also
it is not knowledge of an object that is other than itself, but
is rather that
300
object knowing itself. But (c) the problem of inconsistency
arises once again the instant one attempts to see how such a
knowledge achieves its object in a manner that escapes error.20
The reason for this is that the instant one attempts to
import any content into the abstract formula "the identity of
subject and object" the air becomes murky with the gaseous
remains of previous errors. The object of philosophy, he says,
is not that of art, religion, science or history because each of
these forms of knowledge intend an object that is assumed in
some sense to be independent of the subject (SM, 30609). Yet
what the philosophical form of consciousness reflects on is
nothing other than the succession of worlds created by art,
religion, etc.
In an immediate and direct way the mind can never know

itself: it can only know itself through the mediation of an
external world, know that what it sees in the external world
is its own reflection. Hence the construction of external
worlds . . . is the only way by which the mind can possibly
come to that self-knowledge which is its end. (SM, 315).
Absolute knowledge, therefore, consists in nothing more than a
survey of the succession of errors by subordinate states,
recognized as such by
____________________
20 In fairness to Collingwood it should be pointed out
that he says that absolute knowledge is not secure from error,
but rather it is called absolute because "in it there is no
element of necessary and insurmountable error" (SM, 295).
301
a form of consciousness that takes its object to be nothing
other than itself. But each of these subordinate forms of
consciousness when recognized as illusory (erroneous portrayals
of an object separate from the mind which contemplates it) are
absorbed into philosophy: there is no "map of knowledge" because
"there are no autonomous and mutually exclusive forms of
experience" (SM, 306). Philosophy is therefore the consciousness
of something which is also the consciousness of nothing.
The same conclusion follows if one proceeds in another
direction--from a description of absolute mind. Here Colling-
wood's rejection of all possible content is even more sweeping.
He says that whether the life of the spirit be described by a
group of categories, a group of laws, a group of presupposi-
tions, a world of objects, or a series of stages, it is an
erroneous description of absolute knowledge because the de-
scriptive terms are "versions of a single error: the error of
abstraction, of failing to realize that subject and object,
condition and conditioned, ground and consequence, particular
and universal can only be distinctions which fall within one and
the same whole, and that this whole can only be the infinite
fact which is the absolute mind" (SM, 310). But the recognition
of this "infinite fact" is an act which "abolishes the notion of
an external world other than the mind" (SM, 310). But then it
would seem that the infinite world of fact is abolished with the
302
external world, the baby discarded with the bath water, and
absolute knowledge is left with nothing to contemplate.
Collingwood's attempts to evade the contradictory as-
pects of the concept only tend to add to one's puzzlement. For
example he says that the abolition of an external world other
than mind does not imply the abolition of the distinction
between suject and object:
These distinctions are only abolished by the coincidentia

oppositorum which is the suicide of abstract thought, and
conserved by the synthesis of opposites which is the life
of concrete thought . . . . But in abolishing the notion of
an external world other than the mind we do not assert any
of the silly nonsense usually described by unintelligent
critics as idealism. We do not assert that the trees and
hills and people of our world are "unreal" or "mere ideas
in my mind," still less that matter is nothing but a swarm
of mind-particles. The very essence of trees and hills and
people is that they should be not myself but my objects in
perception: they are not subjective but objective, not
states of myself but facts that I know. None the less, my
knowing them is organic to them . . . . They and I alike
are members of one whole, a whole which the destruction of
one part would in a sense destroy throughout . . . . (SM,
310-11).
But then what is this "whole" in which "subject and object are
identified" and which nevertheless is one for which externality
is an illusion; which is an "infinite world of fact" and yet not
"one stupendous whole" (SM, 299)? How does a synthesis of
opposites differ from a coincidence of opposites, unless the
identity that is that synthesis is identifiable, describable,
recognizable? If absolute knowledge is self-knowledge, and if

303
in it subject and object are identified, and if self-knowledge
is only possible through the intermediation of subordinate
forms, then how is it possible for objects of consciousness not
to be "mere ideas in my mind"? Is it any wonder then that in the
"progressive reduction of art, religion, science and history to
philosophy" not only is it the case that "each one of these
lives disappears; but philosophy itself disappears as completely
as any" (SM, 293), leaving not so much as the smile of the
Cheshire cat?
Now one might grant that what Collingwood is dealing
with in these enigmatic passages is a mystery surpassing un-
derstanding, the mystery of self-consciousness and its existence
in a world that appears external to the mind which nonetheless
knows it. One might grant that in grappling with such a mystery
one cannot help but lapse into forms of speech that are
contradictory, the sort of language familiar to mystics and
spiritualists. One might be so lenient with him, were it not for
the fact that Collingwood himself is claiming consistency for
what he is saying, that he uses logical criteria for deciding if
the claims of consciousness can live up to their expressed
performance, and that what he has led the reader of Speculum
Mentis to expect is a coherent account of what it means to be a
fully adequate form of knowledge.

304
Therefore we cannot rest content with accepting this
state of affairs as a fitting conclusion to his search for
rapprochement. We have the right to demand that if a "higher"
form of consciousness is itself inconsistent on logical grounds,
if it cannot live up in its performance to what it had promised,
we must either (a) declare it to be an exception to the rule of
consistency, or (b) posit an even higher state of reflective
consciousness to further repair the damages. But it is easy to
see that so long as a state of consciousness is distinct from
its object, alternative (b) leads to an infinite regress, and
therefore the goal of Speculum Mentis (and progressive
rapprochement) will be forever frustrated, because
self-knowledge will never be possible: it involves a
contradiction in terms. And the first solution (a) acknowledges
that the criteria of consistency (a supposedly higher form of
consciousness--absolute knowledge) be affirmed as itself
inconsistent. But then why not absolve any of the lower forms
from an equivalent necessity to be self-consistent? Why not stop
the series with history, for example? What need is there for
philosophy?
As we shall see in a future chapter, Collingwood's so-
lution was to opt for alternative (a), which is what we might
expect for someone who felt himself to be committed to the So-
cratic view of philosophy as self-knowledge. But in taking this
route Collingwood provided philosophy

305
with one of those rare occurrences--a bonafide metaphysical
paradigm. If the criteria of consistency is itself inconsistent
it may be so either by (1) asserting something to be both true
and false at the same time and in the same manner, or (2) by
asserting something that is neither true nor false. In the Essay
on Metaphysics Collingwood took the latter option, and spelled
out the consequences of this doctrine as a theory of "absolute
presuppositions" which are the basis for truth and falsity
without being themselves true or false (EM, 21-33).
(4) If absolute knowledge collapses for want of a co-
herent object, how are the various forms of consciousness de-
scribed in Speculum Mentis reconciled to one another? Or does
Collingwood's early rapprochement philosophy end in complete
disaster? Is retrogressive identity the only acceptable basis
for rapprochement?
We have had occasion in previous chapters, and this
chapter is an extension of these reflections, to remark on the
peculiar usage that Collingwood has for terms that appear as key
words in all of the contradictory texts cited above-"other
than . . . , separate from . . . , identical with . . . , the
same as . . . ," etc. These terms, and the understanding of them
in context, count greatly toward contributing to the sense or
nonsense of what Collingwood is trying to say in these highly
elliptical and abstract passages. It is also

306
crucial for an understanding of rapprochement philosophy, since
the entire effort of reconciliation is to show that the
reconciliate are not "other than" or "separate from" each other,
but are rather "identical" or "the same as" each other as parts
of a "concrete whole."
Now we recall from section 2 of this chapter that in
Religion and Philosophy four sorts of identity can be dis-
tinguished--abstract, dialectical, concrete, and absolute. The
first two were bound, we noted, by relations of necessity, the
latter two by relations of contingency. We also noted that each
had a corresponding sense of "separate." If we come fresh from
this discussion to the present problem in Speculum Mentis, we
notice several interesting shifts in meaning, and as usual with
Collingwood, a small investment of careful attention to these
shifts yields dividends for the interpretative speculator.
If we return, for example, to the transition from his-
tory to philosophy we notice that the object of history is taken
to be the "concrete universal" and that this "infinite whole of
fact" is taken to be an object that ostensibly satisfies the
mind:
There is thus no feature of experience, no attitude of mind

towards its object, which is alien to history. Art rests on
the ignoring of reality: religion, on the ignoring of
thought: science, on the ignoring of fact; but with the
recognition of fact everything is
307
recognized that is in any sense real. The fact, as

historically determined, is the absolute object. The mark of
the absolute object is individuality, for individuality is
concreteness. The object as individual is the whole of what
exists, and this is concretely articulated into parts each
of which is again individual, and so to infinity . . . . The
object as a system of fact so organized, is objective
throughout, for every part is a true microcosm, and is truly
infinite . . . . This absolute whole is the concrete
universal . . . . It is the system of systems, the world of
worlds . . . The principle of its structure is not
classification, the abstract concept, but the concrete
concept, which is relevance, or implication . . . and the
logic of history is the logic of the concrete universal.
(SM, 218-21).
Aside from the fact that it is hard to see why an absolute mind
knowing such an object would not be "one stupendous whole" (SM,
299) if its object is a "system of systems" and a "world of
worlds," we notice an additional peculiarity in this passage.
The description of the concrete universal corresponds fairly
closely to the description of both dialectical and concrete
identities as we discovered them in Religion and Philosophy. (1)
The relations and relate are connected not as abstract
particulars subsumed under an equally abstract genus, but as
parts of a whole such that the parts reflect the whole and the
whole reflects the parts--therefore a dialectical identity. But
(2) they are also described as concrete in the sense of
individual--and thus (like the identity of cooperation between
minds in Religion and Philosophy, wherein the two minds share
that unity between them which is taken to be the mark of the
individual, and therefore become one mind) a concrete identity.
If Collingwood is truly pressing these two sorts of identity
into doing
308
service in the single notion of the concrete universal, then it
is not surprising that we should find in it relations of both
necessity and contingency, since dialectical identity involved
relations of necessity, and concrete identity involved relations
of contingency. In the concept of the concrete universal, then,
we have an attempt to express the overlapping of two senses of
identity; it is therefore an instance of what Collingwood will
later call the "overlap of classes" in his Essay on
Philosophical Method (EPM, 26-53). Which brings us to our next
question.
(5) Is there any way that absolute identity as the ideal
for progressive reconciliation can be made intelligible?
Unfortunately the suggestion from the last paragraph, i.e. that
the concrete universal as an overlap-concept might be expanded
to become the absolute object, runs headlong into the
subject-object contradictions we have been at pains to concile.
On the one hand the concrete universal turns out to be the
absolute object, thus accounting for the fact that the identity
of history and philosophy, while a "barren abstraction" like all
such identifications, is less misleading than the others (SM,
246). On the other hand the object of history fails to be the
object of philosophy, because the historical consciousness fails
to be fully aware of the identity of subject and object. History
as a separate form of
309
consciousness fails to apprehend the concrete universal (an in-
finite whole of fact) because of a remnant of the original sin
of the mind--the abstract separation of subject and object (SM,
237-38); the historian simply cannot grasp an "infinite given
whole of fact," and therefore in confronting the the panorama of
history and the virtual infinite of historical evidence the
historian must select his materials, arrange them into periods,
etc.--all acts of abstraction.21
It is at this point that Collingwood introduces his
hypothetical identity of subject and object by an act of ab-
solute consciousness: "If therefore the infinite given whole of
fact is the nature of the knowing mind as such, our problem is
solved, and the possibility of knowledge is vindicated" (SM,
241). Once again we are brought to the brink of complete
disaster for Collingwood's voyaging vessel of consciousness: the
fabric of his mainsail is, as we have seen, ripped from top to
bottom by inner contradiction.
In Religion and Philosophy, at least, he left an escape
route open in the form of the Absolute Mind as God, with whom
men may contingently be
____________________
21 Collingwood's description of historical abstractness
is almost Heideggerian: "History, which seems to be essentially
remembrance, is only possible through forgetfulness, a
forgetfulness which in destroying what it takes away makes it
impossible for us ever to understand what is left" (SM, 236).
310
united through identity of purpose with Jesus Christ. In
Speculum Mentis this door also seems to be shut, both by his
explicit assertion that the absolute mind is the mind of
individuals and not that of some "world spirit" which he rejects
as a "myth" (SM, 298-99), and his affirmation that the absolute
mind is the mind of each of us (SM, 298).22
Fortunately, there is more than one way to reach ab-
solute knowledge, and while we have been preoccupied with a
discussion of consciousness and its objects, we have lost sight
temporarily of another approach to the subject—one which
Collingwood himself, in his haste to
_____________________
22 Once again, in fairness to Collingwood it must be
said that in this passage he says that "the mind of which we are
speaking . . . must at least be the mind of each of us . . . ."
(SM, 298), thus leaving the way open for the insertion of a
higher mind which does not share the inherent failing of human
nature. But in a passage where the life of absolute mind is
described by means of the religious metaphor of the fall and
redemption of man, he seems to reject this possibility. The
metaphor likens the fall of man to the loss of absolute
knowledge through an act which forever separates subject and
object, and redemption to the regaining of this knowledge
through an act of divine transcendence--the incarnation. In his
fallen state man fails to achieve self-knowledge: "not knowing
himself as he ought to be, he cannot know himself as he actually
is. His error is implicit just because it is complete" (SM,
269). However Collingwood criticizes the metaphor as having one
flaw: the "transcendence of God." He has also asserted that "no
one can worship the absolute" (SM, 151), and furthermore that
the point of entry of God into a philosophical system marks
unerringly the point at which it breaks down (SM, 269). Finally,
the reader might also recall that Collingwood maintains that
there is no fixed human nature, so that the imagery of the fall
is further flawed insofar as it is inapplicable to man.
311
reach the absolute standpoint, may have overlooked. We noted
above (section 4) that in the beginning of the chapter on
science Collingwood called art and religion "forms not of
thought but of language (SM, 154). We have also seen him carry
out his analysis of the explication of implicit errors of
successive forms of consciousness in terms of their linguistic
forms--expression (art), assertion (religion), etc. Even his
discussion of knowledge as question and answer tends to focus on
the language of conditionals, hypotheticals, assertions, and
implications. In this discussion Collingwood drops an intriguing
hint which is later picked up in his discussion of the nature of
philosophical thinking. Discussing symbol and meaning in his
chapter on religion, Collingwood writes:
To distinguish a symbol from its meaning is to put oneself

in the way of explaining or translating the symbol. Now it
is a matter of common observation that religion never
explains itself . . . . To ask for explanations is the mark
of extreme sophistication; in other words, it is the mark of
the life of explicit thought . . . . Art is untranslatable,
religion cannot translate itself. Art cannot be translated
because it has no meaning except the wholly implicit meaning
submerged, in the form of beauty, in the flood of imagery.
Religion cannot translate itself not because it has no
meaning . . . but because, although it has a meaning and
knows it has a meaning, it thinks it has expressed this
meaning already. And so it has, but only metaphorically; and
this metaphorical self-expression, this fusion of symbol and
meaning, requires translation . . . . For literal language
is only language recognizedly metaphorical . . . .(SM,
128-30).
We have seen that in his apology in Speculum Mentis for
the sins of Religion and Philosophy, Collingwood pointed to the
failure to
312
recognize the distinction between symbol and meaning in religion
as his major oversight in that earlier work (SM, 108, n. 1). And
we also noted that when he originally published Religion and
Philosophy he deliberately declined to deal with the issue of
religious language (RP, xvi). And we have seen that in Speculum
Mentis the distinction appears as the singular way that the
error implicit in religion is made explicit--religion "asserts
the reality of what is only a symbol and thereby treats a symbol
as though it were a concept" (SM, 153). The same might be shown
for all the other forms of consciousness: each manifests its la-
tent error in the form of inadequate or self-contradictory
expressions. Not only religion requires translation, it seems,
but all the forms of consciousness: when science, for example,
achieves explicit thought, it does so by means of its ability to
express meaning in several ways, by means of overlapping
metaphors in the form of alternative hypotheses and equivalent
mathematical expressions, thus overcoming the fixed formulas of
religious dogma (cf. SM, 155-57).
In the chapter on philosophy this line of thought is
picked up again, and provides us with a possible key to un-
scramble many of the paradoxes of absolute thought:
We have hitherto allowed ourselves to say that in art,

religion, and so forth the substance of truth was present,
but was concealed in an inadequate form: that, for instance,
religion actually solved the
313
riddle of life but presented its solution in a mythological

form. This implies that the task of philosophy, regarded as
the philosophy of religion, is the simple translation of
this solution of the riddle of life out of the language of
mythology into that of philosophy . . . . Translation itself
is based on the fact that the meaning takes new colour and
shines with a new light when we express it in different
words. To set the meaning as an abstract self-identity over
against the language makes translation pointless: to swamp
it in a mere immediate union with the language itself makes
translation impossible . . . . (T)hought in its concrete
form is not indifferent to its own choice of language. It
realizes that an unsuitable linguistic form affects its own
inmost being, and that what we have called merely formal
error is in reality material and essential error. Our
distinction between formal error and material error was, in
fact, only an abstract way of stating the very important
fact that no error is wholly erroneous, but is always
capable of a dialectical development into truth by simply
bringing to light what is already implicit in it: what the
thinker, as we paradoxically say, "really means," but "does
not know that he means." This process of translation into
progressively adequate language is simply the dialectical
self-criticism of thought. (SM, 252-53).
I have quoted this passage at length because it is so
important, so capable of being overlooked,23 and so pregnant
with possibilities. For (1) if philosophy is the process of
making explicit what is implicit in other modes of thought, and
(2) if the implicit errors of these modes of consciousness are
only made explicit when expressed or translated into language,
and (3) if the dialectical development of error into truth is a
___________________
23 Collingwood himself seems to have overlooked its
significance, both in the conclusion of Speculum Mentis, where
the suggestion is not followed up, and later in The Principles
of Art, where the significance of language for thought and the
role of philosophy as translation appear as discoveries, rather
than as a development of a line of thought already initiated ten
years earlier. Cf. PA, Chapter XI, pp. 225-69.
314
process of translation into progressively more adequate language
(the dialectical self-criticism of thought), then it is clear
that (4) absolute knowledge in the form of philosophy is the
translation, by dialectical self-criticism, of expressions of
subordinate forms of consciousness, into increasingly more
adequate language. One is tempted to say (as Collingwood does
not) that the object of philosophy is expressed in absolute
language--i.e. language purged not of all error but only of its
element of necessary and insurmountable error.
What such a language would be like is a matter for
speculation, although Collingwood's later writings provide us
with a few clues. It would have a peculiar grammar and the Essay
on Philosophical Method and the methodological chapters of the
Essay on Metaphysics are attempts to provide us with an informal
account of what that grammar is like. It will deal with problems
of the sort provided by art, by science, and by history, and The
Principles of Art, The Idea of Nature, and The Idea of History
are examples of what it would sound like when these problems are
translated by philosophical consciousness into a more adequate
language. And it would attempt to formulate an idea of what an
object would be like that would totally satisfy the mind, and
these reflections are presented in the Essay on
Metaphysics--Collingwood's last word on the religious
foundations of contemporary thought.

315
But aside from what an absolute language would be like,
it is clear that the conception outlined in the quotation above
has the potentiality for clarifying some, if not all, of the
difficulties we found with the notion of absolute knowledge. (1)
If the differentia of philosophy is taken to be the
identification of subject and object, the difficulties with this
conception arise when it is described in terms of consciousness
and its object--the stumbling block always being
self-consciousness. But if absolute knowledge or philosophy is
described not in terms of consciousness but of language, then
this particular difficulty disappears: language is quite capable
of being self-referential, of discussing and describing itself,
and of introducing modifications to overcome errors in the
expression of its more primary forms (e.g. "natural languages").
Not only are subject and object identified insofar as both are
embraced within the same whole--a world of language in which
pronouns, reflexive forms, and self-referential assertions are
all possible; but the subject is also both a receiver of meaning
and a creator of meaning in this world. Therefore a change in
one part (e.g. the creation of a new meaning--a poetic metaphor,
a new scientific hypothesis) necessitates a change in the whole
(the interconnected world of meaning) and in all the other parts
(related meanings). Since the subject here is a user of lang-
uage, he is identified with

316
objects not immediately (he does not actually "become" the
object) but through the mediation of subordinate expressive
acts--through the unreflective use of the language of art,
religion, science, and history. He becomes one with his object
not actually but virtually, symbolically, by an identity of
meaning.
(2) As a world of language the "concrete universal" (as
expressed judgment) could understandably be a "world of worlds"
and a "system of systems." As an expressive form characteristic
of philosophy it contains elements both hypothetical (calling
forth alternative possible meanings) and categorical (in its
referential determination of a given meaning); and in
philosophical discourse the concrete universal can be engaged in
questioning (like art), answering or asserting (like religion),
abstracting (like science) or referring (like history). And it
can claim adequacy at all these tasks without being "one
stupendous whole," since not all entities are linguistic.24 And
even though not all entities are linguistic, there would still
be no externality, no element of necessary and insurmountable
error, insofar as there are no non-linguistic entities
____________________
24 The terms "linguistic" and "linguistic expression"
must be taken as having the widest possible extension--including
not only the utterances of natural languages, but all sorts of
artificial languages as well (including mathematics and logic)
and even works of art (music, painting, dance, etc.). Cf. PA,
252-69.
317
that are not capable of being described, referred to, or
translated into linguistic entities. Therefore none are in
principle "external to" or "separated from" a subject.
Linguistic expression makes the whole world of objects, real and
possible, actual and fictitious, true and false, accessible to
the user of language, the subject.
(3) Philosophy as translational activity preserves the
contingency (and hence the freedom) of the progressive movement
of rapprochement insofar as there is no necessity to translate
something erroneously expressed unless one is committed to the
creation of, or preservation of, a higher mode of expression.
Philosophy as translation aims at consistency, but not strictly
speaking a formal consistency, but rather a consistency of
coherent meanings--meanings which cannot help but overlap in
specifiable ways, rather than abstract meanings which are
mutually exclusive. And even though it aims at consistency,
there is no fixed set of rules which prescribe a one-to-one
translation of one set of terms into another. Translation must
have a certain flexibility, aiming as it does at the
transmission of meaning rather than at mimetic correspondence.
Translation of meaning is impressionistic rather than
photographic: its rules are not necessarily the rules of strict
formal correspondence, but vary from loose metaphor to verbatim
literal and grammatical transformational analysis. One cannot
even say that the same ideal

318
is always adhered to, for in poetic translation a strict,
word-by-word translation would be undesirable, even ridiculous.
But philosophy, if nothing else, interprets itself as a guardian
of meaning rather than a keeper of rules. As such the forward
motion of philosophy is towards progressively more adequate
meaning, and the means to achieving this is, if not the opposite
of necessary inference, at least independent of it. Its
resources are the resources of freedom rather than those of
necessity.
(4) The ultimate identity that would serve as a maximal
ideal for a progressive rapprochement would therefore be an
identity of meaning, where through an identification of meaning
and meant, subject and object (knower and known) are identified.
The world of experience thus reconciled is an absolute built out
of the differences of the world as it finds them, but not by a
principle of abstraction (which leaves something always unsaid)
but by the progressive consolidation of a world of expressed
meanings, of articulated facts, bound together in such a way
that the whole (itself a meaning) could not be what it is
without its parts (which themselves are, or have, meanings) and
vice versa.
But we have been allowing ourselves the license of an
unrestricted flight of Collingwoodean fancy. How much of it
Collingwood might have

319
agreed with, we can only surmise. But there is little doubt that
his rapprochement philosophy is capable of the sort of
flexibility that would include modifications along the lines we
have suggested here; for indeed, for two or more reconciliata to
reach rapprochement means for them to have, and to be shown to
have, not the same objects unaffected by the knowing of them,
but the same meaning.

PART III
THE LATER WRITINGS (1933-1942)

CHAPTER VII
ANTI-REALISM AND THE PHILOSOPHY OF

MIND
Introduction.
In Chapter I we noticed that one of the major issues of
concern to Collingwood's interpreters is his position on the
realism-idealism controversy. Knox argues that he vacillated
from a youthful realism to a mature idealism and then back to a
dogmatic and sceptical realism in his later writings. Donagan
maintains that the youthful idealism propounded in Speculum
Mentis, with its anti-realistic, anti-abstraction principle, was
superceded by a more mature philosophy of mind in which
abstraction is recognized as being essential for thought; the
resulting philosophy of mind, though flawed, is compatible with
realism and contemporary analytical philosophy. Rubinoff denies
that Collingwood was ever anything but an idealist. And Mink
argues that Collingwood's philosophy escapes both idealism and
realism.
In Chapter II we found that Collingwood's
self-interpretation in the Autobiography also stresses the
realism-idealism controversy, but it stresses mainly his
opposition to errors of realism, and only incidentally denies
that his early work (Speculum Mentis) was, as one of his
reviewers
320
321
called it, "the usual idealistic nonsense." On this inter-
pretation we know only that Collingwood was opposed to the
central tenet of realism, as well as to its many pernicious
consequences, but not what Collingwood's commitment was to
idealism. We also saw in this chapter that Collingwood framed
the central tenet of realism in terms of the knower-known re-
lationship, and not in terms of the dependency of matter on mind
or even in terms of an abstraction principle. On the basis of
the autobiographical interpretation, then, we concluded that an
affirmation of the legitimacy of abstraction does not
necessarily constitute an espousal of the doctrine of
realism--not, that is, without evidence that Collingwood himself
drew the inference that the central tenet of realism necessarily
implies the principle of abstraction.
However in Chapter III we saw that the evidence of
Speculum Mentis does seem to point to this inference, since
throughout that early work Collingwood does attack realism by
means of a critique of the principle of abstraction. Realism's
basic error, according to Speculum Mentis, is the assumption
that truth is the result of abstraction, or is contained in the
abstract concept, and the correction of this error, which is the
work of philosophy, is the realization that all abstraction is
falsification, because "to abstract is to consider separately
things that are inseparable" (SM, 160). But both in Chapter III
and in Chapter VI we were

322
at pains to show that there is an ambiguity in the meaning of
terms like "separate" and "identical" which render the formu-
lation of the principle of abstraction, as it is expressed in
Speculum Mentis, equally ambiguous. The ambiguity infects the
discussion of realism itself, which on the one hand is respon-
sible for all the root errors that consciousness makes about
itself, and on the other hand is the particular form of dog-
matism associated with historicism. These ambiguities remain
unresolved in Speculum Mentis, which is not surprising from the
point of view of the Autobiography, since his rapprochement
philosophy was at this point still incomplete (cf. A, 107, 116).
But even though Speculum Mentis presumes that an attack on the
principle of abstraction is an attack on realism, there is
evidence that in the same book the principal doctrine of realism
as described in the Autobiography is itself attacked in an
argument employing the same strategy as the later work, and the
argument does not depend on premises tied to a principle of
abstraction. Collingwood argued that to ask what an object world
is or would be apart from a mind which knows it, presupposes the
false suggestion that we can describe that which by definition
is also unknown (SM, 267-68). This argument does not depend on
any formulation of a "principle of abstraction" as essential to
realism (cf. CEPC, 9-12).

323
What we require at this point, if the autobiographical
interpretation is to be upheld, is evidence that in his later
philosophy Collingwood did not, as some of his critics have
charged, surrender to the realism that he had rejected in his
early writings. If we can show that throughout his later
writings Collingwood continued to deny the principal doctrine of
realism, we shall take it that on this issue, at least, the
autobiographical interpretation is vindicated. Since this is the
main concern of this chapter, we shall attempt to adhere as
closely as we can to the fundamental formula of realism as
Collingwood understood it: that "knowing makes no difference to
the object known." We shall not attempt to follow up on all the
evidence that is available to show that he also sought to
overcome all the mistaken consequences that follow from this
premise--notably those concerning moral philosophy2--although we
may
______________________
1 As we pointed out in Chapter III, the autobiographical
formula for realism is phrased in extremely abstract terms, with
no differentiation made between various senses or cases of
"knowing" and "object," and no exact designation of what sort of
"difference" is made (or not made). Throughout this chapter we
shall assume that Collingwood is talking about an epistemological
doctrine, i.e. one which primarily says something about knowing
rather than something about reality.
2 In the bibliography of the Krausz collection of

Critical Essays on the Philosophy of Collingwood there is a note
that the lectures on ethics, currently being edited by Mrs. K.
Collingwood in collaboration with Mr. J. Rusk, are going to be
deposited in the Bodleian Library. Until they are made
available, Collingwood's views on ethics will probably remain
only partially revealed in the remarks in The New Leviathan.
324
make occasional note of these consequences in passing. We shall
also not be concerned with a detailed discussion of the
principle of abstraction, except as it bears directly on the
discussion of the central tenet of realism in the passage under
discussion--our reason being that the connection of these two
principles is not a direct part of the autobiographical
interpretation.
But does the denial of the principal doctrine of realism
necessarily imply the affirmation of an idealist thesis?
Certainly to deny the proposition that "knowing does not make a
difference to the object known" logically implies the af-
firmation that "knowing makes a difference to the object known,"
but is that all that it implies, and is this idealism? In an
intriguing article, John F. Post suggests that Collingwood's
anti-realism argument in the Autobiography not only shows that
the central tenet of realism is false, but is just as effective
against the idealist position that knowing does make a
difference to the object known, since either premise implies the
meaningless assertion that one can know what is simultaneously
declared to be unknown.3 Taking our cue from
_____________________
Since Collingwood's ethics is not what has caused widespread
interest in his philosophy, or even widespread controversy about
it, there is little problem with leaving it out of consideration
here. We do this knowing full well that for Collingwood, even in
the Autobiography, theory and practice cannot be fully separated
from one another.
3 John F. Post, "Does Knowing Make a Difference to What

Is Known?" Philosophical Quarterly, XV, no. 60 (July, 1965), pp.
220-28.
325
this article, we would like to propose a modification of Post's
argument and use it as an abstract framework for our investi-
gation of the realism-idealism controversy in Collingwood's
later philosophy.
In order to be as clear as possible (but only as clear
as necessary) about what is being affirmed by this negation, we
propose the following schema:
A. (1) Knowing makes a difference to the object known.
(2) (1) is false (realism).
(3) (2) is false, because the denial of (1) is absurd

(anti-realism).
But does A (3) then imply that A (1) is true? And is this
idealism? In order to see if this is so, let us propose a more
complicated version of schema A:
B. (1) Knowing makes a difference to the object known.
(2) The object known makes a difference to knowing.
(3) (1) is false and (2) is true (radical realism).
(4) (1) is true and (2) is false (radical idealism).
(5) Both (1) and (2) are false (radical scepticism).
(6) (3), (4), and (5) are false because (1) and (2) are both
true.
Whether or not there are any historically manifested
representatives for the positions we are calling "radical
realism," "radical idealism," and "radical scepticism," we wish
to know how
326
Collingwood stands with respect to these fundamental options. If
we can show that in his later philosophy Collingwood held to A
(3) we will have shown that on this issue the autobiographical
interpretation is vindicated. But we wish to know further if
Collingwood held to B (4), B (5), or B (6). We do know that
Collingwood claimed that he could never plead guilty to a charge
of scepticism (EPM, 223), and that in some sense of the word
idealism (but we do not know yet if, in the later philosophy, it
is in the radical sense as we are using that term), he claimed
not to be an idealist (A, 56-57).4 But can we really defend B
(6)? Does Collingwood really hold that B (1) and B (2) are not
related as disjuncts but as conjuncts?
Our strategy will be simply to take up the later writ-
ings, beginning with the Essay on Philosophical Method, con-
tinuing with The Principles of Art, and concluding with The New
Leviathan, with
___________________
4 In Chapter III we noted that in his early philosophy
Collingwood rejected "subjective" and "metaphysical" idealism,
and although he appeared to approve of an epistemological
idealism, held that "Absolute Knowledge" was humanly possible,
and ruled out a kind of objective idealism that relied on a
Divine Mind as a locus for absolute knowledge, we found these
passages in Speculum Mentis to be plagued with difficulties. In
the present chapter where "idealism" is used without quali-
fication, the epistemological, non-subjective variety is meant.
But we must remain alert to the possibility that what he says
about "idealism" may not be applicable to other forms of
idealism, particularly objective idealism of an Hegelian
pedigree.
327
side-excursions to the Essay on Metaphysics, The Idea of Nature,
and The Idea of History. We shall try to find out what
Collingwood means by the terms "object," "knowledge," and the
relation of "making a difference." Of necessity this involves
setting out in some detail Collingwood's later philosophy of
mind. But in this chapter we will only be concerned with so much
of it as relates to the question of the principal doctrine of
realism--and consequently the discussion will center on the
issue of the perception of the "real world" and the modification
of the object of perception by thought. In the next chapter we
will take up the "higher" functions of mind--conceptual,
propositional, interrogative, and implicative thinking--and
later we will be concerned with the historical imagination and
the ultimate meaning of "difference" and "identity" for a
rapprochement metaphysics. We refer the reader to other
expositions of the levels of practical consciousness in
Collingwood's later philosophy of mind: since these are
peripheral to our main concern, we shall have little to say
about them in this chapter. (See, e.g., Mink, MHD, 82-92, 117
figure 2).
It will help the reader to bear in mind that in this
chapter we shall try to give Collingwood every opportunity to
make clear what is the status of objects of perception, and
whether there is any possibility for the object to make a dif-
ference to the perceptual knowledge of it. We

328
shall therefore ask the reader's forbearance while we examine
some "circumstantial evidence" to confirm our suspicion that
Collingwood was trying to escape from the idealism-realism
dilemma.
Finally, it may be helpful if for orientation the reader
refer occasionally to Tables 7 and 8. While these summaries do
not correspond exactly to the sequence of topics in the chapter,
they will give the reader an overview of the essential features
of Collingwood's philosophy of mind presupposed in the
discussion of this theory of perception.
2. Empirical Thinking and the Essay on Philosophical Method.
Critics of Collingwood have dismissed the view of
philosophy expounded in the Essay on Philosophical Method
primarily because they saw it as a specimen of idealistic argu-
ment (cf. LPC, 259-60). The grounds for this assertion are
primarily (a) the view of predication presented in the essay--
viz. that all philosophical statements are both categorical and
universal; and (b) the use to which this analysis of predication
is put--viz. the defense of the Anselmian ontological argument.5
We will refrain from a full discussion of both of these issues:
the former will be examined in the next chapter, insofar as it
is best discussed in
_____________________
5 G. Ryle, "Mr. Collingwood and the Ontological Argu-
ment," Mind XLIV, 174 (April, 1935), pp. 137-51; reprinted in
Hick and McGill, eds., op. cit. pp. 246-60.
329
connection with Q-A logic and the nature of hypothetical and
categorical judgments; and the latter issue will arise again in
the chapter on metaphysics and rapprochement (Chapter X), since
the ontological argument seems to be the ultimate test of
whether there is an upper limiting case of unity in the scale of
forms of knowledge.
The question at this point is whether or not the Essay
on_Philosophical Method contains any statements which would
substantiate, modify, or disprove the central anti-realistic
thesis of the Autobiography. But it appears that in this work
Collingwood intends to dodge all responsibility for dealing with
such issues. Although the purpose of the essay is to describe
the nature of philosophy, Collingwood announces at the outset
that he will not expand the essay to include the place of
philosophy among the other forms of thought, the place of
thought among the other activities of the mind, and the relation
of mind to the world (EPM, 7-8). For reasons of both expediency
and specificity, he declines to deal with the broader issues
raised by the essay: "though no doubt the thoughts here
expressed have implications in metaphysics, logic, and the
theory of knowledge, these implications will not be discussed"
(EPM, 8).
On the other hand, Collingwood confessed at the end of
the essay that his original contract with the reader (EPM, 45)
to treat the proposed

330
characteristics of philosophical concepts, judgments, and
arguments as only hypothetically or provisionally affirmed has
been violated throughout the book by appeal to
"experience"--that is, philosophical experience, instances in
which the concepts, judgments, and arguments are displayed as
having these characteristics (EPM, 223-24). Therefore the
interpreter is put in the difficult position of taking the
substantive remarks of the essay as in one sense true for
philosophy (as Collingwood sees it) without qualification, and
in another sense taking such remarks as only provisionally true.
Until the actual standpoint of the essay is spelled out (and it
will be to the best of our ability in the next chapter – see
below, pp 461, 592 ff), we can only look for indirect evidence
to add probability to our contention that the essay does not
represent a departure from the central theses of the
Autobiography.
When we turn to the Essay on Philosophical Method for
evidence to determine whether or not Collingwood held to B (3),
B (4), B (5), or B (6), i.e. whether the negative proposition
which denies the realist's principal doctrine logically entails
the positive proposition that knowing does make a difference to
the object known, the closest we come to such evidence is in a
passage in which Collingwood is discussing the problem of the
starting point in philosophy. Philosophy can justify its own
starting point, he says, "only if the arguments of philosophy,
instead of having an irreversible direction from principles to

conclusion, have a reversible
331
one, the principles establishing the conclusions and the
conclusions reciprocally establishing the principles" (EPM,
160). The escape from the charge of vicious circularity is
premised, Collingwood says, on the realization that philosophy,
unlike empirical and exact science, does not proceed from the
unknown to the known: it relies on the "Socratic principle that
philosophical reasoning leads to no conclusion which we did not
in some sense know already" (EPM, 161).
Every school of philosophical thought has accepted this

principle, recognizing that philosophy does not, like exact
or empirical science, bring us to know things of which we
already knew in some way; . . . for if the species of a
philosophical genus overlap, the distinction between the
known and the unknown, which in a non-philosophical sub-
ject-matter involves a difference between two mutually ex-
clusive classes of truths, in a philosophical subjectmatter
implies that we may both know and not know the same thing;
a paradox which disappears in the light of a notion of a
scale of forms of knowledge, where coming to know means
coming to know in a different and better way. (EPM, 161).
Here we see that Collingwood is employing the two most
central doctrines of the essay--the overlap of classes and the
scale of forms (the technical meanings of which will be explored
in the next chapter) to justify the philosophical sense in which
knowing makes a difference to the object known. For if
philosophical knowledge proceeded in the manner of empirical
knowledge, then it would simply be the case that the object made
a difference to the subject, or to the knowing of it, without
any
332
reciprocal co-effect of knowing on the known. But Collingwood is
here asserting that "every school of philosophy" accepts the
principle that in philosophy it is necessary to anticipate its
conclusions "by an experience that possesses them in substance
before its reasoning begins," so that "in philosophy the
conclusions can be checked by comparing them with these
anticipations" (EPM, 163). The object of philosophical
knowledge, therefore, is not something passively given: its
initial datum is always being modified in the light of its
conclusions (cf. EPM, 94-97).
Collingwood spells out for his readers the ways in which
this initial datum of philosophy differs from that of empirical
science. It is different (1) in its relation to the process of
reasoning (premises and conclusions are irreversibly related in
inductive reasoning, whereas in philosophy the arguments are
reversible); (2) in its own constitution (it consists not of
empirical facts but of universal propositions which form the
material or substance out of which the final system is
constructed); and (3) in the way in which it comes to be
possessed (it is apprehended not by perception but by
thought--the experience of a thinker) (EPM,166-69). (These
principles will be more carefully examined in Chapter VIII). But
the differences do not end there, and Collingwood spells out a
further difference as (4) the principles that in philosophy
knowing makes a difference to what is known:

333
In empirical science we begin by perceiving that the facts

are so, and go on by forming a theory as to why they are
so; but in adding this new theory to the old facts we do
not come to know the facts in a different way, we only come
to have something new in our minds . . . alongside the old
knowledge. The process is a special kind of accumulation.
But in philosophy the knowledge . . . why things are so
makes a difference to the knowledge that they are
so . . . . Our knowledge is not simply accumulating, it is
developing; it is improving as well as increasing; it is
widening and strengthening itself at once. There is
consequently a parallel difference in the result of the
process, the conclusion which the argument estab-
lishes . . . . (I)n philosophy, the theory that emerges
from consideration of the facts is no mere hypothesis, it
is the facts themselves more thoroughly understood. (EPM,
169-70; emphasis mine).
The incautious reader may leap to the conclusion that
because Collingwood is contrasting empirical and philosophical
thinking that it is only in philosophy that the proviso holds
that knowing makes a difference to the object known, and that in
empirical thinking perception of the facts leaves them un-
altered, the process being a "kind of accumulation" rather than
a kind of modification. But such a conclusion ignores Colling-
wood's precautionary remark at the beginning of the essay. He
says there that when he talks about "empirical" or "non-philo-
sophical" concepts he is referring not to the concepts as they
are actually used in science or mathematics, but to concepts
which are mistakenly thought to be philosophical: "what I am
discussing, when I distinguish philosophical method from that of
exact science, is not mathematics itself but a certain method,
often mistakenly used in philosophy, which is believed to be
that of mathematics" (EPM, 9).

334
Collingwood's point is that he is not primarily concerned with

making statements about empirical or exact sciences, but about
philosophy; and therefore it is not essential to this point that
what he asserts as true for the empirical or exact sciences be
true or not as they are actually employed in those fields, but
only that what he is asserting as true for philosophy and
philosophical concepts--empirical and exact science as
alternative methods and concepts for philosophy.6
But then does the provision that knowing makes a dif-

ference to the object known apply to the whole of the relation
between knowledge and its object, or just to that relationship
wherein the "object" is taken as the initial datum of philoso-
phy, and "knowing" is understood as that special relationship to
an object that is mediated by the sort of thinking called
philosophy? To answer this question we clearly must broaden our
discussion from its self-imposed limitations in the Essay on
Philosophical Method. This broader framework is provided by the
philosophy of mind he declined to deal with in the essay on
method.
_________________
6 It would be a misinterpretation to take the remarks in
the Essay on Philosophical Method as evidence that at this time
Collingwood held to the strict autonomy of separate modes of
thought--science, history, philosophy, etc.--or that he held to
the existence of a metaphysics of the one, the true, and the
good. The essay leaves such issues unresolved.
335
3. Sensation, Imagination, and Empirical Thought.
Since Collingwood declared in the Autobiography that The
Principles of Art was the second book in his mature series, one
would hope for some direct link between its contents and those
of the Essay _ Philosophical Method--the first book of the
series of those he planned to write as of 1932. A promising case
of a direct relation of subject matters might be the
characterization of "empirical thinking" that appears in both
(cf. PA, 221 and EPM, 164-70), were it not for the fact that the
essay on method so severely limits the applicability of that
characterization. The linkage between the two works is
there--but it is indirect: it appears in the form of their
styles of argumentation, the overlapping of concepts employed in
the theory of art, the scale of forms of experience and of
knowledge that appear in the central chapters of the later work,
and in the overall relation of conclusions to starting point,
etc. specified in the Essay on Philosophical Method and
exemplified in The Principles of Art.
Even if we had already examined these methodological
criteria, it is not of direct concern to us in this chapter to
spell out this indirect relationship. Nor would it be to our
purpose to examine in detail Collingwood's philosophy of art.'
In keeping with our principle of limited

336
objectives, we shall be looking only for evidence for or against
Collingwood's judgment on the central tenet of realism.8
_________________
7 The Principles of Art is divided into three major
sections. Book I deals with the distinction between art and
pseudo-art, the latter being primarily art-craft and the re-
flection on art based on it, this being the "technical theory of
art." In this group of arts falsely so-called are all forms of
what Collingwood calls representation art, amusement art, and
magical art, all of which presume a distinction between means
and end that is alien to true art. Book I ends with two chapters
which treat of art in the correct sense of the term--as
expression and as imagination. For the purpose of clarifying the
meaning of the terms "expression" and "imagination," Book II
begins afresh with an examination of the general
characterizations of experience as a whole, based on the
structure of experience as exhibited by reflection on mental
functions familiar to anyone who thinks. Book II includes
chapters on thinking and feeling, on sensation and imagination,
on imagination and consciousness, and on language. Book III
outlines a theory of art based on the distinctions and conclu-
sions of Books I and II. The critical chapters in Book III are
those which discuss art as language and art as truth.
8 Although our concern is not directly with the theory

of art, it is noteworthy to observe at this point that the
structure of argument in The Principles of Art makes little
sense apart from its anti-realistic orientation. The "technical
theory of art," against which the whole of Book I is a
continuous argument, is based on the realist's separation of end
(the work of art) from means (the artist's activity), and some
of Collingwood's most acrimonious comments in this section are
reserved for the "realistic aestheticians" (who, like Sam
Alexander, maintain that "beauty" is a subjective feeling
aroused by direct acquaintance or perception of the art-work
(PA, 149). These are exceeded only by his contempt for the
psychologizing art-critics (who, like I. A. Richards, attempt to
account for this subjective feeling on the basis of the
stimulus-response model--PA, 29-36; cf. PA, 262-64), and for the
positivistic anthropologists (who, like Freud, Edward Tylor,
James Frazer, and Levy-Bruhl, wind up talking like "one of
Moliere's prize idiots" when they attempt to derive art from
magical beliefs, while utterly misconceiving the nature of myth
and magic in the first place--PA, 58-64).
The theory of art as expression and as imagination which
Collingwood works out is also fully anti-realistic inasmuch as
instead of the art-object
337
Interestingly enough it is in the section on empirical thinking
in The Principles of Art that we find the clearest indication in
that work of the stance that Collingwood is taking on the
central tenet of realism. In developing a theory of imagination
Collingwood discusses the intermediate position of imagination
in the region of experience between sensation and intellect, and
points out that "sensation must be regarded as a flux of
activity in which . . . as soon as the act is over, the sensum
has
________________________________
it makes the function of imagination and its expression in
language primary and even exclusively essential to true art.
Collingwood goes so far in this direction that he asserts that
"a work of art may be completely created when it has been
created as a thing whose only place is in the artist's mind (PA,
130)--or, as he later puts it, in his head (PA, 151). In Book II
the anti-realistic orientation (one might even call it a bias)
is apparent in his final analysis of art as language or as the
imaginative expression of emotion---mental functions which exist
even at the level of the logical and symbolic levels of thought
as the "emotions of intellect" (PA, 252-69). Further indications
of the anti-realism of Book II will be analyzed in the text of
this section of our chapter, and in Chapter VIII. Finally, in
Book III, while saying little directly against realism (as one
might expect, since anti-realism is his starting point and not
his ultimate conclusion), Collingwood counters the realistic
esthetician's subject-object dichotomy in art by discussing the
truth in art as something not in the art work only or in the
subject only, but as a knowing of oneself which is also a making
of oneself (PA, 291-92), and by discussing the active role of
the artist's audience in the work of art itself (PA, 300-24).
Even in so short a summary as this, it is apparent that to
excise the anti-realistic orientation of the book is to make
nonsense of its argument. That other estheticians have
recognized this is indicated by the fact that they point to his
"art-only-in-the-head" overstatement as representative of
Collingwood's supposed over-intellectualization of the subject
of esthetics as a whole: see e.g. S. Langer, Feeling and Form,
Charles Scribner's Sons (New York, 1953), p. 382. If this is
overstatement, we would like to suggest that it is due to his
anti-realistic bias.
338
vanished, never to return. Its esse is sentiri" (PA, 198). At
this point he paraphrases the sort of criticism that someone
like G. E. Moore (he cites him by name) would raise to such an
overstatement, and gives his own response to it:
"Naturally," it may be said, "we cannot see a colour with-

out seeing it. But what could be more absurd than to argue
that, because we have stopped seeing it, the colour has
ceased to exist? For all we know, colours may perfectly
well go on existing when we are not looking at them." The
objection is an excellent example of "metaphysics" in the
sense in which that word has at various times become a term
of merited abuse . . . . The fairy-tale about the existence
of unsensed sense, no doubt, is believed by the people who
indulge in it to be a piece of philosophical
thinking. . . . But even if the belief in question were
true, propositions of this kind would still be nonsense
unless it were true not merely that a sensum exists apart
from our sensation of it, but that in this state of
apartness it is open to our inspection; we have it before
our mind in such a way that we can appreciate its
qualities, compare them with those of other sense, and so
forth. The question is not one of metaphysics, whether
colours exist or not, when we do not see them; it is a
question of epistemology, whether we can "have them before
our mind" in the above sense otherwise than by seeing them;
and if so, how. If we cannot, propositions of the kind in
question are all nonsensical . . . . (PA, 198-99; emphasis
mine. Cf. PA, 170; NL, 5.31-5.39).
The argument in this passage is an application of the refutation
of the central doctrine of realism as found in the
Autobiography. Its thrust is to demonstrate the absurdity of the
epistemological proposition that "knowing makes no difference to
the object known" by showing that this entails knowing what is
simultaneously declared to be unknown--in this case sensing
something that is unsensed.

339
What Collingwood is maintaining in this passage is that either
(1) what Moore and empiricists in general have been comparing is
a sensum and something else like a sensum, but different from
it--in fact an imaginatum9--or (2) they have been talking
complete nonsense--the nonsense of being the assertion of the
existence of unsensed sensa.l0
_____________________
9 Collingwood does not use the term "imaginatum" or even
"phantasm," but on the analogy of his deliberate usage of the
term "sensation" for the act of sensation and "sensum" for what
is sensed in that act (PA, 173) we feel it to be justified.
Collingwood's determination to "speak with the vulgar and think
with the learned" causes a certain inarticulation and headache
for his interpreters, as we shall see when we come to the
altered terminology of The New Leviathan. Cf. PA, 174, and below.
1O This is the argument in support of the thesis that

the esse of sense is sentiri that Donagan claims he failed to
find in Chapter VIII of The Principles of Art (LPC, 32-34, and
32 n. 1): in Chapter VIII it appears on P.170. These pages are
predictably the most confusing in all of Donagan's book. (a)
Donagan first asserts that he could find no such argument, and
then himself cites the argument quoted above (evidently feeling
that it is a bad argument, and a bad argument is no argument at
all). (b) Then Donagan proceeds to argue that "what
((Collingwood)) took for an 'obvious truth' is a howler . . . .
(I)t is not at all obvious that being sensed may not be merely
an episode in the history of a sensum, even though it is its
very essence as a sensum. Collingwood seems to have confounded
the obvious truth that only while it is sensed does a sensum
exist as a sensum with the pleasantry that when it is not being
sensed it cannot exist as anything whatever" (LPC, 33). But
since Collingwood's whole point is the "obvious truth" that
Donagan grants, and not the different proposition that an
unsensed object cannot exist as anything whatever when it is
unsensed, it appears that the howl is on Donagan: he affirms
what he proposes to refute. (c) Then in addition to this
supposed refutation of what is not and yet is an argument,
Donagan proposes what he calls "some positive evidence" for
maintaining that the esse of a sensum is not sentiri." This he
finds in the following argument: "The strongest argument for
asserting that most of our sense are physical objects,
340
This would seem to let the empiricists off the hook by
allowing them to opt for (1), but to choose this option is, for
a radical empiricist, to surrender his empiricism, since the
radical empiricist is one who maintains that all knowledge
arises from sensation, where knowledge is not considered to be
coeval with sensation as originator, but rather subsequent upon
it--both temporally and logically.ll What Collingwood is
proposing instead is a modified empiricism in which thought as
the deliberate achievement of a thinker is present even at the
level of sensation. Since radical empiricists assume that
knowledge arises from the active impression of "sense data" on
passive organs of sensation, it is clear that Collingwood's
position would not be taken by them as a supporting modifica-
tion.12
____________________________
and so can exist unsensed, is that every natural language in
which men speak of what they see, hear, taste, and smell is a
physical object language" (LPC, 34). But as Donagan himself
notes (but fails to apply to his own "evidence") on such grounds
as this one could justify belief in fairy-tales, for every
natural language is as much a fairy-tale language as it is a
physical object language. Donagan's "evidence" (or is it an
argument?) is not as mute as the rock that Dr. Johnson kicked to
refute Berkeley, but it is just as ineffective.
1l Cf. Errol E. Harris, Nature, Mind, and Modern Science

(New York, 1954), p. 119.
12 Mink writes: "Collingwood is an empiricist, but one

who belongs to no identifiable school . . . . As an empiricist,
Collingwood is a radical empiricist, but he is also a radical
idealist, for whom the originative powers of thought are coeval
with the most rudimentary forms of experience" (MHD, 111). Mink
rightly recognizes that Collingwood escapes the classification
of both realism and idealism, that "the traditional opposition
of 'empiricism' and 'idealism' is not a fruitful subject for
investigation or debate, and that all of the interesting
341
Nevertheless this is Collingwood's program in Book II of
The Principles of Art, and the strategy of argument in this
section makes little sense unless it is viewed as an extended
argument which both (1) rejects the empiricist epistemology on
which realistic esthetics is based, and (2) attempts to save
whatever meaning survives this critique. The argument we have
just cited, for example, appears just after a rather extensive
survey of the British empiricists on the relation of sensation
to imagination, during which his main concern was not with a
mere refutation that there is a valid basis for the distinction
between real and imaginary sense, but with extracting whatever
sense he can from the philosophical tradition on the subject. In
order to understand the modification of empiricism that
Collingwood is proposing we must examine this distinction more
closely.
________________________
problems cluster around what it means to be radical" (ibid.).
Unfortunately Mink does not address himself to this question,
and therefore his paradoxical assertions shed no light on how it
is possible for a position to be both empirical and idealist
when both are asserted in their radical forms--that is, as
positions which exclude their opposites. An index of his
confusion on this issue can be found in his use of the term
"genetic" as it applies to the scale of forms of consciousness:
Mink asserts at one point that Collingwood avoids the "genetic
fallacy" of the pragmatists and the non-genetic fallacy of the
empiricists (MHD, 111), but later he says that for Collingwood
feeling and rationality are linked genetically (MHD, 262).
342
In Chapter VIII Collingwood sets up "the problem of
imagination" against the background of the contrast between
feeling and thought, the "problem" being that imaginary sense
seem to be neither feelings nor thoughts. The difference between
these two "features of our experience" is carried out on the
basis of several contrasting characteristics. (1) In terms of
the acts of feeling and thinking, feeling is simple and thought
bipolarly complex: thinking carries with it the sense of success
or failure, truth or falsity, or in general a reference to
self-imposed standards or criteria by which it judges whether
what it has tried to do has been done well or ill. Feeling may
have oppositions and distinctions within itself, but the sense
of its having been done altogether "deliberately" or of its
having succeeded or failed is absent from it (PA, 157). (2) In
terms of what is felt or thought, feelings are private, and
thought is public: what one feels is not open to inspection in
the same way as is what one thinks (PA, 157-58). (3) Thoughts
can corroborate or contradict each other, but feelings cannot
(PA, 158), and the main reason for this is that (4) feelings are
in perpetual flux in which nothing remains the same (what we
take for permanence or recurrence is only a greater or lesser
degree of resemblance between different feelings), whereas in
thinking we are concerned with something that lasts, that
genuinely recurs as a factor in experience (although it is not
necessary
343
to assert the eternity of all objects of thought as such (PA,
159).13 Finally, (5) feeling has the character of a foundation
upon which thinking is erected as a superstructure--which is why
feeling appears to arise in us independently of thought, in a
part of our experience which functions independently of all
thinking and seems unaffected by it, a level Collingwood calls
"psychical" (PA, 163-64).14
Within feeling Collingwood distinguishes two "kinds" of
experience, sensation and emotion--not as a distinction between
two species of a common genus, but as two aspects of one and the
same experience (like "a terrifying red") which can be
distinguished
________________________
13 In the subsequent analysis not all of the disting-
uishing criteria are given equal weight by Collingwood, but (4)
is taken as essential to feeling--i.e., that it is impermanent,
or in constant flux, while thought is something that recurs and
lasts. In section 4 of this chapter we will see how crucial this
distinction is for the characterization of attention--the major
function of imagination. On the contrast between sensation (or
feeling) and thought in his early writings, see SM, 188-89.
14 “I hold that the proper business of psychology is to

investigate this level of experience," i.e. the psychical level,
the level of experience at which we merely feel sensations
together with their peculiar emotional charges, "and not the
level which is characterized by thought" (PA, 164). Collingwood
refers the reader to a note on p. 171 in which he states that "a
science of feeling must be 'empirical' (i.e. devoted to
ascertaining and classifying 'facts' or things susceptible of
observation), but a science of thought must be 'normative', or
(as I prefer to call it) 'criteriological', i.e. concerned not
only with the 'facts' of thought but also with the 'criteria' or
standards which thought imposes on itself" (PA, 171, note; cf.
IN, 230-31).
344
by thought, but always occur together in experience.15 They are
combined according to a definite structural pattern describable
by saying that sensation takes precedence over emotion. The
priority is not temporal, causal, or logical (as grounds and
consequent), but as analyzably distinct: emotion exists as a
kind of charge on a sensation, a charge which can be stripped
off the sensum--a process Collingwood calls "sterilizing" the
sensum by ignoring its emotional charge (PA, 162). This is
accomplished, as he later explains, by an act of attention (cf.
PA, 162-63, 203-11).
Thinking is also distinguised in two ways. In its
primary form thought is exclusively concerned with feeling as
its sole and universal subject matter. First-order thinking is
the process by which we become aware, by an act of attention, of
certain feelings and go on to think of these feelings as
standing in certain relations to other feelings, remembered as
past or imagined as possible (PA, 164-65). All empirical thought
is of this sort: making empirical statements we express our
thoughts about the relations between sensa, actual or possible --
__________________
15 The same distinction between a sensum and its emo-
tional charge, within the experience of feeling, is made in The
New Leviathan, but (1) Collingwood adds a "diffuse consciousness
of feeling" between first-order thought and the purely psychical
level, and (2) distinguishes feeling as an "apanage" of mind
rather than a constituent (only forms of consciousness are
constituents of mind) (NL, 4.1-4.31). Collingwood does not
define "apanage," but it refers to the ordinary English usage:
"Man as mind is consciousness . . . ; he has feelings . . . . "
(NL, 4.2).
345
relations of attention to, selection of, comparison with, etc.
the sense under consideration (PA, 165-66).
Thus our experience of the world in space and time, the

"world of nature" or "external" world, which means not the
world external to ourselves (for we ourselves are part of
it, in so far as "we" are our bodies; and if "we" are our
minds, there is no sense in speaking of anything as
external to them) but the world of things external to one
another, the world of things scattered in space and time,
is an experience partly sensuous (strictly sensuous-
emotional) and partly intellectual: sense being concerned
with the colours we see, and the sounds we hear, and so
forth; and thought with the relations between these things.
(PA, 166) 16
But thought also has a secondary function, and second-order
thought is thought about thought, thought concerned with the
relations between one act of thinking and another, or between
the contents of such acts (PA, 166-67). Collingwood does not
further characterize second-order thought at this point, other
than to contrast it with first-order thought as reason to un-
derstanding, or philosophy to science, and to insist that the
term "experience" be used to cover not only feeling, but first
and second-order thought as well (PA, 167).
The stage is now set for Collingwood's discussion of
the problem of imagination. If sense are in a state of constant
flux, it does not
_______________
16 To take this passage as a piece of subjective idealism
would be a mistake, since Collingwood's intention is not to
dissolve the "external" world of nature, but to avoid an
"inside-outside" dichotomy with no room for bodily overlap.
346
seem possible (given the characterization of thought and feeling
that he had just given) that they could be the object of thought
of any kind, because relations between sense cannot be fixed if
the sense are constantly being both "given" and "taken away."
The difficulty is epitomized in the empiricist formula, "sense
data," where the term "data" implies that the sense are not only
given but retained, which conflicts essentially with the term
"sense" which refers to something which neither persists nor
recurs (PA, 169). The term "sense datum" is therefore either
absurd or refers to something like a sensum but different from
it, and hence the act which grasps it is not properly termed
sensation (PA, 170-71).
In order to see what the distinction actually means,
Collingwood engages in an examination of the historical phase of
the problem from Descartes to Kant, centering his discussion on
the common-sense distinction between "real" and-"imaginary"
sense as two classes of a common genus. He distinguishes threel7
phases of the problem: (1) the identification
____________________
17 Although there is no particular dialectical magic to
the number three for Collingwood, it is interesting to note that
the three phases to his historical discussion of this problem
correspond to the three ways that the classes "real" and
"imaginary" as species of the genus "sense" can be related: (1)
the classes can be simply identified (17th century ration-
alists); or (2) abstractly distinguished (English empiricists);
or (3) arranged on a scale of overlapping classes (Kant). For
evidence that this is not an accidental arrangement, see PA,
187, n. 1.
347
of sensation with imagination by the 17th century rationalists;
(2) the defense of the common-sense distinction by the English
empiricists; and (3) the Kantian solution to the problem (PA,
187).
(1) In the 17th century, Descartes, Hobbes, and Spinoza
found that by direct inspection, real and imaginary sensa could
not be distinguished. At bottom both were identical: both
exhibit a kind of confused disorder, and it is only by reasoning
that the two can be distinguished (PA, 175-76). But although
Collingwood ultimately agrees with the conclusions (a) that
there is no direct or internal test by which the two classes of
sense can be distinguished, and (b) that it is only on the basis
of some kind of thought that the distinction can be made (PA,
194), he is not content to merely identify them. He therefore
goes on to examine the defense of the commonsense distinction by
the English empiricists.18
__________________
18 CF. NL, 4.83-4.89: "For Plato, sensations and emo-
tions cannot be knowledge because they lack the precision which
knowledge must have. For Leibniz, feeling in general is confusa
cognitio. I do not accept either view in its entirety. Plato
thought that knowledge cannot even rest on a foundation of
feeling, because feeling is too vague; knowledge must be the
work of pure thought operating all by itself. But what a
foundation needs is strength, and strength is what feeling has.
Leibniz thought feeling was confused knowledge, and to clear up
the confusion is to purge it of what makes it feeling and leave
it knowledge. But feeling is not knowledge at all; it is
feeling; and if you could purge it of what makes it feeling
there would be no residue. Yet each was right in saying that
feeling is confused or indistinct. That is why one should not
try to define it or any kind or element of it; but only to give
examples and say: 'This is the sort of thing to which the word
refers."'
348
(2) The common-sense distinction was reinstated by the
English empiricists, Locke, Berkeley, and Hume, but each failed
to adequately justify it. As if recognizing some inadequacy to
the distinction, each put forward a pair of criteria for
discriminating real from imaginary sense. Locke's "real ideas"
are distinguished from "fantastical" ones insofar as (a) the
latter are related as ectype to the archetypal "real
ideas"--these being the "original" or "real being" of the
objects; and (b) by an "introspective" criteria whereby
imaginary sense are the result of a voluntary activity, whereas
real sense are not--the two cases being distinguished by simple
introspection as to how the ideas arise in us (PA, 17677).19
Similarly Berkeley distinguished "ideas of sense" from
________________________
19 Cf. Donagan: "In his philosophy of mind Collingwood
was fundamentally anti-Cartesian; for he . . . repudiated Des-
cartes' doctrine that acts of consciousness are, as it were,
self-illuminating. You cannot know your own mind by turning an
inner eye on its operations, because introspection can do no
more than to bring to mind something of which you have already
become aware" (LPC, 25). "All that introspection could do would
be to reproduce your visual field in a second 'inner' visual
field; and there is no reason to suppose that having a second
visual field or a second auditory field would make you conscious
of your first one" (LPC, 41). It is clear that Donagan has a
different meaning in mind for "introspection" than does
Collingwood. Collingwood attributes the introspection theory to
the empiricists--Locke, Berkeley, and Hume-and means it to refer
to the distinction between real and imaginary sense on the basis
of the degree of control the subject has over such experiences,
the former being relatively involuntary (PA, 177-79). However
there is a secondary sense of the term "introspection" as used
by Collingwood, this being closer to the sense that Donagan has
in mind. At PA, 205 Collingwood writes that introspection as the
method of putting questions to consciousness cannot tell us
anything about the psychical or purely
349
"ideas of imagination" in two ways: (a) by a restatement of
Locke's introspection theory (the "strength and liveliness"
criteria which, Collingwood argues, only makes sense as meaning
that, e.g. "a real sound is heard whether we will or no, whereas
an imaginary one can be summoned up, banished, or replaced by
another at will" (PA, 178-79)); and (b) the "relation theory"
which suggests that ideas of sense are subject to the laws of
nature, but ideas of imagination are not (PA, 179-82). Finally,
Hume dropped Berkeley's relational theory and (a) reinstated the
introspection theory: "impressions" (real sense) are
distinguished from "ideas" (imaginary sense) by the "degree of
force and liveliness with which they strike the mind"--which
Collingwood once again takes to mean
_____________________
sentient level of experience; the correct method for such
knowledge is to be found in a well-grounded behavioral
psychology. But although Collingwood is rejecting introspection
here as a method for inquiring about the purely psychical level,
he is not rejecting it as a method for understanding the
functions of consciousness. Indeed, so far from being
"anti-Cartesian" in this sense, Collingwood's mature philosophy
of mind is nothing but introspective: it consists entirely of
"soundings" at various levels of consciousness (NL, 9.35-9.4),
made by putting questions to (conscious reflection on) lower
levels (cf. PA, 205-06), which might just call on Donagan's
"Principle of Order" as a blessing on the project. It is
therefore hard to see what sense it makes to assert that the
introspective method was rejected by Collingwood, or how in
rejecting it Collingwood was being fundamentally anti-Cartesian.
To be anti-Cartesian Collingwood would have to maintain that
conscious acts cannot be objects to other conscious acts, which
is certainly not what Collingwood maintained. And if being
anti-introspectively anti-Cartesian means rejecting that there
is an "inner eye" that is something other than a conscious act,
then it is probable that Descartes himself was not a Cartesian,
for certainly his introspective acts were acts of consciousness.
350
the inability of our minds to set purpose to, control, excite,
suppress, or modify our sensory experience (PA, 182-83). (b) But
Hume also recognized (but yet minimized) the difficulties in-
volved in cases like dreams, hallucinations, and violent emo-
tional upheavals, in which the criterion does not apply, since
in such cases sense are not subject to voluntary control, but
are nonetheless certainly imaginary. In such cases he therefore
falls back on the relational theory to account for the
distinction (PA, 183-85).
Collingwood argues that none of the empiricists' cri-
teria is successful. In Locke's representative theory the
substitution of the relation of ectype to archetype for the
relation of imaginary sense to real sense is inconsistent with
his causal theory, which describes real sense as caused by ob-
jects (PA, 177)--it assumes the untenable identification of
resemblance and causation which Hume so vigorously refuted. If
real sense cannot be accounted for in this way, their dis-
tinction from imaginary sense collapses. The introspection
theory similarly fails on the grounds that the exceptional cases
of hallucinations, dreams, etc. proves not the rule that real
sense, but not imaginary sense, are involuntary, but rather that
will has little to do with the distinction: imaginary sense are
often less subject to control than are real sensa (PA, 179).
Finally, the relational theory fails because although imaginary
sense may not obey all the laws of nature, they

351
obey some, and are furthermore subject to "psychological laws"
which cannot be thoroughly disentangled from laws of nature.
Furthermore the relational criterion is circular: we can only
know what the laws of nature are by experience, or by studying
real sensa--but to distinguish real sense by appeal to these
laws presumes the distinction to be established (PA, 182).
Finally, (3) Kant corrects the relational theory by
showing that first-order laws of nature imply second-order
"principles of the understanding," and argues effectively that
while a sensum may be "wild" in the sense of not belonging to a
"family" of known natural laws of the first order, it cannot be
wild in the sense of failing to belong to laws of the second
order.
It is a principle of the understanding that every event

must have a cause. No event that comes under our notice can
escape this principle. The furthest length to which it can
go towards wildness in that direction is a failure on our
part to discover what, in particular, its cause is. Thus
Kant's discovery of second-order laws involves the
discovery that there are no wild sensa. At the same time it
enables him to explain what we mean when we say that wild
sensa exist. We are saying that certain sensa, though . . .
we know they must admit of interpretation, have not yet
been actually interpreted . . . . Instead of trying to
conceive real sensa and imaginary sensa as two co-ordinate
species of the same genus, . . . he conceived the
difference between them as a difference of degree. For him,
a real sensum can only mean one which has undergone
interpretation by the understanding, which alone has the
power to confer the title real; an imaginary sensum will
then mean one which has not yet undergone that process.
(PA, 186-87; emphasis mine).
352
It is noteworthy that Collingwood takes the Kantian position as
the concluding phase of the controversy (PA, 187), and does not
subject it to any criticism as he had the previous positions. If
the reader is left in any doubt that Collingwood wishes to
accept this solution to the problem as final, his footnote to
this passage affixes his final seal of approval: he cites the
Essay on Philosophical Method to the effect that what is
involved in Kant's doctrine is an instance of the rule for
philosophical concepts wherein differences of degree always
imply difference in kind (PA, 187, n. 1).20
___________________
20 That Collingwood's philosophy of mind would accept as
final a viewpoint that is ultimately Kantian would not be
accepted by Lionel Rubinoff, who would argue that a further step
is required to raise it to the "absolute standpoint" at which
subject and object are identified. But there is nothing in The
Principles of Art to indicate that Collingwood accepted anything
but the Kantian viewpoint as final on the question of real and
imaginary sense, and that he looked with anything but favor upon
Kant's limitation of thought to the bounds of possible
experience. The Hegelian absolute standpoint is nowhere in
evidence in this work, except insofar as that standpoint is
implicit in the Kantian philosophy. It is therefore going beyond
Collingwood's last word to assign The Principles of Art to the
"third ontological level" at which subject and object are
identified, as Rubinoff does (CRM, 373). On the other hand to
classify Collingwood as a Kantian would be equally misleading,
even if it would be not as misleading as assimilating him to
Hegelianism. (While statistical frequency of citations of
authors does not count for much in philosophical works, it is
interesting to note that Hegel has only one reference in
Collingwood's index to The Principles of Art, and this an
unfavorable one, while Kant has 14, mostly favorable).
Collingwood never accepted any categorial schema of logical
relations as final, even though he approved of the Kantian for-
mulation of them: see NL, 5.66-5.67, 7.34. Nor is there any hint
of a "deduction of the categories" in Collingwood, either in the
Kantian or the Hegelian sense of the term "deduction." And
finally, Collingwood never elevated art into one of the modes of
absolute experience, as Hegel does: cf. IH, 121, 311
353
The Kantian solution also makes it possible to solve a
long-standing problem in the traditional controversy--the
problem of illusory sense, with its two variants, the terms
"appearance" and "image" as used in false theories of percep-
tion. It would seem like a development of the common-sense
theory that illusions are imaginary sense mistaken for real
ones. But to assign illusion to the class of imaginary sense is
misleading, because there is no special quality that sense can
have in virtue of which they are illusory. If there were it
could never be detected, since mistakes of the same general kind
are made about real sense, especially in unfamiliar cir-
cumstances (e.g. the child or savage looking at a mirror for the
first time and thinking the image to be behind the plane of the
mirror, so attempting to reach at it through or behind the
glass). It is not a special quality of imaginary sense, it is a
mistaken interpretation.
We were wrong, therefore, to define illusory sensa as im-

aginary sensa which we mistake for real ones. Illusory
sensa can be defined without referring to the distinction
between imaginary and real. Any sensum is illusory in so
far as we make an error about it. This error does not
consist in mistaking it for a different sensum . . . . All
that there can ever be in a sensum is directly present to
us in the act of sensation. We may be mistaken in believing
that another person in our circumstances would have a
similar one; but we cannot, in seeing a red patch, mistake
it for a blue one. The mistakes we make about our sensa are
mistakes about their relations with other sense, possible
or expected. (PA, 189).
__________________
14 Cf. also Rubinoff, CRM, 210-11. For Rubinoff's interpre-
tation of Collingwood on perception, see CRM, 107-12.
354
Similarly the term "appearance" is not a quality of a sensum
(such as the quality that the short man has who is further away
than this closer tall man, the two being "really" the same
height). On the contrary the term refers to the interpretation
placed on the two sensa, first of all as resembling one another
(as men) and secondly the judgment that the further or future
experience of these sense will or will not continue to show the
same kind of resemblance (when standing next to one another they
will have the same height) (PA, 190-91).
And finally, the term "image" is used to describe the
kind of relation that exists between, for example, a photograph
and the scene photographed, both of which are directly
inspectable and comparable by us in distinct acts of perception.
But if perception itself is explained on this analogy, it
presumes that we can compare our sensa (the images) with their
originals (the real objects), a condition of the analogy which
cannot be fulfilled (PA, 191-92).
Collingwood's concluding summary of the discussion of
the common sense distinction between real and imaginary sensa is
worth quoting:
Sensa cannot be divided, by any test whatever, into real

and imaginary; sensations cannot be divided into real sen-
sations and imaginations. That experience which we call
sensation is of one kind only, and is not amenable to the
distinction between real and unreal, true and false,
viridical and illusory. That which is true
355
or false is thought; and our sensa are called real and

illusory insofar as we think truly or falsely about them.
To think about them is to interpret them, which means
stating the relations in which they stand to other sensa,
actual or possible. A real sensum means a sensum correctly
interpreted; an illusory sensum, one falsely interpreted.
And an imaginary sensum means one which has not been
interpreted at all: either because we have tried to
interpret it and failed, or because we have not tried . . .
. They are sense in respect of which the interpretative
work of thought has been done well, or done ill, or left
undone. (PA, 194).
4. Attention, Freedom and Corrupt Consciousness.
How does this discussion of sensation and imagination
stand with respect to Collingwood's refutation of the principal
doctrine of realism? We stated above that this denial entails
the affirmation that knowing makes a difference to the object
known. We begin to see the way in which this affirmation is made
in Collingwood's philosophy of mind. For first-order thinking,
"knowing" is an act of interpretation, and the "difference" it
makes to the object is the difference between a sensum
uninterpreted, a sensum interpreted, a sensum interpreted
incorrectly, and a sensum interpreted correctly. The "object"
for first order thinking is sensory experience ("feeling" in
Collingwood's terminology), the object insofar as it is felt.
The term "real" in this context is applied or withheld as a
judgment about sensa, and has a double sense: a rainbow, for
example, is "really there" in the sense that I see it--but then
so is the imaginary beast in the dark corner of the room, and
the snakes of delirium tremens. But in another sense the rainbow
is really there as the rain and

356
the sunshine (which I do not see), or the things in terms of
which I interpret my sensa (PA, 192-93).
And this is Collingwood's ultimate reply to the real-
ists' assertion of an object which is unaffected by the knowing
of it. Collingwood's answer, as stated in his reply to Moore, is
that it must be an entity that is imaginary--i.e. the assertion
of a sensum as already interpreted, or the mere assertion of the
possibility of repeating the interpretation of a sensum as "the
same" as a prior one, or as "unchanged" between episodes of
feeling. "And it is imagination, not sensation," Collingwood
writes, "to which appeal is made when empiricists appeal to
'experience"' (PA, 203). But what the radical empiricist wishes
to assert by an object unaffected by the knowing of it is an
object that is an uninterpreted sensum, and about this it would
be ridiculous to say that it would be just the same as it is if
it were not known at all.
But how does one know, on Collingwood's grounds, when a
sensum is correctly interpreted? If he cannot answer this, the
realists' withers remain unwrung. What the realist wishes to
know is how sense-experience, if it is a pure flux of unrelated
and fleeting sense, can give us knowledge of an "external
world.21 Merely denying that there is such a
____________________
21 This is the epistemological question of perception:
see E. Harris, Hypothesis and Perception (London, 1970), p. 237.
357
thing as an external world (as Collingwood seems to be implying
in the passage at PA, 166) does not solve the problem. It merely
shifts the locus of the question from epistemology to meta-
physics: for what then are the colors, sounds, etc. that con-
stitute the objects of acts of sensation? We have seen that
Berkeley's suggestion that they are mere ideas is rejected by
Collingwood as is the suggestion that it is a metaphysical
problem rather than an epistemological one (PA, 199). Colling-
wood does not mean to imply that there is no such thing as a
world of objects, or that "a world of objects external to the
knower and existing independently of him" is an utterly non-
sensical expression. What he is denying is that "external to"
and "independent of" are properties of sensa; they are actually
interpretations of imaginata (i.e. what a sensum would be like
as it exists in an unperceived state).
The real issue as Collingwood sees it is an epistemo-
logical one, and centers on the supposed passivity of sensation.
The thesis to be examined is that, in some way not clearly

defined, imagination contrasts with sensation as something
active with something passive, something we do with some-
thing we undergo, something under our control with some-
thing we cannot help, a making with a receiving . . . .
((But)) it is not a distinction between activity and pas-
sivity as such. Sensation is an activity. Even if we do it
only because we are stimulated to it by forces outside our
control, it is still something we do . . . . Nor is it a
distinction among passivities (things that happen to us, as
distinct from things we do, according as they are done to
us by external bodies impinging on our own, or by changes
arising within our own organism . . . .
358
For sensation, as well as imagination, is on its bodily

side a change arising within our own organism, and due to
the energies of that organism itself . . . . Nor is it a
distinction among activities (things we do) between those
we do of our own choice and those we cannot help doing. It
is in fact easier to stop seeing this paper, by shutting
one's eyes, than to stop imagining the frightful accident
which one saw yesterday. (PA, 196-97; emphasis mine).
The conclusion seems to be that if there is any distinction at
all between imagination and sensation the basis for this dis-
tinction must fall altogether outside the category of activity--
passivity (cf. NL, 5.4-5.49). This category may be at home in
the mechanical realm of matter, where forces are exerted by
pushing and pulling; but in the realm of thought it is not
adequate. Thought is free as matter is not: to be a thought is
to have an essential freedom to succeed or fail--to be "bipolar"
as Collingwood says.
The empiricists' problem of "sense data" is therefore
poorly conceived, because the assumption of sensa as "given"
(from the Latin, datum) is false as a presupposition (PA, 196,
200). The empiricists (including Hume) failed to appreciate
adequately the intermediating function of imagination in the
sensation-intellection relation. The freedom of imagination
falls between the freedom of sensation (which is the freedom of
a spontaneous, living, sentient organism) and that of in-
tellection (which is the freedom of choice between alternatives
consciously conceived) (PA, 197). The freedom of imagination
consists in the alternatives of recognizing a feeling as
belonging to oneself or
359
in refusing to so recognize it (PA, 224). This refusal or the
process of "disowning" feelings is what Collingwood calls
"corrupt consciousness" (PA, 217-19).
But to understand the intermediation of imagination, its
freedom, and the corruption of consciousness it is necessary to
identify and characterize the function of imagination which
transforms sensa into imaginata. This function is "attention"
(alternatively called "awareness" or simply "consciousness"--PA,
206).
Thought . . . detects "relations between sensa" . . . But

in order that we may detect resemblances or any other
relations between things, we must first identify each of
them: distinguish each as a thing by itself and appreciate
its qualities as those qualities we find it to possess,
even though as yet (not having determined their relations
to qualities found elsewhere) we are not in a position to
name them . . . . This act of appreciating something, just
as it stands, before I can begin to classify it is what we
call attending to it. (PA, 203).
Collingwood is careful to distinguish attention from both in-
tellection and sensation. As distinct from intellection, at-
tention selects and divides, but it does not abstract: abstrac-
tion is a higher function involved in the formation of concepts
("redness" as opposed to "this red patch"), whereas attention
"appreciates" a sensible quality "as it presents itself to us, a
concrete individual" (PA, 204). Nonetheless attention is
"thought in its
360
absolutely fundamental and original shape" (PA, 216).22 At the
merely psychical level, the distinction between conscious and
unconscious does not exist. But the instant there is added the
activity of attention, "the block of feeling present to the mind
is split in two" the conscious part being that part of feeling
attended to, the unconscious part being the ignored remainder of
the field, the negative counterpart of attention, or its
penumbra (PA, 204-05).23
____________________
22 Since attention is both (a) thought in its minimal
sense, and (b) not abstract, Donagan's contention that Colling-
wood's later philosophy of mind (which starts with The
Principles of Art) is founded on the recognition that "all forms
of thinking-from the highest to the lowest, are conceptual,
and . . . all concepts are abstract" (LPC, 14) is obviously
false. Since an essentially similar characterization of the
function of attention is made in The New Leviathan (see NL,
4.18, 4.24, 4.5-4.6, 7.2-7.39) it is doubtful whether Donagan's
contention holds even for The New Leviathan alone (in section 6
of this chapter we shall see that The New Leviathan poses
special problems for any consistent interpretation of
Collingwood's philosophy of mind). In The Principles of Art, at
least, it is clear that "thinking" is a term whose extension is
broader than Donagan is willing to allow: it includes
imagination, which is not abstract, but concrete. In a later
passage Collingwood contrasts analytic and abstract thought with
imagination, and writes: "These ((i.e. analytic and abstract
thought)) are not the only kinds of thought . . . . They are
given merely as examples of what ((intellect)) does which im-
agination, never analytic and never abstract, cannot do" (PA,
254; cf. PA, 287). As we shall shortly see, attention is a
function of imagination.
23 Collingwood recognized that there is a proper func-

tion for psychology in the study of "unconscious feeling" at the
psychical level, but insisted that it still rested on a more
general philosophical principle--this principle being that
consciousness has a double object including not only sensation
but the object of sensation (PA, 206). Collingwood's attacks on
psychology all have this intent behind them as an unspoken
premise: empirical psychology cannot absorb the labor of
philosophical psychology because it depends for its existence
361
On the other side attention is distinguished from feeling in
having the bipolarity that is proper to thought:
Attention . . . has a double object where sentience has a

single. What we hear, for example, is a sound. What we
attend to is two things at once: a sound, and our act of
hearing it. The act of sensation is not present to itself,
but it is present, together with its own sensum, to the act
of attention. This is, in fact, the special significance of
the con- in the word consciousness; it indicates the
togetherness of the two things, sensation and sensum, both
of which are present to the conscious mind . . . . Thus the
difference between seeing and looking, or hearing and
listening, is that a person who is said to be looking is
described as aware of his own seeing as well as of the
thing he sees. (PA, 206).
We see here a clearer statement of what Collingwood meant in
Religion and Philosophy when he said that the esse of mind is
not cogitare simply, but de hac re cogitare (RP, 100; FR, 172).
What is essential to the act is not only sensation or its ob-
ject, but both taken together. Collingwood's use of the term
"consciousness" in connection with higher processes of thought
indicates that the same is true for all the successive levels of
thought (cf. IH, 291, 306).24
__________________
on the philosophical analysis of consciousness. But Col-
lingwood's insistence on this point should not blind the reader
to the complimentary affirmation that philosophical psychology
cannot absorb empirical psychology, because the psychical level
(at which consciousness and unconscious are not distinguished)
merges with the physiological and is not directly analyzable by
the phenomenological analysis of consciousness.
24 Because sensation proper (as distinct from percep-

tion, which is sensation attended to) is "not present to itself"
and has a single object rather than a double, it cannot properly
be called "consciousness" in the technical sense of the term as
defined at PA, 206. Sensation proper exists
362
Because attention is thought, and because thought is
bipolar, it is possible for attention to make mistakes--and this
is the sense in which it is free in the minimal sense of that
term as applicable to thought. A conscious being is not free to
decide what feelings he shall have, but he is free to decide
what feelings he shall place in the focus of his consciousness.
This is not a mere response to stimulus (which is the level of
freedom of a sentient organism), but it is also not yet the
freedom of choice of alternative plans of action. It is
intermediate between the two (PA, 207-08). The attention to
feeling modifies it by "domesticating" it, allowing the one who
attends (the rudimentary self) to dominate the feeling and to
perpetuate it indefinitely (PA, 20609). But it does so only at a
certain risk--the risk inherent in the freedom it possesses. The
risk is the danger of becoming a corrupt consciousness:
As thought ((attention)) must have that bipolarity which

belongs to thought as such. It is an activity which may be
well or ill done;
____________________
at the psychical level of pure feeling, at which the distinction
between conscious and unconscious does not yet exist. The
objects of sensation are merely feelings present to conscious-
ness; by attending to them, consciousness perpetuates them and
thus prepares them for further acts of interpretation. However
as we shall see in section 6, in The New Leviathan Collingwood
introduces a "diffuse consciousness of feeling" at a level below
that of attention proper, and assigns "selective attention" to
the second level of thought--thought about thought--as one of
its practical functions. It is clear that "consciousness" is not
being used there in its technical sense.
363
what it thinks may be true or false. But this seems

paradoxical; for since it is not concerned with the
relations between things, and hence does not think in terms
of concepts or generalizations, it cannot err, as intellect
can, by referring things to the wrong concepts . . . . But
the statement "This is how I feel" does imply bipolarity.
It has an opposite: "This is not how I feel;" and to assert
it is to deny this opposite . . . . ((Consciousness)) lives
by rejecting error. A true consciousness is the confession
to ourselves of our feelings; a false consciousness would
be disowning them, i.e. thinking about one of them "That
feeling is not mine" . . . . I call this the "corruption"
of consciousness, because consciousness permits itself to
be bribed or corrupted in the discharge of its
functions . . . . (PA, 216-17).
The corruption of consciousness (also called bad-faith, insin-
cerity, self-deception, or in psychological terms, repression,
projection, dissociation, and fantasy-building) belongs to
neither of the commonly recognized species of untruths—errors or
lies; these belong to higher levels of thought (PA, 218-19; cf.
PA, 115, 283). But it is an example of untruth in its minimal
sense, and hence of evil; not exactly a crime or vice (because
not fully a choice) and not exactly a disease (because not
suffered passively), corrupt consciousness is "a kind of sheer
or undifferentiated evil, evil in itself, as yet
undifferentiated into evil suffered or misfortune and evil done
or wickedness" (PA, 219-20).25
______________________
25 It is not difficult to see why Collingwood regarded
realists as arch-propagandists of a coming fascism (A, 167).
Fascism is a celebration of irrationalism and is hence the re-
sult of a kind of corruption of consciousness: it disowns ra-
tionality and its attendant emotions. Realism, by dissociating
theory and practice, and denying that knowledge has any effect
on its object, and furthermore by treating moral subjects in a
364
We have seen the sense in which Collingwood regards the
act of attention as intermediate between sensation and
intellection, the extent of its freedom, and its rule in the
corruption of consciousness. It remains to determine the way in
which attention is to be regarded as a function of imagination.
The act of attention, Collingwood has been insisting, is the act
which converts "impressions" (feelings) into "ideas"
(imaginata). In so doing it modifies the feelings attended to:
feeling is no longer something in complete flux, it is dominated
and perpetuated by the act of attention. The modified feeling is
no longer a "bare feeling" it is a feeling of which we have
become conscious (PA, 209, 213). To say that a feeling is
attended to is to say that it is an uninterpreted sensum that is
ready for interpretation, so that "regarded as names for a
certain kind or level of experience, the words consciousness
((or attention)) and imagination are synonymous" (PA, 215). But
within a single experience they can be distinguished. Attention
is the act which converts a bare sensum into one ready for
interpretation, and imagination is the result of that
conversion. "Imagination is thus the new form which feeling
_____________________
purely theoretical way, opens a path for disowning rationality
by denying its applicability--and hence its expression in moral
acts. Whatever moves passions is whatever causes acts, and
rationality (ex hypothesi) does not move passions. This paves
the way for a political movement that plays directly on passions
without regard for reason. Cf. EM, 133-42; NL, 35.43-35.44.
365
takes when transformed by the activity of consciousness" (PA,

215).26
But once again we must raise the empiricist's question:
if an imaginatum is a sensum that is ready for interpretation,
is that readiness something that is found in the sensum or
something given to it by an act of consciousness? It is clear
that Collingwood is straining to make consciousness, or thought
in its minimum sense, speak with an active voice; but it is not
clear whether there is any contribution to this conversation
from the side of the sensa, or from a world of nature beyond
them. It is also clear that Collingwood is maintaining his
assertion that imagination is indifferent to the reality or
unreality of its objects (PA, 136) --an assertion that dates
back to Speculum Mentis (SM, 60, 112). But if feeling is
ambiguous with respect to the category of activity-passivity
_____________________
26 The term, "imagination," like the terms "thought" and
"consciousness," has a very broad extension in Collingwood's
philosophy. Although in The Principles of Art his discussion
tends to confine it to a level of thought between sensation and
intellection, it is capable of operating at higher order levels
of thought as well. In The Idea of History he speaks of an "a
priori" imagination (and we may recall that "a priori" is a term
he uses to distinguish thought from feeling). This a priori
imagination (a) does the entire work of historical
reconstruction, (b) operates in artistic creation, and (c)
functions in perception by "supplementing and consolidating the
data of perception in the way so well analyzed by Kant, by
presenting to us objects of possible perception which are not
actually perceived" (IH, 241-42). If a priori imagination
operates even at the level of historical reconstruction, it
clearly escapes confinement to a strictly intermediate level in
the scale of forms of knowledge. It is simply, in Kantian terms,
the faculty of re-presentation.
366
(and in The New Leviathan Collingwood explicitly extends this
ambiguity to all the other Kantian categories), and if
imagination is indifferent to the distinction between real and
unreal, the reader is left wondering how any interpretation can
be lifted off the face of an utterly indeterminate flux of
sensa. And if it cannot, the obvious question is how knowledge
of an "external world" (i.e. the world extrinsic to our organs
of sense)--in Collingwood's terms, the "not-self," is possible
at all.
5. Idealism and the Limitations of Phenomenology.
In The Principles of Art (and to a lesser extent even in
The New Leviathan) Collingwood seems to deliberately choose
language that refuses to take sides in the realism-idealism
controversy. As Louis O. Mink points out, there are some lethal
booby-traps left for anyone who tries to decide the issue from
Collingwood's texts (MHD, 112). A case in point is the following
passage, in which Collingwood is trying to reconcile the
characterization of imaginata (a) as uninterpreted sensa, and
(b) as feeling dominated and perpetuated by consciousness:
Now it has already been argued . . . that the work of de-

termining relations between things must depend on something
prior to it, namely having these things held before the
mind in such a way that we can compare them with one
another, and so become able to see how they resemble one
another, and so forth. We must know what each is in itself
before we can decide how they are related. To know what a
given thing is in itself is not, of course, the same as
367
knowing what kind of a thing it is . . . . Our knowledge of

what it is in itself, if we try to express that in words,
will be stated in some such phrase as "this is what I see,"
or, since to call my act one of seeing is already to
distinguish, "this is how I feel." . . . And we become able
to say this, not through bare sensation, but through
consciousness of sensation. What makes us able to say it is
that we have, by the work of attention, at once selected
and perpetuated some element which we find in the field of
sensation, and some corresponding element in the sensory
act. (PA, 212-13).
Collingwood apparently finds no difficulty in asserting of one
and the same sensum that it has elements which are at once se-
lected, perpetuated, and found, and that corresponding elements
exist in the sensory act. A few pages later a similar land mine
is planted for explorers at a higher level: "The work of
intellect," Collingwood writes, "is to apprehend or construct
relations" (PA, 216, 255-56). Which is it? The reader will not
get any direct satisfaction from Collingwood in answer to this
question.27 The reason for such deliberate ambiguity may be
either that Collingwood
____________________
27 Cf. Mink, MHD, 112-13: "Collingwood does not
want to decide whether thought 'apprehends' or 'constructs,'
'finds' or 'puts.' In his view, it does both. As the activity of
converting implicit differences into explicit distinctions, it
seems to itself, at any level, to be apprehending. But as an
object of consciousness to a higher level, it seems to be con-
structing. Experience, one might say, is the realist, reflection
on experience the idealist. Neither is false; what is false is
the presumption that there is irreducible logical in-
compatibility between the theories expressive of each. In the
dialectic of theories, realism states the viewpoint of any level
of consciousness from its own standpoint, idealism viewpoint of
any level of consciousness in its reflection on a lower level.
The theory of levels accounts for each by assigning to each a
function which cannot be usurped by the other." Cf. Rubinoff,
CRM, 29, 59, 116, 136 ff.
368
wished to leave the question open, or else that Collingwood held
to the simultaneous truths that (even at the level of
perception) knowing and the known interact, each making a
difference to the other.
There would be no way to decide between these two al-
ternatives were it not for a few pieces of circumstantial
evidence. We have already seen how Collingwood argues directly
against the realists' thesis that knowing (in this case, con-
sciousness) makes no difference to the object known (the sen-
sum): it is carried out directly in his attack on Moore (in the
doctrine of the unsensed sensum) as well as indirectly in his
discussion of the modification of feeling by attention. We have
also seen how Collingwood's attempt to save whatever sense he
could from the empiricists' tradition concerning "sense data"
led to his characterization of imagination (in the form of
selective attention) as intermediary between sensation and
intellection, and his dismissal of "sense data" as interpreted
imaginata masquerading as bare sensa. We have furthermore argued
that the entire discussion about empiricism would be a
non-sequitur and a diversionary aside if it were not that he
hoped to save something from the tradition. What he saved was
not, certainly, the ambiguous phrase, "sense data." Nor was it
the disjunctive proposition that either knowing affects the
known or the object affects the knower (but not both):

369
this disjunction is eliminated with the argument that
activity-passivity is a category inapplicable to bare,
uninterpreted sensation. What is saved is the role of thought as
active down to the level of sensation, where it is present as
the act of attention which alters feeling by perpetuating and
dominating it. But what is it that is dominated and perpetuated?
Here we must enter our circumstantial evidence.
A.--In the theory of art propounded in The Principles of
Art (see notes 7 and 8 above) Collingwood directly attacks the
realistic estheticians for distinguishing too sharply between
what is "found" in a work of art and what is "brought" to it by
an observer. The position of the realistic estheticians, who
maintain that "beauty" is something subjective that is "imputed"
to the art-object, is an attempt to rehabilitate an ancient
technical theory of art based on the false analogy between art
and craft. Collingwood describes what he takes to be their
position as follows:
(T)he technical theory depends on distinguishing what we

find in the work of art, its actual sensuous qualities, as
put there by the artist, from something else which we do
not strictly find in it, but rather import into it from our
own stores of experience and powers of imagination. The
first is conceived as objective, really belonging to the
work of art: the second as subjective, belonging not to it
but to activities which go on in us when we contemplate it.
The peculiar value of this contemplation, then, is
conceived as lying not in the first thing but in the
second. Any one having the use of his senses could see all
the colours and shapes that a picture contains . . . but
((to enjoy an esthetic experience)) he must use his
imagination, and so proceed from the first part of the
370
experience, which is given in sensation, to the second

part, which is imaginatively reconstructed. This seems to
be the position of the "realistic" philosophers who
maintain that what they call "beauty" is subjective.
148-49).
Such distinctions as used by realistic estheticians like Samual
Alexander do not do justice to what Collingwood thinks the true
situation to be:
The distinction between what we find and what we bring is

altogether too naive . . . . If ((an artist)) paints his
picture in such a way that we, when we look at it using our
imagination, find ourselves enjoying an imaginary ex-
perience of total activity like that which he enjoyed when
painting it, there is not much sense in saying that we
bring this experience with us to the picture and do not
find it there . . . . No doubt there is a sense in which we
bring it with us. Our finding of it is not something that
merely happens to us, it is something we do . . . . The
imaginary experience . . . is the kind of experience we are
capable of having. Thus the two parts of the experience are
not contrasted in the way in which we fancied them to be.
There is no justification for saying that the sensuous part
of it is something we find and the imaginary part something
we bring . . . . We bring our powers of vision with us, and
find what they reveal. Similarly, we bring our imaginative
powers with us, and find what they reveal: namely an
imaginary experience of total activity which we find in the
picture because the painter has put it there. (PA, 149-51).
Now if this is true for a work of art, it must be true in some
sense for perception in general, and also for scientific ob-
servation, inasmuch as the same sensation and imagination are at
work--even if one does not assume that the experience of nature
is like an imaginary experience of total activity which is "put"
there by an artist. What Collingwood is saying is that sensation
itself is an active process of an agent, and this

371
is what he means when he says that it "makes" its sensa. He does
not mean that the objects extrinsic to our sensory organs are
"created" ex nihilo by our minds. But he is also saying that our
powers of vision find what they reveal--and the revelation
involved is the process whereby somehow (and this is not ever
specified by Collingwood) information is passed from objects
within our field of experience as a presentation of
uninterpreted sensum. It is the difference between "seeing" and
"looking" or between "hearing" and" listening" that Collingwood
is referring to when he adds the act of attention to this field
of felt experience, and not the difference between a nothing and
a something. Or to put it another way, the difference that
knowing (as perception) makes is not the difference between
there being something where there was formerly nothing at all;
it is rather the difference between something present but
indefinite (an uninterpreted sensum) and something present and
now more definite (a sensum ready for interpretation--present
and sustained). But Collingwood is not denying the presence of
something--on this The Principles of Art is clear:
Theoretically, the artist is a person who comes to know

himself, to know his own emotion . . . . But this knowing
of himself is a making of himself . . . . Moreover, his
knowing of this . . . world is also the making of the new
world which he is coming to know. The world he has come to
know is a world consisting of language; a world where
everything has the property of expressing
372
emotion. In so far as this world is thus expressive or

significant, it is he that has made it so. He has not, of
course, made it "out of nothing." He is not God, but a
finite mind still at a very elementary stage in the
development of its powers. He has made it "out of” what is
presented-to him in the stil1 more elementary stage of
purely psychical experience: colors, sounds, and so forth.
(PA, 291-92; emphasis mine).
B.--As if in confirmation of this conclusion that some-
thing must be presented to consciousness that is not created ex
nihilo by that consciousness, Collingwood at one point even
adopts the metaphorical terminology of matter and form to de-
scribe the levels of mental functions:
(I)n the relation between any one level of experience and

the next above it . . . the higher level differs from the
lower in having a new principle of organization; this does
not supersede the old, it is superimposed on it. The lower
type of experience is perpetuated in the higher type in a
way in which a pre-existing matter is perpetuated when a
new form is imposed on it . . . . In this metaphorical
sense of the words, any new and higher level of experience
can be described in either of two ways. Formally, it is
something quite new and unique, capable of being described
only in terms of itself. Materially, it is only a peculiar
combination of elements already existing at the lower
level, and susceptible of description in terms of these
lower elements. Consciousness . . . is formally
unique . . . . Materially, it is only a certain new
arrangement of psychical experiences. (PA, 233).
Would Collingwood accept this metaphor as applicable at
the level of sensation? Is it not possible that there is a
similar relationship existing at the interface between sensation
and the world of objects extrinsic to the organs of sensation,
such that sensation can be regarded as formally "something quite
new and unique, capable of being described

373
only in terms of itself," but materially it is "only a certain
new arrangement" of material entities--cells composed of organic
and inorganic molecules, functioning together as an organ in
response to physical stimuli from the environment? (Cf. NL,
9.5-9.56).
From the point of view restricted by the self-imposed
limitations of a science of mind, we must answer (as Collingwood
does in The New Leviathan) that we do not know (see NL,
5.2-5.4). But before involving ourselves in a discussion of the
presuppositions of a phenomenology of consciousness, we must
enter further pieces of evidence from The Principles of Art in
which Collingwood lets himself escape for a moment from the
confines of these presuppositions.
C.--Where Collingwood is arguing that the distinction
between real and imaginary sense is not a distinction between
activity and passivity, nor among passivities, nor among ac-
tivities, in the passages deleted from our above quotation
Collingwood argues as follows:
Response to stimulus is in some sense passive, in so far as

it cannot arise without a stimulus; but it is also active,
in so far as it is a response. If I am a kind of factory
for converting wave-lengths into colours, air disturbances
into sounds, and so forth, as the materialists
believe . . . there is work done in that conversion; the
machinery is active, even if it is controlled by no manager
or foreman . . . . For sensation, as well as imagination,
is on its bodily side a change arising within our own
organism, and due to the energies of that organism itself.
The afferent nerves through whose activity we feel a
pressure on a finger-tip are not
374
solid rods conveying that pressure itself to the brain;

they are functioning in their own way as a special kind of
living tissue; if they ceased to function in that way, no
amount of pressure on the finger could give rise to
sensation. (PA, 196-97).
Now if Collingwood meant to come up with a radical idealist
explanation for sensation, and one which hence eliminated the
need for asserting that objects make a difference to the knowing
of them, it is difficult to see why he would be considering such
physiological mechanisms at all. No doubt, he is arguing that
thought (as attention) dominates the "feeling" provided within
the psychic level of experience, and that even at the psychical
level the organs of senses are not utterly passive. But he is
not denying that the physiological mechanisms provide something
for thought to dominate, or that the organs of sensation are
"stimulated" by forms of physical energy.
D.--Several pages later when Collingwood is discussing
the way in which feeling modified by attention (and notice that
the modification is only stated to be in terms of time and
use--attention perpetuates and dominates feeling-and not in
terms of content: it does not modify a red into a blue, or blue
into loud, or sweet into hard; cf. PA, 189) he talks about the
psychological phenomenon of color fading, which is compensated
by the act of attention: 28
_________________
28 In The New Leviathan Collingwood uses this identical
example, but gives it precisely the opposite interpretation:
"colours themselves as
375
In the flux of sensation, one pattern of the total sensory

field is being replaced by another. Attention now focuses
itself on one
element in that field: for example this scarlet patch. As I
look, the red is actually fading; it is being obscured by
the superimposition of its own after-image, which dulls the
scarlet moment by moment. But by attending to the scarlet
and neglecting everything else I create a kind of
compensation for this fading ((By this)) progressive
refocusing of attention . . . we do not lift any sensa, as
such, out of their native flux; but we obtain a new kind of
experience by moving as it were with the flux, so that the
self, and the object are (so to speak) at rest relatively
to each other for an appreciable time. What we have done is
. . . liberated ourselves for a moment from the flux of
sensation and kept something before us long enough to get a
fair sight of it. (PA, 210).
Collingwood's hesitation to accept fully these metaphors ("as it
were. . . so to speak") should not blind the reader to the
affirmation of the point at issue. It is a temporal modification
of feeling that is being made, and by engaging in acts of
attention "something" is kept before us, and for long enough
time a "fair sight of it" may be gotten--not by lifting off any
sense (attention is not abstraction) but by running with it (so
to speak) so that it appears stable compared to its moving
background.
_______________________
we actually see them are vague; and so with sounds, smells,
emotions, etc. We never see anything exactly any colour. However
carefully we look at a colour it remains ambiguous. Indeed,
looking at it carefully creates a new ambiguity; for the eye
becomes fatigued and a complimentary after-image interposes
itself between the eye and the colour at which one is looking,
so that the mere looking at a colour dims it" (NL, 5.71). If
"looking carefully" means "attending to" the colour, it seems
that this process adds to the fading of it rather than
compensating for it.
376
E.--When Collingwood raises the question about how long
a period of time a sensum has to be perpetuated by attention in
order to be retained, he replies that (from the perspective of
The Principles of Art) no definite answer can be given (PA,
210). But The Idea of Nature provides us with the start of an
answer, and at the same time suggests a reason for the use of
terms which express qualified confidence in the physiological
descriptions in the previously cited passages. In discussing the
contemporary view of nature, Collingwood argues that when the
impact of evolutionary science began to be thought out it had
already, in the 19th century methodology of history, a model for
dealing with a world of constant change:
History had by now established itself as a science, that

is, a progressive inquiry in which conclusions are solidly
and demonstratively established. It had thus been proved by
experiment that scientific knowledge was possible con-
cerning objects that were constantly changing . . . . The
historical conception of scientifically knowable change or
process was applied, under the name of evolution, to the
natural world (IH, 13).
One of the consequences of the evolutionary view of nature is
that, as Whitehead put it, "there is no nature at an instant,"
and therefore the study of natural forms of motion requires a
notion of "minimum space" and "minimum time." In the latter case
this issues in the principle that "different orders of
substances take different orders of time-lapse to exist" (IN,
22; cf. EM, 266-67). But this principle does not warrant
adoption of an idealistic stance.

