Unity, Identity, and Explanation in Aristotle's Metaphysics

Unity, Identity, and
Explanation in
Aristotle's Metaphysics
Edited by
T. SCALTSAS, D. CHARLES,
and
M. L. GILL
CLARENDON PRESS · OXFORD

1994
Oxford University Press, Walton Street, Oxford ox2 6DP
Oxford New York Toronto
Delhi Bombay Calcutta Madras Karachi
Kuala Lumpur Singapore Hong Kong Tokyo
Nairobi Dar es Salaam Cape Town
Melbourne Auckland Madrid
and associated companies in
Berlin Ibadan
Oxford is a trade mark of Oxford University Press
Published in the United States

by Oxford University Press Inc., New York
©the several contributors and in this volume Oxford University Press 1994
All right.s reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press.
Within the UK, exceptions are allowed in respect of any fair dealing for the
purpose of research or private study, or criticism or review, as permitted
under the Copyright, Designs and Patents Act, 1988, or in the case of
reprographic reproduction in accordance with the terms of the licences
issued by the Copyright Licensing Agency. Enquiries concerning
reproduction outside these terms and in other countries should be
sent to the Rights Department, Oxford University Press,
at the address above
British Library Cataloguing in Publication Data

Data available
Library of Congress Cataloging-in-Publication Data

Unity and identity in Aristotle's metaphysics I edited by T. Scaltsas,
D. Charles, and M. L. Gill.
Papers from a conference on Aristotle's metaphysics held at Oriel
College, Oxford, in July 1989.
Includes bibliographical references.
ISBN o-19-824o67-8:
l. Aristotle-Contributions in metaphysics-Congresses.
2. Substance (Phi/osophy)-Congresses. 3. Metaphysics-Congresses.
4. Ontology-Congresses. 5. Aristotle. Metaphysics-Congresses.
I. Scaltsas, T. (Theodore) II. Charles, David (David Owain Maurice)
Ill. Gill, Mary Louise, l950-
B491. M4U55 1994 338.4'767'0971-dc20 94-157
I 3579 10 864 2
Typeset by Best-set Typesetter Ltd, Hong Kong

Printed in Great Britain
on acid-free paper by
Biddies Ltd
Guildford and King's Lynn
PREFACE
Earlier drafts of the papers in this volume were presented at a Con-
ference on 'Aristotle's Metaphysics', organized by the editors, which
took place at Oriel College, Oxford, in July 1989. We would like to
extend our thanks to the following organizations for supporting the
aims of the Conference and for their assistance towards the expenses
of the Conference:
The Greek Ministry of Culture
The Hellenic Foundation
The Leventis Foundation
The Marc Fitch Foundation and the British Academy
The Radcliffe Trust
The Rutgers University Endowment for Ancient Philosophy.
We would also like to thank Jeffrey Carr for his valuable assistance
in copy-editing the volume and creating the indices.
One of the participants at the Conference was Montgomery Furth,
who was a very active member of our conversations, and presented a
chapter of his book which was the subject of considerable discussion.
This was one of Monty's last major scholarly appearances before his
untimely death. As a friend and as a reader of Aristotle, he was
important to all of us. We cherish his memory and value the wealth he
left us in his work.
Days after Michael Woods submitted the final version of his contri-
bution to this volume, he died, while still dedicated to his various
projects in progress. Michael was delightful to be with. His modesty,
sincere disposition, and charitable nature strengthened our admiration
for his knowledge of ancient Greek and for the depth of his under-
standing of Plato and Aristotle. He will be dearly missed by all who
knew him.
We dedicate this volume to the memory of Montgomery Furth and
Michael Woods.
T.S., D.C., M.L.G.
CONTENTS
Notes on Contributors ix
Introduction 1
THEODORE SCALTSAS, DAVID CHARLES,
AND MARY LOUISE GILL
I. The Identity of Composite Substance 11
I. A Puzzle Concerning Matter and Form 13

KIT FINE
2. Aristotle on Identity 41
WILLIAM CHARLTON
3. Individuals and Individuation in Aristotle 55
MARY LOUISE GILL
II. The Unity of Composite Substance 73

4. Matter and Form: Unity, Persistence, and Identity 75
DAVID CHARLES
5. Substantial Holism 107
THEODORE SCALTSAS
6. Parts, Compounds, and Substantial Unity 129
SALLY HASLANGER
III. The Potential and the Actual 171
7. Aristotle's Notion of Potentiality in Metaphysics e 173

MICHAEL FREDE
8. The Activity of Being in Aristotle's Metaphysics 195
ARYEH KOSMAN
9. The Priority of Actuality in Aristotle 215
CHARLOTTE WITT
IO. Essences, Powers, and Generic Propositions 229
JULIUS MORAVCSIK
IV. Matter and Form 245

l I. Aristotle on the Relation between a Thing and its Matter 247
FRANK A. LEWIS
Vlll Contents
12. The Essence of a Human Being and the Individual Soul 279
in Metaphysics Z and H
MICHAEL WOODS
13. The Definition of Generated Composites in Aristotle's 291
Metaphysics
MICHAEL FEREJOHN
V. Principles of Aristotle's Metaphysics 319

14. Aristotle's Conception of Metaphysics as a Science 321
ROBERT BOLTON
Bibliography 355
Index Locorum 363
General Index 375
NOTES ON CONTRIBUTORS
ROBERT BOLTON is Professor of Philosophy at Rutgers University, and

the author of many articles on Plato and Aristotle.
DAVID CHARLES is Fellow and Tutor of Oriel College, Oxford, and
University Lecturer in Philosophy. He is the author of Aristotle's
Philosophy of Action, and co-editor of Reduction, Explanation and
Realism. He has published articles on Plato, Aristotle, and philosophy
of mind.
WILLIAM CHARLTON is a Senior Lecturer in Philosophy at the Univer-
sity of Edinburgh. He is the author of many articles and a number of
books including Aristotle's Physics I and II (1970), Aesthetics, An
Introduction (1970), Philosophy and Christian Belief (1988), Weakness
of Will (1988), Philoponus on Aristotle on the Intellect (1991), and The
Analytic Ambition ( 1991).
MICHAEL FEREJOHN is Associate Professor of Philosophy at Duke
University. He is the author of The Origins of Aristotelian Science
(Yale University Press, 1991), as well as of a number of articles on
Aristotelian logic and metaphysics and early Platonic ethics.
Kn FINE is the Flint Professor of Philosophy at UCLA. He has pub-
lished in a variety of fields, including logic, philosophy of language,
and metaphysics. He is the author of Reasoning with Arbitrary Objects
and co-author (with A. N. Prio;) of Worlds, Times and Selves.
MICHAEL FREDE is Professor of the History of Philosophy at the Uni-
versity of Oxford. He is the author of Priidikation und Existenzaussage,
of Aristoteles 'Metaphysik Z', and of Essays in Ancient Philosophy,
and has published many articles on ancient philosophy.
MARY LomsE GILL is Associate Professor of Classics and Philosophy
at the University of Pittsburgh. She is the author of Aristotle on
Substance: The Paradox of Unity, co-editor of Self-Motion: From
Aristotle to Newton, and has written a number of articles on Plato and
Aristotle.
SALLY HASLANGER Associate Professor, University of Michigan, Ann
Arbor. She has published articles on issues in contemporary meta-
physics, focusing in particular on the problem of persistence through
change. She also works and publishes in feminist theory.
ARYEH KosMAN has taught at various institutions in the United States;
since 1962 he has been at Haverford College, where he is John
Whitehead Professor of Philosophy. He is the author of a number of
x Notes on Contributors
essays in the history of philosophy, mostly on Plato and Aristotle. The
present essay is taken from a longer work in process to be entitled The
Activity of Being: A Study in Aristotle's Ontology.
FRANK A. LEWIS is Professor of Philosophy at the University of
Southern California. He has published papers on Plato's later meta-
physics and on various topics in Aristotle's metaphysics, and is the
author of Substance and Predication in Aristotle (Cambridge, 1991).
Juuus MoRAVCSIK is Professor of Philosophy at Stanford University.
He has written many articles on ancient Greek philosophy, ethics, and
the philosophy of language. His most recent books are: Thought and
Language (Routledge, 1990), and Plato and Platonism (Blackwell,
1992).
THEODORE ScALTSAS is Lecturer in the Philosophy Department of
the University of Edinburgh. He is the author of Substances and Uni-
versals in Aristotle's Metaphysics, and of The Golden Age of Virtue:
Aristotle's Ethics. He is the editor of Aristotelian Realism, and has
written articles on Plato, Aristotle, and contemporary metaphysics.
He is Director of Project Archelogos, the creation of electronic com-
mentaries on ancient Greek philosophical texts.
CHARLOTTE WITT is Associate Professor of Philosophy at the University
of New Hampshire. She is the author of Substance and Essence in
Aristotle (Cornell University Press, 1989), and has published widely
on topics in Aristotle's metaphysics. Currently she is working on the
concepts of potentiality and actuality in Book e of the Metaphysics.
MICHAEL WooDs was a Fellow and Tutor of Brasenose College,
Oxford, and University Lecturer in Philosophy. He is the author of
Aristotle Eudemian Ethics, Books One, Two and Eight with trans-
lation and commentary, in the Clarendon Aristotle Series (2nd edn.
1992), and was the General Editor of the Clarendon Plato Series. He
is the author of articles on Plato and Aristotle and on topics in
metaphysics and philosophical logic.
Introduction
THEODORE SCALTSAS, DAVID CHARLES,
AND MARY LOUISE GILL
Prevailing movements. in contemporary metaphysics have fuelled

debate on the central books of Aristotle's Metaphysics. For instance,
recent work on essentialism by Saul Kripke and Hilary Putnam has
motivated a rereading of Aristotle to see whether, or to what extent,
his account of essentialism corresponds to theirs. But the return to
Aristotle has also revealed that understanding his theory can contri-
bute new insights to current discussion. The papers in this volume
examine the nature of essences, how they differ from other com-
ponents of a substance, and how they are related to these other
constituents, both ontologically and definitionally. Addressing these
questions is fundamental for understanding the unity of composite
substances and their identity over time.
In his analysis of a substances, Aristotle uses the concepts of genus
and differentiae, matter and form, and potentiality and actuality,
often appealing to one pair to explain the others. If potentiality
and actuality are taken as explanatorily basic, they should explain
how matter and form are unified, and how they contribute to the
identity of a substance. These concepts will provide a conceptual
bridge enabling one to explore within Aristotle's metaphysical space
. problems concerning the constitution, development, and persistence
of composite substances.
What does Aristotle mean by 'potentiality' and 'actuality', and how
do these concepts explicate matter and form? Aristotle distinguishes
various levels of potentiality and actuality. How, if at all, are matter
and form accommodated by these distinctions? What in substance is
reducible to material potentialities, and do the material potentialities
contribute to its essence? Must they be mentioned in its definition-
the account of its essence? How are they related to the actuality of a
substance? Are they distinct from it, or determined as what they are
by it? If they are distinct, can composites be unities? If composites
consist of a plurality of parts, can the parts be bound together into a
2 T. Scaltsas, D. Charles, and M. L. Gill
single entity of paradigmatic unity and cohesion? What is the role of
actuality in accounting for the unity and identity of substance? Similar
questions about the unity and identity of substances can be raised if
one takes genus and differentiae or matter and form as the basic
explanatory concepts. The metaphysical puzzles about substances also
raise epistemological questions about how one grasps the essence of a
substance and how, if at all, metaphysics is linked to Aristotle's
conception of science in the Analytics. This range of issues is investi-
gated by the papers in the present volume, which centres on the role
of potentiality and actuality in determining the identity and unity of
substances.
The papers in this volume were first presented at a conference we
organized at Oriel College, Oxford, in July 1989, entitled 'Aristotle's
Metaphysics'. Although the title of the Conference did not restrict the
domain of topics to be discussed, the conference presentations dis-
played a remarkable overlap of interest, and we found the discussion
converging on a core of questions concerning the nature of matter and
form and their contribution to the unity and identity of composite
substances. The first part of this volume treats questions concerning
the identity of substances. The papers reveal the limitations of matter
and form in determining identity and individuation, and they also
relate questions of identity to those of substantial unity, which is the
topic of the second part. The papers in the second part consider both
the definitional and ontological unity of substances, and offer alter-
native solutions. Crucial tools in this analysis are Aristotle's concepts
of potentiality and actuality. Questions surrounding these notions and
their relation are the focus of the papers in Part III. The papers in the
fourth part examine various models to explain the relation between
matter and form, drawing from hylomorphic accounts, and from
artefact models. Finally, the paper in the last part turns to metho-
dological considerations, arguing that on the epistemological level,
even in his metaphysical theory, Aristotle is guided by empirical,
rather than a priori, principles.
The volume begins with a puzzle. In the first part, 'The Identity of
Composite Substance', in his paper 'A Puzzle Concerning Matter and
Form', Kit Fine observes that, according to the Aristotelian position,
the matter of one substance can become the matter of another; two
substances can have the same form; and finally, a substance is the
composite of matter and form. But if all this is so, the matter that
composes Socrates at a given time might come to compose Callias,
with the result that Socrates, who consists of the same matter and
form as Callias, would be identical to him. But this is absurd. Fi~e
Pmlnres various wa~s out of the puzzle and points out the difficulties
Introduction 3
associated with each purported solution. He divides the solutions into
three basic groups, each of which rejects one of the three premisses of
the puzzle. He first investigates the rejection of the claim that matter
can migrate from one substance to another and shows that this is
neither plausible, especially for the neo-Aristotelian, nor well sup-
ported by the texts. The rejection of the second premiss, that Socrates
and Callias have the same form, is the view that Aristotle endorsed
individual forms. Fine argues that an Aristotelian would have to
accept a very thoroughgoing version of the view to evade the problem,
and could not thereby avoid certain other difficulties in the composi-
tional account. Finally, Fine explores alternatives to identifying a
substance with the composite of matter and form, including those that
introduce time as a differentiating factor between two substances. But
he concludes that none of the alternatives is fully acceptable to an
Aristotelian or neo-Aristotelian. To that degree, the puzzle remains.
The identity of substances is also the topic of William Charlton's
contribution, 'Aristotle on Identity'. Charlton considers Aristotle's
statement that 'you and your neighbour' are non-identical because 'it
is not the case that both the matter and the account are one', but
claims that, according to Aristotle, your matter fails to be one with
your neighbour's only because it is discontinuous with it. He argues
against the traditional position that substances derive their identity
from prior identifiables which have identity in their own right, whether
these are taken to be quantities of matter or spatio-temporal locations.
The identity of entities like human beings is primitive; there is nothing
which 'makes' you identical with yourself or other than your neigh-
bour; but substances have the status of primary identifiables because
they have form.
In 'Individuals and Individuation in Aristotle', Mary Louise Gill
focuses on the question: What makes an Aristotelian individual the
·individual that it is, and how is one such individual differentiated from
others, which may be qualitatively indistinguishable? The traditional
interpretative dispute has been whether Aristotle regarded matter or
form as the principle of individuation. Gill presents a dilemma to
show that, whatever Aristotle's solution was, the project of individua-
tion cannot succeed, if it is meant to locate something that explains
individuality. A reasonable strategy to avoid the dilemma is to accept
certain entities as the basic particulars and to individuate other things
with reference to them. Such a strategy is not explanatory, because
the particularity of the basic objects is assumed in advance. Gill calls
this strategy 'weak individuation'. She argues that the evidence fails to
support claims that Aristotle regarded either matter or form as indivi-
duator in the strong, explanatory sense, but that passages traditionally
taken to show that matter is a principle of individuation do support
the weaker thesis that matter is a source of plurality. This view is
inadequate, however, because two objects with the same form might
at different times share the same matter. They would then be identical.
Aristotle in fact recognized the problem, and his solution to it indi-
cates that the passages supporting the claim that matter is a weak
individuator do not reflect his considered view.
The unity of matter and form is the topic of the second part, 'The
Unity of Composite Substance', and the papers in this part address
. questions concerning the definitional and ontological unity of com-
posite substances. David Charles, in 'Matter and Form: Unity, Per-
sistence, and Identity', outlines two different ways to understand
Aristotle's account of the unity of a composite substance at a time and
over time. One approach represents him as taking the notion of the
unity of a substance as basic, and introducing his concepts of matter
and form, potentiality and actuality by abstraction from it. On this
view, his concepts of the matter and the form of a composite substance
will be essentially defined as the matter or form of that substance. The
alternative sees Aristotle as taking one of these latter concepts (e.g.
form, or actuality) as basic, and as aiming to explain the unity of a
composite substance in terms of it. Charles examines the second
approach, taking as his starting-point Aristotle's remarks in Met. Z. 17
which involve aspects of the Analytics' method for establishing the
unity of a kind. This leads to the suggestion that in H. 2 and H. 6,
matter and form, or more fundamentally, potentiality and actuality,
are the subject-matter of explanatorily basic propositions (aµeao1
nporaam;) which underwrite the unity of a composite substance.
Charles's aim is further to investigate the extent to which Aristotle in
these chapters is using ideas developed in the Analytics to establish
the teleologically based unity of a composite substance. His conclu-
sion is that while .Aristotle does develop this type of explanatory
approach to the unity of substances in the Metaphysics, his methods
and results show significant, and understaqdable, changes from the
simple Analytics model.
Theodore Scaltsas's position in 'Substantial Holism' can be encapsu-
lated in the claim that a substance is not a cluster of distinct entities
connected by relational bridges between them. Substantial holism is
the doctrine that the homonymy principle applies to all the com-
ponents of a substance unified by the form; namely, incorporation
into the substance involves the reidentification of the incorporated
(concrete or abstract) components in terms of their role in the whole,
as determined by the form. According to Scaltsas, this is what Aristotle
is pointing to wht}n he explains the unity of the components of a
Introduction 5
substance through the potential-actual distinction. The potential is
only homonymously the actual, being named after the actual, without
being the actual. Its actualization is its reidentification in terms of the
actual. Resemblance, and physical continuity between substances, are
not explained in terms of the same components existing in actuality
within the substances but in terms of the same components emerging
by dividing the substances up physically or by abstraction. Abstraction
of components is therefore no more ontologically neutral than physical
division.
In her paper 'Parts, Compounds, and Substantial Unity', Sally
Haslanger argues against the view that the matter and form are iden-
tical, and also against the view that the matter and form are distinct
posterior parts of a substance, unified by both being aspects of the
single substance. Instead, she develops a schematic account of sub-
stantial unity that treats the matter and form of a sensible substance as
distinct prior parts of it. Drawing on the idea that a plurality of items
may be unified by virtue of their relation(s) to a privileged member of
the plurality, she proposes that form, as actuality, functions as the
privileged member of the sensible composite, and matter is unified
with it by virtue of its potentiality for the form.
The papers in Part III, 'The Potential and the Actual', explicate
Aristotle's notions of potentiality and actuality, concepts crucial to the
analysis of unity explored in the previous part. In 'Aristotle's Notion
of Potentiality in Metaphysics e', Michael Frede, taking up the claim
that Aristotle uses the notions potentiality and actuality to explicate
matter and form, turns to Met. e, where Aristotle proposes to clarify
ovvaµu; and tvtpyeza. Aristotle does so by cataloguing various uses of
the expressions. Frede argues that it is a mistake to suppose that
potentiality is one of the kinds of t5vvap1c; that Aristotle here distin-
. guishes. According to Frede, potentiality is any of the various kinds of
ovva1ac; understood in a certain way. The basic kind of t51!vaµ1c; is a
power to produce change in other things. From this basic notion
Aristotle derives various derivative notions, including the capacity to
undergo change, as well as natures and faculties constitutive of them.
These various kinds of t5vvaµ1c; can be understood as each in an
analogous way conferring a certain degree of reality on its bearer.
Taken in this way, a t5Vvaµ1c; is a potentiality. To each kind of t5r5vaµ1c;
corresponds an actuality. Each kind of actuality also confers in an
analogous way a degree of reality on its bearer. In so far as an
actuality is regarded in this way it is called an 'actuality' in the sense
relevant to substances.
In 'The Activity of Being in Aristotle's Metaphysics', Aryeh Kosman
argues that for Aristotle, oiJ(Jfa (substance being) is evtpyeza, under-
stood as active being. Kosman considers Aristotle's distinction in De