377
The principle . . . opens no door to subjective idealism.

One might express it by saying that how the world of nature
appears to us depends on how long we take to observe it . .
. . This, though true, would be misleading . . . . How the
natural world appears to us does certainly depend on how
long we take to observe it; but that is because when we
observe it for a certain length of time we observe the
processes which require that length of time in order to
occur. (IN, 23). 29
But just as we noticed Chapter VI that historical knowledge is
limited in one direction by the "facts" at the historian's dis-
posal, so observation in science is limited in physical situ-
ations in an exactly analogous fashion:
Our experimental knowledge of the natural world is based on

our acquaintance with those natural processes which we can
observe experimentally. This acquaintance is limited
downwards in space and time by our inability to observe any
process that occupies less than a certain amount of space
or a certain lapse of time, and upwards by the impos-
sibility of observing any process that occupies more space
or more time than the range of human vision or the time
covered by human records . . . These limits, upper and
lower, of our observations in space and time have been
greatly enlarged by the apparatus of the modern scientist,
but they still exist, and are ultimately imposed on us by
our constitution as animals of a definite size and living
at a definite rate . . . . The natural world which human
scientists can study by observation and experiment is an
anthropocentric world; it consists only of those natural
processes whose time-phase and space-range are within the
limits of our observation. (IN, 24).
___________________________
29 Cf. IN, 84: "Nature stays put, and is the same
whether we understand it or not;" and IH, 133: "In
science, . . . the facts are empirical facts, perceived as they
occur. In history . . . fact . . . is not immediately given. It
is arrived at inferentially by a process of interpreting data."
378
In these passages Collingwood is not writing as if the
"natural world" external to our sense organs can have no effect
on our knowing of it. In fact it is a presupposition of natural
science that there is such a world and that it can make itself
known to us (cf. IN, 175; EM, 222-23). The statement that the
natural world observed by science is an anthropocentric world
only echoes the Kantian restriction of scientific knowledge to
the bounds of experience possible for a human being of limited
capacities; it does not deny that there is something there to be
known.
F.--Confirmation that we are on the right track comes,
indirectly, from the Essay on Metaphysics. In a passage in which
Collingwood is discussing the extent to which Aristotle's
metaphysics is compatible with contemporary scientific
presuppositions, Collingwood singles out as a rejected presup-
position the belief that the existence of nature is an observed
fact:
Aristotle thought . . . that by merely using our senses we

learn that a natural world exists. He did not realize that
the use of our senses can never inform us that what we
perceive by using them is a world of things that happen of
themselves and are not subject to control by our own art or
any one else's ((but is rather)) the first and fundamental
presupposition, on which alone any science of nature can
arise . . . . For when we speak of the existence of natural
things we mean (as Aristotle very truly says) the existence
of things that move of themselves or events that happen of
themselves. The idea of movement or happening . . . is
contained in the idea of a natural world. The idea of
motion, therefore (for if the world of nature is a world of
bodies all the events in nature are motions), cannot be an
idea which we obtain, as the Greeks thought we obtained it,
through the use of our senses. It is an idea which we bring
with us in
379
the shape of an absolute presupposition to the work of

interpreting what we get by using our senses. The
proposition that there is motion in nature is a
metaphysical proposition. (EM, 215, 217).
Collingwood is not saying that we do not get something by the
use of our senses, but only that through sensation we do not get
the concept of motion or of an existent world of nature; what we
get are uninterpreted sensa. And he is not saying that a world
of nature and of natural objects does not exist, but only that
assertion of its existence is not justifiable as a simple
product of sensual experience, but rather through the correct
interpretation of sensa. And finally, Collingwood is not ruling
out the possibility that the "data" presented to us in sensation
can be given and retained, but that "real, empirical knowledge"
of the natural world (or of any world of fact, including the
historical world of evidence) absolutely presupposes the
existence of a world of nature.
G.--These passages also shed light on our problem in-
sofar as they provide us with a clue to understanding why, in
The Principles of Art, Collingwood hedges in his assent to the
legitimacy of the physical basis for psychical sense. The reason
is not that he refuses to take sides on the realism-idealism
debate, but rather because he is bound by his own
presuppositions concerning the proper limits to the methods of
mental science. In The New Leviathan this rule is stated in
negative fashion
380
as the Fallacy of Swapping Horses:
I have mentioned two approaches to the problem of self--

knowledge: the natural sciences and the sciences of man . .
. . Each is valid. Each is a search for truth, and neither
goes unrewarded. Each, therefore, has its own problems and
must solve them by its own methods . . . . Of these two
different forms of science, the one that has started a hare
must catch it. The reason is plain. You can only solve a
problem which you recognize to be a problem. The same
methods, therefore, which led to the asking of a question
must lead to the answering of it . . . . No amount of
admiration for some other horse must betray you into the
Fallacy of Swapping Horses. If the wretched horse called
Mental Science has stuck you in midstream you can flog him,
or you can coax him, or you can get out and lead him; or
you can drown, as better men than you have drowned before.
But you must not swap him even for the infinitely superior
horse called Natural Science. (NL, 2.6-2.73).
As we shall see in section six of this chapter, Collingwood's
efforts to stay strictly within the bounds of this rule in The
New Leviathan lead him into difficulties concerning his char-
acterization of feeling. In The Principles of Art, however, he
occasionally strays across the border, as in the above passages
where he appeals to physiological descriptions. But in one very
important passage he gives us a glimpse of how natural science
(in the form of behaviorism) and mental science are related. The
passage is cryptic, shimmering with interpretative possibilities
and ambiguities, and therefore we must quote it at length:
At the merely psychical level, the distinction between

conscious and unconscious does not exist . . . . The mind
here exists only in the shape of sentience . . . . When the
light of consciousness falls on such occupations, they
change their character; what
381
was sentience becomes imagination. Hence we cannot study

psychical experience, or even assure ourselves that it
exists, by inquiring of our own consciousness; that can
only tell us . . . of the things to which it
attends . . . . Those which are utterly outside its ken
must be studied by other methods. But what are these
methods to be? Behaviorism has dealt with the problem, and
gone some way towards a correct solution, by dismissing
"introspection," that is, inquiry made of consciousness, as
futile, and identifying the psychical with the
physiological. The method thus devised is perfectly sound,
but for one flaw. Unless we had independent knowledge both
that there is such a thing as psychical experience and what
kind of thing it is, the problem which the behaviorist
solves by his method could never arise. This independent
knowledge is derived neither from observing bodily
"behavior" nor from questioning consciousness, but from
analyzing consciousness, and thus discovering its relation
to a more elementary kind of experience which it
presupposes. The principle of this analysis depends on the
fact that attention (or as we may now indifferently call
it, consciousness or awareness) has a double object where
sentience has a single . . . . ((A)) person who is said to
be looking is described as aware of his own seeing as well
as of the thing he sees. (PA, 205-06).
Let us at the risk of some repetition try to be clear about what
this important passage is asserting. (1) Psychical experience
(the level of pure sensation) cannot be studied or shown to
exist by inquiring of (putting questions to) consciousness, (2)
the reason being that consciousness alters sensation.
(3)Therefore "introspection" (not in the sense used in his
discussion of empiricism, but as the putting of questions to
consciousness) is ruled out as a method for studying psychical
experience (but not as a method for studying
consciousness)--some other method is required. (4) Behaviorism
(a natural science) has devised a method perfectly sound for
this purpose--a method which (a) rejects "introspection," and
(b) identifies the psychical

382
with the physiological (and therefore subject to the
investigative methods of the physical sciences). (5) The only
flaw with behaviorism as a method is that it fails to recognize
that unless it presupposes (a) that psychical experience exists,
and (b) it is a certain kind of thing (namely a physiological
phenomenon but present to conscious beings), it would have
nothing to investigate. (6) This knowledge is established
independently of both "introspection" and behavioral study; it
is established by the analysis of consciousness. (7) The
principle of this analysis is that consciousness has a double
object, whereas sentience has a single.
But what then is this "analysis of consciousness"?
Judging by the "principle" of its analysis, it is nothing other
than phenomenology: the principle cited is merely a restatement
of the principle of intentionality. But then phenomenological
analysis must not only be able to tell us that psychical
experience exists (because sentience has an object?) but also
what kind of a thing it is (something present to us as a
concrete sensum, to be attended to).30 This passage is
____________________
30 This does not solve all the interpretative difficul-
ties in the passage under consideration. It is not clear, for
example, how the analysis of consciousness can discover the re-
lationship of consciousness to a more elementary kind of ex-
perience (the psychical level) without altering it in the pro-
cess, because analysis must itself be a kind of consciousness.
383
therefore isomorphic with all those other remarks about psy-
chology scattered through his writings, from his rejection of
the psychologistic reduction of religion in Religion and
Philosophy to his assault on psychologism as a metaphysical
pseudo-science in the Essay on Metaphysics (cf. EM, 112-42). In
them he is arguing not that psychology (in this case,
behaviorism) does not have a legitimate field of investigation,
but that it presupposes philosophical psychology (in this case,
phenomenological analysis).31
But once the "flaw" in behaviorism is made good by ac-
ceptance of the presuppositions established by phenomenological
analysis, its methods are sound for examining the psychical
level of a natural phenomenon, a level at which "introspection"
is futile. And in The Idea of Nature Collingwood is not bound by
the limitations of the inquiry into consciousness by the methods
of introspection. The Principles of Art occupies an intermediary
position--but Collingwood's sense of the limitations presupposed
in a given piece of thinking prevents him from giving a complete
account of perception, and from giving full assent to the
____________________
31 By calling Collingwood's "analysis of consciousness"
phenomenological we do not mean to imply a conscious identifi-
cation of Collingwood with the continental phenomenological
movement begun by Husserl. If there is any such connection
between Collingwood and the phenomenologists, it is nowhere
explicit in his writings. Furthermore Collingwood would not
accept phenomenological analysis as purely descriptive; he in-
sisted always that it was "normative" or "criteriological." Cf.
EM, 109; PA, 171, note.
384
physiological aspects of perception--the level at which objects
external to our sense organs initiate changes in those organs.
H.--We find an actual instance of an "analysis of con-
sciousness" in one of the epilegomena to The Idea of History
--"History as the Re-Enactment of Past Experience." The analysis
is dialectical: it takes the form of a number of objections
raised by hypothetical realists and idealists to the assertion
that historical knowledge is possible only on the condition that
the historian can re-enact in his own mind past acts of thought.
We shall take up this argument in Chapter IX, so we shall not at
this point enter into the details of this dialectical analysis.
It suffices to say that in this epilegomenon we have an actual
case in which Collingwood argues against both the realist's view
that acts of thought and their objects are independent of one
another (and therefore that although two acts of thought may
have the same object, they are not the same thoughts because
thoughts are bound to subjects and the subjects are different
persons) and the idealist's view that objects are ultimately
dependent on thought (and therefore that in being thought an
object becomes subjective).32 On either view, he argues, history
is not possible (IH, 289-90).
___________________
32 Once again it is worth noting that part of this com-
plex argument is aimed at particular idealistic views of history
(Croce and Oakeshott), so the use of the term "idealism" here
should not be
385
Collingwood's answer to these objectors is ultimately that
thought is neither purely subjective nor purely objective, but
always both (IH, 292). But in answering the idealist objector
Collingwood goes one step further--a step that requires stating
what we have just seen him call the principle of the analysis of
consciousness:
Why did ((the idealist objector)) think that the act of

thought, by becoming subjective, ceased to be
objective? . . . It is because he understood by
subjectivity not the act of thinking, but simply
consciousness as a flow of immediate states. Subjectivity
for him means not the subjectivity of thought but only
the subjectivity of feeling or immediate experience. Even
immediate experience has an object, for in every feeling
there is something felt and in every sensation there is
something sensed: but in seeing a colour what we see is
the colour, not our act of seeing the colour . . . . The
subjectivity of immediate experience is thus a pure or
mere subjectivity; it is never objective to itself: the
experiencing never experiences itself as experiencing.
If, then, there were an experience from which all thought
were excluded . . . the active or subjective element in
that experience could never be an object to itself, and
if all experience were of the same kind it could never be
an object at all. (IH, 294-95).
In this essay the term "experience" refers to the immediacy of
something to consciousness, a term consistent with his usage
throughout his writings. It also provides him with a means for
distinguishing three senses of the term "awareness" which we
know to be a synonym for
___________________________
understood to be applicable to all forms of it--in particular
objective (Hegelian) idealism. It is possible that an objective
idealist would not feel the argument here cited as damaging. We
shall have more to say on this interesting variant of the
autobiographical anti-realist argument at IH, 288, in Chapter IX.
386
"attention" and for "consciousness." It means (1) the immediate
experience of feeling, (2)self-consciousness, and (3)
perception. Collingwood says the term should be confined to its
second sense: "I am aware of my act not only as an experience
but as my experience, and an experience of a determinate kind"
(IH, 291).
But then if awareness is self-consciousness how can
thought be anything but subjective? How can thought be an object
to itself? In answering this Collingwood expands the term
"awareness" even further:
((The objector)) will perhaps say that one act of thought

may be an object to another act, but not to itself. But
this . . . needs modification, for any object is properly
the object not of an act but of an agent, the mind that
performs the act. True, a mind is nothing except its own
activities; but it is all these activities together, not
any one separately. The question is, then whether a per-
son who performs an act of knowing can also know that he
is performing or has performed that act. Admittedly he
can, or no one would know that there were such acts, and
so no one could have called them subjective; but to call
them merely subjective, and not objective too, is to deny
that admission while yet continuing to assume its truth.
The act of thinking . . . has to be studied as it
actually exists, that is to say, as an act . . . . This
study . . . is self-knowledge . . ; it is the object of a
self-knowledge which differs from mere consciousness in
being selfconsciousness or awareness, and differs from
being mere self-consciousness in being self-knowledge;
the critical study of one's own thought, not the mere
awareness of that thought as one's own. (IH, 292).33
__________________
33 Cf. Donagan's "Principle of Order," which is one of
the four major presuppositions which Donagan claims hold for
Collingwood's mature philosophy of mind: "If a man is conscious
of one of his own acts of consciousness, then it is not by that
act itself, but by another act of consciousness which may be
said to be of a higher order" (LPC,28). Any
387
So now it appears that the schema is complete; in addition to
(1) the immediate experience of sensation, (2) mere conscious-
ness, and (3) self-consciousness, we have also (4) critical
self-consciousness or self-knowledge, the critical study of
one's own thought. For an analysis of consciousness to be com-
plete it must be carried through to this final stage, which we
have often seen Collingwood call philosophy. In a later chapter
we shall see that the critical study of one's own thought is the
study of the presuppositions of that thought, or metaphysics,
and that the methodology of metaphysics necessitates the
methodology of history. But for now we wish to point out that in
an actual specimen of the analysis of consciousness Collingwood
has in fact appealed to the principle of intentionality in the
sense that all consciousness (even consciousness of sensation or
perception) implies the presence of an object.
Now if we pull these various strands of circumstantial
evidence together into a final summary concerning the relation
of knower to known, what emerges from the evidence of The
Principles of Art, when taken in conjunction with the evidence
from remarks in The Idea of Nature and The Idea of History, is
this:
___________________
interpretation which accepts an overly-strict adherence to this
principle must ultimately come to grips with the passage just
cited, which states that this principle "needs modification" and
defends the thesis that at least one act of consciousness can be
self-illuminating, namely self-knowledge.
388
TABLE 7
THE KNOWER-KNOWN RELATIONSHIP IN PERCEPTION
1. The knower-known relationship as it occurs in the act of

perception is not adequately described by simple assignment
of activity or passivity to either knowing or its object.
(PA, 196-97)
a. Knowing does not create its object ex nihilo (i.e.,

knowing is not totally active with respect to an object);
(PA, 291-92)
b. nor does the object create knowledge of itself ex nihilo

in the knower (knowing is not totally passive in the act
of perception). (PA, 196-97)
c. Although knowing is active (in the form of sentience)

even at the level of sensation, for there to be an object
of sensation something must be present to it (and
therefore active in this minimal sense)--i.e. it must
present itself to the knower's field of sentience. (PA,
210)
2. In perception the object affects knowledge by being present

to sensation; its being an object for sensation is the same
as its being present to sensation. Its esse is sentiri. (PA,
189, 8-24; PA, 198; cf. PA, 206)
3. At the level of sensation, knowing affects the object by

perpetuating it and domesticating it by an act of attention.
This act makes the sensum one's own (an item of our own
experience) and maintains its presence in experience for
sufficient time to allow further acts of thought concerning
it to occur (discrimination, comparison, abstraction,
measurement, etc.). (PA, 212-13)
___________________
NOTE: In this table, "object" is to be taken as "object of
sensation"--i.e., whatever is sensed, or present to sensation.
"Knowing" is, unless qualified, to mean "empirical
knowledge"--i.e., knowledge by perception, and has the same
broad extension as "thought"--including first order attention.
389
4. These further acts of thought are interpretations of a

sensum.
5. Limitations on observation come (a) from the side of the

object (how long a process takes to occur, how large or
small it must be in order to exist) as well as (b) from the
side of the subject (the limits of duration of attention,
spatial discriminatory capacities, etc.). These are (a) the
presupposition concerning minimum space and time in nature,
and (b) the limits of possible experience in subjects,
respectively. (IN, 23-24)
6. The assertion of the existence of a world of nature external

to our organs of sensation and capable of presenting
information to us by acts of these organs, is not a matter
of direct perception, but the result of an interpretation of
sense. This assertion is a presupposition of natural
science. (EM, 213, 217)
7. Natural science overlaps the study of mind at the psychical

level, the physiological aspects of which are studied by
behaviorism, using the methods of natural science, and not
using the methods of conscious inquiry or introspection.
(NL, 2.6-2.73; PA, 205-06)
8. That there is such a level of experience, and that it

consists of concrete sense present to sensation as something
to be attended to (i.e. as uninterpreted sense) is
presupposed by behavioral psychology, but is demonstrated by
the analysis of consciousness, on the basis of the principle
of intentionality. (PA, 205-06)
390
6. The New Leviathan: Attention as a Linguistic Act.
Unfortunately we are not yet finished with our
survey of Collingwood's writings on the issue of realism. We
come at last to his final work, The New Leviathan--the last work
published during his lifetime, and his final word on the subject
of the philosophy of mind.34 As with The Principles of Art, we
are looking for evidence concerning Collingwood's evaluation of
the principle doctrine of realism,35 and
____________________
34 The New Leviathan is an odd book in many ways, when
compared to Collingwood's other works. It is written in a highly
aphoristic and Olympian style, with the pseudo-mathematical
device of numbered paragraphs and sub-numbered sentences. Its
rhetoric is more hectoring and bombastic than usual, with racial
slurs directed at whole nations or peoples--especially the
Germans and Turks (cf. NL, 12.4-12.42, 33.47-33.75, 42.142.74,
44.1-44.9, and 45.1-45.96). Some of these oddities may be
written off as due to the highly emotional circumstances under
which the book was written (he writes that he concluded it
during a Nazi bombardment of London), as well as the trials of
his struggle with his rapidly failing health (cf. NL, v).
Specific oddities in terms of doctrine will be dealt with di-
rectly in our discussion. But an interpreter should be cautious
about the weight he puts on specific and unparalleled statements
that he finds in this work--statements that do not appear
supported elsewhere in Collingwood's other writings. In our own
work we have therefore approached it with caution, and treat it
last rather than first; to reverse this procedure (as Donagan
and Mink do) and deal with The New Leviathan as a foundation for
understanding his mature philosophy of mind is not, in our
opinion, sound strategy (cf. Donagan, LPC, 19).
35 The New Leviathan is subtitled "Man, Society, Civi-

lization, and Barbarism," and these correspond to the four parts
into which the book is divided. Although there are scattered
remarks about perception, thinking, and science throughout the
latter three parts, our main concern will be only with the first
part, in which the philosophy of mind is expounded that serves
as a foundation for parts II-IV. That the argument of The New
391
especially as that evaluation affects his analysis of
perception.As it happens there is no direct argument in The New
Leviathan that has the form of the autobiographical refutation
of realism. But there is a discussion of the issue of empiricism
that parallels (but definitely does not duplicate) that given in
The Principles of Art. As in the earlier work, this discussion
arises in the context of the distinction between feeling and
thought (NL, 1.61, 4.13-4.19). Within feeling Collingwood
distinguishes a sensuous element and an emotional charge (NL,
4.1). But unlike the earlier work he calls thought a
"constituent" of mind and feeling an "apanage": feeling belongs
to mind not the way a man belongs to a family or a plank to a
boat (as a constituent or an element) but in the way in which an
estate belongs to a family or a mooring to a boat-
___________________
Leviathan as a whole is fundamentally anti-realistic hinges on
the denial that theory and practice can be separated (or that
theoretical reason can be separated from practical reason)--a
denial made at the very beginning (NL, 1.66) and reinforced
throughout parts I-IV in various forms (cf. NL, 7.22,
14.3-14.31, 18.13, 19.25, 20.21, 27.55, 36.25-36.7,
41.32-41.33). Since The New Leviathan is primarily an ethical
treatise (or perhaps more properly a treatise on social and
political philosophy) the stress on the pernicious consequences
of realistic philosophy (cf. A, 47-48, 147) is quite
understandable. But throughout the work there are other direct
assaults on what Collingwood had recognized to be realistic
doctrines: e.g. the rejection of the reduction of the term
"society" to the abstract notion of class (NL, 19.37, 19.7); a
defense of the notion of a common good and a rejection of the
view that all "goods" are private (NL, 20.1220.22, 21.27-21.65,
36.25-36.55); and the description of the body politic as a
dialectical entity (NL, 24.52-24.75, 29.129.75, 39.1-39.15).
392
i.e. as something it has but not as something it is (NL, 4.13-
4.19). With this distinction in mind, Collingwood addresses the
question of radical empiricism:
There is a saying, nihil est in intellectu nisi quod prius

fuerit in sensu. If this were true, the precision or de-
finiteness which is characteristic of thought would already
be characteristic of feeling. Many people try to persuade
themselves that it is; but they are mistaken. They regard
feeling as a constituent of knowledge; but it is only an
apanage of knowledge: an indispensable apanage, but an
apanage and no more. Are there objects of feeling or not? I
do not know. Nobody knows. Some have said there are, some
have said there are not. As the question is unanswerable on
positive grounds I answer it on methodological
grounds . . . . Entia non sunt multiplicanda (runs Occam's
Razor) praeter necessitatem. Following this rule I answer
the question: "Are there objects of feeling or not" by what
I call a methodological negative. Feeling must on any view
have modes . . . . The question is whether a theory of
feeling needs objects as well as modes. The Lockian theory
does; the Cartesian does not. By Occam's Razor the
Cartesian theory is preferable. (NL, 5.19-5.2, 5.39).
Is Collingwood taking a different tack here than he did
in The Principles of Art? It would certainly seem so, since on
the strength of the above quotation a reader could certainly say
that the "matter-form" analogy accepted in the earlier work is
certainly not applicable to something that is an "apanage" and
not a "constituent." If feeling is "an apanage and no more" (and
we do not ask how it can be an "indispensable apanage" and still
not be a constituent) it cannot stand with respect to higher
levels of experience as something capable of being perpetuated
as a "certain new arrangement of psychical experiences" or as a

393
"peculiar combination of elements" preserved by an act of
consciousness (PA, 233).
But before analyzing his argument to justify this "me-
thodological negative," we must notice that Collingwood is
putting a question to consciousness, finding that it is
unanswerable on "positive" grounds, and therefore answering it
on methodological grounds. We recognize this to be a case of the
"inquiry into consciousness" that Collingwood had said cannot
tell us what sort of thing the psychical level is or even whe-
ther it exists. A non-positive answer is therefore not sur-
prising on the grounds of the positive evidence offered by
consciousness alone. It is also not surprising that methodology
enters into the discussion, since Collingwood had asserted that
what the method of inquiry into consciousness could not
establish could be settled by the methods of behavioral psy-
chology as corrected by the "analysis of consciousness" using
the principle of intentionality. What is surprising is that the
"methodological answer" to the question turns out to be
negative: there are no objects of feeling. The "methodological"
criteria therefore cannot refer to the methods of behavioral
psychology (there is no experimental evidence cited), nor is
there any "analysis of consciousness" here based on the
principle that consciousness has a double object while sensation
has a single. In fact it appears that the criterion being used
does not allow sensation to have even a single object: it has
only modes. The criterion appealed to must therefore be the

principles involved in
394
an "inquiry into consciousness" which (as expected) cannot
answer such questions.
But why then is the question answered in the negative?
Should it not be simply left open (as it is at NL, 5.2, where he
says that he does not know and nobody else does either)? Should
it not merely be stated that feeling is ambiguous with respect
to this category as it is with the other Kantian categories (NL,
5.66)? In order to get a handle on why Collingwood reaches such
a conclusion, let us look at the argument offered between the
question and the answer, deleted from the above-quoted passage.
The passages are long and difficult, and are best dealt with by
paraphrase. The argument proceeds as follows:
(1) Locke asserted that there are proper objects
(colours, sounds, etc.) of acts of feeling (seeing, hearing,
etc.); the general name for these objects is "sense data" (NL,
5.21-5.23). Since data are first-order objects to sensation, and
hence second-order objects to simple consciousness (con-
sciousness in its most primitive form).36 The immediate
_____________________
36 There is a lack of terminological uniformity between
The Principles of Art and The New Leviathan which adds to the
difficulty of comparing the views of mind presented in each – a
confusion complicated by the alternation between the purely
relative use of terms (like "first-order" and "second-order"
objects, where the same object may be a first-order or immediate
object to one level of consciousness and a second-order or
mediated object to a second level reflecting on the first) and
the use of similar terms to refer to concrete levels of
395
or first-order object of simple consciousness is the activity of
sensation itself (NL, 5.24). (2) Berkeley asked, "What is the
status of objects of sense-perception?" and answered, "Their esse
is percipi," meaning that the being of colour is its being seen,
the being of sound its being heard, etc.--the sounds, colours,
etc. being apanages of mind or products of the activities of
seeing, hearing, etc. (NL, 5.27-5.29). (3) G. E. Moore and
others (obviously contemporary empiricists) beginning with the
same assumption (i.e. that sense-data are objects of sensation)
reach the opposite conclusion, i.e. that any object is precisely
what it would be if we were not aware of it (NL, 5.31;
Collingwood quotes from "The Refutation of Idealism"). (4) The
Cartesian answer is that there are modes of feeling (blue, hard,
loud, etc.) but no objects. This does not mean that Descartes
denied that there were such things as blue colors, loud sounds,
etc., but that "Descartes denied the blue colour to be the
object of a transitive verb to see, as a dog may be the object
of a transitive verb to kick. It means that for Descartes the
grammar of the sentence 'I see a blue colour' is not like the
grammar of 'I kick a bad dog' but like the grammar of 'I feel a
transient melancholy' or 'I go a fast walk.' The colour, the
melancholy, the walk, are not objects of an action, they are
modes of an action; their names have an adverbial function in
the sentences in
_____________________
mind (like "first-order" and "second-order" consciousness).
Where there is apparent lack of synonymy we shall indicate the
fact. In this case Collingwood's use of "simple consciousness"
is synonymous with first-order attention from The Principles of
Art.
396
which they occur" (NL, 5.34-5.35). On the Cartesian theory
feeling as sensuous or emotional has modes, but neither
sensations nor emotions have objects: therefore "in neither case
is there anything of which it is other than idiotic to ask
whether its esse is percipi or not" (NL, 5.37). (5) If the
Cartesian theory is accepted, then the question which (a)
Berkeley answered one way ("sensa are mind-dependent") and Moore
and others answered in the opposite way ("sensa are not
mind-dependent") is a nonsense question; it is "a question to
which no possible answer is right because it arises logically
from an assumption that is not made" (NL, 5.36).37 (6) By
Occam's Razor, the infinity of entities required by the theory
that says that feeling has objects as well as modes is ruled out
as unnecessary. All one needs for an account of feeling is
modes. Therefore
_________________
37 The argument here is confusing, but the following

schema seems likely:
Question: Are sensa mind-dependent?

Answer (1): Sensa are mind-dependent (Berkeley).
Answer (2): Sensa are not mind-dependent (Moore).
Presupposition: Mind-dependency is something that sensa can
have or be.
The implication seems to be that this presupposition (if that is

indeed the presupposition that Collingwood has in mind) is not
being made on the Cartesian theory because to be mind-dependent
something must be a constituent of mind or of consciousness, and
feelings are not constituents of mind but apanages (and apanages
are not constituents). Another way of saying this is that
feelings cannot have 2nd-order objects because they are not
conscious acts; therefore whether apanages have 2nd-order
objects (objects as well as modes) is a nonsense question.
397
the Cartesian theory is preferable on methodological grounds.
(NL, 5.39). (7) But either view (i.e. that there are both modes
and objects of feelings or that there are only modes) fits the
facts, and neither is inherently nonsensical (NL, 5.38).
We seem to be on familiar territory--this review par-
allels that given in The Principles of Art on the question of
real and imaginary sense. But there are several important dif-
ferences in the two accounts. (a) Although Collingwood winds up
stating, as we have already noticed, that feeling is "ambiguous"
with respect to the Kantian categories (and hence have the
status of uninterpreted sense) he concludes here with a Cartesian
rather than a Kantian answer to the question, "Are there objects
of feelings?" (b) In opting for this answer Collingwood
introduces an essentially linguistic argument (i.e. that the
terms referring to sense function as adverbs rather than as
nouns) that is absent in the previous account.38 (c) The
Cartesian answer, which he accepts, renders it
_____________________
38 The linguistic argument, while interesting, is both
(a) non-essential to the main argument, and (b) unreliable. (a)
To demonstrate its inessentiality it is only necessary to read
the above summary of the argument without it--it still makes
perfect sense. The use of Occam's Razor does not require the
adverbial function of sense-object terms; it hinges on the
non-necessity of positing two sets of entities, sense-modes and
sense-objects, when all one needs are modes. (b) But this
highlights the unreliability of the argument, since what
Descartes (and Collingwood) seem to be proposing is to
substitute the term "mode" for the term "object." Now it is true
that sense-modes sometimes function adverbially (e.g. "He felt
the train moving, but visually nothing changed"), but then the
distinction refers to the manner in which things are felt--by
touch, sight, taste, etc.--rather than what is felt by these
modes. For if terms like blue, sweet, loud, etc. function
adverbially in a sentence, then it
398
"idiotic" to assert that there is anything of which it is true
to assert that its esse is percipi (which runs counter to his
previous assertion that the ease of a sensum is sentiri). (d)
And yet he states that neither the assertion that sensa are
mind-dependent (Berkeley) nor the assertion that they are not
(Moore) is inherently nonsensical, and that both fit the facts.
This contradicts his earlier argument that Moore's
position (or the position of any empiricist who argues that
there are entities describable as "sense data" which are objects
of the act of sensation) implies that something absurdly
describable as "unsensed
_________________________
should be proper to say that "He tasted sweetly" or "He looked
bluely" or "He heard loudly"--all of which are semantically
peculiar. The more acceptable usage is not adverbial but
adjectival: "He tasted a sweet pastry," or "He gazed at a blue
feather," or "He heard a loud siren," etc. On the
Descartes-Collingwood theory it would seem that each "object" of
feeling is only another mode, so that we would have to posit
separate senses for every color, taste, texture, etc.--each
being a "mode" or modification of the general term, "feeling."
But linguistic legislation cannot settle the problem of what it
would mean if sense-data terms were made proper objects of verbs
of sensation: "I see blue," "I hear loud," etc.--which appear to
be truncated or incomplete sentences. Completion of their sense
is made possible by intentional or sense-giving acts of
interpretation. The substitution of the term "mode" for "object"
merely postpones the inevitable question, are such entities
present to us or do we present them to ourselves? But if
Collingwood's linguistic argument is taken not as a piece of
verbal legislation but as an illustration of an epistemological
point he wishes to make, then his denial that there are objects
of feeling may come down to nothing more than the denial of
sense-data as the perception of substances rather than of
processes of nature. It would thus be consistent with The Idea
of Nature, which states that one of the consequences of the
modern view of nature is the resolution of substance into
function (IN, 16-17).
399
sensa” can exist. And finally, (e) there is no mention here, or
in what follows it in The New Leviathan, of anything to do with
imagination or imaginary sense.
Some, but not all, of these discrepancies may be re-
solved by paying careful attention to a subtle shift of levels
between the two works, The New Leviathan being strictly bound by
the negative limitations of "The Fallacy of Swapping Horses,"
and The Principles of Art sitting a little looser in the saddle.
In the later work the inquiry proceeds by putting questions to
consciousness and answering them by taking "soundings"
(consciously reflecting or introspecting) at various levels of
consciousness. Since this is the Cartesian method,39 it is
predictable that the same method will yield similar conclusions:
consciousness does not need to assume that sensation has objects
as well as modes. The methodological negative therefore merely
allows the inquiry to proceed by not inhibiting it with
questions that it cannot (on its presuppositions) answer. Since
one of its main presuppositions is that the only constituents of
mind are acts of consciousness, and that with respect to
______________________
39 Collingwood's starting point in The New Leviathan is
even Cartesian: "Of all the things we know or have been told
about Man, which is the one thing that concerns us at the pres-
ent stage of our inquiry? I answer: The division between body
and Mind" (NL, 1.21). But of course, Collingwood's aim is not to
construct a mathematically secure science, nor to proceed only
by way of clear and distinct ideas. Nevertheless these
concessions to Descartes should alone be enough to render sus-
pect any account of Collingwood's philosophy of mind that begins
with the assertion that it is fundamentally anti-Cartesian
(Donagan, LPC, 25).
400
these essential elements feelings are merely apanages, it is not
a matter for consciousness to decide what the relationship may
be between feelings and what they are directed towards. Whatever
these objects may be, they are not first-order objects for
consciousness. Therefore when Collingwood rejects the assertion
that sensation has objects he may be doing nothing more than
what he does in The Principles of_Art when he criticizes the
notion of "sense data": he may be rejecting the absurd notion of
"unsensed sensa." In other words, the terms "sense-object,"
"sense datum" and "unsensed sensum" may be synonyms: all may
refer to a sensum as a datum--something really present only as
re-presented, an imaginatum.
But then the reader is stumped for an explanation as to
why Moore's position is not inherently nonsensical, and why the
notion of "sense-object" is not resolved into an imaginatum, as
it is in The Principles of Art; in The New Leviathan the entire
issue of imaginative representation is never mentioned. Can we
assume that this is because he had said all he cared to on the
subject in The Principles of Art (cf. Mink, MHD, 81)? That might
be the case were it not for some remarkable direct discrepancies
between the two works on the issue of feeling, discrepancies not
easily dismissed on the above grounds.

401
These discrepancies become more obvious the instant we
begin to expand the discussion of feeling. In The New Leviathan
Collingwood characterizes feeling and its relation to thought as
follows. (1) It is an apanage of mind rather than a constituent
(NL, 4.19-4.2). (2) It consists of sensuous elements and
emotional charges which interpenetrate all over the field of
feeling (NL, 4.1, 4.4). (3) Feelings are evanescent--they begin
to perish as soon as they begin to exist (NL, 5.5). (4) Feelings
are indefinite (NL, 4.8) but strong (NL, 4.86). (5) The strength
that feeling has is (a) vividness ("compression
strength"--basically, intensity) and (b) tenacity (the quality a
feeling has which makes it "linger in the mind, be slow to
vanish, and be easily revived when the occasion permits") (NL,
5.14-5.17). (6) Within the here-and-now field of feeling (NL,
4.4) there are place-differences, time-differences ("it has
spatial and temporal bulk"), and intensity differences (louder
and softer sounds, brighter and dimmer colors, etc.) (NL, 4.4,
4.43). (7) Within this field there is also a focal region where
precision and intensity are greatest, and a penumbral region
where they decrease in every direction until in some outer zone
precision and intensity recede into dimness and confusion (NL,
4.44).
Thus far the description of feeling is not significantly
different from that given in The Principles of Art. But
Collingwood adds several more statements about feeling in
relation to thought, statements

402
that lead us into unfamiliar territory. (8) Feeling has no
edges--its spatial and temporal penumbra fade away (NL, 4.45).
(9) Feeling is the proper object of simple consciousness and is
immediately given to that consciousness (NL, 4.19, 4.24, 4.71).
Without this simple consciousness and the field of feeling
present to it there could be no higher-order acts of selective
attention (NL, 4.5). (10) Distinctions within the field of
feeling are made by an act of selective attention--a practical
act of second-order consciousness in which one's attention is
turned one way or another, "creating . . . a situation in
which . . . consciousness is concentrated on one object (one
feeling) or another" (NL, 4.5). (11) The act of attention really
makes the edges of anything distinguished within the field of
feeling (NL, 4.52 -4.6)--distinctions such as positional
distinctions (distinctions of place and time), qualitative
distinctions, distinctions between sensations and their
emotional charges, between different sensings (seeing, hearing,
smelling), etc. (NL, 4.61-4.63). (12) Any characteristics that
feelings may have are discoverable by simply reflecting on that
consciousness; to attempt to argue about it is to commit the
Fallacy of Misplaced Argument (the fallacy of arguing about any
object immediately given to consciousness) (NL, 4.71-4.73). (13)
There is no generalizing about feelings ("that is, no framing
universal propositions about them and assuring oneself that
these are
403
true, or (alternatively) omitting to do so because one is too
lazy"), although it is possible to think inductively about
feeling ("to think inductively is to assume, because this x has
(or some x's have) a certain characteristic, that other x's have
or would have the same characteristic") (NL, 5.55-5.56). (14)
Feelings have modes but not objects (NL, 5.2-5.39). (15) Feeling
is ambiguous with respect to activity (something I do) or
passivity (something I undergo) (NL, 5.45.49). (16) Feeling is
ambiguous with regard to the Kantian categories of quantity,
quality, relation, and modality (NL, 5.6-5.66). (17) It is also
ambiguous with regard to the Kantian category of quality,
although this ambiguity has limits (NL, 5.67-5.72). (18)
Feelings cannot be remembered, although propositions about
feelings can be remembered (NL, 5.54). (19) Feelings may be
preconscious until they are reflected upon by an act of
consciousness (NL, 5.9-5.91), or unconscious insofar as they are
repressed (NL, 5.86-5.87). (20) Feeling remains preconscious
until it is named (NL, 6.28). Naming is a linguistic act of
expression, which includes everything from the language of
gesture (e.g. an expressive shiver to the cold) to the language
of speech (saying "it is cold") (NL, 6.1, 6.25). "Language in
its simplest form is the language of consciousness in its
simplest form; the mere 'register' of feelings, as wild and mad
as those feelings themselves; irrational, unorganized,
unplanned, unconscious" (NL, 6.58). (21) A language is an
abstraction from discourse, which is the activity

404
by which a man means anything (NL, 6.11). To discourse is to
mean something by the gesture (expression) you make; a language
is a system of gestures (sounds or the like) as having meanings.
Discourse is the activity of meaning something (a) by something
else (b), where meaning (a) is an act of theoretical
consciousness, and (b) is a practical activity, the production
in oneself or others of sounds, etc. which serve as the vehicle
of that meaning (NL, 6.18-6.19). (22) Finally, selective
attention is an act of practical thinking, at the level of
conceptual thought (NL, 7.2-7.21); it is a doing of something to
oneself (focusing consciousness on part of the field of feeling)
and also a doing of something to the object (circumscribing it,
drawing a line between it and the rest of the field) (NL, 7.3).
Leaving the higher functions of consciousness for later
discussion (we will have more to say about language, logic, and
conceptual and propositional thinking in Chapter VIII), we must
now ask how all this stands with respect to the fundamental
tenet of realism. Once again we meet with Collingwood's
intransigent insistence on the activity of thinking and its role
even at the primary level of perception. And once again we find
Collingwood anxious to reject the view of mind that makes it a
passive partner in the knower-known relationship. In his final
published work, then, it is clear that Collingwood has retained
his opposition to the principal doctrine of realism that the
known is unaffected by
405
the knowing of it. To this extent the autobiographical
interpretation is vindicated.
But what about our further question about whether the
object has any effect on the knower? And is this description of
the functions of consciousness consistent with that given in The
Principles of Art, and how does it compare with the
circumstantial evidence that we educed from that earlier work
and from his other late writings? Although we would like to
defend Collingwood's later philosophy of mind as a brave attempt
to reconcile idealism and realism, we find instead that in The
New Leviathan he appears to be so intent on overcoming the
errors of realism that he leaves little room for interpreting
the knower-known relationship in anything but idealistic terms:
not only does knowing make a difference to the object known even
to the level of perception, but there seems to be no way that
the object can make a difference to the knowing of it.
The crux of the matter is, of course, sensation itself.
If there is no way that sensation can present to us a world of
sensa that is something more than in a state of constant flux,
utterly diversified, and without any "edges" at all, then it is
impossible to see (1) why an exercise of the "circumsribing"
function of attention is not utterly free to make any sort of
arbitrary pattern of fancies out of the chaos of sensa on which
it operates; (2) why there should be any sort of correlation

406
between various modes of sensation with respect to one and the
same object of sensation; and (3) how it is possible to
disambiguate feeling at all by any act of interpretation. We
will take up each of these points separately.
(1) We have seen that Collingwood described the field of
feeling as containing positional and qualitative distinctions,
place-differences and time-differences, intensity differences,
and even focal and penumbral regions. But then we saw him go
even further and assert that there are no edges within the field
of feeling, and that all distinctions within it are made by acts
of selective attention--an act of practical consciousness. He
goes so far as to say that not even so much as a "red patch" is
immediately given to consciousness, because "the red is actually
given in feeling to consciousness as a quality transfusing all
the rest of the same field" (NL, 5.65; emphasis mine). The sensa
given in feeling "interpenetrate" all over the field of feeling
(NL, 5.62), and only the activity of consciousness as selective
attention can "cut up" such a field into distinct feelings
("sensations distinct from emotions, visual sensations distinct
from auditory sensations, red patches distinct from green
patches, and so on ad infinitum") (NL, 5.63). The field of
sensation is utterly ambiguous with regard to all the Kantian
categories, and only one small postscript indicates that this
ambiguity has any limits: with

407
respect to the category of quality Collingwood says that "colour
may be indeterminate, but it falls between points on a
colour-scale. We can always fix these limits as closely as we
need . . . but the ambiguity is only restricted, it can never be
removed" (NL, 5.72).
But how is it possible to "restrict" an ambiguity that
(a) includes the impossibility of distinguishing even so much as
a red patch on a green field, and that (2) excludes sensa from
having any "edges" at all? And how is it possible for there to
be differences (qualitative, quantitative, etc.) within a field
if everything within that field "interpenetrates" and
"transfuses" and "begins to perish as soon as it begins to
exist"? And how can there be such differences present in the
field if it is only consciousness that puts such distinctions
there?
Lacking answers to these questions we must conclude that
something has gone wrong with Collingwood's description of
feeling, and especially of sensation. It is not consistent with
itself within The New Leviathan, and it certainly is not
consistent with his remarks in other works, about scientific
observation and its presuppositions. In order that there be
something "there" for perception to attend to, we must assume
that edges are not only "made" by acts of selective attention,
they are also "found" to be there when something is attended
to--once, twice, or however many times one chooses. And although
within a given field of sensation--vision,

408
for example--there may be ambiguities (colors at the periphery
of my vision begin to interpenetrate, or even within the direct
foreground of my visual field if I let my focus blur), it is
nonetheless equally an item of experience that for a given
position within that field something is not always red, green,
blue, etc. all over at the same time, and to the same
degree--this much at least he admitted in The Principles of Art,
when he argued that "we cannot, in seeing a red patch, mistake
it for a blue one" (PA, 189). It is sometimes predominately one
color or another, or it is mottled, or whatever, but certainly
not always totally indefinite. And if it is true that I become
aware of edges by letting my focus wander back and forth from a
red patch to its surrounding green field and back again, and
thereby re-establishing the distinctness of the red patch from
its background, it is not a matter of utterly creating this edge
ex nihilo in the visual field, it is a matter of re-creating or
re-presenting it for myself in the act of perception. If it were
not then it would be nonsense to assert that observation of
nature yields any information at all. Selective attention would
just as surely create the patterns of nature as a painter
creates a visual scene on his canvas: it is a reverse "tabula
rasa" with the messages written on the blank sheet of sensation
by the perceiver's consciousness rather than by natural objects.

409
(2) So also with the formation of percepts which the
medieval scholastics would subsume under the faculty of "common
sense" (not the body of assumed rules or customs or attitudes
towards the world and others, but rather the result of the
operation of a faculty related to imagination, and which holds
together sense from various organs of sensation--colors and
shapes with sounds, textures, smells, tastes,etc.--to form a
"common sensible"). Collingwood's sensuous field as he describes
it cannot be distinct even with respect to modes, if he is to be
consistent, since "visual sensations distinct from auditory
sensations . . . and so on ad infinitum" (NL, 5.63) are only
distinguished by acts of selective attention. But then it
becomes utterly arbitrary whether I associate this color brown
to which I am attending with this shape which I am holding,
where the shape happens to be a cup I am holding under the table
and color is that of the table between my eyes and the cup. Or
even worse, what is to stop my act of attention from taking this
loud noise for this sweet taste? Collingwood wishes us to
withhold such distinctions until we reach a higher level of
discrimination, but by then it is too late: no amount of
discriminating is going to inform me of an error of this sort
unless some original act actually grants me information on the
basis of which I can correct my error. On Collingwood's grounds
(at least in The New Leviathan) it is impossible, using the
findings of mental
410
science, to ever have any such original act. Sensation as such
is essentially and irretrievably ambiguous.
(3) But the problems that Collingwood gets himself into
in The New Leviathan are not due solely to his assertion that
the field of sensation is ambiguous with respect to one or
another of the Kantian categories (which is true even for his
position in The Principles of Art insofar as uninterpreted sense
are ambiguous until interpreted by application of one or more
categories) but rather they are also due to the fact that he
goes beyond this point and makes statements about the field of
feeling that are actually interpretations of it--generalizations
about feeling that his own principles do not allow him to make.
Thus although he says that no generalizations can be made about
feeling (which is itself a generalization about feeling) he
proceeds to say that within a visual field all colors
interpenetrate, that they have no edges, that within the field
of feeling there are place-differences and time differences,
that they cannot be remembered, etc. What are these but
generalizations about feeling? And how does a reader begin to
reconcile the statement that there can be no generalization
about feeling with the statement that immediately follows it,
that it is possible to "think inductively" about feeling, where
"to think inductively is to assume that because this x has . . .
a certain characteristic, that other x's have or would have the
same characteristic" (NL, 5.55-5.66)? For what else could
inductive thinking,
411
on this description of it, be but generalization?
Now it is true that in The Principles of Art many of
these same generalizations were made about feeling, but they
were merely provisionally assumed as a way of distinguishing
feeling from thought, and in order to establish the intermediate
role of imagination. Once imagination as attention is understood
as the minimum specification of the genus thought, the realm of
feeling is redefined as the psychical level of experience on
which attention operates, and through which attention "finds and
makes" the patterns it requires of the world extrinsic to our
sense organs. The methodology of mental science has not been
violated, because consciousness as attention is used to define
feeling (feeling is what attention is conscious of; sensa are
its second-order objects).
All this is demolished by the overstatements of The New
Leviathan. Does the later work then represent a change of mind
in The Principles of Art? And if Collingwood is saying that the
experience of distinctions as given to us in sensation is not his
experience is he not "disowning" a feeling, and is this not the
"corruption of consciousness" that he sought to warn us against?
We cannot answer these questions on the basis of the existing
evidence. We can only try to discern what Collingwood was trying
to say--what he intended to mean rather than what he merely
said. Now
412
it is clear that Collingwood wishes to say, in The New
Leviathan, not only (1) that feeling prior to reflection upon it
is lacking in the distinctions proper to thought alone, and
hence has only a potentiality for categorial predication, but
also (2) that it has place-distinctions and time-distinctions
and intensity-distinctions (even in the pre-reflective state),
and (3) that it nonetheless belongs to us as a field of
experience --as an apanage of mind, or something we have, rather
than a constituent, or something we are.
This is basically a Kantian view on the structure and
function of mind, since (2) is another way of stating what Kant
called the "forms of sensibility"--space and time: for anything
to be an object of possible experience, according to Kant, it
must be spatially extended and temporally successive. This even
seems to be consistent with the view expressed in The Principles
of Art, where Collingwood speaks of the function of attention as
perpetuating sense by keeping them before us long enough to get
a fair sight of them; and also consistent with the view of The
Idea of Nature where Collingwood talks about minimum-space and
minimum-time as limiting conditions for scientific, experimental
observation.
But it has the same problems that the Kantian forms of
sensibility have. Sensa as placed and timed are not utterly
indeterminate, but are rather determined to be someplace and
sometime, and therefore bearing interpretation, in the minimal
sense, already. If they are conditions

413
for the possibility of objects of sensation, then they must be
conditions not only for the subjective act of a percipient, but
also conditions true for the object as well--in Collingwood's
terms, for the sensum. The sensum has to be the "kind of thing"
that meets the requirements of sensation; it has to be a spa-
tially-extended and temporally-persistent thing if perception of
it is to be possible.
The question ultimately comes down to this: if dis-
tinctions, "edges," are not found but made by acts of practical
consciousness, how is it possible to attend to what is not
there--to something (the field of interpenetrating sense) that
is not some thing (nothing determinate at all)? There seems to
be nothing to attend to. Collingwood would say that there is
something there, only not something definite, and that once
selective attention has done its job, it can find the
distinctions that it put there. But the retort occurs
immediately: put where? To distinguish even left and right sides
of a sensory field involves a distinction, which in turn
presupposes spatial distinctions with respect to one's own body
within the sensory field. Are these "put" there as well?
In Collingwood's case the issue is further complicated
by the addition of intensity-distinctions and even a minimal
qualitative
414
distinction, as well as a focal and penumbral structure to the
here-and-now field of sensation (which must also be a
"there-and-then" if the focal-penumbral analogy is to be carried
out). All of this adds up to a field of feeling that is not
altogether indeterminate. But supposing that we can ignore the
contradiction of a field that is determinately indeterminate, a
here-and-now that is also a there-and-then, etc. we still wish
to know how what remains of indeterminacy if such a field is
disambiguated. On this Collingwood leaves the reader in no
doubt: disambiguation of the field of feeling is the work and
product of thought which, by a practical activity called
language, functions as selective attention.
Unfortunately this activity is also not free of am-
biguity, because "language" is not merely the practical side of
second-order thought, as some passages in The New Leviathan
would lead us to believe (e.g. NL, 4.33). It necessarily and
essentially involves a theoretical aspect--the element not only
of producing sounds or gestures or expressions, but of meaning
something by them, where the meaning is an act of theoretical
consciousness (NL, 6.18-6.19). But this description of language
as discourse must be taken as the narrower sense of the term,
"language," which does not include psychical expressions
(grimaces, blushes, tears, etc.) because the latter do not
include thought at all, and hence not theoretical consciousness
and not meaning. How then does one import meaning into the
psychical level of sensation and the chaotic world of sensa?

415
The disambiguation we are especially interested in is
that of the much discussed red patch. If the edges of such a
patch (and the color itself as distinct from its background) are
made by an act of selective attention which is linguistic in the
sense just stated (i.e. including the non-vocal act of meaning,
where meaning is an act of theoretical consciousness), then for
a portion of the field of feeling to be not only present as a
"this" but as an interpreted presence, "this red," and an
interpreted presence related to other such items in a field of
feeling, a "this red patch in a green field," it must be related
to theoretical consciousness through meaning. But once again the
inevitable question arises, is this found in the sensory field
or is it put there? On Collingwood's earlier account of the
matter it is still possible to say, "both found and
put"--leaving the details to behaviorism shored up by the
analysis of consciousness. But in The New Leviathan the
disambiguation of feeling by the elliciting of meaning appears
to be impossible, since one is forbidden to make any
generalizations about feeling. There seems to be no way to
bridge the gap between meaning as an act of theoretical con-
sciousness, and utterly ambiguous sense.
7. Conclusion.
The only way out of these dilemmas is to return to the
position of The Principles of Art, where such intemperate
assertions are not made, and where there is still a possibility
for a "science of feeling" to establish

416
what a "science of mind" cannot--i.e. the manner in which
objects extrinsic to our sense organs can act on them to yield
information about the natural world. This does not mean that we
must ignore The New Leviathan altogether: in the summary with
which we shall conclude this chapter we attempt to retain from
it those elements of the philosophy of mind which are consistent
and compatible with the groundwork laid out in The Principles of
Art. In the matter of language, for example, the later work
shows a distinct improvement (as we shall see in the next
chapter).
But how does it finally stand with Collingwood's view of
the knower-known relationship insofar as the principal doctrine
of realism is concerned? Our final conclusion must be somewhat
disappointing. It is not a matter of faulting Collingwood for
failing to be idealist enough to recognize that reality somehow
exists "for us" and not merely "for itself;" or for not being
realist enough to recognize that there is a world of objects
unaffected by our knowing of them (or, as some would have it,
that the natural language we use to express our thoughts is of
necessity a physical-object language) It is a matter of faulting
Collingwood for failing to be Collingwood enough, for failing to
recognize that the "Law of Primitive Survivals"40 does not apply
only to higher-order mental functions, but (by the logic of the
overlap of
___________________
40 "(A)nother principle that I have assumed throughout
((is)) the LAW OF PRIMITIVE SURVIVALS. It runs as follows. When
A is modified into B there survives in any example of B, side
by side with the function B
417
classes) operates even at the level of sensation, or that no
matter how far one goes down the "scale of forms" one never
reaches a zero-point in the scale (EPM, 31 95., 81-82). These
general principles alone should have led him to recognize that
something is given in sensation other than a riotous chaos of
interpenetrating and utterly ambiguous sensa; that sensation,
although the activity of a not utterly passive agent, is
nonetheless the working-up and further organization of elements
continuous with the world of nature and of nature's energies,
and not something utterly created by the agent and imposed as a
form on a passive and indeterminate matter.
Why Collingwood failed to describe adequately sensation
could only be settled by appeal to evidence which we do not yet
have. Perhaps from the body of his unpublished manuscripts a
theory of perception may yet emerge that can pull together the
paradoxes that are left unresolved in his published philosophy.
But we must notice that the failure is due to the espousal not
of realism, but of something more akin to the radical idealism
as we propositionally formulated it at the beginning of this
chapter. We therefore cannot rest content with any
interpretation that
_____________________
which is the modified form of A, an element of A in its
primitive or unmodified state" (NL, 9.5-9.51). We recognize this
to be the law of dialectical relationship present in
Collingwood's writings ever since Religion and Philosophy, but
obviously not always followed.
418
asserts that the dogmatism of Collingwood's later philosophy was
a reversion to the realism of his youth. On the contrary, it was
a reversion to, and ultimately the dogmatic assertion of, his
youthful idealism. The application of the "Fallacy of Misplaced
Argument" to rule out controversy concerning sensation is as
much an expression of this dogmatism as is the removal of
absolute presuppositions from the criteria of truth and falsity
in the Essay on Metaphysics. But the case of sensation presents
the interpreter with examples of overstatement that tend to
stress the contribution to knowledge made by the knower, rather
than one that stresses the contribution of the object. Therefore
as we already noted, to this extent the autobiographical
interpretation is vindicated.
Collingwood's position of the relationship of knower and
known was from beginning to end one which is developed in direct
opposition to what he understood to be the fundamental tenet of
realism. He never succeeded in utterly freeing himself from this
preoccupation with the errors of realism, and it was perhaps
this preoccupation that ultimately drove him to overstatement.
It may be unfortunate that overstatements were made at all, but
it is at least consistent with the overall interpretation
offered in the Autobiography.
We end this chapter with a final summary of Colling-
wood's core philosophy of mind.

419
TABLE 8
THE PHILOSOPHY OF MIND
1. The field of human experience consists of two kinds of ex-

perience, feeling and thinking. (PA, 157).
a. Feeling is the here-and-now presence of a field of

successive sense, each with its emotional charge. Re-
lative to thought, feeling is simple, private, evane-
scent, perpetually in flux, and appears to arise in us
independently of our willing it. Feeling exists at the
psychical level, is rooted in our physiological
processes, and is (in some sense) the foundation upon
which the superstructure of thought is erected. The
study of feeling is the proper domain of behavioral
psychology. (PA, 157-64, 205).
b. Thought is the deliberate act (or achievement) of an

agent, and can succeed or fail. Relative to feeling,
thought is bipolarly complex, public, capable of genuine
recurrence and contradiction, and is somehow consequent
on feeling. It exists at the mental level, and is not
subject to empirical investigative methods alone, but is
rather the object of criteriological sciences. All
thought presupposes feeling. (PA, 157-60, 164-66, 221).
c. A science of mind proceeds by means of inquiry into

consciousness, or by putting questions to conscious ness
and answering them with data provided by consciousness
itself. From the point of view of a sci ence of mind,
feeling is an apanage of mind (something it has), but a
constituent of experience (something it is); thought is
the only constituent of mind, but it is also a
constituent of human experience. (PA, 20506; NL, 1.61,
4.14-4.2; IH, 291-94).
2. Feeling has a double character, sensation and emotion, united

in any given experience, but distinguishable by acts of
attention. (PA, 160-62; NL, 4.1).
a. Sensation is the activity of seeing, hearing, tasting,

etc. something; and what is sensed in these acts are
420
sensa: colors, sounds, tastes, etc. Sensation is prior

to emotion as capable of being attended to apart from
emotion, but not temporally, logically, or causally
prior. A sensum can be "sterilized" by ignoring its
emotional charge. (PA, 160-62).
b. Emotion is the "charge" accompanying a sensum. Relative

to sensation, emotion is secondary, or capable of being
"stripped off" a sensum and ignored. It is likely that
emotion is present in all experience, and exists in
modified state up to the level of intellect. (PA,
160-62, 169-70, 221, 294).
c. All thought presupposes feeling; and thought alters

feeling, which becomes by successive development due to
thought, imagination, appetite, passion, desire, etc. by
acts of practical thinking. (PA, 230-34; NL, 7.1-7.24,
9.55, 11.1-11.24).
3. Thought is distinguished as primary and secondary, on the

basis of what is being thought about, or as practical or
theoretical depending on what it affects. (PA, 164-68; NL,
1.63-1.68, 10.51).
a. 1st order thought (or empirical thinking) is concerned

exclusively with feeling, relations between sense, be-
tween sense and sensation, between sensation and emotion,
etc. (PA, 164-66; cf. NL, 4.31).
b. 2nd order thought (or intellect) is concerned with other

acts of thought; it is thought about thought. Its acts
may be abstract, analytical, conceptual, propositional,
rational, etc. (PA, 166-68, 253-54; NL, 6.58-6.59).
c. Practical thought is making up one's mind to (do some-

thing to oneself or to one's environment). Thought is
primarily and always practical, because it always has an
effect on oneself or on one's environment. (NL,
1.65-1.68, 7.22, 9.35, 14.3; PA, 289; IH, 310-12).
d. Theoretical thought is making up one's mind that (such

and such is the case, or is not the case, or would be the
case if . . . , etc.). (NL, 1.64-1.67, 14.3, 14.35; PA,
253, 289; IH, 310-12).
421
4. Attention (or awareness or consciousness) is the absolutely

fundamental act of thought which stabilizes and perpetuates
feeling-acts long enough that they may enter into relations
with other feelings and with thought. Attention is the act
of appreciating something present as a concrete individual,
just as it stands, before analyzing it. (PA, 203-04).
a. 1st order attention is an act of first order thought; it

divides the field of feeling into focal and penumbral
regions, perpetuates and domesticates feelings --i.e.
stabilizing feelings and asserting ourselves as the
owners of these feelings and able to dominate them. It
is therefore the free act of an agent, an act free in a
sense intermediate between the freedom of sensation (the
freedom to respond in several ways to a stimulus) and
the freedom of intellect (the true freedom of choice of
alternative plans or courses of action). (PA, 203-08,
215-22).
b. 2nd order attention is an act of 2nd order thought; it

is the practical act which isolates and identifies a
re-presented imaginatum by naming it, or by performing a
linguistic act of meaningful expression. As a 2nd order
act of thought, attention is abstractive or conceptual.
(NL, 4.3-4.68, 6.2-6.36; IH, 242, 291).
5. Imagination is feeling altered by consciousness--feeling as

selected, perpetuated, and domesticated by attention.
Imagination is another name for consciousness, awareness, or
attention, but it is distinguished as the result of at-
tention or the result of the conversion of a sensum unin-
terpreted into one ready for interpretation or into one
interpreted by an act of 2nd order thought. It is an a priori
function of thought which operates at the level of 1st order
thought as perception, and at higher levels as artistic
interpretation, scientific hypothesis formation, and
historical reconstruction (PA, 209-15; IH, 242, 291).
6. There are therefore three stages logically distinguishabl in

the life of a feeling:
a. Bare feeling is something given in sensation, and if

unattended to, is carried away in the flux of feeling.
Its being is its presence to sensation, and it is re-
lated to thought as an uninterpreted sensum.
422
b. Conscious feeling (or feeling as the object of first--

order attention or of simple consciousness) is feeling
perpetuated and dominated by consciousness or imagina-
tion. Its being is its re-presentation to consciousness,
and it is related to thought as a sensum ready for
interpretation.
c. Related feeling (or feeling as the object of second-

order attention, or of consciousness proper) is feeling
placed in its relation to other feelings, something
constructed or apprehended inferentially by the work of
intellect. Its being is relational, and it is related to
thought as an interpreted sensum. (PA,
d. "Sense datum" refers equivocally to all or any of these

above three, but it is properly speaking only (c),
related feeling. (PA, 169-70).
7. The expression of feeling is a bodily act (a practical act)

having levels which correspond to the stages in the life of
a feeling).
a. Psychical expression is the involuntary muscular and

glandular register of feeling, as wild and disorganized
as these feelings themselves. (PA, 228-34, 266; NL,
6.58).
b. Imaginative expression or language in the broad sense is

the activity of discourse, or of meaning something (a)
by something else (b), where meaning (a) is an act of
theoretical consciousness, and (b) is a practical
activity, the production in oneself or others of a
physical vehicle for that meaning. At its lowest level
language is simple naming; this is the level of
conceptual thought. (NL, 6.2-6.28).
c. Intellectual expression is language at the level of

propositional or analytic thought and all levels beyond
this (e.g. rational thought). It is thought expressed as
questions, propositions, inferences, reasons why, etc.
(NL, 6.1-6.29, 6.57-6.59; PA, 221, 224-54).
CHAPTER VIII
LOGIC, LANGUAGE, AND MENTAL

ACTS
1. Introduction.
In Chapter I we found that one of the major
controversies concerning Collingwood's philosophy, as assessed
by his interpreters, is his contention in the Essay on
Metaphysics that metaphysics is an historical science --i.e.
that the proper occupation of the metaphysician is to detect the
presuppositions being made by the scientists of a certain era.
The presuppositions, insofar as they are absolute (that is, not
themselves answers to questions but standing as presuppositions
to all questions in a systematic inquiry), are neither true nor
false, and in discovering them the metaphysician acts as an
historian. His job is not to pass judgment on the truth or
falsity of these absolute presuppositions, but merely to detect
and report them. Since the entire controversy concerning the
historical character of metaphysics turns on the role of
absolute presuppositions, and since the theory of
presuppositions is part of Collingwood's question-and-answer
(Q-A) logic, we concluded that Q-A logic is itself one of the
major concerns in the interpretation of Collingwood's philosophy.
423
424
In Chapter II we found that according to the
Autobiography, Collingwood's attack on the methods of the
realists entailed, for him, the rejection of propositional logic
(F-logic) as an index for correct thinking, since it posited
truth as a property of propositions. In direct opposition to
this Collingwood proposed an alternative in the form of a Q-A
logic, in which truth is taken as a property of propositions
only as answers to questions. By reflecting on the sort of
thinking he was accustomed to doing in his own historical work,
Collingwood found that if one includes in this thought process
the presuppositions on which questions are formulated and from
which they arise, then the presupposition-question-answer
(P-Q-A) complex is not only the unit of truth and falsity in
systematic inquiry but also the index of meaning and validity.
The Q-A logic that he developed from these reflections was
formulated, he felt, in direct opposition to the false methods
of the realistic philosophers, who ignored the necessity of re-
constructing the concrete question for the propositional answer
they were abstractly criticizing--an extension of their
ignorance of historical matters. He went as far as to write up
his Q-A logic and offer it to a publisher, but it was turned
down, and he subsequently destroyed the only manuscript of it.
In Part II we found that in his early philosophy Col-
lingwood does indeed criticize the formal logic of dogmatic
realism, and in
425
Speculum Mentis he does argue that truth is a function of Q-A
complexes. But we also found that in his earliest publication,
Religion and Philosophy, there is no mention of a Q-A logic as
such, but there is a good deal of discussion about dialectic and
dialectical relations. In Speculum Mentis dialectical relations
are even expanded into a logical schema, so that the truth
embodied in a system of thought which shows development is
described as dialectical in form, and dialectic is contrasted
with formal logic as the logic of the concrete and abstract
universal respectively. Since there was some evidence that led
us to suppose that formal logic was held to be of limited
validity in its function as the justification of the consistency
of scientific conclusions, we suggested that although there is
no mention of the fact in the Autobiography, the early writings
show evidence that when Collingwood discusses "logic" we may be
dealing not with one sort of logic but with at least three: (1)
Q-A logic (discussed in Speculum Mentis, partially described in
the Autobiography, and reappearing in semi-formal attire in the
Essay on Metaphysics); (2) F-logic (the formal logic in the
tradition stemming from Aristotle--described and criticized in
Speculum Mentis, called "propositional logic" in the
Autobiography and ascribed to the realists, and discernible as
the logic employed by "exact and empirical sciences" in the
Essay on Philosophical Method); and (3) D-logic (the dialectical
logic present in germ form even in his earliest work, Religion
and Philosophy;
426
described, examplified, and exalted in Speculum Mentis; but
mysteriously unmentioned in his Autobiography).
Our task in this chapter is minimally to examine the
evidence provided by his later writings to test whether or not
he remained true to his autobiographical assertions concerning
his "revolutionary" logic. Since the most notorious and explicit
formulation of this Q-A logic is in the fourth chapter of the
Essay on_Metaphysics, it is clear that we shall have to examine
this chapter in some detail. But since our examination of his
early writings raises questions about the relationship between
Q-A logic and D-logic, and of both of these to F-logic
(questions not even hinted at in the Autobiography) we shall
have to expand our discussion to include these topics as well.
Since the Essay on Philosophical Method describes the sort of
logical relations he had in his earlier works called
"dialectical," it is clear we shall have to examine the major
doctrines of this important work as our primary source for what
Collingwood understood by what we are calling D-logic. And
finally we find that the only locus of discussion of F-logic and
logical relations in the later writings is in connection with
language and the higher-order functions of mental activity in
The Principles of Art and The New Leviathan. This necessitates
picking up the discussion of Collingwood's philosophy of mind
where we left it in Chapter VII--viz. with the analysis of the
functions of intellect.
427
Our strategy is therefore as straight-forward as it was
in the last chapter. We propose to take up these topics in the
order in which we have just reviewed them. In each case we shall
first present a brief summary of the major theses of the texts
we are considering, in tabular form. This will be followed by a
critical discussion of issues relevant to the topics of this
chapter. In the concluding section we shall attempt to pull the
strands of the discussion together in a final evaluation of what
we can reconstruct of Collingwood's thought on the subject of
logic.
2. Questions, Answers, and Presuppositions.
In Chapter IV, where we discussed the form of experience
that Collingwood called "science" in Speculum Mentis (i.e. the
form of experience that habitually takes its object to be an
instance of an abstract universal), we noticed that Q-A logic
was left in an incomplete state, inasmuch as the relationship of
questions to presuppositions was not made explicit. This
incomplete analysis led to the further difficulty of an
equivocation in the use of the term "hypothesis" as he used it
in his description of scientific thinking. This equivocation in
turn rendered ambiguous his discussion of the
hypothetical-categorical distinction on which he based his
judgment of the superiority of history over science. We shall
have more to say in the next section on the issue of categorical
and hypothetical judgments, but at this point we must make good
our promise
428
from Chapter IV to clear up the ambiguity about questions and
hypotheses.
That ambiguity rested on the description of questions
and hypotheses as "non-assertions," the equivocation being that
questions are not merely non-assertions but proto-assertions or
non-assertions which seek completion in assertion. Sentential or
propositional fragments are also non-assertions, but they are
not therefore either questions or hypotheses. Since we shall in
all likelihood never know what the contents of his early
unpublished work, "Truth and Contradiction," had to say on the
subject of Q-A logic, we are left to the surviving fragments of
that doctrine in his published writings as the only evidence we
have for reconstructing his thought on the matter. On the
evidence of Speculum Mentis we can safely assume that as of 1924
he was not employing the distinctions necessary to disentangle
questions from hypotheses. We can also state that on the
evidence of The Idea of Nature the distinction is being clearly
utilized in his lectures of 1934 (cf. IN, v, 29-30). So on the
evidence of these two works alone it appears that he had worked
out the requisite distinctions sometime between 1924 and 1934.1
_______________________
1 Knox found it incredible that Collingwood had devel-
oped his position on Q-A logic and presuppositions prior to
1932, and on this ground alone he felt justified in rejecting
the autobiographical interpretation (IH, x-xi). But as we have
already observed, there is no
429
But the date on which he actually worked out the ne-
cessary distinctions is not as relevant to our discussion as the
fact that he ultimately did work it out to his own satisfaction,
and that we have an indication of his basic ideas on the subject
from the fourth chapter of the Essay on Metaphysics. Since the
summary is given there in semi-formal outline already, we shall
here merely summarize it in tabular form, interspersing his own
notes and comments between the various definitions and
propositions. (His "notes" were numbered; his comments were not.)
________________________
such direct assertion in the Autobiography, and even if there
were it is hardly grounds for rejecting the autobiographical
interpretation without sufficient evidence to the contrary. We
have already stated that we do not accept Knox's authority in
his claim to have had access to such unambiguous and unqualified
evidence in Collingwood's unpublished manuscripts, since to do
so would be guilty of uncritical historiography.
430
TABLE 9
QUESTION-AND-ANSWER LOGIC IN THE ESSAY ON METAPHYSICS: Q-AM
Proposition 1. Every statement that anybody ever makes is made

in answer to a question. (EM, 23).
Comment 1. Statements and questions may be made to someone

else or to oneself (EM, 24).
Comment 2. In proportion as a person thinks scientifically

he knows that his statements are answers to
questions.
Note 1. A question is logically prior to its own answer.
Comment 3. A question is also temporally prior if thinking

is scientific, but the temporal priority is such
that the question does not cease to be a question
when the answer begins, but continues for the
whole duration.
Comment 4. An answered question does not cease being a

question; it only ceases being an unanswered
question. (EM, 24).
Definition 1. Let that which is stated (i.e. that which can be

true or false) be called a proposition, and let
stating it be called propounding it (EM, 25).
Note 2. The use of the word "proposition" is exclusively

limited to that which is stated, and "pro-
pounding" only for proposition. This is not
ordinary usage, which would allow also that a
supposition or question be propounded as well.
Proposition 2. Every question involves a presupposition.
Comment 5. Any given question directly or immediately in-

volves only one presupposition, that being the
presupposition from which it arises. Indirectly
it may involve more than one. (EM, 25).
Comment 6. The answer to any question presupposes whatever

the question presupposes (EM, 63).
431
Definition 2. To say that a question "does not arise" is the

ordinary English way of saying that it involves a
presupposition which is not in fact being made
(EM, 26).
Comment 7. One can make presuppositions without knowing it,

or without knowing what presupposition one is
making ((cf. comments 2 and 5)).
Definition 3. The fact that something causes a certain question

to arise I call the "logical efficacy" of that
thing (EM, 27).
Definition 4. To assume is to suppose by an act of free choice.
Comment 8. All assumptions are suppositions, but not all

suppositions are assumptions.]
Comment 9. Some suppositions are made without being aware

that they are made, or without being aware that
others might possibly be made. (EM, 27).
Proposition 3. The logical efficacy of a supposition does not

depend upon the truth of what is supposed, or
even on its being thought true, but only on its
being supposed (EM, 28).
Comment 10. In scientific thinking it is possible and often

profitable to argue from suppositions which we
know to be false, believe to be false, or neither
know nor believe to be false or true (EM, 28).
Proposition 4. A presupposition is either relative or absolute

(EM, 29).
_____________________
1 Collingwood does not further define "assumption" and
"supposition," but from his remarks it appears that "supposi-
tions" are presuppositions (and may therefore be either absolute
or relative), and that "assumptions" are presuppositions of
which we are conscious. Since we are not always conscious of our
presuppositions (Definition 4 and Comment 9), it follows that
not all suppositions are assumptions.
432
Definition 5. By a relative presupposition I mean one which

stands relatively to one question as its pre-
supposition, and relatively to another question
as its answer (EM, 29).
Comment 11. To question a presupposition is to demand that it

should be verified; that is, to demand that a
question should be asked to which the affirmative
answer would be that presupposition itself, now
in the form of a proposition. Hence to speak of
verifying a presupposition involves supposing
that it is a relative presupposition (EM, 30).
Definition 6. An absolute presupposition is one which stands,

relatively to all questions to which it is re-
lated, as a presupposition, never as an answer
(EM, 31).
Comment 12. Absolute presuppositions are not verifiable(EM,

32).
Comment 13. The use of absolute presuppositions in science is

their logical efficacy; it does not depend on
their being true or false, but only on their
being supposed.
Proposition 5. Absolute presuppositions are not propositions

(EM, 32).
Comment 14. That is because they are never answers to

questions (EM, 32).
Comment 15. The distinction between truth and falsehood does

not apply to absolute presuppositions at all,
that distinction being peculiar to propositions
(EM, 32).
Comment 16. Any question involving the presupposition that an

absolute presupposition is a proposition is a
nonsense question. This includes such questions
as "Is it true?" or "What evidence is there that
it is true?" (EM, 33).
433
Now even though, incredibly enough, Collingwood does not
here define the terms "question," and "presupposition," and
although the term "hypothesis" does not appear in this outline,
it is nonetheless clear that since an hypothesis is a
supposition consciously made within a given inquiry,2 we may
take it that an hypothesis is what Collingwood is calling an
assumption. With the distinction between questions and
assumptions made explicit, the ambiguity from Speculum Mentis
concerning questions and hypotheses as non-assertions disappears.
But we are also confronted with a new problem insofar as
there are several noteworthy differences between the two
versions of Q-A logic, i.e. that offered by the Autobiography
and by the Essay on Metaphysics (which we shall hereafter refer
to as Q-AA and Q-AM respectively). (1) We notice that Q-AA
stressed the fact that Q-A logic differed from propositional
logic inasmuch as it took the Q-A complex as the unit of
meaning, truth, and validity, rather than recognizing these to
be properties of propositions as such (see above, Table 3 nos.
2, 3, and 4). In Table 9 we notice there is no mention of
meaning or validity, and truth or falsity is taken as the
defining characteristic of propositions as such (Def. 1).
_________________________
2 We assume that from the point of view of scientific
inquiry the notion of an unconscious hypothesis would not be
acceptable, since one could never know when the conditions for a
satisfactory demonstration of it were fulfilled. Cf. Table 9 ,
Comment 2.
434
(2) The strict Q-A correlativity of Q-AA is here universalized
and stated in terms of persons (Proposition 1). (3) In
connection with the function of presuppositions in "raising"
questions, which is never defined in Q-AA, Collingwood here
introduces the notion of "logical efficacy," or the property of
"something" (e.g. suppositions) which "causes" questions to
arise (Definition 3, Proposition 3).
At first these differences seem innocuous enough. The
first falls back on the ancient tradition dating back to Plato,
that the minimal unit of truth or falsity is the proposition or
declarative sentence. The second states the Q-AA requirement of
correlativity in terms of persons engaged in acts of questioning
and answering. And the third specifies the priority relationship
of presuppositions to questions and answers. The omission of the
requirement that meaning and validity hold only for Q-A
complexes and not for propositions by themselves might also be
written off as a mere oversight of Q-AM. But honesty requires us
to admit that there are actually major problems involved in
these differences between Q-AA and Q-AM, problems noticed (and
sometimes exploited) by many of Collingwood's interpreters. For
if truth or falsity is proper to propositions as such, how are
we to understand the related truth-value of questions and
presuppositions, not to mention the Q-AA assertion that truth
and falsity do not belong to propositions as such? And if Q-A
correlativity is defined
435
in terms of persons, how can Q-A logic escape the charge of psy-
chologism, which is offensive to logicians precisely because it
is necessarily rooted in the subjective processes of con-
sciousness? And how can one decide which presupposition "causes"
a question to arise within such subjective states of
consciousness?
Since these difficulties all turn in one way or another
on the Q-AA assertion of the priority of the Q-A complex over
the proposition as the unit of meaning, truth, and validity, a
priority inadvertently called into question by Q-AM, we shall
examine each of these logical functions in turn. If the Q-AA
claim cannot be upheld, we shall have to ask whether Q-AM is
sufficient correction of his earlier view to sustain the Q-AA
claim for Q-A logic as an alternative to traditional F-logic.
(1) Let us examine first the claim of the Autobiography
for a Q-A meaning-priority over the meaning of a proposition as
such. In Q-AA Collingwood insisted that (a) in order to find out
what a proposition means the question to which the proposition
is an answer must be known; and also (b) since each answer is to
a certain specific question in the systematic inquiry, it is
possible using the same piece of evidence (presumably the
textual assertion) to argue back from a propositional answer to
its question (see Table 3, nos. 4 a-d, 6 a-b). But these two
requirements seem impossible to fulfill, since it is difficult
to see how one can ever know to which question an answer
belongs, and hence which

436
presupposition-question (P-Q) complex is specifically prior to
it, unless one had independent and prior knowledge of the
meaning of the propositional answer (cf. Mink, MHD, 128).
Supposing the assertion is "Eric is going to Elgin"; how is one
to decide whether this assertion belongs to the question, "Where
is Eric going?" or to the question "Is the moon more or less
than 106 kilometers distant from the earth?" Without knowing
that the proposition about Eric has no evident
meaning-correlation with a question concerning the distance
between planetary bodies, the answer appears unassignable to one
question or P-Q complex rather than another.
Collingwood might well reply that the Q-A complex we are
discussing is an absurd example, because it does not meet the
requirements (a) that they belong to the same systematic
inquiry, and (b) that a question must be relevant to that in-
quiry. Therefore a question cannot be fished out of a
"hyperuranean lucky bag" and stuck together with an utterly un-
related question. But absurdity is what we use as an index for
the passage into incoherence, and we wish to know how Col-
lingwood's Q-A logic prevents us from such a lapse. On his
grounds such a lapse appears not only possible but inevitable,
since we do not have any criteria for deciding what constitutes
"sameness of inquiry." But even laying aside this important
omission, if we take the case of a recognizably systematic

437
inquiry (and in so doing tacitly import requirements not
specified by Collingwood), similar objections arise. Supposing
the inquiry is legal, and we are trying to find out what the
question was that is Q-A related, to an assertion-a slip of
paper bearing the words, "Eric is going to Elgin." Is this the
answer to the question, "Where is Eric going?" or "Who is going
to Elgin?" or "Is anyone going to Elgin?" It appears that in
even a fairly simple case with directly related questions and
answers, an answer may retain its meaning while being Q-A
related in several Q-A complexes. In short, it appears that the
meaning of a proposition may be Q-A complex independent.
Obviously one of the other of the two Q-AA requirements
must be abandoned, and since we wish to retain the notion of a
systematic inquiry for Q-AM, but not necessarily the
meaning-requirement, it may be necessary to drop the latter and
declare Q-AM to be an improved revision. If we adopted this
conclusion we evade the viciously circular argument that "we
cannot know what a statement means unless we know what question
it answers, but we cannot decide which question it answers
unless we know what it means" (Mink, MHD, 128). But we do not
yet know if there is some special sense of the highly ambiguous
term, "meaning" which Collingwood may have in mind, and which
may escape such objections as these.

438
(2) But there is a second logical function attributed to
the Q-A complex in the Q-AA version, but not appearing in
Q-AM--i.e. validity. According to the Autobiography, "no two
propositions can agree with or be contradictory to one another
unless they are answers to the same question" (Table 3, 3).
Again, even setting aside the issue of how one establishes which
question corresponds to a given answer, it appears to fail to
meet objections on the grounds of fairly straightforward
examples. In our example from the previous section the
contradictory of "Eric is going to Elgin" would be, on
propositional grounds alone, "Eric is not going to Elgin," and
it is capable of being constructed without any reference to a
common question, whatever that might be. What is worse, we can
say that these propositions are contradictory even though they
may be answers to different and even opposing questions: the
first may be an answer to "Is Eric going to Elgin?" and the
latter to "Is Eric not going to Elgin?" They are obviously not
the same question, yet they are contradictory answers (cf.
Krausz, CEPC, 225). Furthermore, as we saw in the example from
the previous section, one and the same answer may be the answer
to several different questions, without losing its identity,
i.e. its agreement with itself. It would appear, therefore, that
validity and such validity relations as agreement or
contradiction are also Q-A context independent.