Anima II between two senses of f;vepycia or <:vrel.txeia-first actuality,
which is an achieved dispositional capacity for a certain activity, and
second actuality, which is the realization of such a capacity-and asks
whether or3aia is a first or a second actuality. Kosman points out that
certain considerations point in one direction, others in the other. In
fact, however, the demand for a decision, one way or the other, is
misconceived. According to Kosman, in the case of substance being
there is no such distinction. If the distinction is used at all, ov(Jfa is
both a first and a second actuality. But it cannot be understood as a
first actuality in the sense that it might sometimes fail to be active.
Unlike ordinary activities, such as seeing, which can sometimes fail to
occur, while the capacity for that activity remains, the activity of being
human can never lapse without the destruction of the human being.
So a human being has the capacity to be a human being only if she is
actively being human.
The priority of actuality over potentiality is examined in Charlotte
Witt's paper, 'The Priority of Actuality in Aristotle'. Aristotle dis-
tinguishes different respects in which actual being is prior to potential
being: in time, in definition, and in substance. Witt concentrates on
the third respect, and presents a puzzle generated by it. She shows
that the standard existential or teleological interpretations of priority
will not justify Aristotle's contention that actual being is prior to
potential being in substance. Towards solving this difficulty Witt
proposes that we understand the priority in terms of an asymmetrical
relation between actuality and potentiality. Although both actuality
and potentiality are for an end, potentiality is for the sake of actuality
while actuality is not for the sake of potentiality; so potentiality is
existentially dependent, in some way, on actuality, while actuality is
not dependent on potentiality. This solution, however, gives rise to
a further puzzle: how could the potentiality depend on its future
actuality? Witt responds to this difficulty by proposing that the ac-
tuality on which the potentiality depends is not its own future actuality,
but the actuality of adult members of the species; thus, for example, a
child's potentiality is to become (not the adult, that it may become,
but) a human being.
Julius Moravcsik investigates the semantics of universal generaliza-
tions of the sort used by Aristotle, as well as by contemporary scien-
tists, to describe the nature of members of species-for example,
'humans have eyesight', 'beavers build dams'. Such generic sentences,
Moravcsik argues, cannot be analysed as expressing necessary condi-
tions, nor even probabilistic ones. Furthermore, they do not express
whM i.; 11.;m1l or fretment. but rather what is essential to the members
Introduction 7
of the species. But, he points out, among the sentences that describe
the members of a species, some are exceptionless (e.g. 'the whale is a
mammal'), while others, such as the previously mentioned generics,
are not. He concludes that generics should be understood as expressing
potentialities, which are essential features of members of a species. He
then uses this interpretation to explain the semantics of some of the
Aristotelian sentences that contain the operator w<;; bri ro 1roAu ('usually,
in most cases'). This account requires potentiality to be a primitive
concept, and potential being to be irreducible to actual being.
The papers in Part IV, 'Matter and Form', examine various models
in explaining the emergence and complexity of composite substances.
In 'Aristotle on the Relation between a Thing and its Matter', Frank
Lewis addresses the question whether the proximate matter of a sub-
stance is identical with the substance itself. If some of the matter from
which a substance is generated survives in the finished substance-for
example, if blood is both the 'before' and the concurrent matter of a
living animal-then since part of the blood exists before the animal
does, evidently a case of matter exists that is not identical with the
living animal. It seems to follow that the animal is not identical with
even its proximate matter. Lewis tries to square the non-identity of
the animal and the living body that is its proximate matter with the
idea that the substantial form of the animal is essential not just to the
animal, but also to its living body. He uses the example of blood and
Aristotle's artefact analogy to suggest that the form of the animal is an
external principle of behaviour for its living body, but an internal
principle of behaviour for the animal itself. Because the form is
external to it, the proximate matter resembles an artefact, whose form
is likewise essential but external to it. Responding to the suggestion
that, by the definition of 'uniform', no matter exists below the level
of the uniform parts, beyond the reach of the substantial form, he
appeals to Aristotle's theory of f.li~z<;; (mixture), and argues that the
result of mixture has a concurrent matter distinct from the spatially
determined parts of the thing. The initial ingredients-earth, water,
etc.-exist only potentially in the result of mixture; but they contri-
bute to the object their distinctive properties and powers, and they
make up the actual concurrent matter of the mixture.
Michael Woods examines Aristotle's hylomorphism in 'The Essence
of a Human Being and the Individual Soul in Metaphysics z and H'.
He identifies the objections to the analysis of matter and form in
organisms in terms of the distinction in artefacts, and he devotes his
paper to showing how these objections can be overcome. The first
problem is that in the case of artefacts the form can be instantiated in
different kinds of materials (e.g. a sphere in bronze or wood). Woods
shows that in the case of organisms the form cannot be instantiated in

different kinds of materials, because the form supervenes directly
upon the body of the animal. But this supervenience, he argues,
does not undermine the distinction between matter and form in an
organism, because there is a symmetrical dependence between the
matter and the form. Thus, for example, the human body depends on
the human form for being, and remaining (despite material renewal),
this human body, while the human form depends on the human body
for its discernibility. The different contributions of matter and form to
this symmetrical dependence facilitates the distinction between the
two, a distinction that is required by Aristotle's metaphysical commit-
ments. Woods makes a further distinction between the form of a
substance and the substance's having the form, arguing that the form
does not have instances, but only 'being the form' does. Thus Socrates
and Callias have the same human form, but it has different careers in
its occurrence in their distinct bodies, and for that reason Socrates'
soul is different from Callias' soul.
Michael Ferejohn's paper, 'The Definition of Generated Compo-
sites in Aristotle's Metaphysics', defends the integrity of the organiza-
tion of Met. Z against a recent charge in Frede and Patzig (1988). It
does this by arguing that Aris~otle's 'physical' (<pvmKo\;") discussion of
matter/form composites in Z. 7 contributes importantly to the position
he develops in Z. IO and II on the issue of whether a mention of
matter (in addition to form) in the formula of the composite entails
that the formula is not sufficiently 'unitary' to count as a genuine
definition. Ferejohn begins by offering a novel interpretation of Z. 5,
one of the many passages where Aristotle uses the example of the 'per
se affection' snubness as a stand-in for matter/form composite sub-
stances. According to this interpretation, the passage presents the
following 'logical' (A.oy1K6\;") puzzle, or aporia. On the one hand, it is
not possible to define snubness (or a composite substance) without
mentioning its proper subject, while, on the other hand, explicit
mention of the subject leads to an infinite regress; thus it appears that
such things are not definable at all. Ferejohn then interprets a key
passage in Z. 7 as an endorsement of the thesis that one should not
strictly speaking apply the name of the matter (such as 'brazen'). The
final section of the paper argues that this thesis enables Aristotle to
see in Z. IO and II that it is possible to define a composite by giving a
special .kind of formula that mentions its matter indirectly by defining
its form as 'enmattered', and that these lvv201 26;101 are immune to
the aporia of Z. 5.
The paper in the,,final part, 'Principles of Aristotle's Metaphysics',
... ~... " ~" Pni<:tPmnlotrir~I ouestions raised by Aristotle's metaphysics
Introduction 9
concerning how his metaphysics is linked with his conception of science
in the Analytics. Robert Bolton examines Aristotle's methodology
in establishing metaphysical principles in 'Aristotle's Conception of
Metaphysics as a Science'. Bolton points out that according to Aristotle
metaphysics is a science. He argues, against currently received views,
that for Aristotle, inquiry into first principles in the Metaphysics is
characterized by the features which the Analytics describes as typical
of scientific inquiry, and does not introduce any new procedures.
Bolton offers a detailed examination of Aristotle's proof of the
most fundamental of all metaphysical principles, the principle of non-
contradiction. On Bolton's construal of the proof, Aristotle does not
argue for this principle from a priori grounds, or by use of any new
form of dialectical, or other, proof. Rather, Aristotle develops a
peirastic dialectical argument from f.wfo~a (noted beliefs), which
displays the characteristics of peirastic elenchus, as described by
Aristotle in the Sophistical Refutations. Given Aristotle's view of the
elenchus, Bolton argues, the proof does not by itself provide the
proper basis for genuine scientific knowledge. Bolton argues further
that the proof establishes only an instance of the principle, on em-
pirical grounds. However, Aristotle's proof procedure can be gen-
eralized, thereby providing an inductive procedure for knowledge of
the principle of non-contradiction. Such a procedure would be typical
of what the Analytics describes as the method for reaching scientific
principles, which, according to Bolton, justifies Aristotle's claim that
metaphysical principles are reached 'from experience'.
In explicating Aristotle's theory of substantial unity and identity,
our concern in this volume has been both exegetical and philosophical,
and the papers range over questions of ontology, explanation, and
methodology. Our aim has been, not only to give answers to central
problems in Aristotle's metaphysics, but also to stimulate further
research on the problems that it defines, and around the controversies
that it embodies.
I
The Identity of
Composite Substance
I
A Puzzle Concerning
Matter and Form
KIT FINE
Montgomery Furth has written, 1 'given a suitable pair of individua,ls

... there is no reason of Aristotelian metaphysics why the very fire
and earth that this noon composes Callias and distinguishes him from
Socrates could not, by a set of utterly curious chances, twenty years
from now compose Socrates ... '. He does not specify what these
'curious chances' might be. But we may suppose that Socrates eats
Callias for his lunch and that, owing to the superiority of Callias' flesh
and bone, it is the matter of this which remains in Socrates after the
period of twenty years.
That such an exchange of matter is possible is a point on which
many Aristotelian scholars could agree. However, I wish to argue that
such a case gives rise to a fundamental difficulty; for its possibility
runs into conflict with certain basic metaphysical principles which are
commonly attributed to him and which would also be commonly
accepted.
The problem consequently arises as to how this difficulty is to be
resolved. This problem itself may be regarded in two somewhat dif-
ferent lights. On the one hand, it may be regarded as a difficulty for
Aristotle. The question then is whether one can find a solution which
would be acceptable to him, either in the sense that he would or that
he could accept it. On the other hand, it may be regarded as a
©Kit Fine 1994

1 I should like to dedicate this paper to the memory of Monty Furth, beloved
colleague and friend. The passage cited here originally appeared in his article
(1978: 643), and is repeated in his book (1988: 180-1). I should like to thank the
members of the 1989 conference on Aristotle in Oxford and of the 1989 con-
ference on predication in Irvine for many helpful remarks on the topic of this
paper. I am especially indebted to Gavin Lawrence for discussion of the Greek
text and to Rogers Albritton and Mary Louise Gill for their careful and detailed
comments. The two puzzles which are considered here were originally raised,
though without extensive discussion, in my unpublished paper 'Aristotle· on
Substance' (pp. 27-8).
14 Kit Fine
difficulty for a neo-Aristotelian, i.e. to someone who is sympathetic to
the analysis of things into matter and form. The question then is to
find a solution, regardless of whether or not it would be acceptable
to Aristotle.
For the most part, my concern has been with the exegetical ques-
tion; and even here, my purposes have been somewhat limited. For I
have not attempted to settle on one solution as opposed to another.
My aim has been to map out the exegetical space rather than to locate
the views of Aristotle within it.
However, it should be mentioned that I count myself a neo-Aristo-
telian (and, indeed, it was my own commitment to hylomorphism that
led me to investigate Aristotle's views in the first place). It has there-
fore been of some importance for me to take the purely metaphysical
question into account.
I. THE PUZZLE
Our difficulty is simply stated. Suppose that Socrates has at one time
the same matter as Callias has at another time. Then their matter is
the same; their form is the same; and since each of them is a com-
pound of matter and form, they themselves are the same.
It may help to state the puzzle in more formal terms. Let us use S
for Socrates and C for Callias. Suppose that, in the envisaged situa-
tion, t and t' are the respective times at which Socrates and Callias are
assumed to have the same matter. Let m be the matter of Socrates at
time t and n the matter of Callias at time t'. Let F be the form of
Socrates and G the form of Callias. Given some matter m and a form
F, let Fm (sometimes also written as F(m)) be the compound of m
and F. We then make the following assumptions:
(1) S :ft C;
(2) m = n;
(3) F = G;
(4) S = Fm and C = Gn.
(1) says that Socrates and Callias are distinct, (2) that Socrates' matter
(at t) is the same as Callias' matter (at t'), (3) that Socrates' form is
the same as Callias' form, and (4) that Socrates is the compound of his
matter (at t) and of his form and that, likewise, Callias is the com-
pound of his matter (at t') and of his form. ·
All of these assumptions appear to be reasonable; and yet taken
together they yield a contradiction. For from (2), (3), and (4), it
follows by two appijcations of the Leibniz's Law (the substitutivity of
4 ~ / ~ \
A Puzzle Concerning Matter and Form 15
The above formulation of the argument has the advantage of brevity.
But it has two annoying features. The first of these is that the assump-
tions concern a specific situation, the one in which Socrates eats
Callias. This makes the argument needlessly specific, perhaps danger-
ously so; for we run the risk that the argument may not hold up under
the given choice of a situation, even though it would have held up
under some other choice. One might try to avoid this difficulty by
making the argument suitably general: one supposes only that there is
a possible situation in which the relevant assumptions hold. However,
one then loses the ability separately to assess the premisses upon
which the argument depends, since the assumption of several specific
possibilities has been replaced by that of a single general possibility.
The other defect in the formulation is that it presupposes the intelli-
gibility of de re modal discourse. Thus it talks of the possibilities
for Socrates and for Callias. The formulation also presupposes the
legitimacy of the application of Leibniz's Law within modal contexts.
Thus it is assumed, given that Socrates and Callias are the same in the
given possible situation, that they are in fact the same. I myself would
not question these presuppositions; but it is clearly preferable not to
have to make them.
Fortunately, both of these undesirable aspects of the argument can
be avoided under a longer, but more careful, formulation. This has
three premisses which correspond to the premisses (2), (3), and (4) in
the short form of the argument. Let us say that two things are
cospecific if they are of the same lowest species. The first of the
premisses then says:
Material Migration. It is possible for two cospecific things to ex-
change their matter, i.e. for the matter of one at one time to be the
same as the matter of the other at another time.
The second premiss says:
Common Form. It is necessary that any two cospecific substances
have the same form.
The third says:
Simple Composition. It is necessary that anything enmattered is
identical to the compound of what (at any time) is its form and its
matter.
From these premisses a contradiction then follows in much the same
way as before.
It should be clear that the present formulation removes the defects
noted above: it provides an analysis of the assumptions on which the
argument rests; and it avoids all appeal to de re modal locutions.
Moreover, the present version of the argument is as dialectically
~oPent as the earlier version: anv !:!rounds for reiectine: its oremisses
16 Kit Fine
will be equally good grounds for rejecting the other's premisses. We

will therefore adopt the long form as our official version of the
argument, although we will usually revert to the short form for pur-
poses of exposition.
II. EQUIVOCATION
There is an obvious challenge to the logic of the argument, which

is that it equivocates on the meaning of 'matter'. Perhaps for us
moderns there is no ambiguity in the notion of the matter of a thing.
But for Aristotle, matter comes in 'levels'. Thus there is the level one,
or proximate, matter of something; this itself may have proximate
matter, which is then level two matter for the original thing; and so
on.
Now the matter of which Socrates is a compound is his proximate
matter, while the matter which migrates is presumably some low-level
matter, such as the elements, and certainly not his proximate matter.
Indeed, his proximate matter is his body; and it is clearly not true, in
the envisaged situation, that Socrates' and Callias' bodies are the same.
In response to this charge of equivocation, it should be pointed out
that it is not evident that it is only the proximate matter of something
which combines with the form to produce the corresponding com-
pound. Suppose Socrates is the compound Fm, where m is his proxi-
mate matter (i.e. his body) and F the complementary form; and
suppose m is the compound Gn, where n is the proximate matter of m
and G the complementary form. Then it seems natural to suppose that
Socrates is also the compound·Hn, where His some sort of composi-
tion FG of the forms F and G. 2 So it seems to be possible in principle
that the matter of which Socrates is, in part, a compound should be
the same as the matter which migrates.
Unfortunately, such a possiblility will not help defuse the objection.
For let us grant that Socrates is the compound of some low-level
matter (si;ich as fire and earth) and of some appropriate compositional
form; and similarly for Callias. The componential matter will then be
the same as the migratory matter. However, we will lack the reason
which we had before for supposing that Socrates and Callias have the
same relevant form. For the compositional form is more fine-grained
2 My unpublished paper contains an attempt to develop a theory of composi-
tional forms. It should be noted that the compositional form of a thing can
change if its matter changes. For suppose that F is the form of x and that m = Fn
is the matter of x at one time and m' = F'n' the matter of x at another time. Then
the corresponding comp~sitional forms of x at those times are FG and FG'.
than the predominant form (i.e. the form which is complementary to
the proximate matter); and the cospecificity of Socrates and Callias,
even under a universal conception of form, will be insufficient to
guarantee that the compositional form is the same. 3 Admittedly, the
ambiguity in the use of the term 'matter' will disappear; but it will
have reappeared as an ambiguity in the use of the term 'form'.
There is, however, a more successful way in which the charge of
equivocation may be resisted. We originally described our possible
situation in a way that was meant to make it evident that Socrates and
Callias have the same (predominant) form. We must now redescribe it
so that it is equally evident that the form of their bodies is the same,
and so on all the way through the different levels of matter. ,
It is generally unclear when the matter of two things is of the same
form, even when the things themselves are of the same form. It is
unclear, for example, what are the necessary and sufficient conditions
for two human bodies to be of the same form. All the same, there is a
very general sufficient condition which may be given for two things to
be of the same form. It is that they be qualitatively the same, i.e. that
there be no qualitative differences between them, either of a relational
or of a non-relational sort. For surely part (perhaps all) of what is
implied by saying that the form is universal is that it can be specified
in completely general terms, i.e. without reference to any particular
thing. (This might also be thought to follow from the requirement that
form be definable.)
Thus we may guarantee that the matters of Socrates and Callias are
of the same form by supposing that they are qualitatively the same.
Now, in general, the supposition of qualitative sameness will require
that the universe be eternally cyclic (in both the backward and forward
direction). Socrates and Callias will then be counterparts under two
different cycles. However, such a drastic possibility is probably not
required. For all that we need is that there be no relevant qualitative
difference between Socrates and Callias, i.e. no qualitative difference
which is relevant to them or their matter having different forms.
Of course, one might adopt a Leibnizian view on form: only exact
3 Under this way out, we must allow that two things be cospecific even though
their proximate matter is not; the cospecificity of Socrates and Callias, for example,
will not guarantee that their bodies are cospecific, For if Socrates and each level k
matter mk of Socrates were cospecific, respectively, with Callias and each level k
matter nk of Calli as, then presumably the predominant form F 1 and each level k form
Fk of Socrates would be the same as the predominant form G 1 and each level k
form Gk of Callias. But then the compositional form F 1F2 .• , Fn of Socrates
would be the same as the corresponding compositional form G 1G 2 •. , Gn of
Callias.
r8 Kit Fine
qualitative counterparts have the same form. And one might combine
this with a Leibnizian view on possibility: no two things can be exact
qualitative counterparts. The possibility of Socrates and Callias having
the same form would not then arise. However, Aristotle would not
have adopted such an extreme position, either on qualitative form or
on the existence of qualitative counterparts (see Cael. I. 9 and Met. z.
15); and so he would not have had comparable reasons for rejecting
the possibility.
Indeed, even if he were to reject a qualitative criterion for sameness
of form, it is still hard to believe, given that it is possible for Socrates
and Callias to have the same form, that there would not be some
possible situation in which their proximate matter also had the same
form, the next-to-proximate matter had the same form (if it had any
form at all), and so on all the way down to the penultimate matter.
But assuming such a situation is possible, whether underwritten by a
general qualitative similarity or not, the puzzle can be reinstated. For
Socrates and Callias are distinct. By our assumption, their (predo-
minant) form is the same; and so, by Simple Composition, the only
way they can be distinct is for their respective proximate matters m
and n to be distinct. Again, by the assumption, either m and n both
have no form or their form is t.he same. In the latter case, the only
way for m and n to be distinct is for their proximate matter to be
distinct. Proceeding in this way, we see that the matters of Socrates
and of Callias must be distinct at every level (including the ultimate
level at which they are no longer composite). But also, given that the
same matter cannot exist at different levels, the matters of Socrates
and Callias across levels must be distinct. So no matter at any level of
the one is identical to any matter at any level of the other; and
migration is again ruled out. 4
In this reformulation, it can be supposed that the componential
4 In a rigorous formulation of the argument we could substitute a notion of

relevant similarity for the notion of cospecificity. The analogue of Common Form
would then be: necessarily, relevantly similar things have a common predominant
form (if either has any form at all). It must also be assumed that the proximate
matter of relevantly similar things is also relevantly similar. It may be noted that it
is not necessary to assume, as I have done, that the matter at different levels in
Socrates and Callias is distinct. For let mi be the matter at the i-th level of
Socrates at the one time and ni the matter at the i-th level of Callias at the other
time. Suppose that for some k and /, k < I, mk = n 1. Since mk and nk are
relevantly similar and since nk has something, namely nk + 1 as its matter, mk also
has something, namely mk + i. as its matter. But then mk + 1 is the matter n 1+ 1 of
n1; and the argument may be repeated for mk + 1 and n1+ 1. Proceeding in this way,
we see that there must 1;>,e an infinite descending sequence mi, m 2 . . . of matters
nf <;;nl'r'1tt>~ contrarv to the well-foundedness of the matter-of relation.
matter is only proximate matter, as long as Simple Composition is
taken to have application, not only to sensible things, but to anything
enmattered. Thus Simple Composition becomes the assumption that,
necessarily, anything enmattered is the compound of its form and its
proximate matter. On the other hand, we may allow the migratory
matter, i.e. the matter involved in the assumption of migration, to be
non-proximate. It is worth noting, though, that if some non-proximate
matter migrates in the situation above, then so does some proximate
matter. For take the highest level at which the matter of Socrates
migrates to Callias (perhaps this is the level below that of flesh and
bone). Then at the next level up, we will have two distinct but
identically formed things between which there is the migration of
proximate matter. So for the things at this level, something like our
original formulation of the puzzle could be given, with the com-
ponential and the migratory matter both now being proximate matter.
Given the failure to fault the argument on grounds of equivocation,
let us turn to those solutions which challenge its premisses. There are
three such solutions in all, one for each of the three premisses. Let us
consider each in turn.
III. COMMON FORM
The common form assumption states that it is necessary that cospecific

things have a common form or, under the more refined formulation,
that it is necessary that relevantly similar things have a common form.
This assumption might, of course, be denied on the grounds that
forms are individual. For presumably, in that case, it would be im-
possible for two distinct things to have the same form (even if it were
possible for something different to have had the form in the first
place).
It is not my intention here to enter into the debate concerning in-
dividual form. But I do want to make some remarks on the relevance
of the debate to the resolution of the puzzle. It should be noted, in
the first place, that it is a lot easier to attribute the belief in individual
forms to Aristotle than to hold it oneself. For Aristotle seems to have
a possible basis for the belief, namely that forms are real and active
principles in the world, which is denied to any right-minded modern.
Thus in the absence of an alternative conception of individual form,
the neo-Aristotelian must find some other solution to the puzzle.
In regard to the exegetical question, it should be observed that the
issue of whether Aristotle believed in the mere existence of individual
or universal forms is relatively uninteresting. For granted that he
20 Kit Fine
believed in universal forms, there would be no difficulty in supposing

that he believed in individual forms as some sort of indexed version
of the universal forms (something which we might represent as an
ordered pair of a universal form and a thing which had that form);
and given that he believed in individual forms, there would be no
difficulty in supposing that he believed in universal forms as some sort
of abstraction from individual forms (something which we might
represent as an appropriate equivalence class of individual forms).
The interesting question is the role of individual and universal forms
in his thought and whether, in particular, one rather than the other is
to be accepted as the 'essence' of a sensible thing.
Similarly, what is relevant to the resolution of our puzzle is not the
existence of universal or individual forms, but the status of the forms
which enter into compounds. Can these be universal? Or must they be
individual? The philosopher who would solve our puzzle by appeal to
individual forms cannot be content with their existence. He must
maintain that it is these forms, rather than their universal counter-
parts, which enter into a compound; and this can only be maintained,
I assume, on the grounds that it is the individual, rather than the
universal, forms which constitute the essence of things.
It is also important to distinguish between a partial and a full
advocacy of individual forms. A full advocate will maintain that
anything with a form has an individual form (which serves as its
essence); a partial advocate will maintain that this is only true for
some of the things with form, perhaps for all living things or for all
things which need not themselves be the matter of anything. The full
advocate of individual forms has a general solution to the puzzle. But
the partial advocate may not; for there may be applications of the
puzzle to some of the things which he concedes are without individual
form.
Two cases are of special interest. The first is that in which certain
kinds of artefact are used in place of people. Thus it may be that one
ship acq~ires the matter of another, much as ip the Ship of Theseus
puzzle. 5 The advocate of individual forms may then say that the ships
have their 'own individual form (which is Frede's line) or must find
5 It is worth pointing out that this version of our puzzle is not the same as the
Ship of Theseus puzzle, even though the two are based upon similar possibi-
lities. For in the latter case, the puzzle concerns a conflict in our criteria for
identity over time; whereas in our own case, the puzzle concerns a conflict
between our intuitive judgements of distinctness, on the one hand, and certain
principles from Aristotle's hylomorphic theory of substance, on the other. Thus
the Ship of Theseus puzzle poses a problem for anyone; whereas our puzzle only
poses a problem for the a~~erent of hylomorphism.
some other solution to the puzzle. Such a case may not be too serious,
however, for someone who was unwilling to extend individual form to
artefacts; for he could always dispute that the matter migrated to a
distinct ship.
The other case is more serious in this regard. For as we have
observed, if the low-level non-proximate matter of a man can migrate,
then the proximate matter of something like flesh must also be able to
migrate; and hence the puzzle can be restated for flesh or whatever.
In such a case the partial advocate of individual forms cannot dispute
migration. So he must either dispute Simple Composition, which
presumably would provide him with an alternative general solution to
the problem, or he must concede that flesh and the like have individual
form. Thus it seems that the doctrine of individual form cannot be
confined to sensible substances but must be extended to their matter,
the matter of that matter, and so on.
We should note, finally, that even the full advocate may run into
problems if he regards his advocacy of individual forms as a way of
saving Simple Composition. For as we shall later see, there is another
puzzle whose resolution would seem to require him to give up that
assumption.
IV. MIGRATION
Let us consider whether the puzzle might be solved by rejecting