439
With this line of argument Collingwood would be even
less tolerant then the former. The sense of sameness he is
talking about, he would reply, is not that of an abstract class
or universal, or even of propositions constructed on these
grounds. "Sameness of question" is determined on the basis of
sameness of an historical process, the process of a concrete
inquiry, where the researcher holds before his mind a question
the central meaning of which is continually being modified in
the light of new data or partial answers to it (cf. Table 3, 3
a-b).
Since this reply borders on the relationship of Q-A and
D-logic, we shall have to postpone a more satisfactory
consideration of it until we have had a chance to examine the
latter more closely (which we shall do in the next section). But
we may note here that it still fails to meet the objection. From
the point of view of the F-logic that Q-A logic is supposed to
replace, it is inadequate: F-logic makes the claim that in the
case of certain propositions it is possible to decide on their
validity on structural or syntactic grounds alone, i.e. on the
placement and grouping of their logical constants, irrespective
of its location in a Q-A complex or of the specific meaning of
its terms. The assertion, "Eric is going to Elgin and it is not
the case that Eric is going to Elgin" is invalid on formal
grounds alone, because there is a self-contradiction involved in
the statement. Such a self-contradiction would even render a
question
440
invalid, insofar as questions of the form, "What is the A such
that A and not-A?" can have no valid answer for any possible
substitution for A in the question. And if Collingwood were to
point out that this is what he meant when he said that a ques-
tion is "right" if it enables the inquiry to proceed (Table 3 4
c), we should have to reply that a contradictory answer stops
the inquiry from proceeding not because it is the "wrong" answer
but because it is self-contradictory, which it would be no
matter which Q-A complex it appeared in.
Perhaps we are flogging a dead horse, insofar as there
is no mention of contradiction or validity relations in Q-AM.
But then what sort of logical relations are there between pro-
positions, questions, and presuppositions? We shall take this up
again when we examine the consistency claims for Q-A logic.
However, we have one further logical function of Q-A complexes
to attend to, one which appears in both Q-AA and Q-AM.
(3) Although Collingwood states in Q-AM that the
proposition is that which can be true or false (Table 9, De-
finition 1), he also states that every proposition is an answer
to a question (Table 9, Proposition 1), so that it appears that
he wishes to retain the Q-AA requirement of strict Q-A
correlativity with respect to truth claims (Table 3, 4). While
it is true that Q-AM does not explicitly state that the presence
441
of a question (with or without its presupposition) is a
necessary condition for the truth or falsity of propositional
answer in a P-Q-A complex, this was clearly his intention in
Q-AA, so we wish to examine the sense in which a proposition is
dependent for its truth on the P-Q complex antecedant to it. Now
in both Q-AA and Q-AM Collingwood specifically states that the
presupposition of a Q-A complex need not be true, but only
supposed; and he never suggests that questions by themselves
have any assignable truth value aside from their answers or
presuppositions, so that it is not unreasonable to assume that
questions are neutral with respect to truth or falsity. But then
if both questions and presuppositions need not have any
assignable truth value, from whence comes the truth or falsity
of the answer?
Collingwood would no doubt reply that he never said that
an answer is deducible from a question and its presupposition;
in fact he would say that its not being deducible is essential
to the relationship of presupposition. If there were no need for
inquiry, then there would be no need for presuppositions, since
questions would not "arise" and propositions could simply be
deduced from one another as in a purely formal system.
Collingwood is speaking of the sort of inquiry which is not
closed or complete with respect to propositions and their
relations. He is discussing on-going inquiries in which
discoveries are being made, real questions are arising, answers
are being verified in ways appropriate to

442
the systematic inquiry at hand, and assumptions are being called
for as an aid to deciding alternative answers to questions.
But then the truth value of a propositional answer is
being decided by verificational processes outside the Q-A com-
plex itself, and we are faced with the alternative of adding a
further requirement to the truth claim for the Q-A complex (i.e.
that a Q-A complex is true or false only if the answer is
verifiable or falsifiable), or dropping the claim altogether
(i.e. that the truth or falsity of a proposition is dependent on
its being an element in a Q-A or P-Q-A complex). But in either
case it appears that truth is something that is Q-A complex
independent. In our previous example, if "Eric is going to
Elgin" is true only if Eric is indeed going to Elgin and false
if he is not, this appears to be the case whether or not "Eric
is going to Elgin" belongs to the question, "Where is Eric
going?" or "Who is going to Elgin?" Unless Collingwood has some
special sense in mind for the term "truth," we are forced to
admit that even this requirement fails to hold for even a simple
example.
But perhaps we have disproven too much. There is cer-
tainly some sort of logical connection between a proposition and
its presupposition, and there is also a logical relation between
a question and its presupposition. And we can say that a
proposition that is a valid answer to a question presupposes
whatever the question presupposes, so that

443
there is some sense in which answer, question, and
presupposition form a logical complex. But it appears that in
attempting to sort out the logical relations of the terms of
this complex Collingwood conflated a number of distinct
characteristics of linguistic entities--viz. meaning, truth, and
validity--and then assigned them all to the Q-A complex. But (1)
meaning is a linguistic function that applies minimally to terms
(individual words, or phrases and propositions regarded
precisely as terms, and capable of being used as the subject of
an assertion--e.g. of the form, "A means B"); (2) truth is a
function minimally applicable to propositions (whatever says
something (a predicate) about something else (a subject)); and
(3) validity is a function of complexes of propositions bound
together by logical connectives. On the face of it, it appears
unlikely that anything like a Q-A or P-Q-A complex can satisfy
logically appropriate conditions for being the minimal unit for
all three of these functions.
However it also seems unlikely that an Oxford professor
of philosophy, respected enough to be named to the distinguished
chair of Waynflete Professor of Metaphysics, would be wholly
ignorant of such an elementary logical distinction. We therefore
suspect that something has escaped our notice. Perhaps
Collingwood does have a special usage in mind for each of these
functions (meaning, truth, and validity), so

444
that they may after all escape some of these objections.
Pursuing this lead will take us into a discussion of
Collingwood's views on language and intellectual consciousness,
which will occupy us in section 4 of this chapter, immediately
following our analysis (in section 3) of his Essay on
Philosophical Method, the distinctions of which are required to
make sense of his definitions both of language and of mind (i.e.
definition by means of what he calls a "scale of forms").
But we have one last claim for Q-A logic to attend to
before pursuing this line of inquiry. According to the
Autobiography, Q-A logic is supposed to be an alternative to
propositional F-logic, which he says it replaces. Minimally this
must mean that it can do the same things that F-logic does, and
satisfy the claims that F-logic makes for itself. And there is
much to recommend a direct parallel between Q-A and F-logic. If
we wish to understand what sort of logical relations are
involved in P-Q-A complexes we must begin by assuming that it is
a function more akin to validity than to truth or meaning, since
it involves a relationship between elements that are on a level
with propositions rather than with terms, and it relates these
elements within a complex which resembles a group of
propositions linked by logical connectives. Secondly one can say
that there are certain truth-functional relationships that exist
between the elements of a P-Q-A complex, although it is not
straightforwardly one
445
of entailment:3 questions alone nor questions with their
presuppositions nor presuppositions alone do not entail their
answers, even though a propositional answer alone can be shown
to entail certain presuppositions (e.g. (a) "There is a noon
train to Elgin," (b) "Who is taking the noon train to Elgin?"
and (c) “Erik is taking the noon train to Elgin,” where (a) does
not entail (c), but (c) does entail (a)).4 Similarly within a
P-Q-A complex one can say there are certain logical
relationships that exist between the question and the other
elements: for example that if a propositional answer and one of
its direct presuppositions are known, a valid question can be
reconstructed and interposed between them
__________________
3 What the exact truth-functional relationship is be-
tween a proposition and its presupposition has been a matter of
lively contemporary debate into which we shall not enter here.
But we may note in passing that most of the contending
definitions of presupposing take into account the difference
between presupposition and entailment. For a good discussion of
the subject and an excellent bibliography, see Nuel D. Belnap
and Thomas B. Steel, The Logic of Questions and Answers (New
Haven, 1976).
4 Cf. P. F. Strawson, Introduction to Logical Theory

(London, 1952), pp. 174-75: "If a statement S presupposes a
statement S' in the sense that the truth of S' is a precondition
of the truth-or-falsity of S. then of course there will be a
kind of logical absurdity in conjoining S with the denial of
S' . . . . But we must distinguish this kind of logical ab-
surdity from straight-forward self-contradiction. It is self-
contradictory to conjoin S with the denial of S' if S' is a
necessary condition of the truth, simply, of S. It is a dif-
ferent kind of logical absurdity to conjoin S with the denial of
S' if S' is a necessary condition of the truth or falsity of S.
The relation between S and S' in the first case is that S
entails S'. We need a different name for the relation between S
and S' in the second case; let us say, as above, that S
presupposes S'." Notice that Strawson's definition holds
directly between statements or propositions, and does not in-
volve questions at all. Cf. note 5, below.
446
(in our above example, (b) is constructable directly from (c)
and (a)--assuming that we know that (c) is an answer to a
question, and that (a) is a common presupposition to both the
question and its answer); that from a propositional answer and
an appropriate question one or more presuppositions can be
constructed; that from a question and its presupposition(s) the
range of alternative possible answers is limited; etc.5 And
finally, one can say that the terms used in any given P-Q-A
complex cannot be used with shifting meanings without destroying
the sense (and consequently the logical relations) of the P-Q-A
complex itself (in our above example question (b) and answer (a)
are compatible with presupposition (a') "There are no passengers
on the noon train to Elgin," only if "taking the train" in (b)
means "driving the train," and Eric is an engineer; otherwise a
P-Q-A complex consisting of (a')-(b)-(c) would involve a
self-contradiction, and therefore would be nonsensical).
__________________________
5 Cf. P. T. Geach, Logic Matters (Berkeley, 1972), p.
82: "I say that P presupposes Q when, if Q is not true of an
object x, the question does not arise whether or not we ought to
predicate P of x, and thus neither P nor its negation is true of
x . . . . This relation of presupposition is quite different
from entailment . . . ." Cf. also Belnap and Steel, p. 113: "A
question, q, presupposes a statement, A, if and only if the
truth of A is a logically necessary condition for there being
some true answer to q. Evidently it is a consequence of this
definition that A is a presupposition of q, if and only if every
direct answer to q logically implies A." Both of these
definitions introduce the notion of a question directly into the
definition of presupposition. in contrast to Strawson (above,
note 4).
447
Thus it is not difficult to show that there are
specifiable logical relations involved in Q-A logic with its
distinctive P-Q-A complexes; and we can also say with some con-
fidence that meaning, truth, and validity do have a bearing on
these P-Q-A complexes, inasmuch as they require an identity of
meaning of their terms, have truth-functional relationships
between the elements of the complex, and validity or entailment
relationships between propositional answers and their
presuppositions. But this seems to reverse what Collingwood was
trying to say about Q-A logic (at least in Q-AA): instead of Q-A
complexes determining the meaning, truth, and validity of
propositions, it appears that all three are established
independently of those complexes, and have a logical priority
over them. Without some sort of deductive inference structure
how could one decide whether a question legitimately arises from
a given presupposition, and consequently what counts for a real
answer in a P-Q-A complex? And without some means of
establishing the semantic identity of the terms, we have no way,
within Q-A logic, to determine the limited range of answers
and/or presuppositions allowable for a given question. It
appears, therefore, that far from being an alternative to
F-logic, Q-A logic presupposes it.
But the final blow to the thesis that Q-A and F-logic
are alternative logics is that Q-A logic fails to meet the
claims of
448
consistency, completeness, and formality that F-logic not only
claims but demonstrates for itself--at least for the sort of
axiomatic F-logic with which Collingwood was familiar.6 Now in
some respects Q-A logic as Collingwood discusses it has the
appearance of a formal system: it states some criteria for what
counts as an element within that system (questions, answers,
presuppositions--although as we noted, questions and
presuppositions are never defined), as well as how they are to
be introduced or eliminated (assumptions are made by acts of
free choice, questions arise in accordance with the logical
efficacy of a presupposition, questions are eliminated if their
presuppositions are not made, or if they are raised with respect
to absolute presuppositions, etc.). But this appearance of
system breaks down when it is examined more closely for its
consistency, completeness and formality.
(1) Concerning consistency we have perhaps already said
more than enough. In summary we may note that if it is possible
in some cases (i.e. complex propositions bound by logical
connectives) to decide on the truth or falsity of a proposition
on formal grounds alone, without knowing anything about the
question which it answers, or without requiring that
___________________________
6 Cf. I. M. Bochenski, The Methods of Contemporary
Thought (New York, 1968), PP. 70-72. On Collingwood’s notion of
F-logic and its sources, see EPM, 148n, 153; Krausz, CEPC, 3n.
449
the variables within a proposition be specified semantically,7
then it seems not only that the Q-AA contention that meaning,
truth, and validity are properties not of propositions alone but
of Q-A (or perhaps P-Q-A) complexes is false, but so also is the
Q-AM assertion that some presuppositions (viz. absolute
presuppositions) cannot be true or false because they are not
(or cannot be) answers to questions. On formal grounds alone we
may argue that any assertion is false if it contains a formal
contradiction.
In the sequel we may find that Collingwood might have
meant that the question of formal coherence belongs to a dif-
ferent systematic inquiry (i.e. part of a Q-A system where the
questions raised are precisely about the coherence of state-
ments). But we may still observe that formal questions cannot
ever be irrelevant in any systematic inquiry to which Q-A logic
is applicable: insofar as an inquiry is systematic (or claims to
be so) its coherence is presumed.
(2) Concerning completeness, we must ask, first of all,
what is the full set of expressions of which questions, answers,
and presuppositions are members? Collingwood says that they are
the elements of a systematic inquiry, but that is hardly enough:
a systematic inquiry may also involve digging in the earth or
flying to the moon, but these activities are not part of a P-Q-A
complex. The elements of such a complex
_________________
7 Bochenski, p. 33; cf. IH, 253.
450
obviously belong to a set of expressions in a language of some
sort, a set at once broader than that of F-logic (because
F-logic is ordinarily confined to the set of all propositions
that can be true or false, whereas Q-A logic includes questions
and presuppositions, which are not propositions nor
proposition-fragments and yet have logical functions), and
narrower than it (because Q-A logic is only concerned with those
P-Q-A complexes that are part of a systematic inquiry, and not
the set of all possible propositions, whether they are part of
such an inquiry or not). Furthermore it can only include the set
of all meaningful expressions in a systematic inquiry, since
presuppositions and questions, while truth-value neutral, are
nonetheless bearers of meaning at the level of complexity of a
statement (to be a presupposition, for example, a statement need
not be true or false, but it must be intelligible--i.e. a
well-formed grammatical sentence).
It is not just a matter of having a group name for these
elements of Q-A logic, it is a matter of deciding (or having
grounds for deciding) what counts for a well-constructed formula
(a meaningful expression) in it and ruling out all expressions
that are not. It is clear that Collingwood has some such
criteria in mind, because he indicates that not just any
combination of terms with a question mark at the end could count
as a question (cf. Table 3, 4 a), nor can any arbitrary grouping
of words
451
count as a supposition or proposition just because it ends with
a period. Candidates for membership in Q-A logic are evidently
grammatically sound and formally coherent questions and
assertions in a language of some sort, presumably the language
accepted and employed by the inquiry in progress. But this means
that Q-A logic relies on extrinsic factors for what counts for a
candidate for membership in the Q-A system, and we must presume
such grammatical and formal criteria as essential to it. If
Collingwood presupposes all these rules in the construction of
his Q-A logic, how can we concur with his suggestion in the
Autobiography that Q-A logic was intended as an alternative to
F-logic, or that in logic he was a revolutionary?
If there is anything revolutionary in his Q-A logic it
is not that it supplants grammar and formal logic, but rather
that, if anything, it extends them or adds something to them in
their application within organized inquiry or
knowledge-acquisition, something essential that was ignored in
the reduction of knowledge to a set of propositions which merely
conform to the rules of grammar and logic. (In section 4 of this
chapter we shall take this matter up again in connection with
Collingwood's discussion of language and logic in The Principles
of Art.)
(3) Finally, concerning formality, it is clear from the
very
452
first proposition of Q-AM that Q-A logic is not staying within
the bounds of a formal system: Proposition 1 appears to be
stating a matter of psychological rather than purely logical
fact. If proposition 1 had read: "From every statement a
question can be constructed," it might have generated some
interesting formal problems--e.g. what the interrogative
function entails for propositions cast into this mode; how they
might function in relation to declarative propositions of
various sorts; etc. But instead Collingwood says that "every
statement that anybody ever makes is made in answer to a
question," where personal agency is brought into what might
otherwise be a purely logical analysis of the relationship
between questions and answers. Since the thrust of 20th century
logic is precisely in the opposite, non-psychologistic
direction, it is clear that this way of stating the relationship
of question and answer would not be regarded by formal logicians
as acceptable.9
______________________
8 I. M. Bochenski, Contemporary European Philosophy
(Berkeley, 1969), p. 253.
9 Even sympathetic critics have taken offence at this

approach: see J. F. Post, "A Defense of Collingwood's Theory of
Presuppositions," Inquiry, VIII (1965), pp. 332-54: "The reader
is first struck by the psychologism of what purports to be a
purely logical investigation . . . . The psychologism runs
throughout" (p. 333). Post sets out to eliminate the formal
absurdities from Collingwood's theory of presuppositions, as
does Krausz in his article for the book of critical essays on
Collingwood which he edited: see CEPC, 222-40. The
reconstructions by both Post and Krausz are admirable, but fail
to take into account what Collingwood had in mind
453
But aside from this there is still a question of whether
Proposition 1 (and notice that it is called a proposition rather
than a definition) is factually true. Is it reasonable to assume
that every statement that is ever made by anyone is made in
answer to a question? What about commands, warnings, and
exclamations; petitions filed in court; classroom lectures,
declarations of religious faith, sentences read during a
filibuster in Congress, etc. (cf. Mink, MHD, 127-28)? But even
aside from exceptions, we are certainly not aware that this is
always the case. Collingwood's assurance that this priority of
questions is purely logical (Table 9, Note 1) may soften the
claim a bit, but it does not render it any less uncertain. If a
question is logically prior to its answer, how are we to
understand this priority? Collingwood discusses it in terms of
causality, since a question arises as a result of the causal
operation of what he calls the "logical efficacy" of a
supposition (Table 9, Definition 3). Even if we could set aside
our hesitation to accept something like an efficient cause
operating at the logical level of thought, we are still
_____________________
for Q-A logic. He gave no indication that he ever intended it to
be reducible to, or incorporatable into, a formal system of
logic. A formal "erotetic logic" treats questions just as formal
logic treats propositions--as so many finished products. It is
not concerned with the process of asking questions, but only
with the formal properties of questions once they are already
asked. But Collingwood was interested precisely in the process
of ongoing inquiry, where questions are arising; he insisted
that the process was completely reducible neither to
psychological feeling states nor to the rules of formal logic,
and yet had a "logic" or rationality of its own.
454
left wondering about the priority of questions to answers. If it
is a priority in thought we wish to know why it is not merely
carried away in the stream of immediate consciousness, or why it
is not "merely psychological" or subjective (cf. IH, 292). What
is the objective dimension to a question--i.e. that aspect of it
which is not carried away in the stream of subjective
consciousness? Is it not merely subjective and hence merely psy-
chological unless one can also specify this objective component?
This is where the formalist would press him hardest for
an answer, and it is difficult to see how he would reply. Every
time Collingwood uses the term "logical" to describe the
priority of Q-A complexes, or of presuppositions to questions,
or of questions to answers, he is tacitly appealing to criteria
which he simultaneously appears to dismiss—i.e. to the objective
aspect wherein the relationship does not depend on mere
co-location in subjective consciousness. Without such objective
criteria we could hardly know what assumptions were uniquely
efficacious in causing a certain question to arise, since in our
subjective consciousness we are seldom aware of what all the
assumptions are that we are making relevant to a given
situation. But to say that anyone employing Q-A logic is bound
by the relationships he is describing is to say that the rules
of Q-A logic hold independently of subjective thought-contexts,
and are
455
binding directly on the related elements of the P-Q-A complex.
Without such an objective specification Collingwood's Q-A logic
is open to the sort of criticism that one of his sympathetic
critics leveled at him: "Collingwood says that 'every thought we
find ourselves thinking is the answer to a question' ((EM, 36)).
Surely every presupposition, whether it is an AP ((absolute
presupposition)) or not, may be a thought we find ourselves
thinking. Hence every presupposition may be the answer to a
question, and it seems that there can be no AP's. Short of radi-
cal revision of the whole chapter, there is no way out of this
contradiction."l0
* * * * * * * * *
From the foregoing it is clear that on the face of it
there are serious difficulties with Q-A logic either in the Q-AA
or Q-AM versions. The P-Q-A complex does not succeed in standing
up to the autobiographical claim that it is the unit of meaning,
truth, and validity, nor does Q-A logic meet the criteria of
consistency, completeness, and formality required of it if it is
to serve as an alternative to F-logic. Unless we can find
mitigating arguments in the later writings to reconcile some
_____________________
1O Post, "A Defense of Collingwood's Theory of Presup-
positions," p. 336. Post also raises an interesting objection to
the suggestion that there is a class of assertions (absolute
presuppositions) which have no truth value: "A sentence with no
truth value would be implied by every statement, since there
would be no way for the latter to be true and the former to be
false" (ibid.).
456
of the more glaring of these difficulties we shall find it
impossible to uphold the Autobiography on this issue: it fails
to be coherent on its own much less as an interpretation of
anything else.
But perhaps misled by Collingwood's autobiographical
claim that Q-A logic is to be understood as an alternative to
replace F-logic, we have been asking the wrong questions about
Q-A logic, expecting something more of its explicit performance
than its implicit promise is capable of delivering. Rather than
being a substitute for F-logic, suppose Q-A logic is regarded as
complementary to it, perhaps even tacitly presupposing it?
Suppose it addresses not a body of determinate propositions but
rather a somewhat indeterminate situation from which definite
propositions are to be extracted? In such a situation knowledge
would be something yet to be achieved, and the asking of
questions based on presuppositions (which may state the basic
structural framework or relevant features of the situation
within which such knowledge is to be determined) would be one of
the means that bring determinacy to it. Questions would
therefore be one of the signs of the difference which knowing
introduces within the knower-known situation, signifying the
presence of consciousness or awareness at the linguistic plane
of knowledge. And the P-Q complexes would therefore

457
have a heuristic priority not reducible to, but presupposing,
F-logic and its relevant requirements within inquiry.ll
But if we try to interpret Collingwood's Q-A logic in
strictly formal terms we are committed to ruling out any "psy-
chologisms" that may appear in it, which threatens to gut the
theory as Collingwood proposes it: he would not ultimately
warrant the requirement of an F-logic which would call for the
elimination of the personal dimension to inquiry precisely be-
cause that is what he claims all knowledge is for--i.e. self-
knowledge. Therefore it is clear that just as in Speculum Mentis
he had argued against the abstract universal, so too he presents
Q-A logic as the thought of an agent employed in the acquisition
of knowledge. The combined weight of the evidence we have
considered thus far indicates that Collingwood would refuse to
give up this requirement.
But then this involves using terms like "question,"
"answer," and "presupposition" and even "meaning" and "truth,"
in something other than a univocal, F-logical way. They may not
be the merely formal entities they appear to be, but rather they
may include in their meanings a reference to their entertainment
within a personal consciousness. They may be, in short,
epistemological entities rather than merely or purely logical
ones, and as epistemological entities they must be regarded as
displaying
______________________
1l Cf. Donald S. Mackay, "On Supposing and Presuppos-
ing," The Review of Metaphysics (Vol. II, no. 5, 1948), pp. 1-20.
458
the properties of mental activities better expressed by the use
of verb-forms rather than noun forms: questioning, propounding,
answering, presupposing, etc.--the very forms that Collingwood
prefers to use.
In the next section we shall explore the requirements
that Collingwood finds necessary for dealing with the structure
of concepts of this sort--concepts that we have already
witnessed him calling "dialectical."
3. Dialectical Logic and Philosophical Methodology.
The one work in which Collingwood makes the sort of
distinctions required for a logic that can serve as an al-
ternative to traditional F-logic is in the Essay on
Philosophical Method--a work which in his autobiographical self-
estimate he called "my best book in matter; in style, I may call
it my only book, for it is the only one I ever had the time to
finish as well as I know how" (A, 118). In the Essay the subject
matter (philosophical logic or methodology) is distributed in a
manner which corresponds to the traditional (Aristotelian)
division of logic into terms, judgments, and inferences.
Chapters II-IV deal with terms--philosophical concepts or
universals (the overlap of classes, the scale of forms, and
definition and description). These chapters also discuss
classification and division, but not extensionality

459
or the distinction between the denotation and connotation of
terms. Chapters V and VI deal with the quantity and quality of
judgments, but say nothing directly about predication, the
distribution of terms, or the analytic-synthetic distinction.
Chapter VIII deals with deductive and inductive inference, but
not with the modes of valid or invalid deductive inference. And
Chapter IX even discusses the systematic claims of completeness
and consistency, but says nothing about formality (cf. Mink,
MHD, 61-62). It would therefore appear that if we are ever to
find a clue for understanding Collingwood's view of "logic" it
would be in this work.
However we have already had occasion to note that there
are formidable problems in trying to ascertain what the
standpoint of the Essay is, or how the reader is supposed to
take its substantive remarks about non-philosophical methods or
even non-methodological philosophical issues. We may recall from
that discussion that in the Essay he states that (a) the manner
in which non-philosophical concepts, judgments, and forms of
reasoning are characterized is not meant to say anything
explicitly about mathematical or empirical science, but is
rather only meant to contrast methods commonly thought to be
employed by these modes of thought and mistakenly applied to
philosophical subjects (EPM, 9, 151); and (b) he confesses in
the final chapter that his initial agreement with the reader to
treat as an hypothesis the assumption that philosophical

460
concepts have a peculiar logical structure, has been violated
throughout by appeal to philosophical experience (EPM, 222-23).
In the context of our present discussion these two difficulties
put obstacles in our path to understanding Collingwood's views
on logic. The first renders suspect any interpretation of the
Essay which takes as literally true its remarks on subjects
other than philosophy; and the second raises the issue of the
circularity of the argument of the Essay insofar as the defining
characteristics he proposes as applying to philosophy itself are
first advanced hypothetically and then "proven" categorically by
appeal to the experience of philosophers. We might get around
the first obstacle by pointing out that the Essay itself argues
that logic is a part of philosophy, and "what philosophy is" is
the very subject of the book (EPH, 2-3, 7). But this is further
complicated by the fact that Collingwood states that he will not
discuss the implications of the Essay for metaphysics, logic, or
the theory of knowledge (EPM, 7-8)--a promise which is also
violated throughout, e.g. by his discussion of the ontological
argument and the subject matter of metaphysics in the sixth
chapter, by his discussion of philosophy as a branch of
literature in the tenth, as well as his use of examples from
epistemology, logic, and ethics throughout the book.

461
Until we have had a chance to survey the key concepts
and the central argument of the Essay we cannot resolve these
difficulties of the standpoint of the Essay. But from even a
superficial reading of the introductory chapter one can draw
several conclusions about how Collingwood intended the Essay to
be taken. (1) The Essay is concerned with discussing the nature
of philosophy as an activity of thought, exemplified as
philosophical experience and expressed on certain occasions in
characteristic or typical ways (EPM, 2-3, 7); (2) Concerning
this activity certain generalizations are possible which are
capable of serving as ideal principles of method, and to which
we are bound when we are trying to think philosophically (EPM,
3-4). (3) The strategy of discussing is basically Kantian in the
sense that (a) certain "facts" (or experiences) are brought to
our attention, (b) and then described and analyzed by comparing
them with similar examples from non-philosophical (or
pre-philosophical) equivalents, (c) the comparison exhibiting a
peculiar but typical structure to philosophical concepts,
judgments, and forms of reasoning, (d) which is only possible
under certain necessary conditions (EPM, 3-4, 222-23). At the
very least, then, we are entitled to say that Collingwood's
overriding concern in the Essay is to say something that is true
for a certain kind of thought, namely philosophy. This will have
to suffice as a provisional sense of the standpoint of the Essay
until we examine the issue of its argumentative circularity.

462
As usual we will begin with a brief outline of the key
concepts and conclusions of the Essay before proceeding to a
discussion of them and an analysis of the main line of argument
in the Essay.
463
TABLE 10
PHILOSOPHICAL METHOD
1. The overlap of classes: the concepts or universals of

philosophy are exemplified in overlapping classes instead of
mutually exclusive classes. A universal concept is one which
unites a number of different instances in a class-either a
plurality of individual instances in a general concept or a
plurality of specific differentiations into a generic
concept. The overlap of species of a philosophical genus
consists in the inclusion of part of the meaning of one of
the concepts in the meaning of another (i.e. it is a
relation of intension between concepts rather than a
relation of extension between classes). Alternative ways of
stating the overlap are: (a) two or more specifically
differing concepts may be exemplified in the same instances
(i.e. they may be conceptually different but instantially
the same), so that any distinction in philosophy may be a
distinction in concepts without a difference in instances;
and (b) Aristotle's formula for the overlap, i.e. that two
concepts may be the same but their being is not the same
(being an instance of the one is not the same as being an
instance of the other). (EPM, 27, 31, 40, 50, 91). Ignoring
the first rule of philosophical method (i.e. that the
specific classes of a philosophical concept are always
liable to overlap) leads to three related fallacies:
a. The fallacy of false disjunction (ffd) states that when a

generic concept is divided into its species there is a
corresponding division of its instances into mutually
exclusive classes.
b. The fallacy of precarious margins (the positive ap-

plication of the ffd) states that if there is a dis-
tinction between two concepts there must be a difference
between their instances (or, the margins between species
of a philosophical genus can be trusted not to spread to
include all the instances of co-ordinate species).
c. The fallacy of identified coincidents (the negative

application of the ffd) states that where there is no
difference in the extension of two concepts, there is no
distinction between the concepts themselves. (EPM,
48-50).
464
2. The scale of forms: the specification of a philosophical

genus (or concept) is such that its species differ from one
another both in degree and in kind, and are related to one
another both by opposition and by distinction. They therefore
constitute a scale of overlapping forms in which (a) whenever
the variable increases or decreases, certain critical points
on the scale are reached at which one specific form
disappears and is replaced by another; (b) the variable of
each species (its specific difference) is identified with the
generic essence; (c) every species is a realization of the
genus (there is no zero-point to the scale where the generic
essence is altogether absent, but rather it starts with unity
as the minimum realization of the genus); and (d) any point
on the scale sums up (or intentionally includes) the whole
scale from the minimum specification to that point. The scale
of forms explains the overlap of classes (i.e. shows how it
is possible) insofar as successive species of a philosophical
genus intentionally include the positive content (or meshing)
of all the forms preceding it in the scale. (EPM, 56-60, 81,
89, 91). Ignoring the second rule of philosophical method
(i.e. that the specification of a philosophical concept
constitutes a scale of forms) leads to several related
fallacies, all arising from the ffd applied to differences of
degree and kind and relations of opposition and distinction:
a. The fallacy of calculation assumes that because there are

differences of degree in philosophical concepts they are
susceptible of measurement and calculation (i.e. that
they are differences of degree only and not also
differences of kind).
b. The fallacy of indifference assumes that because the

species of a philosophical genus differ in kind, they
exhibit no differences of degree.
c. The fallacy of the false positive assumes that where

terms are related in a philosophical series the relation
is not of opposition but of mere distinction (so that it
makes positive terms out of negative ones).
d. The fallacy of null opposition asserts that the terms

related in a philosophical series are opposites but not
distincts (so that the negation of positive terms have
no actuality). (EPM, 80-81, 85-86).
3. Philosophical definition: the definition of a philosophical

concept is by means of a scale of forms, beginning with a
rudimentary minimum definition (the generic essence)
465
and adding qualitatively new determinations which gradually

alter the original definition and improve it as a statement
of the concept's essence. Philosophical definitions are
descriptive (i.e. they aim at a complete listing of
attributes, but not by mere enumeration but by exhibiting
their connections); (b) are real rather than verbal (i.e.
they seek to state the essence of a concept rather than the
meaning of a word); and (c) are relative rather than
absolute (i.e. they seek to make more precise what was to
some extent definite already, rather than stating a
definition which makes the difference between knowing
something and not knowing it at all). (EPM, 92, 94-98,
100-01). The phases of a philosophical definition are:
a. the minimum specification of a concept--the lower end of

the scale, or the necessary rudimentary expression of
the genus without which it would fail to be a genus at
all;
b. the intermediate specifications--later modifications of

the minimum specification by adding new qualifying
determinations; and
c. the final specification--the phase reached when the

definition explicitly states all that can be found in
the concept and is adequate to the thing defined. (EPM,
100-01).
4. The principles of concrete affirmation and negation:

philosophical judgments are so related that every negation
involves an affirmation, and every affirmation involves a
specific (not indiscriminate) negation. Whenever a phi-
losophical assertion is made it affirms something definite
and also denies something definite. In philosophy the af-
firmative judgment (S is P and not Q) and the negative
judgment (S is not Q but P) are equally definite and specific
answers to the same question (what is S?). (EPM, 10708, 110).
Ignoring these two principles leads to two associated
fallacies:
a. The fallacy of abstract negation assumes that the

rejection of one account of a philosophical matter does
not require giving a better account of it, or that it is
possible in philosophy to negate without affirming.
b. The fallacy of abstract affirmation assumes that in

philosophy it is possible to affirm (i.e. predicate a
concept in a judgment) without denying anything definite.
(EPM, 105-07).
466
5. Philosophical judgments as universal: philosophical

judgments are universal (i.e. of the form, all S is P.), but
rather than excluding this includes the subordinate forms,
the particular judgment (every S-type is P) and the singular
judgment (every individual S is P) as well as the purely
universal judgment (S as such is P). All three types of
structure are to be found in the philosophical judgment; what
is not found is any sufficiency of one to the exclusion of
the rest. (EPM, 111-12, 115).
6. Philosophical judgments as categorical: philosophical

judgments are not merely hypothetical (i.e. of the form, if S
is P it is Q) but also are essentially categorical in
intention (i.e. of the form, S is P). As categorical,
philosophical judgments are never devoid of objective or
ontological reference. ( EPM, 121, 125).
7. Philosophical inference as reversible: the arguments of

philosophy, instead of having an irreversible direction from
principles to conclusion (as in deduction) or from data via
principles to conclusion (as in induction), have a
reversible direction, the principles establishing the
conclusion and the conclusions reciprocally establishing the
principles. Philosophical arguments escape vicious
circularity insofar as the conclusions in philosophy are
already something known, and establishing them by means of
principles means making them known in a different and better
way. Philosophical conclusions are anticipated by an
experience that possesses them in substance before its
reasoning begins, and its conclusions can be checked by
comparing them with these anticipations. In a philosophical
scale of forms, the terms "experience" and "conclusion" are
applicable to any two successive stages (or levels): the
higher level "explains" the lower. (EPM, 160, 163, 172).
8. Philosophy as systematic: it is the ideal of philosophical

discourse to be constructive or systematic--that is, as
final, complete, objective, and unified as it can be.
Although none of these ideals can be met absolutely, in
viewing philosophical science itself as a scale of forms,
its topics can be arranged as a hierarchy of overlapping
concepts, differing in degree and kind, and related as op-
posites and distincts. (EPM, 177, 186-93).
_________________________________________________________
467
No doubt this outline does scant justice to the argu-
mentative structure and detail of example in the Essay, and
perhaps raises more questions than it resolves. We cannot deal
with all of Collingwood's arguments and illustrations without
duplicating in length the Essay itself. But we shall have
occasion to discuss several of the examples in what follows, in
connection with our examination of the structure of his main
argument. But perhaps we can settle several questions with some
comments about each of the key concepts we have just summarized.
(1) The first misconception to be avoided is that
Collingwood's Essay is concerned with describing characteristics
of philosophical concepts, judgments, and inferences that belong
exclusively to philosophy. Although each of these logical units
is contrasted with its non-philosophical equivalent, the
contrast does not make an absolute distinction between the two
sorts of usages. In each case he is careful to point out that
while such characteristics are occasionally but exceptionally
encountered in non-philosophical experience, in their
philosophical employment they are typical, important, and
essential. Thus when he discusses the overlap of classes, for
example, he points out that it occurs even in empirical science:
exceptional or paradoxical border-line cases provide obstacles
in any attempt to carry out a schema of classification based on
an arrangement of species of a genus into

468
mutually exclusive classes which exhaust the membership of the
genus between them. Collingwood points out the instance of the
classification of animals (a problem discussed with insight by
Aristotle), in which amphibians overlap the classes of fish and
reptile, having characteristics of both (i.e. they are
vertebrates with lungs, like reptiles, but capable of underwater
breathing, like fish) (EPM, 30). But where such an overlap is
exceptional in empirical science and ruled out by the a priori
divisions of exact sciences,l2 in philosophy it is typical and
essential. As examples Collingwood points out the predicability
of the transcendental attributes, unity, truth, and goodness,
under all the categories (EPM, 32-33); the overlap of judgment
and inference as two species of the genus thought, in logic
(EPM, 36); the overlap of thought and action as characteristics
studied by logic and ethics respectively (EPM, 43-44); and in
ethics the overlap of good acts across the divisions of goods
____________________
12 Collingwood seemed to think that the difficulties of
border-line cases arises only in empirical science but not exact
sciences, since in the latter (i.e. in mathematics) the
divisions can be carried out a priori by stipulative defini-
tions, and hence the exclusiveness and exhaustiveness of the
species are assured (EPM, 30-31). But insofar as the overlap of
classes is defined by the scale of forms as being a relation of
overlap-by-inclusion, the system of classification of numbers
seems to fit his description of an overlap of species of a
genus, as Donagan points out (CEPC, 5-6).
469
into the species pleasant, expedient and right (EPM, 41).13
It is interesting to note in connection with this dis-
cussion that Collingwood relates the non-philosophical and
philosophical usages of a term with the pre-philosophical and
philosophical phases of a concept--which is reminiscent of Q-AA
logic, in which the sameness of two questions is said to be the
sameness of an historical process (EPM, 33; cf. above, Table 3,
no. 3a). He writes:
There are words which are used in two different ways, a

philosophical and a scientific; but the words are not on
that account equivocal; they undergo a regular and uniform
change in meaning when they pass from one sphere to the
other, and this change leaves something fundamental in
their meaning unaltered, so that it is more appropriate to
speak of two phases of a concept then two senses of a word.
For example, matter is a word used both in physics and
metaphysics . . . . Such cases are common. Mind, for the
scientist, in this case the psychologist, is the name of
one limited class of things outside which lie things of
other kinds; for the spiritualistic philosophy, it is a
name . . . for all reality . . . . Even in concepts that
have no strictly scientific phase, a similar distinction
can often be traced between a philosophical phase and a
non-philosophical. (EPM, 33-35).
_____________________
13 Mink points out that it is possible and relatively
easy to work out a system of mutually exclusive species for this
example from ethics, simply by employing the mutually exclusive
and exhaustive classes, (a) pleasant and expedient and right,
(b) pleasant and expedient, but not right, (c) pleasant, but not
expedient and not right, (d) expedient, but not pleasant and not
right, etc. (MHD, 65). One might extend Mink's suggestion and
defend the claim that by a rigorous use of stipulative
definitions, a system of classification can always be
constructed that employs mutually exclusive and exhaustive
classes. But as Mink notes, most of the difficulties in what
Collingwood says in the chapter on the overlap of classes, many
of which arise from examples which seem to presume an overlap of
extension between classes, are cleared up in the chapter on the
scale of forms (MHD, 66, 70).
470
In a later chapter in which Collingwood is comparing the li-
terary language of philosophy to the technical (i.e. artifi-
cially symbolic) language of science, he points out that the
subject matter of philosophy demands a vocabulary with groups of
words "nearly but not quite synonymous, differentiated by shades
of meaning which for some purposes can be ignored and for others
become important," and single words which have various senses
according to the ways they are used, but without being utterly
equivocal (EPM, 206-07). It is precisely in such a vocabulary as
this, and in cases of overlapping meanings to which such a
vocabulary applies, that philosophy is interested, and which
gives it its typical characteristics.
But we must also notice, before passing on to consider
further parallels between philosophical and non-philosophical
concepts, that here as elsewhere in the Essay Collingwood does
not provide us with clear guidelines for the limits of this
distinction, or with any criteria for deciding when a concept is
rightfully being employed in its scientific phase and when in
its philosophical one. Furthermore, (as even his comment in the
above quotation shows a dim awareness), lumping pre- and
non-philosophical concepts together may overlook important
distinctions which may turn out to be pertinent when the
overlapping meanings of the term in its philosophical employment
are being unpacked. Collingwood may be guilty of falling prey to
the "genetic fallacy"--i.e. assuming that an earlier stage in
the development of a concept is a necessary condition

471
for the understanding of the mature concept--in short, confusing
temporal and logical priority.
(2) Another example of the manifestation of a phil-
osophical characteristic in a non-philosophical context is the
scale of forms. Collingwood points out that the "fusion" of
differences of degree with differences of kind in a scale of
forms is not unique to philosophy, but is also familiar to
"common sense" or ordinary experience, as in social structures
which distinguish between nobility and gentry, or in criminal
codes which distinguish between capital and other degrees of
punishment (EPM, 73). It is also familiar in empirical science,
where the double criteria of degree and kind are operative in
scales of forms in which different forms are modified and
replaced by others: Collingwood cites the example of the states
of matter, e.g. ice, water, and steam, which differ from each
other in both degree (colder and hotter) and kind (solid,
liquid, and gaseous states). Other examples are the periodic
table of elements and the stages of organismic development (e.g.
embroyogenesis) (EPM, 59). He even insists on this as a way of
defining a scale of forms:
The combination of differences in degree with differences

in kind implies that a generic concept is specified in a
somewhat peculiar way. The species into which it is divided
are so related that each not only embodies the generic
essence in a specific manner, but also embodies some
variable attribute in a specific degree. In respect of the
variable, each specific form of the concept differs from
the rest in degree; in respect of the manner in which the
generic essence is specified, each differs from the rest in
kind. In such a system
472
of specifications the two sets of differences are so

connected that whenever the variable, increasing or
decreasing, reaches certain critical points on the scale,
one specific form disappears and is replaced by another. A
breaking strain, a freezing point, a minimum taxable
income, are examples of such critical points on a scale of
degrees where a new specific form suddenly comes into
being. A system of this kind I propose to call a scale of
forms. (EPM, 57).
Just from the examples cited here it is clear that there are
non-philosophical as well as philosophical scales of forms; but
Collingwood remarks that it is just in such cases where
differences of degree and kind exist in combination that phil-
osophical thought is primarily interested (EPM, 56-57).
But in this case he goes one step further, and tries to
distinguish philosophical from non-philosophical scales of
forms. In the latter the variable is something extraneous to the
generic essence (e.g. heat, which accounts for the variation of
water in its transformations of state, does not enter into the
generic essence of water as expressed in the formula, H2O). But
in a philosophical scale of forms the variable is identified
with the generic essence itself (e.g. in Plato's forms of
knowledge the variable is given as definiteness or truth, which
is an essential characteristic of all knowledge as such) (EPM,
59-60).
However, since Collingwood does not specify for his
readers what he means by the terms, "generic essence," "vari-
able," and "specific forms," rather than clarifying the issue
this distinction raises

473
a hornets nest of further difficulties. What does it mean to say
that the variable is identified with the generic essence? It is
not a matter of demanding technical definitions for these terms,
since on Collingwood's grounds this may involve him in an
infinite regress (see Table 10, 3: philosophical definitions are
by means of a scale of forms, so that the concept of a scale of
forms is defined by means of a scale of forms). What we wish to
know is minimally how he is using these terms, even if this can
be stated only in mere verbal equivalences.
From scattered remarks throughout the Essay we may offer
the following as a first approximation to this requirement. (a)
A generic essence is whatever is included in the definition of a
genus, where (b) a definition is anything which fixes limits,
discriminates, distinguishes, makes clearer or more precise, or
removes ambiguities, and (c) a genus is a universal or a
concept--i.e. that which unites a number of different things,
either as (d) a general concept unites a plurality of individual
instances into a class, or (e) a generic concept unites a
plurality of specific differentiations into a universal (EPM,
26-28, 94-95, 98-100). (f) A variable is whatever is being used
as the index of difference in degree, or that in respect of
which each specific form (species) of a concept (genus) differs
from the other specific forms in degree (EPM, 57). (g) A
specific form is any further modification, qualification,

474
distinction or determination of a generic essence, general term,
or concept (EPM, 94-95, 100). (h) Essence and property are two
species of attribute, to which correspond definitions and
theorems as two species of exposition (EPM, 95). (i) An
attribute is anything which serves to qualify, modify,
distinguish or determine a genus.
Armed with these preliminary definitions we may now
return to the distinction between a philosophical and non-
philosophical scale of forms: in the former the variable was
said to be identified with the generic essence, in the latter it
was something extraneous. Paraphrasing this distinction using
our preliminary definitions yields statements something like
these: in the specification of a concept one species differs in
degree from another species in virtue of some attribute which in
the case of a non-philosophical concept is not included in the
essential definition of the genus, whereas in a philosophical
concept it is so included. In short, it turns out to greatly
resemble the analytic-synthetic distinction. Is Collingwood
distinguishing between a philosophical and a non-philosophical
scale of forms on the grounds that judgments stating such a
scale are either analytical (if philosophical) or synthetic (if
non-philosophical)?
(3) Once more we are up against the problem of Col-
lingwood's use of the term "identity," since we do not know what
"identified" means in
475
the expression "the variable is identified with the generic
essence."14 We may suppose that he minimally means that in
philosophical genera defined by a scale of forms there are real
and necessary connections between the variable attribute and the
genus which is being specified in differing degrees (EPM,
100).15 Whether this means that philosophical assertions are
analytic (in the sense of the tradition deriving from
Leibniz--viz. that the predicate is "contained in" the subject)
is not ever directly stated in the Essay itself. But based on
the discussion
_______________________
14 Cf. Immanuel Kant, Critique of Pure Reason, trans. by
Norman Kemp Smith (London, 1929), p. 48: "Either the predicate B
belongs to the subject A, as something which is (covertly)
contained in this concept A; or B lies outside the concept A,
although it does indeed stand in connection with it. In the one
case I entitle the judgement analytic, in the other synthetic.
Analytic judgements (affirmative) are therefore those in which
the connection of the predicate with the subject is thought
through identity; those in which this connection is thought
without identity should be entitled synthetic."
15 By way of explanation of what Collingwood means by

the "differentiation" involved in the specification of a generic
concept in a scale of forms, Mink writes: "Two things may be
differentiated as distinct from each other or as opposite to
each other. Opposition ‘is a relation subsisting between a
positive term and its own mere negation or absence’ (EPM, 75);
distinction is, apparently, any relation of difference between
two positive terms which differ in meaning" (MHD, 67). On this
interpretation one might also say that "difference in kind" is
also a relation between two positive and different attributes,
and "difference in degree" is a relation between a positive term
or attribute and its relative negation or absence. Unfortunately
this interpretation is weakened by the qualifying preface to the
remark that Mink quotes, which reads: "In its non-philosophical
phase, opposition is a relation subsisting between a positive
term and its own mere negation or absence" (EPM, 75--emphasis
mine).
476
of philosophical judgments as categorical-universal and the con-
troversy this generated with Gilbert Ryle,16 we may infer that
he does not mean us to understand the scale of forms in a
strictly analytical sense. To these two pieces of evidence we
must now turn.
(a) Collingwood insists that philosophical judgments are
universal, but also that they are not therefore merely
hypothetical l7--they are also essentially categorical, i.e. the
subject of the
___________________________
16 There are three unpublished letters between Colling-
wood and Gilbert Ryle, presently in the Bodleian Library, Ox-
ford. CRC-I is dated 4 May, 1935, and is from Collingwood to
Ryle; CRC-II is from Ryle to Collingwood, dated 21 May 1935; and
CRC-III is from Collingwood to Ryle, dated 6 June 1935. All are
related to the article written by Ryle, entitled "Mr.
Collingwood and the Ontological Argument," Mind, Vol. XLIV, no.
174 (April, 1935), pp. 137-51--basically a discussion of Chapter
VI of the Essay on Philosophical Method. In CRC-I, 14,
Collingwood points out that EPM, 63 (where Collingwood takes up
an objection to the view that opposition and distinction are
fused in a scale of forms, so that the scale of forms explains
the overlap of classes) is directed against a view of Croce. In
The Idea of History Collingwood expands on this point. Croce
criticized-Hegel's philosophy of history for confusing
opposition and distinction, since opposites refer to concepts
and distinction to individuals. But Collingwood argues that both
are applicable to historical events insofar as events have an
inside consisting of thought, and thought is conceptual and
hence oppositional. Opposition is a dialectical term, as is
reconciliation or synthesis (IH, 119).
17 Cf. Bernard Bosanquet, Knowledge and Reality (London,

1892), pp. 1-58. Collingwood, like Bosanquet, took issue with
Bradley s use of the categorical-hypothetical distinction, which
Bosanquet summarized as follows: "The universal judgement, if
bona fide universal, and in sense singular or collective,
cannot, so Mr. Bradley maintains ((in The Principles of Logic)),
be categorical. A categorical judgement
477
assertions being made is no mere ens rationis but something
actually existing (EPM, 125). We shall have more to say about
the categorical aspect of philosophical judgments in a moment,
but we may point out that in stating it he is declaring his
departure from at least the narrower versions of the
analytic-synthetic distinction, since he indicates that the
universal statements of philosophy have objective reference and
are not merely tautologies that indicate only the equivalent use
of words.
(b) This is confirmed in the second and final of his
letters in his correspondence with Ryle concerning his logical
views expressed in the Essay. In it Collingwood states that the
most important difference between them is Ryle's disbelief in
synthetic a priori propositions, a disbelief which he says he
does not share, and is connected with the possibility of
metaphysics (CRC-III, 1, 7). In stating his dissent from Ryle's
insistence that all universal propositions are hypothetical, and
hence that "no
________________________
affirms the existence of its elements, and enunciates some mat-
ters, conveyed by an idea, as true directly of Reality. But a
universal or abstract judgement does not affirm the existence of
its elements, and may be true though none of them exist or are
even possible of Reality'' (p. 5). Bosanquet insists that some
judgements, like "Heat is a mode of motion," and "Gases have a
spectrum consisting of lines," have subjects that "are thought
generally, through abstract characteristics, and are not
individually known," but are nevertheless fact: "They are
universal fact, and to say this is impossible seems to me a flat
denial of the commonest experience" (pp. 14-15). Cf. CRC-I, 13.
478
universal propositions are categorical" is a tautology (CRC-II,
5), Collingwood replies that tautologies are always merely
verbal. On his own view of logic, universal propositions deal
with real propositions and not just verbal ones--that is, they
deal not merely with affirmations and negations about words, but
about things (CRC-II, 5-7).
In his first letter to Ryle, Collingwood located the
problem at a different level:
It seems probable to me that the fundamental point at issue

between us is concerned with the way in which we answer the
question "what is a universal?" . . . . It looks to me as
if, in your logic, this question was answered by saying "a
universal is a class"; i.e. that whenever we make a
statement (assent to a proposition) about "all x" we are
really making n statements about the n instances of x which
exist. The theory of universals is thus, so to speak,
resolved into the theory of classes. It seems to me that
this analysis . . . (following Russell) . . . would
represent a line of thought more or less identical with
logical nominalism. In my own view, this line of thought is
so far from satisfactory that it inverts the necessary
order of analysis and is thus a case of obscurum per
obscurius. I am disposed to think that what makes a number
of things instances of a class is their common possession
of some common nature, and that this common nature (the
so-called "universal") is thus the ratio essendi of the
class as such. Instead of resolving the theory of
universals into the theory of classes, I should therefore
be inclined to the opposite line, of resolving the theory
of classes into the theory of universals. This is of course
akin to logical realism. (CRC-I, 28).
In the conclusion of the letter Collingwood summarizes the
matter in the statement that "the question which most
fundamentally seems to divide us appears to me to be question:
Is a universal simply a class, or is it that

479
which makes a class a class?--where you take the first
alternative and I the second" (CRC-I, 29).
It does not take much imagination to see the connection
between what Collingwood is here calling simply a "universal"
and what he had called a "concrete universal" in Speculum
Mentis--a similarity which is all the more striking for the
absence of that small qualifier, "concrete." It will be recalled
from Chapters V and VII (above, pp. 5-3, 5-16, 7-22, 7-28, 7-36,
7-39 and 7-47) that in Speculum Mentis Collingwood (a)
contrasted the "concrete universal," or one to which the
differences between its particulars are relevant," to the
"abstract universal," which is "indifferent to the variation of
its own particulars" or "to their own exemplification in this
particular or that, and differentiated only in their spe-
cifications" (SM, 162-63). He also (b) contrasted the structure
of a classificatory system based on abstract classes with the
structure of a dialectical series of progressively inclusive
("overlapping") universals (SM, 55, 162-64, 206-07); and (c) he
contrasted the formal logic built on the notion of the abstract
universal, or class-concept, with the dialectical logic which
employs the concrete universal (SM, 49, 195, 279). And finally,
(d) he argued that a dialectical logic which employs the
concrete universal "destroyed any distinction between a logic of
opposition and a logic of difference" (SM, 244).

480
It is clear that the Essay continues and develops this
line of thinking first sketched out in Speculum Mentis: (a) is
another way of stating the distinction between the philosophical
and the non-philosophical concept; (b) is transparently a
description of a "scale of forms;" (c) is the contrast he is
drawing between his own logic and that of Ryle, in his corres-
pondence with Ryle, his own being built on the notion of a
universal concept as described in the Essay; and (d) is the
"fusion" of "relations of opposition and distinction" and
"differences of degree and of kind" as described in the Essay
concerning the scale of forms. The logic described in the Essay
is therefore the dialectical logic discussed in Speculum Mentis.
If this is so there is one interesting corollary on the
Essay's thesis that the "variable is identified with the generic
essence"--about which we may speculate. Does Collingwood have in
mind the fact that the F-logic built on the notion of a class,
the substitution of one or another instance for a "variable"
takes for granted that no difference occurs, as a result of the
substitution, to the formula in which the variable occurs? And
if so, would this situation not be drastically altered where the
substitution-instances are specifications of a philosophical
concept having the structure of a "scale of forms" as he
describes it--i.e. one in which on instance differs from another
not only in degree but also in

481
kind, and has relations with it not only of distinction but of
opposition as well? We should like to think that this is the
case, so that when Collingwood speaks of an abstract universal
as being "indifferent to the variation of its particulars" we
may understand him to mean that when an abstract term is being
used as a simple class concept the relationship it has with its
instances is a relationship of a variable with its replacement
instances. But when a concept is so structured that it has a
layering of overlapping meanings, such a relationship of
sustitutivity or replacement is not possible without doing
violence to its structure of meaning, or without reducing it to
being an abstract class-concept.
Remaining alert to this possibility, but also aware that
it is not explicitly warranted by any texts of the Essay, we
turn to his discussion of judgments--the point of contact
between D-logic and Q-A logic.
(4) As entries 4, 5, and 6 of Table 10 indicate, Col-
lingwood maintained that all philosophical judgments are uni-
versal, categorical, and affirm or deny concretely. Collingwood
nowhere insists that this characterization of the philosophical
judgment is complete, but he does argue that it is minimally
necessary. All three of these requirements pose difficulties for
Q-A logic. The "principle of concrete affirmation and negation,"
for instance, applies to the class of expressions that in Q-A
logic are called propositions,

482
namely that which can be true or false. But what about absolute
presuppositions? Collingwood's formulation of the principle
appears to relate it to the Q-A complex:
The affirmative judgement in philosophy runs thus: S is P

and not Q; the negative thus: S is not Q but P; where P and
Q are equally definite and specific answers to the same
question: what is S? The peculiarity of the philosophical
judgement in respect of quality, then, lies in the peculiar
intimacy of the relation between its affirmative and
negative elements, which is of such a kind that P cannot be
validly affirmed while Q is left indeterminate, nor Q
validly denied while P is left indeterminate. (EPM, 110-11).
But this leads to the curious result that absolute
presuppositions, which are neither true nor false because not
answers to questions, are not philosophical because they do not
affirm or deny; either that or there are some philosophical
propositions (namely absolute presuppositions) which neither
affirm nor deny (or perhaps affirm without denying or deny
without affirming)--which violates the principle of concrete
affirmation and negation.
What we do not know is whether Collingwood would agree
that whatever judgments affirm or deny must be also true or
false, or whether there is a class of judgments which can affirm
or deny, and do so concretely, without being either true or
false. It is clear that from the point of view of Q-A logic, the
latter case is clearly possible. It is also clear that D-logic
makes affirmation and negation essential to judgment as such:
Collingwood writes in the Essay that "Any judgement

483
predicates a concept, and whenever we affirm one specific
concept we deny the other specifications of the same genus"
(EPM, 107). But what of truth or falsity in D-logic?
(5) Truth and falsity in D-logic appear to have some-
thing to do with the universal-categorical aspect of phil-
osophical judgments. We shall take up the universality aspect in
this section and the categorical aspect in the next. According
to the Essay just as philosophical concepts are
characteristically universal, so also are philosophical judg-
ments, but Collingwood seems to assume that his readers will
understand what he means by this. (cf. EPM, 111). From his
example of a universal judgment ("all men are mortal") we may
assume that he means by it a judgment of the form "All S is P."
or one which affirms or denies a predicate concept of a
universal subject. But after having told us that the relation
between concepts in philosophy is an intentional "overlap"
rather than an extensional one, he describes the universal
judgments of philosophy in extensional terms:
The species universal, particular, and singular overlap;

the universal judgement that all men are mortal does not
exclude, it includes the particular judgement that some men
are mortal and the singular judgement that this individual
man Socrates is mortal. These three elements introduce
differentiations into its significance, even considered as
a universal judgement: as a pure universal, it means that
man as such is mortal; as a universal of particulars, it
means that every kind of man is mortal; as a universal of
singulars it means that every individual man is mortal.
These are not so much three kinds of universal judgement as
three elements present in every universal judgement whether
in philosophy or anywhere else. (EPM, 111).
484
In the ensuing discussion Collingwood tries to dis-
tinguish the universality of philosophical judgments from that
of others, but in the process appears to badly conflate what is
here given in extensional terms with (a) what is emphasized in a
judgment, (b) the sequence in which judgments are made, and (c)
the priority of meaning of judgments. Thus he says that (a) the
type of judgment in which the "determining element" is the
singular (in what he calls induction by simple enumeration) is
such that "each individual instance of S is found on examination
to be P and is called a generalization;" but (b) a second type
"begins not from the singular but from the particular" and "goes
on to judge" that since each particular kind of S is P. S as
such is P (which he calls an empirical generalization); and (c)
in the third type the universal element is taken as primary: "we
begin by thinking that S as such is P. and this is seen to
involve the particular, that any specific kind of S is P. and
the singular, that each instance of S is P" (called the
universal judgments of exact science or mathematics) (EPM,
111-12; cf. EPH, 135-36). In these cases the manner in which the
universal judgment is reached, or its sequence in the way it is
advanced, is taken as what determines the type or meaning of
universal judgment it is, and the type of universal judgment
indicates what is being emphasized in that judgment.

485
The compression of these three "types" into one judg-
ment--the universal judgment of philosophy--is evidently not a
source of embarrassment for Collingwood, and the best indication
of this is in his response to Ryle on the matter. In the first
letter of their correspondence he states that in his view logic
is concerned with "real thinking" which "contains within itself
every kind of proposition, simultaneously thought together and
having a logically necessary structure” -- these forming for any
universal judgment its logical context, which consists of
"things which if we didn't think we couldn't think what ex
hypothesi we are thinking" (CRC-I, 16-19).
Further: I believe that this logical context must have a

certain logical structure. I believe, for example, that
there are always in it affirmative and negative elements, .
. . categorical and hypothetical elements . . . ((and))
propositional and inferential elements. I believe that in
any example of real thinking all these elements are pre-
sent; although . . . they are certainly not all ordinarily
(or perhaps ever) expressed in words . . . . When I say
"all," I am making a big assumption: I am assuming that
there is a certain complex of element-types which forms a
whole, and that every example of real thinking contains an
example of every element-type contained in this whole.
(CRC-I, 17-18).
Collingwood confesses that he cannot give a complete account of
the whole of element-types that he is discussing, but he admits
that what he has in mind is (a) "the old-fashioned formal logic"
which holds that any proposition whatever must have a certain
quality, quantity, relation, and modality--any complete whole
consisting of these four element-types with its respective
alternative variants; and (b) he says that this

486
"old-fashioned formal logic" was modified by Kant who disting-
uished three element-types within each of the four character-
istics of a proposition, bringing the number of element-types to
twelve. Without subscribing to this as a final number,
Collingwood indicates his assent to the existence of a "good
many" constant element-types found in any example of real
thinking.
He then concludes this remarkable synopsis of his
beliefs concerning logic with the following sketch:
I think that what . . . I have called kinds of proposi-

tions, may be one of two things. They may be what I am here
calling element-types, and in that case any example of real
thinking will contain within itself an example of every
kind of proposition. Or they may be . . . varieties of
propositions which are not element-types, but are merely
variants or alternative forms of this or that one
element-type. In the former case, every kind of proposition
(or inference for that matter) which the logician studies
must be an element-type of which an instance is present in
the real thinking done by himself as a logician. In the
latter case, of course, this need not happen. Lastly I
believe that the main task of logical theory is to
ascertain, so far as one can, what I have called the
logical structure of real thinking and the element-types
involved in that structure. I expect that you would
entirely deny this, and maintain that when the logician
speaks about kinds of propositions he means not
element-types but alternative varieties. I will not go on
to argue that point; I will only observe that logical
atomism, although I recognize the very important work which
it can do in the analysis of propositions taken singly,
begins by begging the question which I am here declining to
argue, in assuming that when we simultaneously assent to a
number of propositions . . . we are not thinking
simultaneously a complex of thoughts which must be thought
together if it is to be thought at all, i.e. a complex
having a logically necessary structure which is itself an
important subject for logical study, but merely an
assemblage of thoughts each of which presents to the
logician only the problems arising out of itself taken by
itself. (CRC-I, 18-20).
487
This very important letter is the only place, so far as
I know, that we find Collingwood, as he says, "giving myself
away rather completely" on the subject of his view of logic. We
shall therefore be forced to rely on it heavily for an
understanding of what Collingwood is attempting to say in the
Essay.
But in the context of our present discussion we notice
especially that it is not accidental when we find Collingwood
engaged in a compression of several logical functions into one
act, since "real thinking" (which must be what is being
expressed in a philosophical judgment) "contains within itself"
all the "element-types" on which logical judgment is based; and
further that this is a complex of thoughts which "must be
thought together if it is to be thought at all, i.e. a complex
having a logically necessary structure." But we shall have to be
on our guard against accepting any manifestation of "real
thinking" as a Deus ex machina explanatory device to resolve all
logical discrepancies we may encounter. "Real thinking,"
whatever else it may be, has a "logically necessary structure,"
so that wherever this structure fails to appear we may assume
that real thinking has not successfully occurred.18
____________________
18 Collingwood's first letter to Ryle also confirms from
a different direction our earlier provisional characterization
of the Kantian standpoint of the Essay: his own statement to
Ryle that the "element-types" of "real thinking" correspond
roughly to the Kantian schematized categories gives
488
(6) We are still on the trail of what Collingwood's
view in the Essay, and hence in D-logic, of truth and falsity
is, and we come now to the chapter on philosophical judgments as
categorical--perhaps the most controversial chapter in the
entire Essay, primarily because of the use of the ontological
argument and its extension to the domains of logic and ethics.
As we have already noted, it was this chapter that set off the
exchange between Collingwood and Ryle.
Once again under the rubric of "categorical"
Collingwood seems to be compressing into a single notion a
number of functions which are logically distinct,l9 and in
Ryle's criticism of this chapter it is this compressive
indistinctness that causes the greatest distress. Both in his
review article and in his correspondence with Collingwood,
________________________
us grounds for saying that what Collingwood is attempting in the
Essay is a brief survey of some salient features of what Kant
would call "transcendental logic" (although there is not much to
be gained by overemphasizing this point).
19 At various points in Chapter VI of the Essay Col-

lingwood says about categorical philosophical judgments (a) that
the subject of categorical judgments is something actually
existing (EPM, 117); (b) that the subject of such judgments
cannot be conceived except as actual (or its essence implies
existence) (EPM, 131, 133); (c) that the subject of such
judgments actually provides an instance of itself (EPM, 130);
(d) that in such judgments we declare ourselves committed to
believe that (a) ~PM, 127); and (e) that anything describable by
such a judgment is obliged to produce an instance of itself
(EPM, 130).
489
Ryle accuses Collingwood of failing to distinguish between
propositions which assert concrete matters of fact (particular
or existence propositions, which Ryle holds are the same--all
existence propositions being particular propositions and
vice-versa) and universal propositions which assert an abstract
relationship between properties, but do not say anything about
something which exists (CRC-II, 2, 4, 5; cf. MFA, 250-51). And
indeed, Collingwood's formulation of the rule for categorical -
universal judgments in philosophy seems to tee designed to cut
directly across this distinction. In discussing the Anselmian
ontological argument as an instance of the rule Collingwood
writes:
Divesting his argument of all specially religious or the-

ological colouring, one might state it by saying that
thought, when it follows its own bent most completely and
sets itself the task of thinking out the idea of an object
that shall completely satisfy the demands of reason, may
appear to be constructing a mere ens rationis, but in fact
is never devoid of objective or ontological reference . . .
. (I)n effect his argument amounts to this, that in the
special case of metaphysical thinking the distinction
between conceiving something and thinking it to exist is a
distinction without a difference . . . . Reflection on the
history of the Ontological Proof thus offers us a view of
philosophy as a form of thought in which essence and
existence, however clearly distinguished, are conceived as
inseparable. On this view, unlike mathematics or empirical
science, philosophy stands committed to maintaining that
its subject matter is no mere hypothesis, but something
actually existing. (EPM, 124-25, 127).
In his response to Collingwood's Essay Ryle denied that
the ontological argument proves anything because it is formally
fallacious. Hence it does not prove that there is anything about
which it is true
490
to say that its essence implies its existence, because a
statement of essence is a universal proposition and hence
hypothetical. From a hypothetical statement alone or in
conjunction with others no concrete matter of fact or existence
proposition can be inferred. In order to infer the existence of
something, Ryle argues, one needs at least one genuine singular
proposition, i.e. one which either (a) embodies at least one
logically proper name, or (b) has at least one definite
description which in fact describes something. Unless a
universal proposition rests on or contains a logically genuine
singular proposition, it does not successfully refer. (CRC-II,
2, 5).
In his reply to Ryle Collingwood counterattacked by
denying the distinction on which Ryle based his disproof. It is,
he says, fallacious: not all universal propositions are
hypothetical, and it is not only singular and particular pro-
positions that can assert existence:
You seem to me . . . to be arguing on the assumption . .

that any general proposition must belong to either one or
the other of two classes . . . "any-propositions" and
"every-propositions." I wish to maintain that there are
also what I will call "all-propositions." (a)
any-propositions. These are what you call general
hypotheticals, which do not depend for their truth on the
existence . . . of any of the things to which they
apply . . . . (Y)ou accept my account of these and my
description of arithmetical and geometrical propositions as
belonging to this type. (b) every-propositions. These I
think you are assuming . . . as a kind, and the only
possible kind, of general categorical. They apply to every
instance of a certain class, and depend for their truth
upon the existence . . . of these instances
491
. . . (Y)ou believe . . .that the only way in which a

general proposition can be categorical is by being
enumerative. With this logical doctrine is bound up a
metaphysical doctrine to which . . . you also subscribe:
viz. that what exists is an assemblage, or various
assemblages, of "particular matters of fact." . . . But in
fact I am quite consciously, in my essay, attacking these
assumptions. I will admit for the sake of argument that
there are any-propositions and every-propositions; but I
contend that there are also all-propositions or (as some
logicians have called them) "true universals" having a
categorical character, i.e. they are not enumeratives, and
yet they are not indifferent to the existence of the things
to which they apply, but are of such a kind that their
truth depends on that existence. I regard such propositions
as especially characteristic of philosophy. (CRC-I, 24-25).
This is as close as we shall ever come to understanding what the
doctrine on truth is in the Essay, and since this issue (at
least on Collingwood's view of it--cf. CRC-I, 7) crosses over
the boundary of logic and enters that of metaphysics, we must
postpone further discussion of this until Chapter X.
But we must note here that in order to defend his case,
Collingwood must produce at least one acceptable example of what
is to count for a valid categorical universal judgment, or a
synthetic a priori proposition. As we shall see in Chapter X, in
the Essay on Metaphysics Collingwood provides us with several
candidates for such judgments--"God exists" being one of them,
but others are offered in his discussion of the transcendental
analytic section of Kant's first Critique. In the present
instance Collingwood sticks to existential propositions, as in
his reply to Ryle:

492
I believe that such propositions as "God exists," "mind

exists," "matter exists," and their contradictories, do not
assert or deny particular matters of fact; nor do I believe
that they assert or deny anything which can be adequately
described as collections or classes of matters of fact. To
assert or deny propositions of this kind, with reasons
given for the assertion or denial, seems to me the business
of constructive or destructive metaphysics . . . . I do
think that a philosophical proposition may be e.g. about
thought or matter; and . . . if I learnt to use your
language, I could call thought or matter a designated
entity; but I could never allow that it was either a
particular matter of fact or a mere collection of
particular matters of fact. (CRC-I, 4, 6).
But the most peculiar and idiosyncratic instance that
Collingwood gives of the rule that philosophical judgments are
never devoid of objective or ontological reference is not the
Ontological Proof of Anselm, but his discussion of logic. He
paraphrases this view in his letter to Ryle:
Logic not only discusses, it also contains reasoning:

consequently, whenever a logician argues a point in the
theory of inference, he is producing an instance of the
thing under discussion; and, since he cannot discuss with-
out arguing, he cannot discuss any point in the theory of
inference without doing so. Consequently, in so far as it
necessarily contains reasoning, the theory of reasoning
((i.e. logic)) cannot be indifferent to the existence of
its own subject-matter; in other words, the propositions
which constitute the body of that part of logic cannot be
in substance hypothetical. For example, if a logician could
believe that no valid reasoning anywhere existed, he would
merely be disbelieving his own logical theory. (CRC-I,
11-12; cf. EPM, 130).
The logician, to paraphrase a paraphrase, not only mentions
logic he also uses it in the construction of his logical system.
But both this example

493
and the example he gives from ethics20 seem to illustrate a
different point than the one he was making about the Ontological
Proof of Anselm. Concerning logic and ethics he seems to be
saying that such concepts are self-instantiating--logic being
logical and ethics being ethical; and that the denial of such
self-instantiating concepts leads to a different sort of
absurdity--what might be called categorical nonsense (or perhaps
what Austin has called "performative absurdity"--cf. Rubinoff,
CRM, 184). Thus if one were to set out to formally demonstrate
that there is no such thing as logical demonstration, or if one
were to argue that one ought not to say what one should or
should not do, he would be engaging in a categorical (or
performative) absurdity: what he says he is doing is
incompatible with what he is doing. But certainly the
Ontological Proof does not produce or provide an instance of
itself. Although it presumes a commitment to the definition of
God as a being than which none greater can be conceived, if one
were to deny that "a being none greater than which can be
conceived" has ontological reference he would not be doing
_____________________
20 Collingwood argues that ethics cannot be merely de-
scriptive nor merely normative any more than can logic. Ethics
describes not action as opposed to ideas of action, but the
moral consciousness; and this it is forced to describe as al-
ready being in some sense what it ought to be. Conversely "this
in turn will affect the account which it gives of action; for no
theory of moral ideals is conceivable which does not admit that
to some extent moral ideas affect action" (EPM, 132). Ethics, in
short, must be ethical.
494
something incompatible with what he is saying he is doing. The
Ontological Proof does not seem to meet the description he is
offering for categorical judgments as self-instantiating, and at
this point we do not have a clear idea of what Collingwood
understands to be the uniting concept for these two sets of
examples.
A candidate for a common concept might be "having on-
tological reference." But then what sort of "ontological re-
ference" does logic have? If logic has ontological reference in
the sense that it occurs as a thought-process in someone's mind
who is thinking logically, then it is no less true that
mathematics and empirical science have a similar reference. But
the entire point of using the Ontological Proof as an example
seems to indicate that the "ontological reference" that he has
in mind is not merely to "second intentions" or to mental
events: it is not, he says, a mere ens rationis but something
actually existing. We therefore seem to be left with the
suggestion that there are some judgments which necessarily imply
the existence of what they describe. But then this narrows the
field to the one special case being discussed in Anselm's
argument, as Collingwood himself admits: "What it does prove is
that essence involves existence, not always, but in one special
case, the case of God in the metaphysical sense: the Deus sive
nature of Spinoza, the Good of Plato, the Being of Aristotle:
the object of metaphysical thought: (EPM, 127).

495
It is true that Collingwood attempts to extend this "object" to
philosophical thought in general" by adding that
"metaphysics . . . is not unique in its objective reference or
in its logical structure; all philosophical thought is of the
same kind, and . . . partakes of the nature of metaphysics,
which is not a separate philosophical science but a special
study of the existential aspect of that same subject-matter
whose aspect of truth is studied by logic, and its aspect as
goodness by ethics" (EPM, 127). But the extension fails to
convince the reader, for there is no good reason provided for
dropping the limitation of the "essence involves existence"
maxim to that "one special case" being discussed in the
ontological proof.
But here we once again encounter Collingwood's advance
warning that he will not pursue the consequences of his Essay
into the realm of metaphysics (much as we are tempted to admit
that he has already done so), and we must leave the matter in
this unsatisfactory state (as he himself does) until we have an
opportunity to examine his explicit views on metaphysics in
greater detail.21
_______________________
21 The reader is left at this point with the very uneasy
suspicion that Collingwood has stated a very important
philosophical truth, but expressed it very badly, and left it in
a highly ambiguous state. Everyone knows that philosophers have
claimed to be stating truths that have ontological reference,
and further that they have employed arguments to reinforce these
truths. What is in question (at least since the time of Hume) is
whether there is any necessary validity to that claim, and
whether their arguments succeed in demonstrating what their
authors believe they do. Collingwood seems to by saying that the
real
496
(7) If the chapter on categorical thinking is the most
controversial one in the Essay, the chapter on philosophical
inference22 is the most axial: it is the chapter that more than
any other displays what is unique in a philosophical methodology
that ideally unifies the specific topics in philosophy; and it
is the chapter in which Collingwood completes the discussion of
the work of his predecessors on the subject.23 In Collingwood's
view the development of the idea of philosophical
________________________
tradition in philosophy has always been committed to maintaining
this claim to ontological reference, and that to abandon it is
to abandon philosophy. But to adequately argue that the
ontological reference claims of metaphysicians is successful he
would have to rely on something more than the evidence provided
by the ontological argument. What is required is a fully
developed theory of meaning and reference. In the sequel we
shall see that he did make a start on such a theory in his
discussion of language and mental acts.
22 Part of the material of Chapter VIII of the Essay was

presented in our own Chapter VII, Section 2. On Collingwood's
view of inference, cf. IH, 253-56.
23 In the first chapter Collingwood reviewed the stages

of development of the notion of philosophical method, which we
saw falling into four major phases: (a) the Socratic phase, in
which the function of dialectical questioning was vigorously
employed to make implicit thought explicit in definitions; (b)
the Platonic phase, in which dialectical method is expanded to
include arguments which directionally proceed from
quasi-mathematical, hypothetical definitions to non-hypothetical
first principles of thought, primarily by removal of
hypothetical restrictions as the argument proceeds; (c) the
Cartesian phase, in which mathematics is taken as an explicit
model for philosophical method (but continually violated in
actual practice by Descartes and his followers by employing
arguments that are not strictly speaking deductive and do not
employ first principles which are self-evident); and finally,
(4) the Kantian phase, in which philosophy is freed from math-
ematical methodology by recognizing
497
methodology culminates in the demand for a self-justifying kind
of thinking. Justification in a philosophical context requires
argument, argument means inference, and inference suggests
inductive and deductive reasoning. Since Collingwood has all
along been comparing philosophical method with the method of
exact and empirical sciences, it is not surprising that the
culmination of this comparison occurs in a chapter which deals
with inference. If philosophy is to escape scepticism and
dogmatism it must show how its arguments can be self-justifying
without
_____________________
that in philosophy there are no definitions, no axioms, and no
demonstrations of a sort essentially mathematical (EPM, 4-6,
10-25, 155-56). The first chapter ends with a critical
discussion of Kantian methodology, which, Collingwood says,
fails to adequately reconcile the conflicting claims of the
critical method (a) as a propaedeutic to philosophy proper (i.e.
to metaphysics) and (b) as philosophy itself (EPM, 20-24). In
Chapter VII ("Two Sceptical Positions"--a chapter which, like
Chapter X, appears to be an aside but is not) Collingwood
resumes his discussion of critical philosophy which he takes to
be one of two related sceptical positions (the other being
analytic philosophy) which dominate the contemporary
philosophical scene. In essence his reply to the attempt to
reduce philosophy to the function of criticism is that it
assumes positive standards of consistency from which it finds
its subject diverging, and unless it undertakes to defend these
standards it fails to justify itself and assumes a dogmatic
stance with respect to its subject-matter (EPM, 140-41).
Similarly the "analytic view" (of which he finds Moore and
Russell representative) cannot exempt the positive principles it
assumes as true from common sense and/or science from its
analysis without self-contradiction (EPM, 143-46; cf. 138-39,
142 n. 1). The upshot of this discussion of critical and
analytic philosophy is that if philosophy is to avoid both
scepticism (the result of an unbridled exercise of destructive
criticism) and dogmatism (the result of assuming certain
principles to be true without justification), it must present
positive grounds for its own activity; that is, it must be
self-justifying.
498
being circular. We might rightfully expect this to be the most
crucial chapter in the Essay.
Noting that (a) "in its demand for close and cogent
reasoning, philosophy resembles exact science," i.e. "each alike
works on the principle that no conclusions may be asserted for
which valid and sufficient reason cannot be given" (EPM, 154);
but also that (b) it resembles empirical science insofar as "it
is supported throughout its texture by crossreferences to
experience (EPM, 164), Collingwood raises the question whether
and how philosophical inference can be deductive (like exact
science) and/or inductive (like empirical science--EPM, 151; cf.
IN, 254-55). For the purpose of distinguishing philosophical
inference from both deduction and induction, Collingwood
analyzes inference into three components: the data from which
the argument proceeds, the principles according to which
inference takes place, and the conclusions to which the argument
leads (EPM, 151). It is on these three that he compares the
inferences of mathematics, empirical science, and philosophy.
In exact science the data are suppositions, the prin-
ciples are axioms, and the conclusions are inferred in the sense
of being demonstrated or shown to follow with perfect logical
rigor from the data according to the principles. The axioms are
of two sorts: those of logic, which do not belong to the body of
mathematics proper and are properly

499
speaking not suppositions but presuppositions "in the sense that
unless they were true the science could not take a single step
in advance;" and special axioms, which are part of exact sci-
ence, but are "self-evident" or indemonstrable, and known to be
true by a kind of intuition (EPM, 151-52).24 The arguments of
exact science are therefore irreversible--that is, "the
conclusions are logically dependent on the axioms; there is no
reciprocal dependence of the axioms on the conclusions" (EPM,
153).
Empirical science, on the other hand, relies on a sort
of inference known as inductive reasoning, in which individual
facts empirically known by perception, or the historical record
of perceptions in the past, are the data from which conclusions
in the form of universal
____________________
24 As is clear from the example he uses, Collingwood has
Euclidean geometry in mind as a model for exact science. But he
covers himself from attack on the grounds that no contemporary
geometrician would accept "self-evident" as descriptive of
geometrical axioms. Collingwood writes that "the main lines of
this view are not, for our purposes, affected if it is
maintained that the special axioms are not known to be true, but
only assumed . . . ((because)) in that case we shall have to say
that the entire body of the science consists of assumptions, but
that these fall into two classes: primary or fundamental
assumptions, the so-called special axioms, and secondary or
derivative assumptions, the so-called conclusions." In either
case "the logical axioms cannot be merely assumed, for . . . we
cannot think as if the principles of thinking were true, for if
they were not we should not be thinking" (EPM, 153). The
contemporary geometer would have to agree that some sort of
deductive inference-structure is presupposed, and that this is
not a part of the content ("body" as Collingwood puts it) of
geometry itself.
500
speaking to be derived (EPM, 165). The principles of induction
are, like exact science, of two sorts: logical principles pre-
sumed as certain, and assumptions with the degree of certitude
only that they are known not to be untrue, and that it is
expedient to use them--a certitude which increases in pro-
bability as the inductive inquiry proceeds, but never reverts to
deductive certitude (EPM, 165-67). Comparing deductive and
inductive inference, Collingwood concludes:
The logical movement of inductive thought is therefore

irreversible in the same sense as that of exact science.
The principles on which induction rests receive in return
no support from the inductive process itself . . . . The
process of thought in exact science, though irreversible as
regards its principles, may be reversible as regards its
data . . . . In this respect, inductive argument is not
reversible; for its data are what they are because they
enjoy the status of facts vouched for by perception; and
although we can infer the existence of an unobserved fact
from reasons inductively established by the study of
similar facts, we only infer it (where to infer, as always
in the context of inductive thought, means to establish as
probable) and do not perceive it. (EPM, 167).
Philosophical inference differs from both induction and
deduction which as species of the philosophical genus,
inference, overlap. (a) In philosophical inference there is no
division of axioms into those belonging to the science and those
belonging to logic, since logical principles are part of
philosophical thought itself. As a consequence philosophy cannot
neglect its own logical presuppositions (EPM, 154-55). (b) What
appears in philosophy as an axiom or indemonstrable proposition
serving as a starting point is, as philosophical argument
proceeds,
501
justified by what follows it. In this procedure philosophical
inference has the form of what Kant had called a "transcendental
deduction." Thus the Cartesian cogito is neither a self-evident
truth nor an assumption, but rather "in Kantian language the
principle cogito ergo sum is . . . transcendentally deduced,
that is, shown to be the condition on which experience as it
actually exists . . . is alone possible" (EPM, 156). (c) This
means that philosophy is obliged to justify its own starting
point. (EPM, 159). But (d):
This can be done only if the arguments of philosophy, in-

stead of having an irreversible direction from principles
to conclusions, have a reversible one, the principles es-
tablishing the conclusion, and the conclusions reciprocally
establishing the principles. But an argument of this
kind . . . is a vicious circle. The solution of the
dilemma lies . . . in the Socratic principle that
philosophical reasoning leads to no conclusions which we
did not in some sense know already . . . . Establishing a
proposition in philosophy, then, means not transferring it
from the class of things unknown to the class of things
known, but making it known in a different and better way.
(EPM, 160-61).
(e) Philosophical knowing in this way differs from that of exact
science inasmuch as the conclusions are in some sense known
without any proof at all: the arguments of philosophy exhibit as
a reasoned and ordered whole of interconnected knowledge what
was already in substance known before the work of philosophical
inference occurred at all:
If philosophy differs from exact science in this way—

the anticipation . . . of its conclusions by an
experience that
502
possesses them in substance before its reasoning be-

gins--other differences will follow, the chief being that
in philosophy the conclusions can be checked by comparing
them with these anticipations, and that by this checking
the principles at work in the reasoning can be verified. If
this is so the direction of the argument in respect of
principles and conclusions is reversible, each being es-
tablished by appeal to the other; but this is not a vicious
circle because the word "established" here means raised to
a higher grade of knowledge: what was a mere observation is
now not merely observed but understood; what was merely an
abstract principle is verified by appeal to facts. (EPM,
163).
(f) Philosophy differs from empirical science insofar as the
initial data of philosophy does not consist of individual facts
apprehended by perception, but of universal propositions appre-
hended in the experience of thinking; and the conclusions of
philosophical inference are not something new, but are the facts
themselves more thoroughly understood (EPM, 168-69). "(I)n
philosophy the knowledge . . . why things are so makes a
difference to the knowledge that they are so" (EPM, 169). (g)
The "experience" on which "conclusions" are philosophically
based, and by appeal to which they are checked, is only
"non-philosophical" or "pre-philosophical" in a relative sense:
These two phrases ((i.e. experience and conclusions)) are

names for any two successive stages in the scale of forms
of philosophical knowledge. What is called experience may
be any stage in this scale; in itself, as all human ex-
perience must be, permeated through and through by philo-
sophical elements; but relatively crude and irrational as
compared with the next stage above it, in which these phil-
osophical elements are more fully developed . . . . But
what is asked of the higher is not simply that it should
agree with the lower, but rather that it should explain
it . . . . The
503
accomplishment of this task is only the continuation of a

process already begun; it was only by thinking that we
reached the point at which we stand, for the experience
upon which we philosophize is already a rational experience
. . . . But the new and intenser thinking must be thinking
of a new kind; new principles are appearing in it, and
these give a criterion by which the principles involved in
the last step are superseded. (EPM, 172-73).
The ultimate response to the charge of vicious circularity in
philosophical reasoning is therefore that, since all experience
is already somewhat systematic and rational, whenever its sys-
tematic connections are exhibited, the experience is "estab-
lished" in a higher sense of the term. (In the language of The
Principles of Art, one might say that it is not merely a datum
but an interpreted datum.) The "reciprocal" establishment in
philosophical inference is therefore not truly reciprocal or
circular, but more like the coiled meanings in a scale of forms.
(h) But this response seems to get us out of the Charybdis of
vicious circularity only be dashing us against the Scylla of
vacuous philosophical inference: for did Collingwood not just
tell us that the distinguishing mark of philosophical inference
was that it is reversible as deductive and inductive inference
are not? Collingwood's reply is that philosophical inference is
always deductive and inductive, just as it is always critical
and analytical, but always with a difference. As deductive it is
a complete system based on principles and connected throughout
by strictly logical bonds; but as such its principles are open
to criticism on the
504
grounds that they must succeed in explaining our experience. As
inductive it seeks rational universality; but only by finding it
already present in the activity of philosophizing which is
already occurring. As critical it always seeks to refute theory;
but only in order to leave an experience standing which can be
interpreted in the light of the principles implied in the
critical process itself. And as analytical it begins with a
knowledge-datum and seeks to explain what this knowledge means;
but in so doing we come to know that datum in a different way,
and therefore as modified in the knowing of it. (EPM, 173-75).
There is much left unexplained in this account of
philosophical inference. For example we would like to know
precisely how philosophy employs the logical principles which
are a part of it, as stated in (a); and we would like to know
more about what the conditional possibility in a "transcendental
deduction" in (b) consists of, and in what sense necessity is
involved in such a deduction, and if this is the same as
"justification" in (c); and we would like to know how it is
possible to "know already" in (c) an experience which in (f) is
called universal and non-perceptual, and in (g) is called
rational; and most of all we would like to know what it means
for this experience, and indeed all experience, to be permeated
through and through with philosophical elements as stated in (g).

505
What appears to be presupposed in all this is a full--
blown epistemology that is only hinted at rather than explicitly
stated. Some of the outlines of this epistemology will appear in
the next and succeeding section of this chapter when we come to
examine the intellectual functions of mind and the analysis of
language, where we shall find some hints towards the resolution
of several of these problems (e.g. in the notion of a
pre-reflective act of meaning at the first level of
consciousness, and in the view that rationality it-
self--including rational inference--is an extension of con-
sciousness' demand for meaning, in this case what might be
called "rational meaning" at the level of reasoned discourse).
But before coming to this discussion and the implications for
Q-A logic, we have one more "key concept" from the Essay to deal
with.
(8) It is in Chapter IX that Collingwood comes closest
to evaluating the dialectical logic of the Essay in terms of the
criteria of consistency, completeness, and formality which we
discussed in connection with Q-A logic in the preceding section.
In this chapter Collingwood examines the claim of philosophy to
be systematic and constructive rather than piecemeal and
analytic--a claim that has been assaulted since the 19th century
on the grounds that a system of thought claims for itself
finality, completeness, objectivity, and unity, and philosophy
has none of
506
these. Our philosophical knowledge is always open to future
developments, is beyond the terminable survey of any individual,
is the expression of a personal and private point of view, and
effectively solves its problems not by mass generalities, but by
handling each problem individually (EPM, 176-78).
Collingwood's reply in essence is that even if one
concedes each of these points to the critic, the ideal of a
system remains undaunted. (a) While not absolutely final, all
knowledge in order to advance must take account of where it
presently stands, and therefore retains a synoptic or sum-
marizing element. In philosophy this record of progress is
intensified insofar as any new development does not merely add
to permanent and unaltered assets, but requires that previous
conclusions be re-examined in the light of new developments
(EPM, 179-81). And (b) while not absolutely complete, every
philosophy is a borrowing from past philosophies and a colla-
boration with present philosophical thought, so that there is
not and never has been any such thing as a private, personal,
self-contained system of philosophy (EPM, 181-82).25 And (c)
although each philosopher contributes only a part of a total
_______________________
25 This obviously does not answer the objection: some-
thing may be no less incomplete for having taken account of
previous and contemporary discussion of an issue, or even for
having taken into account all such discussion, since all pre-
vious discussion may have missed the essential point. What must
be shown is that all the relevant or pertinent issues are dealt
with satisfactorily.
507
system, and hence his work is not that system but only a portion
of it, in philosophy that contribution includes a theory of the
whole of what philosophy is, and thus anticipates its objective
conclusions (EPM, 182-84). And finally, (d) only an "anarchist
of the mind" would fail to see that even a methodical avoidance
of rigidity is itself a system or method, and although a
philosopher should avoid bondage to any readymade or rigid
formula, a willingness to revise one's principles in the light
of one's conclusions does not necessitate the rejection of all
principles or of all conclusions (EPM, 184-85).
But Collingwood feels that the real answer to all these
objections to systematic philosophy is made possible by
regarding the topics or constituents of philosophy as comprising
a scale of overlapping forms, wherein each topic differs from
the rest not only in kind but also in degree. The universal
subject-matter of philosophy differs intensively and
specifically whether the genus is divided into the species of
logic, metaphysics, epistemology, ethics, etc.; or in the
division of the subject into historical phases (ancient,
medieval, modern); or at the contemporaneous level in the
division between what may minimally pass for philosophy, what
retains its coherence in being grouped into one or more schools
of thought, and the final "common spirit" of the present age; or
finally in one's own philosophy, in the division between
unexamined
508
opinions, criticized and defined assertions held with greater
conviction, and a fully articulated and systematic whole of
philosophical judgments (EPM, 19498). In any of these
arrangements of philosophical topics a systematic ideal is
present as a scale of forms, which at each point sums up the
scale to that point, showing how subordinate positions are
opposed to and distinct from that point, and in error by
comparison to a higher point on the scale (EPM, 190). In a scale
of forms of philosophical knowledge any particular position is a
non-final summary, an incomplete termination, a subjective
necessity, and uniformly or methodically flexible (EPM, 191-92).
Each approximates to an ideal of a perfectly philosophical
subject matter treated by a perfectly philosophical method (EPM,
192).
Now although in this discussion of system in Chapter IX
of the Essay there is no clear correspondence with the criteria
for systematicity which we examined under the titles of
consistency, completeness, and formality, in connection with our
comparison of Q-A and F-logics, we can nevertheless see a rough
correspondence between the description of a system which
Collingwood here employs and those used to evaluate formal
axiomatic systems in F-logic. Thus the criterion of consistency
(that within the system propositions containing formal
contradictions are not provable) roughly corresponds to the
claim of unity in a philosophical

509
system; completeness (that all true statements within the system
are derivable from the axioms and rules of method within the
system) correspond roughly (but with considerable overlap) to
Collingwood's terms "finality" and "completeness"; and formality
(that the construction of well-formed statements within a system
is carried out by the strict application of its rules, and not
by the interpretation of specific meanings of the terms within
the statement--i.e. that there be a purely formal way to decide
which statements are well formed within the system) roughly
corresponds to what Collingwood is calling "objectivity."26
But the points of difference are more striking than the
points of superficial similarity. It is obvious from the
discussion of philosophical inference in Chapter VIII of the
Essay that, as Collingwood conceives it, D-logic is not simply a
deductive system, and therefore to apply axiomatic criteria for
evaluating it as a formal, consistent, and complete system would
not be altogether appropriate. A system in which principles are
revisable in the light of conclusions obviously transgresses the
requirements of an axiomatic system of F-logic, in which the
axioms or principles are stated in an object-language and the
rules in a meta-language, in which the rules and principles are
never revisable in
____________________
26 Cf. Bochenski, The Methods of Contemporary Thought,
510
the light of conclusions. For (a) if the meaning of terms within
the system can form scales of overlapping forms, it is difficult
to see how a criterion of strict consistency can be upheld,
since the prerequisite of semantic identity of terms is violated
every times such a term definable by a scale of forms is
introduced. And (b) if the highest term in a dialectical system
is summational or synoptic but not final, it is clear that a
higher term (a statement that first appears as a denial but
implies the positive principles on which the criticism is based)
can always be introduced which is not strictly inferable from
within the system (since its positive principles do not yet lie
within the system)--in short, the system is always incomplete.
And (c) if experience is the touchstone of a revision (or
reversal) of an argument from conclusions to principles, even if
this experience is that of a thinker, it is clear that the ideal
of a purely formal validity cannot be maintained.
This brings us back to the question of circularity,
which is all the more vicious in that it appears to escape ob-
jectivity altogether: for whatever "experience" is appealed to,
it will always be merely someone's property unless what it means
can be communicated in objective form. Collingwood's appeal to
experience in both the discussion of philosophical inference in
Chapter VIII and in his discussion of systematic philosophy in
Chapter IX casts a suspicion of radical subjectivity on the
whole
511
enterprise of the Essay. The issue raised by the entire argument
of the Essay is this: if philosophical arguments are reversible
on the basis of experience (and therefore its principles are
corrigible), how does one know when one has successfully
inferred in a philosophical sense? Collingwood wants to say that
a successful philosophical inference is one which increases our
understanding of experience at the same time it changes it (or
makes a difference to it). But if the criterion of success
appealed to here is the fact of "philosophical experience" or
"the experience of a thinker," then the argument of Chapter VIII
is truly and merely circular: we know that philosophical
inference is successful because we know we draw successful
philosophical inferences. But the appeal to experience is
inadequate as a criterion of success, since not all or just any
philosophical inference will do. Some are better than others,
some contradict others, some (as in Plato's dialogues) are meant
to be understood as logically fallacious.
Another way of stating the problem is that if the cri-
terion of successful philosophical inference is epistemological
(one form of knowledge explaining another), how are we to decide
what is good epistemology? Will any old theory of knowledge do?
What about a realist's epistemology? We have seen that
Collingwood argues that a realist epistemology is unsuccessful,
but he argues that it fails not because it

512
is not epistemology (which it is: it purports to explain what
occurs in knowing--that is, it makes knowing an object of know-
ledge), nor because it does not explain experience (which it
does: it puts forward a theory of how objects are known), but
because it fails to be good epistemology. It contradicts itself
by talking formal nonsense about unsensed sense, knowing what is
simultaneously declared to be unknown, etc. On the "experience"
of realist epistemology alone, on Collingwood's grounds, we
could not say that we have successfully inferred in a
philosophical sense.
But if we take seriously Collingwood's remark that when
we talk about "establishing principles" it is in a different and
higher sense of the term "establish" than when we talk about
"establishing experience," then we may have a way out of this
vicious circularity while retaining a legitimate sense of
philosophical inference--and this is the only hope that the
entire argument of the Essay has. Collingwood's commitment to
the assertion that philosophical concepts are arranged as a
scale of overlapping forms, also commits him to the conclusion
that the principles employed in philosophical argument are
increasing both in generality (taking in more of our experience)
and in intentional reference (stating more of what is in that
experience in terms of significance). This "higher knowledge" is
systematic knowledge as articulated and related in a system of
discourse bound by logical relations and (inasmuch as it is not

513
merely formal) categorically referential. If this "higher
knowledge" is the same thing as what he meant by "real thinking"
in his correspondence with Ryle, then it is possible that the
criterion of success in philosophical inference may be the same
as the criterion of a successful epistemology; it is the
articulation of our own knowledge as a system of discourse,
logically related and categorically referential. In the next
section we shall pursue this lead into Collingwood's philosophy
of mind and his analysis of language, and in so doing catch
sight of logic as an extension of meaning at the rational level.
But before proceeding with this analysis we must make
good a previous promise. We are now in as good a position as we
shall ever be to evaluate the standpoint of the Essay. We noted
in a previous chapter and also at the beginning of this section
that (a) the Essay claims both conditional and unconditional
validity for itself, and (b) that its treatment of pre- and
non-philosophical methods cannot be taken as literally true for
mathematics and empirical science. But within this present
section we have also seen that Collingwood (c) extended
philosophical concepts to include all experience (all human
experience is "permeated through and through by philosophical
elements" (EPM, 172); (d) declared that philosophical thought is
never devoid of ontological reference (EPM, 125); (e) criticized
analytical and critical philosophy for failing

514
to justify themselves by presenting the positive principles from
which their critical and analytical methods proceed (EPM, 145,
148-49); (f) insisted that philosophy is obliged to produce a
theory of itself and to be constructive and systematic (EPM, 1,
198); and finally, (g) stated that what he is doing in the Essay
is to discuss the nature of philosophy by discussing not only
what it is or has been, but also what it is trying to or ought
to be (EPM, 2, 4, 7).
From (e), (I), and (g) it follows that the Essay is
stating a theory of philosophy that is nonetheless itself
philosophy, and is stating what philosophy is and what it ought
to be by presenting its positive principles. From these and (d)
it follows that, ideally speaking, what is true of philosophy
must also be true to some extent for reality (i.e. philosophy
cannot be content with constructing a mere ens rationis or
formal system, but must have ontological reference). And from
these together with (c) it follows that (once more within
unspecified bounds) what is true for philosophy is true to some
extent for all experience.
Now the reason that we must hedge these conclusions with
qualifiers like "to some extent" and "ideally speaking" is that
we cannot simply apply the concepts of the Essay in a wholesale
manner either to experience in general or to reality. Obviously
not all concepts overlap in a scale of forms, nor is everything
in reality arranged in hierarchical fashion.

515
Furthermore if the classificatory ideal of exact and empirical
science is retained as a legitimate way to organize experience
(which is never denied in the Essay), then either one must deny
to it the title of true experience, which is then reserved for
philosophy alone, or else state that there are valid experiences
not subject to the characteristics of philosophical experience.
But if we approach the Essay from the perspective pro-
vided by Speculum Mentis we can say that Collingwood would
choose the first alternative: although it is quite possible to
adopt the scientific point of view on experience, such a
viewpoint cannot be ultimately and finally true (i.e. true un-
conditionally or absolutely), but only true in an abstract way.
If, as Speculum Mentis proposes, all knowledge is truly or-
ganized in a scale of forms, and if the doctrine of the Essay
concerning the relative standing of successive terms in this
scale as "experience" and "conclusion" be taken seriously, then
it follows that the relative position of empirical science and
philosophy are such that empirical science is "experience" to
philosophy's "conclusion." Furthermore if this conclusion is
already present as an experience which "anticipates" its
conclusion before reasoning about it begins, then empirical
science is an experience which anticipates the conclusions that
are first stated explicitly by philosophy. Philosophy "explains"
science only by incorporating its positive content, and
rejecting its negative

516
aspect (i.e. its abstractness). From the point of view of
Speculum Mentis, then, the closest that we can come to saying
what the Essay's evaluation of exact and empirical science is,
is that they are not what philosophy explicitly states itself to
be. Philosophy is self-justifying; exact and empirical sciences
are not. (Cf. Rubinoff, CRM, 26-27).
Unfortunately this gets us into difficulties concerning
our original statement that-the Essay utilizes an essentially
Kantian strategy--i.e. it assumes the fact of certain
experiences and goes on to ask on what necessary and sufficient
conditions such an experience is alone possible. We are left, as
we were in Chapter VII above, confronting something resembling
the "absolute knowledge" of an Hegelian sort rather than
confronting conditional schematized categories without which
experience would not be intelligible. The Essay leaves us in
unresolved puzzlement about this matter, except to the extent
that Collingwood makes it clear that he is not content with
leaving philosophical method where Kant had left it in the
Critique of Pure Reason: philosophy cannot be content with
viewing itself as merely critical--it must go further and state
what the positive grounds are from which criticism proceeds. But
the positive grounds that he provides, as we shall see in a
moment, are not a repetition of his conflicting remarks about
"absolute knowledge" in
517
Speculum Mentis, but a development of the theme of language that
we first discovered in germ in Speculum Mentis.
But where does all this leave us in our discussion of
Q-A logic? In the previous section of this chapter we found that
although Collingwood proposed Q-A logic as an alternative to
F-logic, on examination it turns out either to presuppose it or
to presuppose what F-logic presupposes; meaning, validity, and
truth appear to be establishable independently of Q-A or P-Q-A
complexes, and Q-A logic (either as Q-AA or Q-AM) fails to meet
criteria of consistency, completeness, or formality. In this
section we have been primarily concerned with discussing D-logic
and its relationship to F-logic, from which it differs in a
number of important ways (employing overlapping classes,
reversible inferences, etc.). Q-A logic has hardly entered into
the discussion. In the next section we shall see that rather
than being an oversight on Collingwood's part, it is an
indication that in his view these three logics are related
primarily by locating them in an epistemological context--i.e.
through the intellective, linguistic functions of mind.
As we noted above, the clue to the discovery of the
nature of philosophical thinking in the Essay is the overlap of
classes, and this is described in essentially semantic terms: it
is an overlap of intensional meaning, and meaning is a function
of mind at the linguistic level of

518
consciousness. The final step in our analysis of Collingwood's
views on logic must therefore take us back to the philosophy of
mind in The Principles of Art and The New Leviathan. After
surveying the functions of intellect and reason in these two
works, we shall attempt a final evaluation of the roles of the
three logics we have been discussing.
4. Language and Logic in The Principles of Art.
In our exploration of Collingwood's views on logic we
have noted not only a certain informality to his presentation of
both Q-A and D-logics, but also a tendency to describe these
logics in epistemological terms--concepts, judgments,
suppositions made by acts of choice, Q-A correlativity defined
in terms of persons, etc.--in short, terms which retain a
reference to mental activities as part of their essential
meaning. While any practitioner of contemporary formal logic
would find this to be archaic flaw, we find Collingwood in his
correspondence with Ryle insisting on the propriety of this
epistemological informality, and even suggesting (as in the
passage where he states that he understands the universal to be
that which makes a class a class) that there is somehow a
priority of the epistemological to the formal: that there can be
an F-logic depends on the fact that there are certain
"element-types" of which thought is capable, and it is on the
basis of these that F-logical systems are constructible.

519
Now all this leans heavily on a philosophy of language
and mind for its justification, and we left our examination of
Collingwood's philosophy of mind at exactly the point where we
now wish to resume the discussion. Whatever else Q-A logic,
F-logic, and D-logic may have in common, they are varieties of
discourse, and hence linguistic functions of mind at the level
of intellect. Consequently we must return to The Principles
of_Art (and in the next section to The New Leviathan) to examine
as best we can the functions of mind at the conceptual,
propositional, and rational levels, in order to see if we can
find a clue to unscrambling some of the puzzles we have thus far
encountered in our examination of the three logics with which we
are presently concerned. This will eventually lead us into a
discussion of Collingwood's views on abstraction, which has been
such a sensitive issue for Collingwood's interpreters, and which
we have thus far avoided.
Once more, at the risk of oversimplification, we shall
first present an outline of Collingwood's major conclusions on
language in The Principles of Art.

520
TABLE 11
LANGUAGE AND EXPRESSION IN THE PRINCIPLES OF ART
1. Expression is the fundamental practical act of mind in which

emotions are manifested by bodily acts. It is by the same act
that we express an emotion and become conscious of it. When a
man expresses an emotion what is meant is that he finds
himself in a feeling-situation of emotional excitement from
which he attempts to extricate himself by doing something,
where that doing is both a conscious activity and a
linguistic act. Expression in its primitive state is
non-descriptive because it is not a process of
conceptualization or classification: it is completely
individualized and oriented to the felt situation. It is also
primarily addressed by the person to himself, and secondarily
to an audience of persons like himself, and in both cases the
expression is intended to make his audience (including
himself) understand how he feels. (PA, 109-14).
2. Psychical expression (e.g. grimacing, blushing, cringings)

are primitive expressions at the psychical level. It consists
in the doing of involuntary and perhaps even totally
unconscious bodily acts, related in a single experience but
analyzable into the elements of a sensum, its emotional
charge, and the expression of that emotion. There are no
unexpressed emotions, because every kind and shade of emotion
at this level of experience has its expression in some change
of the muscular or circulatory or glandular system of the
organism. (PA, 228-30, 238).
3. Language in its wider sense, or imaginative expression,

consists of bodily actions expressing certain emotions
insofar as we are conscious that controlling them is our way
of expressing these emotions. These emotions are minimally
those which arise only through the consciousness of self
(e.g. hatred, love, anger, shame). Language in its widest
sense is the bodily expression of emotion dominated by
thought in its primitive form as consciousness. Within such
a system there is a synthesis of material elements
consisting of psychical expressions, but organized according
to a formal principle provided by the corresponding mode of
consciousness (i.e. primitive consciousness of self). (PA,
231-35).
521
4. Speech (vocal language or language proper) is a system of

gestures having the peculiarity that each gesture produces a
characteristic sound. Every kind or order of speech is an
offshoot from an original language of total bodily gesture,
that is, a language in which every movement and stationary
poise of every part of the body has the same kind of
significance which movements of the vocal organs possess in
a spoken language. This total bodily gesture is the one and
only real language: it is the motor side of our total
imaginative experience (or total activity of imaginative
consciousness). Speech is a function of self-consciousness:
the discovery of myself as a person is the discovery that I
can speak. But this self discovery is also the experience of
myself as a listener, and the consciousness of our own
existence is also the consciousness of the existence of
other persons as speakers in the community of language.
Understanding the speech of others involves an act of
imaginative, reconstructive consciousness in which mistakes
can occur, both in expressing one's own emotions (the
corruption of consciousness) or in mistaking the speech of
others (misunderstanding or misinterpretation). (PA, 242-51).
5. Intellectualized speech is language specifically suited to

expressing thought and its attendant intellectual emotions.
It differs from imaginative expression in the same way that
the object of imagination differs from that of intellect,
i.e. as something presented as one, indivisible,
self-contained and complete in itself (imagination) and as a
manifold of such objects with determinate relations between
them (analytic thought), or again as a relation between
something determinate and something indeterminate (abstract
thought). Intellectualized language, even when modified by
the grammatical and logical analysis of language, never loses
its emotive expressiveness. In its final form as artificial
symbolism invented for a purely scientific purpose, it has
both expressiveness and meaning, whereas in its imaginative
form it has expressiveness but not meaning. (PA, 252-61,
268-69).
522
As in the previous sections of this chapter, we will
group our comments around these summarized topics.
(1) We notice first that in his discussion, language is
located on a scale of forms, the minimal form of which is what
he calls "expression," and which consists of species that are
themselves concrete (i.e. having their own form or principle of
organization) but are further determined by successive forms
(i.e. are "matter" or "experience" for the "synthesis" of their
elementary parts by a higher mental act) (PA, 23034).
Collingwood even calls attention to this sort of structure when
he insists that "each level must organize itself according to
its own principles before a transition can be made to the next,
for until that has been done, the raw material needed for the
creation of the next is not forthcoming" (PA, 234). Thus the
emotions of consciousness (emotions that presuppose a
consciousness of self) must be formally or linguistically
expressed before a transition can be made to the level of
intellect.
One consequence of defining language by means of a scale
of forms is that successive levels incorporate the positive
content of all that precedes them in the scale--each is a
summary of what went before it. Consequently, true to his
description of such a scale in the Essay, Collingwood insists
that the element of expressiveness is never lost when

523
language develops into the forms expressive of intellectual
functions. Consistent with his Law of Primitive Survivals,
Collingwood therefore insists that there exists something he
wishes to call "emotions of intellect," which intellectualized
language, even in its symbolic form as mathematics or logic,
continues to express: intellectualized language has both ex-
pressiveness and meaning, whereas imaginative language has
expressiveness, but not meaning as distinct from it. Unfor-
tunately, aside from his brief suggestions that this is what is
involved in the excitement of intellectual discovery, and that
it is intellectual emotions that are involved when one values a
thought to the point that he feels it important enough to utter
in a given situation (PA, 164, 267), Collingwood does not
develop this interesting suggestion in The Principles of Art--at
least not beyond the point necessary for the elaboration of his
esthetic theory (it plays an important role in his discussion of
esthetic truth--cf. PA, 282-88).27 As we shall see in the
conclusion to this chapter,
____________________
27 This is not the only frustration for an interpreter
seeking to understand Collingwood's idea of language in The
Principles of Art: the chapter on language is the most exas-
perating chapter in the whole book. While it is the apex of the
entire argument of his esthetics (art is ultimately defined as
imaginative expression or language) and laced with pregnant
suggestions, it never fulfills its promise in a satisfying
analysis of the phenomenon of language. And what is worse is
that it presents examples which are not only misleading but
downright abusive. As an example of the emotive expressiveness
of intellectualized language Collingwood conjures up for his
readers the image of the "fastidious Cambridge mouth" of I. A.
Richards (whose theory of language in art he
524
there is concealed in the vagaries of this discussion an im-
portant clue to the understanding of his Q-A logic and its
relation to F-logic and D-logic--i.e. the heuristic intellective
function of questions as anticipations of answers.
(2) A second consequence of defining language by means
of a scale of forms is that each level presupposes the materials
presented to it by a lower level: psychical expression is
presupposed by language as imaginative expression; the language
of imaginative gesture is presupposed
______________________
clearly opposes); and he compares grammarians to primitive
African butchers who slice steaks from living animals (PA, 259,
264). Worst of all, he relies for many of his major conclusions
on arguments that are utterly contingent, gratuitous, and
unconvincing. The most notable instance of this is his argument
in support of the thesis that all language, even in the
symbolic, intellectualized language of the mathematician, is
emotively expressive. He states baldly that "every mathematician
knows" that a symbolism re-acquires the emotional expressiveness
of language proper. He then goes as far as to say that in
expressing a perfectly definite intellectual emotion a perfectly
definite act of thought is expressed too, so that a visiting
physicist seeing Archimedes racing naked from his bath down the
streets of Syracuse shouting "Eureka!" would, if he had himself
made Archimedes' discovery and knew how to read this expression
of intellectualized emotion, be able to understand the whole
scene, even that it was the discovery of specific gravity that
caused Archimedes' excitement (PA, 267-68). It is no wonder that
intelligent readers like Susanne Langer accuse him of "philo-
sophical malpractice" (Feeling and Form, p. 384), since ar-
guments like these are utterly unconvincing. On such grounds as
these one might prove that the earth is flat and that precise
scientific information may be passed from mind to mind by mental
telepathy. Unfortunately such diversions as these may also
distract the reader from supplying his own convincing examples
and arguments for the serious suggestions and principles he is
advocating.
525
by speech, and speech is presupposed by symbolic language. It is
therefore just a consequence of defining language by a scale of
forms when Collingwood states that one of the presuppositions of
F-logic (which occurs at the level of symbolic expression) is
the "propositional assumption" which presupposes the grammatical
and lexicographical units (words, phrases, sentences)
constructed at the level of speech: propositions or statements
(that which can be true or false) are isolated from a group of
expressions called sentences which are presumed to be already
grammatically well-formed. (Cf. PA, 260).
Unfortunately, for his readers, Collingwood appears so
bent upon emphasizing the artificiality of forcing the living
language of imaginative expression into the categories devised
by the grammatical and logical analysis of language that the
positive point that he is making is submerged under what appears
to be an abusive assault on grammarians and logicians for doing
what he admits to be their proper jobs. But if Collingwood's
description of grammarians as "butchers" is meant to be merely
abusive, why would he bother analyzing this abusiveness any
further? But he does, distinguishing between the grammatical
functions of lexicography (the cataloguing of recurrent units of
speech, called words, and the listing of their respective
meanings as relations of synonymy), accidence (rules governing
word inflection), and syntax (rules governing words as
functional units in sentences) (PA, 256-57)? And further,

526
why would he feel obliged to describe the three stages (actually
three presuppositions) in the grammatical analysis of speech,
viz. (a) the reduction of language as an activity to the product
of this activity, "speech" or "discourse"; (b) the division of
this product into units (words, idioms, phrases, sentences,
etc.); and (c) the devising of a schema of relations between
these units (syntax or the rules of grammar)--these presupposing
respectively the belief in a "metaphysical fiction" of the
finished product of speech-language, the existence of
self-sustaining atomic word-units that comprise this product,
and fictional rules governing their relationships (PA, 254-56)?
May we not safely assume that having taught himself half a dozen
languages and demonstrated his proficiency both as a translator
and as a celebrated user of English, that he was well aware of
the importance of such grammatical analysis and of lexicography
and the science of linguistics in general, but that he felt
constrained to warn his readers against a certain reductionistic
tendency which results from an overtly idolatrous attitude
towards the teachings of such sciences? And are his remarks in
opposition to this attitude not indicative of his continuing
struggle against what he understood to be the realistic
thesis28--in this case, the treatment of
______________________
28 This is clear from a passage in which he first de-
scribes the activity of the grammarian as a cutting up of a
presumed "thing" called language into pieces called words. "Some
readers will object to this
527
language as as a "thing"?
(3) Of more direct concern to this chapter is his
discussion of "the logical analysis of language"--by which he
understands "a certain technique, first systematically
expounded. . . in Aristotle's Organon" and subsequently
developed by a long line of logicians culminating in "the
logical analysts and positivists of the present day" (PA, 259).
The aim of this technique, he says, is to make language into a
perfect vehicle for the expression of thought--that is, to
remove the
_________________
phrase on the ground that I have used a verb of acting when I
ought to have used a verb of thinking; a dangerous habit, they
will remind me, because when you get to the point of saying
'thought constructs the world' when you mean 'some one thinks
how the world is constructed,' you have slipped into idealism
through mere looseness of language; and that, they will add, is
the way idealists are made. There is much that might be said in
answer to this objection; as, that philosophical controversies
are not to be settled by a kind of police-regulation governing
people's choice of words, and that a school of thought . . .
which depends for its existence on enforcing a particular jargon
is a school which I neither respect nor fear. But I prefer to
reply merely that I said cut because I meant cut. The division
of the 'thing' known as language into words is a division not
discovered, but devised, in the process of analyzing it" (PA,
255). In the immediately succeeding passage, Collingwood also
declares his refusal "to be frightened by the bogy of idealism"
when he states that grammatical rules are devised rather than
discovered (PA, 256). We might add that the reason that
Collingwood meant to say "cut" when he could have said "dissect,
and "butcher" when he could have said "anatomist" (for after
all, grammarians do cut up language "at its joints" as Aristotle
would say, rather than hacking them across muscle bundles for
sale as steaks at a market), is that his intention was polemical
as well as analytical. He wished to warn us of what he regarded
as a dangerous tendency. What else would this be but realism?
528
frustrations of expressions due to the inaccuracies and am-
biguities in speech, and to do so by replacing them with logical
forms (either of the subject-predicate form of Aristotelian
logic or the propositional forms of what Collingwood calls the
"modern analytic school" of logic) (PA, 259-60).
Once again Collingwood distracts the reader from his
positive treatment of the subject by his emphasis on the ar-
tificiality of the logician's project, i.e. proposing a modi-
fication of language rather than a theory of it. What a con-
temporary logician would be surprised at is not the accusation
that he is proposing a modification of language (for after all,
most contemporary logicians would agree that what they are doing
is constructing formal systems, which may or may not have a
similarity or applicability to natural languages). On the
contrary he would be surprised by the suggestion that what he
might be mistaken to be doing is the construction of a theory of
any natural language (which is taken to be the province of a
science of linguistics).
But he might further not only be surprised by, but also
take issue with, Collingwood's list of the "assumptions" of the
logical analysis of language. In addition to assuming that the
grammatical transformation of language has been successfully
accomplished, the logician, Collingwood says, makes three
further assumptions:
529
First comes what I shall call the propositional assumption.

This is the assumption that, among the various "sentences"
already distinguished by grammarians, there are some which,
instead of expressing emotions, make statements. It is to
these that the logician confines his attention. Second, the
principle of homolingual translation. This is an assumption
about sentences corresponding to the lexicographer's
assumptions about words (or . . . lexicographical units)
when he "defines the meaning" of a given word by equating
it with that of another, or of a group of words taken
together. According to the principle of homolingual
translation, one sentence may have precisely the same
meaning as another single sentence, or group of sentences
taken together, in the same language, so that one may be
substituted for the other without change of meaning. The
third assumption is that of logical preferability: namely
that, of two sentences or sentence-groups having the same
meaning, one may be preferable, from a logician's point of
view, to the other . . . . The preferred version is
preferred because it is one which the rules of the
logician's technique enable him to manipulate. (PA, 260-61;
cf. A, 3536).
We suspect that a contemporary logician would object to this
description of the assumptions he makes in his logical in-
quiries: (a) he would object to the confinement of his attention
to statements alone, and worse still to statements as sentences
already distinguished by grammarians; (b) he would say that
homolingual translation or substitutability is not dependent on
identity of meaning alone, but on the logical form or structure
of language as well; and (c) he would argue that logical
preferability is not something dependent solely on the
logician's point of view, but rather is presupposed by the
logical employment of sentences in a language and merely
displayed by translation into formal structures. But rather than
presuming to speak for contemporary logicians, we are more

530
interested in what Collingwood has to say about these
assumptions, and in this connection it is noteworthy that he
never denies that they are valid assumptions. He says only that
as proposals for the modification of language, they can never be
carried out in their entirety. Language proper, even as an
artificial symbolism, can never lose its emotive-expressive
aspect. Once mastered, a symbolic language, invented to serve a
technical purpose, reacquires the emotional expressiveness of
language proper (PA, 262, 268).
The principles of homolingual translation and logical
preferability provide us with some supporting evidence for the
suggestion we made in considering the "variable" (which the
Essay on Philosophical Method declared to be extrinsic to the
generic essence in non-philosophical concepts, but identical
with it in philosophical concepts) in terms of the substi-
tution-relations of a logic based on abstract class concepts. As
we noted in our discussion of the Essay, when Speculum Mentis
described the abstract universal of logic as one which is
"indifferent" to the variation of its instances, and when the
Essay described the classes of F-logic as one in which the
variable is not identified with the generic essence, what he
seems to have in mind is the fact that in F-logic replacement
instances of a class (whether this be of terms or propositions)
are indifferent to whatever other properties these instances may
have: their identity is based solely on the criteria for

531
their membership in the class. They are therefore substitutable
one for the other. In Speculum Mentis and in the Essay
Collingwood deliberately contrasted this sort of replacement-
relation with another in which the differences of the instances
of a class are relevant to the class itself. In Speculum Mentis
this was called a "concrete universal," or one in which the
variable element between the instances of the universal is not
ignored, but makes a difference to the generic universal itself;
in the Essay this was expressed by saying that for philosophical
concepts, the variable (presumably what allows one instance of a
concept to differ from another instance of the same concept) is
identified with the generic essence (i.e. is related to the
meaning of the concept in a necessary and essential manner). We
suggested that if such differences between individual instances
of a universal are essential to the identity of the concept as
such, then they must be regarded as part of that concept's
meaning--which brought us to the brink of a discussion of the
analytic-synthetic distinction.
What we wish to point out here is that in an explicit
discussion of the assumptions of what is undoubtedly F-logic,
Collingwood states the replacement-relation of substitutability
of instances for one another in a class as essential to the F-
logical analysis of language.
(4) Another interesting corollary of the definition of
language by means of a scale of forms is that in the
specification of the genus

532
"expression" into the species "symbolic language," the essential
functions of emotive expression and intentional meaning are both
retained in the same act. This has a peculiar outcome: since no
matter how artifactual a symbolic language becomes it retains or
"reacquires" the emotional expressiveness of language proper, in
Collingwood's view it never loses contact with the emotional
life of the speaker.29
As his remarks about interpersonal self-consciousness
and the community of speakers and hearers indicates (PA, 247--
52), he does not believe that such an admission leads to a form
of linguistic solipsism in which a "private language" is
employed which no one can understand but the speaker (since no
one can "read" the emotional life internal to a user of
___________________
29 Computer analysis and programmers reading this today
might be startled by such a conclusion, since machine-
communication languages like COBAL seem to work quite effec-
tively for the tasks which they are designed to perform. What
Collingwood might say about the development of such languages as
these is hard to say, since it would be peculiar to predicate
"emotional expressiveness" to a computer print-out, while one
would also be forced to admit that there is some sort of
communication occurring between man and machine by means of such
languages. Our guess is that Collingwood would not take
computer-language to be language at all, but a sort of book-
keeping aided by mechanical and electronic devices, not any more
expressive of thought than the noises emitted from a
tape-recorder or the typescript emerging from a typewriter. The
computer operator is still engaging in linguistic expression at
the input and interpretation ends of the process of machine
communication: the rest is automatic book-keeping, not thought.
533
a private language). On the contrary Collingwood goes to the op-
posite extreme and declares that at a primitive level total
communication can and does occur--as in the spread of emotion
like a kind of contagion from person to person, e.g. panic in a
crowd.30 Although this occurs directly only at the psychical
level (and this is the level at which even animals can be said
to communicate), at levels above this, in which emotions require
consciousness of self, such communications occur usually only
through "language" in its broadest sense (as totally bodily
gesture--cf. PA, 238, where Collingwood baldly states that there
are no unexpressed emotions).
Collingwood does not spell out in detail how the in-
tentional meanings of a speaker are bound to his expression of
emotion, but it is clear that he wishes to retain both functions
in all levels of linguistic activity. For Collingwood meaning
never loses its aspect of being
___________________
30 In speaking of psychical emotions which are always
expressed as some change in the muscular or circulatory or
glandular system, Collingwood writes that "nothing but lack of
skill ((in observing such changes and correctly interpreting
them)) prevents us from reading like an open book the psychical
emotions of every one with whom we have to do. But observing and
interpreting is an intellectual process; and this is not the
only way in which psychical expression conveys a meaning. There
is a kind of emotional contagion which takes effect without any
intellectual activity; without the presence even of
consciousness" (PA, 230). Collingwood cites, in addition to the
spread of panic through a crowd, the examples of the sympathetic
feeling of pain or joy, and the terror transmitted from prey to
predator.
534
something that a speaker does (where "speech," of course, is
taken in a broad sense to include not only vocal utterance but
any suitable substitute for it--e.g. the sign language of mutes;
it includes anything in which a physical gesture can be tied to
a meaning). Meanings are bound to what have more recently been
called "illocutionary acts," or what a speaker intends to convey
by what he says. They are not something contained in a
dictionary, with fixed relations of synonymy, but are part of a
living process linked to the emotional life of speakers, whose
intentions to convey meanings have a career in which meanings
are born, develop in a context of relations, and may even die.
Meanings, on Collingwood's view, are not something words have,
but something that speakers do with words. (cf. PA, 269).
This is as far as Collingwood takes us on the path to
language. Whether or not a case can be made for a language that
is purely symbolic or has meaning without being expressive, such
a thing would not be regarded as a true language by Collingwood.
(5) One final aspect of Collingwood's discussion of
language needs to be mentioned before going on to relate lang-
uage to the levels of consciousness, and that is the global or
comprehensive aspect of each level of the linguistic scale of
forms. Each has what Collingwood calls its own principle or
organization or form, although he is not always

535
careful to make explicit what that principle is. In the final
analysis it might not be possible to do so, since any one level
may be capable of a virtual infinity of forms with no apparent
unifying communality between them: imaginative expression in
art, for example, encompasses everything from dancing
(Collingwood calls dance the mother of language) to drama to
painting--the list is only arbitrarily broken off, every work of
art being a "monad" and even the everyday acts of human life
being to some extent works of art.
But in attempting to characterize them as essential
structures, Collingwood points continually toward a completed,
global activity--in the case under discussion, the "language of
total bodily gesture," which he calls the only real (imag-
inative) language (PA, 246-47). He correlates this "total bodily
gesture" language with the corresponding mental function of
"total imaginative activity" (AP, 247). There is a similarity
here to what we have seen Collingwood advocating concerning the
organizational unity of the "logical structure of real thinking"
in his correspondence with Ryle: in any example of real thinking
there is contained an example of every kind of proposition as so
many "element-types" (CRC-I, 18). If any level of consciousness
has its own organizational unity, and if this unity must be
complete before a transition to a higher level of consciousness
may occur (cf. PA, 233-34), then the explicit articulation

536
of the "logic" or coherence of any level of consciousness is a
matter of making explicit or expressing what is already a
completed whole, insofar as that unity is viewed from a higher
level.
We may speculate, in terms of our discussion of logic,
that the systematic unity of F-logic (or what we may now call a
symbolic language in which what we say can be distinguished--but
not separated--from what we mean, and artificial symbols can be
devised and stipulated for a particular meaning) is exhibited
only by the assumption of a higher viewpoint (what might be
called today a "meta-language") for which F-logic is an object.
In Speculum Mentis we observed Collingwood making just such an
argument: F-logic is the unifying principle of a form of
experience which he called "scientific thinking," in which the
distinction between what we say and what we mean is first made
manifest (SM, 128-30, 154-57; cf. PA, 269); but if the principle
of organization of such an activity is not already somehow
complete, the transition could not be made to a higher viewpoint
for which the lower is regarded as an object. Philosophy is such
a higher viewpoint, according to Speculum Mentis, and its logic
is dialectical. Can we conclude for Collingwood that D-logic is
a metalanguage for the discussion of objects that have F-logic
as their unifying principle?

537
Perhaps. But once more we must point out where Col-
lingwood leaves the discussion and where we resume it. He never
called D-logic a meta-language, nor did he ever explicitly work
out the relationship between D-logic and F-logic other than to
point out in the Essay the salient ways in which D-logic differs
from F-logic in their respective ideals. We have also tried to
remain alert to the fact that the existence of something which
meets the description Collingwood gives of D-logic depends on
his ability to provide us with concrete, irreducible, and
convincing instances of such a logic. So far the instances of
D-logic that we have encountered are those which stress the
overlapping layers of meaning in certain terms used by
philosophers, the unity of meaning and reference in universal
categorical philosophical judgments, and the reversible
inference structures employed in typical philosophical
arguments. But in the final analysis the issue of whether there
is anything which meets the criteria for D-logic described by
the Essay rests, as we saw in our analysis of the Essay, on
whether or not there can be a kind of knowledge which is
self-justifying.
But in the meantime we have already seen that for Col-
lingwood the issue of logic is always bound to that of language,
and language is a function of a conscious mind. Before
concluding our survey of the topic we therefore must attend to
final discussion of logical mental functions in The New
Leviathan.
538
5. Language and Mind: The New Leviathan.
We are now in as good a position as we shall ever be to
mount a final assault on the heights of Collingwood's philosophy
of mind--i.e. the functions of intellect and reason. We must
remind ourselves once again of the difficulties we encountered
in Chapter VIII in reconciling the conflicting terminology
between The Principles of Art and The New Leviathan.31 Some of
these terminological ambiguities, we noted, can be cleared up by
recognizing wider and narrower senses that Collingwood allows
for terms like "consciousness" and "thought;" and some of them
can be reconciled by paying careful attention to the differing
approaches of the two works, which helps put such unannounced
shifts in the comprehension of its terms in their methodological
contexts.
Although we did not call attention to this fact in the
immediately proceeding section of this present chapter, this
shift in approach between the two works is even apparent in the
discussion of language in each of the two works. The Principles
of Art considers language in terms of the
____________________
31 The principal discrepancy is that in The Principles
of Art consciousness, even in its most primitive-form, is called
thought, and its function as attention is selective without
being abstractive; in The New Leviathan "simple consciousness of
feeling" is not considered to be thought, and selective
attention functions only at the secondary level of consciousness
and is abstractive (cf. PA, 204-06, 215-17; NL, 4.13-4.5).
539
dialectical array of forms of expressiveness from psychical
expression in physiological changes to symbolic language; The
New Leviathan considers language as an abstraction from, or re-
striction of, "discourse" or "the activity by which a man means
anything" (NL, 6.1-6.11). An unwary interpreter might conclude
from a superficial comparison of these two treatments of
language that Collingwood had "changed his mind" or "repudiated"
his earlier treatment of the subject. Our own inclination is to
see the later work in the light of the earlier one, and hence as
an analytic discussion, from a restricted point of view, of a
topic which has a broader context and represents one portion of
a "scale of forms"--in this case, the scale of forms of
expression.
However this does not resolve all the interpretative
difficulties posed by The New Leviathan on the issue of language
and logic. We shall attend to some of the more important of
these problems after our summary of the doctrines of The New
Leviathan on language and the levels of mental functions.

540
TABLE 12
LANGUAGE, LOGIC, AND MENTAL ACTS IN THE NEW LEVIATHAN
1. Discourse is the activity by which a man means anything: it

is the activity of meaning something (a) by something else
(b), where meaning (a) is an act of theoretical con-
sciousness, and (b) is a practical activity, the production
in oneself or others of a flow of sounds or the like which
serve as the vehicle of that meaning. Discourse is
continuous, so that even the rests or pauses of silence,
immobility, or the like which punctuate it are significant
parts of it, not interruptions of it. But selective at-
tention breaks it up into words; vocal words if it is
speech, gesture-words if in gesture, etc. Once discourse is
broken up into units it is the lexicographer's business to
determine the meaning or various meanings which a given word
bears whenever it is used. The phonetic or other vehicle of
discourse is physical in the sense of being a succession of
feelings or sensations with their emotional charges produced
by the activity of speech or the like. (NL, 6.1-6.19).
2. Language is an abstraction from discourse: it is the system

adopted, the means employed, the rules followed in the
activity of discourse. It is a system of sounds or the like
as having meanings, these meanings being what a person using
that word means by making that sound or gesture. Language is
not a device whereby knowledge already existing in one man's
mind is communicated to another's, but an activity prior to
knowledge itself, without which knowledge could never come
into existence. As consciousness develops, language develops
with it. (NL, 6.11, 6.18, 6.41, 6.58).
a. Language in its simplest form is the language of con-

sciousness in its simplest form; the mere register of
feelings, irrational, unplanned, unorganized, uncon-
scious. At this level of consciousness thought is merely
apprehensive, or capable of taking what is "given" to
it. (NL, 6.58, 10.51).
b. When consciousness becomes conceptual thought, language

develops abstract terms. At this level of consciousness
thought is capable of framing abstractions from what is
given. Selective attention breaks up continuous
discourse into words. A word is a lin-
541
guistic habit of the community using it; the habit of
conveying a special meaning by using any member of a
certain class of auditory and visual vehicles, of which
any member is an example of that word. (NL, 6.12, 6.17,
6.58, 10.52, 34.123.
c. When consciousness becomes propositional thought, lang-

uage develops the indicative sentence as the standard
verbal form in which to state the proposition. At this
level of consciousness thought is capable of dis-
criminating truth from error.(NL, 6.59, 10.51).
d. When consciousness becomes reason language becomes de-

monstrative discourse wherein sentences are so linked
together as to state verbally the consequences of one
proposition in relation to another. At this level of
consciousness thought is capable of understanding both
itself and other things.(NL,6.59,10.51).
3. Conceptual thinking or selective attention is an act of

practical consciousness by which a man emerges from the state
of feeling and simple consciousness of feeling. A man makes
himself conscious of his feelings by naming them, either by
the language of gesture (e.g. by an expressive shiver) or by
the language of speech (saying "cold"). To name the feeling
awakens consciousness of the feeling: until it is named, the
feeling is preconscious; when it is named it becomes
conscious. (NL, 6.2-6.28, 7.2-7.23).
a. Selective attention is the act in which a person attends

to some element or group of elements within the a
confused mass of feeling. The act of attending is doing
something to oneself as well as doing something to the
object being attended to; it is focusing one's own
consciousness on a certain part of the field of feeling
and repressing the rest; but it is also circumscribing
it, drawing a line between it and the rest of the field
(NL, 7.23-7.24).
b. The act of classifying is the practical activity of

consciousness involved in "drawing the line" between
objects (sense) in a field of feeling, i.e. the point at
which one decides to stop calling the color seen "red"
and begin calling it "purple" or whatever. On acts like
this classes depend for their existence; all classes
being artifacts, depending on practical activities for
their existence and depending for their publicity as
between various persons on these persons performing
practical activities of similar kinds; for many classes
are private to the persons who made them. A class is a
collection of many things into one, in virtue of their
resemblance. The distinguishing mark of a class is that
it is a whole whose parts, its "members," are mutually
related by way of resemblance.
542
Membership in a given class demands a certain kind and

degree of resemblance. The settlement of what the degree
and kind of resemblance is for membership in a class is
an act of practical consciousness called classifying.
(NL, 19. 22-19.35).
c. Evocative thinking is the act of arousing in oneself by

the work of thought feelings not found as given in
oneself; it is an act which goes with the act of se-
lective attention. Feelings thus aroused, called
evocations, form a context inseparable from any
selection, and are connected with it by logical
relations. Evocations are feelings felt but not given,
produced by the same act of practical consciousness that
produced the selection or abstraction from the given
field of conscious feeling. (NL, 7.32-7.38).
d. A concept (notion) is a selection together with its

context of evocations. Any logical relation may preside
over the birth, from a given selection, of an evocation
forming part of its context. It is not possible to
compile a list of such logical relations, although
contrast and comparison are two such relations. (NL,
7.34-7.36, 7.39).
e. Abstractions are determinate in some ways, indeterminate

in others (as in a triangle, which is determinate in
having three sides, but except for what is implied in
this, it is indeterminate in everything else). Ab-
stractions are only second-order objects made by the
mind out of its immediate or first-order objects. Ab-
straction is a necessary part of thought; an abstraction
is false if the elements abstracted are judged to be
mutually independent entities. (NL, 7.56-7.57, 7.67,
26.18-26.19).
4. Propositional thinking is the set of mental acts that are

involved in asking and answering questions, and disting-
uishing truth from error (and good from evil). Asking a
question implies contemplating alternatives. A question that
offers no alternatives is a bogus question, from the point
of view of knowledge proper. The technique of knowing proper
or scientific method depends on replacing vague or confused
questions, which are unanswerable, with real questions which
have a precise answer. (NL, 4.344.35, 11.12).
a. Let a man have a certain form of consciousness, C1. To

that form of consciousness let x, y, z be immediate
objects. Let him call into being in himself another form
of consciousness, C2, the consciousness of C1. Then to
C2, C1 is a first-order object, and x, y, z are
second-order objects. For the form of consciousness, C2,
the second-order objects x, y, z, are abstractions from
the first-order object C1 made by C2. and the knowledge
of these abstractions is mediated by C1.(NL, 5.26,
11.34).
543
b. The subject of a proposition or what the proposition is

about is never a first-order object. (NL, 11.34).
c. The predicate of a proposition is never a first-order

object, but always a concept (NL, 11.35).
d. Logic applies to propositions because the predicate of a

proposition is a concept, and logic applies to concepts.
Because the predicate of a proposition is a concept, any
proposition may theoretically involve a mistake, though
there are mistakes that people do not make. (NL, 11.35).
5. Rational thinking, or reason as a mental function or form of

consciousness, is thinking one thing x because you think
another thing y, where y is your reason or ground for
thinking x. A piece of rational thinking involves at least
two propositions standing to each other as ground and
consequent. Rational thinking begins when a man accustomed
to propositional thinking starts making a distinction not
made in propositional thinking as such: the distinction
between "the that" and "the why." The distinction is
preconscious until it is reflected upon. (NL, 14.1-14.2).
a. Simple knowledge is the knowledge that arises when a man

reflects upon a piece of propositional thinking and asks
himself whether he has really done it, and answers in
the affirmative. Such a judgment is fallible, i.e.
errors can be made about it. (NL, 14.2114.25).
b. Reflection on the fallibility of such a "that" judgment

prompts one to seek out a second proposition which
offers reassurance of the trustworthiness of the first.
It is the practical act of trying to alleviate the
distress caused me by the untrustworthiness of my
knowledge that gives rise to the distinction between
"the that" and "the why." (NL, 14.25-14.29).
c. Practical reason (making up one's mind to, or forming

what a moralist calls reasons for an intention) comes
into existence when a man forms an intention, reflects
on it, and asks himself whether he really means it. In
this reflection he seeks another intention y to confirm
the original intention x, something from which the x may
follow as a necessary consequence, or stands in relation
to it as ground to consequent. Reason is essentially
practical and hence prior to theoretical reason, because
to be reasonable means to be interested in questions
beginning with "why"; and this happens because people
crave for reassurance against the fallibility of their
knowledge. (NL, 14.31-14.5).
544
d. Theoretical reason (making up one's mind that, or seek-

ing reasons for a proposition) comes into existence when
a man first, by propositional thinking, makes up his
mind that something is so, and then, seeking to confirm
this piece of propositional thinking, looks for a reason
why he should think so. Theoretical reason is based on
the presupposition that a certain kind of propositional
thinking, viz. that about which questions beginning with
"why" can be legitimately asked, is a matter of free
will (i.e. it is not the mere acceptance of something as
given, but is a voluntary decision to think this and not
that). (NL, 14.35-14.37).
e.Questions beginning with "why" cannot be legitimately

asked about first-order objects, but only about objects
of the second and higher orders (i.e. abstractions). If x
is an intention, any ground for it, A, must be another
intention. One intention supporting another both form
part of the same intention, which includes them both and
perhaps other things. Let us call this larger intention
I. Similarly if s is a proposition about whose truth
someone desires reassurance, it follows that t, the
ground of that reassurance, must be a proposition of
whose truth he is satisfied; and s and t are here
abstractions from a first-order proposition P which
includes them both and perhaps other things as well. As
long as I and P are first-order objects they are matters
of immediate conviction or resolution. To demand
confirmation of either would be to place it in a context
of other intentions or other propositions that might
afford grounds for it; that is, to reduce it to the
level of, or to make of it, an abstraction. An intention
is made into an abstraction by surrounding it with a
context of other intentions; a proposition by
surrounding it with a context of other propositions.
(NL, 14.39-14.44).
____________________________________________________________
545
With the advance warning that this section will contain,
of necessity, some rather thorny exegetical discussion, which we
shall try to confine so far as possible to the doctrines
summarized in this table, we shall again group our comments
around these topics.
(1) Coming to The New Leviathan fresh from the dis-
cussion of language in The Principles of Art, one cannot help
but notice some evident similarities which render the apparent
differences in the discussion of consciousness in the later work
less paradoxical. (a) In both works we find a field of feeling
at which a primary consciousness (called "simple consciousness"
in the earlier work and "diffuse consciousness" in the later) is
directed. In this "first state of mental life" consciousness is
merely "apprehensive" (NL, 10.51) or "appreciates" (PA, 203)
what is presented to it in feeling. The expression of this level
of consciousness is loosely called "language" but it is of an
illogical, ejaculatory sort of gesture language and not yet
speech. (b) The second stage of mental life begins with
selective attention, an act of second-order consciousness (i.e.
the consciousness of first-order consciousness) in which
attention is directed towards some sense and away from others in
a field of firstorder, conscious feelings. Second-order
consciousness is expressed in an act (pointing, speaking, making
designating gestures, etc.) which is the same as the act of
consciousness by which he makes himself aware

546
of some designated entity or feeling as that which he is
conscious of, or what he "means" (in the widest sense of the
term "meaning" which includes the most rudimentary act of
reference). This is the level of conceptual thought. (c) At
higher stages of mental life the expression of conscious
functions include what we know as language proper--indicative
and interrogative sentences at the level of propositional
thought, demonstrative discourse for rational consciousness. The
terms "intellect" and "knowledge" generally apply to the second
level of consciousness and the mental functions above it, which
as forms of reflective consciousness presuppose primary
consciousness without which it would have nothing to reflect on
(NL, 9.54).
In all this we seem to be on familiar territory, just as
we are in the discussion of discourse as a continuous system of
bodily movements, only later broken up by "selective attention"
and made into words that are subsequently catalogued by a
lexicographer (NL, 6.1-6.18). Is this not what we have become
acquainted with in The Principles of Art as the "one real
language of total bodily gesture" which is transformed into
speech and then "cut up" by the "butchers" who practice
lexicography and grammar? Even Collingwood's penchant to phrase
all his definitions in terms of conscious acts of human agents
(especially noteworthy in his description of reasoning, where
the terms "inference" and "implica-

547
tion" are conspicuous by their absence) takes us back to the
beginning of the present chapter, where we found his definition
of "logical efficacy" and Q-A correlativity formally ob-
jectionable precisely because they introduce this personal and
epistemological dimension.
But there are also some important differences which
cannot be overlooked, these tending to group themselves around
the three major intellective functions of conceptual, propo-
sitional, and rational thinking. Leaving aside the latter two
for a moment, we notice first that there is no precise analog in
The Principles of Art to correspond to the discussion of
"naming" in The New Leviathan. In fact in the later work
Collingwood appears to be putting forward a radical linguistic
thesis that it is by naming something that we become selectively
aware of it: language is "prior to knowledge itself" and
knowledge could not come into existence without it. How are we
to reconcile this with The Principles of Art, where Collingwood
says that simple or first-order attention is the "act of
appreciating something, just as it stands, before I can begin to
classify it," and that this is done by "identifying" something
as having just the qualities we find it to possess before naming
it (PA, 203)? And what of the passage from the Essay on
Philosophical Method where he writes:
It is only in some dark and half-conscious way that we know

our thoughts before we come to express them. Yet in that
obscure fashion they are already within us; and, rising
into full consciousness as we find the words to utter them,
it is they that determine the words, not vice-versa. (EPM,
200).
548
The astute reader will note that if indeed what he calls
the "diffuse consciousness" of feeling in The New Leviathan is
the same mental function as "simple consciousness" and
first-order attention in The Principles of Art, then there is no
insoluble conflict: first-order attention divides a field of
feeling into focal and penumbral regions, but it does not
abstract (PA, 204). Therefore second-order attention could be
abstractive and denominative, without violating first-order
functions of consciousness. And surely this parallel is
indicated when Collingwood writes that "thought is at first
merely apprehensive, capable of taking what is 'given' to it,
and then merely conceptual, capable of framing abstractions from
what is 'given"' (NL, 10.51). Similar discrepancies in the use
of the term "attention" in The New Leviathan can often be
resolved by adding appropriate qualifiers in passages where the
unqualified term "attention" is used without specifying whether
it is referring to the first-order function of consciousness
(which divides without abstracting) or second-order
consciousness (which is selective and abstractive). Thus when he
writes (in a passage which is directly proceeded by a discussion
of selective attention) that "The act of attending is not merely
a doing something to yourself, focusing your consciousness . . .
it is also a
549
doing something to the object: circumscribing it, drawing a line
between it and the rest of the field" (NL, 7.3; cf. 4.51), he
appears to be locating both division and selective abstraction
in the attentive act of second-order consciousness, but "act of
attending" here may be referring to both levels, just as
"consciousness" may refer to a given level or to that level and
all below it.
(2) Unfortunately sorting out labels for levels of
consciousness and adding qualifiers for functions of conscious-
ness does not settle the issue of how it is possible to "divide"
a field of feeling or to "circumscribe" something within it
without being "selective" about it, or contrariwise how
first-order consciousness if it is not selective, can divide
without abstracting. Granted that Collingwood's stated intention
in The New Leviathan is to take successive "soundings" at
various levels of consciousness (NL, 9.4-9.42), and that he is
therefore not committed to stating why one level gives rise to
the next, or how they come into existence, or even how certain
mental functions are possible at all; nonetheless difficulties
like these pose problems for the reader even in understanding
the descriptive sense of these "soundings." In this case the
problem is one of understanding the functions of division and
abstraction in discourse, and especially as these pertain to the
process of "naming."
550
Particularly confusing is Collingwood's discussion of
naming as a mediating act of consciousness, which view he re-
jects with the following argument:
Until you name it, the feeling is preconscious. When you

name it, it becomes conscious. This does not mean that the
act of naming it becomes conscious; it does not, either as
an act of your own or even merely as the sound of your
voice or the like. It remains preconscious until you
reflect upon it. . . . "If a man becomes conscious of a
feeling only through finding a name for it, is not that a
way of saying that his consciousness of the feeling is not
immediate, as you said (4.22), but mediated through
language?" The consciousness of B is mediate if you can
only be conscious of B as an abstraction from something
else, A, of which you are conscious. Let A be something of
which you are immediately conscious; then A is a first-
order object and B. the abstraction from it, a second-order
object (5.25), and the consciousness of B is mediated
through the consciousness of A . . . . But the feeling is
not an abstraction from the name of the feeling. The man
who names his feeling thereby becomes immediately conscious
of it; he is not conscious of his name for it until he
reflects on the act of naming it, and he proceeds to think
of the name he has uttered in abstraction from that act.
(NL, 6.28-6.39).
It would appear from this passage that either (a) there is ab-
solutely no function to the "diffuse consciousness" of feeling
(since feeling presents itself directly to second-order
consciousness and its function of selective attention by nam-
ing), or (b) naming it (a second-order function) makes us di-
rectly conscious not of the feeling but of the first-order
consciousness of a feeling (that is, a feeling as "apprehended"
by first-order consciousness--or, as The Principles of Art puts
it, a feeling as "appreciated" by being perpetuated and
domesticated by first-order
551
attention). But if (b) is the case, then what are we to make of
the remark that feeling is not an abstraction from the name of
the feeling--and especially when he defines abstraction in this
context as the attainment of a second-order object by means of a
first-order object, where the first-order object is what is
immediately given to consciousness?
Since (b) is the most likely alternative for making
sense of both The New Leviathan and The Principles of Art on the
issue of first and second-order consciousness, we must settle
the problem of the peculiar usage of the term "abstrac-
tion"--which we shall in (although this possibility is not
made explicit in The New Leviathan) that this may not be an
"either-or" situation; i.e. it may be possible that even though
second-order consciousness is consciousness of first-order
consciousness, it may also be directly conscious of the object
of first-order consciousness as well, that object being feeling.
This would be consistent with his remark in The Principles of Art
that "attention (or . . . consciousness or awareness) has a
double object where sentience has a single," that is, "a person
who is said to be looking is described as aware of his own
seeing as well as of the thing he sees," or again "(w)hat we
attend to is two things at once: a sound, and our act of hearing
it" (PA, 206). It is also consistent with his "Law of Primitive
Survivals"
552
(which we will also deal with in a moment) in The New Leviathan,
which states that in any higher form of consciousness there must
survive elements of a previous function in its primitive or
unmodified state (NL, 6.52); in this case we suggest that what
he may be saying is that feeling is still directly present in
first-order consciousness as a "primitive survival," and hence
still an immediate object to second-order consciousness (the
distinction between "apanage" and "constituent" to the contrary
notwithstanding).
Secondly we wish to call attention to the fact that at
the very lowest level of consciousness "meaning" is introduced:
at the "first state of mental life" a man is said to be con-
scious of a confused mass of feeling because he has found a
language of some kind by which he can "mean" it, albeit a very
primitive, illogical, ejaculatory sort of language (NL, 7.24).
Therefore abstractive or selective attention, the "second stage
of mental life," presupposes the presence of some primitive,
pre-reflective "meaning" on which the naming function operates.
But if "meaning" here refers to the level of gesture-language
(which is included in what Collingwood calls "discourse"), i.e.
the "system of bodily movements, not necessarily vocal, whereby
the men who make them mean or signify anything" (NL, 6.1), then
in its primitive function consciousness can mean something
without naming it. Using "discursive" as an adjective for
Collingwood's use of "discourse"

553
in this broad sense, we can say that such pre-reflective "mean-
ing" in first-order consciousness is discursive without being
properly speaking linguistic (cf. NL, 6.11). The conscious cry
of a distressed infant (recognizably distinct from the random
cries of the newborn--cf. PA, 235-37) might be an example of an
expression that means something (the infant is consciously
signaling his distress) without naming it; it is an act that is
discursive (carries meaning) without being linguistic (using
designated words in a spoken language). But such meanings are
individualized expressions, and not generalized descriptions:
the infant's cry of distress is an expression of his present
emotional state, and not in any way a descriptive generalization
about a feeling or feelings not present, or involving relations
with these other feelings (PA, 112; cf. IH, 314, 330).
If we are on the right track, then when Collingwood
writes that a man makes himself conscious of a feeling by naming
it (NL, 6.25), and adds that this is an act of conceptual
thought (NL, 7.23), and adds further that with this act goes an
act of evocative thinking (defined as "the act of arousing in
yourself by the work of thought feelings you do not find as
'given' in yourself"--NL, 7.32), he is not by-passing or
revising the function of first-order consciousness and its
primitive, discursive
554
meaning-act, but rather he is presupposing it.32 The very first
act of consciousness (and it must be remembered that it is only
second-order consciousness that is reflective and hence properly
speaking thought) is an act of designating meaning, and this
assignment of meaning is coextensive (and one might even say co-
intensive) with consciousness itself. If the question were put
to Collingwood, "Can an analysis of mind and consciousness get
beyond or behind language?" we suggest that he would answer,
"Yes, but not by language proper; and if you were to succeed in
getting there, what you would find would be meanings too
primitive to express in words." But as Collingwood pointed out,
reaching this primitive level is an experiment not easily made
(NL, 6.56), so the function of
_____________________
32 The interpretation we are offering of Collingwood's
description of the functions of first- and second-order con-
sciousness differs sharply with that offered by Alan Donagan.
Donagan argues for a reversal between The Principles of Art and
The New Leviathan on the issue of selective attention, and
postulates that between the writing of the earlier and the later
works Collingwood changed his mind: he came to hold that all
thinking is conceptual and all concepts are abstract (LPC, pp.
14, 48-49). We have already argued that this view fails to hold
without qualification for The Principles of Art, and we are now
arguing that it does not hold even for The New Leviathan. We
maintain that Donagan failed to appreciate the function of
first-order consciousness, and relies uncritically on what he
(but not Collingwood) calls "The Principle of Order" (LPC, 28,
52, 93, 105, 167-68). Cf. W. van Leyden in Krausz, CEPC, 27-29.
555
assigning meaning is mostly unreflective and pre-conscious (cf.
IH, 330; Donagan, LPC, 43)33
Obviously this account is very vague, and leaves
much to be said about the fundamental relations of language and
consciousness. We do not know how to characterize this primitive
sort of pre-reflective, discursive meaning-act, or how language
proper with its highly inflective,
_________________
33 Collingwood's remarks in The New Leviathan on first- and
second-order consciousness pose as many problems for an
interpreter as do his remarks on feeling. Our own interpretation
can make sense of many passages, but then others present
problems. For example, Collingwood writes that Hobbes was right
when he said that experience teaches us that it is a vulgar
error to believe that you must first be conscious of a feeling
before you can fit it with a name (NL, 6.56). Unless he means
something idiosyncratic by "naming," this passage (as well as
the discussion surrounding it--NL, 6.42-6.57) appears to support
a radical linguistic thesis like the one Donagan proposes. But
Collingwood provides us with enough clues to overcome his
overstatement. He points out (a) that language is not always
rational (NL, 6.57), and (b) that Hobbes' doctrine is that
"language has become the pre-condition and foundation of
knowledge so far as knowledge is scientific" (NL, 6.47; emphasis
mine). In The Principles of Art, where Collingwood was more
careful in distinguishing symbolic or intellectualized language
from its more primitive variety, he had pointed out this
restriction on Hobbes' discussion of language: "When
Hobbes . . . says that the primary use of speech is for
'acquisition of science,' for which purpose 'the right
definition of names' is the first requisite, clearly, he is
identifying language in general with intellectualized language
or symbolism" (PA, 226). Therefore if one supplies the qualifier
"intellectualized" before "language," in the passage under
discussion from The New Leviathan (NL, 6.56), and "second-order"
before "conscious," the passage can be brought into conformity
to our own interpretation. Cf. IH, 314, 330, where Collingwood
speaks of artistic expression and "unreflective experience."
556
highly differentiated grouping of sounds and meanings can arise
from it, even though he leaves us with the hint that it has
something to do with poesis or "art" (NL, 6.29). We have already
noted that Collingwood's program of "taking soundings" of
consciousness at various depths does not commit him to any
genetic explanation of how one level of consciousness gives rise
to another. In fact in an important chapter he insists that
there can be no laws for the progressive development of mind.
This is stated in the chapter entitled "Retrospect" (Chapter
IX), in which Collingwood sets forth the four principles that he
says he has assumed throughout his discussion of mental func-
tions, one of which (the "Law of Primitive Survivals") we have
already encountered. Before continuing our discussion it would
be best to summarize these principles.

557
TABLE 13
THE PRESUPPOSITIONS OF THE ANALYSIS OF MENTAL FUNCTIONS
1. The Law of Contingency: the earlier terms in a series of

mental functions do not determine the later. The terms in
this case are mental functions called first-order con-
sciousness, second-order consciousness, etc. "Mental func-
tion: means not a single act but a type of activity, and
"series" means an arrangement in which each term is a mo-
dification of the one before it. (NL, 9.3, 9.36, 9.48).
((This law can be restated and clarified in two variant
forms:))
a. The series of mental functions is an irregular series,

i.e. it is not one in which the-development of its terms
is governed by a rule. All mental processes have an
asymptotic or approximative character; they do not have
real initial and terminal points, but always begin with a
first term with a mixture of the second, and end at a
second term qualified by the first. Every case of mental
"being" turns out on examination to be a case of mental
"becoming". (NL, 9.4, 34.58, 34.63).
b. The development of mind is not predictable. The logical

development of a series of mental functions is in-
dependent of its temporal development. In the logical
development of the series of mental functions ABCD (for
first-order consciousness, second-order consciousness,
etc.), the relation between any two successive terms A
and B is such that B renders A necessary or “presupposes”
A as that out of which it develops, while A does not
render B necessary. In a temporal development of such a
series A comes into existence at one time and its
modification B at a subsequent time. But while the
existence of the series ABCD presupposes
___________________________________________________________
NOTE: We take the liberty of assuming that the comments
from Chapter XXXIV cited in (la) are a clarification of the
expression, "irregular series," which cannot consistently mean
"not governed by any rule whatsoever," since at least one rule
is applicable to it, namely that each term is a modification of
the one before.
558
the existence of each of its terms, and each successor

presupposes its predecessor, it is not the case that
given the successor B. A must have come to exist before
it in time, since A and B may have come to exist
together or coexist in the same act. But whether mind
develops from function A to B. or from AB to C, or ABC
to D (a progressive development), or whether it
degenerates from ABCD to ABC, etc. (called regression)
is not predictable; it depends on the practical energy
available to a mind at any given time. (NL, 9.43-9.48).
2. The Law of Primitive Survivals: When ((a mental function)) A

is modified into B there survives in any example of B. side by
side with the function B which is the modified form of A, an
element of A in its primitive or unmodified state. If A be
consciousness and B second-order consciousness or reflection,
then reflection is a modification of consciousness, and unless
there were a primitive survival of mere consciousness there
would be nothing to reflect on, and no reflection would occur.
(NL, 9.5, 9.51, 9.54).1
____________________________________________________________
1 Notice that in this summary of the principles stated in
Collingwood's "Retrospect" there is nothing to correspond with
Donagan's "Principle of Order"; in fact one might argue that
Collingwood's explicit statement that a series of mental
functions can coexist in a single act (NL, 9.48) essentially
conflicts with it. The best evidence that Donagan can cite in
support of his "Principle of Order" is NL, 5.91-5.92, which is
extracted from a discussion of Freud on whether feelings can be
unconscious (NL, 5.8), and is an account of the extent to which
Collingwood agrees with Freud's usage of the terms "conscious"
and "preconscious." Collingwood writes that "no man is conscious
of any given form of consciousness, even though it is operating
in him, until he 'reflects' on it or calls into being in himself
another form of consciousness, C2, the consciousness of C1'
((NL, 1.73)) the form of consciousness with which we started.
Any form of consciousness, practical or theoretical, call it Cx,
exists in what Freud calls a preconscious condition unless and
until it has been reflected upon by the operation of a form
Cx+l'' (NL, 5.915.92), cited by Donagan by references at LPC, 28
and 108). Notice that Collingwood says "form of consciousness."
Now compare this with Donagan's formulation of this subsidiary
principle, which he elevates into one of the four main pre
suppositions underlying
559
___________________________
Collingwood's entire philosophy of mind (LPC), 27): "if a man is
conscious of one of his own acts of consciousness, then it is
not by that act itself, but by another act of consciousness
which may be said to be of a higher order" (LPC, 28). Notice
that Donagan has substituted "act of consciousness" for "form of
consciousness." In our view this is not a legitimate
substitution; what may be true for a relation between forms or
whole orders of consciousness (first-order, second-order, etc.)
may not be true for individual acts of consciousness which may
be at the same level. On Donagan's "Principle of Order" it is
difficult if not impossible to see how there might be a concept
of a concept, or a proposition about a proposition, or inference
about inference; and certainly not without one of them being
reduced to pre-consciousness. But Donagan might have referred to
IH, 292, where Collingwood does say an act of thought may be an
object to another act but not to itself. This might lend support
to his "Principle of Order" were it not that Collingwood
explicitly denies it in the next sentence.
560
It takes no great interpretative powers to recognize that this
description of the relations of the series of mental functions
is a direct application of the "scale of forms" described in the
Essay on_Philosophical Method, and a fairly clear echo of the
description of the dialectical relations between the "forms of
experience" in Speculum Mentis. With this discussion of the
dialectical structure of mind as a scale of forms of mental
functions we find ourselves again confronting the relationship
between F-logic, with its abstract class-concepts, and D-logic
with its overlapping universals.34 From Table 12, 3 a-b, we see
that the abstract class-concept involved in second-order
attention and naming is the sort that is used in
classifying--the practical activity of "drawing the line"
between objects in a field of feeling (or presumably that field
as prepared by first-order consciousness). From Table 13 we see
that the
_______________________
34 It is interesting to note that Collingwood positions
this "retrospective" chapter in the midst of his analysis of the
level of mind he calls "conceptual thinking." Why did he place
it here and not after his chapters on propositional and rational
thinking (i.e. between Part I on Man and Part II on Society)? Is
it not because the "scale of forms" is a discussion of the
relations of concepts in D-logic, and he felt obliged to call
attention to the fact that the series of mental functions (and
notice that first-order consciousness, second-order
consciousness, etc. are concepts or "names" and not judgments,
propositions, or inferences) are dialectically related as a
scale of forms and not merely as abstract class-concepts?
561
level of mental function at which F-logical class-concepts
operate is a whole (i.e. taken as an order or form of
consciousness) related to other levels of mental functioning as
one element in a scale of forms, where each term is a
modification of the one before it, but contains a "primitive
survival" of that previous form.
Since Collingwood indicates that he is not restricting
his discussion to mind only in its capacity as philosophical
thinking, but rather states directly that he is talking about
"the modern European mind" or the mind of a European man (or
that which "has produced in itself the thing called modern
European civilization"),35 we may finally offer a general answer
to a long-standing question concerning the applicability of what
we have been calling the D-logic of the Essay on Philosophical
Method. Insofar as the several orders of functioning of the
human, civilized mind are arranged in a scale of forms, D-logic
is applicable to the relations between these levels. Since
Collingwood asserts in no
______________________
35 Presumably Collingwood would regard other civiliza-
tions (e.g. American and some Far Eastern countries) as exten-
sions of "the modern European mind," since in the later parts of
The New Leviathan he extends the term "civilization" to
include-any manifestation of civil behavior or what he calls
"civility" (NL, 35.63; cf. 34.4-34.51, 34.7-34.79, and 35.22-
35.44)--essentially, the approximation to the ideal of re-
fraining from the use of force in relations with one's fellow
man, and the spirit of agreement to teach and be taught (NL,
35.44, 36.46-36.51). Therefore whatever civilization strives for
such an ideal would be classifiable as possessing what
Collingwood is calling the "modern European mind," which is not
meant to be a primarily geographical epithet.
562
uncertain terms both in The Principles of Art and The New
Leviathan that the series of mental functions are so arranged,
he would not regard this assertion as a tautology (i.e. that
something having the structure of a scale of forms has the
logical relations of a scale of forms). It is what he would call
a universal categorical judgment (in this case, this means that
it is self-instantiating).
What we do not know is what Collingwood would accept as
adequate restrictive conditions for the applicability of D-logic
and F-logic respectively. That there are such restrictions is
evident from the recognition that within a given level of mental
functioning (e.g. propositional thinking) rules which would not
be applicable for the relationship between that global level and
its successor would be applicable for the elements within that
level. Propositional logic and grammatical rules do not employ
the rules of dialectical logic per se, and yet the former are
legitimate and applicable within certain levels of consciousness
(cf. NL, 11.35).
(3) But we are not yet finished with the issue of
"naming," and we must now come to terms with the major dif-
ference between the analysis of mind offered in The New Leviathan
and all his previous writings on the subject. The major focus of
this apparent reversal of opinion is what we may call the
"linguistic-abstraction" thesis, i.e. that at the second

563
level of consciousness and above (everything included by the
term "reflective consciousness") all thought is the result or
product of linguistic activity, and that all linguistic activity
is necessarily abstract.
We have already had to qualify a more radical version of
this thesis (viz. one which extends the linguistic-abstraction
thesis to all consciousness) by pointing out that at the level
of first-order consciousness there is a prereflective,
discursive, meaning-function, so that second order "naming" does
not create meaning ex nihilo, it identifies them by means of
denominating gestures of some sort (speech or the
like)--gestures that call attention (one's own as well as that
of one's audience) to some specifically meant feeling. Compared
to second-order selective attention, first order consciousness
is not a function in which one act is compared or contrasted
with another as a present one compared to one not present, or as
two presented feelings which are compared as what one means as
opposed to what one does not mean by a given expression. And we
have also had occasion to note that Collingwood's account of
language in The Principles of Art leads one to conclude that for
him all meaning is intentional, whereas all symbolic meaning is
both intentional and conventional. But in The New Leviathan we
find him adding that all meaningful language (i.e. speech) is
abstract, and so are also all the expressions of consciousness
which are founded on it.

564
At issue here is what appears to be a major reversal in
Collingwood's final estimate of the nature of thought. In
Speculum Mentis, it may be recalled, Collingwood made ab-
straction to be the ultimate source of all the errors that the
mind makes about itself, and consequently also about its
relation to the world: to abstract is to falsify (SM, 160, 288).
But in The New Leviathan abstraction is seemingly elevated to
the highest rank of thought: conceptual thinking and all thought
above it (viz. propositional and rational thinking) is
abstractive (NL, 7.63). In confronting this pair of claims it
appears that the only two alternatives are either total
scepticism concerning knowledge claims (if all abstraction is
falsification and all intellectual thinking is abstract, then
all intellectual thinking is falsification), or a complete
reversal in Collingwood's mature philosophy (rejection of the
claim that all abstraction is falsification). Since Collingwood
was steadfastly non-sceptical, the latter is the most obvious
alternative (cf. Donagan in Krausz, CEPC, 9-13).
But as we have seen several times over in the course of
our examination of Collingwood's philosophy, the easy and
obvious interpretation is not always the correct one. The
evidence we are confronted with is an apparent contradiction,
not a documented "repudiation"; and an escape from the con-
tradiction may be found, as in other such cases, by a careful
analysis of the meaning of the

565
mediating term--in this case, "abstraction." Our experience thus
far in interpreting Collingwood would support our anticipation
that the term is not being used univocally, and that it has a
layering of overlapping meanings not always specified
explicitly, but discernable from careful comparison of the
opposing uses of the term.
In Speculum Mentis a concept is called abstract if it is
used as a class-unifying term, and if it only defines membership
in the class on the basis of a specifying criterion (the
determining element) but leaves out of consideration all other
qualifying characteristics of the members themselves (the
indeterminate elements)--and this is the sense in which
Collingwood means that something is abstract if it is thought as
"separate" from its members and their individuating differences.
It is the unity of identity of a class "in spite of" (or
indifferent to) the differences of its members. "Physical
object", for example, includes rocks, fish, birds, men, planets,
stars, etc. indifferently, and laws governing the motion of
"physical objects" takes such a disparate and differently
organized group in their most general, generic aspect--as dead
things, pushed and pulled by mechanical forces. But a live bird
dropped from a leaning tower will not behave as a dead rock
dropped from the same height. But for the purpose of the
abstract statement of the law which governs the free fall of
physical objects, the differences between the

566
bird and the rock are ignored: the self-moving capacities
involved in the flight of a bird are not taken into account in
the law.
This is not to say that such a motion cannot be "ex-
plained" in a much more complicated treatment of this situation
(e.g. one which balances the forces exerted by the flying
activities of the bird, generated by biochemical energies
released and used by the mechanical movements of muscles and
bones, etc., against the forces of gravity); but the closer one
gets to an adequate explanation of the differences between the
linear downward acceleration of the rock and the curvilinear
deceleration of the bird, the closer one gets to taking into
account the specific differences between the "dead" rock and the
living bird--and thus approximating to the ideal of a
non-abstract or "concrete" universal, i.e. one in which
differences are essential.
But consider now Collingwood's use of the term "ab-
stract" in The New Leviathan--in this case in reference to
concepts:
The act of attending is not merely a doing something to

yourself, focusing your consciousness on a certain part of
the field and repressing . . . the rest; it is also a doing
something to the object: circumscribing it, drawing a line
between it and the rest of the field . . . . With the
delimiting of the patch or other selection . . goes the act
of evocative thinking: the act of arousing in yourself by
the work of thought feelings you do not find as "given" in
yourself. These I call evocations: they form a context
567
inseparable from any selection and are connected with it by

logical relations, logic being the science which studies
the structure of concepts or, which is the same thing, the
relations between them . . . . As not given but abstracted
from the given, a selection is a product of practical
consciousness; . . . ((it)) is nothing found, it is
something made . . . . A selection together with its
context of evocations is a concept (notion) or a number of
them. (NL, 7.3-7.39).
What sort of "concept" meets the description here given--the
"abstract concept" of Speculum Mentis or something closer to the
"concrete universal"? Is it not the latter that is thought to be
"inseparable" from the context from which it is selected and
bound to it by logical relations? And is it not the abstract
class-concept which is thought of as "separate" or "indifferent"
to its contextual circumstances?
It is possible, if a bit muddle-headed, to decide as an
interpretative comment on this passage that in The New Leviathan
Collingwood had broken with his earlier views on abstraction and
conceptualization, and then foolishly allowed himself to
backslide into "associationism" by defining a concept as "a
selection together with its context of evocations" --which is
associationism insofar as it confuses what is selected with the
concept by which it is selected (Donagan, LPC, 54-55). But this
is no different than assuming that all planetary motion is
circular and then inventing "epicycles" to explain away
retrograde orbits. Not only would the ordinary association
pyschologist not say that

568
a selection is connected with an evoked context by means of
logical relations which preside over the birth of a concept; he
would on the contrary tend to derive all logical relations from
the purely external and coincidental "constant conjunction"
between a selected item and another, and to reduce logical
relations to this constant conjunction.36 Quite the contrary, by
defining “concept" in the way he does, Collingwood is stating as
clearly as he can, without using the term (and hence avoiding
calling attention to its previous use by Bradley, Bosanquet, and
other idealists), that the concept he had in mind was one in
which a universal is bound to a context by logical relations of
necessity, or one to which differences are essential, or in
short the concrete universal.
But even if this were a "lapse" on Collingwood's part,
one would have several other similar regressions to explain in
his discussion of the
_____________________
36 Cf. Howard C. Warren, A History of Association Psy-
chology (New York, 1921), pp. 6-7: "All the writers belonging to
the association school admit the rise of ideas following
sensations, according to the same laws of association that hold
where the antecedent is an idea. Some go further and regard as a
form of association the simultaneous presence of two or more
sensations in consciousness, such as occurs in the act of
perception. Others merely assume a nexus in such experiences
without explicitly classing them as instances of association. On
the other hand, all agree in denying that one sensation can
bring up another sensation by association; it is generally
admitted that the rise of sensations depends on something
outside of consciousness, or at least on something apart from
the individual human experience.”
569
terms "abstract" and "abstraction" in The New Leviathan. In his
discussion of rational thinking, for example, he states that to
place an intention in a context of other intentions that afford
grounds for it (where "grounds" are taken as an intention from
which a consequence follows by necessity) is to make of that
intention an abstraction (NL, 14.32, 14.43). And what holds for
abstraction at the conceptual and rational levels also holds for
the proposition level: "An intention is made into an abstraction
by surrounding it with a context of other intentions; a proposi-
tion by surrounding it with a context of other propositions"
(NL, 14.44).37 Concepts, propositions, and intentions that are
bound by logical relations to contexts of other concepts,
propositions and intentions surely cannot be "abstract" in the
sense that Collingwood had condemned as falsification in
Speculum Mentis.
Unfortunately rejecting one erroneous interpretation of
these passages on abstraction does not mean that we have
successfully resolved all the interpretative difficulties, nor
that we can even be satisfied that we have understood (much less
confirmed) what Collingwood
___________________
37 Cf. IN, 130: "By a real abstraction I mean a real
phase in a real process, in itself, and apart from the subse-
quent phase to which it is leading . . . . Bud and leaf are thus
phases of one process, and the bud in itself is an abstraction
from that process, but an abstraction made by nature, which
everywhere works in this way through successive phases of the
process, doing one thing before it goes on to the next."
570
means by these dark sayings. We suspect that all of this has
something to do with what Collingwood called "real thinking" in
his correspondence with Ryle, as well as the "total imaginative
activity" and "language of total bodily gesture" in The
Principles of Art: all of these are various systematic wholes
presupposed by the elements or "element-types" that go to make
them up. But as with the earlier instances we found of these
systematic wholes, we do not know in anything more than a vague
and abstract way what the relations are between their elements,
how many such elements there are, if they are just variations of
a single form, etc.
In particular we do not know what the limits of de-
terminacy are for abstractions at any given intellectual level.
Collingwood tells us in The New Leviathan that abstractions (by
which we may understand him to mean the result of acts of
selective attention) (a) are determinate in some ways and in-
determinate in others (NL, 7.56, 10.16; cf. PA, 254); (b) are
always indeterminate--i.e. are never wholly determined (NL,
11.54); and (c) are essential for thought beyond the primary
level of consciousness (NL, 7.63).38 Unfortunately he also adds
that
_______________________
38 Cf. Mink, MHD, 108: "Now the characteristics of 'ab-
straction' which Collingwood refers to in different places are
three: selectivity, indeterminacy, and self-determination As a
process, abstraction is the activity of consciousness (also
called 'attention') directed on an object in such a way as to
spot-light some features of the object and neglect others;
571
(d) "abstractions are only second-order objects made by the mind
out of its immediate or first-order objects as naturally and as
unconsciously as bees make honey out of flowers" (NL, 7.67).
What can "naturally and unconsciously" mean here? Collingwood
would probably answer that we cannot know what an abstraction is
until we have made it, and then we can examine the product of
the practical activity which produced it, but only be engaging
in another conscious act (e.g. examining a concept by
formulating propositions about it, and therefore engaging in an
act of third-order consciousness). But isn't "selective
attention" a deliberate act of consciousness, and is it not
selective attention that is the very process by which
abstraction is achieved? How can it then be "unconsciously"
done? And if "classifying" is the practical activity of
consciousness involved
___________________
these features are thus 'abstracted' from the object, and they
can be named and themselves made the objects of other and
repeated acts of consciousness without the repeated presentation
of the object . . . . However because second-order objects have
been selectively abstracted from the totality of activity, they
are indeterminate in all respects except those by which they
have been constituted . . . . Finally, . . . the activity of
consciousness in abstracting is a free activity, determined by
nothing except . . . by itself in abstracting these features
rather than those from its object. . . . . ((But)) it might be
said that the very existence of a verbal language determines the
limits of abstraction: we can attend only to those
distinguishable features of experience for which we have names."
Mink fails to see any paradox in saying that abstraction can be
free and indeterminate and yet determined and limited by names
in language.
572
in "drawing the line" in abstraction, did he not tell us that
all classes are artifacts (NL, 19.22-19.35; see above, Table 12,
3 b)? How then can abstractions be make as "naturally . . . as
bees make honey"?
The difficulty that dogs Collingwood's heels is the same
as the difficulty that plagues any of the usual accounts of
abstraction: as a theory of how we arrive at concepts, it
appears to presuppose what it attempts to explain. Ordinarily
abstraction is described as a process of "leaving out" inci-
dental or extraneous elements in a field of what is being
attended to, until only what is "essential" remains, this being
the abstracted concept.39 Whether this "leaving out" be regarded
as extracting "form" from its original mixture with "matter," or
a process of "separating in knowledge what is inseparable in
fact," the essential act is still a negative or eliminative one
and this raises serious difficulties. What allows the process to
come to a halt at a certain point? Does this not presuppose that
one already has the concept by
____________________
39 Peter Geach, Mental Acts (London, 1957), p. 18: "I
shall use 'abstraction' as a name for the doctrine that a con-
cept is acquired by a process of singling out in attention some
one feature given in direct experience--abstracting it-and
ignoring the other features simultaneously given--abstracting
from them. The abstractionist would wish to maintain that all
acts of judgment are to be accounted for as exercises of
concepts got by abstraction . . . . My own view is that ab-
stractionism is wholly mistaken; that no concept at all is ac-
quired by the supposed process of abstraction."
573
comparison to which the one emerging from this eliminative
process is recognized as comparable and equivalent? And lacking
this prior concept, what is to stop the eliminative process at
anything short of pure nothingness? Do we not. have to
distinguish the selected item from the criterion by which it is
selected (Donagan, LPC, 54-55)? Or to press the juice out of the
matter-form metaphor, if we are "given" something in sensation
or feeling, do we not have to possess something by which we can
accept or apprehend or take hold of the gift?
The closest that Collingwood comes to meeting these
objections is when he engages in what appears to be a frankly
pragmatic maneuver. His basic response to the charge of elim-
inative abstractionism would be to point out that concepts are
not "found" (as naive empiricists might say), they are "made"
(as subjective idealists say), but only by an act. of practical
consciousness prior to theoretical consciousness, and based on
interest or practical concern with the object attended to (as
pragmatists modify idealism) (cf. NL, 7.22, 18.13).40 In taking
a basically pragmatic view of the purposive character of
concepts, Collingwood reaps the benefit
______________________
40 Cf. H. S. Thayer, Meaning and Action: A Critical
History of Pragmatism (New York, 156877 pp. 429-31; and Mink,
MHD, 7-8, 12, 111, 138. Collingwood would probably not be
flattered by the comparison with pragmatism, but in latter years
he appears to be less hostile to this line of thought: cf. IN,
300.
574
of the pragmatic maneuver, which is precisely to eliminate the
need to posit theoretical and fictitious extramental entities
with which the mind is supposed to conform, on intramental
entities with which it is supposed to be innately furnished. But
in contrast to the pragmatists Collingwood appears to be saying
that even though concepts are made for practical purposes, they
nonetheless have logical relations which preside over their
birth within their "context of evocations"; and he nowhere says
that such logical relations are matters of scientific expediency
or economic utility (cf. SM, 182).
This is as far as Collingwood takes us, and we must note
that it may put us down the road a bit from naive empiricism,
but it does not take us the whole way towards a satisfactory
account of abstraction. Saying that concepts are made does not
explain why a concept is well-tailored to the need--or why one
concept "works" while another does not. And we are still not
enlightened as to what these "logical relations" are that
preside over the birth of concepts (although we are told that
"resemblance" is one, and "contrast." and "comparison" are
others (cf. NL, 7.34-7.36), which bodes ill for our defense of
Collingwood against the charge of associationism).
But as we pursue the development of meaning through the
higher levels of consciousness, we shall try to watch for the
way in which a relatively indeterminate meaning is modified by
conceptual,
575
propositional, and rational thinking--or the way in which a
meaning becomes more determinate as it becomes more "abstract,"
just as in The Principles of Art the uninterpreted sensum
becomes reedy for interpretation by being perpetuated and
domesticated by imaginative consciousness, and finally fully
interpreted by acts of intellectual consciousness. We do not
know, on the basis of The Principles of Art and The New
Leviathan, if Collingwood would accept that there could be any
such thing as a concept that is totally determinate, but if
there could be it would not be something abstract, because
something abstract always has an element of indeterminacy (NL,
11.54).
Finally, we shall use the expression "real abstraction"
to refer to Collingwood's description of abstraction in The New
Leviathan, in order to distinguish it from other theories of
abstraction, including the "false abstraction" he condemned in
Speculum Mentis.41 By "real abstraction" we shall understand the
doctrine that abstraction is the process of consciousness by
which a selection is located in its "context of evocations," or
a proposition in a context of other propositions,
_________________________
41 There is an echo of Speculum Mentis in a later chap-
ter of The New Leviathan, where Collingwood writes: "Abstraction
is a necessary part of thought. In thinking of a process of
change you must think of its positive and negative elements in
abstraction from the process. False abstraction is the same
thing complicated by a falsehood, namely, that these two
opposite elements are mutually independent and hostile entities"
(NL, 26.18-26.19).
576
an intention in a context of other grounding intentions, etc.
Having little to go on from Collingwood's cryptic remarks about
this process, we cannot, except by employing inventive energies,
amplify this weak signal concerning abstraction into a more
articulate program for public broadcast. We are particularly
annoyed by the distortion in the original signal insofar as
relating a concept. or selection to other concepts or feelings,
a proposition to other propositions, etc. is not
straightforwardly the same process as subsuming concrete
instances under a universal, or relating a proposition to its
referrent, etc. We still have no adequate criteria for
distinguishing between the referring and meaning functions of
language and consciousness. But. even lacking such criteria for
this distinction we are still forced to recognize a significant
difference between the simple or "false" abstraction rejected in
Speculum Mentis, and the contextual or "real" abstraction
advocated in The New Leviathan.
(4) Collingwood also provides us with very few infor-
mative statements about what he means by "propositional think-
ing" in The New Leviathan. We are told (a) that it is at this
level that truth and error are distinguished; (b) that in
engaging in it we are involved in asking and answering
questions; (c) that asking questions implies contemplating al-
ternatives; (d) that the subject of a proposition is never a

577
first-order object, and (e) neither is the predicate, which is
always a concept. Since Collingwood's primary concern in The New
Leviathan is with the emotional and practical sides of this
level of thinking, it is not too surprising that the theoretical
aspects of this order of consciousness are somewhat neglected.
But for our purposes we cannot ignore it, and therefore we are
obliged to flesh-out this skeleton to the best of our ability.
The way this can be done is to make use of three clues
which he has provided for us: the discussion of the levels of
consciousness as a scale of forms in the "retrospective" chap-
ter; his discussion of "emotions of intellect" in The Principles
of Art; and the placement of both of these in the context of
language as expressive discourse. Making use of these three
clues we may reconstruct what we believe to be a modest version
of what he intended us to understand about this level of
consciousness.
If we begin with the "first stage of mental life" in
which a presented feeling is "apprehended" by a pre-reflective
act of first-order consciousness, the "Law of Primitive Sur-
vivals" allows us to conclude that the primitive act of "mean-
ing" that is involved (what he might have described in The
Principles of Art as a "sensum prepared for interpretation") is
preserved and modified by second-order consciousness

578
by its act of selective (abstractive) attention. In its modified
form, first-order meaning is a designated and named entity (the
first state of reflective consciousness, or what might be called
the first stage in the actual interpretation of a sensum). At
the propositional level meaning is again modified to include the
unity of meaning in an assertion, i.e. the unity of meaning of a
subject and its predicate, where both parts of the assertion are
assumed to be themselves meanings derived from second-order
consciousness. Hence the subject of a proposition or what it is
about is not merely an object as presented to first-order
consciousness, but that object as modified by first- and
second-order consciousness, and now further determined as the
subject of a third-order propositional act.42
__________________________
42 To the great dismay of the reader of The New Le-
viathan, Collingwood is not always as careful as he might be to
indicate exactly what he means when he uses terms like "object"
in "object of consciousness." In the present instance concerning
propositional thought, this leads to considerable unnecessary
confusion. (a) Collingwood first described each level of
reflective consciousness as being conscious of its immediately
lower form, so that first-order consciousness, C1, is the
"object" of second-order consciousness, C2, just as C2 is to C3,
etc. (NL, 1.73, 5.91). Furthermore, (b) he defined "first-order
object" and "second-order object" in purely relative terms, so
that C1 is a "first-order object" to C2, as C2 is to C3; and C1
is a "second-order object" to C3 (NL, 5.26, 5.91); and (c) he
defined the terms "mediation" and "abstraction" (in one of its
senses) in terms of this relative schema, insofar as the
consciousness of a second-order object is mediated by the
consciousness of a first-order object (NL, 6.31-6.34). But then
he muddies the waters when he discusses each successive level of
consciousness. (d) At the level of second-order (conceptual)
consciousness he says that feeling is not an abstraction from
the name of the feeling (NL, 6.35). (e) At the level of
third-order (propositional) consciousness he states that the
subject and the predicate of a proposition are never first-order
objects,
579
_________________________
the predicate (at least) always being a concept. And (f) for
fourth-order (rational) consciousness, first-order objects are
things about which questions beginning "why" must not be asked,
since questions beginning "why" are legitimately asked only
about objects of the second and higher orders, or abstractions
(NL, 14.39). In each of the cases (d) through (f) Collingwood
seems to be saying that the immediately lower level of
consciousness is not the immediate or first-order object to its
successor-which-contradicts (a). Driven by such paradoxical
statements the reader may not be blamed for deciding that
Collingwood is making one of his unannounced shifts of meaning
between his defined, relative sense of the terms "first- and
second-order objects" (b), and what we may call (g) the absolute
sense of these terms, i.e. the sense in which a "first-order
object" would be the object of first-order consciousness, a
"second-order object" for second-order consciousness, etc. But
even if the absolute sense of these terms may make passing sense
of several confusing passages (e.g. the misleading examples
offered at NL, 11.34), we think it would be a mistake. A better
interpretation, in our estimation, would be to recall that in
The Principles of Art Collingwood told us that consciousness has
a double object, both of which are "present" to the conscious
mind (PA, 206). Thus it is possible that a form of
consciousness, say C3 (propositional thinking) may be conscious
both of a lower form C2 (conceptual thinking) and of what he is
calling its "first-order object"--in this case a proposition.
Thus when he writes that the subject of a proposition is never a
first-order object, the reason is that the first-order object of
an act of propositional thought is a whole proposition, and not
its subject. The subject of a proposition is, like the
predicate, a concept, the product of a practical act of
second-order consciousness, and hence a second-order object for
third-order (propositional) consciousness. Finally, it is
important to point out once again that there is a difference
between a "form of consciousness" like propositional thought,
and an "act of consciousness," like the practical act of putting
together a particular sentence. Having said all this, honesty
requires that we admit that we find difficulties on any
interpretation of these difficult passages from The New
Leviathan; e.g. if we invoke the principle of the double object
of consciousness, both of which are present to it, then we are
faced with difficulties in understanding not only what
"abstraction" means (e.g. at NL, 14.214.27) but also our earlier
use of the "double object of consciousness" principle to account
for the radical linguistic claim that naming a feeling makes us
directly conscious of the feeling. If it is possible for a
higher-order consciousness (say C3 to be directly and
immediately conscious not only of
580
On this line of thinking a sentence would be the expression of
the unity of meaning of an act of propositional thought, and
truth and falsity apply minimally to this level of propositional
meaning.
Unfortunately this leaves us without any way to account
for the presence of questions at this level, since questions do
not assert a predicate of a subject, nor are they true or false.
But if questions are interrogative sentences and have something
of a subject-predicate form, should they not belong to the level
of propositional thought? Here we must take up our second clue
from The Principles of Art. Just as we found that there are
broader and narrower senses of the terms "thought" and
"language" for Collingwood, so also we are now obliged to
recognize broader and narrower senses for the terms "truth" and
"falsity," the narrower sense being confined to the level of
propositional thinking and those above it, and the broader sense
extending not only to the level of conceptual thought
(second-order consciousness) but even to the first stage of
mental life, "apprehension" or simple attention. And here we
note that Collingwood not only breaks with the traditional logic
on the subject of truth and falsity, but even from his own
earlier doctrine in Speculum
___________________________
its own object (for C3, propositions), but also of a lower level
of consciousness and its object (C2 and its concepts), then how
can there be a second-order objects (abstractions) at all?
581
Mentis that art as imaginative expression has nothing to do with
truth (or is "indifferent" to the distinction between truth and
falsehood):
This utterance ((of an artist)), so far from being in-

different to the distinction between truth and falsehood,
is necessarily an attempt to state the truth . . . . Now,
if any one thinks . . . that intellect is the only possible
form of thought, he will think that whatever does not
contain arguments cannot be a form of thought, and there-
fore cannot be concerned with truth. Observing that art
does not argue, he will infer that art has nothing to do
with the truth . . . . It is hardly worth while to refute
this argument, by pointing out that truthfulness about
one's emotions is still truthfulness. . . . Art is not in-
different to truth; it is essentially the pursuit of truth.
But the truth it pursues is not a truth of relation, it is
a truth of individual fact. The truths art discovers are
those single and self-contained individualities which from
the intellectual point of view become the "terms" between
which it is the business of intellect to establish or
apprehend relations. Each of these individualities, as art
discovers it, is a perfectly concrete individual, one from
which nothing has yet been abstracted by the work of
intellect. Each is an experience in which the distinction
between what is due to myself and what is due to my world
has not yet been made. (PA, 28788).
In a footnote to this passage Collingwood adds that "I am not so
much criticizing anybody else, as doing penance for youthful
follies of my own," and informs the reader to see his Outline of
a Philosophy of Art and Speculum Mentis.
Now even bearing in mind that in The New Leviathan
Collingwood distinguishes between "linguistic activity in
general" and "literature or poetry or in general art" (NL,
6.29), where the latter is conscious not only of names for
feelings but of the act of naming

582
them, we are still obliged to recognize that there is a
legitimate extension of the term "truth" below the level of
propositional consciousness. If we recall his discussion of
"corrupt consciousness" at this point it is not difficult to see
that "truthfulness" at the first level of consciousness consists
of apprehending feelings as one's own, rather than "disowning"
them, the habit of which is the corruption of consciousness. At
the second level of consciousness, at which "naming" occurs,
truth is conceptual but non-relational; it is concerned with
"the 'terms' between which it is the business of intellect to
establish or apprehend relations. If truth has, as it appears
from The Principles of Art, the structure of a scale of forms,
then we may say that truth, like meaning, appears in the
expression of all higher-order functions of consciousness, so
that there are pre-reflective, conceptual, propositional and
rational truths.
But then what sort of a "truth" does a question have? We
suspect that it is something between the fully propositional and
conceptual levels of truth, but here we are forced to engage in
a bit of reconstructive speculation. Collingwood tells us that
"real" questions are those which offer alternatives, and "bogus"
questions are those which do not; he also tells us that
scientific method depends on replacing unanswerable vague or
confused questions with real questions which can have a precise
answer (NL, 11.12). The "real-bogus" distinction for questions
is something
583
like the "true-false" distinction for propositions, but with
obvious differences which Collingwood does not specify.
"Offering alternatives" is something that propositions may do as
well as questions, but the proposition "X is either Y or Z"
expresses something quite different than the question "Is X Y or
is it Z?" The alternatives offered in a question are not offered
as affirmatively or negatively stated, but rather as
alternatives demanding or requesting or looking forward to some
further linguistic or mental act which chooses between
them--something that will satisfy the question or complete it in
a propositional sense. With a question one is conscious of the
fact. that something is being left unresolved, and a completing
act is being called for.
But just as propositions, like "real" questions, may
offer alternatives,43 so also propositions may be incomplete or
call for resolution: "Eric is going to . . ." is an incomplete
sentence, since the proposition "(x) is going to(y)" contains a
predicate which calls for
_______________________
43 What Collingwood also desperately needs to make sense
of questions, but steadfastly refuses to provide, is some sort
of categorial schema, or categorial designators, to set the
limits of sense and nonsense for alternatives at the
propositional level. "Which of the following two alternatives is
true, A, B. or C?" is a nonsense question, as is "What sort of
train is this, red, yesterday, or singing 'Celeste Aida'?" but.
not for the same reasons. Both appear to be well-formed
interrogative English sentences, but the alternatives offered in
the first case conflict with the conditions presupposed by the
question, and in the latter case conflict with categorial schema
ordinarily presupposed in ordinary English usage.
584
the presence of some value for the second variable (y) in order
to make sense of the whole propositional sentence. But an
incomplete sentence is not a question--it may merely be a
sentence fragment. An incomplete sentence does not express the
speaker's intention to put an interrogative sentence. What is
required is some syntactic and/or semantic function which allows
the incomplete proposition to be a completed interrogative
sentence, so that a hearer (who may be the speaker himself) may
know that completion is not only lacking but known to be
lacking--i.e. that he has expressed his intended meaning in the
form of a question. In a natural language like English such
completing interrogative functions are available either by the
use of "token" pronouns (who, what, how, when, whether, etc.) or
by alternations in word order, as when we put a verb in the
first position in a sentence ("Is x y?" "Does x y?" "Can x y?"
etc.)44
_________________________
44 There may ultimately be no foolproof way of desig-
nating linguistic structures which unequivocally indicate a
questioning intention. "What fools are these mortals?" does not
require an answer, and neither does "How many have fallen in
this battle" when uttered in a circumstance which makes it clear
that no answer is being sought (e.g. in a funeral oration).
Furthermore a questioning intention may be indicated by nothing
more than an alteration in tone of voice, or by stressing
certain words or syllables: "These apples are fifty cents each?"
Of course we have no way of knowing, but it may be
considerations like these that inclined Collingwood to emphasize
questioning as an intentional or conscious act rather than a
linguistic entity in his Q-A logic.
585
Although such considerations as these are essential to
the description of interrogative sentences in a language (which
some may find to be a more fruitful line of inquiry),
Collingwood would probably relegate it to the "grammatical
analysis of language" about which he had such unflattering
things to say in The Principles of Art. That his concern was not
in this direction is indicated by the fact that he says nothing
about such syntactical and semantical requirements in any of his
writings. His interest was in what we might call the heuristic
aspect of questioning as an intellectual function of
consciousness. In this respect his main concern was to show that
questions have not only a logical relation to assertions,
presuppositions, and other questions in a systematic inquiry
(explored by Q-A logic) but that the reason questions can stand
in such logical relations is that they are expressions of a
certain level of consciousness, and therefore "constituents" of
mind, achievements of thinking agents, and part of the emotional
life of persons.
Therefore when he says that "real" questions are those
which offer alternatives, this is not the entire story. A "real"
question is also a "truthful" question which truly expresses the
intellectual emotions of a conscious agent--emotions like
curiousity, wonder, interest, etc.; the emotions of inquiry. A
question therefore stands not as an atomic entity, but in a
multiplicity of relations--logical, epistemological, emotional,

586
and perhaps others. Its "truth" or truthfulness is a complex
function depending on which context it is seen to belong in. Q-A
logic is an exploration of questions in the context of other
questions, presuppositions, and answers as employed in
systematic inquiry. The requirement of offering alternatives
therefore merely states what a question must be to be fully
determinate at the level of propositional thought--i.e. for a
question regarded as an epistemological entity. Other
requirements may appear from the grammatical analysis of
language, and still more from the logical analysis of language,
the latter governed by the rules which preside (to extend
Collingwood's metaphor) over the birth of questions and
propositions by selection from a context of propositional
evocations--P-Q-A complexes bound together in a systematic
inquiry.45 (Cf. NL, 4.3-4.36).
(5) If this reconstruction of Collingwood's intended
analysis of third-order consciousness is correct, there is
nothing to prevent us from extending it to fourth-order con-
sciousness as well. By the "Law of Primitive Survivals" meaning
survives even at the level of rational discourse, but at this
level the meaning-seeking function of consciousness
______________________
45 In Q-AM Collingwood is careful to point out that in
scientific inquiry when a question is answered it does not cease
to be a question, but only an unanswered question (EM, 24; see
above, Table 9, Comment 4). We might add here that an
intellectual emotion expressed in a question does not cease to
be an emotion once the question is answered, but rather ceases
only to be an unsatisfied emotion.
587
is transformed into a unifying act in which one propositional
meaning is related to another as ground to consequent--i.e.
validity in inference. And since questioning survives at this
level also, the questioning function at the level of rational
discourse is the expression of the anticipation of meaning
fulfillment in a validity relationship, i.e. that one
propositional meaning actually will be related to another as
ground and consequent.
Fortunately we do not have to rely quite so heavily on
our own reconstruction at this level, since Collingwood provides
us with a few more informative remarks than he does for
propositional thinking. Even a cursory reading of his discussion
(Table 12, 5) makes clear that Collingwood is not reducing
rational discourse to the F-logical functions of implication: it
is not a "truth-functional" relation that exists between
propositional forms that he is concerned with, but rather the
reason or reasons for thinking that something is the case or
that something should be done--i.e. intentional inference rather
than strictly formal implication. Collingwood unmistakably makes
the point when he insists that "reason is always essentially
practical; because to be reasonable means to be interested in
questions beginning with 'why'; and this happens because people
crave reassurance against the

588
fallibility of their knowledge" (NL, 14.31).46
But how do we know when one proposition or
intention is the ground for another, or that one is a
consequence of the other? Collingwood's answer brings us once
again before his idiosyncratic usage of "abstraction." "One
intention supporting another both form part of the same
intention, which includes them both, and perhaps other things"
(NL, 14.4). The relation of ground to consequent for Collingwood
therefore appears to be the relation of part to whole, and is a
relation
___________________
46 In the passage just proceeding the one cited Collingwood
asserts that "first-order objects are things about which
questions beginning 'why?' must not be asked," since "such
questions are legitimately asked only about objects of the se-
cond and higher orders (abstractions)" (NL, 14.39). Once again
we are up against the application of his "Fallacy of Misplaced
Argument" (NL, 4.73) which forbids us from arguing about any
object immediately given to consciousness. Unless we bear in
mind that the "first-order object" of fourth-order (rational)
consciousness is a complex intention or proposition containing
at least two intentions or propositions, one being the ground
and the other its consequent, the proscription about
why-questions seems paradoxical; since fourth-order con-
sciousness is reflection on third-order (propositional) con-
sciousness, it seems that it would be perfectly legitimate to
ask why-questions concerning first-order objects of this sort
--e.g. "Why should I mow the lawn?" is a question about the
proposition "I should mow the lawn." But a proposition is a
first-order object for third-order consciousness, and a second-
order object for fourth-order consciousness, so the rule holds.
But notice that the "objects of the second and higher orders"
puts "abstraction" in a downward occurring attitude ("higher"
merely indicating the numerical ordinals for orders of conscious
objects--second-order object, third, etc.). For fourth-order
consciousness, therefore, a proposition is an abstraction from
(or second-order object to) its first-order object, which is a
ground-consequent complex proposition, just as for third-order
consciousness a concept is an abstraction from its first-order
object, a (simple) proposition. (Cf. NL, 7.67).
589
of inclusion (cf. NL, 14.41). To demand confirmation of a
larger, inclusive intention or proposition would be, he says,
"to place it in a context of other intentions or other pro-
positions that might afford grounds for it; that is, to reduce
it to the level of an abstraction," or in other words, "to make
it an abstraction" (NL, 14.43). It would be, to extend the
terminology drawn from the level of conceptual thinking, to
place an intention or proposition in a "context of evocations"
of other grounding intentions or propositions within which it is
included and to which it is related as part to whole.
And here once again we touch on not only what Colling-
wood had called, in his correspondence with Ryle, "real think-
ing," in any example of which would be included all the
"element-types" on which logical relations are founded, but also
the pragmatic aspect we have already noticed in his analysis of
the functions of mind (e.g. when concepts are "made" by an act
of practical consciousness). For fourth-order consciousness both
of these features are brought into play in a chapter on
"Theoretical Reason":
In all forms of rational thinking a distinction is made

between the self and the not-self. Such thinking is pri-
marily practical; its first function is to ask and answer
the question: "Why am I doing this?" It has, however, a
secondary function, to ask and answer questions about what
is not myself. These may be called "theoretical" questions;
but they are never purely theoretical . . . . They arise
out of practical problems concerning the self and other
things . . .
590
Consider the place of experiment in natural science. An

experiment means an interference by a natural scientist
with some process of nature. The "experimental method" in
natural science is the method wherein a scientist comes to
understand a natural process by interfering with it . . . .
Is there nowhere such a thing as "purely theoretical
thinking"? There is; but it is not real thinking, and it.
does not lead to real knowing . . . . Real thinking is
always to some extent experimental in its method; it always
starts from practice and returns to practice; for it is
based on "interest" in the thing thought about; that is, on
a practical concern with it . . . . A man will have a
different theoretical attitude towards things other than
himself according as his practical attitude towards them is
different; and his practical attitude towards them will be
different according to differences in his attitude towards
his own actions. (NL, 18.1-18.2).
There are many avenues to explore in this extremely interesting
passage and the chapter from which it is drawn--e.g. the
"experimental" aspect to all rational thought and its aspect of
"interference," which reminds us of his remark from the
Autobiography concerning his interest in "obscure provinces"
because of the challenge to invent new methods for studying them
(A, 86). But we wish to note here that we have a new dimension
to add to the logical aspect of "real thinking" that Collingwood
had divulged in his letter to Ryle: in addition to (a)
containing in any example of itself all the "element-types" on
which logical relations are founded, real thinking (b) is always
to some extent experimental in its method, (c) starts from
practice and returns to practice, and (d) is based on interest
or practical concern with the thing thought about. Furthermore
one can dimly discern that (e) it depends on differences in
"attitude" towards
591
--such differences as are involved in the three different ways
that a modern European answers the question, "Why am I doing
this?"--viz. because it is useful, right, or my duty (NL,
14.65-14.69). Since each of these gives rise to a particular
world view (utilitarian thinking giving rise to teleological
Greek science, regularian thinking to classical modern science
with its laws of nature, and duty to historical consciousness
and its transformation of contemporary science (NL, 18.318.92),
it is not difficult to see that such "attitudes" are what
Collingwood calls "absolute presuppositions" in the Essay on
Metaphysics.
Since this gets us beyond logic and into Collingwood's
views on metaphysics, we must stop at this point, with the
promise that w% shall resume the discussion in Chapter X, where
we shall do our best to clarify these views. But we must. make
one final remark before bringing this section, and chapter, to a
close. We wish to call attention to the fact that. questioning
functions at both the third- and fourth-order levels of
consciousness, and that at the fourth level (as might be
expected, given the Law of Contingency for mental functions) it
becomes a why-question which presupposes both that there is a
level of propositional thinking (and all that this presupposes),
and also that there is a relation of ground to consequent that
can be established for propositions within this level. This is
an important observation for any one's own actions

592
evaluation of Q-A logic, since questioning traverses the dis-
tinction between simple propositions (minimal units of truth and
falsity in the narrower sense) and complex propositions
(propositions related by logical connectives, and minimal units
of truth-functional validity).
6. Conclusion.
We have come a long way since initiating our discussion
of Q-A logic at the beginning of this chapter, and it is now
time to try to pull the strands of our investigation together.
Unfortunately to try to summarize further what is already a
summary would be to virtually repeat the chapter, so instead we
shall limit our remarks to several observations on what we think
are the more important central features of Collingwood's
discussion of logic.
(1) In the introductory section we recalled the work of
previous chapters in which we showed that Q-A logic is re-
cognized not only by Collingwood but also by his principal
interpreters to be one of the major features of his overall
philosophical outlook. This is explicitly stated by Collingwood
in his Autobiography, and in subsequent interpretation of his
thought one of the major controversies concerning his philosophy
concerns the role of absolute presuppositions in metaphysics
--presuppositions being part of the unique P-Q-A complex defined
in Q-A logic.
593
We also recalled that in previous chapters we found that
although in the Autobiography Collingwood presented Q-A logic as
an alternative to F-logic and fails to mention D-logic at all,
his early writings contrast F-logic not with Q-A logic but with
D-logic. In this chapter we proposed to examine the later
writings to see not only if there is evidence supporting Q-A
logic as described in the Autobiography, but also to find
whatever enlightenment we could about the relationship between
these three logics in Collingwood's philosophy.
(2) When we took a close look at Q-A logic as it appears
in the Essay on Metaphysics (Q-AM) we noticed that the P-Q-A
complex remained central to his observations about logic, and
that the elements of this complex retained a logical
relationship to each other that is not merely reducible to a
relationship of mere co-location in subjective consciousness.
However we noticed also that there were several differences
between Q-AA and Q-AM, the most notable being the fact that Q-AA
described the Q-A complex (or perhaps that complex as extended
to include presuppositions, since every question has at least
one presupposition) is the unit of meaning, truth, and validity,
but in Q-AM meaning and validity are not discussed, and truth or
falsity is assigned primarily to propositions (a point which
seems to directly contradict Q-AA, which denies that truth or
falsity is a property of propositions as such).