Material Migration. It is clear that this is not an option for the neo-
Aristotelian, for there seems to be nothing to prevent the molecules
which now comprise me from later comprising you. But whether it is
an option for Aristotle is not so clear, since it cannot be taken for
granted that he would have adopted anything like our modern scientific
conception of matter.
If form is taken to be universal, then Migration can be rendered as
the claim that it is possible for something to be the matter of two
distinct things with the same form; and this is the version we shall use.
The negation of Migration, which we call Entrapment, then states
that, necessarily, things with the same matter and form are the same;
an 'escape' of the matter to something of the same form is impossible.
It is important to distinguish Entrapment from two stronger claims.
The first of these, which we may call Strong Entrapment, says that
things with the same matter are the same. Thus Entrapment does not
allow the matter of anything to be the matter of anything else with the
same form, while Strong Entrapment prevents it being the matter of
anything else, whether or not the form is the same. The second
22 Kit Fine
strengthening, which we may call (Material) Individuation, says that

two identically formed things are the same in virtue of their matter
being the same. Thus Individuation adds to Entrapment the require-
ment that the identity of the matter should be explanatory of the
identity of the things. 6
There is some textual evidence which seems directly to support the
view that Aristotle held to some form of Entrapment. The three main
passages are from the Metaphysics, namely A. 6, 1016b31-3, Z. 8,
1034a5-8, and/. 3, 1054b16-17. Under an innocent reading, the first
of these supports Strong Entrapment, the second supports Individua-
tion, and the third supports Entrapment.
Some effort has gone into showing that the support for Individua-
tion, or even some form of Entrapment, is illusory. I myself am
inclined to favour the innocent readings. I am also inclined to think
that Aristotle's remarks on the unificatory role of form in Z. 17 and H.
6 provide some indirect support for Entrapment. 7
However, it is not clear to me, even under the innocent readings
and with the indirect support, that the present solution remains open
to Aristotle. This is partly because there is some further subtlety in
how Entrapment is to be interpreted. For if the puzzle is to be solved,
then Entrapment must at least mean either that the proximate matter
of anything enmattered is entrapped or that the proximate and non-
proximate matter of any sensible thing is entrapped. But the above
passages could be taken to support the restriction of Entrapment to
the higher-level matter of sensible things.
More significantly, there is the possibility that Aristotle is guilty of
an inconsistency on this point. For (as I will later propose) he might
believe in Entrapment because of his views on the nature of unifi-
cation and he might believe in Migration because of his views on the
nature of matter, without being aware of the inconsistency between
the two. In any case, we need to consider the evidence on the other
side, in favour of the possibility of migration.
It will be helpful to organize the considemtion of this question
around certain apparent instances of migration. We shall present
three kinds of example (in order of increasing cogency); and in each
6 There is also a principle of individuation corresponding to Strong Entrap-

ment. Moreover, there are even stronger versions of Entrapment. Thus one might
rule out the possibility that the matter of one thing could ever exist except as the
matter of that thing; and one might also rule out the possibility that part of the
matter of something could ever be part of the matter of something else or exist
except as part of the matter of the thing.
7 Charlton (1972) and,,Cohen (1984) examine the evidence for individuation;
th<" rpm::.rh on 11nificMiort in 7. 17 and H. 6 will be discussed later in the paper.
case, we shall consider whether there are any textual or intuitively
compelling reasons to suppose that Aristotle would have accepted the
example.
But before so doing, let us make some general remarks on how the
example and the responses to them are to be characterized. Any
apparent case of migration may be described in neutral terms, terms
to which both the opponent and the proponent of Migration can
agree. Thus they will agree that certain things continue as certain
other things; and they will agree as to the form of these things. But
they will differ on questions of identity. The proponent of Migration
will hold that the things, as described in the case, are distinct but that
their matters, at the respective times, are the same. The opponent
must hold either that the things are the same or the matters distinct.
For the proponent of Migration, the prima-facie evidence for the
claim of identity is the existence of a material continuity-the matter
of the one particular continues, in a suitable way, as the matter of the
other; and the prima-facie evidence for the claim of distinctness is the
existence of a formal discontinuity-the form of the one particular
does not continue, in a suitable way, as the form of the other.
Thus the opponent of Migration must question the prima-facie
evidence for either the claim of identity of matter or the claim of
distinctness of form. In the one case, he must maintain that the
underlying material continuity is not sufficient to convey the matter
from the one particular to the other; somehow the matter gets trapped
on the way (and we have what might be called an instance of entrap-
ment). In the other case, he must maintain that the formal discon-
tinuity is not sufficient to prevent the first particular's reappearance;
somehow it gets reinstated (and we have what might be called an
instance of reinstatement).
Let us now turn to the examples. 8 The first kind arises from the
possibility of fission and fusion. Fission (which Aristotle accepts for
plants) occurs when one thing of a given form changes into several of
that form, fusion when several things of a given form change into one
thing of that form. Thus both are cases of intra-formal change; the
'break' in the form consists, not in its being different, but in its not
being a single thing which has the form.
It seems plausible to suppose that fission is possible just in case
fusion is. However, combinations of the two appear to lead to cases of
migration. For suppose an amoeba xy splits into the amoebas x and y,
which then fuse back into the amoeba (xy). Then will not (xy) be a
8 The framework within which these questions have been considered could be
made much more systematic and precise; but this is not something that I will
attempt here.
24 Kit Fine
distinct particular with the same form and with the same matter, at
some level, as xy?
The opponent of Migration could deny that fission and fusion are,
in the required sense, possible. He need not dispute the apparent
facts. He could concede, for example, that, in a neutral sense of
the term, the one amoeba continues as two. But he could deny that
the matter persists, i.e. that the matter of the two is the same as the
matter of the one.
The appeal to entrapment in this case, though, is not very plausible
(and, as we shall later see, may not even help). What is more plausible
is to deny that the final amoeba (xy) is different from the first xy. The
matter, at some appropriate level, remains the same, but so does the
amoeba; and hence we have a case of reinstatement. 9
However, other apparent cases of migration which arise from the
possibility of fission and fusion are not so readily disposed of. We are
inclined to think that something can survive a fission or a fusion: that
if an amoeba z combines with a sufficiently small or insignificant
amoeba then the resulting amoeba can still be z; and that if a suf-
ficiently small or insignificant amoeba splits off from an amoeba z,
then the remaining amoeba can still be z.
But our opponent of Migration is unable to hold such a view. For
much as in the Ship of Theseus case, an amoeba z with matter m may
split into a large and small amoeba, surviving as the large part; it may
then fuse with a small amoeba, surviving as the fusion; and so on until
the resulting amoeba z' possesses none of the original matter m. In
the meantime, the small amoebas that it has shed may fuse into an
amoeba z" whose matter ism. Our opponent is then required to make
z identical with z', on account of its survival, and also to make z
identical with z" on account of their common matter and form.
It therefore appears that anyone who would wish to solve the puzzle
by denying Migration should maintain the following views concerning
fission and fusion. First, the matter (at some level) survives these
processes. Second, the things themselves do hot survive, no matter
how modest the change; all three things which are party to a fission or
a fusion must be distinct. Third, anything which ceases to exist as the
result of such a process can have its existence restored as a result of
such processes. Thus the thing is not completely destroyed; for it is
still capable of existing. And the processes are not completely destruc-
tive in their effect; for what they destroy they can also restore.
A similar kind of case arises with artefacts. Suppose that a house is
9 It is interesting to nqte that in this case we have the migration of part of the
form hut not the migratiOil Of the Whole Of
m~+tar .,, ~ r!;ct;nf't thincr fulth thP. ~:Jmp
wrecked or a statue melted down. Then can we not create a formally
identical house or a statue with the very same matter?
How one treats such a case depends very much on how one under-
stands Aristotle's dismissive remarks concerning the substantial status
of artefacts (e.g. in Met. H. 2, 1043a4-5). Under an extreme interpre-
tation, artefacts are only treated as things in their own right for the
purposes of analogy. Strictly speaking, the statue is the same as the
bronze, the house the same as its materials. The puzzle would not
then arise, since the required distinction between the thing and its
matter could not be drawn.
However, it is hard to see on such an interpretation how the bronze
could be understood, even by way of analogy, to be the matter of the
statue (which is, after all, identical to the bronze). A less extreme,
and more plausible, view is that the statue and the bronze are distinct
but that they only have the relation of matter to compound by way of
analogy. Perhaps strictly speaking, the statue is an accidental unity of
the bronze and its accidental shape, in much the same way that
'musical Coriscus' is an accidental unity of Coriscus and musicality (cf.
Met. LI. 1015°16-19). A related view is that the statue literally has the
bronze as its matter and hence literally has something, such as a
shape, as its form. However, the form and the thing are not literally
substantial, but only substantial by way of analogy. 10
On either of these last two views, the puzzle will still arise, but with
the analogues in place of the real notions. The same kind of responses
can therefore be given. Thus one line is that the same matter does not
persist through the processes of disintegration and reintegration.
However, this is neither plausible, especially for the case of accidental
unities, nor is it well supported by the text. (For example, in Met. LI. 4,
1014b30-1, Aristotle writes, 'for when a product [he has in mind
something like a statue or utensil] is made out of these materials, the
first matter is preserved throughout'.)
A more plausible line is that under suitable reintegration the original
house or statue is reinstated (the form must then of course amount to
more than just being a house or a statue). Thus the processes of
disintegration and reintegration would be seen as analogous to the
processes of fission and fusion. They might even be thought to ·be
essentially the same kind of process, differing only with respect to
whether the origin of the process has the same form as the result. 11
10 To the extent that substance admits of degree, this interpretation is more
plausible than the other; for it is easier to conceive of the substantiality of a form
as being a matter of degree than anything corresponding to the relation of matter
to form.
11 Note that in this kind of case the whole of the matter may migrate to
~nmP.th1no with :::i: rliffPrPnt fnrm nr PUPn tn C.{"'tmPth1no urithnnt ~"" fnrthPr fnrm
26 Kit Fine
We come now to the last and most cogent putative instance of

migration. This is the case in which the apparent transfer of the
matter is from one living thing to another and takes place by means of
normal biological or chemical processes. For the purposes of the case,
it is probably better not to use the example of people, as we did at the
introduction. For the degree of relevant similarity which is required
for the individuals to have the same form might also endow them with
similar memories, similar personality, and so on. They will also enjoy
whatever causal connection is required for the matter to apparently
migrate. Can we be sure, in such circumstances, that the two indivi-
duals are genuinely distinct?
To avoid such problems, it is better to take the case to concern
trees, let us say, rather than people. For I take it that the relevant
similarities and causal connections would then provide no ground for
supposing that the two individuals are the same. An oak that emerges,
perhaps over millions of years, from the decomposed remains of an
existing oak will surely not be the same tree, similarities in structure,
origin, or whatever notwithstanding.
We may therefore suppose that Socrates and Callias are trees and
that as we 'track' the elemental matter of one at a given time, we find
that it coincides with the elemental matter of the other at a later time.
For migration genuinely to take place, the matter must 'get out' of the
one tree and it must 'get in' to the other tree; the matter must also
'get across' from the one tree to the other. It seems reasonable to
suppose that if the matter can get out and in, then it can also get
across, if only because the processes involved in its getting across are
essentially no different from those involving in its getting out or in.
Now there are only two ways the matter could get out: either at the
individual's cessation, and then of course the whole of the matter
might get out; or before its cessation, in which case it is plausible to
suppose that only part of the matter can get out. Likewise, there are
only two ways the matter can get in: either at the individual's incep-
tion, and then of course the whole of the matter might get in; or after
its inception, in which case it is plausible to suppose that only part of
the matter can get in. Thus what is at issue are Aristotle's views on
substantial change and on growth and diminution.
I hope to deal more fully with his views on growth and diminution
elsewhere. But there is one comment of his which is specially relevant
to our present concerns. In GC I. 5, 32 I 025 he writes, 'This is how
the matter of flesh grows: an addition is not made to each and every
part, but some flows away and some comes in new.' Now the natural
way of understanding the phrase 'some comes in new' is that some
matter that is not a part of the matter of the flesh at one time is a part
of the matter of the flesh at a later time. This then strongly suggests
that migration could take place through the complementary processes
of growth and diminution.
However, there are at least two ways in which this conclusion might
be questioned. First, the phrase 'some comes in new' might be inter-
preted to mean that some matter m that is not a part of the matter of
the flesh becomes some matter n (not necessarily the same as m)
which is a part of the matter of the flesh at the later time. The original
matter m would not then literally enter into the flesh. Second, even if
it were granted that the original matter later became part of the
matter of the flesh, it still might be supposed that it was only in a
time-relative sense that it was part of the matter of the flesh. Thus
even in the circumstances that were most propitious to migration, the
most that could be concluded would be that the matter of the flesh of
one individual was comprised at one time of the same material parts
as the matter of the flesh of the other individual at the later time. But
the fact that the two fleshy matters were comprised of the same parts
at the two respective times would not be a sufficient ground for
concluding that the two matters were the same. Which of these inter-
pretations is to be adopted depends, I think, upon a fuller under-
standing of his views.
Let us now turn to Aristotle's views on substantial change. Accord-
ing to the traditional interpretation of Aristotle on the topic, migration
would then be possible; for the prime matter of the one tree would
persist through all the changes and thereby become the prime matter
of the other tree. In fact, nothing quite as strong as the traditional
interpretation is necessary. For we do not require that prime matter
persists, but only matter at some level; and we do not require that it
persists through all changes but only those involved in the apparent
case of migration. One might hold, for example, a variant of the
traditional view according to which elemental matter persisted through
non-elemental change, i.e. change which was not from one element
to another. Migration could then be upheld under the reasonable
assumption that the changes involved in some apparent case of migra-
tion might all be non-elemental. 12
But the traditional line of interpretation has been questioned in
recent times; and this might then seem to leave open the possibility of
blocking the apparent cases of migration. I myself am disinclined
12 In Met. H 5, 1045"3-6, Aristotle writes: 'If from a corpse is produced an

animal, the corpse first goes back to its matter, and only then becomes an animal.'
The matter here is presumably elemental matter; and there is no suggestion that
elemental changes are required.
Kit Fine
to give up the traditional interpretation, especially in its relevantly

weaker forms. But, all the same, let us see what follows in its absence.
Take something which is not elemental matter (nor prime matter
should there be such). It may then be supposed that it contains
elemental matter. For this thing, like any other enmattered thing, will
submit to hylomorphic analysis: it will have a proximate matter (and a
complementary form); the proximate matter will have its own proxi-
mate matter (and a corresponding form); and so on, all the way down
to something which has no matter. But surely at the ultimate level of
this analysis (if there is no prime matter) or at the penultimate level
(if there is prime matter), we will reach elemental matter.
Consider now the first tree at the juncture at which it dies. Then the
living tree at this juncture contains certain elemental matter; and so
does the dead tree. The opponent of Migration must then contend
that this elemental matter is not numerically the same. 13 For suppose
it were. Then the death of the tree would be a case of a change in
which the elemental matter stays the same. But if the death of the tree
is such a case, then presumably the same could be true of the other
changes involved in the apparent case of migration.
If the earths, let us say, of the living tree and the dead tree are not
the same, then it seems to follow that the forms of the earths are not
the same. 14 For if the forms were the same, then, given that the one
matter continues as the other, it would be hard to see why the earths
themselves would not be the same. Moreover, their difference in form
can be no ordinary difference-it cannot consist, for example, in the
ratio of the contraries making up the earth; for there is no reason why
the ratio would have to be altered by the change. Thus we have to
distinguish two types of earth, one of which is 'potential' by nature
and the other of which is 'actual'.
What evidence is there that Aristotle would have been prepared to
accept such a bifurcation in the form of the elements? He does hold a
doctrine of 'homonymy', according to which a dead hand, for example,
13 One might adopt a radical position according to which elemental or other

low-level matter is cross-temporally indeterminate (and in this sense not a 'this').
Under such a view, it would be incorrect to say either that the earth of the dead
tree was numerically the same as that of the living tree or that it was numerically
distinct. There would be earth in both trees; but whether it was the same earth
would be a question which simply did not arise. Such a view might serve to
somewhat soften the impact of my arguments; but I have not attempted to take it
into account.
14 If there is no matter underlying the elements, then they will have no form in
the sense that is complementary to matter. But they will still have form in the
looser sense of kind or sp~i7s.
is not the same, or of the same form, as a living hand. He is even
prepared to extend the doctrine to the matter of the bodily parts, i.e.
flesh and bone, and perhaps further. (See, for example, GA II. 1,
734b24-735a26.) But there is no real indication that he would have
been willing to extend it all the way down to the level of the elements;
and it is odd, if he had countenanced the possibility of homonymy
for the lower levels of matter, that this possibility was never made
explicit. 15
However, let us grant that the elemental matters of the living and
the dead tree are not numerically the same (and we may suppose that,
if the elemental matter is not the same, then neither is the matter at
any level). All the same, a difficulty remains. For even though the
matter of the living tree does not actually exist when the tree dies, it
may still potentially exist, i.e. be capable of reinstatement. But
presumably the conditions for reinstatement are as favourable as they
could be in the case of the second tree. So given that the matter is
capable of reinstatement, it will actually be reinstated in the second
tree; and migration will again be secured (though the matter will only
get out in the sense of having its potential existence preserved).
Let us say that something is destroyed if it ceases to exist either
potentially or actually; it is and can be no more. The opponent of
Migration must then hold that the matter of the tree is destroyed on
its death. Indeed, what is typical of death in this regard is that it is a
change in which the individual is destroyed. Thus our opponent must
hold quite generally that when something is destroyed then so is its
matter at every level. He must extend the homonymy doctrine 'out' as
well as down; it must be maintained that no subsequent matter can
be numerically identical to any of the matter of something that is
destroyed. This extension is reasonable for bodily parts and the like; a
living hand, after all, is different both from a dead hand and from any
subsequently reconstituted living hand. But it is even more proble-
matic than the original doctrine in application to the lower reaches of
matter.
In conclusion, we see that there are various different grounds upon
which Aristotle's commitment to Migration might be defended. There
are the difficulties over fission and fusion; there are his views on
growth and diminution; and there is the evidence from the traditional
15 The passage in Meteor. IV. 12, 389h29-390°1 cannot very reasonably be
used to support the extended homonymy claim. The point there is to show that
certain things are what they are in virtue of a certain power of action or passion.
This is revealed, for some things, by the fact that they Jose their power when
someone dies and thereby cease to exist; but there is no clear implication that all
of the other things would also Jose their power upon someone's death.
30 Kit Fine
interpretation of substantial change. But even without a substratum

for all change, it may be supposed that the elements are a substratum
for non-elemental change; and even without a substratum for non-
elemental change, it may be supposed that migration can be secured
through reinstatement.
Given the implausibility of the case against Migration for living
things, one might think of combining the two solutions so far given.
The puzzle would be solved for animate things on the grounds that
their form is unique (and hence cannot be shared); it would be solved
for inanimate things on the grounds that their matter is unique (and
hence cannot migrate). Thus on this view there would be two fun-
damentally different kinds of substance (and two correspondingly
different kinds of substantial form): those which are life-like and
individuated by their form; and those which are matter-like and indi-
viduated, within their specific form, by their matter.
It has to be conceded that, once a differential stand on individual
forms is granted, there is a great deal to be said in favour of a
differential solution. But the case for a differential stand is somewhat
mixed. On the one hand, the 'soul' is plausibly taken to be the
individual form of a living thing; and there is nothing quite like the
soul which is plausibly taken to play the same role in regard to a non-
living thing. On the other hand, it is unclear what it is about the
general nature of forms which would enable the forms of some things
to be individual and require the forms of other things to be specific.
Granted that the form of a living thing is its soul, there would have to
be something in a non-living thing which was sufficiently unlike the
soul to be specific (as opposed to individual) and still sufficiently like
the soul to be a form. But it is unclear what such a thing would be.
V. COMPOSITION
I now wish to consider the question of whe.ther the puzzle can be