594
Using this shift as a starting point for our subsequent
investigation, we set out to examine the extent to which Q-A
logic in either of its versions could meet the Q-AA claim to be
an alternative to F-logic. We found that when analyzed with this
claim in mind, far from being an alternative to F-logic, Q-A
logic seems on the contrary to presuppose it, or to presuppose
whatever F-logic presupposes; meaning, truth, and validity are
establishable independently of the P-Q-A complex, and this
independence is a necessary condition for deciding what counts
for something to be a meaningful element in that complex, what
can be true or false in it, and whether its elements are related
in a valid or invalid way. Since from a logical point of view
meaning is minimally a function of terms, truth a function of
propositions, and validity a function of complex propositions
linked by logical connectives, we deemed it unlikely that the
P-Q-A complex could minimally meet the specifications of all
three as Q-AA claimed it could. Furthermore Q-A logic fails to
sustain itself as a systematic structure (at least in its Q-AA
and Q-AM versions) since it fails to meet the requirements for
such a system--viz. consistency, completeness, and formality.
But if Q-A logic fails as an alternative to F-logic, and
can never hope to replace it as Collingwood proposes it, we find
that it has
595
nevertheless a unique logical structure of its own, something we
barely began to explore rather than conclusively demonstrated.
Since Collingwood recognized at least a significant part of this
structure (the central P-Q-A complex, the non-deductive
relationship existing between presuppositions and their
questions and answers, etc.) we speculated that perhaps the
comparison with F-logic was misleading, and that he had a
different intention in mind in describing it as an alternative
logic. Since so much of it is described in epistemological
terms, we proposed to explore Q-A logic in its relationship to
mental acts as described in his later philosophy of mind--the
three acts involved in grasping the meaning of terms, the truth
or falsity of propositions, and the validity of inferences.
(3) The first step in this direction was the examination
of D-logic as explicated in the Essay on Philosophical
Method--which in its format suggested itself to us as a better
candidate than Q-A logic for being an alternative to F-logic. We
found that the Essay described characteristics of the
philosophical concept, judgment, and inference vis-a-vis the
non-philosophical concept, judgment and inference, and in so
doing it described the structure of a D-logic as opposed to what
we recognized to be an F-logic of an Aristotelean pedigree. With
its overlapping classes, related in a scale of forms in which
there is a "fusion" of differences of degree, differences of
kind, relations of
596
opposition and relations of distinction, the philosophical
concept of universal differs from the class-concept employed by
"exact and empirical sciences," or from that concept in its non-
or prephilosophical employment. We noted that Collingwood makes
a point of not stating this distinction in a way which excludes
alternative presence of concepts which meet D-logical require-
ments, e.g. scalar overlap: concepts with such a structure ap-
pear in ordinary experience as well as exceptionally in science.
So also with philosophical judgments which, like those of
science, are universal, but are also "categorical" (or have
objective reference) rather than being merely hypothetical, and
inferences, which are "reversible" rather than being merely
deductive or inductive.
Unfortunately we found several obstacles in the way of
understanding Collingwood's discussion of universal categorical
judgments and reversible inferences, and as an aid to
understanding the views of the Essay on these matters we turned
to the letters between Collingwood and Ryle on the subject of
logic, and made some remarkable discoveries. We found
Collingwood confessing to a view of logic that sets it in a
frankly epistemological context, so that the deductive struc-
tures of F-logic are dependent upon or presuppose what he calls
"real thinking," in any example of which is contained all the
"element-types" of logical relations explored by F-logic, and
which the class-concepts of F-logic presuppose.

597
In the Essay and the ensuing correspondence with Ryle, then, we
found Collingwood arguing in something of the fashion of a
later-day Kantian, defending "transcendental deduction" in
arguments, judgments that are synthetic and a priori without
being merely hypothetically true, and universal concepts that
appear to be similar to Kant's schematized categories.
Noting that to pursue this discussion further would take
us into Collingwood's views on metaphysics, we observed that the
argument of the Essay culminates in a crucial chapter on
philosophical inference, which maintains that philosophical
arguments can be reversible without being viciously circular--a
view which puts great pressure on the assertion that philosophy
is a "self-justifying" enterprise. But this is only possible
insofar as the conclusions of philosophy are "established" by an
experience that anticipates them beforehand, so that the
anticipating experience establishes the principles used to draw
such conclusions, and the principles reciprocally "establish"
the conclusions. The non-circularity of philosophical inference
depends on the fact that the "establishing" of conclusions by
principles is done by a higher form of thinking, one for which
the lower states of thinking are "experience" to its
"conclusion." Philosophy (which must be one form of what he had
called "real thinking" in his correspondence with Ryle)
presupposes that experience is already somewhat systematic, and
it is only because of this

598
that its arguments (and the argument of the Essay as a whole)
escapes vacuous circularity.
Furthermore we noticed that although D-logic in the
Essay has the appearance of meeting the systematic requirements
of F-logic for consistency, and completeness (discussed in the
Essay under the rubric of "system") but not formality, the
claims of consistency and completeness are premissed on
Collingwood's providing convincing and unambiguous instances of
D-logical analysis. And once again we noted that D-logic, just
as had Q-A logic (and we can now say, on the strength of the
Essay and the Collingwood-Ryle correspondence, F-logic as well)
depends on a full-blown epistemology which is not provided by
the Essay itself. This we proposed to examine in the next two
sections in connection with the higher functions of
consciousness in the philosophy of mind as discussed in The
Principles of Art and The New Leviathan.
(4) Taking up The Principles of Art first, we noted that
in his discussion of levels of consciousness, the function of
expression is minimally essential, and language, discourse, and
the grammatical and logical analysis of language are all located
on a "scale of forms" of expression. Now although there is
nothing in The Principles of Art that could pass for a
thoroughgoing examination of the relation of language and logic,
and although the higher regions of intellectual consciousness
remain in this work largely unexplored, we found that defining
language
599
by means of a scale of forms of expression has important
consequences for Collingwood's view of logic.
The first of these consequences is that Collingwood
insists that the minimum specification of the genus, i.e. ex-
pression of emotion, is never entirely lost when it reaches the
level of symbolic language. Unfortunately Collingwood does not
support this point with convincing evidence or argument, and
does not develop it very far. But we recognize that Collingwood
struck a rich vein when he proposed that there are "emotions of
intellect" which are expressed by intellectualized
language--emotions which we recognized (through his example of
Archimedes--improbably as he stated it) to be those involved in
intellectual activities involved in the process of discovery,
for example--emotions like curiousity, interest, wonder, etc.
And we suggested that the "anticipation" of answers in question
is an expression of an intellectual emotion, but that this also
remains unspecified by Collingwood. Even though this theme
remains undeveloped in his writings, we noted that even at the
highest level of language (which would have to include the
grammatical analysis and F-logical extension of language) it
retains its continuity with the emotional life of a conscious
agent--a thesis which appears to run directly counter to the
formalistic claims to autonomy by most contemporary logicians.

600
We also examined several other secondary consequences of
defining language within a scale of forms of expressiveness. One
is the location of the grammatical and logical analysis of
language in a context which renders the more extreme forms of
their claims less credible. The belief that language is a
completed "thing" which can be sustaining pieces (words,
phrases, sentences, etc.)i.e. the assumption made in the
"grammatical analysis of language,” is a mythical claim that is
discredited when it is recognized that the "language of total
bodily gesture" escapes grammatical reduction to lexicographical
entries and grammatical rules. Similarly the "logical analysis
of language" relies on certain presuppositions (the
propositional assumption, the principle of homolingual
translation, and the principle of logical preferability), which
may suit the logician's purpose to transform language into a
perfect vehicle for the expression of thought, but must be
recognized as actually a proposal for the modification of
language rather than a statement of what language actually is.
In this context F-logic appears in the guise of what might be
called today an "ideal language" constructed for the purpose of
modifying a natural language to rule out the frustrations of its
obstructing and misleading meanings.
Finally, we noticed that there is a global or compre-
hensive aspect to each level of the scale of forms of expression
as Collingwood describes
601
it in The Principles of Art, so that each has its own principle
of organization which is presupposed by its successor on the
scale. We saw a connection between this global or summarizing
aspect. and what. he had called the "logical structure of real
thinking" in his correspondence with Ryle, as well as with the
"total imaginative activity" he discussed in connection with the
mental life of the artist--in each of which there is a
"whole-part." dialectical relationship. We suggested that
perhaps logic is a systematic study of the relations existing
within a given level of this scale, so that there is a logic of
concepts, propositions, and inferences, which is nonetheless
distinct from the logic of the overall relationships between
these levels themselves --the former explored by F-logic, the
latter by D-logic. (In a later section we noted that Q-A logic
may transect both F-logic and D-logic--but about this
Collingwood says nothing.)
(5) The last phase of our examination of Collingwood's
views on logic brought us to The New Leviathan, where we pursued
the relationship between language and logic into Collingwood's
analysis of the levels of mental functions. In so doing we found
that here too a "scale of forms" appears in which various orders
of consciousness are related in the manner that the forms of
expressiveness had been in The Principles of Art, and the
division of these levels (conceptual, propositional, and
rational
602
thinking) parallels the division of the subject-matter in the
Essay on Philosophical Method. Logic, language, and
consciousness appeared, therefore, each to have a parallel
structure that meets the D-logical description of a "scale of
overlapping forms" described in the Essay.
But we also found certain obstacles and peculiarities
which make an exact parallel problematic. The most noteworthy of
these obstacles is what we called the radical linguistic
thesis--that from the level of conceptual thought and above all
thinking is a linguistic function, so that language determines
thought and not vice versa. Some of the more glaring
contradictions on this issue between the assertions of The New
Leviathan and his previous writings can be cleared up, we found,
by careful attention to functions of consciousness which are
more fully described in his earlier writings--e.g. the function
of first-order consciousness in "apprehending" or "appreciating"
feelings before naming or classifying them. Still other
discrepancies concerning language are similarly softened by
attention to broader and narrower senses of terms like
"language," "thought," "knowledge," etc., each of which appears
at various levels in a scale of forms in which it appears. In
passing we noted that even "truth" and "meaning" are defined by
scales of forms, and consequently some of the apparent
discrepancy between Q-AA and Q-AM can be cleared up by
recognizing that there is "truth" even in concepts, so that it
is the narrower sense of the term that applies to proposi-

603
tions alone (as in Q-AM); and there is some sense even in the
Q-AA assignment of meaning and truth to Q-A complexes (for even
questions have their "truthfulness").
Other problems are less easily resolved, and in dis-
cussing them we found several novel concepts in The New Leviathan
which are revealing, but remain, like some other themes we have
turned up on this chapter, undeveloped. Two of these were (a)
the relationship between what we called prereflective "meanings"
at the level of first-order consciousness, and the function of
naming at the second level; and (b) the paradoxical notion which
we took the liberty of calling real (or concrete) abstraction
which appears at the second and all higher levels of
consciousness. What emerges from this discussion is an
idiosyncratic but interesting view of the relationship between
the various levels of consciousness which his "Retrospect"
chapter explicitly relates to the D-logical structure of a scale
of overlapping forms. Therefore rather than "repudiating" his
earlier views on the subject of thinking, consciousness, and
abstraction in Speculum Mentis (as Donagan maintains) we find
that much of his later work is an expansion and further
examination of the dialectical views expressed in that earlier
work (as Rubinoff and Mink have argued). Everywhere we find
D-logical structures and arguments, and everywhere we find
Collingwood pointing out the limitations of an abstract,
class-concept oriented F-logic in the analysis of language and
conscious mental acts.

604
Even his discussion of the linguistic aspect of the analysis of
mind is the development of a hint we found him (in Chapter VI)
making as early as Speculum Mentis, where philosophy is viewed
as "translation" into a language where error is not eliminated,
but reduced by recognizing the metaphorical nature of language.
But in the final reckoning, what can we say about the
autobiographical interpretation on the issue of Q-A logic? We
have found much that supports Collingwood's claim that from an
early date he argued that Q-A logic is a necessary antidote to
the tendency of realists to interpret everything in the
F-logical terms of what he calls the abstract class concept, and
in this sense he sought even in his later writings to defend an
"alternative" point of view, one which employed the principles
of Q-A logic with its characteristic P-Q-A complex. But we have
also found that he modified his logic as he developed his
thinking, and the guiding principles shaping the major lines of
this development were not those of Q-A logic either in its Q-AA
or final Q-AM version, but rather those spelled out as D-logic
in the Essay on Philosophical Method.
Does this mean that we must admit that the Autobiography
on this issue fails to be vindicated as an interpretation of his
later philosophy? Our answer in all honesty must be ambiguous:
yes and no. Yes, in

605
the sense that the Autobiography fails to provide us with a
clear sense of the actual relationship in Collingwood's thought
and writings between what we have been calling Q-A logic,
F-logic, and D-logic, all of which appear to operate in his
thinking and are, in one way or another, legitimate logics (in
some sense of that abused word, logic). But also no, in the
sense that in the Autobiography Collingwood (a) pointed to the
Essay on Philosophical Method as his "best work in matter" (a
clear indication that even at the time of writing of the
Autobiography he felt that the D-logical structures described in
the Essay were sound), and (b) related F-logic to realism,
against which a good part of his thinking was directed.
Thus although we cannot reasonably expect that the
Autobiography should say everything that could or should be said
about his views on logic and their role in his own philosophical
development, we believe that in the matter of Q-A logic it
remains in general (if not in every detail) a valid
interpretation of what he subsequently achieved, and certainly
more faithful to that achievement than any of the non--
autobiographical interpretations that have been offered thus
far. We suspect that if he had lived longer he may have pursued
a resolution of the remaining ambiguities about meaning, truth,
and validity; he may have developed the notion of the
"categorical" or referential aspect of judgment more fully,

606
perhaps distinguishing it from the meaning function of language;
and he may have written a more detailed epistemology to provide
us with a key to unlocking the Pandora's box of element-types
contained in his "real thinking." These themes and more remain
dangling threads in the unfinished carpet in which interpreters
like Mink and Rubinoff see a common figure.
But as we have not yet exhausted all that Collingwood
has said of significance on these topics in his later writings,
we may yet find more light shed on what sort of thinking he
accepts as "real." One candidate may be provided from the
writings that resulted from his lifelong defense of history as a
philosophical subject matter; a second from his heretical views
on metaphysics. The first is involved in his rapprochement
between history and philosophy which in the Autobiography he
claimed to be his major goal as a philosopher-historian; the
second is the extension of this endeavor at reconciliation to
the central battleground of philosophy--metaphysics. In Chapter
IX we shall examine the first, in Chapter X, the second.

CHAPTER IX
THE PHILOSOPHY OF HISTORY
1. Introduction.
In our traversal of the later writings we come now to
the third great theme of the Autobiography, and the pattern of
the last six chapters indicates sufficiently what our next
general tasks must be. First, we seek evidence that
Collingwood's autobiographical interpretation of his development
is in its main lines carried out in his later writings on the
issue of the relationship of philosophy and history. Secondly,
we must take account of the major developments of his later
thought insofar as these impinge on our present subject--i.e.
the positions on logic and mind that we have been examining in
the previous chapters of Part III. Thirdly we must take up the
problems left unresolved on this subject from his earlier and
later writings, paying special attention to those aspects of it
which give the appearance of major reversal of opinion either in
conclusions or presuppositions, but especially the latter. We
may expect that in pursuing these general goals we shall see how
Collingwood responds to some of the classical issues that arise
in any discussion of the philosophy of history.
607
608
But in carrying out these tasks we encounter a special
textual problem that does not arise prior to this point in our
investigation. When it comes to the philosophy of history we
cannot rely on simply outlining the conclusions provided by the
evidence of his later published writings and then comparing them
to those of the Autobiography. The problem is the nature of the
evidence. Collingwood did not live long enough to complete The
Principles of History, which he had hoped to be his major
contribution to philosophy (IH, v-vi). What we inherit instead
is the posthumous publication, The Idea of_History, which is the
result of the editorial labors of T. M. Knox. In his preface to
that work Knox states that he assembled the resulting manuscript
out of materials from three sources: (a) a set of thirty-two
lectures on the philosophy of history, written during the first
six months of 1936 and revised in 1940 in preparation for
publication (note that this revision is after he wrote the
Autobiography); (b) various lectures and essays written by
Collingwood between 1934 and 1939; and (c) a 1939 manuscript
consisting of roughly the first third only of his incomplete
work, The Principles of History. As this material is passed to
us through Knox in the form of The Idea of History, from (a) we
are left with an Introduction and Parts I-IV, which consists of
the revised 1936 lecture survey of historiography from ancient
times to the early 20th century; and from (b) and (c) and some
material from (a) we get Part V, the "Epilegomena," a set of
terminal essays outlining

609
Collingwood's own philosophy of history. What remains is
Collingwood's work, but as arranged and selected by Knox. The
question arises, does this editorial arrangement reflect
Collingwood's autobiographical interpretation or Knox's
editorial reinterpretation?
Our problem is that we do not have access to the same
material as did both Knox when he wrote his preface and Col-
lingwood when he wrote the Autobiography. Although Collingwood
had authorized publication of the whole of (c), Knox saw fit to
include only those three excerpts that appear in The Idea of
History (IH, vi), along with a few very controversial quotes
from it in Knox's preface. The rest is now lost. We have already
found Knox's judgment faulty on this matter, not only because it
is inconsistent with his own editorial policy in publishing the
companion volume, The Idea of Nature, but also because it
conflicts with Collingwood's explicit statements in the
Autobiography. Critical historiography would demand that when
two authorities with access to the same materials disagree, one
is obliged to suspend judgment until further evidence or
convincing argument can be found to resolve the dispute. On
these grounds alone it seems we are forced to fall back on more
probable narrative arguments to reconstruct his thinking on the
matter.
610
But we are not obliged to remain at this impasse, and
this is precisely where our first general task blends with our
second and third. In our own survey of Collingwood's development
we have found ample indication to support his thesis in the
Autobiography that his starting point was located in his
criticism of the errors of realism. In Chapter III we found
Collingwood linking the errors of realists to the neglect of
history. These errors involved (among others) (a) the view of
the past as so many dead, atomic events--the "data" of history;
(b) a positivistic application of the methods of natural science
to history, based on the false analogy between historical and
physical events; (c) the assimilation of historical knowledge to
natural science (culminating in historical positivism and
philosophical psychologist); and (d) an abstract separation of
knowledge of fact from knowledge of self. The correction of
these erroneous views involved developing his thinking to
embrace (a) the notion of the past as a living process of
becoming, leading to the present and surviving as elements in
that present (ultimately becoming his doctrines of re-enactment,
Encapsulation, and historical evidence); (b) the concept of
history as the science of human affairs, (c) with methods of its
own not reducible to abstract deductive logic, but rather
employing a Baconian Q-A logic for testing evidence; and (d) a
philosophical rapprochement with historical understanding
insofar as knowledge achieved by historical inquiry is knowledge
of the historian's own situation which is at the same

611
time knowledge of himself, i.e. philosophy. In this issue we
therefore find joined together all of the principle themes we
have found to be the guiding ideas in the autobiographical
interpretation of Collingwood's philosophy--themes which we have
thus far confirmed to be the leading ideas in all his early and
later published writings.]
In criticizing the errors of realism concerning history
we also noticed at least one of the major issues from his later
philosophy beginning to emerge. In Chapter VI we saw a gradual
refinement of the
__________________
1 If we can ever expect to find a suitable instance of a
form of thought displaying the "real thinking" which binds
together in a single example all that Collingwood leads us to
expect of it, we have good reason to suspect that it will be
found minimally exemplified at the historical level in the scale
of forms of knowledge. Cf. per contra, Mink and Rubinoff. Mink's
"figure in the carpet" (roughly the overall point of view which
unites Collingwood's philosophy) is first encountered in art,
which is "the basic form of the figure in the carpet to which
every part of the rest of the design is related" (MHD, 237). For
Rubinoff this would probably not be art but religion, because it
is only with religion and especially the Christian religion with
its doctrine of redemption, that theory and practice are unified
and the theme of reconciliation appears explicitly (CEPC, 106;
cf. IH, 314-15). While each of them makes an impressive case in
defense of their own interpretation, we wish to note that (a)
art and religion hardly come in for mention in the
Autobiography, whereas there is a sizable discussion of the
rapprochement of history and philosophy; (b) even in the early
writings like Speculum Mentis, "concrete thought" (thought
dealing with individual facts and employing the concrete
universal) does not occur prior to historical thinking; and (c)
in his later writings the most noteworthy exercise of Q-A logic
and the use of philosophical method of the sort he described as
"real thinking" takes place in his discussion of history.
612
concept of history from its more realistic form in Religion and
Philosophy (history dealing with "facts" independent of anyone's
understanding of them) to the sort of view he sketched in the
Autobiography. We noted that in his early writings we could find
direct evidence that the principles of history we called LG and
ARCH 1, 2, and 3 are actually formulated in his essays on the
philosophy of history. But although some of the principles we
called HIST 1-4 were indirectly detectable or exemplified in his
earlier writings, they are explicit and literal only in his 1936
British Academy lecture, "Human Nature and Human History," at
which point the development of his thought on the subject as
outlined in the Autobiography may be said to be complete.
Between these two points we noted that there was a gradual re-
finement of the concept of the object of history from one which
(a) (in Religion and Philosophy) merely identified its object
with factual becoming or the whole of changing reality (hence
identifying history and philosophy as "the same thing"), to (b)
(in Speculum Mentis) a form of knowledge directed on an object
grasped as "the concrete universal," expressed in categorical
singular judgments, but limited insofar as it sought to grasp an
infinite world of fact which lies outside the consciousness of
the historian; and finally, to (c) (in the Essays on the
Philosophy of History written between 1921 and 1930) a concept
of history
613
as a multi-leveled structure in which, in its most precise form,
its object is specified as the deeds of men, done in the past,
known by critical evaluation of present evidence, and limited
only by the historian's own understanding of this evidence. Our
question from Chapter VI is whether there is any room in this
final concept of history as a scale of forms for the earliest
stage in this development, i.e. the realistic element of "fact"
as something which exists independently of anyone's knowledge of
it (RP, 49; FR, 83).
But if the work of reconciliation of philosophy and
history is complete when we have shown that the object of his-
tory displays the structure of a philosophical scale of forms,
why did Collingwood proceed, after his essays of the twenties,
to think and write on the subject? Obviously he had something
further to say, and we assume that this must have involved not
only the question of the multiple layers of meaning to the
philosophical concept of history, but also the question of the
sort of truths that are embodied in history and the form of
inference that historical thinking employs. In pursuing our
third general task we therefore contact Collingwood's handling
of the classic contemporary issues in the philosophy of history,
viz. issues concerning (a) the meaning of history, (b)
historical truth, and (c) the nature of historical

614
explanation.2 We have already met with Collingwood's response to
the first issue, and we shall have more to say about it in a
moment. We know at least that whatever else history might be, if
it is a philosophical concept its meaning has the structure of a
scale of overlapping forms. But thus far we have not found
Collingwood saying much about the nature of historical truth and
inference, and we anticipate that in The Idea of History these
will be the issues that occupy the foreground of his concern.
Therefore we expect to see that the task of reconciling history
and philosophy is carried forward in his later writings by
extending the dialectical logic of his Essay on Philosophical
Method to history at the propositional and inferential levels of
thought.
These considerations provide some substance to the
general tasks we have set for ourselves in this chapter. Our
immediate concern in the next section will be to show, first,
that in his overall reflection on the
______________________
2 Cf. W. H. Dray, "Philosophy of History," in The En-
cyclopedia of Philosophy, ed. by Paul Edwards (New York, 1967),
VI, 247-54. In his article Dray distinguishes between
speculative and critical philosophies of history, and within the
latter identifies the three major problem areas as those dealing
with historical explanation, the historical individual, and
historical objectivity. Since historical objectivity is
basically the question of the nature of historical truth, two of
the three of these issues in the critical philosophy of history
are directly addressed by Collingwood; the third is present, but
in slightly altered form. As we shall see Collingwood takes as
his historical individual the meaning of an historical act, so
that in this issue are bound both the speculative and critical
aspects of the philosophy of history.
615
nature of historical thinking Collingwood's orientation is
always displayed by a critique of the fundamental presupposition
which he called "realism." This critique depends on a definition
of history as a philosophical rather than a formal-logical
concept--i.e. one having the structure of a scale of forms
rather than that of a formally abstract class-concept. Secondly,
this critique leads directly to the formulation of his famous
"inside-outside" theory of historical events, i.e. that his-
torical events differ from physical events insofar as they are
essentially expressions of thought which the latter are not.
Thirdly, since the subject matter of history is thought,
Collingwood develops a minimal notion of what is to be included
in that concept and, fourthly, how it is grasped by an of
historical re-enactment. Fifthly, re-enactment presupposes the
activity of an a priori imagination, the functions of which he
sketches. And finally, this excursion from historical meaning to
mental functions climaxes in a remarkable debate in which
Collingwood's critique is extended beyond realism to the errors
of idealism as well, showing how on either presupposition
historical thinking, and hence historical truth, is not possible
at all.
It is at this point that we begin to see how Colling-
wood's later philosophy of history moves beyond the essays of
the twenties. We encounter arguments that exhibit the sort of
structure that he had described in the Essay on Philosophical
Method as peculiarly philosophical

616
--arguments that are reversible and conclusions that are
reciprocally established. We also are made increasingly aware of
the degree of his reliance on the theory of mental functions
that we were at pains to describe in the previous chapter. And
we see how, as might be expected, the issue of historical truth
leads directly to the problem of historical inference. In
section three we shall take up the topics of historical
methodology, Q-A logic, and historical inference, trying once
again to watch for hints that will help us to get a clearer idea
of what Collingwood understood by "real thinking" and the rela-
tionship between Q-A, F-, and D-logics.
Textually, what we are attempting is a reconstruction of
The Idea of History. In section two we will take up the
Introduction to The Idea of History and the first, second,
fourth and fifth of the seven Epilegomena--all of these being
works dating from around 1934-36 and proceeding his work on the
unfinished Principles of History. In section three we shall be
concerned with the remaining Epilegomena, but primarily the
third (which deals with historical inference)--all of these
being extracts from the 1939 draft of the Principles of History.
Our survey thus will have the effect of a counterbalance to
Knox's editorial arrangement of these essays, since it
approximates their actual chronological

617
order. Whether it also establishes their logical order as well
we leave to the judgment of the reader.
2. Anti-Realism and History: (a) The Definition of History.
The Introduction to The Idea of History presents, in
admirably concise language, a sketch of the nature, subject
matter, method, and aim of the philosophy of history--an ex-
position of which any scholastic philosopher would be proud.
Collingwood sets out to answer the questions, what is history?
what is it about? how does it proceed? and what is it for? (or
alternately, what is its value?)--and his answers are as
follows. (1) History is a kind of research or inquiry belonging
to the sciences, that is, "the forms of thought whereby we ask
questions and try to answer them"--an activity of "fastening
upon something we do not know and trying to discover it." (2)
The object of history or the kind of thing it finds out is "res
gestae: actions of human beings that have been done in the
past." (3) History proceeds by the interpretation of evidence,
where evidence is a collective name for things existing here and
now and of such
____________________
3 It is interesting to note that in spite of Colling-
wood's assertion in the Autobiography that in his early years as
a tutor and lecturer at Oxford he became something of a
specialist in Aristotle (the first lectures he gave were on De
Anima) (A, 27), few of his interpreters have followed up on this
lead. A case in point is the characterization of history as a
science, given in the Introduction, and another Aristotelianism
appears in his use of the potency-act distinction in connection
with historical evidence.
618
a kind that the historian, by thinking about it, can get answers
to the questions he asks about past events. And (4) the aim of
history, its value, is human self-knowledge, where knowing
oneself means "knowing, first, what it is to be a man; secondly,
knowing what it is to be the kind of man you are; and thirdly,
knowing what it is to be man you are and nobody else is"--a
knowledge achieved by knowing what one can do. (IH, 9-10).
History is, to put it shortly, scientific, humanistic, rational,
and self-revelatory (IH, 18).
Here in brief form is the sum, if not the entire sub-
stance, of Collingwood's philosophy of history, and the re-
mainder of The Idea of History consists of first a description
of how this complete idea came to take shape in the minds of
working historians (Parts I-IV), and secondly a commentary on
what each of these points means (Part V-The Epilegomena). Coming
to these four characteristics from the viewpoint of our recent
study of Collingwood's development to this point it is not
difficult to interpolate for ourselves how they are formulated
in opposition to realism. (a) When Collingwood calls history a
science and then describes science as "asking questions and
seeking answers" we recognize behind this assertion
Collingwood's assault on formal logic as a static relationship
between a completed set of propositions, and his "alternative"
proposal for a Q-A logic of discovery as the true instrument of
historical science.
619
(b) When Collingwood says that the object of history is res ges-
tae, we recognize that he is calling attention not only to the
fact that it is not knowable by direct observation or
"acquaintance" (on which realist epistemology is based), but
also that it requires a different description of knowing with an
epistemology of its own, one for which the basic model of
"explanation" by subsumption of particulars under a universal
law is inappropriate. (c) When Collingwood states that history
proceeds by the interpretation of evidence we cannot help
recalling the critique of "sense-datum" empiricism in The
Principles of Art. The way in which historical events of the
past are present to us requires the conscious act of
interpretation even more necessarily than does the perception of
objects which are immediately present before us. Although in
perception we can apprehend or appreciate something just as it
stands, before beginning to classify it (PA, 203), historical
evidence is only actual evidence insofar as it is known to be
what it is, i.e. insofar as it is historically interpreted (IH,
247, 280). And (d) when Collingwood insists that the aim of
history is for human self-knowledge, we hear the refrain of his
lifelong theme of reconciliation between the ancient
philosophical imperative to "know thyself" and the contemporary
view of history as the critical knowledge of human events. In
Collingwood's view the only way to effect this rapprochement is
by ridding philosophy of the notion that

620
it can live up to the sort of objectivity proposed for it by the
analogy with exact or empirical sciences. The reconciliation is
complete when both these errors are corrected, and both
philosophy and history are understood to be forms of
self-knowledge.
Lest we be accused of reading too much into The Idea
of_History we hasten to add that this aspect of rapprochement is
explicit in the Introduction where Collingwood writes that a
subject matter is philosophical insofar as it deals with "the
organized and scientific development of self-consciousness" (IH,
4), and adds that it is history rather than other forms of
knowledge that raises the peculiarly modern problems that shed
new light on this development.
The past, consisting of particular events in space and time

which are no longer happening, cannot be apprehended by
mathematical thinking, because mathematical thinking
apprehends objects that have no special location in space
and time . . . . Nor can the past be apprehended by
theological thinking, because the object of that thinking
is a single infinite object, and historical events are
finite and plural. Nor by ((empirical)) scientific
thinking, because the truths which science discovers are
known to be true by being found through observation and
experiment exemplified in what we actually perceive,
whereas the past has vanished and our ideas about it can
never be verified as we verify our scientific hypotheses.
Theories of knowledge designed to account for mathematical
and theological and scientific knowledge thus do not touch
on the special problems of historical knowledge; and if
they offer themselves as complete accounts of knowledge
they actually imply that historical knowledge is
impossible. (IH, 5).
621
It is noteworthy that although there is an indistinct
echo of the various forms of knowledge from Speculum Mentis,
there is no explicit connection made between history as itself a
philosophical scale of forms and as a form of self-knowledge
along with others (art, science, etc.). Thus we are faced from
the outset with a dilemma, if not an outright reversal from the
standpoint of his earlier writings: it appears that one can
either effect a rapprochement by pointing to a common element of
self-knowledge in both history and philosophy, thereby defining
both in terms of a universal and ignoring their structure as a
scale of forms, or define history in terms of its object as part
of a scale of forms of knowledge and lose its rapprochement with
philosophy as both forms of self-knowledge.
But this is less a problem than it appears. Collingwood
has not abandoned his requirement from the Essay on
Philosophical Method of defining a philosophical concept by
means of overlapping forms. In the Introduction he has merely
given us a brief sketch of the highest exemplification of the
concept (viz. scientific history) prior to showing how this
concept has developed to this point from its prior stages. It is
the burden of Parts I-IV of The Idea of History to trace this
development through its manifestations in historiography--a
remarkable overview which we shall not attempt to examine here
in detail. We wish to point out only

622
that both in Parts I-IV and in the Epilegomena of Part V,
Collingwood's strategy is to show that history, in its
development and in its present state, is the expression of mind
in its several overlapping functions--presentative perception,
re-presentative imagination, critical understanding and
reconciling reason. Each of these functions gives rise to a form
of historiography in which they form the guiding idea. Thus
Collingwood traces the idea of history (a) from its pre-historic
beginnings in mythology and sacred literature, through (b) its
first concrete manifestations in Greek histories with their
emphasis on eyewitness accounts, (c) subsequently developed in
the Christian era as "scissors and paste" history which relies
uncritically on "authorities" who in turn rest on eyewitness
accounts; then (d) to critical historiography originating in the
Renaissance but coming into its own in the 18th century with
philosophers like Vico who recognized that the historian can and
must employ his own reasoning to adjudicate conflicting accounts
of historical authorities; and finally (e) to the present era of
scientific historiography, in which the active role of the
historian is thoroughly recognized as essential to the very
process of historical thinking. It is not difficult to recognize
in these stages of the concept of history a schema we have
already encountered in The Principles of Art and The New
Leviathan in the process of thought in general. Thus we have (a)
history as the confused potential for bearing meaning (mythical
and
623
theological history), (b) as events perpetuated and domesticated
in perception by eye-witnesses (early Greco-Roman
historiography), (c) as chronicles represented in imagination
and memory ("scissors and paste"--e.g. Medieval historiography),
(d) as critically constructed into coherent narratives by
judgmental understanding (critical historiography), and (e) as
fully related to self-conscious reason by recognizing and
integrating the active role of the historian in the construction
of his narrative account (scientific historiography). It is in
this way that he offers recompense for the neglect of history by
the realists, and it is in this way that he shows that history
is a form of self-knowledge.
This overview of the argument of Part I is not totally
free of the inaccuracies that arise whenever oversimplification
occurs. Greco-Roman history, for example, is not completely or
even primarily based on eye-witness accounts: much of it is
mixed with legend, myth, and even outright inventions by the
historian (orations in Thucydides, for example). In Speculum
Mentis Collingwood made a point of describing how early
historiography is mixed with dramatic and religious elements, so
that there is some confusion in the mind of an historian like
Herodotus between the ideals of factual history per se and the
artistic ideals of drama (SM, 211-16). Collingwood was no less
aware of such an overlap when he wrote Part I of The Idea
of_History: but he was not concerned with

624
showing the relation of history to other forms of experience
(e.g. art), but rather with showing how the clear idea of
scientific history grew out of a confused idea of it. Our point
in bringing out the analogy between the development of
scientific history and the levels of mental functions is meant
to call attention to the shift in ideals of historiography, and
to show how in that shift a scale of forms is generated.
If the four elements of Collingwood's definition of
history are formulated in opposition to realism, how do they
form a challenge to the central realistic dogma on the rela-
tionship of knower and known? If the object known is not af-
fected by the knowing of it, then it appears that objectivity is
retained for history, but self-knowledge is not; but if the
events of history are altered in the process of coming to know
them, then historical objectivity is itself threatened. In
showing how the concept of history is preserved in the truths of
history, Collingwood will argue, as he already had in The
Principles of Art for perception, that historical events are not
created ex nihilo by historical imagination, but they are
preserved and prolonged by historical consciousness, and this is
based on the historian's operating presuppositions. As a first
step in showing how he works this out in detail in the
Epilegomena essays, we must clarify the notion of an historical
event, the minimum exemplification of the

625
concept of history, and the res gestae of his definition in the
Introduction.
(b) The Outside and Inside of Historical Events.—It is
interesting to note that the first two of these Epilegomena,
both lectures, are directed against views of history by realists
and idealists respectively. The first is Collingwood's 1936
British Academy lecture, "Human Nature and Human History," and
it packs a double anti-realistic punch. It makes as its major
thesis that "the science of human nature was a false
attempt--falsified by the analogy of natural science--to
understand the mind itself, and that, whereas the right way of
investigating nature is by the methods called scientific, the
right way of investigating the mind is by the methods of
history" (IH, 209). But rather than remaining content to refute
an 18th century error, Collingwood carries the argument forward
from empiricism to contemporary realism. He takes as his
proximate target the essay "The Historicity of Things" by the
avowed realist, Samuel Alexander.4 Alexander's essay identifies
historicity with the universal "timefulness" of things, a tactic
which results in resolving all knowledge to history (IH, 210).
In order to preserve the autonomy of both history and science,
Collingwood makes it a point to distinguish history not only
_____________________
4 In Philosophy and History, The Ernst Cassirer Fest-
schrift, ed. by Raymond-Klibansky and H. J. Paton (Oxford,
1936), pp. 11-25.
626
from physical change but also from "timefulness"--both resulting
from the confusion of the scientific conception of nature and
the historical conception of mind.
First he tackles an older concept:
Since the time of Heraclitus and Plato, it has been a

commonplace that things natural, no less than things human,
are in constant change, and that the entire world of nature
is a world of "process" or "becoming." But this is not what
is meant by the historicity of things; for change and
history are not at all the same. According to this
old-established conception, the specific forms of natural
things constitute a changeless repertory of fixed types,
and the process of nature is a process by which instances
of these forms . . . come into existence and pass out of it
again. Now in human affairs . . . there is no such fixed
repertory of specific forms . . . . (H)uman history shows a
change not only in the individual cases in which these
ideals are realized or partially realized, but in the
ideals themselves. (IH, 210-11).
Collingwood recognized that the evolutionary conception of na-
ture has replaced the older concept of a physical universe of
unaltered species, so that in the newer view, as expressed by
Whitehead, "the very possession of its attributes by a natural
thing takes time" and "the historicity of things" is proven by
the fact that there can be no such thing as nature at an instant
(IN, 212). But this presents an even subtler danger, since
history is still not reducible to science nor historical events
to scientific ones:
These modern views of nature do, no doubt, "take time

seriously." But just as history is not the same thing as
change, so it is not the same thing as "timefulness," whe-
ther that means evolution or an existence which takes time.
. . . According to him ((i.e. the
627
historian)), all history properly so called is the history

of human affairs . . . . There is a certain analogy between
the archeologist's interpretation of a stratified site and
the geologist's interpretation of rock-horizons with their
associated fossils; but the difference is no less clear
than the similarity. The archeologist's use of his
stratified relics depends on his conceiving them as
artifacts serving human purposes and thus expressing a
particular way in which men have thought about their own
life; and from his point of view the paleontologist,
arranging his fossils in a time-series, is not working as
an historian, but only as a scientist thinking in a way
which can at most be described as quasi-historical. (IH,
212).
It is in order to distinguish these two classes of events and
these two sorts of thinking activities that Collingwood in-
troduces his "inside-outside" theory of human acts.
He distinguishes between physical events and human acts
on the basis that physical events have only an "outside"
consisting of "everything belonging to it which can be described
in terms of bodies and their movements," whereas human acts have
in addition an "inside" or "that which can only be described in
terms of thought" (IH, 213). If this appears to be a mere
restatement of a kind of Cartesian dualism, Collingwood is
anxious to add that it is not the dualism of a pair of mutually
exclusive classes. The historian, he insists, is never concerned
with either of these to the exclusion of the other. "He is
investigating not mere events (where by a mere event I mean one
which has only an outside and no inside) but actions, and an
action is the unity of the outside and inside of an event"
(ibid.). While it is true that the natural

628
scientist goes beyond the immediate event by observing its
relations to others and bringing them under a general formula or
law of nature, to the scientist nature is always merely a
"phenomenon" or spectacle presented to his intelligent
observation. But the events of history are never mere phenomena
or spectacles, but "things which the historian looks, not at,
but through, to discern the thought within them" (IH, 214).
This last remark underscores the importance of viewing
Collingwood's inside-outside theory in the light of his
philosophy of mind. When he says that the historian looks "not
at but through" the outside of an event he is saying that the
historian takes such acts as expressions of thoughts, i.e. as
conveyers of meaning exactly like language (where language is
taken in its extended sense as including that "gesture language"
that Collingwood recognized as more basic than spoken language).
Acts, like sounds, ink-marks, gesture-signs, etc. are physical
"bearers of meaning"--they express the intentions of an
historical agent in acts as his words do in sounds. It is
because they have this character that they have the capacity or
potential to be evidence to the historian; but it is also the
reason that they require interpretation by an historian. To
paraphrase the situation for Collingwood, a physical event may
or may not be whatever it is without human interpretation; but
an historical event cannot. Sounds may or may not be bearers of
meanings,
629
but human actions, like words, do not have this alternative.
Unfortunately Collingwood's inside-outside doctrine has
generated a fair amount of misunderstanding, even among his more
sympathetic interpreters. Some of this appears to be the result
of his own overstatement--as when he writes that "the processes
of events which constitute the world of nature are altogether
different in kind from the processes of thought which constitute
the world of history" (IH, 217). It has led to the charges both
that he over-intellectualizes history and that he draws an
overly strict distinction between history and nature. Once again
the first charge seems to be directly supported by Collingwood.
Having said that history is not the same as change, natural
process or timefulness, he further limits the range of
historical research by insisting that it does not even include
all human activities:
It does not follow that all human actions are subject-

matter for history; and indeed historians are agreed that
they are not. But when they are asked how the distinction
is to be made between historical and non-historical human
actions, they are at a loss how to reply. From our present
point of view we can offer an answer; so far as man's
conduct is determined by what may be called his animal na-
ture, his impulses and appetites, it is non-historical; the
process of these activities is a natural process. Thus, the
historian is not interested in the fact that men eat and
sleep and make love and thus satisfy their natural
appetites; but he is interested in the social customs which
they create by their thought as a framework within which
these appetites find satisfaction in ways sanctioned by
convention and morality. (IH, 216; cf. IH, 315).
630
The question immediately arises, does this not
rule out a great deal of what is usually taken to be the
subject-matter of history? What about the rages of Ivan the
Terrible, Napoleon, or Hitler, or natural disasters like the
Lisbon and San Francisco earthquakes? We shall see in the next
section what Collingwood takes to be the true subject-matter of
history, but we see from the above quote what his general answer
to the objection is. Natural events are of interest to the
historian only to the extent that they impinge on human events
in the proper sense, or are incorporated into them as an aspect
of thought. Passions, natural disasters and other physical
phenomena are historical only to the extent that people
thinkingly react to them--Hitler's rages as instrumental or
obstructive to the discharge of Nazi warplans, the Lisbon
earthquake causing decisions on relocation of populations or
influencing government financial policies, etc. But regarded as
entities in themselves, i.e. as not the expression of human
thought, passions and feelings are the subject-matter for psy-
chology, land-mass shifts for geology, etc.
The second objection is more serious, and takes a bit of
sleuthing to uncover Collingwood's thought on the matter. We
have already suggested that while historical events necessarily
involve thought, since they express human intentions in
purposive acts, natural events do not. Does this mean they do
not as a matter of fact, or as a matter of necessity?

631
Collingwood appears to leave the question open, commenting only
that "The only condition on which there could be a history of
nature is ((on the assumption)) that the events of nature are
actions on the part of some thinking being or beings, and that
by studying these actions we could discover what were the
thoughts which they expressed and think these thoughts for our-
selves. This is a condition," he adds cryptically, "which
probably no one will claim is fulfilled" (IH, 302).5 What sort
of
_____________________
5 Collingwood could not have anticipated that someone
after his death would, in fact, make such a claim--but such is
the argument of J. Blachowicz in "History and Nature in Col-
lingwood's Dialectic," Idealistic Studies, VI, 2 (January,
1976), pp. 49-61. Although sympathetic to Collingwood's overall
aims, Blachowicz finds that Collingwood failed to extend his
dialectical analysis to nature itself, and is therefore guilty
of drawing an overly strict distinction between events of nature
and those of history--thus being open to an accusation of
falling prey to his own "Fallacy of False Disjunction.
Blachowicz argues that in order to extend his dialectical an-
alysis to nature Collingwood merely had to see that the "pre-
suppositions of nature ((may)) be disclosed by way of access
into its 'inside' as well" (p. 56). But for Blachowicz the
"inside" of physical events consist of their teleological as-
pect, which is contained in the fundamental parameter of com-
plexity or organization. "As the variable of internal organi-
zation assumes different values, the generic essence is quali-
tatively altered, generating the scale of forms which is the
scale of nature," so that "the presuppositional logic is fully
applicable" to development in nature" (p. 57). Now while we
agree with Blachowicz in bemoaning Collingwood's failure to
achieve a clearer rapprochement between Q-A and D-logics, we is
alien to Collingwood's purpose to attribute to physical,
non-human events. Presuppositions, whatever else they may be,
are essentially a part of a P-Q-A complex and essentially
“linguistic” in the sense of expressing the thoughts of persons.
Physical events do not ask questions (although they may, on
human interpretation, "raise" questions in the minds of
observers); therefore they do not "have” presuppositions in
Collingwood's sense. But Blachowicz is right to point out
Collingwood's curious neglect of the applicability of the scale
of forms to nature--a deficiency not shared by some other
contemporary cosmologies.
632
"thinking" Collingwood has in mind in the expression "thinking
beings" we shall see in a moment, but we have already seen that
he has in mind the sort of problem-solving choices that embody
the unification of theoretical and practical activities--what he
has elsewhere called "real thinking" or "concrete thought." The
outside and inside of an historical act are not themselves two
separate events, but two sides to the same act--exactly like
expressive language and thought. Thus when the historical agent
is mistaken in his theoretical assessment of the situation in
which he finds himself called upon to act, this mistaken thought
cannot be ignored by the historian who attempts to understand
the act: his mistaken thought is essential to understanding the
significance of the agent's act (IH, 316-17).
Now this point is crucial in understanding the inside-
outside theory: while Collingwood is willing to use the term
"event" for both history and nature, he is not able to bring
himself to do so for the term "action" because the latter term
requires the essential element of self-consciousness in the
awareness of alternative possibilities (IH, 215). Collingwood
finds it highly questionable, if not downright repugnant, to
predicate this state of affairs of nature. This becomes apparent
in an earlier portion of The Idea of History, in a discussion of
teleology--that aspect of "acting for an end" that is teetering
on the brink of the distinction between history and nature. The
issue arises in
633
Collingwood's discussion of Kant's distinction between natural
and human affairs, viz. that nature acts in accordance with law,
but only man acts in accordance with the concept or
consciousness of law. Nature, therefore, is the sum of processes
governed by laws blindly obeyed, but the world of human affairs
is governed not simply by law but by the consciousness of law
(IH, 92). He then cites with apparent approval Kant's
demonstration of why there should be such a thing as history.
Nature's purpose in creating any of her creatures is, of

course, the existence of that creature, the realization of
its essence. The teleology of nature is an internal
teleology, not an external: she does not make grass to feed
cows, and cows to feed men; she makes grass in order that
there would be grass, and so on. Man's essence is his
reason; therefore she makes men in order that they should
be rational . . . . Man is an animal that has the peculiar
faculty of profiting by the experience of others; and he
has this faculty because he is rational, for reason is a
kind of experience in which this is possible. . . .
Consequently the purpose of nature for the development of
man's reason is a purpose that can be fully realized only
in the history of the human race and not in an individual
life. (IH, 98).
Now it is worth noting that while Collingwood faults Kant for
locating the activating force for the plan of human history,
i.e. progress in rationality in human irrationality, i.e. pas-
sion, ignorance, and selfishness (IH, 103), he does not fault
him for the basic distinction between nature and the world of
human affairs. Such is not the case with his discussion of
Croce, whom he charges with blurring the distinction

634
completely. Collingwood quotes a passage in which Croce
challenges his reader to an experiment: if you wish to
understand the true history of a neolithic man, become a
neolithic in your mind; and if you wish to understand the true
history of a blade of grass, become that grass in your mind--but
if you cannot, content yourself with describing and arranging
artifacts or mechanisms in an external way. Then Collingwood
responds to Croce's challenge:
As concerns neolithic man, the advice is obviously good. If

you can enter into his mind and make his thoughts your own,
you can write his history, and not otherwise . . . . When
he made a certain implement, he had a purpose in mind; the
implement came into being as an expression of his spirit,
and if you treat it as non-spiritual that is only because
of the failure of your historical insight. But is this true
of a blade of grass? Is its articulation and growth an
expression of its own spiritual life? I am not so sure. And
when we come to a crystal, or a stalactite, my skepticism
reaches the point of rebellion. The process by which these
things form themselves appears to me to be a process in
which, through no lack of our own historical sympathy, we
look in vain for any expression of thought. It is an event;
it has individuality; but it seems to lack that inwardness
which, according to . . . Croce, is made (and I think,
rightly made) the criterion of historicity. (IH, 199-200).
The upshot of this discussion is that while Collingwood
is willing to grant an "internal teleology" to nature, he is not
willing to abandon the distinction between natural and his-
torical events, because the latter cannot be understood except
as the expression of human thought. It is less a matter of ne-
cessity than it is one of fact that nature is not so constitu-
ted; but it is a matter of necessity that history is, since,

635
on Collingwood's view, an historical event necessarily is an
action of a human agent (one is tempted to say "by definition"),
and this requires awareness of alternatives, or reflective
thought. And it is this difference in essential constitution
that he has in mind when he says that they are "altogether
different in kind" (IH, 217--notice the wording: the processes
of events which constitute the world of nature are altogether
different in kind from the processes of thought which constitute
the world of history).
The "inside-outside" terminology is therefore a meta-
phorical expression which Collingwood finds useful for de-
scribing what he takes to be an essential difference between
historical and physical events--a distinction that he maintains
both against positivistic6 historians who assimilate history to
science as if the nature of the "facts" or events were basically
the same (IH, 132-33), and against the idealistic historians who
assimilate nature to history by assuming that one can re-think a
physical event in the same way that one can re-enact an
______________________
6 The successors to the English empiricist tradition in
history are the positivistic historians, just as the successors
to that tradition in philosophy are the realists of Oxford and
Cambridge (IH, 126-33, 142, 163-64, 173-74). As we have already
noted, empiricism, naturalism, realism, positivism, and even
psychologism are products of the same frame of mind, and
Collingwood often uses the terms synonymously. He recognized in
all the root error of realism, i.e. the assumption that
knowledge consists of the confrontation of a mind with an object
outside the mind, where the latter remains unaffected by the
knowing of it.
636
historical one. The philosophy of history which is argued in the
Epilegomena is hardly intelligible unless it is understood
against the background of this opposition between realistic or
positivistic history and its idealistic counterpart--a theme we
shall encounter again in the discussion of historical
re-enactment.
(c) Individuality, Universality, and the Subject Matter
of History.-- In analyzing the res gestae of history Collingwood
has been more concerned with the differences between historical
acts and physical events than with the positive characteristics
of the acts themselves. Thus far we know minimally that history
is concerned with acts of men done in the past, acts which have
an "inside" consisting of thought. We know also that these acts
are done self-consciously or "on purpose," and that they involve
the practical resolution of situational problems expressing a
thoughtful choice between alternatives (IH, 215, 283). And
finally, they are experiential--they are, as he phrases it, "not
spectacles to be watched, but experiences to be lived through"
(IH, 218).
In the fifth Epilegomenon, "The Subject Matter of His-
tory," Collingwood reiterates these characteristics and adds
another which sheds some light not only on the minimal unit
presupposed for understanding historical processes, but also on
our previous discussion of universal concepts and abstract
classes. Reversing the order of the approach taken

637
in the first Epilegomenon, Collingwood this time proceeds from
the subjective to the objective. The subject matter of history
must first of all be experience and not the mere object of ex-
perience, i.e. the processes of nature that are the subject
matter for natural science. But it is not even experience as
such that the historian seeks, since immediate experience is a
"mere flow of consciousness consisting of sensations, feelings
and the like" which are carried away in the flux of sensuous
experience (cf. IH, 233). Nor is it even thought in its
immediacy, i.e. "the unique act of thought with its unique
context in the life of an individual thinker," since if this
were so, in thinking an historical subject matter "the historian
would be the person about whom he thinks, living over again
((acts)) in all respects the same" (IH, 302-03). What the
historian apprehends is not the individual in all its in-
dividuality, but something essentially universal:
The historian cannot apprehend the individual act of

thought in its individuality, just as it actually happened
((or in Ranke's words, wie es eigentlich gewesen--IH,
130)). What he apprehends of that individual is only some-
thing that it might have shared with other acts of thought
and actually has shared with his own. But this something is
not an abstraction, in the sense of a common characteristic
shared by different individuals and considered apart from
the individuals that share it. It is the act of thought
itself, in its survival and revival at different times and
in different persons; once in the historian's own life,
once in the life of the person whose history he is
narrating. (IH, 303).
Thus far we recognize Collingwood's epistemological setting
for the notion of a universal concept which "survives and

638
revives" in mental activities; but he here goes on to give it a
different perspective than we have heretofore encountered. The
passage continues:
Thus the vague phrase ((of Croce--IH, 199)) that history is

knowledge of the individual claims for it a field at once
too wide and too narrow: too wide, because the individuality
of perceived objects and natural facts and immediate
experience falls outside its sphere . . . ; too narrow,
because it would exclude universality, and it is just the
universality of an event or character that makes it a proper
and possible object of historical study, if by universality
we mean something that oversteps the limits of merely local
and temporal existence and possesses a significance valid
for all men at all times . . . . (T)hought, transcending its
own immediacy, survives and revives in other contexts; and .
. . individual acts and persons appear in history not in
virtue of their individuality as such, but because that
individuality is the vehicle of a thought which, because it
was actually theirs, is potentially everyone 's. (IH, 303;
cf. IN, 247, 280).
In this subtle and neglected theme we find Collingwood's un-
derstanding of universal concepts lit up from several directions
simultaneously. (a) Not only does it contrast the individual
(what he had called, in Speculum Mentis, the concrete universal)
with the abstract universal (or the class-concept of F-logic);
and (b) not only does this universality take part in the
survival and revival of a concept in the context of different
mental acts, thus transcending the local and temporal limits of
immediate feelings; (c) but also it binds the universality of a
concept to the individuality of the meaning of an historical act
which expresses a thought which, as he says, is potentially
everyone's because it was actually theirs. The universality of
historical thinking is bound by

639
Collingwood to the notion of meaning expressed in historical
acts, and this is the individuality with which history is
minimally concerned. This theme will be renewed when we come to
Collingwood's treatment of re-enactment, where it re-appears in
the critique of the distinction between acts of thought and
their objects.
It is also here that we encounter the self-consciousness
that is essential not only to historical understanding but in
the historical act of an agent in the first place. Both in his
definition of history as consisting of acts of historical agents
with their inside consisting of thought, and in his definition
of historical re-enactment on the part of the historian, the key
term is, of course, "thought." Collingwood does not back away
from stating what is included in this term.
How much or how little is meant to be included under the

term "thought"? The term "thought" . . . has stood for a
certain form of experience or mental activity whose pe-
culiarity may be negatively described by saying that it is
not merely immediate, and therefore is not carried away by
the flow of consciousness. The positive peculiarity which
distinguishes thought from mere consciousness is its power
of recognizing the activity of the self as a single
activity persisting through the diversity of its own acts .
. . . The peculiarity of thought, then, is that it is not
mere consciousness but self-consciousness. The self, as
merely conscious, is a flow of consciousness, a series of
immediate sensations and feelings; but as merely conscious
it is not aware of itself as such a flow; it is ignorant of
its own continuity through the succession of experiences.
The activity of becoming aware of this continuity is what
is called thinking. (IN, 306; cf. IN, 222).
640
Collingwood follows this continuity-achieving activity briefly
through its manifestations in perception, memory, and imagin-
ation--the stages of recognizing feelings as "mine," prolonging
them, etc. The account is a brief tour of the stages of mental
functions that we have already encountered in The Principles of
Art and The New Leviathan. As before, it emphasizes the active
role of thought at all levels of mental activity. But in this
case Collingwood is anxious to draw the distinction between
reflectively self-conscious thought and non-reflective or
"unconscious" thinking, and in so doing we again see it from a
new angle than we have previously.
(T)he thinking which we do in memory or perception as such

may be called unconscious thinking, not because we can do
it without being conscious, for in order to do it we must
be not only conscious but self-conscious, but because we do
it without being conscious that we are doing it. To be
conscious that I am thinking is to think in a new way,
which may be called reflecting. Historical thinking is al-
ways reflection; for reflection is thinking about the act
of thinking . . . . But what kind of thinking can be its
object? . . . In order . . . that any particular act of
thought should become subject-matter for history, it must
be an act not only of thought but of reflective thought,
that is, one which is performed in the consciousness that
it is being performed, and is constituted by that con-
sciousness. (IH, 307-08).
Notice that Collingwood does not find the expression "uncon-
scious thinking" self-contradictory (cf. per contra, Donagan,
LPC, 271) and that he accepts without hesitation the presence of
the activity of self-consciousness even at the level of
perception--a point to which we called attention earlier. Nor is
it a great surprise that he finds

641
historical thinking to be self-consciously reflective, although
we find him for the first time stating that it is "constituted"
by that consciousness. But then he continues:
The effort to do it ((i.e. think reflectively)) must be

more than a merely conscious effort . . . ; it must be a
reflective effort, the effort to do something of which we
have a conception before we do it. A reflective activity is
one in which we know what it is that we are trying to do,
so that when it is done we know that it is done by seeing
that it has conformed to the standard or criterion which
was our initial conception of it. It is therefore an act
which are enabled to perform by knowing in advance how to
perform it. (IH, 308).
At this point it should be perfectly obvious why, on Colling-
wood's grounds, physical events cannot be regarded as historical
actions: they are not acts performed with the foreknowledge of a
"standard or criterion" which is their initial conception of it.
To use Collingwood's lowly example, the blade of grass grows not
in order that the cow should eat it, or with the conception of
being food for herbivorous animals, but that there should be
grass. But even that is not correct, for it has no conception at
all of what it is to be grass rather than some alternative. But
historical acts cannot be thus regarded; their universal
character is part of their essential constitution:
An act is more than a mere unique individual; it is some-

thing having a universal character; and in the case of a
reflective or deliberate act (an act which we not only do,
but intend to do before doing it) this universal character
is the plan or idea of the act which we conceive in our
thought before doing the act itself, and the criterion by
reference to which, when we have done it, we know that
642
we have done what we meant to do . . . . Reflective acts

may be roughly described as the acts which we do on pur-
pose, and these are the only acts which can become the
subject matter of history. (IH, 309).
Now although Collingwood has phrased the distinction in temporal
terms, ("an act which we . . . intend to do before doing
it . . .") it is clear enough from the rest of the paragraph
that what he has in mind is what is also reflected in the basic
Kantian distinction between actions done in accordance with a
law and actions done in accordance with the concept of a law;
historical acts fall into the latter category, physical events
in the former. What Collingwood has added is the observation
that all such deliberate reflective acts are historical, that
they involve criteria by which success or failure may be
assessed, and that therefore these acts have aspects that are
both individual (as experientially one's own) and universal (as
reflectively intended).
Here we encounter a problem. In historical thinking (in
this case we mean the thinking of agents involved in the doing
of historical deeds) we find the sort of thinking Collingwood
called "concrete" in Speculum Mentis and "real thinking" in The
New Leviathan and his correspondence with Ryle concerning the
Essay on Philosophical Method. We recall from our discussion of
his philosophy of mind that all real thinking has both
theoretical and practical dimensions, and that the latter is
foundational to the former. In saying that historical acts are
those which we plan to

643
do before doing them Collingwood appears to be reversing this
priority --conceiving a plan being a theoretical activity and
executing it a practical one. Collingwood anticipated this
objection and even drew from it the logical conclusion that
acting is the only thing one can do on purpose, since thinking
on purpose would involve conceiving your own act of thought
before executing it, and having done so you would have executed
it already. Theoretical activities, it follows, can only be
non-purposive or, as he puts it, "done in the dark, with no
conception of what is to come from engaging in them" (IH, 311).
The reader familiar with the debate concerning Collingwood's
views on absolute presuppositions (i.e. that they are not
propositions, that one can make them without knowing that they
are being made, etc.) will realize that we are on the verge of a
crucial issue not only in his philosophy of history, but in his
view of metaphysics.
We know already what part of his answer to this objec-
tion is: any theoretical activity is already practical insofar
as it employs concepts that are made by acts of selective
attention. But his actual reply is interesting insofar as it
suggests a connection with Q-A logic which we shall have to bear
in mind when we come to examine his thesis that metaphysics is
an historical science.
644
Today it is no longer necessary to argue that art, science,

religion, philosophy, and so forth are proper subjects of
historical study; the fact of their being studied
historically is too familiar. But it is necessary to ask
why this is so . . . . In the first place, it is not true
that a person engaged in purely theoretical thinking is
acting without a purpose . . . . (E)very actual inquiry
starts from a certain problem, and the purpose of the
inquiry is to solve that problem; the plan of the
discovery, therefore, is already known and formulated by
saying that, whatever the discovery may be, it must be such
as to satisfy the terms of the problem . . . . In the
second place, the difference between conceiving and
executing a purpose was not correctly described as the
difference between a theoretical act and a practical one.
To conceive a purpose or form an intention is already a
practical activity. It is not thought forming an anteroom
to action; it is action itself in its initial stage. (IH,
311-12).
Collingwood's reply, therefore, is based on his Q-A logic as
grounded in his epistemology: thinking on purpose does involve
conceiving one's own act of thought before executing it, but as
a plan of inquiry based on the presence of a problem situation
or a question, whereas the execution of the plan involves
answering the question, solving the problem, or satisfying the
terms of the inquiry-initiator.
It is not difficult to make the application to histor-
ical inquiry. In order for an event to be of interest to an
historian it must be one which expresses the thought of an
historical agent in a situation in which alternative courses of
action are open to him, and in which he is responsible for
determining the event by acting according to his consciousness
of a "plan" or "idea"--the alternative meanings he not only finds
in the event-situation but gives to it by his chosen

645
actions. It will be helpful to keep this in mind as a sort of
archetype of what an historical event is--candidates for such
events occurring in politics, warfare, economics, morals, art,
science, religion, and philosophy (IH, 309-15)-the list is
stated as exemplary rather than exhaustive. The difficulty for
the historian in each of these cases is to identify the problem
from which the act proceeded, and to reconstruct the steps by
which its solution was attempted (IH, 312-13)--and not merely to
repeat the conclusion. Until this step has been taken the
historian cannot be sure that he has grasped the past at
all--that is, he cannot be certain that historical thinking in
the form of re-enactment or re-thinking has occurred at all.
But re-enactment raises us to a new level of thinking,
and to this issue we must now turn.
(d) Historical Re-enactment.--The deeds of men which
form the subject-matter of history must not only be known by
seeing them to be expressions of thought, or seeing "through"
them, they must be known as past deeds. The historian must
therefore be aware of himself as distinct from the past he
studies, while he is yet able to revive that past as thought in
his own mind (cf. IH, 174). In order to meet this requirement
Collingwood introduces his widely misunderstood doctrine of
historical re-enactment. The choice of the term "reenactment" to
describe the relationship between a historical event and a
historiographical
646
reconstruction of that event appears to be deliberately it calls
attention to the central concept of a purposive act as the
irreducible unit of history, while yet locating it in an
epistemological structure similar to the
presentation-representation schema of perception (the difference
being, of course, that the act of an historical agent can only
be re-presented through the intermediation of evidence and
interpretative argument). Collingwood is even careful to vary
his terms to bring out various aspects of reenactment, using
synonyms like re-create (IH, 97), reconstruct (IH, 65), relive
(IH, 172, 175), re-think (IH, 215), and revive (IH, 164)--an
assortment of expressions which can hardly be accidental (cf.
EPM, 205-07 on the in-appropriateness of technical terminology
in philosophy).
In this classic statement of re-enactment Collingwood's
approach is via the inside-outside doctrine:
The processes of nature can therefore be properly described

as sequences of mere events, but those of history
cannot.They are not processes of mere events, but processes
of actions, which have an inner side consisting of
processes of thought . . . . All history is
_______________________
7 Collingwood's first use of the term "re-enactment" in
The Idea of History occurs in Part I during his criticism of
Tacitus, where it is also linked to the discussion of the out-
side and inside of an event (IH, 39). It reappears in his
discussions of Vico (IH, 65) and Hegel (IH, 97), both of whom
are praised for recognizing the principle. In Part V it is
re-introduced in the first Epilegomenon along with the inside-
outside theory of human acts (IH, 215).
647
the history of thought. But how does the historian discern

the thoughts which he is trying to discover? There is only
one way in which it can be done: by re-thinking them in his
own mind. The historian of philosophy, reading Plato, is
trying to know what Plato thought when he expressed himself
in certain words. The only way in which he can do this is
by thinking it for himself. This, in fact, is what we mean
when we speak of "understanding" the words. So the
historian of politics or warfare, presented with an account
of certain actions of Julius Caesar, tries to understand
these actions, that is, to discover what thoughts in
Caesar's mind determined him to do them. This implies
envisaging for himself the situation in which Caesar stood,
and thinking for himself what Caesar thought about the
situation and the possible ways of dealing with it. The
history of thought, and therefore all history is the
re-enactment of past thought in the historian's own mind.
This re-enactment . . . is not a passive surrender . . . ;
it is a labour of active and therefore critical thinking.
The historian not only re-enacts past thought, he re-enacts
it in the context of his own knowledge and therefore, in
re-enacting it, forms his own judgment of its value,
corrects whatever errors he can discern in it (IH, 215;
emphasis mine).
The lines to which we have called attention in the above passage
are the principles of history cited in the Autobiography and
outlined in our chapter on Collingwood's autobiographical
interpretation as HIST -1, -2, and -3 (Table 4 ) . Since these
principles are explicitly stated in this essay, and the fourth
(HIST-4) is the major thesis of the first Epilegomenon itself,
it is clear that when Collingwood wrote that portion of the
Autobiography he had these essays before him.
But as if anticipating that his doctrine of re-enactment
would be misconstrued by assimilating it to an overly concrete
meaning of one of its secondary senses, Collingwood takes pains
to state what re-enactment is not before trying to exhibit what
it is. (1) We have already

648
quoted the passage from "The Subject Matter of History" where he
rejects a dramatic view of history in which "the historian would
be the person about whom he thinks, living over again in all
respects the same" (IH, 303). The historian in re-enacting the
event is not restaging it by repeating its outside motions; he
reenacts it by re-thinking that universal element that consti-
tutes its inside meaning. (2) But he is also not advocating an
intuitionist view of history which relies on a mysterious union
between the mind of an historian and that of the agent he is
studying. In spite of contrary appearances, as when Collingwood
uses expressions like "reliving" past experiences and "plunging
below the surface" of our minds where we "become" the person
whose acts we are discovering (A, 113), Collingwood has
something more direct in mind than what these dramatic
expressions might lead one to believe. Whatever else
re-enactment might involve, it is not concerned with immediate
experiences, nor is it a "passive surrender" but rather a
"labour of active and therefore critical thinking," which en-
tails envisioning for oneself the situation in which the agent
is called upon to act (IH, 215, 316; cf. A, 100). (3) This is
also the reason that re-enactment is not a simple exercise of
the memory. The past with which the historian is concerned is
not simply recalled, since there is an essential difference
between memory and the exercise of historical thinking. What
Caesar's memory may provide for him are a series

649
of outstanding images, but these he must present in a coherent
account, which involves not just the images but reconstruction
of the events. History is "a wholly reasoned knowledge of what
is transient and concrete" (IH, 234), whereas "memory is not
history, because history is a certain kind of organized or in-
ferential knowledge, and memory is not organized, not inferen-
tial at all" (IH, 252).
But if re-enactment is neither a dramatic restaging of
past events (1), nor an uncritical passive intuition of past
thinking (2), nor a replay of the graphic record of the
remembered past on the blank screen of the historian's mind (3),
it is also not (4) the creation ex nihilo of a fictitious past.
In the third Epilegomenon ("Historical Evidence"--a fragment
from The Principles of History) he writes that the business of
the historian is not to invent anything, it is to discover
something (IH, 251); and in the second Epilegomenon ("The
Historical Imagination") he writes:
As works of imagination, the historian's work and the

novelists do not differ. Where they do differ is that the
historian's picture is meant to be true. The novelist has a
single task only: to construct a coherent picture, one that
makes sense. The historian has a double task: he has both
to do this, and to construct a picture of things as they
really were and of events as they really happened. (IH,
246).
Just as, in The Principles of Art, Collingwood rejected any
radical theory of sensation which would allow for a creation ex
nihilo of sense,
650
so in The Idea of History he stops short of a creationist view
of historical events. Re-enactment presupposes that something
remains of the original event, and this something (we have
already identified it as the universal meaning of an individual
act) is not simply created by an historian's re-enactment.
Historical reference to past events presupposes an object that
retains some identity in both contexts, the original situation
in which it occurred, and the historian's re-enactment of that
event in the context of his own thoughts.
The question is, is anything important altered by re-
enactment, and if so, how is historical objectivity possible?
Since the issue is crucial not only for Collingwood's philosophy
of history, but also for rapprochement in general and his view
of metaphysics in particular,8 we must take care to present the
issue carefully and completely. We must go beyond saying what
re-enactment is not by giving a preliminary sketch of what it is
(which we are about to do), and then showing both that it
exercises some of the functions of a priori imagination and
_________________________
8 It is crucial not only because re-enactment is the
central thesis of his philosophy of history as described both in
the Autobiography and in The Idea of History, but also because
it is the key to unscrambling the central dilemma of the Essay
on Metaphysics. If metaphysics is an historical science, and if
history is the re-enactment of past thought, then metaphysics is
the re-enactment of past thought. But then, "reenactment" must
be taken in the sense that Collingwood meant for it, which does
not rule out critical thinking, as we have already seen.
651
that it rests on the central insight of his entire philosophy of
mind as a resolution of the realist-idealist dilemma (this will
occupy us in the two succeeding sections).
The preliminary sketch is presented as much by example
as by general description. The historian who sets himself the
task of understanding the Theodosian Code not only undertakes to
understand the document that he has inherited from the past,
i.e. the written words, but also to discover what the person who
wrote them meant by them, where the latter task also entails
envisaging for oneself the situation with which the emperor was
trying to deal: "he must see for himself, just as if the
emperor's situation were his own, how such a situation might be
dealt with; he must see the possible alternatives, and the
reasons for choosing one rather than another; and thus he must
go through the process which the emperor went through in
deciding this particular course" (IH, 283). Until he is able to
do this Collingwood insists that he cannot say that he has any
historical knowledge of the meaning of the edict.
In this description we have all the elements with which
we have become familiar--the original historical event with its
outside (the written code) and inside (the emperor's intentional
meanings), the representation of this event in the historian's
understanding by
652
re-thinking the universal element (the meaning of the code
as-understood by its author), and the necessity of
reconstructing that thought by understanding the context in
which it occurred (the emperor's situation with its possible
alternatives) and the agent's reasons for choosing this mode of
action rather than another (the emperor's choice of a legal code
to establish order rather than the use of military force under
his direct command, for example). Notice that this sort of
thinking is concrete (it never loses sight of its point of
reference--the decision of the emperor expressed in the written
code) at the same time that it is universal (it is concerned
throughout with the meaning of the code as a work of human
intelligence and purpose). Notice also that it proceeds by
locating its universal elements (the array of alternative
meaningful acts available in the situation the emperor was
facing). Are these not the embodiment of what we have
encountered in previous chapters as the "concrete universal"
(Speculum Mentis) or "categorical universal" (Essay on
Philosophical Method) in the first case, and the doctrine we
called "concrete abstraction" (from The New Leviathan) in the
second?
Although we have not yet examined the grounds for Col-
lingwood's remark that it is only on the condition that history
is the re-enactment of past thought that historical un-
derstanding is possible, we get a glimmering of his meaning in
this brief description. When he writes

653
that unless and until the event can be reconstructed, it cannot
be said to be historically understood at all, he is saying that
re-enactment is the necessary and sufficient condition for
thinking historically: necessary because without thinking out
the meaning of the event in its context of situational
alternatives, the significance of the event as historical (i.e.
as the action expressing the intentions of an agent) is
impossible to grasp; and sufficient because once that is done
nothing further is necessary--so that "historical explanation"
by subsumption of the event under a law governing other similar
events, as if it were a natural occurrence abstractly like
others of a class, marks the point at which the event loses both
its individuality as this event at this time (viz. the emperor's
situation) and its concrete significance (why the emperor chose
this course of action rather than another).
No doubt this leaves many loose ends, only some of which
we shall gather together in succeeding sections. A1though we
have not yet seen how Collingwood avoids the errors of the
idealist's view of history as he moves in his accustomed way
within the realism-idealism polarity, we foresee that unless he
expands on the role of the historian in viewing events a tergo,
and therefore capable of seeing more in them than did the
historical agent himself, his doctrine of re-enactment is in
danger of
654
appearing as nothing more than a repetition of the thoughts of
historical protagonists.9 In emphasizing the role of thought in
both the historical event and the historiographical re-enactment
Collingwood may have succeeded in damaging the realist's notion
of events unaffected by thinking, but he has done this by
emphasizing the way in which thoughts of agents and historians
are the same, which leaves the serious problem of how they
differ. This difference is apparent in their respective
linguistic modes of address: Caesar says, "I shall cross the
Rubicon;" the historian writes, Caesar said that he would cross
the Rubicon." He is aware of his own thought as distinct from
that of Caesar.
What is required is the doctrine of "encapsulation"
which the Autobiography would have us believe he had worked out
as the third "proposition" concerning history (A, 114), but
which is not mentioned by that name at all in The Idea of
History. If Encapsulation refers to the manner in which the
historian preserves a past thought within his own consciousness
without losing its aspect as past, and if this requires initial
__________________________
9 J. B. Bury could retrospectively see, for example,
that the Theodosian code and the founding of a university at
Constantinople were the two most important acts of Theodosius
II, because Bury saw in them the foundation that held civili-
zation together in the Eastern empire, while the Western fell
under successive invasions by barbarian tribes (A History of the
Later Roman Empire, quoted in The Historian's History of the
World, ed. Henry Smith Williams (New York, 1907), vol.VII, p.
45). It is unlikely that such an event was anticipated by
Theodosius, who nevertheless did not write the code--it was the
work of a committee of nine named by him for the purpose.
655
self-awareness, then it is clearly a function of what he calls
the a priori historical imagination, the notion of which we must
therefore consider next, bearing in mind that what Collingwood
must do is define the sense in which the historian's thought
retains its autonomy vis-a-vis the object of his historical
consciousness.
(e) The A-Priori Imagination.--Just as in the Kantian
critical philosophy, Collingwood felt that the first phase in
answering the question, "How is historical knowledge possible?"
is to show that there is a form of consciousness able to
re-present or re-think past thoughts, but not as simple
perception nor as memory nor even as abstract reasoning. From
our survey of Collingwood's philosophy of mind we know that any
sort of conscious re-presentation requires the functioning of
imagination. The topic of the historical imagination was chosen
by Collingwood as the subject of his Innaugural Lecture on being
appointed to the post of Waynflete Professor of Metaphysics in
1935, the text of that lecture being published by Knox as the
second Epilegomenon in The Idea of History.
We have already pointed out that while the first Ep-
ilegomenon (also a public lecture) is aimed at rejecting the
errors of contemporary realists concerning a "science of human
nature" on the natural model; the second is just as concerned
with correcting the errors of idealists on the subject--notably
those of Bradley. Collingwood praised Bradley as the

656
leader of the movement in England away from the positivistic
accumulation of facts and towards "scientific history," i.e.
history as a form of knowledge aware of itself as distinct from
natural science and yet valid in its own right (IH, 134-35). But
in his essay, The Presuppositions of Critical History 1O Bradley
mistakenly assumed that the criterion of history is experience
as informed by the knowledge of the laws of nature--a relic of
the positivism he sought to overcome. Induction of laws of
nature from observation can never give anything more than
probable laws, which fails to serve as a universal criterion for
what can or cannot happen in history, since much of history
deals with improbabilities (IH, 139). Collingwood points out
that Bradley's proposed criterion (a) does not adequately
distinguish history from fiction, since it claims to decide not
what did happen but only what could happen--which applies
equally well to fictitious narratives. (b) It leaves the
historian completely reliant on authorities, so long as their
accounts satisfy the negative criterion of being possible, which
hence leaves critical historiography unachieved. Finally, (c) it
leaves the historian unable to accept any unusual or improbable
experience not consistent with his own--e.g. odd social customs,
heroic deeds, improbable events, etc. (IH, 239-40; cf. IH, 139).
____________________
1O Recently reissued with introduction and commentary by
Lionel Rubinoff: Quadrangle Books (Chicago, 1968).
657
The thrust of Collingwood's Waynflete Inaugural Lecture
is therefore to propose an alternative criterion to that offered
by Bradley. It takes the form of an autonomous activity of
historical imagination different from perception (aisthesis) on
the one hand and scientific understanding (noesis) on the other,
and yet not reducible to the sort of history proposed by "common
sense" realists. Collingwood begins by distinguishing history
from perception, primarily to show that accounts of knowledge
that are based on the model of perception (e.g. the acquaintance
theory) make history impossible.
No doubt, historical thought is in one way like perception.

Each has for its proper object something individual . . . .
But what I perceive is always the this, the here, the now .
. . . Historical thought is of something which can never be
a this, because it is never a here and now. Its objects are
events which have finished happening, and conditions no
longer in existence. Only when they are no longer
perceptible to they become objects for historical thought.
Hence all theories of knowledge that conceive it as a
transaction or relation between a subject and an object
both actually existing, and confronting or compresent to
one another, theories that take acquaintance as the essence
of knowledge, make history impossible. (IN, 233).
We have been told in Part IV of The Idea of History that the
acquaintance theory of knowledge accepted one horn of the di-
lemma that Bradley bequeathed to his English successors: "either
reality is the immediate flow of subjective life, in which case
it is subjective but not objective, it is enjoyed but cannot be
known, or else it is that which we know, in which case it is
objective and not subjective, it is a world of real things
outside the subjective life of our mind and outside each other"
658
(IH, 141). Bradley accepted the former horn and realists (Cook
Wilson and Oxford realism on the one hand and Bertrand Russell
and Cambridge realism on the other) embraced the latter (IH,
141-42). Collingwood writes that Samuel Alexander admirably
expressed the acquaintance theory when he wrote that knowledge
is a relation between two things, a mind and its object, and the
mind therefore does not know itself, it only enjoys itself.
Collingwood hastens to add that such a view makes history as the
self-knowledge of mind (i.e. the philosophical concept of
history) impossible (IH, 142).
But notice that while he rejects this view he does not
deny that the object of history is something individual; he
merely neglects to tell us what that individuality is. We have
already found this designated as the individual meaning of an
historical event expressive of the thought of an agent He is
also not denying that there is some sense in which the objects
of history are events which are present, viz. as evidence. To
adopt a terminology that is not Collingwood's, evidence is not
something in itself, it is only evidence for another, i.e. for
an historian who recognizes it as such, so in this sense
historical evidence is not something co-present with a mind, but
is rather something dependent on it.
If it is not "aisthesis" historical knowledge is also
not "noesis," and in countering the latter Collingwood opposes

659
another tendency, one more prevalent today, in the guise of the
covering law model" of Popper and Hempel,ll than it was even in
Collingwood's day. Collingwood rejects the thesis that
historical events are "explained" like natural events by
deducing them from antecedent conditions and general laws.
In another way history resembles science: for in each of

them knowledge is inferential or reasoned. But whereas
science lives in a world of abstract universals, which are
in one sense everywhere and in another nowhere, in one
sense everywhere and in another nowhere, in one sense at
all times and in another at no time, the things about which
the historian reasons are not abstract but concrete, not
universal but individual, not indifferent to space and time
but having a where and when of their own, though the where
need not be here and the when cannot be now. History,
therefore, cannot be made to square with theories according
to which the object of knowledge is abstract and
changeless, a logical entity towards which the mind may
take up various attitudes. (IH, 234).
Notice again that Collingwood is not saying that history is not
reasoned or inferential, but that its subject matter is not
something abstract: it is something concrete and individual, an
act performed at a certain time and place. Later in this chapter
we shall examine what Collingwood understands by both historical
evidence and historical inference, but here we need to recognize
only that historical understanding is neither "aisthesis" nor
noesis" nor a combination of the two, but rather "a third
thing," i.e. "wholly reasoned knowledge of what is transient and
concrete" (IH, 234).
________________________
1l See Theories of History, ed. by Patrick Gardiner (New
York, 1959) pp. 276~85, 344-56, 428-43.
660
Such a view of historical thinking also breaks with the
"common sense" view of history, in which memory and belief in
authoritative testimony are taken as the essential functions of
the historian's thought. The scientific historian does not
merely repeat his acquaintance with the events first perceived
by a witness, then remembered, recollected, and repeated to
someone else and believed to be true. The contemporary historian
is aware that he must tamper with his authorities by selecting
what he considers important in the accounts of his authorities,
by interpolating in them things which are not explicitly said,
and by criticizing, rejecting, or amending what he recognizes to
be misinformation or outright falsehood (IH, 235). In this work
of selection, construction, and criticism the historian exhibits
his own autonomy, showing that his thought possesses "a
criterion to which his so-called authorities must conform and by
reference to which they are criticized" (IH, 236). And in all
three of these functions he gives evidence of the working of an
autonomous or a priori imagination--that is, a form of
consciousness that is a priori in the Kantian sense of being
indispensable or necessary (IH, 240).
Collingwood's portrait of the functions of the a priori
imagination is interesting for the light it sheds on the cri-
tical aspects of historical re-enactment. He points out that
historical imagination differs from two other functions of a
priori imagination, namely artistic

661
and perceptual imagination, not in being a priori, since the
artist's work has its own inner necessity, and in perception one
cannot help supplementing the "data" of perception by presenting
objects of possible perception which are not actually perceived
(e.g. the underside of a table, the back of a cube, the inside
of an unopened egg--imaginative functions well analyzed by
Kant), but "in having as its special task to imagine the past:
not an object of our thought" (IH, 242).
In this exercise the historian stretches his web of
imaginative re-construction between points which are at first
assumed to be fixed--the statements of "authorities" he uses as
his sources (IH, 242). But on analysis these fixed points also
resolve themselves into achievements of historical thinking
itself; statements first accepted as settled for the purposes of
a given inquiry are themselves called into question in other
contexts (IH, 244). It then appears that the fixed points are
not accepted facts or authoritative statements but criteria used
to justify the use of such authorities. But Collingwood insists
that these criteria are also provided by the a priori historical
imagination.
The a priori imagination which does the work of historical

construction supplies the means of historical criticism as
well. Freed from its dependence on fixed points supplied
from without, the historian's picture of the past is thus
in every detail an imaginary picture, and its necessity is
at every point the necessity of the a priori imagination.
Whatever goes into it, goes into it not because
662
his imagination passively accepts it, but because it

actively demands it. The resemblance here between the
historian and the novelist . . . here reaches its culmin-
ation. Each of them makes it his business to construct a
picture which is partly a narrative of events, partly a
description of situations, exhibition of motives, analysis
of characters. Each aims at making his picture a coherent
whole . . . . The novel and the history must both of them
make sense; nothing is admissible in either except what is
necessary, and the judge of this necessity is in both cases
the imagination . . . . Where they differ is that the
historian's picture is meant to be true. (IH, 245-46).12
With the raising of the issues of historical coherence
and truth it is clear that we have passed beyond discussing the
concept of history as a scale of forms. Collingwood immediately
points out that "being true" in this context means that the
historian abides by three "rules of method" in the construction
of his historical narrative: (1) his picture of the past must be
localized in space and time (i.e. it must refer to a
_____________________
12 Collingwood has been accused of emphasizing the ac-
tive role of the historian to the neglect of his passive func-
tion of relating received narratives--see W. B. Gallie,
Philosophy and the Historical Understanding (New York, 1964),
pp. 18-19. But Collingwood anticipated this objection. "The his-
torian, generally speaking, works at a subject which others have
studied before him. In proportion as he is more of a novice,
either in this particular subject or in history as a whole, his
forerunners are, relatively to his incompetence, authorities;
and in the limiting case where his incompetence and ignorance
were absolute, they could be called authorities without
qualification. As he becomes more and more master of his craft
and his subject, they become less and less his authorities, more
and more his fellow students, to be treated with respect or
contempt according to their deserts" (IH, 238). As the above
passage also makes clear, Collingwood's use of the term
"picture" of the past (part of which consists in a narrative of
events) is obviously a synonym for Gallie's "narrative;" cf. IH,
242, 245.
663
particular time and place); (2) it must be consistent with
itself; and (3) it must stand in a relation to evidence such
that the evidence, consisting of something here and now
perceptible (written documents, artifacts, ruins, etc.), can be
bound to the events by chains of historical inference (IH,
246-47). All three of these are rules to which the novelist need
not subscribe, and all three are criteria which are nowhere
"given" to the historian along with his "facts." They are not
among the statements of his authorities, but are rather criteria
for accepting such statements as "historical" in the first
place. They are, like Kantian categories, conditions for the
possibility of there being historical data at all.
Several comments are in order here. First, Collingwood
calls these three statements "rules of method," but it is clear
that they are presuppositions of historical inquiry rather than
procedural imperatives. They stand, relative to all progressive
questions in an historical inquiry, as their ultimate
presuppositions, and they are revealed only when the direction
of the inquiry is reversed. The question of who was the victor
of the battle of Waterloo arises only on the presupposition
(among others) that there was such a battle at a certain time
and place, as indicated by interpretation of certain evidence.
But if the historian were challenged (perhaps by a novelist who
is free to cancel the event in his imagination and pursue the
dramatic consequences) to state his reasons for

664
assuming that there was such a battle he would haul out his
evidence and begin constructing an elaborate argument aimed at
the conclusion that it did indeed occur. Now it would take
considerable Socratic cheek to force out the further question
about why he presumes that a battle must either have occurred or
not occurred, but if the interlocutor managed to do so, and in
the process completely reversing the direction of the inquiry,
the historian would (if he replied at all) say that history must
after all make "sense" (i.e. be consistent with itself). He
would, in short, be driven to display his presuppositions.
Secondly, it is noteworthy that these presuppositions
occur in a discussion of the autonomous a priori imagination,
even though it involves the relationship between propositions,
questions, and presuppositions, and involves the necessary re-
lations between them. This is not accidental; it is what he
means when he says that "the a priori imagination which does the
work of historical construction supplies the means of historical
criticism itself" (IH, 245), and it is what is involved when he
insists that in his thought the historian is engaging in an
autonomous activity of a priori imagination (IH, 249). This
confirms what we suggested earlier, that the functions of
imagination are not limited to the conceptual level of mental
acts, or to the pre-propositional level of linguistic
expression, but extend to higher intellectual functions as well.

665
Thirdly, they are a priori criteria, and this term ap-
pears to encompass a number of meanings. Insofar as historical
imagination is a priori it is said to be something not em-
pirically received (IH, 248) but rather "innate" or original to
the mind (IH, 247); it is not arbitrary, but necessary-something
clear, rational, and universal (IH, 240, 242, 248-49); it is a
criterion brought to bear in judgments concerning matters of
fact or evidence, and is used in interpolations and
constructions about these matters of fact (IH, 138, 240-41,
248); it is an original and fundamental activity of mind itself,
and is therefore a self-determining, self-justifying,
self-dependent form of thought (IH, 247, 249); and yet the
principles it employs are not finally fixed, but are capable of
change (IH, 248).
Finally, we must note that in The Idea of History we are
left largely unenlightened about the exact relationship between
historical presuppositions and the a priori imagination. We do
not know if the presuppositions of history can be regarded as a
priori in the same sense that historical imagination is, or if
on the contrary they can be regarded as products of the latter
(and hence at least in this sense posterior to its activity). It
is not difficult to see that if history is to be an autonomous
science, its presuppositions must be clear and rational,
universal and necessary (in some senses of these terms
acceptable to Collingwood). It is also clear that they are not
empirically received and

666
that they are used as criteria in the judgments of history
concerning matters of fact. But we do not know in what sense
Collingwood sees them as "innate" or original to the mind, or
how they are part of a self-determining, self-justifying,
self-dependent form of thought.
Nor is this an extraneous issue for Collingwood's
philosophy of history. We noted at the beginning of this section
that in this essay one of Collingwood's major goals was to state
the "criteria" of history alternative to those proposed by
Bradley (IH, 238-89). The criteria he proposes turn out to be
strangely enigmatic. On the one hand he says that the criterion
of historical truth is "the idea of history itself: the idea of
an imaginary picture of the past," which he calls innate in the
Cartesian sense and a priori in the Kantian sense (IH, 248). But
as if he knew that this idea was insufficient to say not only
how that idea differed from that of the novelist (who also has
an "imaginary picture of the past"), but how one uses such an
idea to judge historical evidence, he also proposes his three
"rules of method" or presuppositions of history as a science. It
is only the latter that serves to correct Bradley's earlier
attempt to state such a criterion, and it is only such
presuppositions which can in any real sense be said to "change."
Once again, Collingwood seems to stop just short of the sort of
questions that are"metaphysical"

667
in the sense of being ultimate: how are activities of
imagination responsible for presuppositions or vice versa? How
can presuppositions be both necessary and changeable? Why are
facts and evidence necessary for an "imaginary picture of the
past"?
(f) Re-enactment: Beyond Realism and Idealism.--Although
in Knox's editorialized version of The Idea of History we are
left at this point without direct answers to these questions,
there is an heroic assault on the question that Collingwood
regarded as the ultimate task of a philosophy of history to
answer. Just as Kant in his Prolegomenon assumes that a certain
kind or kinds of knowledge exists and goes on to ask the further
question how it is possible, Collingwood in his Epilegomena
assumes that his survey of historiography in Parts I-IV has
sufficiently demonstrated that historical knowledge exists, and
goes on to ask how it is possible. Hence the fourth
Epilegomenon, entitled "History as Re-enactment of Past
Experience," begins with the question, "How, or on what condi-
tions, can the historian know the past?" (IH, 282). His answer
to this question is his reasoned justification of historical
thinking. In the fourth Epilegomenon this takes the form of a
series of arguments put forward as answers to objections about
re-enactment from realistic and idealistic standpoints.

668
In proceeding sections and chapters we have tried to
indicate why Collingwood believed that historical thinking is
not possible on realistic presuppositions. From his earliest
publications Collingwood maintained that any epistemology which
tried to justify historical knowledge by assimilating it to
perception, as if historical events were present to con-
sciousness in the same manner as perceived objects, overly
concretizes the past and is therefore utterly misleading. In the
second sentence of the present essay he calls our attention once
again to the fact that the past as past simply does not
exist--its events are not present but past: "the past is never a
given fact which we can apprehend empirically by perception"
(IH, 282). Nor does the historian merely repeat what is said by
an eyewitness of the facts, since the facts he is concerned with
are past events mediated by critical evaluation of evidence,
involving inference on the part of the historian --inference
about the value and reliability of his evidence. His knowledge
is not direct or immediate in an empirical or perceptual sense,
but rather is mediated or inferential.
But if the past known by an historian is mediated by
historical thinking, it is not simply absorbed by that thinking,
and in the attempt to preserve historical objectivity we find
Collingwood taking the unusual tack of granting the realistic
devil his due. Collingwood agrees that the past must be
preserved in its aspect as past, and not utterly assimilated

669
into the present. He therefore finds it necessary to take a
stand opposed to views of history offered by some of his
idealistic contemporaries. Although he agrees that a more
fruitful line of inquiry is initiated when it is recognized that
our imaginary picture of the past is an a priori idea, so that
justification of historical thinking takes the form of showing
on what grounds such an idea is possible, he finds himself
unable to accept those proposed by some idealists. We have just
seen why Collingwood rejected Bradley's presuppositions of
critical history as inadequate for grounding scientific history.
In his review of historiography in Parts IIV Collingwood also
criticized idealistic philosophers of history like Croce and
Oakeshott for absorbing the past into the present: by
overemphasizing the role of historical thinking in the
construction of historical narratives they made the past merely
an aspect of the present thinking of the historian (IN, 154-55,
202, 289).
So where the realist is forced, by his assumption that
all true knowledge is based on immediately perceived objects, to
overly concretize the past and underplay the role of historical
thought in re-creating that past, the idealist makes the
complementary error of dissolving the past altogether by
reducing it to an aspect of present consciousness. What makes
the essay we are presently concerned with such a remarkable
piece of thinking is Collingwood's care to avoid both of these
errors. The key to

670
understanding the complex argument of this essay is to bear in
mind that in it he is trying to overcome the dilemma that
Bradley bequeathed to his successors and which, in his treatment
of scientific history in Part IV, Collingwood had singled out as
crucial (IH, 141). His escape from this dilemma takes the form
of three arguments in the fourth Epilegomenon --one overall
argument aimed at overcoming Bradley's dilemma, and two
subsidiary arguments aimed at refuting objections to
re-enactment by realists and idealists respectively. All three
arguments draw on Collingwood's theory of mind and his analysis
of acts of consciousness, and brilliantly illuminate its central
insight.
Since the arguments are complex, and since in the essay
the main argument is reserved until the end, we shall give here
a brief synopsis of all three arguments, and then take them up
again in the order that Collingwood gives them in the text.
Bradley's dilemma, as Collingwood states it, is that
reality is either the flow of subjective life, in which case it
is subjective but not objective, or else it is that which we
know, in which case it is objective but not subjective, a world
of real things outside the subjective life of our minds.
Collingwood answers that historical reality escapes this dis-
junctive description because it is both subjective and objec-
tive. This is so because the object of history is itself
thought, i.e. the intentional

671
meaning of an historical agent expressed as an event localized
in space and time (its "outside") and having an "inside"
consisting of thought. Because this object is thought it can be
re-thought or revived or reenacted in further acts of thought by
the agent himself or by another, e.g. by an historian critically
interpreting evidence.
But to be truly historical a re-enacted thought must be
known to be both the same thought as that of the historical
agent, and yet distinguishable from that of the historian; for
if it were not the same then the historian could never be sure
that he has grasped his object (and hence historical objectivity
would be sacrificed), and if it were not different he could not
distinguish it from his own thought (and the past would be
absorbed into the present thought of the historian). The realist
denies that two acts of thought can be the same (for no two acts
of thought can be literally identical) and the idealist denies
that they can be different. On either premise historical
thinking is impossible.
But an act of thought is not something wholly immediate,
like the flow of consciousness involved in mere feeling. Our
experience as thinkers shows us that an act of thought is
inseparable from its object, i.e. meanings as expressed in
physical acts, statements, arguments, etc. An act of thought is
something someone does deliberately, and it is not irretrievably
carried off in the flow of feelings; on the contrary it is that

672
which has the power to survive and revive over time (two
characteristics of all thought from his philosophy of mind), and
in various contexts. Nor is an act of thought something wholly
mediate, totally dependent on its context for its meaning. An
act of thought is capable of retaining its identity of meaning
in various contexts. Therefore thought is both mediate and
immediate. Historical thought as re-enactment is thus possible
because the historian is concerned with an object which has the
power to revive in other contexts, and the historian has the
power to re-think that thought in the context of his own
thoughts. Historical thinking is therefore both objective (the
historian can distinguish himself from his object because the
object can retain its identity in various thought-contexts) and
subjective (the historian's re-enactment occurs in the context
of his own thinking and he knows it as his own experience).
History is possible because re-enactment of thought is possible,
and re-enactment presupposes an object that is both mediate and
immediate, objective and subjective. Such an object is within
our own experience as thinkers, and such a reality escapes
Bradley's dilemma.
Now one cannot help but notice from this synopsis that
there are several key concepts on which the whole argument is
hinged, and that the success of the argument depends on how
Collingwood handles these concepts.