solved by rejecting Simple Composition. The standard interpretation
of Aristotle (and one which I myself am inclined to hold) is that he is
committed to both Common Form and Migration. The standard view
(i.e. the content of this standard interpretation) is also one to which
the neo-Aristotelian is likely to subscribe. The rejection of Simple
Composition is therefore of special interest under this interpretation
or view, since it constitutes the only way of rescuing Aristotle or the
neo-Aristotelian from contradiction.
It is clear that Aristotle believes, in some sense, that any substance
or anything enmatte~ed is a compound of matter and form; for he
repeatedly refers to such things as compounds of matter and form,
and is often quite explicit in saying that they are such compounds. Let
us call this unconstrued belief Composition. Then the question is how
this belief is to be construed.
The most straightforward construal is in terms of Simple Composi-
tion; for one most naturally takes the compound to be the com-
pound of the matter and form at a given time, and one most naturally
takes the relation between the thing and the compound to be one of
identity.
The naturalness of the construal is a strong presumption in its
favour. However, there are two quite separate considerations that go
against it. The first is that it makes the operation of compounding
unduly selective or unduly conflationary. For granted that the matter
of something can be different at different times, one may wonder how
the composition of the form with just a single one of the matters is
capable of yielding the given thing, and one may wonder how the
composition of the form with different matters is capable of yielding
the same thing.
The other consideration arises from the puzzle; for this shows that,
under the standard interpretation of Aristotle, he cannot consistently
endorse Simple Composition. However, the force of this consideration
is somewhat problematic. For the fact that an interpretation renders a
philosopher's views inconsistent or subject to some other kind of
difficulty is not in itself a reason for rejecting the interpretation. It is a
reason only to the extent that it is plausible to suppose that the
philosopher himself would have recognized the difficulty. Now in the
present case, it is not at all implausible to suppose that Aristotle was
unaware of the difficulty. For many of Aristotle's commentators have
failed to see the corresponding difficulty in their own interpretation of
him; and it is not ungenerous, one hopes, to suppose that he himself
had no greater logical acumen than his interpreters in this regard.
Nor is it difficult to see what might have led Aristotle or his
interpreters to overlook the inconsistency. First, they overlooked the
somewhat subtle argument that if the non-proximate matter of a
substance can migrate then the proximate matter of something non-
substantial can also migrate. Second, Aristotle had few clear examples
within his metaphysics of things whose proximate matter could migrate.
Certainly, this would not be possible for living things or the higher
forms of matter; nor would it be possible for the lower forms of
matter. His general line on artefacts was not worked out; and so,
arguably, migration would be possible at only one level within the
hylomorphic hierarchy, namely the one below the lowest level at
which homonymy still held. But Aristotle never identified what kind of
32 Kit Fine
matter lay at this level and so was never forced to confront a specific
case of what he would recognize as the migration of proximate matter.
However, even if it is granted that Aristotle was inconsistent on this
point, it is still important, under a more normative conception of the
exegetical task, to ascertain what Aristotle might have said in response
to the puzzle, i.e. to ascertain which solution, if any, does least
violence to his views. So let us consider, both for this reason and for
the ones stated earlier, the various other ways in which Composition
might be construed.
One possible reading is to relativize the relation of identity to a
time (or to substitute a relativized relation of coincidence for identity).
Something enmattered is then taken to be identical, or to coincide, at
a given time to the compound of its matter and form at that time. The
puzzle would then be avoided by not requiring that the object be
absolutely (i.e. numerically) the same as the compound with which it
coincides. Thus on this view the enmattered things and the com-
pounds would in general be distinct, though they would enjoy the
peculiar feature that any object of one sort would always be found in
the company of an object of the other sort. The compound would
enjoy a kind of errant existence-now being associated with this
thing, now with that.
The main difficulty with this view is that there is no real evidence
that Aristotle thought there were compounds in addition to the en-
mattered things with which they coincided. For example, in the battle
for substantiality which is depicted in the central books of the Meta-
physics, it is not as if the compounds were a contender in addition to
the sensible things. It is clear that Aristotle regards the compounds
and the sensible things as one and the same.
A second possibility is that Aristotle takes Composition to mean
Simple Composition and yet does not take Simple Composition to be
actually true but only something that would be true under the sim-
plifying assumption of a static universe. However, it is odd that he
never wakes the assumption explicit or considers the question of what
is actually true; and it is especially odd for someone so sensitive to the
phenomenon of change.
The third reading, which we may call Plural Composition, takes
something to be the compound of its form and its various matters over
time. Thus a thing x is taken to be identical to the compound F(m 11
m2 , ..• mk), where F is the form of x and m 1, m2 , •.. mk are its
proximate matters in order of temporal occurrence.
This proposal avoids the ontological excesses of the earlier ones;
the ontology is the familiar one of substance, matter and form. It
is also quite accep~~le from a linguistic standpoint; for it is not
unnatural to suppose that the reference to matter is plural reference.
However, the proposal is still unable to rescue Aristotle from con-
tradiction. For our puzzle reappears under the modest extension of
Migration to include the possibility of Socrates and Callias having the
same matter, not only at two times, but over the whole period of their
existence.
The last reading, which we may call Relative Composition, relativizes
the notion of a compound to a time. Anything enmattered is identical,
in an absolute sense, to a compound. But the compound is not, in an
absolute sense, the compound of matter and form, but only relative to
a time. Thus the enmattered thing x is taken to be identical to the
compound Ft(m), where F is the form of x and m is the proximate
matter of x at t; or, under a combination of the plural and relative
readings, x is identical to F1(m 1 , . . . mk), where t is the period of time
throughout which x exists and m 1, ... mk are its proximate matters in
order of appearance.
This proposal employs the familiar ontology; and again, it is not
unnatural to suppose that in the reference to the compound there is
an implicit reference to time. The proposal also saves Aristotle from
contradiction. For even though Socrates and Callias have the same
matter and form at the respective times, the compounds can be dif-
ferent on account of the times being different.
The first two readings can be ruled out on grounds of intrinsic
implausibility. This leaves the original construal in terms of Simple
Composition and the remaining readings in terms of Plural and Rela-
tive Composition. Of these, only the third avoids inconsistency (at
least under the standard view) and only the second, or the second in
combination with the third, avoids the problems of selectivity and
conflation. Thus it might appear, from the evidence reviewed so far,
that the preferable reading is in terms of a relativized version of Plural
Composition.
However, the exegetical situation is complicated by the fact that,
for Aristotle, the question of composition cannot be considered in
isolation from his views on the unifying role of form. This role is
described with great lucidity in books Z and H of the Metaphysics. In
H. 6, 1045a7- IO, he writes, 'To return to the difficulty which has been
stated with respect both to definitions and to numbers, what is the
cause of their unity? In the case of all things which have several
parts 16 and in which the totality is not, as it were, a mere heap, but
16 The requirement that the unity have several parts is probably stronger than
Aristotle intends. For although the problem of unity is most evident when the
thing has several parts, it still arises when the thing has only one part, as with the
body being the sole (proximate) part of the man.
34 Kit Fine
the whole is something besides the parts, there is a cause.' This cause,
which he identifies with a form, is something which makes the parts
into the unity; it makes bricks and stones, for example, into a house
(rn41a26).
Now there is an intimate connection between unification and com-
position. This may be roughly described by saying that the operation
of composition must render unification possible; it is by their jointly
entering into the compound, in their own distinctive manner, that the
form is able to make a unity of the matter (but in a way, one hopes,
that is compatible with Aristotle's obscure remarks at rn45h7-24).
It should be noted that this connection makes all the more intelli-
gible our earlier animadversion against selectivity and conflation. It
would be going too far to insist that all of the matter which is unified
should be componential matter, i.e. that it should directly figure in
any correct account of the compound. In the expression 'cat', for
example, it might be supposed that all of the subexpressions 'c', 'a',
't', 'ca', and 'at' are unified by means of the form of juxtaposition; and
yet either of 'c', 'a', and 't', or 'c' and 'at', or 'ca' and 't' could be made
the matter of a corresponding compound. But in such cases, it must
be apparent how in compounding certain of the matter the rest gets
unified; and since this is not apparent for things whose matter varies
over time, it would seem to follow that all of their matter will enter
into the compound.
Indeed, it is a general requirement on any satisfactory account of
composition that it should make clear, or at least be compatible with,
the unifying role of form. However, it is hard to see how Relative
Composition, under either the simple or the plural version, is so
compatible. The difficulty is to understand how the time by which the
compound is indexed is relevant to the unificatory process. It cannot
be that time is one of the elements that is unified; for the time at or
during which the matter exists is not in the requisite way a part of the
resulting unity. Nor can it be that unification is relative to a time; for
how can a time, as such, affect the manner whereby the form makes
some given matter into one thing rather than another?
The only reasonable view seems to be that the process of unification
is not many-one, but many-many. The form does not generally make
any given matter into a single thing; rather, it makes it into several.
Thus the temporal index serves as a kind of external device by which
one of the resulting plurality of unities is picked out.
However, it is hard to make sense of such a conception of unifica-
tion. For how can the form make the matter into something that is
ontologically one, i.e. a unity, without thereby making it into some-
thing that is numerically one? Certainly, the analogy of form with
structure (which Aristotle so often uses) is of no help in this regard.
Juxtaposition, for example, makes the letters 'c', 'a', and 't' into the
single word 'cat'; and there appears to be no reasonable understanding
of how it (or an allied operation) could make those letters into several
such words.
It therefore appears, under the additional assumption that unifica-
tion should be many-one, that the relativized versions of Composition
are no better able than the unrelativized ones to avoid the charge of
inconsistency. We might also note that, under this assumption, the
puzzle may be formulated without bringing in Composition at all. For
instead of concluding that the two compounds are the same, we may
suppose that their common form must unify their common matter into
one and the same thing.
We are therefore left, under the standard view, with no consistent
interpretation of Aristotle (and with no consistent option for the neo-
Aristotelian who accepts the unifying role of form). Moreover, the
contradiction runs deep; it depends only upon fundamental assump-
tions, and there is no obvious way by which it should be removed.
All the same, it might be thought, even when no attempt is made to
remove the contradiction, that the plural reading of Composition is
still to be preferred to the simple one; for it avoids the problems of
selectivity and conflation; and it thereby co-ordinates better with what
Aristotle took to be the unifying role of form. However, given our
own previous diagnosis of the contradiction, this is not such a plausible
line to take. For it may be surmised that Aristotle was led to overlook
the possibility that the proximate matter of something could vary over
time in much the same way that he was led to overlook the possibility
that it could migrate. It could be conceded that he envisaged the
possibility that the non-proximate matter of a living thing could vary.
But it would be claimed that he was never forced to confront any
specific cases of variation in the proximate matter and that he over-
looked the proof that variation in non-proximate matter implies
variation in proximate matter. (The proof goes as follows: take the
highest level at which the matter of something varies; then at the next
level up, we will have a single thing whose proximate matter varies.)
We are therefore left with no reason for preferring any of the other
construals to Simple Composition. Perhaps what can be granted is
that the presuppositions behind Composition and Simple Composition
are not the same. For the latter takes a thing to be the compound of
its form and whatever is, at any time, its proximate matter, while the
former takes the thing to be the compound of its proximate matter
and form under the presupposition that the proximate matter is
unique.
Kit Fine
VI. THE MEREOLOGICAL PUZZLE
Let us return to those solutions to our puzzle which require that we

give up either Material Migration or Common Form. It might be
thought that, with such a solution, there would be no need also to give
up Simple Composition. But there is a secondary puzzle, related to
our earlier considerations concerning selectivity, which provides inde-
pendent reasons for rejecting this assumption.
The new puzzle goes as follows. Surely Socrates' elemental matter
m at one time might not be the same as his elemental matter n at
another time. Let F and G be the respective forms of Socrates at
those times which are complementary to m and n. Then by Simple
Composition (stated with respect to matter at any level), Socrates is
both the compound Fm and the compound Gn; and hence the two
compounds are the same. Now n is a part of Fn; and since Fn and Gm
are the same, n is also a part of Fm. But n and F have no part in
common; and so, for n to be a part of Fm, it must be a part of m. By
symmetric reasoning, m is a part of n. So m and n are parts of one
another and hence are the same. A contradiction.17
The two puzzles are rather different. The first depends upon the
possibility of migration, of the same matter being in two things; the
second depends upon the possibility of variation, of different matter
17 Various points about the argument should be noted. ( r) The term 'part' is
used in a broad sense to include the case in which something is a part of itself.
The assumption that mutual parts are the same can be avoided if it is supposed, at
the outset, that one of the two matters, m and n, is not a part of the other. The
argument then establishes that it is a part. (2) It could be supposed that the com-
pound was of the form F 1(F2 ••• (Fk(m) .. .)) rather than of the form (F1F2 .••
Fk)(m). The commitment to compositional forms would thereby be avoided and
Simple Composition could be limited in its application to proximate matter. (3)
The matter m and n has been taken to be elemental (or penultimate) in order to
make it more plausible that it could vary. Variation requires that the matters m
and n he distinct in number and would therefore qe excluded under a position
which took the question of their identity to be without sense. (4) It is important to
the argument that m and n be at the same hylomorphic level; for it is this which
prevents n having any part in common with the form F. Suppose, for example,
that F were the composite form GH. Then the matter n = Hm would be a part of
(GH)m = G(Hm), yet not a part of m. From this example we see that it would be
incorrect to suppose that any material part of Fm is a part of m. Rather, any
material part of Fm is either a part of m or is a compound of m and a part of F.
We also see that the claim that forms are parts is theoretically significant; for
without it the proper formulation of the principles governing the mereological
behaviour of compounds could not be given. Thus Aristotle's view that the form
is part of the compound should not be dismissed as a fanciful extension of the idea
that the matter is a pa~\
being in the same thing. Both puzzles make use of the assumption that
anything enmattered is numerically identical to a compound of what,
at any time, is its matter and form. But the first argument only
exploits the functional aspect of compounding, namely the fact that
for any matter and form there is at most one compound of that matter
and form. The second argument also exploits a mereological aspect of
compounding, namely the fact that the mereological content of the
compound is exhausted by its matter and form. Given their respective
assumptions, the two arguments then move in opposite directions:
under the first, two compounds which should be distinct are shown to
be the same; under the second, two compounds which should be th~
same are shown to be distinct.
Despite their differences, the two puzzles can be seen to have a
common source in the question of unity. A form is meant to make the
parts of something, either some or all of them, into that thing. Two
necessary conditions for a form to unify may be distinguished. The
first is that the form make the parts into a definite thing, i.e. into one
thing rather than two. The second is that the form make the parts into
the requisite whole, i.e. something which has as its parts the parts
of the given thing. We may now think of the first puzzle as a proof
that the first condition cannot be satisfied and the second puzzle as a
proof that the second condition cannot be satisfied.
How is the new puzzle to be solved, on behalf of Aristotle or the
neo-Aristotelian? And to what extent can the solutions to the two
puzzles be combined? One possible solution to the second puzzle,
which is compatible with any solution to the first, is to reject the as-
sumption that matter and form are mereologically exhaustive of the
compound. There are, in fact, two assumptions here. One is that the
matter and form are parts of the compound, the other is that any
material part of the compound Fm must either have a part in common
with F or be a part of m. Now adherence to a hylomorphic view does
not require that the matter or the form be parts of the thing which has
that matter and form. However, it is clear that Aristotle's hylomor-
phism is not of this mereologically neutral sort.
In regard to the first assumption, he repeatedly says that the matter
is part of the compound. There is also some indication that he believed
that the form is part of the compound (and even without this assump-
tion a version of the argument could still be given). For example, the
most natural reading of the passage at rn23b18-22 of the Metaphysics
suggests that the form, as given by the 'angle', is a part of the bronze
cube. Admittedly Aristotle does not take compounds to be wholes in
the same way as heaps; and he does not take the matter and the form
of the compound to be parts in the same sort of way (thus he writes in
Kit Fine
Metaphysics H. 3, 1043b5, 'the syllable does not consist of the letters

+ juxtaposition, nor is the house bricks + juxtaposition'). But these
disanalogies serve, not to undermine the mereological claims, but to
guard against their being misunderstood.
In regard to the second assumption, it is clear that Aristotle must
take the matter and form to be wholly constitutive of the compound.
For what else could make a distinctive contribution? Of course, this is
not to deny that there are any other parts; the 'heap' a + b + c is
constituted by a and b + c, and yet has b, c, a + b, and a + c as other
parts. However, it does mean that something can be another part only
in virtue of an appropriate mereological relationship to the consti-
tutive parts. But given that this is so, it is hard to see, in the present
case, how some matter which is not a part of m and has no part in
common with F could be a part of Fm. For what could the relevant
mereological relationship to F and m be?
Another solution is to deny the assumption that it is possible for the
matter (at any given level) of a thing to be different at different times.
This denial, namely the thesis that the matter of a thing must stay the
same, may be called Rigidity. Since Rigidity is often confused with
some form of Entrapment, it will be helpful to consider, if only
briefly, the connection between the two.
The one assumption is essentially a converse to the other. For
Entrapment tells us that the thing is the same when the matter is the
same, while Rigidity tells us that the matter is the same when the
thing is the same. Or to be more precise, given the condition that mis
the matter of x at one time and n the matter of cospecific y at another
time, Entrapment then tells us that x = y if m = n, while Rigidity tells
us that m = n if x = y.
Even though Entrapment and Rigidity are logically independent,
there is a plausible implication from the second to the first, i.e. from
migration to variation. For suppose that my elemental matter, let us
say, can migrate. Then it can be the matter of a different person. But
there would appear to be no reason in principle why I should not exist
at the time at which the other person has that matter; and so my
elemental matter at that time will then be different from what it was.
However, it seems to me that the implication in the other direction
is not so plausible. 18 For even though my elemental matter changes
over time, this is not necessarily because some elemental matter has
'got in' or 'got out'; since what one takes to be the entry or exit of the
matter may in fact involve a change in its identity.
18 And hence the argument in Furth's book (1978: r8o, lines 15-21) from
Variation to Migration can be resisted.

It seems clear that the neo-Aristotelian would be committed to
Variation and that therefore the present solution would not be open
to him. There is also some evidence that Aristotle is committed to
Variation (and, as far as I know, no evidence that he is committed to
Rigidity). Perhaps the closest direct evidence is contained in the
passage from Generation and Corrufgtion (32rb25) which has already
been cited in support of Migration. 9 He there writes of new matter
becoming part of the matter of the flesh. But again, the evidence is
not decisive. For let us grant that m is a part of the matter x of the
flesh at one time and that it is not a part of the matter y of the flesh
at another time. Then it would follow that x was not numerically
identical to y if it were in a time-independent sense of part that m was
a part of x and not a part of y. But it may be or fail to be a part in a
time-dependent sense; and, in that case, the non-identity of x and y
could not be inferred. Nor can it be argued, should x and y be the
same, that their matter will change on the grounds that their parts
change. For the sense in which m is a part of x is presumably one
which puts it at the same hylomorphic level as x and so is not relevant
to the identity of its matter.
Apart from the direct evidence, there is some indirect evidence that
Aristotle is committed to Variation. As I have mentioned, there is a
plausible implication from migration to variation; and so the evidence
for the one will, to that extent, carry over to the other. In the case of
the evidence from homonymy, we may argue more directly. For
suppose we ask on what reasonable basis it could be maintained that
my elemental matter remains the same over time? The only plausible
view seems to be that the matter, in such cases, is not something
which we take to vary according to the biological circumstances but
something whose essence is to be the matter of the thing in question.
Thus my elemental matter will be whatever fulfils a certain role at a
sufficiently low level of my make-up; and that matter will grow and
diminish, staying the same throughout, as I grow and diminish.
On such a view, Aristotle would presumably have been willing to
countenance the extension of the doctrine of homonymy to the lower
levels of matter; and so the fact that he appears not to be so willing
suggests that this was not his view and that variation was for him a
real possibility.
But although the textual evidence seems to point in the direction of
Variation, it should be remarked that Rigidity is a perfectly coherent
doctrine and that it has the possible advantage over Variation of not
19 Above, p. 26. The passage has been used in support of Variation by
Anscombe (r953: 83) and by Hartman (I97T 59-60).