673
Ultimately, as we shall see, they depend on his description of
the several functions of mental activity. (1) In particular the
success of the argument depends on the meaning of the
expression, "act of thought," and in the first objection to
re-enactment Collingwood calls attention to the controversy
surrounding this phrase. The first objector (invented by
Collingwood, but evidently a realist) assaults re-enactment as
an ambiguous concept which ignores the act-object distinction:
To re-enact an experience or re-think a thought, he might

argue, may mean either of two things. Either it means en-
acting an experience or performing an act of thought re-
sembling the first, or it means enacting an experience or
performing an act of thought literally identical with the
first. But no one experience can be literally identical
with another, therefore presumably the relation intended is
one of resemblance only. But in that case the doctrine that
we know the past by re-enacting it is only a version of the
familiar and discredited copy-theory of knowledge . . . .
In every experience, at any rate so far as it is cognitive,
there is an act and an object; and two different acts may
have the same object . . . . The two acts are different
acts but acts of the same kind. They thus resemble one
another, and . . . hence the conclusion that the doctrine
we are considering is a case of the copy theory of
knowledge. (IH, 283-85).
As an example of a thought that can be re-enacted Collingwood
suggests Euclid's proposition that the angles at the base of an
isosceles triangle are equal l3--which we shall abbreviate as
EP. An objector upholding the
________________________
13 Collingwood's examples are not always felicitous, and
in this case one might, with some justice, accuse him of
choosing a judgment that prejudices the argument by dealing with
mathematical entities. Why not a more typical case like Caesar's
decision to cross the Rubicon? In Collingwood's defense it might
be pointed out that the case is not
674
act-object distinction, he argues, would naturally enough
understand re-enactment as a case of a relation between two
thoughts such that the enacted thought and its re-enactment are
numerically different but specifically identical. Since they
have the same object in mind (EP) they are instances of a single
species or class (the class of all instances of EP); but since
they occurred at two different times or to two different
persons, they are numerically different. But this is not the
only sort of identity-in-difference relation that can exist
between mental acts, and in order to demonstrate another variety
of this relation Collingwood suggests considering three cases:
Case 1: I think EP for five seconds continuously.
Case 2: I think EP for five seconds continuously, then

cease thinking about it for three seconds, then
think EP.
Case 3: Euclid thinks EP and after a lapse of several

centuries I think EP.
Collingwood argues on the basis of the first case that the unity
of the
____________________
exceptional, since an historian of geometry could very well be
rethinking EP in rethinking Euclid's contribution, and in Col-
lingwood's view, intellectual history is just what history
is--all history is the history of thought. Furthermore the
example is chosen for its appeal to just those who regard ma-
thematical propositions as exemplary of the act-object dis-
tinction (Frege, Russell, etc.). Finally, it would not be
difficult to rephrase the argument using an example of a more
recognizably typical historical act instead of EP, as he does in
the fourth Epilegomenon, where he uses the example of the
Theodosian Code.
675
act of thinking EP is not dependent on the duration of time
during which it is thought. "There is no more reason to
correlate the unity of a single act of thought with the
time-lapse of one second, or a quarter of a second, than with
any other. The only possible answer is that the act of thought
is one sustained through five seconds"--that is, it has the
"identity of a continuant" (IH, 286). From the second case
Collingwood concludes that if one is to say that EP is the same
act of thought then one is forced to say that the act of thought
is not merely sustained, but revived after an interval.
(Presumably, to deny this would involve one in an infinite
regress similar to Plato's "third man" argument.)
Now where the act-object distinction fails adequately to
describe the sort of experiences we have as thinkers in the
first two cases, its defenders would argue that the third dif-
fers significantly from the previous two, and for this case
Collingwood puts forward his anti-realist argument in the mo-
dified form in which we stated it in Chapter VII:
The objector . . . maintains that although the object of

two people's acts of thought may be the same, the acts
themselves are different. But in order that this should be
said, it is necessary to know "what someone else is
thinking" not only in the sense of knowing the same object
that he knows, but in the further sense of knowing the act
by which he knows it: for the statement rests on a claim to
know not only my own act of knowing but someone else's
also, and compare them. But what makes such comparison
possible? Anyone who can perform the comparison must be
able to reflect "my act of knowledge is this"-and then he
repeats it: "from the way he talks, I can see that his act
676
is this:--and then he repeats it. Unless that can be done,

the comparison can never be made. But to do this involves
the repetition by one mind of another's act of thought: not
one like it (that would be the copy theory of knowledge
with a vengeance) but the act itself. (IH, 288).
Therefore in order to affirm that the third case is not like the
first two, one would have to presuppose that one simultaneously
claimed to deny, i.e. that one act of thought can be shared by
two different people separated by a span of time. Collingwood
concludes that thought can never be a mere object, since to know
what someone is thinking or has thought involves thinking it for
oneself, and to deny that this is possible is to back onself
into a solipsistic corner in which it only possible to think
one's own thoughts (IH, 288-89). Collingwood is content to leave
his realistic objector parked in this corner.
Before taking up the objection to re-enactment by the
idealist, we would like to pause for several observations about
the course of this argument so far.
Notice first that Collingwood does not further pursue
the issue of abstract universality (specific identity with nu-
merical difference), nor does the argument based on the three
cases of re-enacting acts of thought really depend for its
conclusion on this incomplete discussion. What Collingwood left
unsaid was that the kind of identity-in-difference that is
involved in re-enactment is a case wherein the second thought is
not
677
merely a copy of the first, nor is it another replacement
instance of the class of thoughts that includes the first, but
rather the second includes the first as part of its content. It
is almost trivially true to note that when as an historian I
assert "Euclid thought that EP" the second proposition (HP)
includes the first (EP) as part of its content. What Collingwood
is trying non-trivially to point out is that when someone asks
how this expression is possible and answers it, as the realist
does, by denying that the EP used in the first assertion can be
identical to the EP mentioned in HP, he has explicitly
contradicted himself.14
But of course the realist does not assert that they are
unqualifiedly not identical, but only that they are not
identical acts of thought; they are identical in having iden-
tical objects. Therefore Collingwood bears down on the phrase
"set of thought," and our second observation pertains to this
phrase. When the realist insists on the abstract distinction
between an act of thought and its object he opens himself to the
criticism that the act of thought as a mere event has no direct
relationship to the object of the act, and Collingwood's
______________________
14 Collingwood gives no indication that he is cognizant
of the use-mention distinction as such, but the argument ad-
dressed to the realist objector presumes that the historian
thinking HP does not merely mention EP but also uses it in the
sense of enacting the proposition for himself. Although the
historian is not using EP to make a geometrical point as part of
a systematic inquiry in that subject, he must understand it
sufficiently to grasp its meaning for the development of geo-
metry, the history of which he is narrating.
678
strategy is to call attention to this epistemic gap and its
necessary presupposition that somehow the event and the object
both belong to the same thought. In the ensuing argument with
the idealist objector Collingwood will make an equivalent and
complimentary point about the object of an act of thought, and
displaying the unity of these two sides to thought--as an act
and as an object--is the way in which he resolves the Bradleyan
dilemma.
Thirdly, we wish to note in passing that Collingwood
appears to be employing the sort of argument that he had char-
acterized as philosophical in the Essay on Philosophical Method.
Even though the argument is incomplete at this point we are
aware that, for example in the three cases of a reenacted
thought, a conclusion (the impossibility of re-enactment) is
being checked by comparing it to experiences that should
anticipate it (the three cases of our experience as thinkers);
and the principle from which the conclusion is drawn (the
act-object distinction) is as a premise being revised (it is
not, as an abstractly universal principle, defensible vis-a-vis
our actual or concrete thinking, and is in need therefore of
revision). We are witnessing an argument in the process of
reversing itself (revision of the act-object distinction)--not
an abstract reductio ad absurdum that merely negates a premise,
but one which aims at revising a starting point in the light of
its conclusions so that it can explain

679
how experience as it actually exists is alone possible (EPM,
156). How the premise is revised we are about to see.
Finally we have to call attention to Collingwood's
phrasing of the anti-realist argument in the quotation above,
where he says that the assumption that two acts of thought can-
not be the same rests on the ability of the objector to perform
a comparison which involves asserting in addition to one's own
act the further one, that "from the way he talks, I can see that
his act is this" which he then repeats (IH, 288). Is there not
some significant qualification overlooked in this deceptively
simple argument, and is this not disguised under the simple
phrase, "from the way he talks I can see that . . . ."? It
appears that I can know what I mean by an act of thought, my own
thought, by simply repeating it and examining it; calling it an
"act of thought" in the context of Collingwood's philosophy of
mind precisely means that it is something that I can do on
purpose--it is an achievement. But how is it possible to achieve
someone else's act of thought, or do his thinking "on purpose"?
It seems I can only know what someone else means by his act of
thought by interpreting what he says he means, and in that
interpretation I may be mistaken, since meaning for Collingwood
is intentional. Can I miss another's intention in a way that I
cannot miss my own?
We must keep this in mind as we proceed, but if there is
an assumption here it is the same assumption that is made,
Collingwood would
680
surely say, in re-thinking my own thoughts, since I may be
mistaken in interpreting my present intentional meaning as the
same one I had several moments ago. Collingwood's point is that
to say that it is not the same is to presume that one knows what
that act was, i.e. "not this," and for that difference to be
sustainable one must see it as a meaning difference, one
graspable by a comparing act of thought which holds both
together and declares them to be non-identical or different in
some discernible or meaningful fashion. The mere temporal
counting is not a meaningful difference between EP as thought by
Euclid and as re-thought by me. To say it is the same meaning is
to say something essential about the act of thought, not its
accidental temporal correlation.
Nonetheless, one wishes to persist against Collingwood,
is there not some significant difference in the fact that Euclid
thinks EP and I think EP? For this we turn to the next phase of
the argument.
(2) Having disposed of one aspect of the act-object
distinction, Collingwood proceeds to take up the cudgels against
the idealist objector to re-enactment. Where the realist would
claim that re-enactment proved too little (because it failed to
show how two acts of thought can be identical), the idealist
would claim that it proved too much.

681
It has shown that an act of thought can be not only per-

formed at an instant but sustained over a lapse of time;
not only sustained but revived; not only revived in the
experience of the same mind but (on pain of solipsism) re-
enacted in another 's. But this does not prove the possi-
bility of history. For that, we must be able not only to
re-enact another's thought but also to know that the
thought we are re-enacting is his. But so far as we reenact
it, it becomes our own; it is merely as our own that we
perform it and are aware of it in the performance; it has
become subjective, but for that very reason it has ceased
to be objective; become present, and therefore ceased to be
past. (IH, 289).
In order to leave no doubt in the reader's mind that this is an
idealistic objection, Collingwood points out that this is just
what leading idealistic philosophers of history were maintaining
when they asserted that history is one's own experience arranged
sub specie praeteritorum (Oakeshott) and that all history is
contemporary history (Croce) (cf. IN, 151-59, 190-204).
What is at issue here is the very past-ness of the past.
Collingwood is willing to grant the idealist's demand that in
order for a re-enacted thought to be historical it must be known
that it was thought not only by myself but originally by the
historical agent. Propositionally this is the demand that the
difference between the assertions of EP and HP must be
preserved, since merely to re-think EP would be to think
geometrically rather than historically. Collingwood goes even
further down the path toward idealism insofar as he affirms that
thinking historically is a self-conscious process: "unless he
knows that he is thinking historically, he is not thinking
historically. Historical thinking

682
is an activity . . . which is a function of self-consciousness,
a form of thought possible only to a mind which knows itself to
be thinking in that way" (IH, 289). This much he had to admit on
the grounds of his own philosophy of mind, which argued that
reflective thought is criteriological, i.e. contains within
itself the criterion for judging itself to be a successful or
unsuccessful piece of thinking.
But Collingwood insists that the idealist objector is
making a second point, that this necessary condition for his-
torical knowledge cannot be fulfilled if historical thinking is
re-enactment because re-thinking a thought for oneself prevents
one from recognizing it as another 's. On the basis of our
abbreviated propositions, the idealist appears to be saying that
because one asserts EP one cannot also assert HP, because
thinking EP prevents one from doing so (in becoming my thought
it becomes subjective and therefore ceases to be objective or
something shared with someone else). Collingwood adds that if
the idealist were maintaining that one of the conditions for
historical knowledge to occur is that the historian first
mistakes his agent's thinking for his own and then overcomes
this error by recognizing that his own thought is distinguished
from that of Euclid's, then it might be of interest as showing
one way that re-enactment could be a pre-condition for
historical thought. "But the re-enactment of past thought is not
a pre-condition
683
of historical knowledge, but an integral element in it; the
effect of the ((idealist's)) contention, therefore, is to make
such knowledge impossible" (IH, 290).15 If we cannot know
ourselves to be engaged in re-enacting the thought of historical
agents, we cannot be said to be thinking historically, and the
idealist's contention is that we cannot have such knowledge.
Collingwood sees the objection as hinging on the
impossibility of an act of thought being both subjective and ob-
jective. Being subjective means being aware of an experience as
my own; therefore, since "experience" may mean several things,
"being aware of an experience"
_____________________
15 Collingwood is here tacitly rejecting the view he
held in Speculum Mentis concerning the way in which the mind
attains to truth. In the earlier work he argued that history,
like all other forms of experience, proceeds through the via
dolorosa of error to truth. "The progressive alienation of the
mind from its object is in history complete. The world is
triumphantly unified as ((an individual, concrete, infinite))
object, only to find itself separated from the mind by a gulf
which no thought can traverse. But in this process, which seems
to travel at every step further from that intimacy of subject
with object which constitutes knowledge, the indispensable
condition of knowledge is progressively and inversely realized .
. . . For an infinite given whole of fact cannot at any point be
grasped by the mind." (SM, 238-39). Although he adds that this
gulf is bridged from both sides (which foreshadows a doctrine
about evidence in The Idea of History, as we shall see), these
passages seem to point to the zero-line of complete scepticism,
saved only (so the early Collingwood thinks) by the existence of
absolute knowledge. But in The Idea of History in general,
Collingwood gives not the slightest indication that re-enactment
is an exercise of the "absolute standpoint," or that absolute
idealism had found the solution to the problem of necessary
historical error. Cf. Rubinoff, CRM, 292-306.
684
is also an equivocal expression. It may mean (1) the immediate
feeling of something such as a pain; or (2) the perception of
objects; or (3) the self-conscious awareness that is involved in
acts like being aware of losing one's temper--the third sense
being the most proper English use of the term (IH, 291). We here
recognize Collingwood's analysis of consciousness as a scale of
forms of mental acts. In this case he is interested in the sense
of the term "awareness" as it applies to historical knowledge,
where at least that degree of self-consciousness must be present
which allows the historian to distinguish his own thought from
that of the historical agent he is studying. Historical thought
must have at least that degree of subjectivity that allows one
to recognize an act of thought as not merely an experience, but
my experience; and it must have that degree of objectivity that
allows one to recognize this same act of thought as having a
certain cognitive character, a determinate meaning, etc. (IH,
291). Collingwood's justification of this kind of thinking takes
the form of exploring the contradictions that arise from
attempting to deny that a thought can be both subjective and
objective.
Indeed to say that would be to contradict oneself. ((1)) To

say that an act of thought cannot be objective is to say
that it cannot be known; but anyone who said this would be
claiming thereby to state his knowledge of such acts. ((2))
He must therefore modify it, and will perhaps say that one
act of thought may be an object to another act, but not to
itself. But this again needs modification, for any object
is properly the object not of an act but of an agent, the
685
mind that performs the act. True, a mind is nothing except

its own activities, but it is all these activities
together, not any one separately. ((3)) The questions is,
then, whether a person who performs an act of knowing can
also know that he is performing or has performed that act.
Admittedly he can, or no one would know that there were
such acts, and so no one could have called them subjective;
but to call them merely subjective, and not objective too,
is to deny that admission while yet continuing to assume
its truth. (IH, 291-92).
Notice that in both (l) and (3) Collingwood tacitly
appeals to his anti-realism argument and uses it against an
idealist objector--thus supporting Post's suggestion that the
argument cuts both ways. But the "together" in (2) calls for
special comment, since it is revealing about Collingwood's grasp
of the relationship between individual acts of consciousness and
the contexts in which they occur. In Collingwood's philosophy of
mind it is axiomatic that "thought is not mere immediate
experience but always reflection or self-knowledge, the
knowledge of oneself as living in these activities" (IH, 297)--a
principle quite contrary to any view of conscious acts which
attempts to analyze them as atomically distinct from one
another. Nevertheless, Collingwood insists, they do retain a
certain identity. Although an act of thought occurs at a given
time, and in a context of certain other acts of thought, of
emotions and sensations and memories, etc., the peculiarity of
thought is that it can sustain itself through a change of
context and revive in a different one. As true as it may be that
the immediacy of thought in its given context cannot be
re-enacted, the self-identity of an act of thinking is

686
independent not only of its context of feelings and emotions but
also its context of other thoughts.
This has sometimes been denied. It has been said that

anything torn from its context is thereby mutilated and
falsified . . . . Others . . . have embraced the opposite
doctrine that makes it both easy and legitimate to detach
them from their context; for there is no context; there is
only a juxtaposition of things standing to one another in
merely external relations. On this view, the unity of a
body of knowledge is only that kind of unity which belongs
to a collection: and this is true both of science, or
system of things known, and of a mind, or system of acts of
knowing. (IH, 298-99).
We do not have to be told that the view that holds that
"anything torn from its context is thereby mutilated and
falsified" is that of the idealists from which Collingwood
wishes to distinguish himself. In criticizing this view he
indicates as plainly as one might expect that he could no longer
affirm a view accepted in unregenerate days by himself (cf. IH,
299)--for this in fact is the position he maintained in Speculum
Mentis, where abstraction and falsification are identified.16
But notice
__________________________
Cf. Rubinoff, CRM, 112-13, 147-49, 297-99. Rubinoff
recognizes the problem of the difference between Speculum
Mentis, in which history is a knowledge of facts independent of
mind and therefore abstract and false, and The Idea of History,
in which history is the "knowledge of a part of mind itself"
(CRM, 297). But Rubinoff's solution strikes us as fantastic and
ungrounded: he sees both works as part of his grand scheme of
the "three ontological levels of consciousness" and therefore
bound together in the necessity of "the dialectic." He writes:
"This dialectical requirement is a result of the inherent
contradiction which . . . pervades the whole of Speculum Mentis,
between the presuppositions of realism (for which there is a
distinction between subject and object) and the presuppositions
of idealism (for which this distinction is overcome). When this
conflict is dialectically resolved, consciousness elevates itself
687
that he immediately castigates the opposite view that all acts
of thought are atomically distinct from one another. He writes
that such a view "by substituting logical analysis for attention
to experience overlooks the immediacy of thought, and converts
the act of thinking, from a subjective experience, into an
objective spectacle" (IH, 299).17 Whatever Collingwood's
familiarity was with logical atomism and
___________________________
to the third ontological level of existence. Once having arrived
at this level, history reconstitutes itself, this time under the
influence of philosophy which is the final consummation of the
rapprochement between subject and object" (CRM, 298). Aside from
the unrestricted use of the misleadingly florid language of
Hegelianism in passages like this one (in talk about the
dialectic, consciousness elevating itself or arriving somewhere
where history reconstitutes itself and reaches its final
consummation --metaphors inadequately resolved by Rubinoff and
nowhere found in The Idea of History), one wonders why, if
Collingwood had such a scheme in mind, he nowhere said so
explicitly--a remark which Knox, himself a Hegel scholar, would
certainly have passed on to us. We find it more likely that the
reconciliation between idealism and realism which Collingwood,
judging from the essay we are considering, surely aimed to
achieve, did not occur by absorbing the one into the other or
both into the "absolute standpoint (itself an abstraction) or a
"third ontological level" but rather by the correct analysis of
historical thought as the re-enactment of expressions of acts of
thought in which meaning grounds both its subjective and
objective dimensions--its aspects as my experience and as
universally comprehensible language. In fairness to Rubinoff we
must add that on occasion he makes it clear that his "tran-
sition" resolving the "dialectical opposition" between Speculum
Mentis and The Idea of History is his own reconstruction (as at
CRM, 297), but the distinction between Collingwood's assertions
and Rubinoff's reconstruction is not always clearly maintained.
17 Donagan's "Principle of Order" (LPC, 28) commits him,

in our opinion, to interpreting Collingwood's philosophy of mind
in an atomistic fashion, since Donagan appears to take an act of
thought as an event occurring in the mind: in order that it be
an object of consciousness (on Donagan's "Principle of Order")
it requires another act--a second event--of a higher order, for
no act of consciousness is self-illuminating. A literal
adherence to this principle would render
688
its successor, logical positivism (and the Essay on Metaphysics
indicates that he was at least knowledge able with Ayer's
version of the latter, which he traced to its pedigree in John
Stuart Mill--see EM, 143, 163), we know he rejected its approach
not only to historical experience but to the reductionistic
analysis of mental acts as well.
Collingwood's rejection of rigid adherence to either an
atomistic or an equally untenable context-bound view of mental
acts raises once again the problem of "encapsulation" from the
Autobiography. For it appears that not only is the term not used
in The Idea of History, but on the present account of
re-enactment there appears to be no need for a critique (or any
other argument) to "incapsulate" or confine the historical past
from the superficial or obvious present of the historian
_______________________
re-enactment impossible; it is simply epistemological atomism.
We have already argued that Donagan's "Principle of Order" as he
phrases it (a) is an illegitimate modification of what Col-
lingwood has said about the hierarchy of forms of mental acts,
and (b) that it is not itself a tenable principle concerning
mental acts, because it appears to rule out the possibility of
having a concept about concepts, propositions about propositions
(since the act which reflects must be of a higher order). We
wish to add now that (c) it cannot be regarded as a pre-
supposition of Collingwood's entire philosophy of mind, because
it ignores the sort of knowledge Collingwood calls self--
knowledge, the "knowledge of oneself as living in these activi-
ties" of thought (IN, 297), as well as his explicit assertion
that it is by that act itself that one knows that he is per-
forming it (IH, 292). Although Collingwood is saying that it is
the person who knows it, he has stated plainly that the mind of
that person is all the acts of thought together, not separately.
Such a holistic assertion is incompatible with epistemological
atomism.
689
(what might be called by other philosophers the world of the
historian's lived experience or his Lebenswelt). This would be
the attempt to mediate an immediacy, the immediacy of experience
as one's own, or subjective. It appears, in short, to be an
example of what Collingwood called the Fallacy of Misplaced
Argument, since there is no initial confusion between one's own
experience and that of another in the first place, nor is there
any initial loss of identity of the enacted meaning or the
re-enacted thought in its several contexts.
But as Collingwood stated in the Autobiography, in-
capsulation is not an occult entity; it refers to the retention
of an identity of meaning (or what he calls an "unconverted
residue") in more inclusive acts (A, 141)--in our model, the
unaltered presence of EP in HP. Therefore incapsulation is not
excluded from the theory or re-enactment, but is rather its
negative side. It is merely a short-hand way of calling
attention to the necessary bi-valence of all thought (a) as acts
of meaning embedded in the immediate experience of one's own
life and forming part of the continuity of the personal
consciousness of that life, and (b) as capable of retaining an
identity of meaning in other contexts, and especially in the
thoughts of historians. To press the juice from the metaphor,
the "capsule" is the preservation of the objective identity of
meaning (usually the plan of the historical agent in its context

690
of alternatives) from being simply absorbed into the subjective
life of the historian by an awareness that that meaning as
historically re-enacted is not merely the solution to a problem
in the personal life of the historian.
(3) Several pages later Collingwood sums up his major
argument against both realists and idealists and relates them to
Bradley's dilemma:
To disentangle ourselves from these two complimentary er-

rors, we must attack the false dilemma from which they both
spring. That dilemma rests on the disjunction that thought
is either pure immediacy, in which case it is inextricably
involved in the flow of consciousness, or pure mediation,
in which case it is utterly detached from that flow.
Actually it is both immediacy and mediation. Every act of
thought, as it actually happens, happens in a context out
of which it arises and in which it lives, like any other
experience, as an organic part of the thinker's life. Its
relations with its context are not those of an item in a
collection, but those of a special function in the total
activity of an organism. So far not only is the doctrine of
the so-called idealists correct, but even that of the
pragmatists who have developed that side of it to an
extreme. But an act of thought, in addition to actually
happening, is capable of sustaining itself and being
revived or repeated without loss of identity. So far, those
who have opposed the "idealists" are in the right, when
they maintain that what we think is not altered by
alterations of the context in which we think it . . . .
Because it is a thought and not a mere feeling or sensa-
tion, it can exist in both these contexts ((i.e. the con-
text of my own thoughts and those of another)) without
losing its identity, although without some appropriate con-
text it could never exist. (IH, 300-01).
The example that Collingwood uses for a thought that is both
mediate and immediate is Plato's argument in the Theatetus
against the view that knowledge is merely sensation (a pregnant
example, incidentally). The argument as it can be developed
either in Plato's mind or mine or

691
anyone's is what he calls thought in its mediation; the argument
in the context of discussion and theory as Plato placed it in
his own thinking is the thought in its immediacy--i.e. the
immediate thought-context in which Plato conceived it as part of
his own project of thinking out the problem (IH, 301). We assume
for Collingwood that it is possible to re-enact a proposition or
a concept as well as an argument, so that the meaning being
re-enacted extends not only to the inferential but also to the
propositional and conceptual levels of thought as well.
Here we have Collingwood's final reply to both realists
and idealists on the act-object distinction. It rests firmly on
an insight foundational not only to his philosophy of history
but his entire philosophy of mind, the insight that meaning is
the irreducible and indivisible unity of an act of thought, and
that both an act of consciousness and the object of the act are
two aspects to, or contexts for, one and the same entity.
Calling the enactment or re-enactment of a meaning an "act"
means calling attention to its personal. experiential, immediate
presence in the mental life of a person (something that someone
does "on purpose"), whereas calling attention to this same
meaning as social, communicable, and mediate means calling it an
"object" of consciousness. It is not a matter of trying to find
a way of putting together an "event" occurring in the mind, an
event called an "act of

692
consciousness," and an object of such a point-occurrence; they
are two sides to one and the same reality--a meaning whose
identity survives alteration in context and is capable of being
revived (or re-enacted) after a lapse of time. Collingwood's
rejection of both an atomistic realism and an absolute idealism,
as we have just encountered it, does not display any need for an
absolute identity of subject and object, nor does it retreat
into an endless order of acts of consciousness external to and
atomically distinct from one another--neither an absolute
experience nor a schizophrenic hall of mirrors. On the contrary
it takes its stand on the unity of meaning of singular acts of
consciousness, and its ability to survive and revive in various
contexts, so long as that context is part of the life of thought
of someone capable of doing things "on purpose."
But successfully defending re-enactment against attacks
from hypothetical objectors is not the same as demonstrating
that it is the necessary and sufficient condition for historical
thinking, and therefore we have not finished with Collingwood's
demonstration of how history as a science is possible. Suppose
we accept his argument that for history to be possible as
knowledge the past must be capable of being re-thought or
re-enacted, and that re-enactment presupposes an invariance of
meaning in several contexts without loss of experiential imme-
diacy. Is this sufficient to show that historical knowledge is
reliable? Obviously not,

693
since we do not yet know under what conditions we can say that a
re-enactment has been successfully carried out, historical error
being, we presume, possible.
If we return to Collingwood's three presuppositions for
historical thought, we appear to have three candidates for a
possible sufficiency criterion. A historical account may thus be
said to be a successful re-enactment (a) when it refers to human
acts localized in space and time; (b) when it is a coherent
narrative; and (c) when it stands in a relation to evidence such
that it can be bound to historical events by chains of
historical inference. Both (a) and (b) have been more or less
specified by Collingwood: (a) refers to the res gestae that form
the subject matter of history as distinct from the study of
merely physical events in nature; and (b) is a restatement of
the requirement that the subject matter of (a) be handled as
part of a systematic inquiry. But these alone would fail to
distinguish history from good fiction. In order that there be
that degree of certitude necessary to call historical narrative
knowledge, and not the literary expression of someone's artistic
fantasy, there must be a way to decide when (a) succeeds in
actually referring to a publicly accessible past. What
distinguishes history from romance is (c), and to this double
issue of historical evidence and historical inference we must
now turn.
694
3. Evidence, Inference, and Necessity: (a) Historical Evidence.
Why is it a double issue? From our present perspective
this should not be a difficult question to answer. It depends on
the view of history that it being presupposed by someone for
whom evidence is one thing and thought about it is another, and
their conjunction requires an argument. It is the sort of
question which assumes that historical thinking occurs when a
mass of facts, contained in documents called "testimony"
(ultimately resting on eyewitness accounts) is examined with a
view to constructing a written history of the period or event
under consideration. The work of the historian would then be one
of selecting relevant testimonial assertions, arranging these
propositions into a coherent narrative, and engaging in
reasoning only when there is a break in the continuity of the
narrative--conflicting testimony, gaps to be filled, or his-
torical peculiarities which require the intervention of the
narrator to "explain" something unusual to the reader.
This is the historiography that Collingwood disparag-
ingly calls "scissors and paste" (IH, 257), and "evidence" on
this view of history is merely the mass of records or "sources"
pertaining to the period the historian is considering (IN, 278).
In this sort of historical thinking the "data" are thus external
to the mind itself: they are contained in the set of all
relevant documents preserved in archives, libraries, and
museums. The need for inference arises only when various
authorities
695
disagree, and the only issue is whether to accept or reject a
piece of evidence as a true account of what happened. But even
though the audience to which Collingwood's remarks are addressed
is one which assumes that historical thinking and its object are
independent of one another--hence exhibiting the realistic
outlook on the relation of mental acts to their objects--it is
also an audience Collingwood hopes to convince that the role of
historical reasoning is far more intimate and interdependent
than is being presumed.
We can therefore anticipate that in the essay entitled
"Historical Evidence" (a fragment from The Principles of History
and the longest of the seven Epilegomena) his strategy will be
to propose an alternative account of historical reasoning, and
this in fact is what he does. Instead of beginning with a
definition and classification of historical evidence--i.e.
starting out as if one began with a ready-made ob-
ject--Collingwood first characterizes scientific historical
thinking and then describes its special relation to evidence,
where evidence is decidedly the derivative concept. The thrust
of his argument in this essay is to show that history is a
science in the sense of being an organized body of knowledge.
But it is organized in a way different from the exact or
empirical sciences, insofar as it approaches its data and
constructs its conclusions using a kind of inference that is
unique to history and different from that of the other
sciences, while yet being compelling

696
in its conclusions and having a necessity of its own. Evidence
in this schema of historical inference is correlative to the
historian's questioning activity and is not simply a property of
certain sets of propositions. Evidence is not the set of "facts"
which verify or falsify our historical propositions by
"corresponding" to them; it is whatever serves as a locus for
the historian to answer the questions which allow him to
re-enact past acts of thought.
Now although we have made the case in previous chapters
that Collingwood's use of Q-A logic suggests that it is more
favorably compared to a logic of discovery than to a formal
deductive system (in spite of his own misleading rhetoric that
Q-A logic is an "alternative" to F-logic), it is noteworthy that
in the Introduction to The Idea of History when he asks how
history proceeds or what its method is, his answer is not "by
the application of Q-A logic" but rather "by the interpretation
of evidence" (IH, 9-10). Q-A logic appears in the answer to the
first question he asks, i.e. "What is the nature of history?"
the answer to which is "a kind of research or inquiry"--a form
of thought whereby we ask questions and try to answer them (IN,
9). What any science does is to employ Q-A logic; what
historical science does is answer questions by the
interpretation of evidence.
697
Since we already have some idea of what he takes his-
torical thinking to be, it may be well briefly to characterize
what Collingwood has to say about evidence in this and the other
Epilegomena. We recall that the universality of an historical
event is an act expressing an intentional meaning in the context
of an array of situational alternatives confronting an
historical agent, and that this universal meaning is the
individuality sought by the historian who would re-enact the
event. Furthermore in describing the process of re-thinking the
past Collingwood leaves room for a contribution made by the
object--in this case the past thought. In discussing history as
re-enactment in the fourth Epilegomenon, for example, he writes
that "historical knowledge is that special case of memory where
the object of present thought is past thought, the gap between
present and past being bridged not only by the power of present
thought to think of the past, but also by the power of past
thought to reawaken itself in the present" (IH, 294). Nor is
this an isolated remark; it is a recurrent theme in The Idea of
History. In discussing the subject matter of history in the
fifth Epilegomenon we have already quoted the striking passage
where he speaks of the individuality of a particular historical
act being something that "oversteps the limits of merely local
and temporal existence and possesses a significance valid for
all men at all times," and adds that this act transcends its
immediacy because "that

698
individuality is the vehicle of a thought which, because it was
actually theirs, is potentially everyone's" (IH, 303). And in
the second Epilegomenon Collingwood expands further on this
"potential-actual" distinction as it pertains to evidence:
Everything is evidence which the historian can use as evi-

dence. But what can he so use? It must be something here
and now perceptible to him: this written page, this spoken
utterance, this building, this fingerprint. . . . The whole
perceptible world, then, is potentially and in principle
evidence to the historian. It becomes actual evidence in so
far as he can use it ((as evidence on some question)) . . .
. Evidence is evidence only when someone contemplates it
historically. Otherwise it is merely perceived fact,
historically dumb. (IH, 247; cf. 203).
It should be clear at this point why Collingwood begins
a discussion on evidence with a discussion of the nature of
historical thinking: it is the thinking that determines what is
evidence, and not the other way around. It should also be clear
that while historical thinking makes the difference between a
perceived fact being merely that or historical evidence, there
is a contribution being made by the "perceived fact" inasmuch as
it has a "power to re-awaken itself" in the present. Once again
the difference that historical thinking makes to its object is
not creational but constitutional; it does not create the fact
but determines it as evidence. "His business is not to invent
anything, it is to discover something" (IH, 251).
In the third Epilegomenon this useful distinction be-
tween actual and potential evidence is redefined in terms of
scientific history.
699
If history means scissors-and-paste history . . . a source

is a text containing a statement or statements about the
subject . . . . If history means scientific history, for
"source" we must read "evidence." And when we try to define
evidence in the same spirit in which we defined "sources,"
we find it very difficult . . . . In scientific history
anything is evidence which is used as evidence . . . . Let
us put this by saying that in scissors-and-paste history,
if we allow ourselves to describe testimony--loosely, I
admit--by the name of evidence, there is potential evidence
and there is actual evidence. The potential evidence about
a subject is all the extant statements about it. The actual
evidence is that part of these statements which we decide
to accept. But in scientific history the idea of potential
evidence disappears; . . . everything in the world is
potential evidence for any subject whatever. (IH, 278-80).
The obvious question is, how does the historian manage
to restrict this infinite domain? Once again Collingwood seems
to be overstating his case. It is clear that his intention is to
display the active role of the thinking historian in
constructing his narrative. To this extent there is an exact
parallel between Collingwood's historian, whose task is to
re-enact past acts of thought in his own mind, and Collingwood's
artist, for whom the act of expressive imagination takes place
"in his head." But just as the latter tends to render art
inaccessible to public examination (how do you examine the
inside of someone else's head?), so also the scientific
historian's "evidence" appears to be not only thought-dependent
but inseparable from the infinite array of everyday artifacts.

700
Collingwood's reply is that this is only a problem for
the scissors-and-paste historian, who relies on there being a
manageably small amount of testimony (IH, 278). For the
scientific historian evidence is strictly correlative to the
questioning activity, and it is his arrangement of questions
that guides what the historian chooses to consider as evidence.
In an extended example (one that has often been excerpted in
collections of essays on the philosophy of history) Collingwood
uses the fictional example of an inspector investigating a
murder ("Who Killed John Doe?") as a paradigm for what a
scientific historian does. The example is too familiar to
require repeating here, but it is important to note the lessons
that Collingwood derives from it. (1) Each step of the
investigation is dependent on the asking of a question, and each
question is asked in the right order (IH, 273). One assumes that
"the right order" here, based on the example, is one in which
the answer to one question is the presumed basis for the next
question, and that the whole series is governed by the intention
to solve a single problem. (2) The questions are put by the
investigator to himself--a process which calls attention to the
autonomy of the historian's inquiry, "where by autonomy I mean
the condition of being one's own authority, making statements or
taking action on one's own initiative and not because these
statements or actions are authorized or prescribed by anyone
else" (IH, 274-75).

701
What distinguishes the scientific historian from his
predecessors is precisely this ability to act on his own without
waiting for statements to be made to him. In this the historian
is in the position of a Baconian cross-examiner, who puts
evidence to the test (IH, 269), and it is this activity which
determines what actual evidence is.
Question and evidence, in ((scientific)) history, are cor-

relative. Anything is evidence which enables you to answer
your question--the question you are asking now. A sensible
question (the only kind of question that a scientifically
competent man will ask) is a question which you think you
have or are going to have the evidence for answering. If
you think you have it here and now, the question is an
actual question . . . . If you think you are going to have
it the question is a deferred question . . . . (IH, 281).
One might say for Collingwood that in this context actual evi-
dence is determined by actual questions, potential evidence by
deferred questions.
Once again, this account leaves serious questions un-
answered. What counts for a relevant question in this series,
and what dictates the choice of observable evidence? Colling-
wood's response tends to be circular: the historian chooses
those questions which his evidence permits him to answer (IH,
281). But since the evidence is chosen on the basis of the
questions, the process seems to be one of chasing one's own
tail. What makes the circularity all the more glaring is that
two pages previous to this discussion of the correlativity of
question and evidence Collingwood had charged that the
scissors-and-paste man, like the nineteenth century

702
landscape painter, protects himself by choosing subjects that he
is able to "get away with" (IH, 279). In either
case--Collingwood's or the sub-scientific scissors-and-paste
man--one chooses one's questions to suit the evidence, and the
evidence to suit the question.
In his zeal to put forward his thesis that scientific
thinking is more of a question and answer process than it is one
of deriving logical inferences from ready-made statements, has
Collingwood failed to make important distinctions--e.g. between
the way a piece of evidence first presents itself and the way it
is subsequently treated in an investigation? Has Collingwood
once again lost sight of the real contribution of the object in
his efforts to counteract the presumed realistic bias of his
readers? Or have we overlooked something?
Now it is in a section entitled "Statement and Evidence"
that Collingwood provides us with a hint of a missing element in
this entire discussion. The way out of the circularity of the
correlativity of evidence and question has to do with the sort
of question that the scientific historian asks as opposed to
that of the scissors-and-paste man.
Confronted with a ready-made statement about the subject he

is studying, the scientific historian never asks himself:
"Is this statement true or false?" . . . The question he
asks himself is: "What does this statement mean?" And this
is not equivalent to the question "What did the person who
made it mean by it?" although that
703
is doubtless a question that the historian must ask, and

must be able to answer. It is equivalent rather to the
question "What light is thrown on the subject in which I am
interested by the fact that this person made this
statement, meaning by it what he did mean?" This might be
expressed by saying that the scientific historian does not
treat statements as statements but as evidence: not as true
or false accounts of the facts of which they profess to be
accounts, but as other facts which, if he knows the right
questions to ask about them, may throw light on those
facts. (IH, 275).
When Collingwood writes that the historian treats
statements as other facts the reader is obliged to recall that
the irreducible datum of history is the universal meaning of an
intended event--the "inside" of an action. If the historian is
limited by the truthfulness of the statements of his sources,
then his efforts to construct a coherent narrative by stringing
together true propositions about the past are seriously
threatened by testimony which contains conflicting accounts of
the same event (IH, 257). What Collingwood is arguing is that if
history is not bound by this presupposition it is not in danger
of such a loss of coherence. By treating testimony as factual in
the sense of seeking interpretation of it just as one does of
all evidence, i.e. by asking what it historically means, the
historian may preserve the coherence of his inquiry while at the
same time preserving evidence as evidence rather than simply
discarding it as useless. In Collingwood's view this is
precisely what makes history scientific: instead of being the
passive spectator--a condition in which thought is barely
occurring, as Collingwood often says --the historian is
providing a measure of control over his subject matter.

704
The historian's laboratory is his own a priori imagination, and
he is only limited by his ability to exercise that imagination
in the form of revealing questions aimed at eliciting meaning
from present evidence.
(b) Historical Inference.--If historical thinking occurs
by means of an ordered sequence of questions, what is the role
of inferential reasoning? If we were correct in our
reconstruction of Collingwood's functions of higher-order con-
sciousness in the last chapter, then can we find confirmation
here that by an application of the Law of Primitive Survivals
the meaning-seeking function of consciousness is extended to the
level of inferential thought in the form in which one pro-
positional meaning is related to another as ground to consequent?
At first glance this does not seem to be the function
that Collingwood assigns to it. What historical inference does
is to form a bridge between present evidence and past
events--the former being question-correlative and perceptible,
the latter being answer-correlative and imperceptible, but
re-thinkable (IH, 251-52). But at the same time that he is
preoccupied with saying what historical inference is not, he
lets slip the opportunity to say more explicitly what it is. We
shall first let Collingwood tell us what it is not, and then try
to reconstruct what it is.

705
History has this in common with every other science: that

the historian is not allowed to claim any single piece of
knowledge, except where he can justify his claim by exhib-
iting . . . the grounds upon which it is based. This is
what was meant . . . by describing history as
inferential. . . . Different kinds of science are organized
in different ways; and it should follow . . . that
different kinds of science are characterized by different
kinds of inference. The way in which knowledge is related
to the grounds upon which it is based is in fact not one
and the same for all kinds of knowledge. That . . . a
person who has studied the nature of inference as such--let
us call him a logician--can correctly judge the validity of
an inference purely by attending to its form, although he
has no special knowledge of its subject-matter, is a
doctrine of Aristotle; but it is a delusion, although it is
still believed by many very able persons who have been
trained exclusively in the Aristotelean logic and the
logics that depend upon it for their chief doctrines. (IH,
252-53).
What is worth noting in this passage is the positive
assertion disguised as a negative one, i.e. Collingwood's in-
sistence that it is a delusion that knowledge of the subject-
matter has no bearing on the validity of scientific inference
and that it depends solely on its form. Positively stated this
means that the subject-matter does have a bearing on determining
the validity of inference. It is also worth noting that
throughout this and the subsequent discussion of inference,
everything is phrased in epistemological terms; inference is a
kind of thinking rather than a formal relationship between
propositions, assertions, or the like. Finally it must be
recalled that when Collingwood refers to the logic of Aristotle
"and the logics that depend upon it for their chief doctrines"
he has in mind all F-logics of a propositional pedigree, as we
noted in the previous chapter.

706
Now in this inferential process of justifying knowledge
by exhibiting the grounds upon which it is based, Collingwood
contrasts historical inference to deductive and inductive
inference in a manner very reminiscent of the Essay on
Philosophical Method, but this time with emphasis on the
"compulsion" with which the conclusion follows from the pre-
misses. Proof, he writes,
might be either compulsive, as in exact science, where the

nature of inference is such that nobody can affirm the
premises without being obliged to affirm the conclusion
also, or permissive, as in "inductive" science, where all a
proof can do is to justify the thinker in affirming its
conclusion, granted that he wishes to do so. An inductive
argument with a negative conclusion is compulsive, that is
to say it absolutely forbids the thinker from affirming
what he wishes to affirm; with a positive conclusion it is
never more than permissive. If history means
scissors-and-paste history, the only kind of proof known to
the historian is of this latter kind . . . . If criticism
leads him to a negative conclusion, viz. that the statement
of its author is untrustworthy, this forbids him to accept
it, just as a negative result in an "inductive"
argument . . . forbids the inductive scientist to affirm
the view he hoped to affirm. If criticism leads him to a
positive conclusion, the most it gives him is a nihil
obstat. (IH, 261).
One has to wince at calling a conclusion to a proof a view
someone "hoped to affirm," for why could a proof not be one that
concludes to a view that the author either has no hope for
whatever, or even is one he hopes to deny? Yet with some
correction (e.g. replacing dispositional terms like "hoping"
with logically relational or at least more neutral terms) the
line of thought is clear enough, and poses no particular problem
to the reader. Since he has just been told what constitutes sub-
scientific
707
historical inference, he is led to expect that he will next be
enlightened about actual scientific inference. What he gets
instead are more dispositional terms--the "compulsion" that is
contrasted with "permissive" inductive arguments.
One hears it said that history is "not an exact science."

The meaning of this I take to be that no historical argu-
ment ever proves its conclusion with that compulsive force
which is characteristic of exact science. Historical in-
ference, the saying seems to mean, is never compulsive, it
is at best permissive; or . . . it never leads to cer-
tainty, only to probability. Many historians . . . must be
able to recollect their excitement on first discovering
that it was wholly untrue, and that they were actually
holding in their hands an historical argument which left
nothing to caprice, and admitted of no alternative conclu-
sion, but proved its point as conclusively as a demonstra-
tion in mathematics. (IH, 262).
An argument that admits of no alternative conclusion and proves
its point conclusively is one whose compulsive force must be
that of necessity. Once again the reader anticipates that he
will be enlightened about the special nature of the necessity
involved in historical inference. What follows next is utterly
astonishing.
If any reader wishes to rise here on a point of order and

protest that a philosophical question, which ought there-
fore to be settled by reasoning, is being illegitimately
disposed of by reference to the authority of historians,
and quote against me the good old story about the man who
said "I'm not arguing, I'm telling you," I can only admit
that the cap fits. I am not arguing; I am telling him. Is
this wrong of me? The question I want settled is whether an
inference of the kind used in scientific history, as
distinct from scissors-and-paste history, yields compulsion
or only permission to embrace its conclusion (T)he only way
of knowing whether a given type
708
of argument is cogent or not is to learn how to argue that

way, and find out. Meanwhile, the second best thing is to
take the word of people who have done so for themselves.
(IH, 263).
If this is the manner in which the historian exhibits the
grounds upon which his knowledge is based, then surely one may
be excused from affirming the claim that history is a science.
For how can it provide anything but an authoritative criterion
for successful historical inference? On such grounds as these
one could excuse anything that even pretends to be an inference;
if it is true that whatever I do is historically valid
inference, then any inference is valid so long as it is what I
do. One could hardly ask for a more blatant example of what Knox
called Collingwood's dogmatism.
What has happened to the role of philosophy
which, in Speculum Mentis, judges whether a particular form of
knowledge explicitly lives up to its implicit promise? Do we not
rather appear to be in the full grip of a form of thought that
is attempting to justify itself from within--what he had once
himself called historical dogmatism? Has Collingwood abandoned
the role of philosophy as outlined in the Essay on Philosophical
Method of being a self-justifying form of thought which avoids
dogmatism or assuming certain principles to be true without
justification? Are we not now in the more cloistered atmosphere
of The New Leviathan, where the "Fallacy of Misplaced Argument"
forbids us to argue about objects given directly to
consciousness?
709
If we try again to look past what Collingwood said to
what he meant, where that meaning is informed by what is un-
deniably his overall intention in these essays, a mitigating
argument can be put forward to soften the apparent intransigence
of this remark. We have noted in passing the epistemological
orientation of the entire discussion of inference-just as in the
Essay on Philosophical Method. When Collingwood calls history a
science and then defines science as an organized body of
knowledge, and finally states that knowledge is organized when
it can exhibit the grounds on which it is based, one cannot
expect that a formal account of historical inference is what
will follow. If one were to succeed in formalizing an historical
inference Collingwood would surely say that not only was it
historically unenlightening to focus on the function of logical
connectives and their formal meanings (if-then, and, or, all,
some, etc.) rather than the content of the expressions being
linked together, but also that what was peculiarly historical
about it would be completely left out in the interest of showing
how it was isomorphic with other forms of inference. The sort of
coherence aimed at in historical thinking is not formal validity
but narrative continuity. This continuity is demonstrated when
the meaning of an act is related to the meaning of another act
as historical ground and consequent.

710
Furthermore Collingwood often repeats the point that in
order to be history it is a necessary condition that particular
events be related in a coherent fashion, and not as instances of
a class of similar events. If this requirement is added to the
previous observation it follows that an historical inference is
the establishment of coherence in the form of narrative
continuity between the meaning of one particular historical act
and another such that the one is the historical ground and the
other its consequence. What Collingwood's dogmatic assertion
amounts to is a statement of an historian's autonomy in this
context, so that in order to find out if such a narrative has a
coherent continuity one has to repeat the process of re-thinking
it for oneself, which necessitates re-thinking the meaning of
this act as grounds for the meaning of another particular act,
not another like it in its class. Rehearsing another argument
like it will not do either, for the argument is
meaning-dependent. If a would-be historian is unable or
unwilling to re-enact the argument for himself, he can only rely
on the testimony of another (which means that he ceases to be a
scientific historian at this point), in which case Collingwood,
the Roman-British historian, announces that his testimony is as
good as any other, and he finds that such narrative coherence
does exist, and is compelling.

711
There are many difficulties with this reconstruction, as
well as with Collingwood's entire handling of the question of
historical inference. For starters there is the need for a
clearer distinction between the kind of inference that the
historian uses to link evidence to his proposed or received
narrative events, and the inference that occurs when the his-
torian links events to other events within his narrative. All of
this borders on a discussion of the issue of historical ex-
planation, around which has grown an extensive secondary liter-
ature for the last thirty years (cf. CEPC, 331-48). To deal with
this in even a cursory way would require a separate treatise.
Our concern here is first with evaluating the philosophy of
history expressed in The Idea of History as legitimately
interpreted in the Autobiography; and secondly understanding
this philosophy of history in the light of developments of his
ideas in the other later writings, especially concerning
epistemological logic and the philosophy of mind.
But it is precisely here that our major difficulty with
Collingwood's handling of the issue of historical inference
arises. For just as the Essay on Philosophical Method leaves us
wondering when a philosophical inference is unsuccessful, so The
Idea of History does not provide us with grounds for determining
when historical inference fails to achieve coherent narrative
continuity. For if, lacking Collingwood's assurance on any
particular historical inference, we decided to be scientific
historians ourselves, what assurance do we have that the

712
"compulsion" to a given conclusion is not an utterly subjective
conviction that is merely shored up by the evidence we choose to
consider as governed by the questions we ask, which are in turn
based on our own presuppositions? And if the shade of
Collingwood were to retort, "what assurance do you require?" the
reply is the quite Collingwoodean requirement that it be first
and foremost a self-assurance. How is that achieved?
(c) Historical Necessity.---The issue is one of valid
vs. invalid historical inference, and such a de Jure issue is
not settled by a de facto exemplification or description (we are
not asking Collingwood, we are telling him). The answer that
seems appropriate to the argument as Collingwood has been
developing it is that historical inference succeeds when it is
grounded in evidence. But we have just seen that there is an
apparent circularity here inasmuch as evidence and questions
are, in Collingwood's view, strictly correlative. It simply
postpones the inevitable need to state the missing criteria,
since not just any sequence of completed questions and answers
nor any haphazard choice of evidence will satisfy an intelligent
historian.
Although the challenge may seem unfair, it would be
enlightening in this regard to conduct a thought-experiment in
which Collingwood's canons of scientific history would be
applied to an outrageously bad piece

713
of historiography, such as Erich von Daniken's Chariots of the
Gods? l8 Von Daniken exasperates the learned and delights the
ignorant by declaring that he has sufficient evidence to
demonstrate that ancient astronauts visited the earth in
prehistoric times, and that the legends of gods descending from
the skies in fiery chariots are merely a dim memory of these
events. Von Daniken plies himself and his readers with hundreds
of questions, constructs historical arguments linking his
evidence to these astronautical events, and uses his a priori
historical imagination to construct a narrative of events which
refer to deeds of men (?) done in the past--events he is trying
to re-enact in his own mind. He engages in this historical
reconstruction for the purpose of self-knowledge, and especially
in the light of pre-suppositions drawn from 20th century science
and its expansion by space exploration. He even appears to
cross-examine evidence by putting questions to it, and chooses
evidence that he thinks will allow him to answer the questions
to it, and chooses evidence that he thinks will allow him to
answer the questions that he has. And some--presumably himself
included--would say that his historical arguments are compelling
and admit of no other rational conclusion.
What is missing here? It appears that Collingwood's
criteria of scientific history is superficially well suited to
justify the work of
____________________
18 Translated by Michael Heron (New York, 1969).
714
both respectably orthodox practitioners and disreputably
heretical dilettantes. In order to demonstrate that "history"
like that of Erich van Daniken is spurious and fantastic it
would be necessary to take into account the reaction of the
community of historians and their own work interpreting the same
evidence pertaining to the same "events." Collingwood, of
course, knew this quite well, as indicated by his requirement
that second-order history, the history of historiography, is
itself an integral part of the historian's task. He would also
have found Von Daniken's books beneath contempt and quite
acceptably dismissed by equally amateurish rebuttals like that
of the Christian fundamentalist, Clifford Wilson.l9 Nevertheless
the serious question remains: on Collingwood's grounds, how is
bad history--history that fails to be scientific--eliminated?
Evidently by critical historiography. But then scien-
tific history presupposes critical history rather than replacing
it, just as critical history builds on scissors-and-paste while
yet modifying and correcting it. Collingwood's efforts to
proclaim a historian's version of the Copernican revolution may
have led him to express himself in a way that misleads the
reader into overlooking Collingwood's discovery that history is
a philosophical concept and therefore defined as a scale of
forms in which two successive stages in the scale are related by
opposition and
________________________
19 Crash Go the Chariots (New York, 1972).
distinction, and have differences of both degree and kind.
Scientific history may be opposed to critical history in the
sense that it rejects any attempt to make historical objects
independent of the historian's questioning thought (therefore
ignoring the essential ideality of the past); but it is also
distinct from critical history insofar as it is self-consciously
aware that when it replaces the concept of "evidence" for the
critical historian's "sources" and the scissors-and-paste
historian's "testimony" the concept of evidence includes the
lesser forms in its meaning. It is therefore not only a
different kind of history but is a more perfect embodiment of
the generic concept than is critical history, just as critical
history is a more perfect form than scissors-and-paste.20
But it is not obvious that the same can be said for the
necessity that Collingwood claims for historical inference.
Since Hume it has been natural to suppose that the only sort of
necessity admissible in inference is strict logical neces-
sity--what Collingwood is calling deductive
_____________________
20 What is the minimum specification of the scale of
forms of the concept of history? It cannot be direct observa-
tion, because to be history at all it must deal with res gestae
incapable of present perception. Can it be memory? While
Collingwood rejects memory as grounds for history because his-
tory is organized and memory is not (IH, 252), this is from the
point of view of scientific history--the highest point on the
scale of forms of history. The generic concept must be minimally
specified by an act that re-enacts, and by a thought about
thought. Minimally what else can this be but an act of memory?
That an act of memory is not organized and not inferential means
only that it has the status of a confused potential for
interpretative acts of meaning.
716
compulsion, with some allowance made for inductive inference
(yielding the "virtual compulsion" due to preponderant
probability). But in the strict sense, when the meaning of an
event is related to the meaning of another event as ground to
consequent, how can this be a necessary relation? In the realm
of human affairs it is always possible that the outcome could be
otherwise than the way it was. And if the historian's work of
exhibiting the evidence as grounds for the coherence of his
narrative of these events is also an inferential process, in
what sense is the conclusion--the particular narrative of
events--a necessary conclusion from the evidence? It is clearly
this latter question that occupies the foreground of
Collingwood's concern in The Idea of History, and our question
at this point concerns the nature of a "compulsion" which is
neither deductively necessary nor inductively permissive, but is
yet just as binding in its conclusion as is a mathematical
proof. It is also clearly not Collingwood's intention to reduce
this compulsion to a subjective or psychological state. He means
to assert that historical inference compels like all scientific
inference compels--i.e. with necessity.
When historical inference refers to the ground-conse-
quent relation between narrated events there is a sense of ne-
cessity in which it is understandable to say that past events
cannot be other than the way they are. Necessity has a human-
based usage wherein some event or state of

717
affairs is said to be necessary if it cannot be made by me or
anyone else to be other than the way it is. As we shall see in
the next chapter, this is the way Collingwood analyzes the term
"cause" in the Essay on Metaphysics. In this sense the past is
beyond human causality: it cannot now be made by me or by anyone
else to be other than the way it is. But this is not the
interesting sense of the term so far as the argument of The Idea
of History is concerned. We are interested in the relation of
evidence to reconstructed events of the past, and reconstructing
or re-enacting that past is within the non-necessary realm of
causable or alterable events.
But there is one way that I can see to salvage Colling-
wood's treatment of historical inference. When the question-
evidence correlation came up for discussion in the previous
section we noted that the kind of question that the historian
asks aims not at "the truth" but at the meaning of the evidence.
It is this hermeneutical preoccupation that provides us with a
way of understanding a non-deductive but necessary inferential
compulsion.
The missing link in this entire discussion is a proper
elucidation of the term "interpretation"--an axial process since
it is by means of the interpretation of evidence that the
historian draws his inference about the meaning of past events.
It follows that the correctness of the inference is dependent on
the correctness of the interpreting process.

718
Now in the interpretative process there is one sense of
"necessity" that is both legitimate and appropriate for Col-
lingwood's use in this connection. Collingwood seems to have in
mind a view of inferential necessity that is parallel to his
notion of conceptual truth. Just as truth is not only a property
of judgments or propositions but extends to concepts as well in
the sense that a concept may truly or falsely express an
intended meaning, so also necessity extends to concepts,
statements, and inferences in the sense of "fitness" (like the
elements in a Gestalt are seen as necessary to the
interpretation of the pattern). In the hermeneutical sense
something is said to be necessary when it cannot be other than
the way it is without losing its identity of meaning. In the
historical sense the hermeneutical demand for meaning proposes
an a priori necessity in the sort of fitness that makes a
conclusion necessary only when it makes both narrative and
evidential sense, so that the meaning of both the evidence and
the event are fused together so tightly that nothing essential
is left out of account. The necessity is one of referential
narrative meaning, i.e. by means of this argument (a set of well
interpreted pieces of evidence) such a conclusion (the
re-enacted event) makes historical sense.
This is what Collingwood seems to mean when he says that
the necessity of the historian's picture of the past (i.e. his

719
narrative account) "is at every point the necessity of the a
priori imagination" (IH, 245-46) whose functions are selection,
construction, and criticism (IH, 236).
Conclusion.
In this chapter we have had the opportunity to examine
how Collingwood carried out the rapprochement between philosophy
and history, and although we were left with several unanswered
questions and some serious problems in understanding what
Collingwood meant us to understand by historical inference, it
is clear enough that the autobiographical interpretation in its
main lines is carried forward into The Idea of History, even
though it is a posthumous publication and is the product of the
editorial re-arrangement in its present form. We saw that in
Part I of The Idea of History the scale of forms of mental
functions is reflected in the development of the concept of
history as it is expressed in historiography. We also saw that
the Epilegomena provide some striking illustrations not only of
Collingwood's mature philosophy of mind, but also of the kinds
of judgments and arguments that were described, but not
exemplified, in the Essay on Philosophical Method. And finally
we saw how Collingwood attempted to free himself not only from
the errors of realism but from those of idealism as well. And in
the process we had a glimpse of Collingwood's use of a basically
interpretative approach to his subject matter, an approach that
allowed for the self-conscious control by

720
the historian of his narrative by means of inference based on
the interpretation of evidence. The reconciliation of history
and philosophy does not merely take the form of a reduction of
both to forms of self-consciousness--although that is, in
Collingwood's view, generically the case. It takes the form of a
demonstration that self-knowledge (the goal of philosophy) is
possible only on condition that meanings can be re-enacted in
such a way that they retain both their objectivity (their
identity as meanings) and their subjectivity (as experience).
What we have not shown is what the limitations of this
reconciliational process are. Does history simply absorb
philosophy without remainder in this re-enactment? What is the
independent role of philosophy?

CHAPTER X
METAPHYSICS AND RAPPROCHEMENT
1. Introduction.
With this chapter we come to the end of our journey
through the published writings of R. G. Collingwood, and to the
central paradox of the Essay on Metaphysics, which has been the
cause of so much critical discussion among the students of
Collingwood's philosophy.
In Chapter I we noted that it was this work (and those
portions of the Autobiography that support it) which so
scandalized Knox, who argued that at the end of his life
Collingwood "turned traitor" to his philosophical profession by
absorbing philosophy into history. According to Knox, until 1936
Collingwood still held that metaphysics as a separate study of
the One, the True, and the Good, a study distinct from history,
was still possible; but after 1938 he did not, as evidenced by
the historicism of the Essay on Metaphysics and the
Autobiography. Knox concluded that sometime between 1936 and
1938 Collingwood's philosophical standpoint radically changed,
and while the major evidence for this change is taken by Knox
from unpublished manuscripts, he makes an impressive case for it
from the published writings themselves (IH, x-xi).
721
722
The fact that there is no mention of a major reversal of opinion
in the Autobiography is one of the main reasons Knox rejects it
as a legitimate interpretation of the development of
Collingwood's philosophy.
In Chapter I we also noted that this is the major dif-
ficulty to which most of Collingwood's subsequent interpreters
addressed themselves. Donagan grounds his interpretation of
Collingwood's later philosophy on the disparity that he finds
between the earlier idealistic philosophy, in which all ab-
straction is rejected as falsification, and the later analytic
philosophy of mind which "repudiates" this position by re-
cognizing that all thinking is conceptual and that all concepts
are abstract (LPC, 14, 18). While we have had occasion to
question Donagan's analysis of the abstractness of thought in
Collingwood's later philosophy, it is difficult to escape his
conclusion that the Essay on Metaphysics represents an
abandonment of views expressed earlier, in which philosophy is
said to consist of universal judgments: for if historical
statements concern particular matters of fact, how can meta-
physics as an historical science preserve its universal aspect
as philosophy? Donagan argues that Collingwood came to think
that it was impossible to have a non-abstract idea of pure
being, and therefore attempted to save metaphysics by giving it
an orientation away from ontology and towards the history of
science. But what he proposed was a contradictory and
indefensible science of
723
truth-neutral absolute presuppositions, which is incompatible
with his later philosophy of mind (CEPC, 15).
It is this reformed metaphysics which Walsh, Roten-
streich and others have also found to be neither conceptually
justifiable nor historically accurate, representing what Walsh
has called "metaphysical neutralism"--the refusal to take a
metaphysical position (CEPC, 149, 197). The attempt to extricate
Collingwood from this blind alley has led at least two of
Collingwood's gifted defenders--Mink and Rubinoff--to engage in
feats of interpretative excess, which we spent part of Chapter I
examining. While we have learned more from these studies of
Collingwood than we can accurately and adequately acknowledge,
we decided that there was still room for a study of
Collingwood's public philosophy based on the guidelines provided
by the autobiographical interpretation.
In Chapter II we laid the groundwork for this self--
interpretative study by locating the four major themes around
which Collingwood organized his Autobiography: his opposition to
realism, Q-A logic, the philosophy of history, and the
philosophy of rapprochement. In Parts II and III we have
verified that the autobiographical interpretation is a remark-
ably concise overview of themes that recur constantly throughout
Collingwood's early and later writings, and often enlightens the
strategy of argumentation in
724
these works. We have had occasion to find fault with what
Collingwood maintained at one place or another; we have
witnessed apparent contradictions not only between one
publication and another but within the same treatise; and we
have found themes latent in these writings that help to overcome
puzzling and frustrating obscurities in some difficult
passages--themes that were not explicitly discussed in the
Autobiography. But we have found no evidence thus far of a major
reversal on any of the central issues of the autobiographical
interpretation.
But with the Essay on Metaphysics we face our greatest
challenge. Approaching this work from the background of our
survey of the other published writings we find both striking
parallels and equally striking reversals. We observed in Chapter
III that even before Collingwood's opposition to realism
hardened during World War I, his outlook on many substantial
philosophical issues was already formulated. Thus it is
noteworthy that seven of the ten chapters of Religion and
Philosophy are assembled under the sub-headings, "Religion and
Metaphysics" (Part II) and "From Metaphysics to Theology" (Part
III), and contain discussions about proofs for the existence of
God, materialism and idealism, personality, identity and
difference, the Absolute, the problem of evil, and God's
self-expression in man (RP, viii-xii). In this earliest work
Collingwood writes as if there is a legitimate function of
metaphysical analysis not merely in locating and propounding

absolute presuppositions,
725
but in deciding which of them are acceptable and which are not.
For example, when he examines the three presuppositions (he
calls them "hypotheses") that "(e)ither the world is entirely
material, or it is entirely spiritual, or it is a compound of
the two" (RP, 73), he reaches some conclusion concerning the
issue: he rejects both materialism and a crude idealism in favor
of a modified form of idealism which accepts "mind as the one
reality" in the sense that "the world is the place of freedom
and consciousness, not of blind determinism" (RP, 94-95).
Surely "the world is entirely material" is as much a
metaphysical presupposition as "all events have causes" (EM,
51), and in rejecting this "hypothesis" Collingwood is not
merely stating that it is no longer an absolute presupposition
of science (although he does indicate that this is the case--cf.
RP, 82), but is deciding the issue in favor of idealism--which
is hardly remaining neutral on its truth or falsity. Nor is this
an isolated case. When he analyzes the immanence-transcendence
dualism with respect to God's identity in the universe, he does
not leave the matter in unresolved conflict, but reconciles them
through the "concrete identity of activity"--i.e. "the identity
of two minds which think the same thing" in the sense of sharing
the same thoughts and volitions: "so far, that is, as they know
any truth or will any good"

726
(RP, 117-18). Thus in stating that God cannot be simply immanent
in the world or transcendent to it, but rather present in it in
the same sense that personality is self-identical and yet
inter-personal in action, he is not merely stating what thinking
Christians believe, he is affirming the truth of a theological
presupposition and rejecting two heretical versions.
But it is also in this early work that we found the
following passage in which there is a striking anticipation of
the rejection of metaphysical ontology in the Essay on
Metaphysics:
It is often maintained that ultimate truths are incapable

of proof, and that the existence of God is such an ultimate
truth. But I venture to suggest that the impossibility of
proof attaches not to ultimate truths as such, but only to
the truths of "metaphysics" in the depreciatory sense of
the word; to truths, that is, which have no definite
meaning. We cannot prove that Reality exists, not because
the question is too ultimate (that is, because too much
depends on it), but because it is too empty. Tell us what
you mean by Reality, and we can offer an alternative
meaning and try to discover which is the right one. No one
can prove that God exists, if no definite significance is
attached to the words; not because . . . the reality of God
transcends human knowledge, but because the idea of God
which we claim to have is as yet entirely indeterminate.
(RP, 64).
This is essentially the same argument that we
find Collingwood making in the Essay on Metaphysics to reject an
ontology of "pure being," and the existence of "truths" which
are incapable of proof is remarkably close to the later concept
of absolute presuppositions. It is clear from this early work
that at this time Collingwood accepted a

727
legitimate function to metaphysics and rejected some
metaphysical arguments as vacuous and therefore illegitimate.
Eight years later Collingwood appears to have less hope
for the legitimacy of the enterprise of metaphysical analysis.
In Speculum Mentis metaphysics is described as the scientific
form of experience abstractly (and therefore falsely) extended
into philosophy: it is that form of dogmatic philosophy in which
scientific thinking justifies its objective validity "by showing
that the real world is constructed in such a way that, in
thinking it scientifically, we are thinking of it as it really
is" (SM, 272). But this involves showing that the laws of
thought (logic) are really the laws of being (metaphysics), and
this cannot be done.
Metaphysics is impossible; for its task is to vindicate the

objective validity of the ways in which we think, and if
there are any flaws in our methods of thought, these will
affect our metaphysical theory of reality and introduce into
it the very mistakes which by its help we had hoped to
eradicate. Hence the theory of being as distinct from
thinking (metaphysics) will only be the theory of thinking
as distinct from being (logic) expressed in a different
terminology but subject to the same fatal weakness, namely
that just as logic can never analyze real thinking--the
thinking that, going on in the logician's mind, always lies
behind his analysis--so metaphysics can never analyze real
being, being as it is in itself untainted by thought. (SM,
274).
Thus as we have seen in these two early works, Collingwood's
antipathy to metaphysics defined as a science of "pure being"
has its origin in his early philosophy, and the Essay on
Metaphysics does not represent a departure from this tendency.

728
Where they part company appears to be over the recon-
ciling role of philosophical thought. Speculum Mentis leaves
open the possibility that as the exercise of "absolute knowlege"
philosophy can achieve, by unification of subject and object,
what cannot be achieved by scientific metaphysics. In the Essay
on_Metaphysics the role of philosophy seems to be restricted to
the discovery and analysis of absolute presuppositions by
historical means, and the gap between the metaphysical subject
and his historical object is measureable by the attitude of
natural piety or even numinous terror that surrounds absolute
presuppositions. From Part II we are therefore left with the
problem of trying to understand how history can establish what
science cannot, and how philosophy can be content with leaving
absolute presuppositions in an unreconcilably truth-neutral
state.
Part III has provided us with the means for resolving
this issue at the same time it deepens our understanding of the
problem. On the one hand our view of history from Collingwood's
analysis of it, as we reviewed that analysis in The Idea of
History, helps us to understand why the obstacle of infinite
factuality, which wrecked the historical enterprise in Speculum
Mentis, is no longer a difficulty for the scientific historian.
On the other hand, the scale of forms of conscious acts that
provides the foundation for historical thought seems to leave no
place for the absoluteness of presuppositions. Not only is
there no upper limit

729
to the levels of consciousness to correspond to the "absolute
knowledge" of Speculum Mentis, but the very facticity of
evidence is generated by the questioning activity of the
historian--so that absolute presuppositions appear to be neither
absolute nor, from an historical perspective, truly pre-supposed.
To add to this distressing situation we are confronted
with a statement from the Essay on Philosophical Method, which
has the highest blessings from Collingwood in his Autobiography,
about the nature of metaphysics--a statement which seems to run
directly counter to the rejection of metaphysical ontology both
in the earlier works and in the Essay on Metaphysics. He writes:
metaphysics, even if it is regarded as only one among the

philosophical sciences, is not unique in its objective re-
ference or in its logical structure; all philosophical
thought is of the same kind, and every philosophical science
partakes of the nature of metaphysics, which is not a
separate philosophical science but a special study of the
existential aspect of that same subject-matter whose aspect
as truth is studied by logic, and its aspect as goodness by
ethics. (EPM, 127).
It is also in the Essay on Philosophical Method that the
distinction is made between "the categorical singular judgement
which composes the body of historical thought and the categor-
ical universal of philosophy" (EPM, 136), where categorical
thinking means thought which "in fact is never devoid of ob-
jective or ontological reference" (EPM, 125).

730
It appears that the Autobiography overlooks this distinction and
that it is denied altogether in the Essay on Metaphysics.
Is it possible that Knox was correct on this matter, and
that it was only in his mature middle period that Collingwood
succeeded in freeing himself from the scepticism and dogmatism
of his earlier and later periods? In the light of the evidence
we called forth in Part III this seems hardly credible. On all
the major autobiographical themes that we have examined thus far
Collingwood shows a steady and continuous development.
Nevertheless we cannot easily integrate the positions expressed
in the Essay on Metaphysics with the rest of Collingwood's
philosophy as we have investigated it thus far, nor can we
simply refuse to consider it or relegate it to a lower status
than the rest of his published writings-and especially not in
the light of the agreement between this essay and the
Autobiography. For when we read in the former that "All
metaphysical questions are historical questions, and all
metaphysical propositions are historical propositions'' (EM,
49), and that "The problems of metaphysics are historical
problems; its methods are historical methods" (EM, 62), and even
that "Metaphysics has always been an historical science" (EM,
58), we are certainly inclined to the opinion that Collingwood
has reduced, if not all of philosophy, at least all of
metaphysics to history. And when one considers that it was
Collingwood's judgment that metaphysics is the very heart of
science and civilization,

731
which stand or fall with it (EM, 22, 46, 103, 224, 233), it is
difficult not to draw the conclusion that history has replaced
philosophy as the central reconciling force, and that therefore
the role of philosophy has indeed been "liquidated by being
converted into history" (IH, x).
Even if we were to accept this as Collingwood's complete
and final judgment on metaphysics, our thesis concerning the
autobiographical interpretation as we developed it in Chapter II
would not be entirely wrecked, because the Autobiography largely
agrees with the Essay on Metaphysics on this matter:
This was my answer to the rather threadbare question "how

can metaphysics become a science?" If science means na-
turalistic science, the answer is that it had better not
try. If science means an organized body of knowledge, the
answer is: by becoming what is always has been; that is,
frankly claiming its proper status as an historical
inquiry . . . . (A, 67; cf. EM, 77).
What seems to be called into question is rather the role of
philosophical rapprochement, which plays a large part in the
Autobiography, but is not even mentioned in the Essay on Me-
taphysics. It is for this reason that we have chosen to consider
both of these issues in our final chapter: it is our belief that
when examined together these two issues-- rapprochement and
metaphysics--will illuminate one another in a way that examining
either alone will not.

732
But in pursuing our inquiry to this final point we will
be, in our own way, undergoing an experience not unlike one
which Collingwood describes in the Autobiography. For it was in
our second chapter that we encountered those peculiar passages
in which Collingwood described the turning point in his early
thinking--the closest thing to a "conversion" that is even
hinted at in the Autobiography. In two crucial chapters--the
ones in which Collingwood writes about his discovery of Q-A
logic and his rejection of the realist philosophy he had been
taught at Oxford--he writes about his experience of the Albert
Memorial which, though he describes it in grotesque terms,
became an event the assimilation of which reshaped the direction
of all his future thought.
A year or two after the outbreak of war, I was living in

London and working with a section of the Admiralty Intel-
ligence Division in the rooms of the Royal Geographical
Society. Every day I walked across Kensington Gardens and
past the Albert Memorial. The Albert Memorial began by
degrees to obsess me. Like Wordsworth's Leech-Gatherer, it
took on a strange air of significance; it seemed Like one
whom I had met with in a dream; Or like a man from some far
region sent, To give me human strength, by apt admonishment.
Everything about it was visibly misshapen, corrupt, craw-
ling, verminous; for a time I could not bear to look at it,
and passed with averted eyes; recovering from this weakness,
I forced myself to look, and to face day by day the
question: a thing so obviously, so incontrovertibly, so
indefensibly bad, why had Scott done it? . . . What relation
was there, I began to ask myself, between what he had done
and what he had tried to do? . . . If I found the monument
merely loathsome, was that perhaps my fault? Was I looking
in it for qualities it did not possess, and either ignoring
or despising those it did? (A, 29-30).
733
In a later passage Collingwood relates this experience
to his rejection of the realists' attitude toward the history of
philosophy, leading to his rejection of the view that there are
"eternal problems" in philosophy (A, 60). It led to the
discovery that every philosophical problem has its history in
which the sameness of an issue is not the sameness of a uni-
versal but that of an historical process, and the difference
that between two phases of this process (A, 62). Two pages later
Collingwood writes about his reform of metaphysics based on both
the non-existence of eternal problems and the uncovering of the
absolute presuppositions of science (A, 6667).
Collingwood's Essay on Metaphysics is our Albert Me-
morial. It is a grotesque piece of work, and like the young
Collingwood we must force ourselves not to avert our gaze from
it, and to ask ourselves what relation there is between what he
did and what he was trying to achieve. For one cannot avoid the
embarassment that arises when faced with its seemingly
intemperate and radically simplistic assertions. It seems to be
a work that is indefensibly and incontrovertibly bad, unworthy
of the subtle and sometimes vigorous mind of the Waynflete
Professor of Metaphysics, the archeologist of Hadrian's Wall,
and the author of the Essay on Philosophical Method.

734
But it is here that we must also take our cue
from Collingwood's reflections on the Albert Memorial, for we
must ask ourselves if we are looking in it for qualities it does
not possess, and either ignoring or despising those it has. Thus
we will be wary of accepting even Collingwood's own assessment
of what he is doing in this essay. Our examination of the Essay
on Philosophical Method forewarns us against accepting at face
value any advance disclaimer at the beginning of a work of what
he is not going to do; so when he writes in the Preface to the
Essay on Metaphysics that "This is not so much a book of
metaphysics as a book about metaphysics," and adds that it does
not expound his own metaphysical ideas nor criticize those of
others (EM, vii), we can be reasonably sure that he will violate
this contract with the reader just as he did in his essay on
method. To talk about philosophy in any meaningful way is to
philosophize, and this is no less true of metaphysics, as the
examples in Part III of the book make clear.
Nevertheless to a degree not necessary up to this point
we must rely on the latent principles that we have found coming
to light in our survey of the later writings thus far in Part
III. These themes may yet provide us with a means for uncovering
the consistent core of meaning underlying the paradox of
metaphysics as an historical science. For we may anticipate that
metaphysics, like history, is a philosophical concept and

735
therefore is defined by a scale of forms in which occur
oppositions such as those between a science consisting of
categorical universal statements and the same science defined as
dealing with individual matters of fact. In a scale of forms
such an opposition is reconciled when it is shown that the two
forms are phases of the same generic concept, one being a more
perfect embodiment or expression of the other. Furthermore our
analysis of the functions of consciousness and language in
Collingwood's philosophy of mind leads us to suspect that the
priority and irreducibility of the meaning-seeking role of
unifying consciousness will have considerable impact on any
discussion of absolute presuppositions. And finally, as we saw
the concept of history being refined and re-shaped in the last
chapter, we could not help but recognize the profound importance
of understanding what Collingwood means by "historical thinking"
when considering what it means to say that metaphysics is an
historical science. For if at the heart of historical thinking
is re-enactment, then calling metaphysics an historical science
means calling it a re-enactment with all that that
implies--especially the recasting of the act-object distinction
as Collingwood envisions it.
This provides us with our strategy for the rest of this
chapter. In the next section we will attempt a brief outline of
the Essay on Metaphysics, especially attending to the way in
which it reflects
736
the four major themes of the Autobiography. In section three we
shall continue this "forced look" at the essay by analyzing the
obstacles that Collingwood has placed in the path of first
philosophy in the first two parts of the Essay on Metaphysics.
In section four we shall attempt to reconstruct the unity of the
four autobiographical themes prefatory to returning to the
obstacles to first philosophy, the overcoming of which is the
aim of our fifth and final section.
We must add a final word on the limitations of this
chapter. Our program necessitates restricting our discussion to
the four major autobiographical themes. For this reason we are
not attempting a complete discussion of the contents of the
Essay on Metaphysics, which involves passing over a good deal of
interesting and perhaps even relevant material. A complete
treatment of Collingwood's metaphysics will only be possible
after a thorough study of the unpublished manuscripts,
especially those on cosmology, epistemology, logic, and, of
course, metaphysics itself. This chapter is to be regarded as a
prolegomenon to such a study--as is the entire dissertation.
2. The Autobiographical Themes in the Essay on Metaphysics.
If we are to accomplish even the first stages of this
"forced look" at Collingwood's Essay on Metaphysics we must
disabuse ourselves of
737
several pre-conceptions that we may bring to the
discussion--some of which arise from our own previous
investigation. It is only in this way that we will prevent
ourselves from looking in it for qualities it does not possess.
For if we are expecting that the Essay on Metaphysics will place
a capstone on the arch of thought that we have seen Collingwood
erecting for the last eight chapters, we will be sadly
disappointed. Ontology is dismissed. There is no discussion of
the nature of categorical universal propositions, no analysis of
the nature of "real thinking," and no final synthesis of formal,
dialectical, and Q-A logics. Realism and idealism are left in a
state of unresolved opposition. And while some of the
traditional concepts of metaphysics come up for discussion,
others are ignored altogether. Thus the existence of God is
discussed, but in a way which orthodox Christians would find
alien at best and heretical at worst; and while various senses
of the term "cause" are examined, other concepts like substance,
matter, activity, process, etc. are not. And where we might have
wished for clarification, in "first philosophy" that explicates
the presuppositions of historical science, of terms like event,
evidence, reference, inference and necessity (cf. A, 77), such
hopes are left unfulfilled. In short, the Essay on Metaphysics
is not the First Philosophy of Collingwood we might have hoped
for, and to this extent his warning in the first sentence of the
Preface is precise: it is not a book of metaphysics.

738
As we have already noted, where the lack is most sorely
felt is in the area of rapprochement. Where we might be led to
expect that a work on metaphysics by the author of the
Autobiography might have important things to say about the
identity that is sought by a reconciliational philosophy --what
in another context he might have called an object that would
completely satisfy the mind--there is not a word on such a
desperately needed topic to complete the recapitulation and form
an apotheosis of the four autobiographical themes in the later
writings. A glance at the contents of the essay tends to confirm
this observation insofar as there is considerably reflection of
the other three autobiographical themes-which makes rapprochement
conspicuous by its absence.
The first three chapters are devoted to examining
several senses of the term "metaphysics" as indicated by the
three names that Aristotle gives to "first philosophy" in the
collection of treatises following the physics
(τ α µ ε τ α τ α φ υσ ι κ α ). In these chapters Collingwood
rejects the Aristotelian proposal for metaphysics as ontology on
the grounds that a science of pure being would be so general
that it would be lacking in any determinate subject matter. But
Aristotle's description of metaphysics as a science of the first
principles of physics is defended by Collingwood as the true
subject matter of metaphysics, which is a study of the absolute
presuppositions of science. Chapters IV and V present the
version of Q-AM
739
logic which we examined in detail in our chapter on logic,
language, and mental acts, in which absolute presuppositions
(APs) are located at the apex of P-Q-A complexes in a systematic
inquiry. Chapters VI and VII develop the concept of metaphysics
as an historical science of absolute presuppositions--the
radical thesis of the Essay on Metaphysics. Chapters VIII
through XVII are devoted to what Collingwood calls
anti-metaphysics. Chapter VIII defines anti-metaphysics as
metaphysics done out of a motive of either resentment or fear,
by an amateur thinker whose attitude toward actual metaphysics
appears as hostility to it. Chapters IX through XVI examine two
versions of anti-metaphysics--psychology as the pseudo-science
of mind, and positivism from J. S. Mill to A. J. Ayer. Chapter
XVII, which concludes Part II, contrasts Collingwood's view on
absolute presuppositions with those of the modern realist,
Samuel Alexander, whose attitude of "natural piety" or
unquestioning acceptance toward APs Collingwood approved, but
whose epistemology failed to appreciate the degree to which APs
are historically grounded.
Part III is a collection of examples of metaphysical
analysis of the sort that Collingwood proposes as acceptable,
and presumably illustrating the principles of metaphysical an-
alysis as he outlined them in Part I. Chapters XVIII through XXI
take up the problem of the existence of God, especially as the
statement "God exists" is analyzed by the

740
logical positivists. Collingwood analyzes the statement not as a
verifiable or falsifiable proposition, but as an absolute
presupposition of 20th Century science, which could not exist
without the AP that there is one God or "in other words, that
there is one world of nature with one system of laws running all
through it, and one natural science which investigates it" (EM,
213). Chapters XXII through XXVIII set forth a radical
re-interpretation of Kant's metaphysics as contained in the
Transcendental Analytic section of the Critique of Pure Reason.
According to Collingwood this is "an historical study of the
absolute presuppositions generally recognized by natural
scientists in Kant's own time" and for some time afterwards (EM,
245). And finally, Chapters XXIX through XXXIV are a slightly
altered version of the article on causation that he had
published in the Proceedings of the Aristotelian Society for
1939. In it Collingwood distinguishes three developing senses of
the term "cause": as a motive for the free and deliberate act of
a conscious agent (the historical sense); as the way in which a
natural event can be produced or prevented by human intervention
(the engineering science sense), and as the relation between two
events or states such that one is related causally to another
called its effect (the theoretical science sense) (EM, 285-87).
Collingwood argues that the first is the true sense, the second
a legitimate but restricted extension of the first, and the
third an abandoned presupposition of natural science (EM, 289,
327).
741
From this bird's eye view it is an easy matter to locate
three of the four autobiographical themes. The chapters
rejecting metaphysics as ontology as well as the whole of the
section on anti-metaphysics are directed against current forms
of realism. Ontology as Collingwood describes it is the ultimate
product of universal abstraction--the limiting case when
everything determinate that can be said of an object is left out
of account and a subject matter is reached which is totally
lacking in any distinctions whatever--"pure being." There being
no definite subject matter, there is consequently nothing for
science to investigate (EM, 13-14). We recognize the view of
abstraction here that Collingwood attributes to the realists,
and "pure being" is an object so unaffected by the knowing of it
that it remains entirely indeterminate and hence unknowable.
Similarly the sort of psychology that Collingwood takes to be
anti-metaphysical is the behavioristic attempt to reduce mental
processes to mechanical events--a "pseudoscience of mind" which
adopts a materialistic and mechanistic attitude towards thought
and deliberately ignores its "criteriological" aspect (generally
its truth or falsity)--in short, treating thought as an inert
object (EM, 114-15). And the logical positivism of Ayer is
faulted for its simplistic and erroneous description of
observation in science, which assumes that "facts" are simply
observed without recognizing that observation itself presupposes
interpretation (EM, 143-46, 163). Positivism is guilty of the
error of assuming, as does Samuel Alexander,

742
that the process of knowing is a simple apprehension of
"compresence" of two things, one of which happens to be a mind
(EM, 170, 177)--the root error of realistic epistemology.
It is also not difficult to find many of the pernicious
consequences of realism cited in the Autobiography also
represented in the Essay on Metaphysics. In the form of anti-
metaphysics, realism provides the propaganda for the spread of
irrationalism in science, religion, and civilization (EM, 13342,
234). Where the survival of each of these requires the
responsible agency of human intelligence--the belief that a
person's knowing can have an effect on the outcome of a syste-
matic inquiry or a series of actions--anti-metaphysics denies
such agency, and therefore threatens the survival of the sort of
civilization that Collingwood portrayed in The New Leviathan.
Similarly Q-A logic--the second major autobiographical
theme--is not only made the groundwork for the revision of
metaphysics as a science of absolute presuppositions, but is
formulated in conscious opposition to the pronouncements of the
logical positivists, whose acceptance of an F-logical criterion
for deciding truth claims (at least for analytical statements)
is well recognized. In fact it is apparent that the peculiar
form that Q-A logic takes in the Essay on Metaphysics is due to
the fact that it is being formulated with one eye fixed on the
pronouncements
743
of logical positivism, and if it were not for this polemical
tendency we might have been offered a Q-A logic better able to
stand on its own merits rather than those it has by not being
some other unacceptable thing. Therefore Collingwood tells us
only enough about absolute presuppositions to allow us to escape
the trap laid for metaphysical statements by the logical
positivists--that they are neither tautologous (F-logically
analytical statements) nor empirically verifiably/falsifiable,
and are therefore meaningless. Collingwood charges them with
failing to ask the further question of what else they might be
before declaring them to be meaningless (EM, 162-65), and when
he asks the question for them he concludes that while it is true
that they are not verifiable or falsifiable propositions, they
nonetheless retain an inquiry-dependent function as the grounds
from which questions proceed--the function he calls
presupposition (EM, 29-31). But were it not for this preoccu-
pation with escaping from the positivistic trap Collingwood
might well have gone on to say in what sense absolute presup-
positions can be said to be true or false in a way which does
not mean verifiable or falsifiable in the empirical sense.
Instead we are left with metaphysical statements that are not
propositions (as the term is understood by positivists), not
tautologous, but yet not meaningless--all negative character-
istics. All that is left of a positive nature to absolute
presuppositions is their
744
role in logically grounding questions in a systematic inquiry.
This is precisely the role assigned to them in the Autobiography.
Of course it is the third autobiographical theme that is
dominant in the Essay on Metaphysics, since the reconciliation
of history and philosophy is made the template upon which is
forged a reformed metaphysics. When the questions, propositions,
problems and methods of metaphysics are said to be those of
history, the identification of the two is as plainly stated as
one can expect. In fact the identity is so lacking in
qualification that the reader is obliged to think as far back as
Religion and_Philosophy to find a reconciliational identity of
equivalent unsophistication. But if what is overlooked in the
reconciliational identity is, like the earlier identities of
religion, philosophy, and history, the sense of how they are
related in a process of development or in a scale of forms, is
it not the logical outcome of the line of thought in the
Autobiography whereby Collingwood decided that there are no
eternal problems? For if it is true that there are no timeless
truths, then all truths are indeed historically relative, and
the Essay on Metaphysics is an attempt to explore what that
would mean for scientific inquiry.
Nevertheless we cannot be content with this state of
affairs. In Part II and Part III we have observed how the
earlier version of reconciliation, based on the rather
simplistic orientation of religion,

745
science, history, etc. to a common object ("all of reality"),
gave way to the more sophisticated but unstable dialectical
scale of forms of knowledge which distortedly approximated to an
ideal of Absolute Knowledge, expressed as philosophy. We
followed this progression to the Essay on Philosophical Method,
where opposition and distinction are relations accepted as
normal staging between two phases of a developing concept, one
being not merely different from the other but also a more
perfect embodiment of it. And finally we found that such
hierarchical scales were found exemplified in Collingwood's
analysis of mental functions and expressive linguistic acts. In
the last chapter we found that the concept of history exhibited
this structure, and in reconciling conflicting historiographical
traditions Collingwood locates them on a developing scale
culminating in scientific history.
It was when Collingwood sought to go beyond reconcili-
ation by means of a scale of forms, and to explore the senses in
which higher levels of mental activity were expressed in
historical thinking that he reached both the high and low points
of his philosophical career: the high point being the
demonstration that history is possible only on the basis of
re-enactment--an argument which balances realism and idealism in
a remarkable re-interpretation of the act-object distinction;
and the low point being the collapse of his defense of
scientific history as a
746
separate discipline for lack of a sufficiently unique
characterization of "historical inference." It is exactly at
this critical juncture that we must pick up the thread of our
interpretative reconstruction once again.
Before doing so we must press our "forced look" to its
final stage by extracting the discomfitting features of the
Essay on Metaphysics and confronting the obstacles to first
philosophy which these features present for us. To help us with
this task we end this section with a tabulated synopsis of the
principal ideas of Parts I and II of the essay.

747
TABLE 14
METAPHYSICS AND ANTI-METAPHYSICS
1. A metaphysics of the sort indicated by Aristotle in the

treatises by that name may be regarded either as ontology,
i.e. the science of pure being, or as the organized knowledge
of the presuppositions underlying ordinary science, where
science is a body of systematic or orderly thinking about a
determinate subject matter (EM, 11).
a. But it cannot be ontology, since there cannot be orderly

thinking about an indeterminate subject matter, and pure
being is a completely indeterminate subject matter
because it represents the limiting case of the
abstractive process, where abstraction is the ignoring
of differences between individual things and the
attending only to what they have in common. When ab-
straction is pushed to the limiting case in which
everything determinate is left out, there is nothing
left for science to investigate. (EM, 12-15).
b. Metaphysics may therefore be the science of the pre-

suppositions that are logically prior to, or form the
underlying ground for, ordinary science, where ordinary
means not a constituent part of metaphysics (EM, 11-12,
20).
2. To think scientifically is to be aware that every statement

made is in answer to a question which is logically prior to
its own answer. What is stated is a proposition which can be
either true or false. Every question involves a
presupposition. To say that a question "does not arise" is
the ordinary English way of saying that it involves a
presupposition which is not in fact being made, and the fact
that something causes a certain question to arise is the
logical efficacy of that thing. The logical efficacy of a
supposition does not depend on the truth of what is
supposed, or even on its being thought true, but only on its
being supposed. To assume is to suppose by an act of free
choice. A presupposition is either relative or absolute. A
relative presupposition is one which stands relatively to
one question as its presupposition and relatively to another
question as its answer. An absolute presupposition is one
748
which stands, relatively to all questions to which it is

related, as a presupposition, never as an answer. Absolute
presuppositions are not propositions; the distinction
between truth and falsehood does not apply to absolute
presuppositions at all. (EM, 21-32).
3. The discovery of presuppositions and the distinction between

relative and absolute presuppositions is not accomplished by
low-grade thinking, such as apprehension or intuition (as in
realist theories of knowledge), nor by simple introspection.
It is accomplished by high-grade thinking, a skillful mental
effort which brings about not only a difference of degree in
the intensity of thinking, but also a difference of kind in
its quality. Instead of passive apprehension, high-grade
thinking is an active process of asking questions in a
systematic and orderly fashion. The thinking that is
involved in the disentangling and arranging of questions is
called analysis, and the analysis which detects absolute
presuppositions is called metaphysical analysis. Since
analysis is what gives science its scientific character, and
since in their method of operation science and metaphysics
are the same (i.e. insofar as they both distinguish between
relative and absolute presuppositions), science and
metaphysics stand or fall together. (EM, 34-41, 170-71).
4. All metaphysical questions are historical questions, and all

metaphysical propositions are historical propositions. The
problems of metaphysics are historical problems; its methods
are historical methods. Metaphysics is the attempt to find
out what absolute presuppositions have been made by this or
that person or group of persons, on this or that occasion or
group of occasions, in the course of this or that piece of
thinking. The metaphysical rubric proceeding an absolute
presupposition is: "In such and such a phase of scientific
thought it is (or was) absolutely presupposed that . . . ."
(EM, 47, 49, 55, 62).
5. The historian makes his own statements on his own authority

according to what he finds the evidence in his posession to
prove when he analyzes it with a certain question in mind.
The subject matter of metaphysics is about a certain class
of historical facts, namely absolute presuppositions. Its
methods are the methods of history, i.e. to get at the facts
by the interpretation of evidence. In its form metaphysics
is systematic in the sense in which all historical thought
is systematic, i.e. as exhibited in the clear and orderly
749
manner in which it states its problems and marshals and

interprets evidence for their solution; but it is not
systematic in the sense of being a deductive science. As an
historical science, metaphysics shares the presuppositions
of all history. (EM, 59-65).
6. Absolute presuppositions, like historical facts, do not

occur singly, but in sets or constellations. The presup-
positions in a constellation are logically related as a
single fact; they are made at once in one and the same piece
of thinking, and each is consupponible with all the others.
To be consupponible means that it must be logically possible
for a person who supposes any one of them to suppose
concurrently all the rest. Within the constellation each
presupposition taken separately is also a single historical
fact, and is not deductively related to the others. The
metaphysician's business is not only to study the likenesses
and unlikenesses of several different constellations of
absolute presuppositions but also to find out on what
occasions and by what processes one of them has turned into
another. One phase changes into another because the first
phase was in unstable equilibrium in which its fabric was
under strain, and the historian analyzes the internal
strains to which a given constellation is subjected, and the
means by which it takes up these strains or prevents them
from breaking into pieces. (EM, 66-67, 72-74).
7. Anti-metaphysics is metaphysics undertaken by an amateur who

does not consider it his proper job, but is impelled to do
so for various motives. It takes three forms. (a)
Progressive anti-metaphysics is metaphysics undertaken by
someone whose proper job (science) demands it, but because
it has been neglected by the professionals he has to
undertake it, and his resentment makes him feel himself to
be the professional metaphysician's enemy, and his own work
as an attack on their work. (b) Reactionary antimetaphysics
is metaphysics undertaken because one wishes to do
metaphysics consistent with the principles of obsolete
pseudo-metaphysical doctrines, but inconsistent with
contemporary metaphysics, which he fears as a danger to his
own work, and therefore his fear of this inconsistency makes
him regard his own work as an attack on metaphysics in the
contemporary sense. (c) Irrationalist antimetaphysics is
metaphysics undertaken because one wishes to abolish
scientific thinking itself in order to bring into existence
a form of life in which all the determining factors should
be emotional. (EM, 82-83, 88, 99-100).
750
8. Positivistic (progressive) anti-metaphysics fails because it

presupposes what it simultaneously denies. What it denies is
that metaphysics is possible, because the statements of
metaphysics are actually propositions which are not
empirically verifiable or falsifiable, and are therefore
pseudo-propositions. But the position which states that
metaphysical assertions are true or false propositions is
not the true science of metaphysics (i.e. the historical
science of absolute presuppositions) but pseudo-metaphysics,
and is the result of the logical mistake of confusing
presuppositions with propositions, and of assuming that it
is only propositions that can have logical efficacy in
causing questions to arise in scientific inquiry. What
positivistic anti-metaphysics presupposes is that scientific
thinking is possible and legitimate, which means that
thinking which makes absolute presuppositions (science) is
possible. But since absolute presuppositions are the
assertions that are truly metaphysical, the denial of
metaphysics is the denial of scientific thinking. Therefore
the only way that positivistic metaphysics can survive its
own criticism is by confusing true metaphysics with
pseudo-metaphysics. (EM, 148-49, 162-64, 169-71).
9. Reactionary anti-metaphysics (Samuel Alexander's realism)

fails because it does not recognize that absolute presup-
positions are not timeless truths, but have a history of
their own. This position recognizes both that the subject
matter of metaphysics is absolute presuppositions and that
these are not proven but recognized as facts and held with
an attitude of unquestioning acceptance. But what it fails
to acknowledge is that absolute presuppositions are not
facts apprehended by compresence of a mind with an object,
or even perceived as a pervasive set of characteristics of
everything that exists. They are arrived at by analytical
questioning aimed at discovering what is presupposed in a
particular case of scientific thinking. These
presuppositions are not truths recognized semper, ubique, ad
omnibus, but are presupposed by scientific thinking at a
particular time, and are themselves subject to change and
development. (EM, 172-80).
10. Irrationalist anti-metaphysics (psychology as the pseudo-

science of mine) fails because it confuses thinking with
feeling, and refuses to recognize that all thought is cri-
teriological. metaphysics is one branch of the science of
thought, because it aims at discovering absolute presup-
positions, which are thoughts. If psychology is the science
which tells us how we think, its claim to be the sciency of
absolute presuppositions appears to be legitimate.
751
But psychology treats thought not as criteriological (i.e.

as including as an integral part of itself the thought of a
standard or criterion by reference to which it is judged to
be a successful or unsuccessful, or true or false, piece of
thinking), but as an empirical (non-criteriological)
feeling--a relic of 18th century materialistic epistemology.
But to treat thought without regard to its truth is to treat
scientific thought in the same way, which makes science
itself a meaningless word. Since the conclusions of
empirical psychology with respect to thought are not
established by its methods (experimental observation of
feelings), they must arise elsewhere, and such prejudice
that teaches by precept that thought is only feeling serves
as the propaganda of irrationalism. (EM, 101-03, 107,
111-14, 117, 120, 129, 142).
752
3. Obstacles to Understanding Collingwood's Reformed Metaphysics.
If the architecture of the Essay on Metaphysics presents
us with a design that is at least three-quarters familiar, the
interior is an alienating nightmare. The reader encounters
difficulties understanding the strategy of arguments, the
absence of satisfying conclusions to a line of thought, what
appears as abusive insults rather than convincing critical
elenchi, and failure to define adequately key concepts in a way
that alleviates the reader's growing sense of puzzlement.
(a) The first obstacle one encounters is Collingwood's
unhistorical and unscholarly treatment of Aristotle's meta-
physics, which is misrepresented as a science of "pure being"
--a term that Aristotle did not use, and would not be acceptable
to him as an equivalent to his own expression, "being as being"
(τ ο ο ν η ο ν ). As a prologue to a study of metaphysics as an
historical science it is startling to find Collingwood ignoring
the obvious orientation of Aristotle's metaphysics towards
individual substance (ο υ σ ι α or τ ο δ ε τ ι ο ν ) that has a
separate existence as opposed to an attribute
(π α θ ο s κ α τ α σ υ µ β ε β η κ ο s) which can only exist
secondary to, or dependent upon, a substance.1 Furthermore it is
axiomatic
_____________________
1 Hippocrates G. Apostle, tr., Aristotle's Metaphysics
(Bloomington, 1966), 1017b 23-26, p. 83; cf. 1003b 5-10, p. 54.
753
with Aristotle that universals are incapable of separate
existence, and Collingwood need not have read far into
Aristotle's treatise to find it stated as a major obstacle
(α π ο ρ ι α ) and a definitive objection to Plato's theory of
forms that being cannot be a genus, because to be is to be a this
or something particular but a genus is something general.2 In
fact what Collingwood is describing when he writes of pure being
is what Aristotle would recognize as matter--its only
characteristic being its complete indeterminacy.
The reader is also puzzled when Collingwood, with
slightly more accuracy, cites the three titles Aristotle gave to
the kind of inquiry presented in the collection of treatises
known as the metaphysics, i.e. First Science, Wisdom, and
Theology, and then, after paraphrasing what these terms mean
(presumably for Aristotle), ignores or slightly glosses over the
reasons why Aristotle considered these different names for one
and the same science with the same object.3 Instead he
arbitrarily chooses to consider only two definitions of
metaphysics: as the science of
_______________________
2 Ibid, 998bl5-35, p. 45: "Thus if 'unity' or 'being' is
indeed a genus, no differentia will be either a being or one."
Cf. 1038al-30, p. 128: definition by downward division, i.e. by
differentiae, leads to substance.
3 In fairness to Collingwood it must be added that the

triple orientation of Aristotle's Metaphysics has proven re-
sistant to two millenia of attempts at unification until Fr.
Joseph Owens' monumental study, The Doctrine of Being in the
Aristotelian Metaphysics (Toronto, 1951).
754
pure being, and as the science of the presuppositions underlying
ordinary science (EM, 11). Since it is not transparent that
these two senses are mutually exclusive alternatives, why does
Collingwood assume that rejecting the one, i.e. metaphysics as
ontology, necessarily affirms the other? In fact they seem to
overlap, since it is an absolute presupposition of ordinary
science that its objects exist, and conversely an absolute
presupposition must exist to be a fact. And what of metaphysics
as theology, which is not further discussed in Parts I and II,
but appears again as the first example in Part III ("The
Existence of God")?
(b) Of course Collingwood can be forgiven for what might
turn out to be a creative misinterpretation of Aristotle, as
well as for an arbitrary starting point for his essay, if they
were justified by subsequent maneuvers. If metaphysics is not
the ontological analysis of pure being, perhaps it can still
survive as the grounding science of first principles that
Aristotle thought it could be (EM, 19-20). Unfortunately this
presents us with a second set of obstacles. For Aristotle such a
First Philosophy was grounded on an analytic of scientific
knowledge in general (ε π ι σ τ η µ η ) as presented in the
Organon. For Collingwood this involves first redefining science
as a Q-A process of on-going inquiry rather than a deductive
system of propositions (EM, 22-24), and then identifying the
first principles for such a system as the absolute

755
presuppositions which underly all its questions (EM, 25-28). The
third stage is to distinguish between relative and absolute
presuppositions, where the former are themselves answers to
questions and the latter are not (EM, 29-33). The province of
metaphysics is the discovery, comparison, and organizational
structure of absolute presuppositions (EM, 37-40).
The force of this argument for a science of metaphysics
depends on Collingwood's ability to maintain successfully the
distinction between relative and absolute presuppositions, for
without it there is only an arbitrary stopping point for a
regressive inquiry from answer to question to presupposition to
question to presupposition, . . . (etc.). But if an absolute
presupposition is one which stands, relatively to all questions
to which it is related, as a presupposition and never as an
answer (EM, 31), and if every statement that anybody ever makes
is made in answer to a question (EM, 23), then how can absolute
presuppositions ever be discovered? For surely discoveries are
made by asking questions, in this case of the form, "What is the
absolute presupposition of this and all other questions in this
inquiry?"
Perhaps Collingwood meant us to understand that an
absolute presupposition is never the answer to a question in a
progressive inquiry--something one comes upon along with other
answers to questions in
756
a systematic inquiry. Thus the absolute presupposition that
"every event has a cause" (EM, 50, 52, 179) is not something a
scientist discovers while pursuing the answer to a question
about the cause, for example, of an eclipse. It occurs when he
leaves off investigating particular causes and asks the
metaphysical question, "Why do I assume that there is a cause of
this or that or any event?" This is a retrogressive
question--one which is not progressing toward an answer but
retrogressing from a question to its presupposition. There would
thus be two inquiries, one being that of progressive ordinary
science and the other the retrogressive metaphysical science,
and the restrictions concerning absolute presuppositions would
refer to questions in the former inquiry, not the latter.
If this is what Collingwood meant us to understand. then
we might say that once again he has expressed himself in an
unfortunate manner, and this failure is not mitigated by his
poor choice of examples to illustrate his reformed metaphysics.
The issue of causality, for example, is a metaphysical issue of
great antiquity, yet still a contemporary problem and suitably
general enough to be a subject for reformed metaphysical
analysis. But as an example of an absolute presupposition, as
Collingwood presents it, it creates more problems that it
solves. In Chapter VI Collingwood gives the following three
versions of the principle of causality as exemplary of
Newtonian, Kantian, and Einsteinian sciences

757
respectively: (i) some events have causes; (ii) all events have
causes; and (iii) no events have causes (EM, 51-52). Setting
aside difficulties with accepting this as an historically ac-
curate assessment of the principle of causality for the phases
of physical science mentioned, the reader is still puzzled by
the apparent non-absoluteness of these presuppositions. Surely
any one of them is an answer to the question, "Do all, some, or
no events have causes?"--a question which Aristotle may well
have pondered. And are there not questions of even greater
priority, which Aristotle definitely did ponder, viz. "What is a
cause?" and "Are there any causes?" Is not Collingwood's
analysis of causality in the final example of the Essay on
Metaphysics itself an examination of these two questions,
insofar as (in Chapter XXIX) he discusses the three meanings of
the term "cause" and rejects the third as an abandoned AP of
contemporary (Einsteinian) science?
And even these three senses of "cause" as Collingwood
defines them are not primary, since they can themselves be the
answer to further questions. For example in Sense I, "that which
is 'caused' is the free and deliberate act of a conscious and
responsible agent, and 'causing' him to do it means affording
him a motive for doing it" (EM, 285). If a cause is a kind of
motive, what is a motive? Are there various motives, of which
cause is but one, or are cause and motive equivalent terms in
every sense? Are there free but not deliberate acts of agents,
or not-free but
758
deliberate acts? Etc. If to have an attitude of unquestioning
acceptance toward APs is to stifle as illegitimate questions
such as these, then Collingwood's reform of metaphysics takes on
the appearance of an intellectual purge, an example of an anti--
metaphysics which inhibits scientific inquiry rather than pro-
moting it. And if it is possible for an historian as acute as
Collingwood to be mistaken about the absoluteness of pre-
suppositions in a case as relatively simple as this one, what
about more complex and subtle systems of APs in science? What is
the measure of the absoluteness of presuppositions?
But even accepting some friendly modifications of his
Q-A logic as applied to his metaphysical program, there are
further difficulties with the doctrine of absolute presupposi-
tions. The problems start with the fact that Collingwood never
says what an absolute presupposition is. Is it at least a
statement? One assumes it must be, since it has a sentential
form, and is made in answer to a metaphysical question, or
during a metaphysical analysis. But why does Collingwood then
forbid us, on penalty of lapsing into metaphysical absurdity,
from asking the further question, "Is this statement (we do not
call it a proposition, since propositions are statements made
only in answer to progressive questions in an inquiry) true or
false?" For if it is not true or false in the same sense that a

759
proposition in the progressive inquiry is, does that mean that
it is not true or false in any sense whatever?
The problem becomes acute when it comes to constella-
tions of presuppositions, since Collingwood allows that meta-
physical analysis can uncover absolute presuppositions in groups
called constellations, and can furthermore determine if within
that constellation one or more presuppositions can be
"con-supponible" with the others (EM, 66, 76, 287). If not, then
there will be "strains" in the constellation, which will make it
"unstable" and therefore eligible for the process of
transformation into another constellation (EM, 48 n.l; 74, 76).
But how is it possible to recognize that one AP is not
con-supponible with another without deciding which APs within
the constellation are more stable, or at least less stressful,
than the others? Does that not assume that because one group of
APs are what they are, a given AP or group of APs cannot
consistently be included in the con-supponible group? And is not
consistency one acceptable sense of the term, "true" and
inconsistency of "false"?
And even if we ignore Collingwood's own statement that
the "strains" in a constellation of APs is due to their "mutual
incompatibility" (EM, 287) and suggest that such strains be
analyzed solely on the grounds of their decreased adequacy or
efficacy in causing questions to arise, the problem is not
resolved--only shifted to even more

760
obscure territory. Certainly a presupposition such as the
perfection of circular celestial orbits was "efficacious" in
causing questions to arise, and that is not the sense in which
it caused "strains" in the constellation of presuppositions
which led to the passage from Ptolomaic to Keplerian astronomy.
What we are left with is a situation in which the
metaphysician-historian, lacking adequate definitions of terms
like "absolute presupposition," "constellation," "con-
supponible," "strain" etc., is set a seemingly impossible task.
Lacking any clear idea of his subject matter, we must ask, like
Plato's Socrates, how he is to know when he has succeeded in
finding an instance of what he is looking for?
(c) Which brings us to our third set of problems. Even
if we could accept a non-ontological metaphysics of absolute
presuppositions, the announcement that these are discovered by
historical methods, and that all metaphysical propositions are
historical propositions is an expression of historicism extreme
enough to warrant assigning Collingwood a paradigmatic position
in the checkered career of the subject of radical revisions of
traditional metaphysics. As we have already pointed out, the
doctrine that metaphysics is an historical science is the
positive version of the principle stated negatively in the
Autobiography that in philosophy there are no eternal problems
(A, 60; cf. EM, 64-65). But when coupled with the remarkable
statement that metaphysics has always been an historical

761
science (albeit not fully conscious of the fact and for that
reason never fully scientific) it appears to be a revision not
compatible with even the acceptable version of Aristotle's
metaphysics. For it certainly cannot be denied that when
Aristotle tried to uncover the first principles of physics he
did not have in mind a science bound to 4th century B.C. Greek
culture, but one which would achieve the status of ε π ι σ τ η µ η
--science as a stable understanding of change based on
unchanging first principles. And when Kant set himself the task
of discovering how metaphysics as a science is possible, he did
not think its possibility rested on truth-neutral
presuppositions which would change with time (CEPC, 137). On the
contrary, every indication is that he thought himself to have
discovered all the true principles of scientific thought, and
that synthetic a priori judgments have a necessity that is not
historically contingent but is rooted in the structure of the
knowing mind.
Of course, Collingwood was aware that his predecessors
did not explicitly say that what they were doing was history,
and he merely chides them for not understanding what they were
in fact doing (EM, 18-20, 58). Our difficulty with Collingwood's
historical metaphysics is deeper than a factual error on the
matter of what metaphysics has "always" been. It has to do with
the sort of "facts" that the metaphysician is setting out to
discover. In the last chapter we found Collingwood arguing that

762
what makes history scientific is precisely the degree to which
the historian decides what is to be accepted as a fact, and that
this is dependent on the historian's questioning activity. Now
we find Collingwood taking just the reverse position: it is a
certain kind of fact (AP) that determines what questions are
asked. Approaching the Essay on Metaphysics from this
perspective we are forced to conclude that as fact to be
discovered, an AP must be the answer to a question. Collingwood
appears to acknowledge this himself when he indicates that the
"metaphysical rubric" (i.e. "In this or that piece of scientific
thinking by this person or group of persons at such and such a
time it was absolutely presupposed that . . . ") is necessary
for the factuality of an AP to be established.
But this appears to involve Collingwood in a contra-
diction: if it is an historical fact, an absolute presupposition
is determined by questions; and if it is metaphysically
absolute, a presupposition is not determined by any question,
but is that which specifies which questions sensibly arise in a
systematic inquiry. This time the avenue of escape by pos-
tulating the bi-directionality of inquiry has been cut off by
the fusion of both routes: the methods of metaphysics are the
methods of history, and the presuppositions of metaphysics are
those of history. Therefore if history has presuppositions (and
Collingwood admits that it does), there is at least one case
where absolute presuppositions

763
are the answers to questions in a progressive inquiry, namely
the case of history. Is Collingwood prepared to defend the
thesis that the absolute presuppositions of history are
determined to be facts by the questions of history? Can the
serpent swallow its own tail?
The issue of the presuppositions of history brings us up
before another aspect of the obstacle, and this one joins with
the first to form the beginning of a barricade. In the
Autobiography Collingwood has told us that the re-enacted his-
torical event is prevented from becoming confused with the
present thinking of the historian by being "incapsulated in a
context of present thoughts." In the last chapter we found
Collingwood arguing that the historian is able to keep his own a
priori imagination from becoming unglued, so that the historian
is prevented from confusing himself with Admiral Nelson at the
Battle of Trafalgar. But to deny ontology its day runs the risk
of being incapable of distinguishing such realms as these in
anything but a hypothetical way. By what presuppositions does
the historian distinguish between his own "real" life (the
ordinary world of lived experience or lebenswelt) and the "real"
world of his historical protagonists? Aside from the practical
observation that any bonafide confusion between these realms may
make life-sustaining decisions rather poorly grounded, the
theoretical implications for Collingwood's reformed metaphysics
threaten to be equally tragic. For it will not

764
do to define the domain of history as the "deeds of men done in
the past," and then refuse to deal with the ontological question
of how past and present are distinguished in one's "real"
situation. Somehow the presuppositions of what we are calling
ordinary experience are the ground from which and within which
the presuppositions of history are located. That distinction and
that relationship are not well served by the simple
identification of history and metaphysics.
(d) Having raised the spectre of an a priori imagination
run amuck, we turn to our final group of obstacles, those
centering on his discussion of anti-metaphysics, and especially
his vituperation on psychology. For while he castigates
psychology as the pseudo-science of thought, in the entire
section on anti-metaphysics he himself engages in the crassest
kind of popular psychologizing. When he describes those
philosophical "amateurs" who are thrust into doing metaphysics
in the guise of anti-metaphysics, he says that they take an
attitude opposed to metaphysics because they either fear or
resent the subject matter. On what grounds does he make this
assertion? In this imputation of motives, is not Collingwood
out-psychologizing the psychologists? Completely ignoring any
positive benefit which behavioral science has had on
civilization, Collingwood's assault appears to itself contribute
to the "propaganda of irrationalism" that he attributes to
psychologists: for certainly there is

765
a kind of irrationalism to rationalism--i.e. reason which fails
to provide reason with respect to un-reason or sub-reason--and
this is where psychology has made its greatest inroads, racing
in where philosophy feared to tread.4 But at best Collingwood's
diatribe is anti-metaphysics in the sense that it not only
discovers an absolute presupposition of what is generally
accepted to be a contemporary science (psychology) but also de-
clares this presupposition to be false; for what else would it
mean to call psychology a "pseudo-science" of thought? Does the
Greek prefix succeed in disguising the fact that a false-science
is one which proceeds from a false absolute presupposition?
What is worse, the criticism that Collingwood levels at
psychology appears to be equally applicable to his own reformed
metaphysics--i.e. that it fails to treat thought as
criteriological, meaning that it ignores the aspect of thought
as true or false (EM, 107, 115). For are not
_______________________
4 Is there any place in Collingwood's version of civi-
lization, or in his philosophy of mind, for the less-than-fully
rational man? Surely children have a place, according to The New
Leviathan, since they are in process of becoming fully rational.
But what about those who never will--the mentally handicapped
and the mentally ill? Collingwood comes close to aristocratic
arrogance when he sets standards below which what is occurring
does not deserve the name of thought. Psychology, whatever its
faults, has made advances in the understanding and treatment of
sub-rational behavior, and Collingwood's failure to acknowledge
this progress in the Essay on_Metaphysics is not to his credit.
766
absolute presuppositions thoughts? And treating them as
true/false neutral--does this not mean treating them as "data"
or facts on a par with feeling and its objects? To say that they
may be something of which one may not be aware either dodges the
question or severs absolute presuppositions from their logical
function--their efficacy in causing questions to arise, a
function that Collingwood is anxious to maintain. Once an ab-
solute presupposition is discovered it is invalid to ask if it
is true, or on what evidence it is accepted, or how it is to be
demonstrated, or by what right it is presupposed (EM, 47). In
short, the acceptance of an absolute presupposition as a
truth-neutral fact precludes any question of treating it as a
thought, and therefore renders it on a par with objects of
feeling. It reduces a thought (an absolute presupposition) to
something that is not thought (something that is not cri-
teriological) and is therefore as much a pseudo-science of
thought as psychology. By what right does Collingwood then
declare that metaphysics is one branch of the science of thought
(EM, 101)?
And while he appears to be less vulnerable in his as-
sault on A. J. Ayer and logical positivism, one still cannot
help but notice that Collingwood is guilty of some of the same
charges that he levels against them. The most serious is that
after conceding the victory to them by admitting that absolute
presuppositions are not verifiable or falsifiable,

767
he then counterattacks with the weaker charge that they did not
ask what metaphysical assertions are before declaring them to be
either tautologous or meaningless (EM, 165). But has Collingwood
not made the same mistake? Instead of asking what absolute
presuppositions are if they are not true or false as answers to
questions, he leads the reader to believe that they are not true
or false in any other sense either. He never really tells us
what absolute presuppositions are.
4. The Unity of the Autobiographical Themes.
Is there some standpoint from within the framework
provided by the autobiographical interpretation that will allow
us to overcome these obstacles, or at least to soften some of
these harsh features of the Essay on Metaphysics? Having forced
ourselves to face the paradoxes of Collingwood's reformed
metaphysics, if we are determined to assimilate our experience
of this grotesque memorial we must ask ourselves what relation
there is between what Collingwood did and what he was trying to
achieve. But this requires understanding it not just in the
negative and partial sense in which it is an expression of his
opposition to "realism" as a philosophical movement that
(arguably) threatened to disrupt science and civilization, but
in the positive sense of a work that issued from the unified
center of his thought.

768
For another thinker this might not be necessary or even
possible, since it might not be considered an ideal worth
striving for. A philosopher who conceives of his role in a more
analytic or critical vein, for example, might be content to put
forth his thought in a series of vignettes on various subjects,
and it may be of no concern to him what the relationships are
not only between the subjects chosen but even among the
conclusions reached by his analysis--so long as it was done in a
craftsmanly manner. But for the Collingwood of the
Autobiography, for whom the goal of philosophy is rapprochement,
this is not a satisfactory state of affairs. To this extent
Collingwood was never an analytic philosopher.
Nor did he set out to be. One of Collingwood's first
public addresses was delivered at the 1919 Ruskin Centenary
Conference held near his home in Coniston--where he also spent
his final days. In this address, given at the beginning of his
intellectual career, Collingwood laid out, with his char-
acteristic self-assurance, what is to be expected of a phil-
osopher. Speaking of Ruskin's failure to achieve a fully
philosophical outlook (and bear in mind that Collingwood's
father was Ruskin's biographer and close friend), he then
distinguished between "having a philosophy" and "being a
philosopher." Ruskin had a philosophy, but he was not a phil-
osopher (EPA, 11). What does this mean?

769
When I speak of a man's philosophy, I mean something of

this sort. I see a man living a long and busy life; I see
him doing a large number of different things, or writing a
large number of different books. And I ask myself, do these
actions, or these books, hang together? Is there any reason
why the man who wrote this book should have gone on to
write that one, or is it pure chance? Is there anything
like a constant purpose, or a consistent point of view,
running through all the man's work? Now if you ask these
questions about a particular man, you will generally find
that there are certain central principles which the man
takes as fundamental and incontrovertible, which he assumes
as true in all his thinking and acting. These principles
form, as it were, the nucleus of his whole mental life:
they are the centre from which all his activities radiate.
You may think of them as a kind of ring of solid
thought--something infinitely tough and hard and
resistant--to which everything the man does is attached.
The ring is formed of a number of different ideas or
principles, welded together by some force of mutual
cohesion. This ring of thought . . . is what I mean by a
man's philosophy . . . . The fact seems to be that a man's
deepest convictions are precisely those he never puts into
words . . . . (E)verybody has a philosophy, but only the
philosopher makes it is his business to probe into the mind
and lay bare that recess in which the ultimate beliefs lie
hidden. (EPA, 9-11).
Here we have the basic program for the discovery of absolute
presuppositions that only came to fruition twenty years later in
the Essay on Metaphysics. But can we apply this standard to
Collingwood himself? Are there certain central principles which
he takes as fundamental and incontrovertible, and which he
assumes as true in all his thinking?
Throughout both cycles of our carrousel excursion
through Collingwood's published writings we have repeatedly
tried to catch hold of this "ring of thought" and to measure the
degree to which it conforms to the autobiographical in-
terpretation. We have examined the four

770
autobiographical themes as they occur and re-occur in the
earlier and later writings. But we have examined them separately
rather than as a coherent system, and the weakness of this
strategy is now apparent; for as we confront the obstacles to
first philosophy in the Essay on Metaphysics it is not clear how
the four themes form a "consistent point of view" from which
vantage point we can measure the strength of his self-interpre-
tation by its ability to aid us in untying the knots and laying
claim to Collingwood's metaphysical inheritance for ourselves.
In particular it is not apparent why an anti-realist
position necessarily implies an anti-F-logic, or why the latter
gives rise to either the Q-A logic of the Autobiography and the
Essay on Metaphysics, or the D-logic of Speculum Mentis and the
Essay on Philosophical Method. Nor is it evident how either of
these themes is related to an ideal of rapprochement, which
seems to have more to do with resolving experiential alienation
than with formal contradiction or dialectical opposition. Nor
again is it clear how a reconciliation between history and
philosophy can be regarded as a paradigmatic for rapprochement
philosophy in all its forms, or finally how such a philosophy
can remain opposed to realism root and branch, given its
orientation toward unification of opposing viewpoints like that
of realism and idealism. So while we have steadfastly

771
resisted the temptation to provide the missing "figure in the
carpet" where Collingwood has not done so himself, this
unfortunately leaves us with a badly fractured Collingwood who
remains for the most part true to the autobiographical
self-interpretation, but at the expense of a clear understanding
of the unity of his thought, by which standard he would surely
wish his own philosophy to be judged.
Therefore we pause at this point to take a final overall
look at Collingwood's philosophy as we have examined it to this
point, trusting that by compressing it into the span of a few
pages we can reveal its architectural strengths and shed its
decorative weaknesses.
We began our survey with his earliest published book,
Religion and Philosophy, the highlight of which, Collingwood
recalled (some twenty-two years later in the Autobiography), was
a passage in which he criticized psychological studies of
religion for their treatment of mind as an external phenomenon
or thing. Such an approach to mind renders religious con-
sciousness opaque, because it refuses to participate in the
thought processes it claims to be studying, which leaves "the
cold unreality of thought which is the thought of nothing, ac-
tion with no purpose, and fact with no meaning" (RP, 42; FR,
77). With the unbalanced negativism typical of much of his later
philosophy, he failed to tell his readers in the Autobiography
what the positive phase to this criticism was. This we found

772
for ourselves by examining the argument of Religion and
Philosophy, which pivots on the reality of communication and
inter-personal identity: for it was in this earliest work that
we found Collingwood already arguing that the starting point of
any investigation of mind or any theory of knowledge is the fact
that persons do communicate their knowledge, and this means that
two or more persons can actually share the same knowledge (RP,
98-99; FR, 170-71). He goes so far as to say that since the esse
of mind is de hac re cogitare, when two minds think the same
thought they become actually one mind, sharing between them the
unity of consciousness which is the mark of the individual (RP,
101; FR, 173), and by de hac re it is clear that he did not mean
primarily a material entity but a meaning. His argument about
re-enactment in The Idea of History is the final refinement of
this very same line of thought: re-enacting someone else's act
of thought means thinking that same thought for oneself, not one
like it (IH, 286-92). The proper context for his early
rapprochement philosophy is therefore to be found in the
communicative relationship of minds with one another, the
failure of which (as error) calls forth the effort of
reconciliation.
But so much cannot be said without saying more, since it
is evident that a mind can communicate with itself--or fail to
do so. This becomes the overall concern of his second
philosophical work, Speculum

773
Mentis, which expands the ideal of rapprochement to include the
overcoming not only of the communication gaps in meaning between
persons but also the same kind of disunity within oneself. The
forms under which we interpret our own experience--as imagined,
believed, observed, inferred, etc.--are in the modern world
alienated from one another, which tends to render the unity of
self-consciousness in need of internal rapprochement (SM, 30,
4142). The basis for achieving this inner reconciliation is
modeled on the same process that occurs when failure to com-
municate makes the reconciliation between persons necessary.
This cannot be by the kind of knowing that occurs in art, where
the mind contemplates monadic imaginary objects (SM, 60-61); nor
by religious consciousness, where the mind asserts as true a
sacred object (God) as a symbol, the meaning of which it can
never literally translate (SM, 119, 128-29); nor yet by
scientific thinking, which leaves its object always separated
from mind by the distance measurable by the difference between
universal and particular, law and instance (SM, 185-86). The
kind of reconciling knowledge we seek to repair the torn fabric
of self-consciousness is first fully exemplified in historical
thinking, which is the first truly non-abstract or concrete
knowledge, because it is the first form of experience in which
the object is individually recognizable as something wholly the
same as the mind which seeks it--i.e. thought as an active
agency.
774
While the concept of history and its philosophy has a
long and complex development in Collingwood's career, he never
abandoned this insight. He always returned to the historical act
of an agent as the model for sense-location in thinking, the
active dimension of mind stressed by the Italian and other
post-Kantian idealists, with whom he felt a spiritual kinship.5
With historical thinking there is no feature of experience, no
attitude of mind towards its object which is alien (SM, 218),
because it is all the work of mind itself. And where history
appears to set itself the impossible task of understanding the
infinite whole of fact (SM, 239), it is relieved from this
impossible burden by the limiting reflection of itself in each
of its objects, each being a "mirror of the mind." This is where
historical thought provides a challenge to typically
sensation-bound empiricist epistemologies. Where the "plain
man's metaphysic" places mind and matter out side one another,
and then finds itself unable to bridge the gap that it has
postulated (RP, 73), empiricist epistemology provides an
intermediary in the form of sensation, but is equally baffled by
the lack of unifying meaning in the flux of sensation (SM, 188).
In both cases that which is known is an object presumed to be
indifferent to the act which seeks to grasp it. But in the
_____________________
5 Cf. "The Present Need of a Philosophy," Philosophy, 9
(1934), 262-65: "What is needed to-day is . . . a philosophy
showing that the human will is of a piece with nature in being
genuinely creative, a vera causa, though singular in being
consciously creative."
775
case of history the object is itself a thought which seeks
understanding (IH, 294). Historical thought is saved from the
ultimate frustration of trying to penetrate the infinite and
unknowable realm of the "thing in itself" by the recognition of
itself in each of the objects it studies, each of these being a
"concrete universal" or individual meaning expressed as a
deliberate act--a "part" in which the whole mind is present and
in which the whole agent is reflected (SM, 218-19).6 With
historical understanding the gap between mind and its object is
bridged from both sides, the object reaching out to be
understood (as evidence), and consciousness reaching to grasp
the same act in its aspect as past thought (as historical
interpretation) (IH, 304). Reconciliation at this point is
successful because the mind has found something wholly
intelligible: that which is object is also subject (SM, 242-45,
249).
In the essays on the philosophy of history which Col-
lingwood wrote in the decade of the twenties we find him con-
tinually trying to reverse the naturalistic or empiricistic bias
of his readers, including that of some of his idealistic
contemporaries (like Croce, whose work he admired).
__________________________
6 Caesar's crossing the Rubicon was not the mere physi-
cal passage of a body across water, nor was it the mere act of
his legs. It was also not an imaginary or hypothetical crossing.
It was an expression of the reality that was Caesar, and
expressed that which Caesar represents in history. But it also
is only a "part" of Caesar's life--but a significant part, a
part that reflected the whole of it: his courage and daring, his
foresight and confidence in risking the fortunes of Rome.
776
In these essays it becomes increasingly more obvious that in
making the history-philosophy reconciliation paradigmatic for
all other kinds of rapprochement between forms of experience, he
is suggesting that rather than using the mind-thing or mind--
object polarity as the primary instance of knowing and then
extending this to historical events (which tends to render
events thing-like or objective, but at the expense of always
remaining distantiated from the event by the same attitude of
abstract objectivity) the exact reverse of this situation is
what is called for in 20th Century thought. In historical
thinking the mind-mind polarity is regarded as primary, and as
paradigmatic for thinking; the mind-object situation is
secondary or derivative. Collingwood's entire philosophical
development is an attempt to carry out the program implicit in
this paradigm shift.
But if the mind-object dichotomy is overcome by taking
the mind-mind identity exemplified in historical understanding
as the primary instance of what it is to successfully know
something, and if Collingwood found himself allied with the
Italian idealists in such a fundamental standpoint, his
articulation of this conviction took a direction all its own. It
has both realistic and idealistic dimensions. Its realistic
moment is in the recognition of the factuality of history and
its bonds with evidence; its idealistic moment is in analysis of
the thought processes

777
that are involved in interpreting that evidence, and the
criteria for deciding when the interpretative effort results in
the unification of thought that is being sought. As we traced
his thought through its development in the essays of the
twenties into the period during which he composed the Essay on
Philosophical Method, we found Collingwood showing a growing
awareness of the levels of meaning to the concept of history,
and an increasing sensitivity to the criteria whereby one can
decide not only when minds are reconciled, or even when forms of
consciousness within a mind have attained rapprochement, but
when individual acts of consciousness are unified as an
achievement of the overall nisus of consciousness toward meaning
at all levels of thought. (a) In Religion and Philosophy
reconciliation is regarded as complete when it is shown that two
forms of thought (such as religion and philosophy) intend the
same object, or mutually co-imply (or presuppose) one
another--without recognizing any difference of degree or kind in
the resulting identification (RP, 107-19; FR, 178-88). (b) In
Speculum Mentis this "concrete identity of cooperation" becomes
a reconciliation by means of a developing "dialectical series"
in which each of the reconciliata is a modification of its
predecessor, but in which relations of opposition and
differences of kind do not yet entail relations of distinction
and differences of degree (SM, 55, 206-208). (c) In the Essay on
Philosophical Method
778
Collingwood completed this line of thought in his "scale of
forms" in which the rapprochement (or "overlap") of concepts
entails that between the reconciliata are relations of both
opposition and distinction, and differences of both degree and
kind (EPM, 73-76). The ultimate grounding of rapprochement at
this point is dependent on the inclusive and exclusive
relationships that are involved in the intensional structures of
meaning.
But while we have found that there is an increasingly
more prominent awareness of the meaning-component to knowledge
in Collingwood's later philosophy we have been careful to note
in our survey of the earlier works that this is neither an un-
heralded arrival nor a novel idea. Thus in addition to being his
point in common with the Italian idealists, before that alliance
became explicit we find Collingwood (in Religion and Philosophy)
arguing that the true task of historical theology is to find out
not only what was said by the historical Jesus, but what was
meant (RP, 43; FR, 78), and that proving the existence of matter
or of God requires understanding first what meanings are
attached to the words (RP, 62-63). And in Speculum Mentis
Collingwood carries this thought one step further by arguing
that each of the forms of experience is characterized not only
as a form of consciousness but by its typical form of expression
in language--questioning, asserting, generalizing, referring,
inferring, etc. He went as far as to say that the (unifying)
activity of philosophy is its "translation" of various forms of

779
experience into language literally true--an activity which is
called forth especially where there is a fusion of symbol and
meaning, as in art and religion (SM, 128-30). The process of
translating, interpreting, or developing truth out of the
partiality and abstractness of error is the dialectical
self-criticism of thought, and the task of philosophy as
absolute knowledge (SM, 252-53).
Of course it was only in the Essay on Philosophical
Method that the idea of dialectical development is made explicit
and rapprochement is linked to the overlap of concepts and their
relationship in a developing scale of forms. Here we found
Collingwood contrasting philosophical and non-philosophical
concepts as two phases in the development of an idea, and
stressing how one can recognize when the concept has been
typically worked out philosophically (EPM, 59-61, 73). This
rapprochement of the inner alienation between concepts is now
made the basis for deciding how and when the process of
reconciliation is complete, and carries the unity of
philosophical meaning forward from the conceptual to the
propositional and inferential levels of thought (EPM, 100-101,
161-63). It was the closest that Collingwood would be able to
come to developing a complete philosophical logic based on the
ideal of rapprochement, and the key to understanding it is to
fully grasp the clue that he himself provided at the beginning
of the investigation: philosophical reconciliation is

780
recognized by the overlap of concepts, which is an overlap of
intension between meanings rather than an overlap of extension
between classes.
The Essay on_Philosophical Method called attention to
the characteristic way in which a philosophical concept is de-
fined by means of a scale of overlapping forms, in which any
point on the scale intentionally includes all the meanings in
the scale up to that point (EPM, 90, 101). It also pointed to
the peculiarity that in such a scale there is no zero point--no
point at which a concept's meaning vanishes altogether (EPM,
81). In Part III we have encountered both of these
principles--definition by means of a scale of forms and the
irreducibility of meaning structures--exemplified in his
philosophy of mind at all levels of consciousness. In The
Principles of Art the development of feeling is described first
as the uninterpreted flux of sense, then as prepared for
interpretation by being perpetuated and domesticated by
consciousness or attention, and then finally as explicit
thought, where thinking about a sensum means interpreting it
(PA, 194, 212-13). And in that same work language is defined as
a scale of forms of expression, where psychical expression,
gestures, speech, and intellectualized or symbolic language form
a developing series in which each term sums up and intensionally
includes those which preceed it (PA, 228-61). In The New
Leviathan the fundamental parameter of meaning is re-affirmed
when language is defined

781
as any system of bodily movements whereby the men who make them
mean or signify something (NL, 6.1), and is described as an
abstraction from discourse--a system of sounds or the like as
having meanings (NL, 6.11, 6.18). The ensuing scale of forms of
consciousness is the product of language (NL, 6.4-6.41). And
finally in The Idea of History the meaning of an historical act
is the individuality aimed at in historical understanding (IH,
303), and is at the center of his radical re-interpretation of
the act-object distinction on the grounds of which he showed how
history as re-enactment is alone possible (IH, 282-302).
However it is the realistic moment to this process of
thought that has haunted our reflections on Collingwood's
thought, and this is the ghost that confronts us once again as
we approach the central paradox of historical metaphysics. For
in the previous chapter we wondered if the essential ideality of
the past left any room for factuality as something not only made
but also found, not only accepted but also given. But the
factuality of history is not an abstract, brute given-ness, it
is itself a thought and therefore mediated by meaning. The
reconciliation between ideality and factuality is an extension
of the rapprochement between persons and concepts: it is carried
out by an act of interpretation. Just as the reconciliation
between persons is premissed on the fact of communication, the
rapprochement between the historian and the thoughts

782
which constitute his subject matter is mediated by historical
inference, which takes the form of critical reflection on
evidence (IH, 133). It is by the interpretation of evidence that
we communicate with the past: history is the knowledge of a
significant present (evidence) by means of acts of
interpretative thought informed by principles which govern all
evidence as such (philosophical principles, such as the
distinction between universal judgments of the form, "All S is
P." particular judgments of the form, "Some S is P." and
individual judgments "This S is P." this last being the
interpretative judgment of history (EPH, 136)), and principles
which govern specific groups of evidence (scientific principles,
such as this historical presupposition that events of the past
are localized in space and time, and that historical narrative
must be consistent with itself (IH, 246-47)) (EPH, 136-37). The
act of interpretation is not a process extrinsic to the fact
itself: interpretation is only the historical fact further
specified (RP, 46; FR, 80; EPM, 170), because a common meaning
is the irreducible basis of them both. In this the process of
historical interpretation is not different from the general
interpretative processes of all perception and thought.
Perception is nothing other than the interpretation of sense
(SM, 204-05), and to think a concept is to interpret a fact in
terms of it (EPH, 28-29; PA, 194). Therefore it is but an
extension of the reconciliational process of interpretation when
from potential historical evidence (the whole

783
perceptible world) actual evidence is selected and perpetuated
by means of historical inference based on the exercise of a
priori historical imagination (IH, 246-47, 280-81).
But how is such knowledge achieved? How is evidence
interpreted and individual narrative judgments formulated? Here
again we must return to Collingwood's paradigm for re-
conciliational thinking, the mind-mind polarity and the fact of
inter-personal communication--but this time the trail of
evidence is less distinct. Closest to us are the unmistakable
assertions in The Idea of History that evidence is interpreted
by the systematic placement of intelligent and informed
questions (IH, 273-75), and in the Essay on Metaphysics and the
Autobiography that all scientific knowledge is an application of
Q-A logic (EM, 22-24, 38; A, 30-37). But as we traced this
evidence into Collingwood's earlier philosophy, in the direction
indicated in the Autobiography, we were startled by several
unexpected (and autobiographically unindicated) developments.
Where the Autobiography leads us to believe that Q-A logic was
meant to be an alternative to F-logic, the latter being the only
methodic tool of the realists, our own findings indicated not
only that in his early writings the alternative to F-logic is
not Q-A logic, but dialectic or D-logic, but also that Q-A logic
is not a defensible substitute for F-logic. In the case of
dialectic, the appearance of deliberate concealment in the
Autobiography was dispelled by the recognition of D-logic as the
final
784
form of the criteria for reconciliational completion in the
Essay on Philosophical Method. But what of the indistinctness of
Q-A logic in the early phases of his thought, and how does Q-A
logic relate to the unity of that thought as we have been
developing it in the last few pages?
Part of the answer to these questions is available to us
from the published writings, but part of it remains for us to
reconstruct for ourselves. In our examination of the early works
in Part II we found Collingwood identifying logic with the
theory of knowledge: abstract logic deals with processes of the
subject and does not consider differences of the object (RP, 15;
FR, 53). It is contrasted with concrete or dialectical logic,
which considers thought in its relation to its object (SM,
274-77), and it is the latter that is the true instrument of
philosophy. However if thought is regarded as a process rather
than as a product, the way in which it comes to be is by
questioning or supposal--the process whereby assertion develops
from non-assertive (hypothetical or questioning) thought (SM,
186-89). In Part III we traced this argument forward into the
Q-A logic of the Autobiography and the Essay on Metaphysics
(Q-AA and Q-AM respectively), where the ambiguity of hypothesis
as non-assertive and as pre-assertive is resolved by the
distinction between supposing and questioning, which ultimately
issued in the distinction between presuppositions and questions
(A, 30-39; EM, 27-29). In Part III

785
we also found that when taken as a whole his views on logic
indicate that Q-A logic is not a competitor to, or substitute
for, F-logic, but is rather an informal presentation of some of
the rules that pertain to the elements of rational inquiry as an
application of the reconciliational process of interpretation.
We thus found ourselves with a striking example of Collingwood's
irony: a philosophical logic posing as a work on philosophical
method, and a philosophical methodology described as Q-A logic.
But if we resist the temptation to be misled by such labels, it
is clear (a) that for Collingwood all logic pertains primarily
to the meaning dimension of thought rather than to any sort of
meaning-independent formal system; (b) that the criteria for
deciding when thought has become philosophically consistent or
reconciled with self-knowledge are presented by the D-logic of
the Essay on Philosophical Method; and (c) that Q-A logic is the
methodology for acquiring knowledge, and the means by which the
interpretative processes of thought are directed at achieving
unified meanings.
The part of this process that remains indistinct, and
which we are left to reconstruct for ourselves, is the logical
genesis of these ideas within Collingwood's overall plan for
rapprochement philosophy. Here we have only a few scant hints
from the Autobiography to guide us. We are

786
told there that Q-A logic, the primacy of historical thought as
the rapprochement paradigm for all forms of experience, and the
rejection of realist epistemology are all recalled by him as
having occurred during the period, and as an outcome, of his
reflections on the Albert Memorial (A, 30, 60, 67). But we are
not told how the monument served to focus these lines of
thought, how the coalesced to form a pattern, or what central
idea allowed him to think that the solution to one of these
problems is tied to the solution of all of them. The only clues
he leaves for us are the questions that he began to put to
himself during his daily communings with the Albert Memor-
ial--why Scott had created such an aesthetically offensive
monstrosity, if its true purpose is masked by asking the wrong
questions about it, and if such questions arise due to false
expectations which prevented him from appreciating its true
worth.
Unless there are, in the unpublished writings and
diaries, indications of how these ideas evolved, we shall
probably never know what went through Collingwood's mind during
these encounters. Without knowing what particular features of
the monument he found offensive, or what his expectations were,
or the insights that he had that allowed him to overcome his re
wlsion, we can only surmise a framework for this intellectual
and personal event. But we do know what feelings the Albert
Memorial invoked in him, and we believe we are in possession of
the
787
general way in which he overcame these feelings. The monument
not only was a representation of the alienating distance that
Collingwood felt between himself and Scott, it was also the only
piece of evidence that could serve as a bridge by means of which
Collingwood could reconcile himself with Scott by re-enacting
the thoughts which are embodied in the monument itself. Were the
architect a contemporary the question might be addressed to him,
since (short of third party intermediation) the ordinary way to
overcome the space between estranged people is for them to
converse, to put questions to each other, to elicit answers. But
in this case, as in the case of all historical artifacts, direct
dialogue was not possible, and intermediation is necessary.
Collingwood's reflections on the Albert Memorial must have
focused on this acute problem: how does one communicate with the
past? For the living, traces of the past surviving in the
present are the only expressions by means of which historical
conversation can continue, and in this process of interpretation
it is the historical understanding which asks questions, and it
is present evidence--like the mute yet immensely expresive
Albert Memorial--which the only source of assuring ourselves
that our answers and our interpretative efforts have been
successfully carried out. Somehow the gap is bridged: the
memorial speaks, the observer is moved, the dialogue takes place.

788
How is this possible? On what model of thought can such
an event occur? Collingwood was convinced that it was not
possible on the grounds of empirical or realistic premisses (IH,
208-209), nor even by the principles of subjective idealism. It
is not accomplished by using the tools of objective thinking
employed in the natural sciences, nor by disregarding such
methods altogether and adopting an intuitive or subjective
approach, which still assumes the mind-object polarity, but opts
for the primacy of the subjective pole over the objective one
(IH, 124, 292-97). It is accomplished by adapting for history
the usual method employed to cross the gap between persons in
the mind-mind situation, i.e. questions and answers eliciting
unification through shared meanings. If I wish to learn
something about someone it is not achieved by inventing him like
a character in a novel, nor by observing him like a bug on a
pin, nor by experimenting with him like a ball on an inclined
plane, nor even by seeing in him the symbol of the creator.
While all these approaches may be helpful in one way or another,
they tend to make something of the person that he is not--an
imagined entity, an abstraction, a thing, a creature. And the
same is true in learning about oneself--this much we take to be
the final message of Speculum Mentis. If I wish to find out what
a man is, I have to do so by finding out what he thinks, for
this is what he is (IH, 10, 218-20; NL, 1.61); and the ordinary
way to do this is by asking him

789
questions aimed at responses which will help to find out what he
means, where he stands, what he stands for, what he means to do,
what he means for us to do, etc. And if one wishes to find out
what a man was, one has to do so by asking questions aimed at
eliciting responses from existing evidence which will help to
find out what his intentions were, how he perceived his
situation, what he stood for, what he meant to accomplish, etc.
In the latter case one must overcome the distance of historical
time, and the way this is done is by examining and
cross-examining present evidence. The scientific historian is an
intelligent, inquisitive interpreter of evidence.
And just as a person may minimally be said to be acting
rationally not when he can show that his actions follow
deductively from some general principle, nor when he can relate
it to some other act as a reaction, but when he can in-
telligently answer questions about why he did one thing rather
than another, so also one may minimally be said to know some-
thing not when he can show that it is the conclusion of a de-
ductive inference, nor when he can show that it is disting-
uishable from its opposite, but when he can intelligently answer
questions about it (NL, 4.31-4.35, 11.1-11.12, 14.114.37). Q-A
logic, as Collingwood informally schematizes it, is the
linguistic expression of the living process of seeking out and
establishing meaning, exhibiting both phases in the discovery of
meaning--as consciously known to be incomplete (the question)

790
and as having sense-completion (the answer). To ignore the
aspect of the conscious nisus for meaning by presenting answers
as propositions, as if they were only related by formally
explicit logical connectives, is a falsification of the true,
humanly grounded process of inquiry. In this sense Q-A logic is
the true logic of thought, and "replaces" F-logic, which does
not recognize the process of passing from incomplete to
completed meaning. It is only Q-A logic which calls attention to
the active role of the conscious agent in the process of
inquiry, for it is only persons who ask questions and seek
meaningful answers. Q-A logic is the logic of interpretative
inquiry, and as a meaning-seeking function cuts across all
levels of consciousness; it is not bound to the conceptual,
propositional, or inferential levels of thought.
Finally, historical events themselves have a Q-A logical
structure: in the world of practical affairs the paradigm for
sense-location is the historical act of an agent, in which a
questionable situation is resolved by the answering act of a
protagonist aware of alternatives and conscious of the possible
consequences of his acts. To act is to endow a gesture with
meaning and at the same time to express that meaning as an act
(cf. IN, 212). Because such an act is actually the agent's (his
experience) it is potentially everyone's (a shared

791
experience)--that is, it is capable of being re-enacted (IH,
247, 280, 303). That which can be re-enacted is an historical
event, so that the historian, in discovering what it is to think
historically (to re-enact a meaning) realizes at the same time,
and by this act, what it is to be a man. It achieves the
philosophical mandate given by the oracle at Delphi: it is
knowing himself, i.e. knowing what it is to be a man (an enactor
of meaning), to be the kind of man he is (the enactor of meaning
in the context of his own experience) and to be the kind of man
he is and no one else is (the enactor of meaning in this
situation with these perceived alternatives). Historical action
therefore constitutes meaning by bringing it into the world--a
primacy that supercedes the merely theoretical sense of meaning
as something latent in the world (e.g. verbal or lexicographical
meaning). A historian re-enacts this primal activity, and the
coherence of his narrative is built upon the historical meaning
which is both found by the historian in the evidence of the act,
and made by the historian through interpretation of that
evidence.
If history is a kind of understanding in which no pro-
duct of human action is foreign, and if it is posited as the
science of human nature and therefore as a model for reconcil-
iation of all alienation situations in which understanding is
called forth to mediate between self and not-self, it does not
appear to be alone in this effort. Behavioral science makes

792
the same claim, and from his earliest writings Collingwood re-
cognized it to be his natural enemy (RP, 40-42; FR, 75-77; SM,
274-78; A, 92-95). It claimed to deal with the same subject, but
employed a methodology drawn from an empiricistic epistemology
which takes the mind-object polarity as primary. This is a
vestige of the "naturalistic" viewpoint, which he recognized in
historical positivism, and from which he struggled to free both
himself and his readers (cf. EPH, 12-20, 25, 31; IH, 126-33). It
is characteristic of this viewpoint that facts are independent
of anyone's knowledge of them; that objects are active and mind
is plastic and passive; that in the act of knowing the object is
therefore unaffected by the knowing of it; and that individual
acts of knowledge are both atomically distinct from one another
and capable of being analyzed like objects in the perceptible
world. Collingwood denied all of these assertions, and did so on
the strength of the mind-mind paradigm for reconciliational
philosophy: for historical thinking provides an instance of
successful understanding in which the object is another thought.
It exemplifies a case in which, without the activity of thought,
the object would cease to be what it is--instead of being a
metal arrowhead, a Grecian shard, or an Egyptian manuscript,
without historical interpretation it becomes a piece of metal, a
chunk of clay, black marks on flattened papyrus. The past is
therefore so far from being indifferent to its being known that
it would
793
cease to be past at all if it were not for the exercise
of the historical a priori imagination.
But Collingwood did not stop there, and in generalizing
this principle he came to regard all knowing as entailing an act
of interpretation not basically different from that which is
involved in historical thought. He came to realize that meaning
is never something merely found, it is something made by the act
of attention (PA, 213-16; NL, 4.5, 6.1-6.21, 7.2-7.22). Just as
an ambiguous historical situation is resolved by the decisive
act of an agent, the field of sensual flux is stabilized by acts
of attention which domesticate and perpetuate sense, allowing
them to survive and revive in other contexts (PA, 209-10; IN,
303). Meaning is an achievement of personal consciousness, the
means whereby a person appropriates not only his own experience
but that of others. It is also the source not only of all
perceptual and imaginative unity, but of the continuity of all
unified acts of thought (IH, 306).
The deliberate effort to achieve reconciliation by the
act of interpretation aimed at the unity of meaning, therefore,
runs contrary to the effort to define by atomic dissolution into
non-meaning elements, or by subsumption of such elements into
arbitrary sets or classes viewed as collections of
inter-changeable terms indifferent to their generic essence. It
presupposes that the object of knowledge (meaning) cannot be
unaffected by the knowing of it, because it is only human
knowing that
794
seeks and creates meaning in the world through intentional acts,
without which there is nothing definite to be known. It has,
therefore, an opposite orientation to any philosophy which makes
as its fundamental presupposition the assumption that the object
of knowledge is unaffected by, or indifferent to, the knowing of
it. Such a presupposition has been made by some philosophers who
have been known as realists (Cook Wilson, Samuel Alexander, G.
E. Moore, and Bertrand Russell being some examples). Whatever
places the object of consciousness over against it as something
beyond its ability to absorb creatively by the medium of mean-
ing, is the natural enemy of that consciousness: it is the
originator of non-meaning not as that from which meaning arises,
but as that which is forever beyond meaning, unassimilable, the
"thing-in-itself" which is not a thing, not an it, and not a
self. It is the "pure being" which Collingwood rejected as a
subject matter unsuitable for any science, because it is utterly
undifferentiated and lacking in all specification, and hence
indistinguishable from nothing.
And at this point we arrive where we set out from-with
the critique of realistic epistemology. We also find ourselves
at the doorstep of the Essay on Metaphysics.

795
5. The Rehabilitation of Reformed Metaphysics.
We should now be in a position to clear the obstacles to
first philosophy that we presented in Section three. If our
reconstruction of the unity of Collingwood's thought is correct,
it should remain as a solid ring of thought when brought up
against the barriers that are presented in the Essay on
Metaphysics.
But to begin we must ask ourselves why Collingwood un-
dertook to write on metaphysics at all, for as we have seen he
certainly gives indication in the early writings of a typically
British antipathy to the subject. It appears to be just as
accidental that Collingwood should find himself named to the
chair of Waynflete Professor of Metaphysical Philosophy in 1935
as it was that he should have worked for the British Admiralty
Intelligence in World War I and found himself walking past the
Albert Memorial. Are his reflections on metaphysics as much the
result of a coincidence as his reflections on logic? What is
required is not to account for an accidental feature of his
biography but an essential feature of his thought. The situation
in which any philosopher is placed is always to some extent
fortuitous. But as Collingwood notes, it is not so much what
happens to a figure in history but what he does with what
happens to him that creates historical interest. Collingwood
chose to reflect on the A1bert Memorial, and he chose to reflect
on metaphysics,
796
and our question is how that reflection is a product of the
unity of this thought rather than a response to something that
is thrust upon it, as it were, from without.
For someone committed to the primacy of historical
thinking and the mind-mind paradigm in epistemology, what is it
to do metaphysics? Traditionally metaphysics has always been
thought directed at uncovering the stable basis for change--not
only that which "follows the physics" in the literary sense of a
set of treatises attributed to Aristotle, but also that which
seeks to discover the principles which lie beyond the changing
physical world. Very early in his career Collingwood decided,
with the Italian idealists, that change is a realistic concept
and history an idealistic one, since what changes is a material
thing, but that which historically develops is a mind. If a
philosopher with a strong inclination toward epistemology were
to begin, like Kant did, with the assumption that the mind has a
definite structure which determines its active construction or
synthesis of the world of experience, metaphysics would take the
form of discovering what that unchanging structure of mind is.
But Collingwood also decided early in his career that a mind is
what it does, and therefore that there is no such thing as a
fixed "human nature" for all time and all men, so far as mind is
concerned--so this is not a promising direction in which to find
the object of metaphysics. And

797
finally, we have seen that Collingwood ruled out an ontology of
pure being not only because it is arrived at by abstraction from
the mind-object situation, but also because in his study of
cosmology he found that modern physics had abandoned the notion
of substance altogether, replacing it with the concepts of
function, process, and motion.
What is left? For Collingwood what remains is precisely
the fact of interpersonal communication within the field of
mental phenomena. What grounds change within this domain of
reality is whatever provides the basis for reconciliational
identity between minds, or within mind, and that is whatever is
being presupposed. What grounds change in a world of developing
thought is the assumed shared meanings that make diachronic and
synchronic communication possible, and the reestablishment of
significance by a question-and-answer process is the work of
reconciliational philosophy. A metaphysics of rapprochement is
therefore an inquiry into the presuppositions of questions on
the basis of which inquiry proceeds: metaphysics is the science
of absolute presuppositions.
(a) From Anti-Metaphysics to Reformed Metaphysics.--
Collingwood makes it easy for his readers to grasp the negative
element in his critique of anti-metaphysics (itself a negation
of a negation), but difficult to get beyond that to what the
positive basis is for this critique. It is therefore quite
possible to overlook the fact that

798
his elenchus may be an utter failure without touching the
positive thought from which it springs, and which it ultimately
disguises. From the perspective of rapprochement philosophy we
may now be able to see past the disguise.
Collingwood examined in detail two forms of anti-
metaphysics: psychologism and logical positivism.7 His objection
to psychologism is basically that as "materialistic
epistemology" it treats all thought as if it were a datum-i.e.
as if it were a feeling or an aggregate of feelings, and
consequently as non-criteriological (EM, 109, 114-15); and his
objection to positivism is that it refuses to recognize that
______________________
7 Collingwood distinguished three forms of anti-meta-
physics, which he called progressive, reactionary, and irra-
tionalist (EM, 82-84). If we are correct in assuming that
positivism is the progressive form of anti-metaphysics, and
psychology is the irrationalist form, what is the reactionary
form? Probably the sort of philosophy represented by Samuel
Alexander's realism, examined in Chapter XVII of the Essay on
Metaphysics, since he accuses Alexander of being "influenced by
the quaint, characteristic eighteenth-century dogma" that there
is an underlying or pervasive character of everything which
exists-- expressible as the law of universal causation (EM, 175,
179). This is the way he described reactionary anti-metaphysics
(EM, 93-94). But both of these sections are excessively vague,
full of broad historical generalizations, and fail to satisfy
the reader's desire for a clear presentation of the viewpoint
being opposed. A perfect candidate for reactionary metaphysics
would be twentieth-century Thomism (cf. EM, 91, where the
reactionary anti-metaphysician "embraces x as his own
'doctrine,' claims x was 'right,' and professes himself an
'Xist"'). But astonishingly, Collingwood does not draw a bead on
this elephantine target, but instead takes potshots at Watt's
steam engine and Locke's political views in the 19th century.
799
thinking rests on presuppositions, and assumes that all gener-
alization is based on observation of matters of fact (EM, 146-
47). What are the positive equivalents to these critical re-
marks? In the matter of thought, it is clear that Collingwood
wishes to maintain that all thought is "criteriological," mean-
ing that in every act of thought it is necessary that the
thinker himself should judge its success or failure--a self--
critical judgment that Collingwood distinguishes from the (real
but irrelevant) possibility that one man's thought may be judged
successful or unsuccessful by someone else (EM, 109, 115). In
the matter of presuppositions, it is clear that Collingwood
wishes to maintain that there are such thoughts, and that they
are to be distinguished from empirical generalizations which are
verifiable or falsifiable by induction from observable facts
(EM, 147). Our difficulty arose from trying to conjoin these two
ideas: for if presuppositions are neither true nor false, they
do not appear to be able to retain their criteriological
aspect--i.e. the use of "a criterion, the double notion of truth
or falsehood, by reference to which he judges a thought" (EM,
115); and if all thought is based on presuppositions, then
thinking is based on that which is non-criteriological, i.e.
that which is not thought.
The first step toward resolving this difficulty is one
which we have already suggested: to expand the notion of truth
so that it is not bound by the propositional criterion, i.e. by
that which is verifiably

800
true in answer to a question. Now in Chapters VII and VIII we
took considerable effort to demonstrate that Collingwood
recognized levels of mental functioning that could definitely be
called thought, and were not yet at the propositional level of
truth. While the early Collingwood recognized that questions
themselves were non-assertions, he failed to distinguish various
kinds and degrees of non-assertive thought--a deficiency that
was remedied in his later writings, where concepts are described
as the entertainment of meanings in imagination, having the
status of thoughts which are non-assertive and pre-propositional
in the logical sense. One function of conceptual imagination is
precisely to consider a meaning in isolation from, or
indifferent to, any reference to anything else (i.e. any
predication): 8 imagination is essentially the
_____________________
8 To consider a meaning in isolation from, or indiffer-
ent to, any reference to anything else is to consider it ab-
stractly. Since all thinking is based on pre-suppositions, does
this mean that Collingwood had come to believe that all thinking
is conceptual and hence abstract? Was Donagan right after all?
We think there are several decisive reasons for believing that
this is not the case. (1) Collingwood never denied that it is
possible to think abstractly: he merely argued that such
abstract thinking cannot bear the weight of unconditional truth
that we wish to place on it (SM, 252-53; NL, 26.18). (2) He also
argued that it is only a phase in our thinking (like the bud is
an abstraction of the flower), and that in thinking abstractly
we are not condemned to remain at this point--or as he says, to
live in Ezekial's Ezekiel’s Valley of Dry Bones (NL, 7.65 7.64).
Our release from such confinement occurs in the form of further
acts of propositional thinking, where a proposition is an answer
to a question, and a question offers alternatives (NL, 11.22,
11.34-11.35). (3) Abstraction always occurs in what Collingwood
calls a "context of evocations," over which preside acts of
evocative thinking governed
801
"suspension of the activity of asserting" (SM, 78), and in this
activity it "is indifferent to the distinction between the real
and the unreal" (PA, 136). Yet there is a kind of truth that is
involved in the act of imagination, and this pertains to its
expressiveness--the very function which relates imagination in
an essential way to art. The truth of art is not a truth of
relation but a truth of fact: its truths are concrete individual
experiences, and the truth of art is the truth of its
expressiveness of these experiences (PA, 288). Its opposite--its
untruth--is the "corruption of consciousness" which disowns a
feeling by expressing it misleadingly or self-deceptively (PA,
219-20). The function of imagination is the perpetuation and
domestication of meaning as the ground or material out of which
further acts of consciousness will shape questions,
propositions, inferences, and other expresions of higher-order
thought.
Here we have a relation of presupposition which fits the
description Collingwood has for it in the Essay on Metaphysics.
It is thought, but in the form of imagination; it is
criteriological in the sense that it has its own standard of
success or failure, as the true expression of a felt meaning,9
but not in the sense of true or false
________________________
by logical relations (NL, 7.32-7.39). As we have pointed out in
previous chapters, this is a very different description of ab-
straction than the one he had condemned as falsification in
Speculum Mentis.
9 It should be recalled that Collingwood found nothing

absurd about there being such things as "emotions of intellect"
802
propositions; and it is a function prior to, and the logical
ground of, further acts of higher-order thinking--e.g. the
verificational acts involved in scientific inquiry, within which
absolute presuppositions are embedded as primitive survivals.
Absolute presuppositions are a priori, concrete meaning-concepts
of imagination, not unlike the Kantian categories.l0 This
conclusion evidently relies on the theory of
______________________
(PA, 267), and that there would therefore be nothing absurd
about there being an "emotional charge" on a presupposition. In
fact, Collingwood's remarks about being "ticklish in one's
presuppositions” (EM, 44), as well as about the numinous terror
that may surround them (EM, 46) confirms this observation. What
would be interesting is what Collingwood might have had to say
on the manner in which questions arise due to the causal
efficacy of absolute presuppositions: has this anything to do
with the emotions of intellect?
10 On the interpretation of absolute presuppositions as

a priori concepts there is considerable agreement by Collingwood
s commentators: cf. Mink, MHD, 144-48; Rubinoff, CRM, 234-35;
and Toulmin, CEPC, 205-08. With respect to their similarity to
the Kantian categories, we suspect that not all would
agree--particularly Rubinoff, who adopts a basically Hegelian
interpretation of Collingwood. Our own reading is that
Collingwood had more Kantian leanings: in spite of such passages
as that in The New Leviathan, where he denies the "German"
belief that it is possible to compile a list of the logical
relations which govern conceptualization (NL, 7.34), what he is
describing in that process is the Kantian schematism of the
categories. His rejection of any categorical compilation as
consistent with the passages in the Essay on Metaphysics, where
Kant's transcendental analytic is treated as an attempt to state
what the presuppositions of physics were in Kant's own time (EM,
231-81). Collingwood is not denying that there are such
relations (as even his remarks about the "element types" of
logical thought, in his correspondence with Ryle, would
indicate), but only that Kant's list is not definitive,
exhaustive, or independent of the state of thinking on physics
at that time. On absolute presuppositions as a priori concepts,
see also David Rynin, "Donagan on Collingwood: Absolute
Presuppositions, Truth, and Metaphysics," The Review of
Metaphysics, XVIII (December,
803
meaning that we have extracted from Collingwood's writings only
in bits and pieces, an incomplete but pervasive view of
inter-relating acts of consciousness and linguistic expressions,
some of which we explored in Chapter VIII. At this point we only
claim to have indicated enough of this theory to show that the
relationship of presupposing is one that is not unheralded in
Collingwood's philosophy of mind, nor does it defy the demand
that thought be regarded as "criteriological"--so long as the
definition of "criteriology" be wide enough to include the
double aspect of truth as the truthful expression of meaning and
falsity as the corruption of consciousness.
But if we can congratulate ourselves for having cleared
the main difficulty of our fourth obstacle,ll can we say that we
have resolved all our problems about anti-metaphysics? Not
quite. The psychologizing in which Collingwood engages in this
portion of the Essay on Metaphysics remains for us an unresolved
embarrassment, but one which should not
_______________________
1964), 301-33. It is interesting that Rynin, who claimed to be
the last living logical positivist (ibid., 331), would undertake
a defense of Collingwood's view of metaphysics as a science of
absolute presuppositions, and to mount a scalding attack on
Donagan for treating Collingwood as "a very third rate thinker,
struggling without success under a burden of ineptitude that
would be a source of concern in a not very promising
undergraduate" (ibid., 332). Rynin's own estimate of Collingwood
is considerably higher.
1l We are approaching the obstacles from Section 3 in

reverse order, for reasons that will become obvious as we con-
tinue: the main reason is that we wish to reserve the question
of ontology until last.
804
obstruct our passage toward the positive aspects of his reformed
metaphysics. As universal as were Collingwood's interests, we
cannot expect him to have accurately evaluated every major sci-
entific development of the last century, and his failure to
appreciate the true worth and benefits of behavioral sciences
should not deter us from recognizing in his own work an equally
important observation: that historical thinking can contribute
an essential dimension to the self-understanding of man. From
this point of view his warnings to us about the pretensions to
wisdom of some who look to behavioral science as a "guide to
life" are only a counsel to beware of such a one-sided prejudice
about the nature of thought. It is a warning that he first
issued in Religion and Philosophy and is repeated in no less
than six of his other mature works. And while Collingwood seems
less concerned about the ill effects of logical positivism (he
did not seem to be convinced that they would be taken with much
seriousness by the scientific community), the danger here is
that in attacking metaphysics the positivists would unwittingly
contribute to the attack on reason in the contemporary world,
because an attack on the doctrine that thought has
presuppositions is an attack on the foundations of science, and
an attack on science is an attack on rationality and therefore
on civilization. Collingwood's fear is that we would fall victim
to a kind of belief about our own abilities to think that would
paralyze us by convincing us that we are not able to think for
ourselves, to judge if the products of our

805
own thought are well or ill done, and to behave in general as if
our thinking would not make any difference to what we are trying
to achieve. A belief about our own minds is one that is
reflected in the way we behave, and therefore if we believe that
we are a conglomerate of feelings, we shall surrender the
imperative to know ourselves by giving up the effort that
consciousness must put forward to achieve continuity in
experience. In Collingwood's estimation this is tantamount to
giving up on rationality and on civilization altogether.
Considering what was happening in Europe at the time he was
writing, he did not care if he was regarded as a cackling goose.
(b) Metaphysics as an Historical Science.--If we can say
that an absolute presupposition as an a priori concept of
imagination is a kind of thought (but not propositional
thought), we can also say that it is minimally eligible for
being a subject matter for historical re-enactment. But there is
much that we still do not know about it. We are not yet in a
position to overcome our third obstacle, the unlikely match of a
science traditionally oriented towards universal and timeless
truths, and a kind of research concerned only with particular
and spatio-temporally localized matters of fact. This is the
central paradox of the Essay on Metaphysics with its thesis that
metaphysics is an historical science. In fact it appears to be
doubly paradoxical, because it is puzzling when approached

806
from either direction. For if the methods of history are
designed to "get at the facts" by the interpretation of
observable evidence, why should a science of absolute
presuppositions, for which there can be no directly observable
evidence (not being propositional, and therefore neither
verifiable or falsifiable), be interested in adopting the
methods of history? What sort of a fact is an a priori concept?
And contrariwise, if historical truth is always ad hoc and is
therefore only true for a particular time and place, of what
interest is it to an historian to deal with matters that are a
priori and conceptual? What light is thrown on the relationship
of presupposing by saying that it is brought into awareness by a
kind of analysis which is drawn from the science of history?
To begin with we must first find out why Collingwood
believes that absolute presuppositions constitute an appropriate
subject matter for historical thought. Then we must approach the
other side of the question by asking why the presuppositions,
questions, propositions, problems and methods of metaphysics are
the same as those of history. Now we have already suggested that
absolute presuppositions are thoughts, but on the criteria
offered in The Idea of History, this is not enough. In Chapter
IX we found Collingwood telling us that to be a proper subject
matter for history, an act must not only be an act of thought
but of reflective
807
thought, where reflective thought means an act done in the
consciousness that it is being done (IH, 307-08). While it is
possible for a concept to be the object of an act that is
"conscious of. . . ," what is required for an act that is
"conscious that . . .” is a proposition. But then does this not
imply that to be an object of history an absolute presupposition
must be a propositional act of thought, and therefore that it
cannot be simply a concept? And does it not further imply that
as an act of reflective thought an absolute presupposition
cannot be unconsciously presupposed in any sense that is his-
torically relevant?
These are difficult questions, and have wrecked the
efforts of more than one of Collingwood's interpreters to sal-
vage sense from the Essay on Metaphysics. But if we are to test
the strength of our reconstruction of Collingwood's central
"ring of thought," we must risk the same fate. In that
reconstruction we noted that at the center of the object of
historical interest is the resolution of an indeterminate
situation by the act of an historical agent. This is an act made
consciously, in the face of perceived alternatives, and having
consequences beyond the act itself--consequences which the act
brings about, but not with apodictic necessity. One phase of
Collingwood's reform of metaphysics is carried out by drawing
out the implications from the insight that the same sort of
process is operative in the acts of thought

808
that bring science into being. In spite of what positivists and
other formalists may believe, a science like physics does not
spring like Athena fully armed from the head of Zeus; it grows
in human fashion by the deliberate process of inquiry, the
processes he informally described under the title of Q-A logic.
It grows by putting questions in an orderly sequence and
marshalling intelligence at all levels of consciousness in the
effort to answer them. The sequence of thoughts in a scientific
inquiry therefore parallels the sequence of thoughts which con-
stitute the significant "inside" of historical events. And if
one wishes to understand a scientific event, one must re-enact
the sequence of thoughts which brought the scientist to draw the
conclusions he did, just as one understands an historical event
by re-enacting the thought processes involved in it.
Such understanding is, like all thinking, an interpre-
tative process, and requires that the re-enactor be adept at
forging questions which can overcome the discontinuities that he
encounters. As an historian Collingwood was aware that there are
many of these--that scientific thinking is a discontinuous
process with interruptions, detours, inconsistencies, even
revolutions; and that the manner in which these are reconciled
is not the same for different subjects, different people, even
different places and times. Therefore while the way rapprochement
occurs is by achieving common understanding of shared meaning,
Collingwood's
809
historical experience convinced him that true meaning is always
individuated and concrete, and the way in which shared meaning
is achieved is not the same for all inquiry situations. The kind
of answer that will satisfy a physicist is not the sort of
answer that will satisfy an artist or theologian; and the sort
of answer that would satisfy a classical Greek physicist would
not be the same as an answer that would satisfy his medieval,
renaissance, or modern counterpart--not because the phenomena
have changed, but because the expectations of what the phenomena
can mean have shifted, and this represents a change in the
heuristic presuppositions which ground the questions. This is no
less true at the personal level, for even within the con-
sciousness of a single individual the questions asked at one
point in a person's lifetime will have a different basis than
those asked after his thought has developed beyond that point.
Therefore one cannot begin by assuming that there is an
unchanging, generic sense of "knowing," and an equally univocal
meaning for the answers, questions, and presuppositions of all
inquiry per se. An inquiry about inquiry is an ad hoc affair,
and one should not expect to find the common and unchanging
ground of all possible inquiries, semper, ubique, ad omnibus,
for this would be the epistemological equivalent of a search for
an unchanging substance. Like all efforts of rapprochement, the
way in which continuity of meaning is established in a given
context is by
810
showing how thought develops in that context, i.e. by showing
how particular meanings become progressively more inclusive, how
the variable (knowing in this particular situation) modifies the
generic essence (inquiry) and becomes identified with it.
Rapprochement is complete when that development can be
demonstrated as a scale of forms. In the case of science such a
scale takes the form of a conceptual system, i.e. a nexus of
meanings produced by the exercise of an act of real (or
"concrete) abstraction by which a concept is located in a
context of other concepts.
But while these comments may help us to understand how
an absolute presupposition can be cogently interpreted as an a
priori concept compatible with Collingwood's philosophy of mind,
we do not seem to be much closer to an understanding of how it
can be both a priori and factual, conceptual and reflective. For
this we need to turn to the other aspect of the process of
re-enactment, the thought of the historian. For just as in the
process of discourse by which reconciliation is achieved between
minds there are moments of listening but also moments of
questioning, so also in historical thinking there is a kind of
discourse in which the historian reaches beyond being a passive
listener and takes charge of his subject matter by
cross-examining it. In re-enacting the sequence of thoughts
which form the body of some piece of scientific inquiry the
historian may well ask questions that the scientists never
asked. He may thus ask why the initial scientific question

811
arose--a questions which (if the historian does not want to
distort the nature of thinking itself) is directed at the
objective or logical (rather than subjective or psychological)
basis from which the question arose, i.e. its relation to
meanings already established and unquestioned. It seeks to
uncover what the scientist is assuming that allowed him to ask
the question he did. In pressing this process or
cross-examination to its conclusion the historian reaches a
point where he recognizes that all the questions of a given
systematic inquiry on a particular occasion presuppose a meaning
or set of meanings not given as a result of any prior and
similar questions in the inquiry, but forming the necessary
condition for all questions to arise within that inquiry. But in
stating what those absolute presuppositions are, the re-enactor
of the inquiry sequence is doing something that the original
agent, the scientist, may never have done, i.e. to express an a
priori concept in the form of a statement.
Only someone who expects history to be a scissors-and-
paste affair, a mere reportage, would be put off by this ex-
pansion of the historian's role. For Collingwood it was of the
essence of scientific history that it be autonomous, which means
that the active processes of the historian's thought are always
involved. Such is the active nature of historical reenactment
that it is not bound by the requirement that only

812
what occurred in an historical sequence is what can re-appear in
the continuity of past events, of reconciling himself or his
thought to the thinking of the scientist, he may put questions
that may never have occurred to his protagonist, and he does so
on his own initiative, because he is the only one who can know
what he expects to find out, or when he has achieved that
rapprochement that his narrative seeks to achieve. But in this
autonomous placement of questions we begin to get the sense of
how Collingwood sees a way to overcome what appears to be a
fundamental and irreconcilable difference between two
traditional ways to approach factuality--the diachronic narra-
tive approach of the historian and the synchronic logical ap-
proach of the metaphysician. If the scientific inquirer had to
think some thoughts (absolute presuppositions) in order to think
others (the questions and answers that form the body of his
inquiry), and if the first thoughts are concepts embedded in the
meaning of all the questions of the inquiry itself, then they
can be said to be thoughts which occurred simultaneously
(synchronically) with that inquiry. But this does not mean that
they suddenly drop out of historical interest and become the
province of the metaphysician-as-logician. It is what is
involved in understanding the scientific event itself. In
re-enacting that event historical thinking brings a latent
thought (an a priori concept) into explicit consciousness (an
act of propositional thought) in the form of an absolute pre-
supposition, something that is therefore

813
both found and made by the act of historical understanding.
It is in this sense that absolute presuppositions are
factual, and can be both the result of a question (the active
question of the historian rather than those he receives from the
historian in his inquiry) while yet remaining, as definitions of
a priori concepts, pre-varificationally truth-neutral. Where
such thoughts do not necessarily form a temporal sequence, the
historian puts them into such a sequence (diachronically)
through his narrative reasoning. He is expressing a ground
consequent relationship in narrative form. Nevertheless this is
not mere invention on his part, on pain of ceasing to think
historically: it must be rooted in evidence, such evidence
being, of course, the documentary materials of the scientific
inquiry articulated in language--with all that that implies.
We cannot claim to have accounted for all the problems
that we have raised concerning the historical re-enactment of
absolute presuppositions. Some of these issues shall be dealt
with further in subsequent sections; others are obscured by
shadows we have never succeeded in penetrating with light--the
nature of historical inference, the relationship between Q-A
logic and formal structures in language and thought; and still
others seem both to retain striking echoes from his early re-
conciliational philosophy, and to suggest further thought which
never had the opportunity to take place. What we are left with
is a very general argument for reconciling history

814
and metaphysics--an argument that shows how it is possible to
employ the methods (the interpretation of evidence by the active
and systematic process of questioning), the presuppositions
(that thought is capable of being re-enacted due to the
invariance of meaning in different contexts), and the proposi-
tions (the ad hoc nature of inquiry about inquiry, the use of
evidence, etc.) of history to do the work of metaphysical an-
alysis (bringing to consciousness, or expressing in statements, a
priori concepts, in the form of absolute presuppositions,
incorporated in acts of scientific inquiry).
But this sort of argument employs rapprochement only at
the level that it had reached in the earliest phases of
Collingwood's development: the identification of metaphysics and
history is an abstract rapprochement identity of the sort that
we originally encountered in Religion and Philosophy. Obviously
not all of history is metaphysics, because not all of history is
concerned with the absolute presuppositions of scientific
inquiry; and it would appear that not all of metaphysics is
history, because the logical analysis in which a given
constellation of metaphysical presuppositions is shown to
display incompatibilities is not part of the usual business of
construction of coherent narrative, or at least not to the
extent that it is not a thing of the past. The next step in the
reconciliational process would be to show the way in which
history
815
and metaphysics are defined by a scale of forms. In The Essay on
Metaphysics we find not the slightest gesture in this direction.
(c) The Absoluteness of Presupposing.--So far we have
argued that by making use of our reconstructed "ring of thought"
in our approach to the major obstacles to Collingwood's reformed
metaphysics, it is possible to make limited sense of The Essay
on Metaphysics. We may summarize this argument thus far as
follows. (1) As an application of rapprochement philosophy,
metaphysics seeks to overcome failures in the continuity of
scientific inquiry by establishing the unified ground of shared
meaning upon which systematic inquiry is based. (2) The way this
is done is by applying the interpretative methodology of Q-A
logic to concrete examples of scientific thinking, not in the
forward direction of the inquiry itself, but in the
retrogressive direction--i.e. towards the presuppositions of the
inquiry's questions. (3) When this retrogressive inquiry is
pushed to the limit, what is uncovered are the a priori concepts
which form the logically connected system (or constellation) on
the basis of which all the questions of the inquiry are possible
(i.e. make unified sense), and from which they all arise. (4)
Metaphysical reconciliation is complete when the system of such
concepts is made explicit (or is brought into consciousness) and
articulated in the form of a conceptual framework (a
constellation of absolute presuppositions) having the structure
of a philosophical scale of forms. (5) Such a system

816
of concepts linked by intentionally more inclusive meanings
becomes expressed by acts of interpretative thought which bring
absolutely presupposed concepts into consciousness in the form
of definitions. (6) Such interpretative acts are historical
insofar as history is the sequence of past acts of reflective
thought, reenacted in the context of present thought by the
intelligent cross-examination of evidence--the evidence in this
case being concrete instances of scientific inquiry. (7) In such
a re-enactment historical thinking affects its objects (a priori
concepts) by locating them in a context of other concepts, i.e.
by putting them in propositional form.
We are thus now in a position to understand how someone
could mistake an absolute presupposition for a proposition.
While an absolute presupposition is an act of a priori
imagination at the conceptual level of thought, it is expressed
in a form that appears like a proposition: it is expressed in a
sentence, having the apparent unity of a judgment, and therefore
posing as a candidate for that level of consciousness. But it is
only in the retrogressive, metaphysical inquiry that it takes
such a form, and we now understand that it is actually put there
by the active process of historical thinking, and as a result of
deliberate interpretative questioning.

817
But this glosses over difficulties to which we must now
attend, not the least of which is the problem of the distinction
between absolute and relative presuppositions. We recall that
this distinction is made on the basis of the fact that there are
some presuppositions which are not themselves the answer to any
higher-order question in the progressive systematic inquiry. On
our present line of thinking this is no longer a tenable
distinction; it must be the answer to a question in some inquiry
or it cannot be said to be a thought, or at least not a
reflective thought (which it must be to be of interest to the
historian)--i.e. one made in the consciousness that it is being
made. And if it is the answer to a question in the retrogressive
metaphysical inquiry, how are we to understand the way in which
such a question arises? For to say that it "arises," on
Collingwood's view of the matter, is to say that it has a
logical connection with our previous thought and does not arise
by capricious curiosity. But that would mean that it arises due
to the causal efficacy of another presupposition, in which case
it is no longer absolute but relative.
How can our "ring of thought" get us out of this maze?
Surely the very concept of rapprochement presupposes that there
is something to be reconciled, i.e. two thoughts which do not
share the continuity typical of self-consciousenss. The kind of
consciousness employed to achieve the

818
first stage of reconciliation is questioning--that expression of
the acknowledged incompleteness of an act of propositional
consciousness which anticipates its requisite completion in ar,
act of assertion. Where the discontinuity occurs in a systematic
scientific inquiry, which represents the articulation of a
single question or set of related questions, it indicates a
failure to grasp a prior step in the inquiry, which leads to a
regressive type of question and reverses the direction of the
primary inquiry. Such a reversal and its resultant questions are
the result of a higher stage of consciousness which Collingwood
described as "reason" in The New Leviathan, namely "thinking one
thing, x, because you think another thing, y; where ~ is your
'reason' or, as it is sometimes called, your 'ground' for
thinking x" (NL, 14.1). It is a kind of thinking that arises as
a "practical act of trying to alleviate the distress caused me
by the untrustworthiness of my knowledge" (NL, 14.3), and never
loses this practical aspect. In the case of a scientific inquiry
the "distress" is due to the discontinuity in the inquiry, and
the historical, retrogressive question is an exercise of the
rationality of the historian, the absolute presupposition being
the "ground" or "reason" which will reassure him of the
trustworthiness of a certain line of questioning.
Now such a reflective reversal may be part of the in-
quiry process itself (regression to an hypothesis--e.g. one
which requires further definition) or it may be one which arises
when the entire inquiry itself

819
is called into question. Once again in the latter case this may
occur in the process of normal scientific inquiry (as in the
early stages of scientific revolution), or it may occur due to
failure in the process of trying to re-enact a piece of science.
In the latter case the historian may actually invent the
question for the purpose of eliciting from the inquiry its
presupposed concepts, or its conceptual framework. But these two
cases seem to be very different: if the historian, in exercising
his autonomy in raising questions on his own initiative, is free
to invent a metaphysical question, it is clear he is creating
the discontinuity that the absolute presupposition is meant to
repair; but the scientist in a period of scientific crisis
perceives or feels the discontinuity in the experience of his
own scientific consciousness, and therefore does not create it
but finds it. Once again we find ourselves confronting the
idealistic dimension of Collingwood's philosophy, and uneasy
about what to make of it. If we define the "idealistic turn" as
that point at which a philosopher decides that meaning is
something not found but made, then there can be little doubt
that Collingwood made that turn early in his career, and never
went back on it. But we have repeatedly witnessed Collingwood
arguing that meaning is something both found and made: this is
the celebrated "unity of act and fact" of Gentile and the
Italian idealists, and the basis of Collingwood's later
re-interpretation of the act-object distinction.

820
The distinction between relative and absolute presup-
positions is our present case in point, and we wish to under-
stand if an absolute presupposition is something found or made
or both. Clearly Collingwood wants to say both, but can he do so
consistently? For if it is both, then the absoluteness of
presuppositions threatens to vanish altogether: as something
found its reality is affirmed, but only as contingent to an act
of historical interpretation which establishes its factuality;
and as something made its independence is surrendered to that
which is responsible for its creation. In either case it turns
out that an absolute presupposition is something that is always
dependent on something else, and is therefore only relatively
absolute--a paradoxical conclusion to be sure.
The relativization of absolute presuppositions also
bears down on us form a different direction, for we are forced
to recognize that in calling absolute presuppositions a priori
concepts we have vastly oversimplified the logic of presuppos-
ing. The deciding factor in considering whether a thought is a
concept, a proposition, or an inference is not the form that one
or the other takes when it is expressed in words, but the way it
functions in the process of thinking. Thus one and the same
linguistic entity--the indicative sentence--may express a
concept in the form of a definition, or may predicate one
concept of another in the form of a proposition, or may re late
one concept to another as ground and

821
consequent in the form of an inference. While we have argued
that a presupposed concept may appear in indicative sentential
form because we typically express ourselves in sentences in
order to communicate a complete thought, this does not tell the
whole story. For in much of what Collingwood says about it,
presupposing is treated as a complex relationship rather than as
a simple linguistic or formally logical entity. He would say,
for example, that one form of experience presupposes another (as
science presupposes religion in Speculum Mentis, where "science"
and "religion" are concepts representing whole regions of
experience)--indicating that the relationship was one of logical
dependency but not of deductive entailment. Such a usage is
virtually repeated when he writes in the Essay on Metaphysics
that contemporary science is "monotheistic" in that it presup-
poses that God exists. And again in the "retrospect" chapter of
The New Leviathan he writes that each of the levels of con-
sciousness presupposes its predecessor as that out of which it
develops, while the successor is not rendered necessary by the
form of consciousness which proceeded it (NL, 9.43). In fact it
is this section of The New Leviathan that comes closest to
defining the relationship of presupposing than any other place
in his published works, since it is set in the context of a
description of development which he defines as "a logical pro-
cess in which B 'presupposes' A, C 'presupposes' B. and D 'pre-
supposes' C" without the earlier terms necessitating the later
ones (NL, 9.47-9.48).