Kit Fine
requiring what might seem to be an arbitrary distinction between the

matter which conforms to homonymy and that which fails to conform.
Given the natural affinity of Entrapment to Rigidity, we therefore
have a single docrine about the nature of matter which is capable of
providing a simultaneous solution to both puzzles.
There remains, finally, the question of whether Simple Composition
should be rejected. If it were only a question of solving the present
puzzle, then it is clear that Simple Composition should be dropped in
favour of Plural Composition, since the formerly recalcitrant parts
would thereby be incorporated into the compound. But the original
puzzle also needs to be taken into account.
Under the standard interpetation of Aristotle, we were led to be-
lieve that Composition should be construed as Simple Composition;
and consideration of the present puzzle does not seriously under-
mine this view, for the reasons already given. However, under a non-
standard interpretation, it would be plausible to suppose, in the light
of the present puzzle, that Composition should be construed as Plural
Composition; and this position could then be combined with either
Individual Form or Entrapment.
Still, it should not be thought, in either case, that the difficulties
over understanding unification are thereby removed. To illustrate
these difficulties, let us again use the analogy of form with structure.
Then under the adoption of Individual Form, it would be as if the
structure which moulded the letters 'c', 'a', and 't' into the word 'cat'
were unique to the resulting word; and under the adoption of Entrap-
ment, it would be as if the letters from which the word 'cat' was
constructed were unique to the resulting word. Such a conception of
unity is hard to understand; for it seems as if part of what is meant to
explain the unity, be it the principle or the elements, is itself under-
stood in terms of the unity.
If I am right, the general exegetical situation is rather odd. For the
logical relations among the assumptions of the temporal puzzle are
such that if either Material Migration or Cdmmon Form is denied,
then Simple Composition can be retained; and they are also such that
if Material Migration and Common Form are retained, then Simple
Composition must be denied. But the exegetical relations appear to
be the reverse of these. If it is supposed that Aristotle rejected either
Material Migration or Common Form, then it is plausible to suppose,
given the mereological puzzle, that he also rejected Simple Composi-
tion. On the other hand, if it is supposed that he accepted Material
Migration and Common Form, then it is plausible to suppose, in the
absence of any reasonable consistent alternative, that he also accepted
Simple Composition.~: ,
2
Aristotle on Identity
WILLIAM CHARLTON
In the section of the Treatise entitled 'Of scepticism with regard to the
senses' Hume distinguishes identity and non-identity from unity and
plurality: 'One single object conveys the idea of unity, not of identity.
On the other hand a multiplicity of objects can never convey this idea,
however resembling they may be supposed ... Since, then, both
number and unity are incompatible with the relation of identity, it
must lie in something that is neither of them.' But alas, he goes on, 'at
first sight this seems utterly impossible. Betwixt unity and number
there can be no medium.' Luckily the Humean imagination thrives on
impossibilities. By means of a 'fiction' concerning time it is able to
cobble together 'an idea which is a medium betwixt unity and number,
or more properly speaking is either of them according to the view in
which we take it: And this idea we call that of identity' (r888: 200-1).
Jonathan Bennett (1971: 334) thinks Hume's problem was to see
how there can be statements of identity which are neither trivially
tautologous ('unity') nor self-contradictory ('number'): a problem
which, as he observes, Frege solved by distinguishing sense and refer-
ence. That may contribute to Hume's perplexity, but on the face of it
his point is that while the notions of unity and plurality are clear and
have clear applications, the notion of identity is in some way spurious.
Strictly speaking, he seems to be saying, there is no such thing as what
it is for an object identified on one occasion to be identical with an
object identified on another. I think that Hume's instinct here is both
defensible and Aristotelian. I shall first argue that though Aristotle
has a good deal to say about unity 1 he nowhere tries to say wherein
identity lies. I then try to show that it is indeed wrong to ask 'What
© William Charlton r994
I am grateful for comments on drafts of this paper to the Editors, to the par-
ticipants at the Oxford conference, and to an audience in Budapest that I
addressed in 199r.
1 I shall be concentrating on passages in which Aristotle considers why a
quantity of material is one thing and not several, or is one with another quantity
of material. I shall not be concerned with his answers to two questions which have
held more interst for some of his readers, including David Charles in this volume:
42 William Charlton
makes an Aristotelian substance identified in one way identical with
one identified in another?' but that we can ask the different question
'In virtue of what is a substance identifiable and reidentifiable?' and to
this Aristotle's answer is that it is in virtue of its form. Finally I try to
spike the guns of anyone who argues that Aristotle cannot give this
answer because the form of an Aristotelian substance is a universal.
I
The passage in which Aristotle is generally thought to deal most
explicitly with identity is translated by Christopher Kirwan (1971) as
follows:
Some things are one in respect of number, some in respect of form, some in
respect of genus, some in respect of analogy: in number, those whose matter
is one, in form those whose account is one. (Met. L1, 1016b31-3)
This comes near the end of a chapter on ways in which things are said
to be one; why should we suppose that it bears on identity, that it is
meant to tell us what it is, for example, for the mountain I glimpsed
through parting clouds when walking in the Himalayas to be identical
with the mountain you glimpsed? 2 S. Marc Cohen (1984) offers two
arguments.
The first is that oneness, as against plurality, is a one-place predicate
(x is one), whereas identity as against non-identity is two-place (x is
identical with y). When Aristotle says that things are one in form
'whose account is one' he is talking about a two-place predicate:
Socrates is one in form with Callias. So he should be talking about a
two-place predicate in the words which immediately precede, things
are one 'in number whose matter is one': Aristocles is one in number
with Plato.
This reasoning seems to me quite unpersuasive. For Aristotle is not
considering sentence-schemata (Cohen's term) like 'x is one' and 'x
and y ~re one'. He is just considering the term ' ... one'. This may be
preceded either by a phrase like 'this bundle', 'this water', in which
case the justification for using it would be some kind of continuity, as
explained earlier in the chapter, or by phrases like 'Adeimantus and
(i) Why is the material of an object one with its form? (ii) why are several
different properties or powers united in an object?
2 This question, to which I confine myself here, needs to be distinguished from
another question which is given more prominence in Charles's contribution to this
volume: what determiqes the identity, i.e. the essential nature, of the form (or
matter) of a substance?,
Aristotle on Identity 43
Glaucon', 'dogs and horses', in which case the justification would be
unity of species or genus.
Cohen's second argument is this. At b35-6 Aristotle says that
things which are one in number must be one in form. But he has
earlier (b11- 13) said that things which are one in that they are con-
tinuous need not be one in form. So in the later passage Aristotle
cannot mean by things 'one in number' things which are one by virtue
of continuity without being guilty of 'glaring inconsistency'.
This is to ignore the context of b11-13. The whole passage b11-16
runs:
In a way we say anything is one if it has quantity and continuity, but in a way
not, unless it is a whole, that is, unless it has one form. For example we
should not say the parts of a shoe were so much one if we saw them lying
together anyhow, unless on account of continuity, but only if they are in such
a way as to be a shoe and already have some one form.
At first it may seem odd to think of calling bits of leather lying
together 'anyhow' one thing at all, but Aristotle is probably thinking
that if a shoemaker after cutting out the pieces of leather for a shoe
does not sew them together at once he will tie them into a bundle. In
any case the point of the passage is the one emphasized by Jennifer
Whiting (1986: 263): even the oneness of a continuum is a poor sort of
unity if it depends solely on continuity of matter: an object has the
unity of a continuum most fully when it is an organic unity as well.
Aristotle reverts to this idea in Met. I, 10523 22-5 and I think it lies
behind 1016b35-6: things which have the full unity of continuity must
each have 'some one form' which makes them one in form with
themselves.
If this interpretation seems forced, the traditional interpretation is
harder to sustain than Cohen will acknowledge. It is not just that the
rest of the chapter is about unity and not identity. If these lines are
about identity then the long and careful account of unity through
continuity is overlooked in the recapitulation. Moreover, this chapter
is closely related to L1. 9 and I. I, which both refer back to it. In I. 1
Aristotle again spells out his theory of unity though continuity and
also recurs to another topic of A. 6, units in which we count or
measure; but he does not mention identity or non-identity. In A. 9
there is only one way in which numerical unity of matter can make for
oneness ( 1018"6, alO), and it is surely the continuity of A. 6.
Cohen and others may have been encouraged to think that Aristotle
is talking about identity at 1016°31-3 because of his use of the phrase
'in number'. We ourselves use the phrase 'numerical identity' to
distinguish genuine identity such as exists between the Morning and
the Evening Stars. So when Aristotle talks of numerical oneness we
44 William Charlton
are liable to assume that he is talking about our numerical identity.

But the expression 'numerical' or 'in number' is not well fitted to
express the identity of Aristocles with Plato and his non-identity with
Aristotle. No doubt if Aristocles is non-identical with Aristotle,
Aristotle is a second man and he and Aristocles are two men. But it
sounds odd to say that they are two men 'in number' or that Aristocles
and Plato are 'one man in number'. The great philosophers of the
fourth century are two in number; two is the number of them there
were; but two is not how many of Plato and Aristotle there were.
When Aristotle speaks of oneness in number I think he is always
speaking of unity as against plurality. He sometimes contrasts being
one in number with being two in form, and this oneness in number
coincides with identity. He would say that the marble out of which I
carve a Hermes is one in number with the unshaped thing (Phys. I,
r9obr5, 24) and though his thought would be that they are one thing
and not two, we should say that they are identical. But the notions of
identity and non-identity are expressed in English chiefly by 'same'
and 'other' and the corresponding Greek words are ro avro and
a..t..to or eu:pov (not, according to Met. I, 1054b23-30, bu'upopov since
that implies otherness in properties). Aristotle has three (for him)
fairly full discussions of sameness. In these we see him skirting round
identity, sometimes touching it; but not, I think, saying what identity
for substances is.
In Top. I. 7 he says that things can be the same in genus, like man
and horse, in species, like man and man, and in number, which
happens when 'the names are many but the thing is one' (103 3 9- ro).
The things, he continues, most agreed to be the same are those which
are one (sic) in number, and here too there are three possibilities. (1)
Raiment is the same as apparel: here the 'names' have the same
meaning. (2) Fire is the same as the naturally upwardly mobile. Here
the expressions do not have the same meaning, but one signifies
someihing common and peculiar (Wzov) to what is signified by the
other. (3) Socrates is the same as the seated man 'from what is
incidental' or 'supervenient' (ano rov avµPs/317Koroc;). The last case,
but only the last, is of what we call 'numerical identity'. Aristotle
considers it a less important kind of sameness than (I) or (2) and says
no more about it here than that it is still a genuine sameness in
number ca23-39). But his elaborate account in Met. LI. 6, 1015b16-34,
of incidental oneness, of things one Kara avµf3r,f317K6c; 'by virtue of
supervenience', clearly covers it. The educated is the same as Coriscus
in that the former supervenes on the latter, and the educated is the
same as the just in that both supervene on one substance.
A second chapter ~T1 sameness is Met. LI. 9. Here Aristotle starts
with sameness by virtue of supervenience and his account agrees with
Ll. 6 and Top. I. 7. Things are said to be the same 'of themselves', he
continues, in the same ways in which they are said to be one, that is, if
'their matter is one either in form or in number, or if their substance
is one' (1018a6-7). Their substance is one, we may suppose, if they
are one in form like Callias and Socrates. Their matter is one in form
(or perhaps, rather, in species) if they are composed of wine or water
(1016a17-21, cf. a28). And there is no reason to think that the things
of which the matter is one in number are anything but continua.
Oneness may be a one-place predicate while sameness is two-place.
But as Aristotle says at 1018a8-9, one thing 'may be used as several,
as when we say it is the same as itself' and that which is composed of a
continuous quantity of material is one with itself.
The notion of a thing's being the same as itself reappears in the
third relevant chapter, Met. I. 3:
Same is said in many ways: in one way, when we say that something is the
same in number; in another, if something is one [sic] both in account and in
number, as you are one with youself both in form and in matter; and again, if
the account of the primary substance is one, as equal straight lines are the
same. (1054"32-h2)
And things are other in corresponding ways, for instance:
if it is not the case that both the matter and the account are one, which is why
you and your neighbour are other. (1054b16-17)
Aristotle is certainly speaking here of 'numerical' identity and non-
identity, and he seems to be saying that you are non-identical with
your neighbour because, though you are one with him in account, you
are not one with him in matter. But does he mean that your matter is
non-identical with his or merely that it is discontinuous? Although the
latter may suggest to us the grotesque thought that you and your
neighbour are not Siamese twins, that Aristotle means nothing more
is suggested by his formal definition of oneness as indivisibility at
1052h16, and there is no textual evidence on the other side. Reference
is sometimes made to Met. Z, 1034a5-8:
The whole, this sort of form in this flesh and these bones, is Callias and
Socrates; and it is other on account of the matter-for that is other-but the
same in form-for the form is indivisible.
But again the point could be that the flesh of Callias is not a con-
tinuous lump with the flesh of Socrates (in the way the water of the
Channel is continuous with that of the Atlantic); or that Callias' flesh
is different qualitatively from Socrates', as Aristotle suggests in Met. I.
9, a chapter with linguistic parallels to these lines.
William Charlton
But is Aristotle here saying what it is for a substance identified in

one way to be identical with a substance identified in another? Is he
saying that identity consists in unity of matter and form? At 1052°1-3
he distinguishes the questions 'What sort of things are one?' and
'What is it to be one?' 'To be one is to be indivisible' answers the
second question; 'That which is continuous by nature and a whole' is a
partial answer to the first. If he were to apply this distinction to
sameness, to which question is 'Things which are one in matter and
form are the same' an answer? I shall now argue that he ought to say
it tells us not what identity is but only some of the things that are
identical. Identity for substances is not, as Hume suggests, a kind of
illusion; but (as he might be happy to agree as a second best) it is an
unaccountable brute fact.
II
Aristotle distinguishes, as we saw, between things that are the same
'of themselves' and things which are the same 'incidentally' or 'by
virtue of supervenience'. If A is the same as B by virtue of super-
venience we can say what makes A the same as B in the sense of
'makes' in which being a single girl (as contrasted with love of free-
dom) makes you a spinster. We can find a formal cause, so to speak,
for incidental sameness. This would give Aristotle a means of handling
Fregean identity-statements without distinguishing sense and refer-
ence. To ask if that which shines west of the rising sun is that which
shines east of the setting sun is to ask if the two shinings or positions
supervene on the same planet. But what of planets, as distinct from
bright shiners, or of men, those ontological paradigms? The slayer of
Laius is the same as the father of Antigone if the killing and the beget-
ting supervene on the same man, but if the man who killed Laius is the
same as the man who begot Antigone he is the same, surely, 'of him-
self'; and in that case it is unclear that anything 'makes him' the same.
Aristotle's distinction is similar to, though not identical with, the
distinction. between underived and derivative.· identity. According to
Kant, 'Difference of spatial position at one and the same time is ... an
adequate ground for the numerical difference of the object' (1781: A
263, B 319). He seems to think that objects derive their identity from
places: I am a different man from my twin because the place that I
occupy at any time is different from the place that he occupies at that
time. What about those places themselves? Do they derive their
identity from anything further? Kant does not say, but Newton, whom
he probably has in mind, tells us in the famous Scholium that absolute
times and places 'are,.as it were, the places as well of themselves as of
all other things. All things are placed in time as to order of succession
and in space as to order of situation ... and that the primary places of
things should be movable is absurd.' Kant's view, I think, is that the
places from which objects derive their identity are identical with
themselves and non-identical with one another 'of themselves'. He
feels able to take this strange view because, as he says in the Am-
phiboly, objects are not 'things in themselves' but mere 'appearances'.
A transcendental realist might prefer to say that my place is non-
identical with my twin's because I am non-identical with him: places
derive their identity from their occupants. Aquinas sides with realists
against Kant here, but he does not seem to think that the identity of
substances like human beings is underived. It is derived from thefr
matter: I am non-identical with my twin because my flesh is non-
identical with his, my flesh is non-identical because it is composed of a
different helping of prime matter, and helpings of prime matter, like
absolute places, are non-identical 'of themselves'.
Now if entities of one sort derive their identity from entities of
another, we can explain what constitutes identity for the first in terms
of the second. For Kant occupying the same place, for Aquinas being
composed of the same prime matter, makes the man who killed Laius
the same man as the one who begot Antigone. But nothing can
constitute identity for entities whose identity is underived. If, pace
Kant and Aquinas, human beings are like that, nothing can make the
man who killed Laius identical or non-identical with the man who
begot Antigone: either he just is, or he just isn't.
Why is that? Perhaps we should not be able to say what makes him
the same man; but might there not still be something that makes him
the same? Might not identity be simple and unanalysable, as Moore
(1903: 7) thought goodness and yellowness are?
Moore took goodness and yellowness to be simple, non-relational
predicates. If identity were a simple relational predicate it would be
something which can be had or stood in at a time. When I say 'The
poker was yellow (because heated) on Sunday' I do not say (pace
McTaggart 1927: 14-15) that it timelessly exhibits Sabbatarian yellow-
ness, the fancy property of being yellow-on-Sunday; I say that it on
Sunday exhibited ordinary yellowness. But if I say that the peak
glimpsed by you on Sunday was identical with the peak glimpsed by
me on Monday, I do not say either that it was on Sunday identical or
that it was on Monday identical. Identity is, so to speak, timeless. 3
3 And in this respect differs from composition. Constituents compose what they
compose at a time, and it appears, at least, that the particles which compose one
object at one time, at another could compose another. On this see Kit Fine's
contribution to this volume.
William Charlton
That is not because there is something spooky, what Aristotle