822
This clarifies the relationship at the same time that it
complicates and relativizes it. For if the scale of forms of
consciousness is such that any point on the scale not only
summarizes and completes the terms below it but also presupposes
them, then in what sense can we say that absolute pre-
suppositions are confined to the conceptual level (or should we
say "located" at the conceptual level?) of consciousness? We are
at a loss to say what the criterion is for the absoluteness of
presuppositions. While it would appear that the minimum
determination of meaning is at the conceptual level of thought,
that would seem to leave us with as many absolute
presuppositions as there are conceptual meanings, a situation
which achieves ultimacy at the expense of complete dissipation.
What we seem to be left with is a glimpse into an
epistemological concept of remarkable complexity--one which has
yet to be dealt with in a satisfactory manner by anyone with
whom this author is familiar.l2 We know that Collingwood un-
derstood an absolute presupposition to be pre-propositional and
also pre-interrogative (both in the logical rather than temporal
sense), and that it has efficacy in causing questions
______________________
12 However the issue of conceptual change in constel-
lations of presuppositions has been recognized and discussed
with considerable Insight by Stephen Toulmin, who acknowledges
his debt to Collingwood on this subject. See his Human
Understanding: The Collective Use and Evolution of Concepts
(Princeton, 1972), pp. 52-85.
823
to arise--these being epistemic functions. We also know that it
has certain logical properties which tend to apply to pre-
suppositions regardless of their semantic content--that a pre-
supposition acts as a unifier in a systematic inquiry; that its
relationship to questions and propositions resembles inference,
but yet is not deductively or inductively related to questions
and answers in the P-Q-A complex; that it has relations of
consupponibility with other presuppostions, and with these
others forms a constellation or synthesis which, like a
deductive system, is subject to conflicts or strains due to
inconsistencies. But yet it has some of the properties of
religious belief: it has the character of a decision with
practical consequences; it is held with an attitude of un-
questioning acceptance or natural piety; it can be embraced or
abandoned, but never verified or falsified; and it is sometimes
surrounded by a kind of "numinous terror" ordinarily evoked by
sacred objects. And finally Collingwood called attention to its
historical dimension: the establishment of its factuality by
cross-examination and interpretation of evidence, its
development to full consciousness by rational acts of
metaphysical analysis which shares the premise of historical
understanding that acts of thought are capable of reenactment
without essential loss of universal meaning.
Even a cursory examination of this set of properties
could well keep us busy for several more chapters, raising as
they do questions which

824
touch not only on epistemology, but also on logic and the
philosophy of religion. At this point we can only admit that it
does not seem possible to maintain the unconditional distinction
between absolute and relative presuppositions, where that
distinction pertains to all possible inquiries. but in any given
inquiry such a distinction is not only possible, it is
essential: for without the presuppositions that ground the
entire set of questions, no scientific inquiry can achieve its
objectives--i.e. to fulfill the intention to answer the
questions that arise only due to the causal efficacy of these
presuppositions, and ultimately to be grounded in these
presuppositions as their rational justification. It is not a
complete disaster for rapprochement philosophy, therefore, to
argue that absolute presuppositions are inquiry-dependent
meanings and therefore only relatively absolute. They are at
least inquiry-constant within that inquiry, and this suffices
for Collingwood's purposes, since it is consistent with the ad
hoc character of all historical interpretation.
Our disappointment at Collingwood's failure to satisfy
us on the nature of the relationship of presupposing should not
prevent us from acknowledging our debt to him for what he did
achieve in this area. In marking out the realm of absolute
presuppositions he pointed the way beyond the empiricistic and
positivistic attempts to eliminate general

825
metaphysics. Like Caesar, he crossed the river even if he did
not totally succeed in subduing the territory on the other side.
He showed us a way in which a science of metaphysics is
possible, without actually accomplishing the construction of
that science. Others who have followed him--men like Errol
Harris and Stephen Toulmin--have since taken the banner to
higher ground. It is still a territory left largely unexplored.
(d) Ontology and Reformed Metaphysics.--We come now to
the final obstacle, Collingwood's rejection of ontology as a
fitting subject-matter for metaphysical analysis. Of all our
obstacles, this is the most difficult for us to surmount, not
only because Collingwood was so perfunctory and uncompromising
about it, and not only because it represents an unhistorical and
unscholarly dismissal of a subject matter of great antiquity and
coeval with philosophy itself, but also because it runs contrary
to attitudes that he espoused from an early date, and defended
for many years. For while the rejection of a science of abstract
being is a recurrent theme, so also is the requirement that the
object of philosophy is no mere ens rationis, but one which has
ontological reference (EPM, 125); "philosophy is the theory of
existence; not of existence in the abstract, but of existence in
the concrete; the theory of all that exists" (RP, 16; FR, 54).
Such is also the significance of Anselm's ontological argument,
which is defended in works from all stages in his philosophical
career (RP, 66; EPM, 124-25; EM, 189-90), and such is

826
the essential nature of all philosophical judgment: it is always
both universal and categorical, stating what is both common and
essential in that which exists (EPM, 127, 136). To abandon this
requirement is not merely to reject one erroneous form of
metaphysics, it is to strip philosophy of its distinguishing
feature. To fail to overcome this obstacle is therefore to fail
to defend the role of philosophy as Collingwood had defined it
in the Essay on Philosophical Method, which the Autobiography
singled out as his best and "only" book.
Is there any way that our reconstructed "ring of
thought" can aid us in adopting a viewpoint that will allow us
to see our way beyond this final obstacle? After providing us
with a means to understand how metaphysics is possible as a
science of absolute presuppositions re-enacted by a form of
thought using the methods of historical interpretation, can we
now find a way to re-establish the continuity between reformed
and traditional metaphysics, between scientific history and
Aristotle's metaphysics? For rapprochement philosophy, what is
the meaning shared by a science of absolute presuppositions and
a science of being as being? To reconcile ontology and reformed
historical metaphysics would require redefining ontology in a
way which would allow for a non-abstract concept of being, one
which does not arise at the limit of the abstractive process by
leaving out determining characteristics, but

827
rather takes such individuating marks as essential to it. This
would be a science of "concrete being"--a reformed ontology that
would be a part of reformed metaphysics.
Collingwood himself leaves open the possibility that
rejecting one sense of "ontology" leaves other senses of the
term untouched by the criticism (EM, 16). But our concern is
less with his rejection of the title than it is with his
rejection of the concept which the title represents. For in
spite of his claims about what philosophy should be, in most of
what he writes he refuses to entertain any questions concerning
ontological reference in the sense of a science which says
something directly about reality rather than something about
knowledge of that reality. In Speculum Mentis this refusal takes
the form of a denial that metaphysics can ever analyze real
being, being as it is in itself untainted by thought (SM, 274);
and in the theory of perception offered in The New Leviathan it
is expressed in the denial that feelings have objects as well as
modes (NL, 5.39). In such places as these it has been hard for
us to defend Collingwood against the charge of radical
subjective idealism. But in other passages in his works it is
possible to see a different thought process at work, a process
which allows him (for example in his correspondence with Ryle)
to recognize the metaphysical legitimacy of such statements as
"mind exists," "matter exists," and "God exists"--statements
which are both categorical(or referential) and

828
universal. Can we leap to the conclusion that such statements of
ontological commitment form the body of absolute presuppositions
that we have found so elusive throughout this chapter? They
certainly appear to be ultimate and primitive enough to present
themselves as candidates for inclusion in reformed metaphysics,
and the only sort of question to which they could be the answer
would be equally primitive, viz. what exists? But then why is it
not legitimate for reason to press this retrogressive inquiry
one step further by asking for reassurance for these assertions?
Why should reason be prevented from asking the ontological ques-
tion, "why is there something rather than nothing?"
Perhaps we are once again asking more of the Essay on
Metaphysics than it was intended to achieve. it is nonetheless
not unreasonable to put such questions to Collingwood, since
metaphysics is one branch of the science of thought (EM, 101),
and therefore seeks meaning just as all science and all thought
does; and if it is an absolute presupposition of science not
only that reality is intelligible or conforms to law, but also
that there is a reality to investigate (EM, 213, 222-27), then
it seems that on either view of metaphysics--Collingwood's or
Aristotle's--the examination of the meaning of being is a
legitimate enterprise. Surely we must assume that Collingwood's
intention was not primarily to write a treatise on Aristotle or
his metaphysics, but to write an essay or series

829
of essays showing how a conception of metaphysics on the mind--
mind model is still possible, and in doing so he found suffi-
cient basis in Aristotle for making the points he had to make to
the audience he presumed he had. Collingwood's rejection of
Aristotle's ontology is not so much an attempt to re-write
Aristotle as it is an attempt to show that he was in basic
agreement with those critics of Aristotle who rejected a science
of pure being, and to indicate that such a fruitless inquiry had
its roots in some of the ideas put forward in the treatises on
metaphysics attributed to Aristotle. Presumably his readership
included some of the same critics who found A. J. Ayer's
Language, Truth, and Logic of some interest, and at the
beginning of an essay which by its title threatened to raise
again the metaphysical spectra he was anxious to show this
audience that he was not about to revive this long-dead topic.
But for such a presumed, postivistically influenced audience he
was equally anxious to show that in rejecting one meaning of
metaphysics one has not demonstrated the impossibility of
metaphysics in any sense whatever, and metaphysics as a science
of first principles also has its pedigree in Aristotle's
treatises.
But having said this there is the additional problem
that Collingwood is sidestepping a discussion of what we have
demonstrated beyond all reasonable doubt to be one of the most
important issues in his own philosophy--its basic opposition to
realism. Even if "pure being"

830
is the metaphysical counterpart of an object unaffected by the
knowing of it, it is also the generic concept of a scale of
forms of being that Collingwood could hardly refuse to
investigate as one of the absolute presuppositions of science.
For on his own grounds, he not only accepted in his philosophy
of mind the precepts that the esse of mind is de hac re cogitare
(RP, 100; FR, 172), and in his theory of perception that the
esse of a sensum is sentiri (PA, 198), but also in cosmology
that the modern conception of nature presupposes that being is
becoming. "For an evolutionary science of nature, the esse of
anything in nature is its fieri," which requires that the old
conception of substance that grounded mechanistic physics (and
before that the animistic physics of the Greeks) is resolved
into the concept of function (IN, 15-16). With these clues in
mind, can anything be legitimately said about the meaning of
being that would be acceptable to Collingwood in his role as
metaphysical reformer?
Surely Collingwood would continue to argue that there
can be no science of something as lacking in distinction as pure
abstract being. But at the lowest level of consciousness
something present but indeterminate must survive and be per-
petuated long enough for sensation to occur at all (PA, 212-13).
The first specification of the generic essence of being is
therefore minimally the indeterminate presence of something to
sensation. As this is perpetuated and domesticated in con-
sciousness the meaning

831
of being of the object is refined and modified: from being an
indeterminate presence it becomes a sustained object for
perception, an imagined whole, a coherent imagined whole, a
referred-to imagined whole, etc.--in accordance with the
progressively higher activities of consciousness. At each stage
the meaning of the object's being is determined by the form of
consciousness, and the unique stage of the demand for meaning
made by that form of consciousness on the object.
Is there anything that lies at the summit of this scale?
Is there an object so rich in meaning, so determinate in its
being, that it utterly satisfies the highest demands of the mind
for significance? We have very little to build upon in the
published writings to reconstruct an answer to this question,
since the entire question of concrete ontology is left
unresolved in the Essay on Metaphysics. What we are lacking is
any adequate idea of Collingwood's cosmology-something he
lectured on repeatedly, and which he declared to be one of the
most pressing needs in contemporary philosophy. There is some
indication in The Idea of Nature that Collingwood supported
Hegel's division of reality into the realms of matter, life, and
mind, where each is in the process of turning into the next
higher form, and each is a more perfect embodiment of the
Absolute Idea (IN, 159). But he is also critical of the residue
of mechanism and
832
logicism in Hegel's philosophy of nature (IN, 121-32). The
modern idea of nature, Collingwood argues, recognizes not only
that there is no nature at an instant (as Whitehead phrases it),
but also that a thing is what it does, so that there is no
distinction between motion and that which moves: "modern
physical theory regards matter as possessing its own
characteristics, whether chemical or physical, only because it
moves; time is therefore a factor in its very being, and that
being is fundamentally motion: (IN, 151-52). Furthermore the
older conceptions of space and time--postulated as infinite
extension and succession--"seem to be nothing but abstractions
from the idea of movement" on the one hand, or logical
presuppositions of that idea on the other. But then one
encounters this remarkable passage:
This at any rate seems clear: that since modern science is

now committed to a view of the physical universe as finite,
certainly in space and probably in time, the activity which
this same science identifies with matter cannot be a
self-created or ultimately self-dependent activity. The
world of nature or physical world as a whole, on any such
view, must ultimately depend for its existence on something
other than itself . . . . (M)odern science, after an
experiment with materialism, has come back into line with
the main tradition of European thought, which has always
ascribed to nature an essentially derivative or dependent
status in the general scheme of things. It is true that the
most varied proofs have been offered as to why nature must
be dependent, and the most varied theories as to what it
depended on; but in general, with strikingly few
exceptions, scientists and philosophers have agreed that
the world of nature forms only one part or aspect of all
being, and that in this total realm its place is a
secondary one, one of dependence on something prior to
itself. (IN, 155).
833
But while Collingwood points out that many contemporary cos-
mologists argue, as did their classical and medieval prede-
cessors, that what nature depends upon is God, Collingwood comes
up with a less traditional solution:
Throughout the long tradition of European thought it has

been said . . . that nature, though it is a thing that
really exists, is not a thing that exists in itself or in
its own right, but a thing which depends for its existence
upon something else. I take this to imply that natural
science . . . is not, as the positivists imagined, the only
department or form of human thought . . . and is not even a
self-contained or self-sufficient form of thought, but
depends for its very existence upon some other form of
thought which is different from it and cannot be reduced to
it . . . . What is this other form of thought? I answer,
"History." (IH, 175-76).
What can this mean? What sense of "History" (with a
capital "H") can he have in mind? This brings us face to face
once again with the central paradox of the Essay on Metaphysics,
for while Collingwood has made a fairly convincing case in The
Idea of Nature that the modern idea of nature is modeled on
evolution, and evolution is modeled on the idea of history (IN,
9, 132), at precisely the point where we expect him to complete
the argument on the ontological side he appears to swap horses
and ride out on an epistemological charger, a kind of equus ex
machina. If natural science is dependent upon something else
because nature is dependent on something else, then by force of
the same analogy if what natural science depends upon is history
(in the sense of a kind of knowledge), what nature depends upon
is the ontological referent

834
of historical knowledge. What can this be? Certainly not the
deeds of men done in the past. Is "History" (with a capital "H")
to be taken in the sense of cosmic development, in spite of the
fact that there is nothing in the development of Collingwood's
thought to prepare us for this sudden expansion of the idea of
history, and everything to indicate that it was moving in just
the opposite direction?
Or is there some reality that grounds both nature and
history? Such would seem to be the sense of those early passages
in his writings such as that in Speculum Mentis in which
Descartes' cogito ergo sum is interpreted to mean that "the
concrete historical fact, the fact of my actual present aware-
ness, ((is)) the root of science," and consequently that "Sci-
ence presupposes history and can never go behind history" (SM,
202). Such is the idealistic strand in Collingwood's braided
ring of thought. But opposed to this is the realistic counter--
tendency:
modern thought is disentangling itself from the cobwebs of

subjective idealism . . . . Kant would suggest a very dif-
ferent conclusion: namely that if nature bears on its face
the marks of dependency for its existence on something
else, that something is the human mind . . . . This is bad
philosophy . . . . The most rigorous thought of our own
time, scientific and philosophic alike, has turned
resolutely away from these subjectivist or phenomenalist
doctrines, and agrees that whatever nature depends upon it
does not depend on the human mind. (IN, 156).
835
What is left? Could Collingwood, whose respect for religious
consciousness remained unquestioned throughout all his published
writings, fail to see that a mind unreconciled to the mind of
God has failed in its primary metaphysical obligation? It would
not seem so, judging solely from the following passage from
Chapter XXI of the Essay on Metaphysics, "Quicunque Vult," which
may well be taken as Collingwood's final offering of
rapprochement philosophy:
If metaphysics is our name for the statement of absolute

presuppositions, and if metaphysics and theology are the
same, there are three ways in which the existence of a
world of nature might be made to figure among the doctrines
of theology. 1. It might be a proposition in metaphysics,
as it is for Spinoza, that God and nature are the same. But
this would entail the consequence that natural science is
the same thing as metaphysics: which cannot be right if the
business of metaphysics is to state the absolute pre-
suppositions of natural science. 2. It might be a propo-
sition in metaphysics that the world of nature exists, but
this proposition might be left wholly unrelated to the
proposition that God exists. But then it would not be a
proposition in theology; and therefore, if theology and
metaphysics are the same, not a proposition in metaphysics.
And what about the presupposition of which it was the
statement? The act by which we hold such presuppositions, I
have said elsewhere, is religious faith; and God is that in
which we believe by faith; therefore all our absolute
presuppositions must be presuppositions in holding which we
believe something about God. 3. It might be a proposition
in which the existence of the world of nature was stated in
the form of an attribute or activity of God; and this seems
the only possible alternative. (EM, 215-16).
BIBLIOGRAPHY
Aristotle. Metaphysics. Translated with Commentaries and

glossary by Hippocrates G. Apostle. Bloomington: Indiana
---------. The Metaphysics. With an English translation by Hugh

Tredennick. (The Loeb Classical Library). Two Volumes.
Cambridge: Harvard University Press, 1933.
Belnap, Nuel D., and Thomas B. Steel. The Logic of Questions and
Answers. New Haven: Yale University Press, 1976.
Blachowicz, James. "History and Nature in Collingwood's

Dialectic." Idealistic Studies, VI (January, 1976),
Bochenski, I. M. Contemporary European Philosophy. Translated

from the German by Donald Nicholl and Karl
Aschenbrenner. Berkeley: University of California Press,
1969.
---------. The Methods of Contemporary Thought. Translated from

the German by Peter Caws. New York: Harper and Row,
Publishers, Incorporated, 1968.
Bosanquet, Bernard. Knowledge and Reality. London: S.

Sonnenschein and Company, 1892.
---------. The Principle of Individuality and Value. London:

Kraus Reprint, 1927.
Bradley, F. H. Appearance and Reality. Oxford: Clarendon Press,

1893.
---------. The Presuppositions of Critical History. Edited with

an introduction and commentary by Lionel Rubinoff.
Chicago: Quadrangle Books, 1968.
---------. The Principles of Logic. Two Volumes. Second edition

revised with commentary and terminal essays. London:
Oxford University Press, 1922.
Bury, J. B. A Historv of the Later Roman Empire. London:

Macmillan and Company, Ltd., 1923.
836
837
Chisholm, Roderick M. Realism and the Background of

Phenomenology. New York: The Free Press, 1960.
Collingwood, Kate. Personal letter to the author, dated 26

October 1969.
Collingwood, R. G. An Autobiography. London: Oxford University

Press, 1939.
---------. Correspondence with G. Ryle, 9 May 1935 and 6 June

1935. The Bodleian Library, 1964: Ms. Eng. lett. d. 194.
---------"Economics as a Philosophical Science." International

Journal of Ethics, XXXVI (1925), 162-185.
---------. An Essay on Metaphysics. Oxford: The Clarendon Press,

1940.
---------. An Essay on Philosophical Method. Oxford: The

Clarendon Press, 1933.
---------. Essays in the Philosophv of Art. Edited by Alan

Donagan. Bloomington: Indiana University Press, 1964.
---------. Essays_in the Philosophy of History. Edited by

William Debbins. New York: University of Texas Press,
1965.
---------. Faith and Reason: Essays in the Philosophy of Religion

Edited by Lionel Rubinoff. Chicago: Quadrangle Books,
1968.
---------. The Idea of History Edited by T. M. Knox. Oxford:The

---------. The Idea of Nature. Edited by T. M. Knox. Oxford: The

---------. The New Leviathan. Oxford: The Clarendon Press, 1942.
---------. "Political Action." Proceedings of the Aristotelian

Society, XXIX (1928-1929), 155-176.
---------. "The Present Need of a Philosophy." Philosophy, 9

(1934), 262-265.
---------. The Principles of Art. Oxford: The Clarendon Press,

1938.
838
---------. Religion and Philosophy Company, 1916. London:

Macmillian and Company, 1916.
---------. "Sensation and Thought." Proceedings of the

Aristotelian Society, XXIV (1923-24), 55-76.
---------. Speculum Mentis. Oxford: The Clarendon Press, 1924.
Däniken, Erich van. Chariots of the Gods? Translated by Michael

Heron. New York: G. P. Putnam's Sons, 1970.
Donagan, Alan. The Later Philosophy of R. G. Collingwood.

Oxford: The Clarendon Press, 1962.
Dray, W. H. "Philosophy of History." The Encyclopedia of

Philosophy. Edited by Paul Edwards. New York: Macmillan
Publishing Company, Inc. and The Free Press 1967, VI,
247-254.
Gallie, W. B. Philosophy and the Historical Understanding. New

York: Schocken Books, 1964.
Gardiner, Patrick, editor. Theories of History. Glencoe: The

Free Press, 1959.
Geach, Peter. Mental Acts. 1957. London: Routledge and Kegan

Paul,
Harris, Errol E. "Collingwood on Eternal Problems."

Philosophical Quarterly, I (April, 1951), 228-241.
---------. Hypothesis and Perception. London: George Allen and

Unwin, Ltd., 1970.
---------. "Mr. Ryle and the Ontological Argument." Mind, XLV

October, 1936), 474-480.
---------. Nature, Mind, and Modern Science. London: George

Allen and Unwin, Ltd., 1954.
Hegel, G. W. F. Logic. From The Encyclopaedia of the

Philosophical Sciences. Translated by William Wallace.
Lon. Oxford University Press, 1873.
Hick, John H., and Arthur C. McGill, editors. The Many-faced

Argument. New York: The Macmillan Company, 1967.
Johnston, William M. The Formative Years of R. G. Collingwood.

The Hague: Martinus Nijhoff, 1967.
839
Kant, Immanuel. The Critique of Pure Reason. Translated by N.

Kemp Smith. New York: St. Martin's Press, Inc.,
1929/1961.
Klibansky, Raymond, and H. J. Paton, editors. Philosophy and

History. Oxford: The Clarendon Press, 1936.
Knox, T.M. "Notes on Collingwood's Philosophical Work, with a

Bibliography." Proceedings of the British Academy, XXIX
(1944).
Krausz, Michael, editor. Critical Essays in the Philosophy of R.

G. Collingwood. Oxford: Clarendon Press, 1972.
Langer, Susanne K. Feeling and Form. New York: Charles Scribner

s Sons, 1953.
MacKay, Donald S. "On Supposing and Presupposing.'' The Review

of Metaphysics, II (September, 1948), 1-20.
McCallum, R. B. "Robin George Collingwood." Proceedings of the

British Academy, XXIX (1944), 463-468.
Mink, Louis 0. Mind, History, and Dialectic: The Philosophy of

R. G. Collingwood. Bloomington: Indiana University
Press, 1969.
Owens, Joseph. The Doctrine of Being in the Aristotelian

Metaphysics. Toronto: The Pontifical Institute of
Mediaeval Studies, 1951.
Passmore, John. A Hundred Years of Philosophy. Baltimore:

Penguin Books, 1957/1968.
Paton, H. J. "Fifty Years of Philosophy." Contemporary British

Philosophy, Third Series. Edited by H. D. Lewis. New
York: Humanities Press, 1956/1961.
Post, John F. "A Defense of Collingwood's Theory of

Presuppositions." Inquiry, VIII (1965), 332-354
--------. "Does Knowing Make a Difference to What Is Known?"

Philosophical Quarterly, XV (1965), 220-228.
Robinson, Richard. Plato's Earlier Dialectic. Oxford: The

Rubinoff, Lionel. Collingwood and the Reform of Metaphysics.

Toronto: University of Toronto Press, 1970.
840
Russell, Bertrand. "Logical Atomism." Contemporary British

Philosophy, First Series. Edited by J. H. Muirhead.
London: George Allen and Unwin, Ltd., 1924.
---------. "Meinong's Theory of Complexes and Assumptions."

Mind, XIII (1904).
Ryle, Gilbert. "Mr. Collingwood and the Ontological Argument."

Mind, XLIV (April, 1935), 137-151.
Rynin, David. "Donagan on Collingwood: Absolute Presuppositions,

Truth, and Metaphysics." The Review of Metaphysics, XVIII
(December, 1964), 301-333.
Strawson, P. F. Introduction to Logical Theory. London: Methuen

and Company, Ltd., 1952.
Thayer, H. S. Meaning and Action: A Critical History of

Pragmatism. New York: The Bobbs-Merrill Company, Inc.,
Toulmin, Stephen. Human Understanding: The Collective Use and

Evolution of Concepts. Princeton: Princeton University
Press,-T972.
Warnock, G. J. English Philosophy Since 1900. Oxford: Oxford

Warren, Howard C. A History of Association Psychology. New York:

Charles Scribner's Sons, 1921.
Wilson, Clifford. Crash Go the Chariots. New York: Lancer Books,

Inc. 1972.
BETWEEN THE HORNS:
SOME COLLINGWOOD DILEMMAS AND THEIR SOLUTIONS
By Glenn Shipley
1. The Four Autobiographical Themes and Their Dilemmas.
Many students of R.G. Collingwood report that their interest in his thought was
awakened by An Autobiography, the remarkable testimony of an Oxford don running counter to
the currents of thought of his day and passionately devoted to the archeological, historical and
philosophical life that he led. But reading beyond the Autobiography to his other published
writings forces those of us drawn to his thinking to wonder to what extent his theory and his
practice coincided. We are thrust into a consideration of this matter by dilemmas that arise
directly from his autobiographical self-interpretation. This paper will take a direct look at four
dilemmas arising from Collingwood’s Autobiography, then proceed to an examination of how he
himself approached some related philosophical dilemmas, and finally it will draw on themes
revealed in Collingwood’s actual praxis to point the way towards resolving the dilemmas he left
for us.
I have argued elsewhere that a careful reading of the Autobiography reveals four
dominant themes that indeed carry over into all of his early and later published writings: (1)
Collingwood’s reaction to the Oxford “realist” philosophers; (2) his development of question-
and-answer (Q-A) logic; (3) the theoretical and practical relationship of history and philosophy;
and (4) the overall goal of a rapprochement philosophy that could provide a reconciliational
bridge not only between history and philosophy, but also theory and practice, etc.1
Aside from the Autobiography’s supposed “errors of omission," there remain dilemmas
which still gore those of us who want to make sense of Collingwood’s philosophy. Chief among
them are some arising from these very issues - the idealism-realism controversy, Q-A logic as an
alternative to propositional logic, metaphysics as an historical science of absolute, truth-neutral
presuppositions, and philosophy as absolute or relative. While I will have a few things to say
about all of these issues, my primary concern will be with the issues of anti-realism and
rapprochement philosophy.
(A.) Let us start where Collingwood starts, with the realism-idealism controversy.
Collingwood wants us to be clear that the assertion of an object unaffected by the knowing of it –
the fundamental standpoint of the position he unwaveringly calls “realism” – is the root error he
1
From Realism to Rapprochement: The Autobiographical Interpretation of Collingwood’s Philosophy (FRTR),
Loyola University of Chicago, 1983. This work is available in digital format to anyone who requests it.
has devoted his intellectual life to overcoming. The key argument overturning this philosophical
position is one which demonstrates that the realist's position affirms what it simultaneously
denies: it claims to know what it admits it does not know - i.e. what the object is like outside of
the condition of being known.
The dilemma that emerges is as follows. As John Post points out,2 Collingwood’s anti-
realism argument cuts both ways: it is as effective against the idealist position as it is against the
realist one. To say that one knows that the object is affected by the knowing of it is ruled out by
the same inability to satisfy the conditional premise. The logical outcome of the argument would
then seem to be a confirmation of a basically skeptical position, that one cannot know whether an
object is or is not affected by the knowing of it. The goring occurs because Collingwood
claimed not to be a skeptic.
Now if Collingwood were to construct a comparable argument to rule out skepticism, it

would need to address the inversion of the fundamental realist presupposition to show that
“Knowledge is unaffected by the object known” is absurd. He would need to tread carefully
between the objectivity-subjectivity dilemma. An object totally determined by acts of knowledge
possesses a meaning that is made rather than found, and which affirms subjectivity at the
expense of abandoning any independent claim for objectivity. An object totally determining the
act which knows it has a meaning which is found rather than made, which affirms objectivity at
the expense of subjectivity, which then appears passive, plastic, and dependent.
Even if we are careful to distinguish broader and narrower meanings for terms like
"knowledge," "affected by," and "object," in Collingwood's philosophy of mind it is still difficult
to find any room for the object to have any affect on the act of knowledge. Even though he states
in The Principles of Art that in the case of perception knowledge does not make such a difference
to the object known that it creates the object ex nihilo, and that something must be present to
sensation with sufficient strength and for long enough time for attention to focus on it (PA, 291-
92), nevertheless in The New Leviathan Collingwood abandons even this concession to realism.
He holds that objects of sensation are utterly ambiguous with respect to all interpretative
categories of understanding, going so far as to assert that in a field of sensa a red patch is not
distinct from any other characteristics in that visual field (NL, 4.45-4.63). In fact he claims that
attention “draws the line” that makes such distinctions in a field of sensa possible (NL, 5.62-
5.65). What we are left with is "a flux of things within us and without, and … although this is
actually present to us (nothing else, indeed, is present) it is not intelligible" (PH, 173). If it is an
act of consciousness that makes all perceptual distinctions, then it would appear that perceptual
2
John F. Post, "Does Knowing Make a Difference to What is Known?" Philosophical Quarterly, xv, no. 60 (July,
1965), pp 220-28.
objects are “made” rather than “found,” and Collingwood has exhibited a basically Kantian
orientation and crossed the line into subjective idealism.3
Now while this anti-realist stance may be consistent with his autobiographical
interpretation, it leaves the reader lacking resources to solve the dilemma. It is particularly
vexing in the light of Collingwood’s insistence that scientific history requires the critical
interpretation of evidence, where evidence is not something completely determined by the act of
historical interpretation. In any version of re-enactment, the question arises concerning the
original historical enactment, whose encapsulation, one assumes, has some fabrication-resistant
substance to it in order to avoid historical fraud or charlatanism.4 In essence, then, this is the
recurrent “found” vs. “made” ambiguity, and future interest in Collingwood hinges on the ability
of interpreters to either resolve the paradox or remain impaled on the horns of yet another
version of the realist-idealist dilemma.
(B.) This dilemma carries over into the second theme, Collingwood’s “discovery” of Q-A
logic. Collingwood wants us to see Q-A logic as an alternative to propositional or formal logic,
which he portrays as the realist's only "method" of philosophical analysis. As presented in quasi-
axiomatic form in the Essay on Metaphysics it would appear that since a one-one correspondence
between indicative sentences in a language and propositions in a formal system is not possible,
one might be free to dispense with formal logic altogether, and engage in a free-form (if
Baconian) cross-examination of evidence based on unquestioned presuppositions.
But even at a very rudimentary level it is clear that Q-A logic cannot serve as a
replacement for formal logic. It does not provide any way to decide such simple issues as when
one proposition is validly or invalidly inferred directly from some other proposition, or how one
is to decide issues of logical competence. Interposing a question within an inference adds
nothing to it: from a logical point of view it is unnecessary to the derivative relationship of a
conclusion to its premises.
But we need to feel the tips of the horns of this particular dilemma a bit more acutely. In
both versions of Q-A logic - in the Autobiography and the Essay on Metaphysics, Collingwood
describes the process in person- or thought-dependent terms: "Every statement that anyone ever
makes is made in answer to a question (EM, 23); "'Proposition' denotes an assertive act of
thought or what in those acts of thought is asserted" (A, 30). Yet as "logic" Q-A hopes to
preserve its compulsive, necessary force as "criteriological." A proposition is always a logical
3
Collingwood’s attempt in The New Leviathan to evade this conclusion by calling the denial of sense-objects a
“methodological” negative is not convincing.
4
Grace Simpson’s defense of Collingwood’s archeological practice in the 1937 excavation at King Arthur's Round
Table is interesting in this regard. "I saw Bersu push his trowel to demonstrate to my father how Collingwood 'made
a post-hole'. I heard those shocking words. One does not make a post-hole, one finds one. I have never forgotten
how upset my father was." "Collingwood's Latest Archeology Misreported by Bersu and Richmond," Collingwood
Studies, vol. V (1998), p. 114.
and not merely a linguistic entity - i.e. it states what ought to be the case rather than what merely
is the case concerning assertive acts of thought (A, 31). A question is said to "arise" because
some presupposition has "logical efficacy" to make it arise (EM, 27). Now if what "logic" and
"logical" means is something like "how a real person really thinks when he is being careful and
scientific" this is either a tautology (logic is thinking logically) or it is embedded in personal
consciousness. But if it is thought or person-dependent, a proposition loses its compulsive force
as following necessarily from premises, whether it is in F-logic or Q-A logic.
So the dilemma is that if logic is person or thought-dependent, it lacks necessary

compulsion: "it" (the conclusion or proposition) does not "follow" or "arise," I make it do so. In
other words logic is not psychology, but it is psychology.5 No self-respecting formal logician
would agree to this. Whatever else logic is, it is not epistemology, and it is not dependent on
what any person happens to think. But if logic is neither person nor thought dependent, it ceases
to be a product of human intelligence and becomes an "automatic" or mechanistic process over
which man has no control.
(C.) The third autobiographical theme - the philosophy of history - is the issue that has
drawn the most attention by students of Collingwood. The issue I wish to focus on is the sort of
"reconciliation" that Collingwood finds between these two modes of thought. It is plain that the
remainder of his reconciliational program hinges on this, since Collingwood clearly wanted to
make the history-philosophy rapprochement paradigmatic for all of its other manifestations.
In order to satisfy the conditions of the Autobiography for minimal reconciliation,

Collingwood needs (1) to analyze the concept of history and exhibit philosophy as necessarily
contained within it, and (2) to analyze the concept of philosophy to exhibit history within it.
Philosophy is in history insofar as history is a variety of self-knowledge, is the history of
thought, is the exercise of a priori imagination, deals with a past that is actually an aspect of the
present, reasons to its conclusions using evidence and inference, aims at a kind of truth, etc.
History is in philosophy insofar as reality is a process of change, there are no eternal truths or
unchanging entities, no such thing as human nature, its metaphysical subject matter has
presuppositions rooted in the science dominant in its day, etc. It is not surprising that some
students of Collingwood have found that this relationship puts philosophy in a very negative
light.
5
" Take the mental activity as a self-contained fact; refuse, so far as that is possible, to treat of its metaphysical
aspect, its relations with real things other than itself and you have psychology. Thus in scientific thought as studied
by logic we have a judgment in which the mind knows reality: psychology, treating the judgment as a mere event,
omits its reference to reality, that is to say, does not raise the question whether it is true" (RP, 40). In a footnote to
this passage, Collingwood adds: "The same omission or abstraction is made by Formal Logic, which I take to be a
psychological rather than a philosophical science."
The stance of the Essay on Metaphysics puts a sharper point on this dilemma. If the
metaphysician's task is simply to ferret out the absolute presuppositions of science and in the
process be restricted from showing how one set of absolute presuppositions is better as "truer"
than another, or how one sort of science is better as having more "con-supponible" absolute
presuppositions than another, or how absolute presuppositions in one constellation give rise to
absolute presuppositions in another, or why one science is replaced by another, then the
metaphysician starts to resemble a rather sophisticated scissors-and-paste man. He is producing
a "catalogue" of principles. The kind of historian this metaphysician is reduced to is one who
cannot be truly evaluative or criteriological, who cannot demonstrate what is progressive in
historical process, and whose attitude towards his subject matter is less respectful than
idolatrous. But if on the contrary the metaphysician takes his historical work as seriously as
Collingwood would have him take it in his exhortations elsewhere about scientific history, then
his hold on the truth-neutrality or absoluteness of presuppositions starts to slip. So Collingwood
seems to be saying both that metaphysics is an historical science while simultaneously denying
that there is anything particularly historical about it (it consists in logical analysis of systems of
inquiry to expose their absolute presuppositions - which has nothing transparently to do with
deeds of men done in the past, re-enacted and narratively presented, etc); and that history is a
metaphysical science while simultaneously denying that there is anything metaphysical about it
(it rests on truth-neutral presuppositions, is not grounded in ontology, etc.).
(D) This leads us into our fourth and last autobiographical theme, the issue of
"rapprochement philosophy". The Autobiography is somewhat vague about what that is, even
while it is explicit in what it applies to - primarily theory and practice and issues in moral, social
and political philosophy. But if we are to take the examples of this rapprochement as they occur
in his actual published writings, the meaning of "reconciliation" varies from (1) the mere
sequencing of two modes of thought in a "process of development," to (2) a complementarity
between two modes of thought able to approach one another and yet retain their separate identity,
to (3) a dialectical relationship of sub- and super-ordination in a scale of forms with necessary
relations of opposition and distinction, differences of degree and kind, to (4) a complete unity or
identification of the reconciliata.
While the main issue here is what the ultimate goal is for reconciliational philosophy, the
brunt of the dilemma is felt on the issue of whole vs. part. What is the overall context within
which rapprochement works? Are the reconciliata parts of some larger whole, or is the whole
merely the sum of its parts?
Now there is no apparent reason that there cannot be many forms of reconciliation, and
for different purposes, but for an anti-realist philosopher like Collingwood who repeatedly
directs his readers to seek truth in "the whole" rather than "the part," this does not seem
satisfactory. Is there an Absolute end-point to the reconciliational process? While philosophy is
portrayed as “absolute knowledge” in his early writings, one looks in vain in the Autobiography
for any hint of "the Absolute" or mention of dialectic. In his final published works "the
Absolute" no longer appears as a noun, but only as an adjective - a qualifier regarding
presuppositions, and that only in the Essay on Metaphysics. We wish to know if in his final view
Collingwood abandoned the absolute idealism of his youth, as in the cryptic description of the
"unified life of the mind" sought in Speculum Mentis, or if he became a complete relativist with
no final anchor for any scale of forms of knowledge. In short, if there is no common ground for
reconciliata - no anchor or absolute standpoint - they remain unreconciled.
2. Getting Between the Horns: Theory and Practice
An electronic search on the term "dilemma" across all of Collingwood's published

writings calls up references to the term in virtually every one of them - the Autobiography and
The New Leviathan being the sole exceptions.6 Some of these references are not really to the
point of this paper, but of those that are, the discussion is often revealing of how Collingwood's
mind worked.
1. Collingwood's earliest published work, Religion and Philosophy, is remarkable for the
interesting and undisguised glimpses it gives us into positions that he will elaborate in his later
works. For example in the key chapter on personality he points to the fact and continuing ideal
of inter-personal communication as the fundamental starting point for any theory of knowledge,
and takes it as evidently true that two minds can think the same thought (RP, 98-103). Since the
mind is its thoughts, two minds which think the same thought or will the same action share the
unity of consciousness which is the mark of the individual (RP, 101-2). Not only can two persons
have "one mind" insofar as they share the same thought or will the same action, but two forms of
thought like history and philosophy are called "the same thing" in the sense of referring to the
same reality or the same intentional object - the "one real world" (RP 51). Each presupposes the
other, as the particular presupposes the universal and vice-versa (RP, 49; cf. PH, 352).
In a chapter entitled "matter" Collingwood develops an argument aimed at resolving the

"materialism-idealism" dilemma - namely the disjunction made by the "plain man's metaphysic"
between mind as a non-extended or spaceless reality consisting of thought, and matter which
6
An electronic word search for "dilemma" yields the following references: RP, 71, 88, 120, 122, 129; SM, 21, 66,
232-35, 238, 240-41, 283-86; EPM, 21, 61, 68-69, 80-81, 85, 161: PA, 34-36, 311; A, [none}; EM, 333-35; NL,
[none]; IH, 141-42, 151, 157-58,163-65, 174-75, 278-79, 291-301, 353-54, 482; PH, 14, 33, 197, 200-01, 226-28;
IN, 41-2, 51, 140, 174. Since Collingwood sometimes refers to a dilemma as a paradox, contradiction or
opposition, it would be useful to do electronic searches on these terms as well. "Paradox," for example, appears at
NL, 5.34, 6.52, 6.54, 7.16, 26.61, and 36.91.
occupies space and is subject to the law of causation, but has no consciousness of itself (RP, 72).
The dilemma is escaped once it is shown that matter is inconceivable without ascribing to it
some qualities of mind (it is active as well as passive) and mind without ascribing to it some
qualities of matter (it is passive as well as active) (RP, 94). In these demonstrations it is assumed
that real thought cannot be self-contradictory without being self-destructive (RP, 129), so it
would appear that dilemmas are always abstract and merely apparent.
Collingwood claims that the position he is advocating violates neither realism nor
idealism in their more satisfactory forms (RP, 101, n.1). In the chapter on materialism in Part II
(Religion and Metaphysics) he argues that the materialist position (which he identifies as a kind
of realism - see RP, 73, n 1) leads to a part-whole dilemma ("either know the whole or do not
pretend to know even this one part") and ultimately to skepticism (knowledge of reality as a
whole is knowledge of an endless series of discrete parts, so it is impossible), and that both
problems are escaped by arguing that "knowledge of this single part is already knowledge of the
whole" (RP, 88; cf. PH, 188).7 But in the end, in the choice between a crude materialism (read:
"realism") that reduces all of mind to matter, and an "equally crude idealism" that reduces all of
matter to mind, even the youthful Collingwood will opt for idealism - but only if "mind" is
understood as a concept with layers of meaning. The mind, he argues, is not a mechanical
structure automatically producing thought. It is an activity whose esse is not just simply
cogitare but de hac re cogitare (RP, 100).
To ask whether mind is a form of matter or matter is a form of mind is very largely a
question of words. The important thing is that we should be able to bring the two into
relation at all; that we should hold such a conception of matter as does not prevent us
from admitting truth, morality, and life as a whole to be real facts, and that we should
hold such a conception of mind as does not reduce the world to an illusion and experience
to a dream. The first of these errors is that of a crude materialism, and the second that of
an equally crude idealism. The view for which we are contending would claim the title
of idealism rather than materialism, but only because the current conception of mind
seems a more adequate description of the world than the current conception of matter.
We are laying stress on the fact that the world is the place of freedom and consciousness,
not of blind determinism; at present this can best be conveyed by saying that mind is the
one reality. (RP, 94-95.)
The fact and ideal of complete interpersonal communication, along with the themes of "thought"
as a meaning-seeking activity, the reciprocity of part and whole, the description of mind as
intentionality, the non-mechanical nature of human understanding, and the insistence on
describing knowing in such a manner that determination of the object by mind is not ruled out -
all these are themes that recur throughout his later writings - modified, but never abandoned.
7
This argument re-appears in a different context in the Essay on Philosophical Method as the solution to another
dilemma (the circularity of philosophical argument) by showing how "coming to know means coming to know in a
different and better way" what is to some extent already known (EPM, 161).
2. Speculum Mentis explicitly rejects his earlier solution to the realist part-whole dilemma
(SM, 240), and then attempts an alternative escape by showing that "the infinite given whole of
fact is the nature of the knowing mind as such" (SM, 241). In each of the five forms of
experience that he explores, a distinction and separation is made between subject and object, "a
relation of difference without identity" characteristic of the abstract concept (SM, 243). In all
such distinctions each side - subject and object - represents a false abstraction, and getting
between the horns occurs when both subject and object are shown to be part of the same concrete
universal in all forms of experience. Speculum Mentis also hints at a dialectical development
wherein two concepts are synchronically opposed by ignoring their diachronic development.
This occurs by making explicit what was only implicit in a previous, subordinate, preceding
mode of thought. Philosophy is therefore self-knowledge achieved by correcting errors made by
subordinate forms of thought (art, religion, science, history and the like), each of which is seen as
a philosophical error as a "logical consequence of realism" which continually mistakes the
product of abstraction as something independent of the mind which makes it (SM, 249-52).
But on the way to "absolute knowledge" - the philosophical self-knowledge of mind -

Collingwood writes the following, remarkable passage. Speaking of the transition from religion
to philosophy he says:
the task of philosophy, regarded as the philosophy of religion, is the simple translation of
this solution of the riddle of life out of the language of mythology into that of philosophy.
… Translation itself is based on the fact that the meaning takes new colour and shines
with a new light when we express it in different words. To set the meaning as an abstract
self-identity over against the language makes translation pointless: to swamp it in a mere
immediate union with the language itself makes translation impossible. Meaning and
language are simply the universal and the particular. Abstract logic, the false theory of
science, sets them outside one another. Abstract aesthetic, the false theory of art and
religion, swamps the one in the other. Concrete logic sees them as distinct yet
inseparable. (SM, 253).
The function of philosophy is thus "the process of translation into progressively adequate
language," which "is simply the dialectical self-criticism of thought" (SM, 253-54). Once again
idealism and realism are pitted against one another, but this time Collingwood accuses them both
of "committing the fundamental error of separating the metaphysical inquiry as to what the world
is in itself from the psychological inquiry as to how we come to know it" (SM, 266). Idealism
leaves the opposition of subject and object unreconciled, or else it reconciles them by turning the
object into another mind or a society of minds or an infinite mind - all of which he rejects. (SM,
266-67)8 In its place he proposes a philosophy of absolute knowledge reconciling realism as a
background and idealism as a foreground as products of "abstraction" - the adoption of the
8
In a later section on historical philosophy, Collingwood concedes that it was the achievement of the great German
idealists that they "killed scientific realism - the popular philosophy of today - as dead as a door-nail" (SM, 287).
externality of the object, which in reality is mistaking the imaginary for the real (SM, 291-98).
What remains of his "map of knowledge"9 in the end is summarized as follows:
Our position at the start was wholly realistic, and there is a sense in which it is realistic to
the end. … [But} we did not assume that any one form of experience could be accepted
as already, in its main lines, wholly free from error. Led by this principle, we found that
the real world was implied, but not asserted, in art; asserted, but not thought out, in
religion; thought out, but only subject to fictitious assumptions, in science…. The real
object is the mind itself, as we now know. But in abolishing the notion of an external
world other than the mind we do not assert any of the silly nonsense usually described…
as idealism. … The very essence of trees and hills and people is that they should be not
myself but my objects in perception: they are not subjective but objective, not states of
myself but facts that I know. None the less, my knowing them is organic to them: …
They and I alike are members of one whole…. (SM, 311).
We shall have more to say about the nature of this "one whole" in the next section. But note that
throughout Speculum Mentis the “forms of experience” are characterized by their mode of
expression in language – as metaphorical, implied, asserted (SM 252-53; cf. FRTR, 312-13).
3. The Essay on Philosophical Method works these relationships into a set of criteria for
deciding whether apparently opposing viewpoints (e.g. the horns of a dilemma) are fully
reconcilable or represent a pre-philosophical stage in which two reconciliata are opposed but not
distinguished, or are distinguished but not opposed. When they are seen as fully philosophical,
they are integrated into an overlapping scale of forms in which differences of degree are
coordinated with differences of kind, so that opposites are located at different points on a
hierarchical scale (EPM, 68). One is impaled on the horns of a dilemma by acceptance of a false
disjunction (EPM, 80-81) and all disjunctions are ultimately false, once they are understood
philosophically.
Later in the essay Collingwood takes up the issue of the circularity of philosophical
argument as "justifying its own starting point" (EPM, 160). With a significant glance back at
Kant, he writes:
If the first principles of philosophy are to be justified, they must be justified by that
philosophy itself. This can be done only if the arguments of philosophy, instead of
having an irreversible direction from principles to conclusions, have a reversible one, the
principles establishing the conclusions and the conclusions reciprocally establishing the
principles. But an argument of this kind, in which A rests on B and B rests reciprocally
on A, is a vicious circle. Are to conclude that philosophy is in the dilemma of either
renouncing this characteristic function and conforming to the irreversible pattern of exact
science, or else losing all cogency in a circular argument?
The solution of the dilemma lies in a feature of philosophical thought… the Socratic
principle that philosophical reasoning leads to no conclusions which we did not in some
sense already know already… [F]or if the species of a philosophical genus overlap, the
distinction between the known and the unknown, which in a non-philosophical subject
9
It is interesting that the dual metaphor of “mirror” and “map” are employed in Speculum Mentis - in its title (mirror
of the mind) and its subtitle (“a map of knowledge”). With characteristic compression, Collingwood identifies the
two (SM, 315) without recognizing the radically different representative function of each.
matter involves a difference between two mutually exclusive classes of truths, in a
philosophical subject-matter implies that we may both know and not know the same
thing; a paradox which disappears in the light of the notion of a scale of forms of
knowledge, where coming to know means coming to know in a different and better way.
4. Following the Essay on Philosophical Method, Collingwood worked out a theory of

linguistic expression as part of his philosophy of mind. This is accomplished in The Principles
of Art and The New Leviathan - but only in brief outline, and ignoring other riches in these
domains. In The Principles of Art, for example, Collingwood takes the wraps off some of his
anti-realistic weaponry and takes aim at psychological behaviorism and the "stimulus-response"
model of art appreciation as a form of anti-aesthetics (PA, 34) - issues we shall not pursue here,
except to note that his strategy again is dissolving a dilemma by focusing on the "false
disjunction" from which it arises.10 What I wish to call attention to in The Principles of Art is the
theory of perception and the characterization of language that he develops there. Concerning
perception, he allows room for the analysis of consciousness (basically a phenomenological
method) to develop answers to questions which an analytic philosophy of mind (putting
questions to consciousness) cannot (PA, 205-06; see FRTR, 382-83 & 383, n.1). Concerning
language, he positions it as a further specification of expression of emotion by means of bodily
acts (PA, 109-14), and speech as a further specification of language (PA, 242-51). In the midst
of this discussion is one of Collingwood's wonderfully lucid paragraphs - this about meaning:
Language in its intellectualized form has both expressiveness and meaning. As language,
it expresses a certain emotion. As symbolism, it refers beyond that emotion to the thought
whose emotional charge it is. This is the familiar distinction between 'what we say' and
'what we mean'. 'What we say' is what we immediately express: the eager or reluctant or
triumphant or regretful utterance in which these emotions and the gestures or sounds that
express them are inseparable parts of a single experience. 'What we mean' is the
intellectual activity upon which these are the emotional charge, and towards which the
words expressing the emotions are a kind of finger-post, pointing for ourselves in the
direction from which we have come, and for another in the direction to which he must go
if he wishes to 'understand what we say', that is, to reconstruct for himself and in himself
the intellectual experience which has led us to say what we did. (PA, 269)
In The New Leviathan this positioning of language is carried further, where Collingwood
calls discourse "the activity by which a man means anything;" it is the activity of meaning
something (a) by something else (b), where meaning (a) is an act of theoretical consciousness,
and (b) is a practical activity, the production in oneself or others of a flow of sounds or the like
10
"And we do frequently… impale ourselves on the horns of a dilemma by arguing that either the artistic attitude
towards reality is the right one and the scientific the wrong, or the scientific is right and the artistic wrong; where
right or wrong mean justified or unjustified by the unalterable character of reality itself. But the answer to the
dilemma is that both are right, and that each is wrong if it claims to exclude or supersede the other; because the
opposition between them is like the opposition of unity and plurality – an opposition in which each term is necessary
to the other. As unity and plurality are categories or transcendentals of pure logic, which means that any object of
logical thought must necessarily be thought of as both one and manifold, so art and science are categories or
transcendentals of the mind, which means that any activity or operation of the mind must have the characteristics of
art and also the characteristics of science." - The Idea of a Philosophy of Something, 1927 (IH2, 353).
which serve as the vehicle of that meaning. (NL, 6.1-6.19). Language is an abstraction from
discourse: it is the system adopted, the means employed, the rules followed in the activity of
discourse - a system of sounds or the like as having meanings, these meanings being what a
person using that word means by making that sound or gesture (NL, 6.11, 6.18, 6.41, 6.58). This
is developed into a scale of forms of thinking from conceptual thinking (selective attention,
naming, classification, evocative thinking, formation of concepts, and abstraction) to
propositional thinking (with its subject-predicate form, questioning and answering, truth and
error, etc) to rational thinking (simple knowledge, reflection, theoretical and practical reason) -
all within a "field of feeling" over which ranges "simple consciousness" in the form of attention.
In all this at the very lowest level of consciousness "meaning" is introduced. At the first stage of
mental life a man is said to be conscious of a confused mass of feeling because he has found a
language of some kind by which he can "mean" it (NL, 7.24), where "meaning" here applies
down to the level of gesture-language (included in the genus "discourse"), wherein in its
primitive function consciousness can mean something without naming it (NL, 6.1). Pre-
reflective meaning in first-order consciousness can therefore be discursive without being
properly speaking linguistic (cf. NL, 6.11).
In sum, we find Collingwood developing a theory of meaning and interpretation from his
analysis of language and mind. This has huge implications but largely unexplored for his
handling of philosophical dilemmas and his location or mapping of the logical functions of mind.
5. The climax of what turns out to be Collingwood's re-interpretation of the act-object

distinction in epistemology occurs in The Idea of History. Leading up to it Collingwood
discusses what he came to call Bradley's dilemma - that
Either reality is the immediate flow of subjective life, in which case it is subjective but
not objective, it is enjoyed but cannot be known, or else it is that which we know, in
which case it is objective and not subjective, it is a world of real things outside the
subjective life of our mind and outside each other. Bradley himself accepted the first
horn of the dilemma; but to accept either horn is to be committed to the fundamental error
of conceiving the life of mind as a mere immediate flow of feelings and sensations
devoid of all reflection and self knowledge. So conceived, mind is itself, but it does not
know itself; the being of mind is such as to make self-knowledge impossible. (IH2, 141)
Collingwood credits Michael Oakeshott with overcoming Bradley's dilemma by showing

that experience is not immediate, but contains mediation or thought within itself, so that "the real
is no longer divided into that which 'knows' but cannot be known… and that which is 'known' but
cannot know" (IH2, 152). But Oakeshott falls prey to a further dilemma concerning the dead
past and the living present, assuming there is no third alternative. Collingwood finds that there is
an "escape between the horns" by means of a the concept of a "living past which, because it was
thought and not mere natural event, can be re-enacted in the present and in that re-enactment
known as past" (IH, 158) In other words, he denies the disjunction.
Collingwood builds this into the remarkable discussion of re-enactment in The Idea of
History. There Collingwood states that a re-enacted thought, whether the original thought was
one's own or someone else's, is neither immediate as imbedded in the total flow of consciousness
(as idealists would assert) nor atomic and merely mediate (as "those who have opposed the
idealists" argue), but is both mediate and immediate - immediate in being part of the flow of
experience in a personal life and mediate in being something revived and repeated in separate
instances of being thought (IH2, 300-1). Here the dilemma is dissolved by proposing a "third
alternative" - the "double character of thought" - that denies the disjunction, which had mis-
described the mental reality. In the context of his philosophy of mind, this bi-directional aspect
of thought is what he would refer to as the "double object" of consciousness.
3. The Unity of the Autobiographical Themes
So where is the unity of the Collingwood's philosophy in all of this? What guidance do
these examples give us for solving the autobiographical dilemmas –the "found" vs. "made"
dilemma of realism and idealism, the person or thought dependency or autonomy of logic,
metaphysics as an historical science of absolute presuppositions which are neither metaphysical
nor historical; and rapprochement philosophy which leaves reconciliata unreconciled. Where is
rapprochement when we really need it?
If we were to take his autobiographical self-interpretation as final and complete, we

might have expected to find something like the following scenario carried out in his later
philosophy: The agenda of philosophy is spread out on the axis of the idealism-realism debate,
which takes the form of an epistemological dilemma concerning knowledge, its object, and an
affective relationship between them, and a metaphysical dilemma between subject and object.
The solution to these dilemmas is not by employing formal logic and ruling out one of these or
the other as an isolated proposition, but by employing Q-A logic and exhibiting the
presupposition of the question common to both positions - seeing them as attempts to answer the
same question. There would follow an historical re-enactment of the process of development of
the problem, in which it would become evident that each is necessary to the other and showing
how they form a sequential developmental pattern. Completed reconciliation would exhibit the
structure of the resolved dilemma within a dialectical scale of forms.
But this is not what a careful and honest interpreter of Collingwood will find, at least not
explicitly among the published writings. Concerning the realism-idealism debate, for example,
Q-A logic is seldom if ever in evidence. Clearly there is no treatment of dilemmas by continual
location of a common question, or the exposition of underlying presuppositions of which we are
unaware. Instead we frequently find Collingwood denying the disjunction by finding some
mediating term which the disjunction ignores, or by showing that each horn of dilemma implies
the other (is not truly disjunctional). But sometimes he accepts it as a refutation of a position,
and takes one horn over the other. In short, we find that in his actual praxis, Collingwood
displays the time-honored argumentative skills and logical strategies of the philosopher. And
while there are plenty of instances of historical analysis where idealism is portrayed as the proper
epistemological successor to realism (in one of its guises as empiricism, naturalism or
positivism), no final historical reconciliation between realism and idealism is ever carried out,
and the two positions are never fully exhibited as successors in a process of development or
coordinated concepts in a scale of forms. What we also find is the emergence of a pattern of
thought that is not made explicit in the Autobiography, and at the risk of some repetition, I would
like to take one final look at this territory.
In his earliest writings, Collingwood's philosophical touchstone is the fact and ideal of
inter-personal communication and language, and this is taken as his launching point for resolving
serious dilemmas and complete scepticism. In Religion and Philosophy Collingwood defines
personality in terms of the identity of minds engaged in acts of communication and meaning. He
distinguishes between abstract and concrete identity, portraying dilemmas as abstract and
denying that there can be anything like a concrete contradiction or a real paradox. In this same
work intentional unity is made the basis for reconstructing the "identity" of the separate
disciplines of religion and philosophy, history and science – his earliest version of
rapprochement.
In Speculum Mentis Collingwood says his "mea culpas" for his neglect of the implicit-
explicit distinction of his earlier work and then surveys the same territory with a dialectical
instrument forged from linguistic variations in the various "forms of experience" - wherein the
object of knowledge is expressed metaphorically (art), assertively (religion), as abstract universal
generalizations (science) or as narrative statements of concrete matters of fact (history), each of
which generates its own dilemmas when seen from the point of view of "dogmatic" realistic
philosophy. He then assigns the proper role of philosophy as "absolute knowledge" to
"translation" into more adequate language, i.e. language free of systematic or misleading
(realistic) ambiguities. The Essay on Philosophical Method gives us an example of what that
language would look like (its classes would overlap and form a scale of forms, it would be
expressed in categorical propositions, etc.), and indicates that in philosophy the mutual cross-
justification of starting-point by conclusion and conclusion by starting point escapes vicious
circularity only by recognizing that knowing an object is not a total revelation but rather is
knowing better something which one already knows.
In The Principles of Art and The New Leviathan, language is located within a discursive
context of total bodily expression, and both works present higher levels of mental functioning as
a progressive structuring made possible by the practical extension of intentional meaning into
language and other symbolic expressive forms. This line of thinking reaches its apex in The Idea
of History, where Bradley's dilemma - reality as either subjective or objective - is resolved by
showing that the disjunction is false. Reality is both subjective and objective, because real
thinking has a double character: it is both immediate as embedded in a context of personal
consciousness, but is also mediate in that it is capable of being revived, repeated, or re-enacted as
the content of what is thought about.
When I look at this trajectory of his thought, what I see as a unifying theme is the
growing emphasis by Collingwood on the fundamental and irreducible activity of a person's
mind as a progressively contextualized engagement with meaning in such a way that everything
felt, attended to, imagined, represented, expressed, grasped, supposed, questioned, affirmed or
denied, enacted, re-enacted, narrated, explained, communicated, recommended, instituted,
legalized, and died for, is of the nature of inalienable human meaning, and never merely an
object unaffected and utterly separate from the acts of understanding which constitute the
knowing of it. I find this to be the founding insight of his "re-enactment" argument in the Idea
of History, where "meaning" is the only presumed sense in which two minds can think the same
thought or the same mind think a thought two times in succession or continuously through time.
If that mind-mind re-interpretation of the act-object distinction based on meaning and the fact
and ideal of inter-personal communication is taken as paradigmatic for all other kinds of thought,
we have a strategy for escaping between the horns of some of the autobiographical paradoxes.
And to these we must now turn.
The first dilemma whose horns we are trying to avoid is the idealism-realism controversy.
In particular we wish to know how it is possible to preserve the integrity of an act of knowing
such that it remains true both that an object is affected by the knowing of it (“made”), and the
knowing is affected by the object known (“found”) - in short, one in which both subjectivity and
objectivity are preserved. We need an act of knowledge in which the esse of mind is not merely
cogitare simply but de hac re cogitare
Idealism and realism both tend to pit knower against known, subject against object, as if
these are anything more than "false abstractions" or entities isolated from within the overall field
of total meaning - two fixed objects defined or imagined as confronting one another. We know
that this description fails in the mind-to-mind act-object situation, where subject and object are
unifiable and even reversible. What provides a common matrix for both situations is the concrete
world of meaning. When both poles of these oppositions – knower-known, subject-object - are
located within the horizon of meaning they are no longer disjunctive isolates. Therefore when
one asks how such knowledge is possible, it is only on condition that no object in its
presentational immediacy is so unaffected by the knowing of it that it lacks meaning altogether,
since if it did, it would be in principle inhumanly alien. Nothing touching a human has that
characteristic. Similarly no knowing act is so immune from the intrinsic meaning of its objects as
to be in principle unaffected by them by a failure of shared or re-enacted meaning.
In the scale of forms of knowledge, there is no zero-point on the scale, because the
generic essence that is successively exemplified and concretized is meaning as such. We are
coming to know better (refined meaning) what we in some sense already know (presentational,
uninterpreted, pre-reflective meaning), not creating what we know ex nihilo. But the reverse
situation is also true: no act of knowledge is so independent of the object known that it cannot be
determined in its meaning by the object: the esse of mind is not simply cogitare but de hac re
cogitare. Meaning escapes between the horns of the dilemma insofar as it is an object of
knowledge that is both found and made.
The second autobiographical dilemma concerns Q-A logic as an alternative to formal

logic, and here again the dilemma is resolved by recognizing that, as Mink11 and van der Dussen
pointed out,12 what Collingwood was proposing was not a "logic" at all, at least not in the form
of a mechanical process of relating formal entities like propositions to one another in an
inferential fashion and not as related to human reasoning. “Logic” is part of an essentially
human enterprise of the sort more recently entitled "hermeneutics" or interpretation theory, with
its characteristic emphasis on human understanding, the primacy of textual and contextual
meaning, the "hermeneutic circle," and the historical rootedness of the acts of interpretation.
In his later writings Collingwood increasingly emphasized the linguistic component of

the act of knowing. He locates it within the global range of activities of a personal, active,
knowing mind. This puts him in the company of like-minded philosophers like Gadamer13 and
Ricoeur14 who have argued for a hermeneutic solution to the problems of philosophy, where
11
Louis O. Mink, Mind, History, and Dialectic: The Philosophy of R.G. Collingwood (Bloomington, Indiana
University Press, 1969), 131.
12
Jan van der Dussen, History as a Science: The Philosophy of R.G. Collingwood (The Hague, Martinus Nijhoff
Publishers, 1981), 296.
13
Hans-Georg Gadamer, Philosophical Hermeneutics, translated and edited by David E Linge (Berkeley, University
of California Press, 1976). It is interesting to note that Gadamer explicitly credits Collingwood as "almost the only
person I find a link with" concerning question and answer logic and historicity: Truth and Method (New York, The
Continuum Publishing Company, 1993), 370.
14
Paul Ricoeur, Interpretation Theory: Discourse and the Surplus of Meaning (Fort Worth, The Texas Christian
University Press, 1976).
forms of experience, categorical acts of knowing and the like are viewed as part of a
hermeneutical “horizon of meaning.”
The issue here has to do with the degree of engagement of a mental agent, a thinking
person, with the subject matter he or she is trying to understand. To the extent that it is an on-
going inquiry, the subject is actively involved and utilizes a wide range of intellectual tools,
some of which may include operations with minimal need for contextual reconstruction - e.g.
employing established formal logic or mathematics. Seen in that light, formal logic and even the
sort of dialectical logic that Collingwood works out in the Essay on Philosophical Method, are
tools taken up by understanding for the interpretation of meaning which engages the mind in the
form of questions, problems, puzzles, paradoxes, contradictions - in short, meaning-conflicts or
tears in the fabric of discourse which understanding is challenged to repair.15 But in such cases it
is still the inquiring agent who is responsible for carrying out such operations - picking up the
tool or putting it down - asking and answering these sorts of questions at any point in the inquiry.
There simply is no inquiry without an inquirer.
This answers the second dilemma: Q-A logic is not really "logic" at all, it is a form of
hermeneutics focused on the meaning of whatever "text" requires understanding, and assuming
that both the knower and the known are part of an overall context of meaning. Formal logic and
dialectical logic are both basic linguistic tools that we employ to achieve the understanding we
seek within the questioning process. They are only as "automatic" as we allow them to be, since
we grasp them for the handiness they have in dealing with the meaningful world as logically
structured. Q-A logic is unintelligible and inconsistent as an alternative to formal-propositional
logic, but not as a hermeneutic methodology of an understanding agent.
We are on less secure grounds on the third and fourth autobiographical themes – the
philosophy of history and rapprochement, but we can see an escape route between the horns of
the absolutist-relativist dilemma posed by the Essay on Metaphysics by suggesting a more
flexible meaning for "absolute presuppositions." If an absolute presupposition rather than being
a quasi-propositional entity in a formal system is regarded as an orienting concept for a knowing
subject's placement within the horizon of meaning of a systematic inquiry, the dilemma is
resolvable.
I say "resolvable" rather than "resolved" because there are aspects of the ontology and
epistemology of this placement which are still problematic. It is still not clear what sort of entity
an absolute presupposition is, or what overall status it has concerning knowledge acts, or even if
it is a really describable in such terms at all. Concerning history as a concrete endeavor, the issue
15
In an experimental essay Collingwood goes so far as to say that the basic function of understanding is "seeing
continuities." Reality as History, dated December, 1935 (PH, 184).
is critical. The philosophical problem is with providing an ontological anchor or grounding for
locating any such historical interpretative act of understanding, or in epistemological terms
establishing an actual "reference" which has intrinsic connection with the possible "meaning" or
"sense" or "thought" corresponding to it. Using other clues that Collingwood leaves, one might
describe absolute presuppositions as pre-assertive entities functioning as orientational anchors
for some entire process or sequence of understanding, establishing what will ultimately count as
a meaningful answer to questions within that inquiry. But that grounding must be historically
rooted, for without it two people would not be able to share in the same inquiry, and the same
person would not be able to consistently pursue it over time.
Here Collingwoodians may find helpful the recent work in what he calls "cognitive
science" by the American philosopher, John Searle.16 Searle describes meaning as the name we
give the practical act of imposing intentionality on an expressive but plastic medium like
language (speech acts or some other symbolic system). Just as intentionality is the directedness
of mind to an object, speech acts are a representation of intentional mental acts. Each intentional
state with a "direction of fit" has conditions of satisfaction. Searle describes only a "mind-to-
world" or a "world-to-mind" or "null" directions of fit, which fails to capture what Collingwood
would call the "mind-to-mind" direction of fit. We wonder if Searle would regard absolute
presuppositions as having a "null direction of fit," which would require metaphysical analysis to
determine its "conditions of satisfaction."
Finally, what is not clear in all this is the overall context within which one or any number
of minds could be said to have achieved a final reconciliation overcoming the paradox of their
estrangement from one another or even an estrangement within the same mind between various
forms of experience understood as intersecting horizons of meaning.
This brings us directly up against the fourth and final autobiographical dilemma. Here
Collingwood's treatment of absolute knowledge as a form of translation is interesting, as is the
issue of the circularity of interpreting the part in terms of the whole and the whole in terms of the
part - the "hermeneutic circle." Seeing the meaning of the part in terms of the meaning of the
whole and the meaning of the whole in terms of the meaning of the part is only not circular
insofar as in each of these movements - from part to whole and back again - involves shifting
one's location between them within the horizon of meaning, and in the process get to understand
better that which we already know. Is it possible, then, that reconciliation is this circular weaving
process by human understanding, restoring the continuity in a torn fabric of meaning? And is it
possible that there could be a “space of perspectives” (Collingwood’s version of the “fusion of
16
John Searle, The Philosophy of Mind (The Teaching Company, Springfield, 1998), Lecture Nine: How to Study
Consciousness Scientifically.
horizons”) between persons such that further questions are unnecessary, as expressed in a
language free of systematic errors? 17
Here again we would like to follow up on a fertile suggestion by John Searle. Searle
argues that a science of mind or intentionality requires an explicit acknowledgement that any
particular intention (and a forteriori its expressed meaning in language) is not understandable as
an atomic or isolated unit, but rather requires a large number of other such intentions in a context
he calls "holistic."18 I believe this is precisely what Collingwood was describing in The New
Leviathan when he wrote about the "context of evocations" within which acts of intellectual
abstraction occur (NL, 7.32-7.33). By relocation of a meaning in its originative context of
evocations we restore the part to its context in the whole, and this is a question-and-answer
process.
If achieving an "absolute standpoint" requires that there be a final end to this process, it
would seem that for Collingwood, this is not possible. All that one can aim for is the continuing
efforts at "translation" uniting people of intelligence and good will. How would one know when
this is successful? Only by engaging in further acts of meaningful communication, to see what
questions arise. But in doing so, one must be willing to be surprised by what happens next -
when object-meanings shape further acts of knowledge and understanding.
So even if it is still incomplete, the evidence seems to me to be persuasive that for

Collingwood, reconciliational philosophy is achieved by a hermeneutic approach to experience.
If this is so, we have solved the last of the Collingwoodian dilemmas. It is easy to find texts of
Collingwood where he appears as either a relativist or an absolutist, but in the end he escapes
between the horns, because the goal of his philosophy was to repair the torn fabric of meaning
wherever it occurs. One does not know where such discontinuities occur until questions arise;
and when these questions take the form of a dilemma, philosophical rapprochement is necessary.
4. Conclusion: Collingwood for the 21st Century.
In this paper we have attempted to review how Collingwood’s autobiographical self-

interpretation holds up to his actual practice as illustrated by his handling of philosophical
dilemmas in his published writings. We have found that what the evidence reveals is contrary to
what we might expect from a literal reading of the Autobiography. His treatment of dilemmas
follows traditional modes of philosophical argumentation, and does not reveal the marks of Q-A
17
Collingwood uses the term "perspective" in the manner that hermeneutics uses "horizon," and in several passages
he makes reference to the historian's "space of perspectives" and "the problem of relation between perspectives" -
see EPH, 53-56 and PH, 128-29.
18
Searle, Ibid.
logic as he describes it. And there is no final resolution or rapprochement between realism and
idealism, nor is there an explicit absolute presupposition to serve as a context to resolve the
controversy.
But careful attention to his treatment of dilemmas also reveals a subtle direction to his
thinking that is recognizable to anyone familiar with the continental philosophical movement
known today as “hermeneutics.” Without carrying this analysis forward beyond the themes of
the Autobiography, we have suggested that this approach may help to resolve many if not all of
the dilemmas that Collingwood has left as part of his own legacy to the next generation of
philosophers.
So I suggest that the way to pull Robin George Collingwood through the 20th century
speculum and into the 21st century, to satisfy our present need for a philosophy, is to further
explore and expand the unifying aspect of this hermeneutic dimension of his thought. In so
doing we may be able to complete the reconciliation that he left as part of his unfinished
heritage.
Chicago, June 24, 2001
*An earlier version of this paper was presented at the Collingwood Conference held at St.
Catherine’s, Oxford in July of 2001. It was subsequently revised and published in the journal,
Collingwood and British Idealism Studies,
A Review of
Collingwood’s Essay on Metaphysics – 1998 Revised Edition.
Edited with Introduction by Rex Martin
by Glenn Shipley
A. The Reviewer’s Job.
It is the duty of a book reviewer not only to provide the reader with a preliminary
overview of what he will encounter upon reading the actual text, but to recommend, forewarn,
or even suggest restructuring the reader’s engagement with that text. In what follows I wish to
be clear that this new edition of the Essay on Metaphysicsi is an important publishing event for
students of Collingwood, and I wish to express my respect and gratitude to Rex Martin for the
scholarly work he has done in preparing it for the reading public. For reasons that will become
plain, however, I hope to caution the reader about relying too heavily on the introductory essay
as a guide to what follows. In the sequel I will recommend a different order for reading the text,
which I believe makes the work more approachable than it is in its present form.
Two further remarks about the scope of this review. Rather than embark on an overview
of the entire Essay on Metaphysics, we will focus on the newer material of the Revised Edition,
and to its Editor’s Introduction. Secondly, since even a casual reader of Collingwood’s Essay on
Metaphysics is aware that its two central issues are absolute presuppositions as the primary
subject matter of metaphysics, and the claim that metaphysics is an historical science, in what
follows we will concentrate on these two issues and what Martin wants us to understand about
them.
B. What OUP Hath Wrought.
In issuing a revised and expanded version of the Essay on Metaphysics (EM2), the
delegates of Oxford University Press (OUP) have continued the tradition of both honoring and
obscuring the late, great Waynflete Professor of Metaphysical Philosophy, R.G. Collingwood.
They do so by publishing and keeping his original works in print, while simultaneously adding
further layers of interpretative mystique to his intellectual legacy.
In the present case, the delegates have utilized the services of a well-qualified editor, Rex
Martin, who not only presents new and previously unpublished material to the interested public,
but also offers an interpretation that adds both light and shadows to the portrait of Collingwood’s
legacy. Thus the delegates have continued what was begun a half-century ago, when T.M. Knox
was chosen to edit the manuscripts that were subsequently published as The Idea of History and
The Idea of Nature, and now continues with The Principles of History. These works have
greatly expanded Collingwood’s influence and made available works that have continued to
enlighten, delight, and perplex several generations of readers. So also the Essay on Metaphysics
in its expanded form offers a view of metaphysics that is well on its way to becoming a classic in
its own right, but is overlaid with an interpretation by its editor that is itself not free of
controversy.ii
The bulk of this new edition is still the whole of the original published text of the Essay
on Metaphysics with a few corrections, some of which were indicated by Collingwood in his
own marked copy of the published work. Readers familiar with the original work will find the
same set of topics in the same order: a general overview of the nature of metaphysical study,
beginning with Aristotle; the rejection of metaphysics as an ontology of pure being; metaphysics
as an inquiry concerning the absolute presuppositions of science; inquiry as the exercise of
question-and-answer (Q-A) logic; the reform of metaphysics as an historical science; positivism
and psychology as self-refuting forms of anti-metaphysics; and three specimens of metaphysical
analysis. As in the original, there is no detailed analysis of Bradley’s Appearance and Reality,
practically no discussion of Hegel (8 page references compared to 46 for Kant), no discussion of
mind and its relation to nature (as in the conclusions to his lectures on the subject in 1934, 1935,
and 1937),iii and no explicit connection to his other writings - neither to earlier works like
Religion and Philosophy and Speculum Mentis, nor to the mature positions he had staked out in
the Essay on Philosophical Method and The Principles of Art and the lectures which
subsequently formed the basis for the posthumous publications, The Idea of History and The
Idea of Nature.
Some of these issues are addressed in wholly new materials added to the original text.
Bracketing this text are a 103-page section at the beginning, including Martin's 4-page Preface
and 57-page Editor’s Introduction with notes and a bibliography, and at the end a section which
includes a valuable and accurate set of notes about the composition and publishing history of the
original book. This 84-page “Endnotes” section includes previously unpublished texts,
interspersed with Martin’s helpful comments: "The Nature of Metaphysical Study” (NMS); and
“Function of Metaphysics in Civilization” (F) - the former being the second of two lectures
Collingwood delivered “opening a course of 16 lectures on metaphysics by various speakers, to
be delivered 15 and 17 January, 1934” – five years before completing the manuscript of the
Essay on Metaphysics (April, 1939), and the latter a lecture delivered in 1938. It also includes
the complete letter of A.D. Lindsay, which resulted in Collingwood’s remarkable footnote on p
48 about conceptual change. Martin handles all of this material in a manner which displays his
craftsmanship and respect for the subject matter and its author.
But while the reader is grateful for the appearance of all this new material, one cannot
help but wonder about the peculiar bundling that seems to be occuring. While there was a strong
need for an editorial introduction when the original Idea of History was published because
Collingwood did not live to see this work through the press, this is not true for the Essay on
Metaphysics, which he finished in 1938 and regarded as a complete work (EM2, xviii). Treating
the completed Essay on Metaphysics as if it were an unfinished work requiring additional
materials and an extensive introduction like The Idea of History may thus prove misleading to
the reader.
So in spite of Martin’s explanation (EM2, lxviii) the reader is still curious about the
justification for the choice of materials included in this expanded edition – just as he was puzzled
about Knox’s choices for IH and IN. For example in the “Endnotes” section Martin includes a
wonderful essay entitled “The Nature of Metaphysical Study,” containing an interesting analysis
of F.H. Bradley’s Appearance and Reality, which Collingwood apparently studied with some
care. But then Martin teases the reader (in “A note on provenance” – pp. 376-78) with the
remark that “Bradley’s philosophical contribution is elaborated at greater length in another of
[Collingwood’s] unpublished manuscripts, entitled ‘The Metaphysics of F. H. Bradley: An
Essay on Appearance and Reality’ (dated Christmas, 1933).” As in the case of Knox’s selection
of a completely unsatisfying conclusion to The Idea of Nature, one wonders why Martin agreed
to publish this essay on Bradley, and not the evidently more thorough one from 1933 – or better
yet, both of them. And one wonders why the Delegates of OUP agreed to couple the two
conclusion to Collingwood’s lectures on Nature and Mind with their publication of The
Principles of History rather than a revised edition of The Idea of Nature or perhaps even a
separate volume consisting of these and other of Collingwood’s unpublished manuscripts.iv
C. What Rex Martin Hath Wrought

The answer to some of the questions about selection and omission of textual materials is
addressed by the editors of the posthumous publications. Increasingly it appears to lie not with
Collingwood’s instructions and wishes, but with considerations of editorial interpretation. In
sections 2-6 of his introductory essay to this new edition of the Essay on Metaphysics Martin
struggles with the same difficulty manifested by most other Collingwood scholars: how to
separate what is properly said about the issues that Collingwood was addressing from what is
specified in Collingwood's own presentation of those issues during his productive lifetime. With
respect to the former, one is free to say whatever a commentator is prepared to defend; with
respect to the latter, the reader requires convincing evidence from the texts, and the interpreter is
(or should be) bound by the rules of responsible interpretation.v
And herein lies the interpreter’s field of battle. Where many less cautious commentators
would have tripped over one of the many land mines that Collingwood has left in the war zone of
his later writings, Martin skillfully and bravely rises to the occasion by offering an interpretation
which makes use of Collingwood’s own roadmaps rather than the far riskier plan of using a
ready-made one provided by continental or analytic philosophy, or an antiquated one left by
Hegel or even Croce, or one designed for an entirely alien territory by Ryle or Wittgenstein (or
Nietzsche). Martin is further to be commended for prudently choosing as his primary guide the
Essay on Philosophical Method, which Collingwood himself called his best book and the only
one he had time to finish as well as he knew how (A, 118), and one which he clearly intended (at
least at one time) as a companion piece to the Essay on Metaphysics.
But that does not mean that Martin emerges from the minefield unscathed – that would be
expecting too much. One would have a better chance finding a formula for lasting peace in the
Balkans. Martin even risks scholarly life and limb by including previously unpublished
manuscripts, parts of which are in direct conflict with the interpretation he is offering – thereby
virtually ensuring that he will return from the frontlines, bandaged and limping from assorted
injuries.
Also to his credit, Martin’s Introduction exhibits both a sincere effort to present the
precise details of the composition of the materials subsequently published as the Essay on
Metaphysics, but also an interesting set of observations about the issues Collingwood was
addressing, and suggestions for what approach the reader can take for assimilating that material.
After (1) an enlightening initial section discussing how EM came to be written, Martin devotes
four sections to the topics respectively of (2) absolute presuppositions (APs), (3) metaphysics as
an historical science, (4) the character of metaphysics (as displaying the structure of science as a
scale of overlapping forms), and (5) the tasks of metaphysics (assessing comparative scientific
progress). A final section (6) is devoted to a detailed analysis of Collingwood’s discussion of
causation. Of the three examples of metaphysical analysis offered by Collingwood in EM,
Martin calls the section on causation the only true example of present day metaphysical thinking
(EM2, liv), and so it is the only one for which he offers a commentary.
If I understand him correctly, Martin wants to soften some of the harsher statements
about these primary concerns presented in EM by overlaying them with an interpretation which
isolates Collingwood’s theory of absolute presuppositions from his radical exposition of
question and answer logic (Q-A logic), and separates the concept of metaphysics as an historical
science from the procedure of historical re-enactment.
With respect to the former, Martin convincingly argues that until the appearance of the
Autobiography there is no evidence that Collingwood linked APs to Q-A complexes or to an
organized Q-A logic, even though the concept and function of APs have a much longer history in
Collingwood’s writings – a history independent of the Q-A logic he espoused in both A and EM.
This later radical linkage of APs to Q-A complexes, in fact, would render them not only true-
false neutral, but also meaningless.
With respect to the latter (metaphysics as historical but not as the exercise of re-
enactment), Martin argues that re-enactment only properly applies to deliberate human actions,
whereas APs (as somewhat “unconscious” entities) are not really actions in the true sense of the
term. APs are best understood in the light of the doctrine of a “scale of forms” which
Collingwood had worked out in the Essay on Philosophical Method. In such a scale the higher
forms incorporate and exemplify the lower forms in such a way that differences in kind are
inseparable from differences in degree, so that a constellation of APs at its higher
implementation – in a body of scientific inquiry – will not only differ qualitatively from its lower
manifestations (as Newtonian science differs from Aristotelian, or Copernican astronomy differs
from Ptolomaic), but will also be more adequately what the lower form was inadequately.
Martin links this view of APs with a treatment of historical understanding which views
the object of history not as a spectacle of isolated events and evidence, but as activity or a
process of cumulative development (EM2, xxv). Since history is concerned with processes
(activities) which exhibit such a cumulative development, APs can change and be subject to
analysis which demonstrates their increasing adequacy – and hence, in some sense, their “truth.”
This, Martin argues, allows Collingwood to claim that metaphysics is an historical science
without abandoning its aspect as criteriological. But Martin asserts that Collingwood is not
saying that what a metaphysician is doing is to engage in what he had in IH famously called “re-
enactment”. Martin goes so far as to say that “the idea of reading re-enactment into
Collingwood’s account of metaphysics is largely off the point” (EM2, xxxi, repeated at xxxii).
Minimizing this particular historical dimension to metaphysical thinking would allow one to
maintain the criteriological or normative aspect of metaphysics as philosophy, and thereby avoid
the “demons of scepticism and relativism” that so horrified Knox (EM2, liii, repeated at liv).
D. Problems with What Rex Martin Hath Wrought.
It is at this point that the reader may become suspicious. Encouraged by Collingwood to
turn to original sources to check for oneself if the criticisms are valid, one finds that in his
Preface to the first edition of the Idea of History Knox repeatedly charges Collingwood with the
twin problems of scepticism and dogmatism (IH, xi, xv, xvii, and xix). While the third issue of
"relativism" is raised by Knox (IH1, xiii), Martin assumes the linkage to "skepticism," as if
equating, for Knox, dogmatism with relativism. In the preface to IH1, Knox cites Collingwood’s
admonition to adopt an attitude of “unquestioning acceptance” toward APs, which Knox argues
is a surrender to the dogmatic strain in Collingwood’s philosophy present ever since his youthful
composition of works like Religion and Philosophy. Furthermore the quality of APs as neither
true nor false signaled for Knox the sceptical side, since it denied there was any criterion for
deciding the truth or falsity at the very foundations of scientific inquiry.
Now this might appear as a trivial mis-emphasis on Martin’s part, but it heightens the
attentiveness of the cautious reader, and it is not the only problem I encountered with the Editor's
Introduction. My overall discomfort has to do with the proper limits of responsible
interpretation. There is no problem with finding alternative approaches to the subject matter in
EM, and surely there is merit in Martin’s suggestion that the reader approach the issue of
metaphysics as the science of absolute presuppositions without paying undue deference either to
its controversial “question-and-answer (Q-A) logic” or to the kind of historical thinking specified
by Collingwood in his discussion of historical "re-enactment." But it is another thing altogether
to say that these recommendations conform to Collingwood's settled view of the matter. As we
shall see, they are particularly difficult assertions in view of Collingwood’s explicit statements
that re-enactment is the very condition of the possibility for historical thinking to take place at all
(so that removing re-enactment would completely eviscerate high-grade historical thinking), and
that the metaphysician cannot evaluate one system of APs as better than another.
Now it would be hard to charge Martin with failing to understand either Collingwood’s
Q-A logic or his analysis of history as an exercise of imaginative re-enactment, given not only
Martin’s familiarity with all the published and unpublished manuscripts but also his careful
analysis of re-enactment in his book, Historical Explanation: Re-Enactment and Practical
Inference. Nonetheless, given the entire body of Collingwood’s work as published to date, his
solution is still open to question. As representative of Collingwood’s currently published
thought, Martin’s suggested reinterpretations appear puzzling at best, and perhaps ultimately
insupportable.
(1) Meaning and Truth in Q-A Logic. Martin correctly argues that by identifying APs as
non-propositional insofar as they are defined only by means of the P-Q-A complex in Q-A logic,
APs not only lose the ability to be true or false, but also lose their meaning. The problem here is
that Martin fails to call the reader’s attention to the subtle shift in the two extant versions of Q-A
logic - as presented “for the first time” in the Autobiography (Q-AA) and Q-A logic as it appears
in the Essay on Metaphysics (Q-AM). In point of fact there are significant differences between
these versions, not the least of which has to do with the very issue Martin raises. The reader
approaching the present work without having first read the Autobiography would find Martin’s
assertion puzzling, since it is only in Q-AA that the P-Q-A complex is declared to be the unit not
only of truth, but of meaning and validity (contradiction). In the quasi-axiomatic presentation of
Q-A logic in the Essay on Metaphysics there is no mention at all of the P-Q-A complex as being
the minimal unit of meaning and logical validity.vi
This is no small matter, since the whole purpose of Martin’s separation of APs from Q-A
logic is to avoid the situation in which an AP would not only be T-F neutral but also
meaningless. If it should turn out that Collingwood could successfully maintain by means of a
Q-A logic that an AP could both be T-F neutral but yet retain meaning, there would be little
point in this aspect of Martin’s reconstruction.
(2) Absolute Presuppositions Not A Priori Concepts. Martin flatly rejects the
interpretation of the truth-neutral property of APs as being due to their status as a priori concepts
rather than propositions (EM2, xxvii) - in spite of the fact that there is fairly widespread
agreement by many Collingwood interpreters on this pointvii. There is no convincing argument or
counter-evidence to demonstrate this assertion.viii Certainly no one expects that all bases can be
covered in a 100-page introduction, but this is another key issue concerning the status of APs
and metaphysics, so it can reasonably be expected that it would be rigorously defended.
Furthermore two of the three examples that Collingwood examines in EM, Part III,
suggest that APs are conceptual entities. The concepts Collingwood examines are God,
causality, and the principles Kant discusses in the transcendental analytic section of the first
Critique. Of these Martin chooses only to consider the concept of causality in his Introduction,
claiming that of those examples proposed, only Collingwood’s analysis of causality is an
example of contemporary metaphysics. Whereas Collingwood is proposing all of these concepts
as specimens of metaphysical analysis, no satisfying argument is offered for Martin’s restriction
of these topics to causality alone. Would not a discussion of the Kantian principles have
solidified his argument that APs are proposition-like entities rather than concepts?
(3) Absolute Presuppositions and the Scale of Forms. It is ironic that while Alan
Donagan totally neglects definition by means of a scale of forms (SF) in discussing APs,
Martin’s tendency is to go to the other extreme. On the grounds that the expression “scale of
forms” nowhere occurs in EM, but occurs 20 times in Martin’s Introduction to EM2, one might
be forgiven for leaning towards Donagan on this issue. Martin wishes to dissociate the Q-A
logical treatment that Collingwood offers in EM by substituting for it a definition of APs by
means of SF, without sufficient recognition that in EPM the discussion of the SF occurs in the
context of the traditional epistemological structure of concept-judgement-implication (paralleling
the logical term-proposition-inference schema) of the whole of EPM, and occurs in the section
that deals with concepts or terms, not judgements or propositions. In other words, a SF applies
to philosophical concepts that are contrasted with classes – not primarily to whole theories,
which are more at the level of a complex of propositions related in a scientific exposition.ix
(4) The subject matter of re-enactment. Martin asserts that re-enactment is restricted to
actions of agents, and is not properly applied to beliefs per se (EM2, xxxi). But this ignores the
fact that in IH Collingwood’s prime example of historical re-enactment is an axiom of Euclid’s
geometry (IH 284-85). Now one might argue that Collingwood’s examples are not always
felicitous, but that does not explain why he would choose such an example if the application of
the principle he was trying to explore fell entirely outside its province. It would be hard to
portray an axiom of geometry as the action of an agent in Martin’s restricted sense, and not a
“belief.”
Martin leaves the reader a huge escape-hatch when he later asserts that “unless absolute
presuppositions are actions or can be the more or less direct effects of actions… there is little
point in making re-enactment the focal point of any Collingwoodian assimilation of metaphysics
to history” (EM2, xxxii). Little point, that is, unless one were to take into account that for
Collingwood all thought is action or the result of action (cf. NL, 1.61-1.66, 5.91, 6.19) – even
so-called “unconscious” thought, and that Collingwood argued that re-enactment is the very
condition that makes all historical thought possible (IH, 282).
But even if he were right on this matter, there would then be little point in Martin’s
equating metaphysics qua historical understanding with APs , since even on his own grounds
“mind-as-thought… is completely involved in process” (EM2, xxxv). So if one were to accept
Martin’s approach to them, APs would have to be mental entities, therefore involved in
processes, therefore activities of subjects, and therefore historically re-enactable. Other writers
have confirmed such a linkage, particularly in connection with the mental activity of language.
For example, David Boucher quotes from Collingwood’s “Observations on Language” (an
unpublished mss) that: “language is a mode of conduct, an activity;” that meaning is an essential
attribute of language as activity; and that words have no meaning except as used in context by a
speaker (SPTC, 137). Further, Peter Lewis points out that “For Collingwood and Wittgenstein,
language is characterized, first and foremost, by activity” (CS V, 35). Is it Martin’s position that
APs are not mental, not meaning, and not linguistic? And if he admits that they are, must he not
also admit that they are actions, and as such re-enactable by historical understanding?
(5) Process and progress: Not only in the original EM but also in “The Nature of
Metaphysical Study” which Martin includes in the Endnotes of EM2, Collingwood explicitly
denies that a metaphysician can show that one school of science or one metaphysical system
(presumably of APs) can be demonstrated to be “better” than another (EM, 52-53; EM2, 389-
94). Martin notes this as a “discontinuity” with the later published work (EM2, lxxiv), but the
bridging argument that he offers fails to make the link between the two senses of “right” that
Collingwood proposes for an AP - as a fitting description of what someone actually presupposes
in their thinking, and as one that is better than another as a description of what a superior science
presupposes over its predecessor on a scale of forms. Failing this, how can Collingwood, much
less Martin, successfully defend that metaphysics is normative or criteriological? It appears to
be capable of historical reportage only.
To secure this point, we need to take a minute to examine what Collingwood actually has
to say about the matter in the “Function of Metaphysics in Civilization” (EM2, 379-420). “The
question will certainly be asked, whether, even so, one kind of metaphysical system may not be
better than another,” Collingwood writes, and then sets out two stages in answering the question.
In the first (the intuitionist approach, which would claim to answer it by simply inspecting them
to see which of the two is better), he argues that in such a comparison there is no real standard,x
since the metaphysician is incapable of standing on a neutral ground where both appear equally
inspectable. “Working on this assumption” (he writes) “I can see no way of answering the
question definitely….” (EM2, 391-2). In the second alternative he presents the case as one of
choice: “deciding which of two things is the better is the same as choosing between them.” But
in this form the question is meaningless, since “It could have meaning only if there were
somebody who had before him the option of pursuing whichever kind of science he chose.
There is no such person. The question which kind of science is the better, therefore, is a
nonsense question because to ask it presupposes the existence of a situation which does not
exist.” Collingwood goes on in the same passage to say that it is meaningless to ask which
metaphysics is intrinsically better than the other, the Aristotelian pluralistic metaphysics or the
Kantian monistic metaphysics – as it is meaningless to ask if the system of science resulting from
them, or even the sciences themselves, are better than the other. (EM2, 393-94).
Whether or not one accepts this argument as valid (and I do not),xi nevertheless to this
reader, this passage and one’s similar to it in the original EM represents an immovable obstacle
to Martin’s recommended re-interpretation of Collingwood’s agenda for metaphysics as an
historical science – i.e. as a non-re-enactable, analytic, historical study of the “processes” of AP
transformation, viewed as a progressive scale of forms. It is a wonderful recommendation, but it
simply goes beyond Collingwood’s last words on the subject as we have them.
D. Rending Asunder What Rex Martin Hath Wrought.
Martin addresses some of these issues in the last section of the Introduction, where he
offers a detailed commentary on the unpublished mss which form the Endnotes section of EM2.
By then, the credulity of anyone familiar with Collingwood’s published writings has been
stretched to the breaking point. I suggest, therefore, that readers approach EM2 in an order
different than the one Martin and OUP have presented. To this reviewer, given the problems with
Martin’s Introduction outlined above, it makes more sense to read EM2 in the following order:
1. Martin’s Introduction, Section 1: The Composition of the Essay on Metaphysics (pp.
xv-xxi).
2. Martin’s Endnote A: The publishing history of the Essay on Metaphysics (pp. 344-
48); Endnote B: Emendations made by Collingwood himself (pp. 348-50); and
Endnote C: A Letter from A.D. Lindsay (pp. 350-55).
3. Collingwood’s Essay on Metaphysics (the original published text).
4. The three unpublished pieces by Collingwood.
5. Martin’s Introduction, Sections 2 through 7.
The reason for this suggested rearrangement is that Martin’s thesis makes a great deal
more sense (even if problems remain about accepting it as final) after first understanding how
EM came to be written, then reading the work itself, and then seeing how it appears in the
context of Collingwood’s other works in progress on the same issues. Finally, it is only after
surveying all this material that one can reach conclusions even remotely similar to those Martin
proposes – just as Martin himself did. He cannot spare the reader’s discomfort with this
difficult-to-assimilate work by predigesting it in his Introduction – it simply does not work.
I would also recommend that the reader spend some time reading over the two
conclusions to the lectures on nature and mind, presently published with The Principles of
History.
Finally, I offer this as a personal concluding note: it seems to me that, like
Collingwood’s experience of the Albert Memorial, there remains something irreducible and yet
indigestible about the set of ideas represented in the Essay on Metaphysics. In it Collingwood
recognized that, just as in a work of art, there is something trope-like about the relationship of
APs and the science based on them, but he had difficulty analyzing and expressing that
relationship. His account of them on the basis of Q-A logic appears to be as close as he came to
expressing the insight that an AP is bound in a logical and unifying way to an entire body of
propositions in a scientific inquiry, but is yet itself non-propositional. But while the suggestion
to amplify this relationship in terms of a scale of forms may add significantly to the
understanding of it, simply assimilating APs to this somewhat abstract and still-obscure doctrine
from EPM does not succeed in accounting for it or even adequately expressing it. A complete
discussion of APs would require locating them not only in a logical frame of reference (and
perhaps more than one, since Collingwood recognized not only formal logic but also dialectical
and Q-A logics), but also in a schema which displays their status and structure as ontological,
epistemological and psychological entities.
It also seems to me that Collingwood recognized metaphysics as a unique form of inquiry
which has links both to Aristotle’s “First Philosophy” and to a more contemporary view of
historical understanding (which requires a radical reformation of the original subject matter), and
that the object of this inquiry is a kind of presupposition not affirmed in any conventional or
conditional manner, but held “absolutely” in some sense of that much-maligned word. But to
divorce this understanding from the sort of thinking characterized by Collingwood as re-
enactment is to remove the very process of thought by which a metaphysician would assimilate
the subject for himself.
What this metaphysical inquiry actually is – even after the sixty years since the Essay on
Metaphysics first appeared – remains inadequately described and unsatisfactorily explained.
i
In this review the generally accepted abbreviations are used for Collingwood’s published writings and those of
his major interpreters. An exception is EM2, which I propose for the Revised Edition of the Essay on
Metaphysics in order to distinguish it from the first edition. On the same basis I also suggest IH2 for the revised
edition of the Idea of History, particularly important since IH2 omits Knox’s original Preface and adds an
Introduction by Jan van der Dussen. In addition to these abbreviations, Q-A stands for Question and Answer,
and AP for Absolute Presupposition. FRTR refers to From Realism To Rapprochement: The Autobiographical
Interpretation of Collingwood’s Philosophy – my 1983 doctoral dissertation at Loyola University of Chicago,
available free of charge to anyone who asks for it (send email to glennshipley@comcast.net or
sophologist@poetic.com ).
ii
By comparing Rex Martin’s Introduction with the Preface by T.M. Knox to The Idea of History, I do not mean
to attribute to Martin the same sort of misunderstanding, bias and irrelevancy that Knox imposed on
Collingwood studies.
David Boucher, “The Principles of History and the Cosmology Conclusion to the Idea of Nature,”
iii
Collingwood Studies II (Swansea, 1995), p. 152.
In their introductions to the works in question, Martin, Dray and van der Dussen address the problem of the
iv
unpublished mss and provide reasons for including the manuscripts that they did. The question I am raising has
more to do with exclusions and overall organization of textual materials in their present arrangement.
v
I assembled a set of these in my dissertation, FRTR, Chapter I, pp. 44-46.
vi
Shipley, FRTR, Chapter II, table 3 and pp. 84-89; Chapter VIII, table 9, and pp. 430-458.
Cf. Mink, MHD, 144-48; Rubinoff, CRM, 234-35; Toulmin in Krausz, CEPC, 205-08; and Shipley, FRTR,
vii
802, n 10.
Later in the Introduction (EM2, lxxv) , Martin does make the point that APs are “assertion-like” entities, and
viii
he also acknowledges something like an a priori status for them inasmuch as they are adopted independently of
experience. But I do not see where he presents any argument or evidence against the position that as “non-
propositional” and as mental entities, APs might actually be something closer in structure and function to
concepts (which have discernible meanings, but are not appropriately called propositionally true or false) than
they are to judgements. Furthermore as “descriptive” framework entities, regarding APs as definable concepts
rather than as verifiable propositions or judgements, seems to point in that direction.
ix
In fairness to Martin, at EM2, xl, he does examine the issue of whether there are limits or modifications that
must be made when applying a SF as a definition of an AP. But even as articulated in this discussion, I do not
see how it can escape Collingwood’s own restriction on the non-comparability of various systems of APs.
x
Collingwood makes a similar case against Sir W.D. Ross concerning a scale of forms of goods, on the grounds
that they lack a common standard for measurement. See Boucher, EPP, 34.
xi
The reason I do not is because a de facto lack of instantiation does not resolve a de jure issue of principle.
Collingwood has not demonstrated that it is not possible for someone to choose which science he engages in. To
say, as Martin does, "That is not the way things happen" (EM2, xlviii) fails to wring the withers of someone who
might argue that there is nothing preventing it. In fact it would be hard to understand how science could
significantly change if for someone the process was not stipulatively determined. When Newton wrote the
Principia he was not merely reflecting how the science of his day was done, he was saying how it ought to be
done and providing an instance of it.

FRTR & BH & Rem2

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

FRTR & BH & Rem2

Hochgeladen von

Copyright:

Verfügbare Formate

FROM REALISM TO RAPPROCHEMENT:

THE AUTOBIOGRAPHICAL INTERPRETATION

In writing this dissertation I have received support and

I owe a greater debt than words can express to my wife,

The author, Glenn C. Shipley, is the son of Herman

He graduated from St. Alphonsus Grammar School in 1952,

He began graduate study in philosophy at Loyola

LIST OF TABLES . . . . . . . . . . . . . . . . . . . .. viii

LIST OF IN-TEXT REFERENCES . . . . . . . . . . . . . .. ix

I. INTRODUCTION: COLLINGWOOD AND HIS INTERPRETERS

1. Collingwood: His Life, His Writings, and His Era. . . 1

II. THE AUTOBIOGRAPHICAL INTERPRETATION OF COLLINGWOOD'S

1. The Autobiography as Literary Evidence . . . . . . . 53

III. REALISM AND IDEALISM

IV. QUESTIONS AND ANSWERS, LOGIC AND DIALECTIC

V. HISTORY AND PHILOSOPHY

VI. RAPPROCHEMENT, RELIGION, AND ABSOLUTE KNOWLEDGE

VII. ANTI-REALISM AND THE PHILOSOPHY OF MIND

VIII. LOGIC, LANGUAGE, AND MENTAL ACTS

IX. THE PHILOSOPHY OF HISTORY

X. METAPHYSICS AND RAPPROCHEMENT

1. The Interpretations of Collingwood's Philosophy . . . . 41

2. The Critique of Realism . . . . . . . . . . . . . . . . 77

3. Question and Answer Logic . . . . . . . . . . . . . . . 84

4. The Principles of History . . . . . . . . . . . . . . . 91

5. Rapprochement Philosophy . . . . . . . . . . . . . . . 103

6. The Meanings of History . . . . . . . . . . . . . . . . 240

7. The Knower-Known Relationship in Perception . . . . . . 388

8. The Philosophy of Mind . . . . . . . . . . . . . . . . 419

9. Question-and-Answer logic in The Essay on Metaphysics . 430

10. Philosophical Method . . . . . . . . . . . . . . . . . 463

11. Language and Expression in The Principles of Art . . . 520

13. The Presuppositions of the Analysis of Mental Functions 557

14. Metaphysics and Anti-Metaphysics . . . . . . . . . . . 747

In this dissertation references to Collingwood's pub-

A An Autobiography. London: Oxford University Press, 1939.

EM An Essay on Metaphysics. Oxford. The Clarendon Press, )97~.

EPA Essays in the Philosophy of Art. Edited by Alan Donagan.

EPH Essays in the Philosophy of History. Edited by William

EPM An Essay on Philosohical Method. London: Oxford University

FR Faith and Reason: Essays in the Philosophy of Religion.

IH The Idea of History. T. M. Knox, editor. London: Oxford

IN The Idea of Nature. T. M. Knox, editor. London: Oxford

NL The New Leviathan. London: The Clarendon Press, 1942.

PA The Principles of Art. Oxford: The Clarendon Press, 1938.

RP Religion and Philosophy. London: Macmillian and Company,

SM Speculum Mentis. Oxford: The Clarendon Press, 1924.

CEPC Krausz, Michael, editor. Critical Essays in the Philosophy

CRM Rubinoff, Lionel. Collingwood and the Reform of

FYC Johnston, William M. The Formative Years of R. G.

LPC Donagan, Alan. The Later Philosophy of R. G. Collingwood.

MHD Mink, Louis O. Mind, History, and Dialectic: The Philosophy

During his lifetime at Oxford, the English philosopher,

In Part I the problem of interpreting Collingwood's

In Part II the early writings (1916-1932) are examined

In Part III the later writings (1933-1946) are surveyed

Copyright, 1983: Glenn Shipley, PhD

THE PROBLEM OF INTERPRETING

INTRODUCTION: COLLINGWOOD AND HIS INTERPRETERS

1. Collingwood: His Life, His Writings, and His Era.

Robin George Collingwood (1889-1943) was the youngest of

four children and the only son of W. G. Collingwood (FYC, 6).

His parents were artists, amateur archeologists, and friends of