would call something asµvov (cf. Phys. III, 207a19), about identity. It
is because it is not a predicate. That the peak glimpsed by you is to
the north of the peak glimpsed by me can be part of what I say about
it; that it is non-identical can not. At first that may sound false: what
if I say: 'The peak glimpsed by you was not identical with the peak
glimpsed by me'? We need to consider the context. If the lead-up to
my remark is my saying 'The peak I glimpsed was fully 28,000 feet
high' and you say 'No it wasn't: I took a measurement', then it seems
to me that 'The peak glimpsed by you was not the peak glimpsed by
me' is not primarily about peaks; it is primarily about my earlier
words, and I might have said 'I wasn't talking about your peak'. If the
lead-up is your asking 'Did we glimpse the same peak?' perhaps I do
say something about the peak glimpsed by you, but only that it did
not supervene on it to be glimpsed by me.
The point (noted by Austin 1961: 94) is that there are a number of
things that cannot be part of what we say about an object. One is that
what we are saying about that object is false (or true); another is that
we are saying it about that object and not another. 'Aristotle is the
person we are discussing,' Professor Burnyeat said to me by way of
combined counter-example and reproof when I read the first version
of this paper; but it seemed to me that this succeeded only in being a
comment on the discussion from outside, and not a genuine counter-
instance.
There can be no formal cause, then, of underived identity. Inci-
dental sameness is not derived in the way in which, according to Kant
and Aquinas, the identity of human beings is derived, and there are a
number of differences between the incidental-of itself and derived-
underived distinctions. Moreover, there can (Aristotle holds) be a
formal cause of a thing's being of itself one, and unity is, in fact, a
kind of predicate, though an odd one. We can say that the Mediter-
ranean and the Atlantic now are one and undivided, though millions
of years ago they were divided, discontinuous, find therefore two. But
you are the same as yourself and other than your neighbour in a way
that is not nnly not incidental but, it seems to me, underived.
It can be doubted, of course, whether it so seemed to Aristotle.
Those who think we derive our identity from our matter mostly think
this is Aristotelian doctrine. I have argued that it is not stated in the
passages usually invoked, such as 1016h31-3 and 1034a5-8, and there
are other well-known passages which tell against it. Matter is not ra&
u, not 'this something' (implied at Met. Z, ro29a27-30 ), or at least
not in itself (De An. II, 412a7) or in actuality (Met. H, 1042a27-8).
The phrase ra& u m~Y; sometimes mean rather 'a particular sort of
thing' than 'a particular individual', but these and other passages
strongly suggest that in Aristotole's view the entities which are basic
for purposes of identification cannot be mere continuous expanses of
material as such. And as Jennifer Whiting (1986: 362) and A. C.
Lloyd (1970: 519) insist, that is reasonable. A mere quantity of matter,
whether ordinary or prime, has not enough unity to play the role of
that from which a substance derives its identity. If I were to say 'The
ten stone of flesh glimpsed by you in the river are the same ten stone
as were glimpsed by me', I am not saying just that the flesh glimpsed
by you is continuous with the flesh glimpsed by me or part of the same
swimmer. My claim is that all the flesh you saw is identical with all th~
flesh I saw. But how is this totality to be identified except as the flesh
in the swimmer you saw? Locke (1690: II. xxvii. 3) (and perhaps
Hume too in his less guarded moments: see Treatise, 1888: 255-6)
derives the identity of parcels of matter from that of the constituent
atoms; but Aristotle is not that kind of atomist.
Aristotle holds that 'composite substances' such as living organisms
exist in a primary way vis-a-vis properties and relations (Met. Z,
10283 13-31) and that 'to be separate and r6& n belong to them more
than to anything else' (1029 3 27-8). This ought to mean that human
beings and perhaps houses have underived identity: nothing makes a
man identified in one way identical with a man identified in another.
But he could hold that the identity of a substance is underived and
still ask under what description or in what aspect it is a basic identi-
fiable. What an Aristotelian substance essentially is and how it should
be conceived is expressed in an account of the form 'An s is an f of m'
( r6& AK roD&) where 's' ranges over sorta! terms like 'human being',
'house', 'f' over form-expressions like 'sphere', 'shelter', 'thinker',
and 'm' over matter-expressions like 'flesh', 'stones and tiles in such-
and-such an arrangement'. I may refer to a substance by using any of
these descriptions: 'that angry traveller', 'that flesh' or 'those legs',
and 'that human being', which Aristotle takes to signify matter and
form together, a purposive agent consisting of flesh and bone or arms
and legs. Given that Oedipus is a unified, identifiable existent, under
which description does he scoop the ontological jackpot? Since we
may think of him either as a mass of flesh or as an angry traveller, in
which aspect is his identity underived?
It is important to distinguish this question from the question whether
a substance derives its identity from its matter or its form. 4 To ask
4 In sharply contrasting what he calls an 'explanatory' and a 'non-explanatory'
approach (pp. 76-80 of his contribution to this volume), David Charles seems to
me not to take account of this distinction. My own approach, I think, includes
elements of both of his.
50 William Charlton
that is to imply that its identity is, after all, derived, and that either
its matter or its form is prior for purposes of identification. Some
remarks in Aristotle suggest such a view, for instance 1029a5-T 'If
the form is prior to the matter and more of a thing that is, it will be
prior also to that which consists of both, for the same reason.' But this
involves great difficulty. The notions of matter and form are notions
of what something like a human being or a house consists of and what
these constituents make up; they are therefore abstractions from
ordinary sorta! concepts and depend upon them.
On the other hand, Aristotle has two familiar lines of thought which
lead to the conclusion that an object has underived identity in its
formal aspect or in virtue of (Kara, 412a8) its form. There is his theory
of unity (Met. A. 6, I. 1): an object is a unity if it is a continuum, a
continuum if it is moved by action on a part of it, and so movable by
virtue of its structure or, in the case of a living thing, its soul. This
makes an object's form responsible for its being a basic identifiable
somewhat as causal powers are responsible for change. And there is
his theory of being (Met. r. 2 etc.): the primary existents, the things
that are in the primary way, are things like human beings and horses.
For one of these to be a thing that is, is simply for it to be a human
being or a horse. But it is this because of its form. So its form is
responsible (as a formal cause, naturally) for its being a primary
existent and hence a basic identifiable. Oedipus has underived identity
because (formal cause) he is a substance, and he is a substance
because (formal cause) he has a human soul. (To say that his form is
the formal cause of his having underived identity is not, of course, to
say that it is the formal cause of his identity with the slayer of Laius.)
III
I have now argued that, while Aristotle has plenty to say about unity
and plurality, he nowhere tries to say what constitutes identity for
substanc~s, and that he is right not to attempt this: not because a
substance identified in one way cannot really be identical with a
substance identified in another or because the notion of identity is
confused, but because it is not the notion of a predicable relation, and
identity or non-identity for substances is a brute fact. I have also
suggested, however, that he thinks a substance has identity rather by
virtue of its form than by virtue of its matter. This suggestion might
encounter resistance on various grounds but I shall limit myself to one
particular line of criticism.
The form of an Aristotelian substance, it is argued, is a universal,
something common tp all individuals of the same sort. Hence it
Aristotle on Identity SI
cannot be in virtue of its form than an individual is not identical with
another individual of the same sort.
This is really an objection, not to the interpretation I am proposing,
but to any interpretation according to which the identity of an
Aristotelian substance is derived from its form. It would be an ob-
jection to saying that Aristotle makes forms, so to speak, the material
cause of identifiability, the basic entities that have identity, whereas I
am giving them the kind of responsibility which is enjoyed by formal
causes and sources of change. But I do hold that the concept of
Socrates' form is the concept of him in that aspect in which he is
identical with himself and non-identical with Callias. The idea that
Aristotelian forms are universals seems to cast some doubt on this, so
I should like to examine it.
The inquiry is complicated by the fact that Aristotle (like Plato)
uses both abstract and concrete expressions for forms. Scholars can
divide on whether 'sphere' or 'sphericality', 'thing that perceives' or
'what it would be to be a horse' better express his form-concepts.
Those who favour abstract expressions see the matter-form relation
primarily as a thing-property relation, and this is naturally, construed
as one of particulars to universals they instantiate. Those who prefer
concrete expressions see the matter-form relation rather as one of
constituents to things constituted and this is not so naturally equated
with the relation of particulars to things instantiated. But I have
discussed this aspect of the form-matter relationship on other occa-
sions; here I shall argue that the question whether Aristotelian forms
are particulars or universals is radically confused.
The argument which is used, for instance, by J. H. Lesher (1971),
D. D. Colson (1983), and P. K. Moser (1983) to show that Aristotelian
forms are universals goes like this:
(1) If two individuals are both men, they have the same form.
(2) Socrates and Callias are both men.
(3) So they have the same form.
(4) What is had or shared by several individuals is a universal.
(5) So the form of Socrates is a universal.
This argument has an obvious and non-contingent affinity with the
famous argument about largeness in Parmenides 132a: 'If two individ-
uals are both large, there is some one thing, largeness, by virtue
of which they are large.' Here a statement in which something which
is predicated of two concrete objects is transformed into a statement
in which a relation is said to hold between two concrete objects and
an abstract object. One step takes us off firm ground into a bottomless
swamp. And the same step is made in the argument that Aristotle's
52 William Charlton
forms are universals. For it to work we must suppose that the form of
Socrates, Socrates' form, is the form he shares with Callias; and to
suppose this we must suppose that there is something which is his
form and something he shares with Callias. There is nothing wrong
with the phrases 'Socrates' form' and 'the form Socrates shares with
Callias' in themselves. We can say 'The form Socrates shares with
Callias is the form of a man; Socrates' form is that of a man, not that
of a horse.' But we cannot say 'The form Socrates shares with Callias
is Socrates' form', any more than 'The colour wine shares with blood
is wine's colour'.
'Are forms particulars or universals?' sounds like 'Are whales fish
or mammals?', but forms do not belong to the natural kingdom and
universals and particulars are not natural categories. These things
belong to language and thought, and a realistic approach to them
must be in the formal mode. Saying 'Socrates is brown' is a different
kind of linguistic act from saying 'Socrates is a man'. In the first case I
say what Socrates is like (no16<;;), and 'brown', 11eAa<;;, is a colour- or
quality-word. In the second I say what Socrates is (or what he is like
as a substance: Cat. 3b15-21), and 'man', avepwno<;;, is a form- or
substance-word. But just as it does not follow, because there is such a
thing as presenting a crooked appearance, that there are such things
as appearances which are crooked and presentable, so we cannot
infer, because there is such a linguistic act as expressing the form of
an object, that there are such things as forms which are expressible.
'The speaker gave the form of Socrates, not his colour' does not
specify something the speaker expressed but rather the way in which
he spoke of Socrates; the whole phrase 'to give the form of' should be
taken together to signify a form of predication.
And if words like 'colour' and 'substance' have a primary use in
specifying kinds of linguistic act, 'universal' and 'particular' have a
primary use in saying in a different way how words are used. A word
may signify or express something which is common to, instantiable by,
a number of objects; to that extent it signifies spmething general; or it
may be used to refer to a particular individual:
That being so, while we cannot ask whether a form is a particular or
a universal, we can ask whether form-expressions are used to express
what is general, to refer to particulars, or in both ways. In the light of
passages like Cat. 3brn-24 I think Aristotle's answer would be the
obvious one: they are used in both. In 'I trod on a snake' or 'the
snake bit me' 'snake' refers to a particular individual; but in 'If
anything is a snake it is oviparous' it is used predicatively and signifies
something common.
But here the frien~~.- of abstraction may protest. Neither 'snake',
they may say, nor 'perceiver' is a form-expression: only 'what it would
be to be a snake' or 'the ability to perceive' will express serpentil'le
form; and these linguistic items signify universals.
Perhaps they do, but how are they used? The abstract colour-word
'brownness' has a natural use in 'The Greek word 11D.ar; is used to
predicate brownness'. Substance-words are not, in Aristotle's view,
used predicatively in quite the same way as quality-words: they are
not used to predicate something of something independently identi-
fiable; but he still thinks they can be used predicatively, and would
probably agree that the ability to perceive is part, and what it is to be
a snake is the whole, of what we predicate when we so use the word
orpu;. That being so, it will be correct to say that the form of an
Aristotelian substance is a universal if the notion of a substantial form
is the notion of what we predicate when we use a substance-word
predicatively.
That has certainly been widely held, though I do not think it
matches much of what Aristotle says about matter and form. But that
is a separate issue. I could 'give' someone who wants it this use of
'form', grant that forms in this sense are universals, and still maintain
what I want about identity: namely that Aristotle conceives substances
as having underived identity, and having it in what I call their 'formal'
aspect, the aspect in which they satisfy descriptions like 'shelter', 'per-
ceiver', or 'purposive agent', not descriptions like 'bricks' or 'ftesh'. 5
5 This paper has been an attempt to explore Aristotle's views on the concept
we apply in asking, say, whether the peak you glimpsed is the same as the peak I
glimpsed. I have not considered his views on the use of this concept in reasoning
or on principles like Leibniz' Law. For discussions of these matters the reader
may see N. White, 'Aristotle on sameness and oneness', Philosophical Review So
(r971), pp. r77-97; K. Barnes, 'Aristotle on identity', Phronesis 22 (1977), pp.
48-62; F. J. Pelletier, 'Sameness and referential opacity in Aristotle', Nous 13
(1979) pp. 283-3II; G. B. Matthews, 'Accidental unities' in M. Schofield and
M. C. Nussbaum, ed., Language and Logos, Cambridge University Press 1982,
pp. 223-40; F. Lewis 'Accidental sameness in Aristotle', Philosophical Studies 43
(1982) pp. 1-36.
3
Individuals and Individuation
in Aristotle
MARY LOUISE GILL
What is an Aristotelian individual, and what makes the individual

numerically one? These questions cover several different issues, which
we should begin by setting apart.
I
A first issue, which has been widely discussed in connection with the
Categories, is the nature of an Aristotelian individual as distinct from
a universal. In the Categories, Aristotle mentions an individual man
(b ri:; avOpwno:;) and an individual horse (b ri:; 'inno:;), and he
claims that entities of this sort are neither 'said of' a subject nor
'inhere in' a subject; other things are said of or inhere in them. For
example, in the case of an individual man, things that determine what
he is, such as the species man and genus animal, are said of him;
while things that characterize him, such as the qualities whiteness and
knowledge of grammar, inhere in him. Both types of predicables
depend for their existence on the individuals they determine or charac-
terize; remove the individuals, and everything else is removed as well
(2b1-6). In addition to the ultimate subjects, which Aristotle calls
primary substances, he mentions another group of individuals, which
include the individual white ( ro ri l.t:w<oi') and the individual knowledge
of grammar (1/ ri:; ypaµpauKft). He says that these are not said of
anything else but none the less inhere in something else.
Both types of individuals-substantial and non-substantial-are
described as 'indivisible' (aroµa) and 'one in number' (h
dp10wJ)). Aristotle claims that 'without qualification, things that are
indivisible and one in number are not said of a subject, but nothing
prevents some from being in a subject-for the individual knowledge
of grammar is among the things in a subject' (1 116-9). Scholars debate
whether non-substantial items are individuals as tokens of types or as
©Mary Louise Gill r994

56 Mary Louise Gill
fully determinate types, which are repeatable. 1 In speaking of non-
substantial items as indivisible and numerically one, does Aristotle
mean that they are non-repeatable particulars or that they are merely
immune to further division-that is, not 'said of' any more deter-
minate item?
One side of the debate insists that all individuals are non-repeatable;
the other allows that some entities are individual and numerically one,
because they are fully determinate; but they are none the less re-
peatable because they can inhere in more than one subject. To allow
for the possibility that some Aristotelian individuals are repeatable, I
shall use the word 'indivisible' to specify a fully determinate type that
divides a wider determinable kind, 2 and I shall confine the word
'particular' to things that are non-repeatable. 3
II
In addition to the issue of individuals as opposed to universals, there
is a second topic covered by our original questions: What entities
are substantial particulars, and what makes a substantial particular
numerically one thing and not a mere plurality or heap? Numerical
oneness in this context concerns the unity of something composed of
parts. What unifies the parts of a substance at a time and over time?
In Met. z. 16 Aristotle denies that the four elements-earth, water,
air, and fire-are substances, 'Because', he says, 'none of them is one
thing, but like a heap, until they are worked up (literally 'concocted')
and some one thing comes to be from them' (104ob8-rn). In Z. 17 he
claims that form is responsible for making the constituent materials
into some definite object, for example, these materials into a house or
those into a human being (1041b4-9). The present question concerns
the source of substantial unity, and the discussion in Z. 17 makes
fairly clear that Aristotle's answer is substantial form. If scholars
disagree, the dispute is not whether form or something else accounts
for substantial unity, 4 but how to understand the nature of substantial
1 There is a vast literature defending the view that non-substantial individuals
are tokens of types. For one classic treatment, see Ackrill (1963). Versions of the
alternative view are defended by Owen (1965a) and Frede (1987: 50-63). I have
endorsed Frede's interpretation of individuals in the Categories in Gill (1984:
9-14).
2 On this topic, see Frede (1987: 50-4, 63); cf. Code (1986a: 421 and n. 17).
3 In letting 'individual' cover both non-repeatable and repeatable items and
restricting 'particular' to non-repeatable items, I am invoking a distinction in-
troduced by Code (1986a: 414 and 421).
4 An alternative view, that a material power (vital heat) or stuff (pneuma)
accounts for substantial unity, has, however, recently been defended by Freuden-
th::il (rnoA)
Individuals and Individuation 57
form and its causal role in unifying the parts of a substance and in
maintaining the unity throughout the object's career. 5
III
In this paper I want to focus on a third issue: What makes an
Aristotelian particular the particular that it is, and how is one such
particular differentiated from others, which may be qualitatively in-
distinguishable? On several occasions Aristotle contrasts numerical
unity (or sameness) with specific or generic unity (or sameness). We
are likely to be disappointed, however, if we expect these discussions
to reveal a principle of individuation-something that explains the
uniqueness of the particulars.
In Top. I. 7 Aristotle says that two things are the same in species
that are undifferentiated in species, for instance, two men or two
horses. Things classified under the same species are specifically the
same. Things classified under the same genus, such as a man and
a horse, are the same in genus. In what situations do we ascribe
numerical sameness? Aristotle's treatment of numerical sameness
does not parallel his treatment of sameness in species and genus in
asking what makes two objects numerically the same. Instead, he says
that there is numerical sameness when two or more names or descrip-
tions apply to the same thing. In the first and strictest sense, numerical
sameness is indicated when two synonymous names or a name and a
definition apply to the same thing, as the expressions 'mantle' and
'cloak', or 'man' and 'two-footed land animal'. Second, there is
numerical sameness when a name and the description of a proprium
apply to the same thing, as the expressions 'man' and 'receptive of
knowledge', or 'fire' and 'that which is naturally borne upward'.
Examples of a third sort involve accidental descriptions that apply
to one thing, as the descriptions 'the seated one' and 'the musical'
specify the object named by 'Socrates'. Aristotle's discussion of
numerical sameness does not indicate what accounts for an object's
uniqueness or what makes two objects numerically distinct. And his
examples suggest that numerical sameness is not even confined to
particulars, since the expressions 'two-footed land animal' and 'man',
for instance, apply to a species. 6 This treatment of numerical same-
ness does not bear on individuation.
One example in Top. I. 7 might, however, appear to bear on that
issue. Aristotle considers whether, when drafts of water from the
5 On this topic, see the papers in this volume by Charles, Haslanger, and
Sc~l~as, My views on unity are given in Gill (1989), esp. chs. 5 and 7.
58 Mary Louise Gill
same fountain are called 'the same', the type of sameness corresponds
to one of those already mentioned-sameness in number, species, or
genus. All instances of water, he says, are the same in species because
they share a certain similarity. Although drafts from the same fountain
display a more striking likeness than those from different sources,
they remain instances of specific sameness (103a20-3). Unfortunately,
Aristotle does not tell us how he differentiates the two drafts. Some
Aristotelians would argue that matter accounts for the individuation;
others attribute this function to form.
IV
This is the debate on which I shall focus. Does Aristotle regard matter
or form as the principle of individuation?
I find the debate troubling for two reasons. First, quite apart from
Aristotle's theory, I doubt that we can explain what makes a par-
ticular the particular that it is. An account of particularity aims to
explain why one object, which may be qualitatively indistinguishable
from another, is none the less distinct. Attempts to answer the question
lead to the following dilemma. 7
Consider the example made famous by Max Black. 8 Let us suppose
that the universe contains only two objects-a pair of spheres that are
exactly alike, each made of chemically pure iron, with a diameter a
mile wide, the same temperature, colour, and so on. The two spheres
have all their properties in common. How are they individuated?
There are two spheres, but what distinguishes one from the other?
The problem for any general account of particularity-whether in
terms of matter, form, spatio-temporal location, or whatever-is that
it will apply indifferently to each case. This is so for any general
account. Whatever makes a human being a human being makes
Socrates and Callias human beings for the same reason. What ac-
counts for them as human beings does not distinguish them as par-
ticulars. Similarly, the properties that identify one of Max Black's
spheres also identify the other. So the first side of the dilemma is this:
A general account of particularity, like any general account, specifies
all particulars indifferently without discriminating them.
If an appeal to matter or form is supposed to enable discrimination,
then we need to know how the matter or form of one object differs
7 I thank Paul Coppock for posing the dilemma to me.

8 Black (1952) in Loux (1976: 253-4). The example derives from Jonathan
Edwards ([1754] I95T 387-92). I thank Calvin Normore for calling my attention
to this text.
from that of another. What differentiates the matter of Socrates from
that of Callias? Must we presuppose the particularity of the feature
that accounts for individuation? If we do, we can individuate one
object by means of another, but the account is not explanatory, since
it presupposes the particularity it aims to explain. If the uniqueness of
the individuator is secured by something else, for instance, a par-
ticular quantity of matter by its spatio-temporal location, then matter
is not the principle of individuation, and the new principle-spatio-
temporal location-is subject to the same question. In a relativistic
universe, like Aristotle's and our own, what accounts for the unique-
ness of spatio-temporal locations? Notoriously, the task of individuat-
ing the individuator generates a circle-for instance, a piece of matter
individuated by its spatio-temporal location, and the spatio-temporal
location individuated by the body contained at that location. 9
So the dilemma is this: either the account of particularity is explana-
tory but applies to all particulars indifferently without illuminating
their difference, or the account differentiates the particulars but is not
explanatory, because it presupposes the very thing that it is meant to
explain.
My first objection to the debate about individuation, then, is the
apparent futility of attempts to explain it. My second objection is that
the evidence used in the Aristotelian debate is underdetermined. To
my knowledge, no text demonstrates that Aristotle defended matter
or form as the principle of individuation-something that guarantees
the uniqueness of the object whose matter or form it is.
The evidence does, however, appear to support a weaker thesis that
avoids the dilemma and gives Aristotle an apparent means of in-
dividuating physical objects. If particularity cannot be explained, as
the dilemma suggests, a plausible strategy is to accept certain entities
as the basic particulars and to individuate others with reference to
them. Although this conception of individuation is not explanatory,
since the particularity of the basic objects is assumed in advance, it
allows for progress, because one entity is individuated by another. Let
us call this conception weak individuation. 10 I will argue that the
passages in which Aristotle says that objects that share the same form
are differentiated by their matter, and that unity of matter makes two
such objects numerically one, can best be understood as treating
9 Cf. Black (1952) in Loux (1976: 256).

10 I thank Kit Fine for calling my attention to the significance of this alternative.
The distinction I am using between strong individuation, which is explanatory,
and weak individuation, which is not, is different from Cohen's (1984: 42 n. 3)
distinction between strong and weak versions of the principle of individuation.
60 Mary Louise Gill
matter as an individuator in the weak sense. However, as we shall see,
matter is inadequate for individuation, even on the weak conception,
and Aristotle presents an argument that shows he knew that it was.
Let us now tum to the evidence. We shall first consider the evidence
used to support the claim that matter is the principle of individuation,
then that brought forward in favour of form.
v
Scholars defending matter as Aristotle's principle of individuation
commonly cite a text at the end of Met. Z. 8, where he says:
And that which is already whole, such and such form in these instances of
flesh and bone, is Callias and Socrates; and the whole differs on account of
the matter (for the matter is different), but it is the same in form (for the
form is indivisible). (rn34"5-8)
They also cite a text in Met. L1. 6, which says:
Some things are one in number, some in species, some in genus, and some by
analogy-in number things whose matter is one, in species things whose
account is one, in genus things whose scheme of predication is the same, and
by analogy things related as two further things. (rn16b31-5)
In the first passage Aristotle says that, although Callias and Socrates
share the same indivisible form, they differ on account of their matter,
because their matter is different. This claim can be construed in two
main ways. First, he could be saying that Callias is distinct from
Socrates because his constituent matter differs qualitatively from that
of Socrates. 11 Obviously Aristotle does not mean that their bodies
differ in kind, since he has just said that each whole is a certain form
realized in 'these instances of flesh and bone'. The two individuals
share the sort of matter proper to human beings. If their constituent
materials are qualitatively distinct, the two individuals differ because
of material differences in their flesh and bone-such as texture,
markings, and so on.
Not all individuals can be distinguished by qualitative material dif-
ference, however. Identical twins and Max Black's pair of spheres
have qualitatively indiscriminable matter. 12 In z. 15 Aristotle con-
11 This interpretation is proposed by Charlton (1972: 244-5), who finds support

for it in Met. I. 9, ro58br- II, which treats qualitative material difference. The
proposal is criticized by Cohen (1984: 47-50), and Charlton defends a different
reading in his paper in the present volume.
12 This inadequacy of the qualitative version of the thesis is mentioned by
Whiting (1986: 360).

siders the possibility that a second object might have all the features
of the sun. Clearly, he says, such an object would be a sun (104oa33-4).
This passage suggests that Aristotle thinks that two objects could be
qualitatively exactly alike. If matter individuates such objects, it must
do so on some basis other than qualitative difference.
A second way to construe the claim in Z. 8 is to take it as saying
that Callias' and Socrates' flesh and bones, though qualitatively the
same, are numerically different. The use of the demonstrative and
plurals in ra'ia& rail; aapc) Kai llarol;- appears to support this proposal.
But this alternative can itself be construed in two ways.
On several occasions Aristotle attributes unity to things whose
material parts are continuous. Continuity in an artificial sense is dis-
played by things whose parts are bundled or glued or otherwise
in contact, natural continuity by things whose parts are naturally
continuous and whose proper (Ka(}' a6r6) motion is one (Met. L1. 6,
1015h36-1016a6). 13 Things are distinct that lack such continuity
(IO 17"3-4). In saying that Calli as and Socrates differ because their
matter is different, Aristotle could mean that the matter of the two
individuals is not continuous. 14 But Jack of material continuity, though
enabling us to discriminate two individuals, does not explain their
particularity. If presented with Max Black's imaginary universe,
Aristotle could use material discontinuity to account for the presence
of two spheres rather than one, but this criterion would not enable
him to determine which of the spheres is which. On this interpretation
matter individuates in the weak sense, but it does not explain the
uniqueness of the individuals.
Those defending matter as a principle of individuation require a
stronger notion of numerical unity, such that quantities of matter are
somehow indexed and can therefore explain an object's uniqueness.
But if Aristotle thinks that the numerical distinctness of Callias' and
Socrates' matter explains the particularity of the two men, his theory
is objectionable for the sorts of reasons mentioned in the second half
of the dilemma discussed above. Let us assume that the matter of
Callias and Socrates makes them numerically distinct. Since their
common form is realized in different quantities of matter, they are
discrete particulars. What makes each quantity of matter the particular
quantity that it is? If one answers that each piece is composed of
numerically distinct elemental ingredients, the question can be asked
again: What makes each elemental ingredient the particular that it is?
13 Cf. Met. I. I, 1052a19-21.

14 For this reading of the passage, see Charlton's paper included in this volume
(sect. I). Cf. Furth (1978: 643), who claims that matter is the source of plurality.
Mary Louise Gill
This line of questioning either continues indefinitely, yielding at each

stage a more ultimate matter, or it finally stogs with some ultimate
stuff. Aristotle adopts the second alternative, 5 though scholars dis-
agree about the point at which the analysis ends. 16 If matter is the
principle of individuation, the ultimate stuff, whatever it is, should
account for the particularity of an ordinary physical object. A physical
object will then be a compound of layers of form characterizing the
simple material. On this view the quantity of matter accounts for the
object's particularity, and the object's proper form determines what
the object is.
Yet we can still ask: What makes a particular hunk of ultimate
matter the particular that it is? If Aristotle answers that the two
hunks of ultimate matter travel different spatio-temporal paths, then
the account of individuation through matter is not primitive, since
spatio-temporal continuity rather than matter is the means of dis-
criminating individuals. And this alternative, too, is open to the same
question. How do we individuate spatio-temporal locations? Since
place, on Aristotle's view, is the inner surface of the surrounding body
(Phys. IV. 4, 212a5-6), places are differentiated with reference to
bodies. And since time is simply the measure of change (Phys. IV. I 1,
219br-2), times, too, are ultimately differentiated with reference
to bodies. If Aristotle regards matter as the principle of individua-
tion, capable of explaining an object's uniqueness, his account is
inadequate.
Before attributing this view to Aristotle, we need evidence that he
regarded matter as somehow indexed. I doubt that there is such
evidence. Without it we are left to choose between the other two
interpretations of the passage in Z. 8. On the first interpretation
Aristotle's claim lacks generality. Although Socrates and Callias can
be differentiated by qualitative differences in their matter, this cri-
terion will not work for identical twins, two drafts of water from the
same fountain, or Max Black's pair of spheres, which have qualitatively
identical matter. This lack of generality seems a fairly serious short-
coming. According to the second interpretation, which construes dif-
ference in matter as material discontinuity, Aristotle can differentiate
particulars of the same type, including Max Black's spheres, by
15 See Met. a. 2, 9943 1-h6.

16 The traditional view is that the ultimate stuff is prime matter. This view has
been challenged by King (1956), Charlton (1970: appendix), Charlton (1983), and
Gill (1989: ch. 2 and appendix), who argue that the four elements are Aristotle's
ultimate matter. It is also disputed by Furth (1988: 221-7), who argues that the
ultimate materials are the hot, the cold, the wet, and the dry, which he calls
Aristotle's 'ultrasimple~:.
pomtmg out that their bodies do not compose a single continuous

solid. On this view matter is a weak individuator, because it accounts
for the numerical distinctness of the particulars. But it does not
explain their particularity.
This seems a plausible interpretation of the passage, given the
various places in which Aristotle attributes unity to things whose
matter is continuous and plurality to things whose matter is discon-
tinuous. We should, however, note that the interpretation suggests
that Aristotle regards matter as ontologically basic, since both the
composite and its substantial form will be individuated with reference
to it. Is this a view Aristotle might plausibly have held?
It is at least a view he entertained. 17 Recall that in the Categories
Aristotle treats physical objects, such as a particular man and a par-
ticular horse, as the basic particulars. The existence of everything else
depends on them. But the Categories ignores, or fails to recognize,
the inner complexity of these objects. Once Aristotle addresses the
problem of substantial change in the Physics, he acknowledges that
physical objects are internally complex. Such complexity is needed to
answer the Parmenidean challenge against coming-to-be. Parmenides
had argued that something cannot come to be from nothing, and
Aristotle agreed with his predecessor in excluding such absolute
becoming. Aristotle also recognized that sheer replacement-the sub-
stitut\on of one item for another-is not an adequate answer, since
Parmenides could object that substitution involves the destruction of
the first item into nothing and the emergence of the second out of
nothing. To answer Parmenides, Aristotle proposed that coming-to-be
involves both the replacement of something and a continuant that
survives the replacement. To explain the generation of physical
objects, he posited matter, which pre-exists and survives the emerg-
ence, and form, which emerges with the object and determines what
the emergent object is. The generated product is a composite of
matter and form.
17 I have elsewhere (Gill r989) argued that Aristotle explores this view in Met.
z and H but rejects it in H. 6 and e. Because exposition of the view I attribute to
Aristotle in H. 6 and e would unduly complicate the present paper, I will not here
discuss the implications of those revisions for individuation. Let me simply say
that I take Aristotle ultimately to defend separately existing physical objects as
ontologically basic. They are given in experience as the basic particulars, and the
philosopher's job is to explain various things about them, including what they are
and how they differ from one another. In the project of differentiation, different
criteria can be used in different cases: e.g. human beings and horses are differen-
tiated by their species-forms, Socrates and Callias by differences in their indivi-
sible forms, and identical twins by accidental features.
Mary Louise Gill
By treating matter within a composite as a subject in which sub-

stantial form is temporarily realized, Aristotle apparently transfers to
matter the role played in the Categories by physical objects. The form
of a composite is ontologically dependent on the matter in which it
occurs: remove the matter, and the form is removed as well. On this
view the composite is posterior to both its matter and form, since it
consists of these two more basic components.
Aristotle's theory of generation is clearly relevant to our passage
in z. 8. The passage comes at the end of two chapters in which he
has been discussing the role of matter and form in substantial change.
Z. 8 has been arguing that a composite is generated when a form
is imposed on matter. Since the matter pre-exists the generation
and survives the destruction, it can exist on its own apart from the
form of the object it temporarily constitutes. This treatment gives
Aristotle reason to think that when the same indivisible form is
realized in two discontinuous parcels of matter, the individuals are
distinct because of their matter. It therefore seems reasonable to
construe the passage at the end of Z. 8 as treating matter as the source
of plurality.
Let us now consider the passage quoted above from Met. L1. 6. Here
Aristotle says that things are one in number whose matter is one. He
goes on to say that things one in number are also one in form or
species (1016b36). The passage, though similar to that in Top. I. 7, in
distinguishing numerical, specific, and generic unity, is more helpful,
because it mentions unity of matter as the source of numerical unity.
What sort of material unity does Aristotle have in mind here? Like
material difference, which we discussed in connection with z. 8,
material unity can be understood qualitatively or strictly numerically.
Our passage in L1. 6 is presumably not about qualitative sameness. If it
were, identical twins, whose matter is indistinguishable, would count
as one in number; and two drafts of water from the same fountain,
which were said in Top. I. 7 to be merely one in species, would now
count as numerically the same.
It seems more likely that Aristotle is speaking of numerical oneness.
Still, nothing in the chapter indicates that he thinks that matter
guarantees the uniqueness of the object whose matter it is. Met. L1. 6
talks at length about continuity as a source of unity ( 1015b36- 1016a17)
and mentions lack of continuity as a source of plurality (1017"3-4). 18
So, in saying that things are numerically one whose matter is one,
Aristotle probably means that things identical in form are numerically
one, if their constituent matter displays a natural or artificial con-
tinuity. 19 As in Z. 8, matter appears to individuate in the weak sense,

but it is not a principle of individuation.
But if this is Aristotle's position, it is unsatisfactory. Suppose that a
bronze statue of Socrates is melted down, and the bronze reworked
into a replica of the original statue. 20 If I have characterized Aristotle's
position correctly, the two statues of Socrates, which share the same
indivisible form, will be numerically identical, because the matter of
the first persists as the matter of the second. Aristotle's theory of
change not only allows, but apparently demands, that matter can
migrate from one object to another. Given the possibility of material
migration between objects that share the same form, and assuming
the individuative role of matter, it turns out that two objects, which
Aristotle would presumably want to distinguish, are numerically one
and the same. How would he address this objection? 21
Aristotle does not directly deal with the problem as stated, but he
does address a related issue, concerning non-substantial changes. In
Phys. V. 4 he considers on what conditions changes are one in genus,
form, or number. He says that two changes are one in genus according
to the scheme of the categories. For example, a locomotion is one in
genus with every other locomotion, but different in genus from an
alteration. Two changes are one in form if, in addition to being one
in genus, they are also one in indivisible form (ar6µq1 d&i). For
example, a blackening differs in form from a whitening, but two
whitenings are the same in form. It is not sufficient for formal identity
that the end-points of the change are the same, because motion in a
circle between two points and motion in a straight line between the
same two points would then count as one in form. To be one in
19 This view is defended by Charlton (1972: 242-4), and criticized by Cohen
(1984: 44-7). Charlton addresses these objections in his contribution to this
volume. I agree with Cohen that Aristotle is concerned with the two-place
predicate 'x is one with y in respect of number' (this seems clear from Aristotle's
consideration of oneness in other respects-in species, genus, and by analogy).
But this agreement need not entail rejecting Charlton's view that numerical
oneness concerns the continuity of matter; we need only reject Charlton's example
(about two adjoining seas), which is misleading. Cohen objects that two shoes
joined at the soles would count for Charlton as one in number. The response to
this objection is that, if we say that x and y are one in number, 'x' and 'y' specify
an object in two ways, not two components of it. x and y (e.g. the Morning Star
and the Evening Star) are one in number, if their matter is continuous.
20 Similar examples are discussed by Cohen (1984: 61-2 and n. 33) and Whiting
(1986: 365-9). Aristotle's treatment of the individuation of changes in Phys. V. 4

(see below) tells against Whiting's proposal that the problem can be solved only
by introducing particular forms.
21 For the puzzle, and various ways to solve it, see the paper by Fine in this
volume.
66 Mary Louise Gill
indivisible form, the entire track (called by Aristotle the £v (~)

must be the same.
To explain oneness in number Aristotle appeals to three factors, the
subject of the change (o), the track (ev ~' and the time (on:)
(227b23-4). Two changes are one in number only if the subject, the
track, and the time are the same (227b29-32). This solution requires
clarification, however ( 2283 3- 19). Suppose that Socrates recovers
from ophthalmia in the morning, suffers a relapse at midday, and
recovers again in the afternoon. Are the two recoveries one in
number? The subject is the same, the track is the same, and the time
taken for the recovery is the same. Aristotle does not treat times as
indexed, so he cannot appeal to absolute times to distinguish the two
changes. Time is simply the measure of change. In Phys. V. 4 he says
that a change is numerically one if it occurs in one stretch of time. A
temporal stretch counts as one if the time is uninterrupted (227b30- I),
interruptions being marked by periods of rest (228b3-7). In the case
of Socrates' two recoveries, Aristotle concludes that they are numeri-
cally distinct, because the time is interrupted: between the two re-
coveries the subject undergoes a different change, which breaks the
continuity of time. 22
Although Aristotle does not discuss substantial generation and des-
truction in Phys. V. 4, he could presumably extend the account in that
chapter to cover the case of material migration mentioned above. If
two statues of Socrates are made out of the same bronze at different
times, the statues are distinct because the time during which the
matter constitutes the two is interrupted. In the interval the bronze
survives the destruction of the first statue and the generation of the
second.
If this is Aristotle's answer to the puzzle about material migration,
then continuity of matter is not sufficient even to account for weak
individuation. Continuity of time is also required. The passages in
Met. Z. 8 and -1. 6 cannot represent Aristotle's considered view on
individuation, because his own solution in Phys. V. 4 reveals the
inadequacy of that position.
VI
Was Aristotle then committed to form as the principle of individua-
tion? One immediate objection to this suggestion is the claim in Met.
Z. 8, which we have already discussed, that Callias and Socrates differ
22 I have discus~d Phys. V. 4 in more detail in Gill (1984: 14-21).

on account of their matter but are the same in their indivisible

(aroµov) form. If Callias and Socrates share the same indivisible
form, how can form account for their difference? 23 Still, we should
perhaps ignore this evidence, since we have seen that matter fails to
account even for numerical unity and plurality. Perhaps Aristotle can
do better by appeal to individual forms.
Various doctrines of individual form have been attributed to
Aristotle. Some interpreters think that he is committed to forms that
have been differentiated below the species level. 24 Recent work on his
embryology suggests that to account for the transmission of parental
characteristics in reproduction, he must differentiate the species form
into individual forms. 25 But such forms, though individual in the sense
that they are immune to further division (and thus 'indivisible'), are
still repeatable even if they are not in fact repeated. 26 Presumably, if
Aristotelian reproduction were perfectly successful, the father's form
would be duplicated in his offspring. On this version of the proposal,
form is not a principle of individuation.
A different version of the thesis better suits the claim that form
accounts for the numerical uniqueness of an individual. Some inter-
preters think that your form and mine can be exactly alike yet be
numerically distinct. 27 These commentators regularly defend their
thesis by appeal to passages in the Metaphysics in which Aristotle
refers to forms as rb& n, or 'this'. 28 These references, they say,
23 Frede and Patzig (r988: i. 87) translate cf<h; here as 'Spezies' to neutralize
the passage, but Z. 8 has been discussing the generation of composites through
the imposition of form on matter, and that discussion provides little justification
for introducing a new translation of cl<io:; here.
24 See e.g. Irwin (1988: 252-5).
25 See Balme (1980) and Cooper (1988). At a recent conference at Villanova
University, Alan Code (1993, unpublished) argued that GA IV. 3, which presents
Aristotle's account of inherited characteristics, can be interpreted without appeal
to forms below the species level. In my comment on Code's paper for the same
occasion (Gill 1993, unpublished), I tried to show that Aristotle viewed at least
sexual difference as a formal differentiation below the species level. But Code has
convinced me that making the case for individual forms in the biology is not as
straightforward as some interpreters have thought.
26 Irwin (1988: 574 n. 25; cf. 265) rejects this claim, because he believes
that the spatio-temporal characteristics that make the individual unique and
unrepeatable must also be part of the individual form.
27 Sellars (1957); Hartman (l97T 60-4); Frede (198r esp. 63-71); A. C.
Lloyd (1981: 24-7); Whiting (1986: 367): Frede and Patzig (1988: esp. i. 48-57);
and Witt (1989: 176, 178). An assessment of the arguments for this type of
individual forms is given by Irwin (1988: 569-70 n. 1). For arguments against the
view, see Loux (1991: esp. w2-4, 223-35).
28 Met. LI. 8, w17"24-6; ff. T, ro42"28-9; e. 7, w49"35, ..1. 3, 1070"1T-12 and
"13-18; cf. z. 3, w29"27-30: see also GC I. 3. 318°32-3.

68 Mary Louise Gill
indicate that he regarded forms as particulars. 29 On this view, matter
cannot be the individuating factor, because Aristotle sometimes
denies that matter is a rb& n and allows that, at best, it is a rb& n
in potentiality only. 30 His use of the expression rb& n is taken to
indicate that forms are particular, matter not. 31
Proponents of this view of individual forms often simply assume
that rb& n means 'particular', but the assumption remains disputable. 32
There are two general ways to regard the morphology of the expres-
sion. 33 On one interpretation rb& n is modelled on such phrases
as avOpwnbc; nc; ('a certain man'), with rb& corresponding to avOpwnoc;,
and n to nc;. On the other interpretation rbifa is demonstrative, and n
indicates what the subject is. Neither interpretation requires that the
phrase mean 'particular'. Since the expression that indicates the kind,
whichever of the two it is, can specify a kind at various levels of
generality, the other expression can indicate either a particular that
falls under the kind or a division of that wider kind. For example, the
phrase 'a certain animal' and the phrase 'this animal' can each pick
out a specific type of animal as well as a particular token. So the
meaning of rb& n is not determined by its morphology. Since the
meaning of r6<51 n as 'particular' is open to question, Aristotle's
reference to forms as rb& n is insufficient reason to conclude that they
are particulars. He might mean that forms are determinate types
falling under a wider determinable kind. Forms might thus be
indivisible but not particular.
The favourite passage for the defence of particular forms is Met. A.
5, where Aristotle states that different substances have different causes
and elements. 34 He says: 'The causes and elements of things in the
same species are different, not in species but because the cause of
particulars ( uvv KaO' liKacrwv) is different-your matter and form
and moving cause and mine-but they are the same in universal
account' (rn71a27-9). This passage provides the best evidence avail-
29 Frede (1987: 65); A. C. Lloyd (1981: 38-40); Whiting (1986: 373 n. 2); and
Frede and Patzig (1988: i. 52).
30 Met. H. 1, 1042•27-8, De An. II. 1, 412•7-8; cf. Met. A. 3, 1070•9-10.
3 1. See Whiting (1986: 362-3).
32 For objections see e.g. Zeller (189T i. 369-71 n. 6); Cherniss (1944: 351-2
n. 261); Owens (1978: 389-90); Albritton (195T 701-3); Driscoll (1981: 136-7 n.
39); Lear (198T 150-2); Gill (1989: 31-4); and Loux (1991: 143-6). Irwin (1988:
570 n. 3) responds to some of these objections.
33 See Frede and Patzig (1988: ii. 15); Gill (1989: 31). An earlier discussion
occurs in Smith (1921: 19).
34 Albritton (1957: 700-1; 707); Hartman (197T 63-4); A. C. Lloyd (1981: 6);
Whiting (1986: 373-4 n. 2); Frede (198T 65); Frede and Patzig (1988: i. 52); and
Irwin ( 1988: 253).
able that Aristotle was committed to particular forms. But it is not

very good evidence. 35 Earlier in the chapter he claimed that the
source ( apx~) of particulars is particular ( rb Ka(}' ii1warov), saying
that, while human being is the source of human being generally,
Peleus is the source of Achilles, and your father of you (1071a20-2).
Fathers are sources as moving causes. Peleus and your father are
indeed particulars, just as Achilles and you are, since they are physical
objects. In the later passage, Aristotle does not say that your form
and mine are particulars; he says only that they, along with our matter
and moving cause, are different (aVo). Suppose that you and I are
siblings. Since we share the same father, our moving cause can be
different only in the peculiar sense that your father is yours and my
father is mine. If this is how we must distinguish our moving cause,
our matter and form should presumably be distinguished on the same
grounds. You and I are the particulars, and our causes and elements
are individuated with reference to us. Far from supporting a commit-
ment to forms that are in themselves particular, this passage appears
to support a notion of proprietary matter and form-matter and form
individuated with reference to the objects whose matter and form they
are. Here physical objects are treated as the basic particulars.
Still, let us suppose that Aristotle did regard substantial forms as
in themselves particulars. Are they suited to be principles of indivi-
duation? If form explains an object's individuality, we need to know
how the form of one object differs from that of another. We are
assuming that two particular forms can be qualitatively exactly alike.
Some defenders of the thesis will respond that the forms of Callias
and Socrates differ because they are realized in different parcels of
matter. 36 But then form is not after all the principle of individuation,
since the matter, rather than the form, differentiates the particulars.
Alternatively, one might argue that the spatio-temporal history of the
form determines its distinctness. 37 Again, the form fails to be the
source of individuality. And since, as we saw earlier, the question
remains concerning the individuation of matter and spatio-temporal
locations, none of these proposals explains the numerical discreteness
of the individuals.
Alternatively, we could regard the particularity of form as basic. On
35 My reading of the passage is similar to Lesher's (1971: 174-5). For alterna-
tive interpretations that avoid particular forms, see Modrak (1979: 376-7) and
Loux (1991: 233-4).
36 Hartman (l97T 63-4). Frede adopts this view for synchronic individuation
(r~8T 78).
• 7 See Frede (1987: 68-9), who is particularly interested in diachronic indivi-
duation; cf. Frede (198T 78); and Frede and Patzig (1988: i. 45-6).
Mary Louise Gill
this view, as on the analogous proposal for matter, one thing can be
weakly individuated by another. 38 But if we are to accept this view,
we need better evidence for it than the passage discussed above from
Met. A. 5 and Aristotle's claims that forms are rb& r1. Furthermore,
Met. Z. 8 is not alone in claiming that two objects can share the same
indivisible form (awµov t:fJor;). As we have seen, Aristotle uses
the same phrase in his treatment of changes in Phys. V. 4. Socrates'
two recoveries from ophthalmia share the same indivisible form. Had
Aristotle believed that the two recoveries are differentiated by their
particular forms, why did he appeal instead to the discontinuity of
time?
VII
The Categories indicates that, at some point during his philosophical
career, Aristotle regarded physical objects-things like a particular
man and a particular horse-as ontologically basic. He calls them
primary substances and claims that they are the ultimate subjects on
whose existence that of everything else depends. But outside the
Categories, when he addresses the problem of substantial change in
the Physics, he appears to promote matter to the status enjoyed in the
Categories by physical objects. As the continuant through substantial
change, matter is the basic subject. Both the form and the compound,
in their different ways, depend for their existence on it. Relying on
this conception of matter as basic ontological subject, Aristotle some-
times appeals to difference of matter to differentiate physical objects,
and to unity of matter to identify them.
But as we have seen, because matter serves as continuant in sub-
stantial change, it cannot successfully account for even the numerical
identity and plurality of physical objects. If matter can migrate from
one object to another, and if the two objects share the same in-
divisible form, they will be identical unless something other than
matter accounts for their difference. In the Physics Aristotle differen-
tiates the two items, not by appeal to particular forms, but by appeal
to the discontinuity of time. This treatment suggests that his project of
individuation is more ad hoc than is commonly believed, involving
different criteria in different cases. Horses and men are differentiated
by their forms, Callias and Socrates by the discontinuity of their
matter, and duplicate objects, composed at different times of the
same matter, by the discontinuity of time.
38 This seems to be Whiting's view, since she acknowledges that particular
forms, as she construes them, presuppose the individuality they are meant to
explain (1986: 373).
Individuals and Individuation TI
We may be disappointed by Aristotle's appeal to temporal dis-
continuity to account for recalcitrant cases. But we should be seriously
troubled by the various passages in which he appeals to matter as the
source of numerical unity and plurality. Not only will the proposal not
work, but it implies that matter is ontologically basic in Aristotle's
system. When he considers the ontological primacy of matter in Met.
Z. 3, he expresses serious reservations (1029a26-30). Nevertheless, he
appears to rely on precisely that conception in such passages as the
one we examined at the end of Z. 8. I suggest that the conception of
matter in that passage is not anomalous, but reflects a view whose
implications Aristotle explores throughout Met. Z. 39 I also suggest
that it is a view he ultimately rejects. Met. H. 6 and e appear to rely
on a different conception of matter, based on a reassessment of its
role in substantial change. But that story, and its implications for
individuation, must be the topic for another occasion. 40
39 The conclusion of z. 3 appears to leave open the possibility that matter is

ontologically basic: notice Aristotle's use of the potential optative at 1029a29. The
chapter does not exclude the claims of matter to be substance as subject; it merely
says that on grounds of separation and thisness, the form and the composite
would seem (<51>(D11:1• lh) to have a stronger claim to be substance than matter
does. This tentativeness is important. I discuss this passage and the problems
matter poses for Aristotle's theory of substance in Zin Gill (1989), esp. chs. 1 and
4.
4° For a brief statement of my view, see above, n. 17. See also Gill (forth-
coming). Versions of this paper were read at a conference entitled 'Logic and
Metaphysics in Aristotle and Early Modern Philosophy' at Duke University, at a
Workshop on Ancient Philosophy at the University of Texas at Austin, at Ohio
State University, and at the University of New South Wales, Australia. I thank
those audiences for questions and discussion. I owe special thanks to Paul Coppock
for posing the central dilemma to me. Jim Lennox and Edwin Allaire gave me
many valuable comments. I particularly thank Kit Fine, whose criticisms caused
me to rethink and substantially revise the paper.
II
The Unity of
Composite Substance
4
Matter and Form: Unity,

Persistence, and Identity
DA YID CHARLES
I. INTRODUCTION
In this paper, I shall contrast two ways to interpret the relations be-
tween matter and form, potentiality and actuality in parts of Aristotle's
Metaphysics Z, H, and e. I shall not argue that either of these
approaches successfully resolves all the problems raised by Aristotle's
intricate discussion. Nor do I believe that together, at least in the
formulations I shall consider, they exhaust all the relevant exegetical
possibilities. My aim is simply to examine the distinctive ways in
which they analyse some controversial texts and problematic issues in
this complex·and disputed area.
Aristotle talked of 'matter/form' and 'potentiality/actuality' in dis-
cussing the unity (both at a time and over time ) 1 of particular com-
posite substances. Hence, he must have thought that this terminology
provided the resources to explain, or at least give some insight into,
the unity of a composite substance at a time and its persistence (or
continuity) as a unity over time. The task of the philosophical critic is
to consider how far he was successful in these projects. But this
requires us first to understand how, in Aristotle's view, talk of matter/
form or potentiality/actuality2 helped to achieve these goals. The
©David Charles 1994

1 Aristotle does not always distinguish between unity over time and unity at a
time. 'Being one thing' may mean 'being one thing at a time' or 'being one over
time'. However, issues of persistence over time are discussed in parts of e. 7 and
8 (cf. 1049b7- IO, 1049a14-16). Since the substances in question are (e.g.) men
which develop, flourish, and decline, Aristotle should address the issues of unity
over time as well as unity at a time.
2 Aristotle seeks to explain the relation between genus: differentia in terms of
that between matter and form (1038•3-9, 1045"r4-23, cf. •29), and the latter in
terms of that between potentiality and actuality (1045•23-4, 30-3, br7-23). A full
account of Aristotle's deployment of the potentiality/actuality relation should show
how it explicates that between genus and differentia. This is a major task, far
beyond the scope of the present essay.
David Charles
present essay focuses on the prior task of the interpretation rather
than the evaluation of Aristotle's proposals.
Consideration of Aristotle's aims allows us to formulate two condi-
tions of adequacy for an interpretation of his views. It should seek to
show:
(A) how matter and form, or potentiality and actuality, serve (in
Aristotle's view) to account for (or perspicuously describe) the
unity of particular composite substances at a time;
(B) how matter and form, or potentiality and actuality, serve (in
Aristotle's view) to account for (or perspicuously describe) the
unity of particular composite substances over time, their
persistence and continuity as a unity.
I shall describe (A) as the Unity Requirement, (B) as the Persistence
Requirement. An interpreter might argue that one should aim to
satisfy only one of these requirements, or perhaps neither. But this
would require her to show that Aristotle himself did not wish to meet
one of (A) or (B).
II. TWO TYPES OF ANSWER
One attempt to satisfy the Unity and Persistence Requirements runs

as follows. 'Take the notions of the Unity of a particular composite
substance at a time and over time as basic. The basic concept then
just is that of one unified and persisting composite substance (e.g.
Socrates); the unity or persistence of such substances is not further
analysable.'
If Aristotle adopted this view, matter (potentiality) and form
(actuality) would best be seen as interconnected abstractions derived
from the basic notion of one unified particular substance. They are
what they are in virtue of their contribution to the being and per-
sistence of that unified composite substance. Thus, the matter of this
human being is defined as that which, together with the appropriate
form, yields this unified composite substance; and the form of this
human being is defined as that which, together with the appropriate
matter, yields this unified composite substance. This view offers an
answer to the question:
What is it for this matter to have this potentiality?
along the following lines:
Matter and form are one and the same: they are aspects of one
particular composite substance. Thus, the matter is what is poten-
1
tially this human being, while the form is what is actually this
.
Matter and Form 77
What it is for this matter to have this potentiality is for it to be the
matter of this particular human being.
On this account, the unity of the particular composite substance is
treated as basic, and is not further explained by the relation between
matter and form. More specifically,
(C) the unity of the particular composite substance is not itself
explained by invoking a relation between matter and form, for
that relation itself adverts to the unity of the particular
composite substance;
and
(D) the identity of the relevant matter and form are themselves
fixed by the identity of the one unified substance from which
they are abstracted (what makes this one form/one piece of
matter is its contribution to this one unified substance).
If one adds to (C) and (D) the further claim that the relevant matter
and form of a composite substance are essentially the matter and form
of that composite substance, one arrives at
(E) the matter and the form of a particular composite substance are
only correctly describable as the matter and the form of that
composite substance, and cannot exist save as the matter or the
form of that composite substance.
If (E) is accepted, there will be a radical discontinuity between the
matter of a particular unified substance and the components into
which that substance is dissolved; they cannot be the same individual
matter.
(E) leads naturally to the following view. The proximate matter of a
particular human being is its hands, heart, brain, etc. One cannot
correctly describe what these are without reference to the particular
human being which they constitute, namely as its hands, heart, etc.
Similarly with its form. Further, one cannot specify the non-proximate
material components which make up its blood or flesh without treating
these in turn as abstractions from the particular unified composite
substance, namely as its flesh, its blood, etc. For if one could (and
they were present in this human being), there would be a new problem
of accounting for the unity of its proximate and non-proximate matter.
On this view, all the material components of a particular human being
are tied necessarily 'all the way down' to that composite unified
human being. They are all essentially the matter of this composite
particular man.
Thus far, the discussion has focused on particular substances, like
-~r111·... r~tP'i;! nr thl~ m~n Rnt A rlf;:tntlP t-:::.llrf;: nnt nn lu nf n<Jirtir-11l•lr mon
David Charles
but of particular composites 'taken universally' (1035b27-31, 1037a5-

7).3 Thus, he speaks of man, and not just of this man, where the
former is the latter composite considered in abstraction from its par-
ticularity as this man. There are analogues of (C), (D), and (E) in this
case also. Thus, the unity of the composite taken universally might be
construed as basic, and its matter and form taken universally under-
stood as abstractions from this composite. What makes this type of
form or type of matter the one it is would be its contribution to the
composite man. If so, one will not be able to specify what the form or
matter (considered universally) essentially are except as the form or
matter of man. If these analogies can be sustained, not merely could
this flesh not exist save as the flesh of this man, but flesh itself (and
other matter essentially involved in man) could not exist save as the
matter of a man. One could not fully specify what the matter in the
composite taken universally was except as 'the matter of a man'.
In a suitably striking phrase, all such Aristotelian matter (proximate
and non-proximate alike) would be essentially 'pregnant with,
humanity'-and as such represents a conception which is deeply alien
to us. 4
Composite substances (whether particulars or considered univer-
sally) may be compared (on this view) with sentences when under-
stood (as for example by Frege) as the basic units in semantic theory.
Thus, matter and form would be, in a certain respect, like subject and
predicate as Frege defined them: a predicate is that which together
with a subject yields a sentence, while a subject is that which together
with a predicate yields a sentence. Frege thought that it was a mistake
to try to explain the unity of a sentence by means of an independently
specified type of interconnection between subject and predicate (for
the relevant interconnection is itself dependent on the unity of the
sentence) (C). 5 What it is to be a predicate is to contribute to unity
of a sentence in this way (D). Outside of the context of sentences,
there are no genuine predicates, only 'dead' ones (E). What it is to
be a subject or predicate is to play a given rolt~ in a sentence. Finally,
what makes a predicate one predicate or a subject one subject is the
role they play in one atomic sentence. And it would be a similar
3 In this reading 'as universal' is taken with 'a certain composite' in 1035b30 to
designate the composite 'as universal' which is composed from a universal form
and a type of matter. I have been fortified in this reading of 1035b~,,...31 by
discussion with Myles Burnyeat.
4 For this view, see Burnyeat (1992: 25-6) and Kosman (198T 388-91).
5 This analogy is, of course, only a partial one. Fregean subjects and predicates
are not, unlike Aristotle's matter and form, in some way identical with one
another. The latter pair <lf~ more closely related than the former.
Matter and Form 79
mistake according to this view to try to explain the unity of a compo-
site substance by means of an independently specified type of con-
nection between matter and form (C), independently conceived
(D, E).
Naturally, this view has no difficulty in accommodating the
unity or persistence of one composite substance. Indeed, it takes
these phenomena for granted and does not explain them in any
way. Thus, it is unclear that it can account for the role of matter or
form as Aristotelian starting-points (archai: Phys. 191a10-14); for
the direction of explanation it proposes runs in the opposite direc-
tion. In what follows I shall describe this as the non-explanatory
theory.
This view is also radically non-reductionist. The unity and essential
being of a composite substance are not explained by the unity and
essential being of its constituents. Rather, the latter's unity and essen-
tial being are explained in terms of their contribution to the rele-
vant unified substance. Thus, in the case of the universal composite,
Aristotle would be so far removed from the desert landscape of
twentieth-century materialist reductionism as to revel in the jungles
of the vitalist conception of matter, from which the philosophers
of the seventeenth century strove (fairly successfully) to disentangle
us.
The explanatory approach, by contrast, is one in which at least one
of the pair matter/form (or potentiality/actuality) is taken to be inde-
pendent of, and prior to, the notion of a composite unified substance.
Thus, for example, the type of matter or form (or both) of a human
being is to be specified independently of their being the matter or form
of the composite human being, and their principle of combination be
stated without reference to their being elements in a unified composite
substance. An explanatory account may take some identity claim for
granted (e.g. concerning the identity of a form), provided that this did
not presuppose the identity of a unified composite substance. Such an
account will need to show how matter and form thus specified are
related so as to constitute and maintain a type of unified compo-
site substance. Similarly, at the level of particular substances (like
Socrates), an explanatory approach will aim to show what makes this
one particular unified composite (e.g.) by reference to a given type of
form, or a particular form, being enmattered in particular quan-
tities of matter. Once again, the identity of (e.g.) particular quan-
tities of matter, or a particular form, may be presupposed provided
that the identity of this composite ;ubstance (namely Socrates) is not
assumed.
Elements of explanatory and non-explanatory views are to be found
80 David Charles
in a variety of influential contemporary interpretations of Aristotle's

account of composite substances. 6 While, in the final analysis, the two
approaches need heroes and heroines, we should proceed cautiously
before lining up parties to this dispute. The issues are not always
sufficiently clearly defined for worthwhile progress to be made solely
by challenging the views of a named authority. For they too may lack
a clear perspective on what is at issue. Nor is the controversy itself a
single one. Adoption of an explanatory approach at the level of types
of substance (e.g. man, 'taken universally') is consistent with taking
a non-explanatory standpoint view of the matter or the form of par-
ticular substances such as Socrates. Conversely, adopting an explana-
tory viewpoint concerning particular composite substances is consistent
with thinking that at the level of types of substance there can be no
definition of what the form of a human is without reference to the
composite human being it enforms.
In the next sections I shall consider some of the relevant texts in
Met. Z. 17, H. 2-3, 6, and e. 6-9, and suggest that they provide
some evidence that Aristotle was proposing an explanatory account of
the unity of both particular and types of composite substances. I shall
also claim that, in these chapters, Aristotle is formulating the issues
concerning unity in ways which are influenced by his discussion of
definition, unity, and explanation in the Posterior Analytics, and that
this is in itself philosophically significant. But both theses are presented
in this essay as conjectures for further study rather than as beliefs to
be secured here by conclusive argument.
6 Elements of the non-explanatory view can be detected in claims such as:
'potentiality and actuality are the same thing present together in that full activity
which is nothing other than the manifestation of the one entity that both are,'
Kosman (1984: 144). 'Aristotle compares asking for an explanation of why poten-
tiality and actuality are one with asking for an explanation of why anything is one.
No answer is needed ... because the explanation of the thing is at the same time
an explanation of its being one,' Burnyeat et al. (1979: 44-5). However, neither
Kosman nor the London Group opt wholeheartedly for the non-explanatory view.
Other passages in these works suggest a somewhat different account, even on
occasion some version of the explanatory one (see below, n. 38). Wilfred Sellars
(196T II8) appears to adopt the non-explanatory viewpoint without qualification.
(This and other examples are cited by Frank Lewis in discussing the 'pmj~ctivist
approach' to metaphysics, this volume: 255-6). Elements of the non-explanatory
viewpoint can be detected in Theodore Scaltsas, 'Substantial Holism' (this volume:
127), when he suggests that (in Aristotle's solution) 'the components of a sub-
stance are identity-dependent on what the whole is'. I classify as 'non-explanatory'
any account which defines matter and form in terms of the composite substance
they constitute. Such accounts will remain 'non-explanatory' if one in addition
defines composite substances in terms of matter and form thus understood.
Matter and Form 81
III. THE EXPLANATORY MODEL OF THE

ANALYTICS INTRODUCED AND REFINED:
Z. 17, H. 2-3
In the Posterior Analytics, Aristotle requires that the search for essen-
tial properties be conducted in the following (canonical) way. 7 At
some stage the inquirer knows that (in the case of thunder)
noise/thunder belongs to the clouds
and so knows that thunder exists. 8 At this stage she has a preliminary
(or incomplete) account of thunder which may pick out a necessary
but not yet an essential feature of thunder. 9 To find the latter she
needs to find a cause which makes it the case that
noise/thunder belongs to the clouds.
The cause, in question, must be both necessary and sufficient for
thunder to occur (i.e. for a given type of noise to occur in the clouds).
For if it is not so, she has not found the cause of its occurrence qua
°
thunder but qua thunder of a given type. 1 Further, the relevant cause
must bring it about that
noise/thunder belongs to the clouds
by a recognizable form of necessitating connection (e.g. efficient
causation, teleological necessitation, material groundedness) which
will be asymmetric.11 Thus, it is because fire is quenched that
noise/thunder belongs to the clouds,
7 This section rests on a longer study of the Posterior Analytics which I have
undertaken. Some elements of that study are to be found in my paper ( 1990:
145-54).
8 An. Post. 93"16-20; cf. "29-36.
9 'The essence' (as I employ the term) will be the correct answer to the 'Why?'
question. Before discovering that, one may know that noise occurs in the clouds,
but not why it does (cf. An. Post. 94a4-6). My present formulation is intended to
leave o~en the possibility of one's possessing definitions of what 'thunder' means
without knowing that thunder exists. On this latter issue, see Robert Bolton's
discussion (1976: 514-44).
HI For this mode of argument see An. Post. 98b35-8, 99a30- 07.
11 These are the options available in An. Post. II. II ad init. 'Material
groundedness' appears to apply to all cases where 'given certain things, something
else is necessary' (94a21-2), and is not confined to the material cause. In the
Analytics, Aristotle typically concentrates on the efficient cause (93a30-6, hIO- 14,
etc.). It is not clear in this context whether formal causes can necessitate their
effect via an asymmetric connection without resting on a further pattern of
efficient or final causation or material groundedness.
David Charles
where the efficient cause explains the relevant asymmetry. Further,

the basic cause, constitutive of thunder, is one about which no further
question can be asked of the form:
why does noise/thunder belong to quenching of fire?
At this point she has reached explanatory bedrock. This is an im-
mediate proposition, and no further middle term can be introduced to
explain the connection between thunder and quenching of fire. 12 The
relevant noise is caused (efficiently) by fire being quenched-without
any further intermediate (efficient) causal connection.
The basic cause should explain all the other necessary features of
thunder. This serves to underwrite the unity of thunder: all its nec-
essary properties are explained by one efficient cause. 13 If this were
not so, thunder would not be one unified kind. The unity of thunder
depends on the unity of its cause. At the basis of the Analytics
account there is for every genuine kind some simple feature which is
the cause of its other necessary properties. Thus, the unity of thunder
as a kind depends on the unity of this basic feature. 14
In the Analytics, Aristotle suggests that he will apply this model to
substances as well as substances and their properties (An. Post. II.
1-2), but fails to do so in a systematic way. 15 My exegetical sugges-
tion is this: in Metaphysics z. 17, H, and parts of e, Aristotle is
attempting to do precisely what he had failed to do in the Analytics,
and indicates that he will do in Met. Z. i7, 1041"23-30: apply
his explanatory model to the case of composite substances. In this
passage, he compares his Analytics question about thunder,
why does noise belong to the clouds?
12 For this strategy, see An. Post. 93 3 30-6, bro-14. Immediate propositions
differ from immediate terms (like 'monad') which are the starting-points for whole
sciences (93b21-4). In the immediate proposition, 'noise belongs to quenching of
fire', neither of the terms need be immediates.
13 This is why the method of scientific proof (apoqeixis) resolves one of the
problems which Aristotle has pinpointed in the methods of division and the
'formal syllogism' in An. Post. II. 4-6 (cf. 92 3 29-33): their inability to ej~ablish'
the unity of the definiendum. (On this, see Charles 199w: 215-30.)
14 If there is no such unifying cause, one is dealing with two types of pheno-
menon and not one. Aristotle illustrates this in his discussion of two possible types
of &ride (An. Post. 97b13-25).
· One exception is An. Post. II. 9, where Aristotle seems to distinguish
between substances which have an internal essence (middle term), and those
which are simples (and lack structure): 93b25-7. But he does not develop this
theory in the Analytics any further, nor deploy the matter/form relation required
to apply the explanatory account of An. Post. IL 8. to composite substances. I am
indebted to Kei Chiba fote(~iscussion of these issues.
Matter and Form
with
why are these (e.g. bricks and stones) a house? (1041a27, b6-7),
and notes that one should search in both cases for the cause (in one
the efficient, in the other the teleological). My proposal is that we
should take this comparison as a guiding thread to lead us through the
maze of the remainder of z. 17, H, and parts of e.
Aristotle's answer to his own question about houses in Z. 17 is fairly
uninformative: 16 he writes that these bricks and stones are a house
'because what it is to be a house belongs to them' (1041b6-7), but
does not expand on how this is to be understood. Equally, his com-
ments on the general form of the question are less than fully satisfying.
While he formulates the question as
why is the matter some definite thing? (1041b7-8),
he does not indicate how house/some definite thing is to be under-
stood. But (at least) it seems clear that house and what it is to be a
house must be distinct, and this suggests that while the answer refers
to the form alone (1041b8), 'house' in the question refers to something
else: for example, the composite. However, in Z. 17 Aristotle is
proceeding at a highly abstract level, and these details are not spelled
out.
In H. 2 Aristotle introduces differentiae to distinguish matter in
terms of composition, position, time, etc ... (1042b11-1043a2). In the
case of houses, the relevant differentia is 'arranged thus' ( 1043a7-8),
although the final cause might be added ( 1043a8-9), which in this case
is to be a covering for possessions and bodies (1043a16-18, 32-3).
From these materials we can formulate the basis for a full definition of
house:
house = planks and stones arranged thus for sake of covering
possessions and bodies ( 1043a31-3).
Some of the differentiae are causally more basic, the cause of each
thing being what it is (1043 3 1-3). In this case, the final cause explains
why the stones and planks are arranged in a given way, and why being
a possessions-coverer constitutes the form (1043"33). So, one can
understand the Z. 17 question more precisely as
why are these stones and planks a house, (or better) arranged
thus? 17
which is to be answered by
16I discuss these passages further elsewhere (Charles 199w: 215 ff.).
17I assume that house and arranged thus are substitutable in this context in the
same way as thunder and noise are in An. Post. II. 8. (cf. 93b7-14).
David Charles
because they are so arranged in order to protect possessions i.e.

to be a house.
In this case, the basic explanatory feature is the teleological cause,
and this is what makes it the case that this collection of planks is a
house, i.e. arranged in a certain way. The cause is necessary and
sufficient for this being so, and this connection inherits the asymmetry
of the teleological cause. Further, the presence of this goal (being-a-
possessions-protector) is what explains the other necessary features of
a house: for example, why it needs a roof (in most climates), why it is
made of wood rather than feathers, etc. As the essence this plays the
unifying role of explaining the presence of the other necessary features
of the house. The relevant immediate proposition
being arranged thus belongs to all belongings-protectors
expresses the immediate teleological claim that this arrangement is as
it is in order to protect belongings. No further teleological claim is
required to explain the connection between being arranged thus and
being a protector of belongings and bodies, and no further middle
term is available.
What is the analogous question in the case of human beings? It
appears in Z. 17 as
why is this body with this feature a man? (1041b6-7),
which might be formulated in the Analytics style as
why does man belong to a body with this feature?
In Z. 17 Aristotle does not comment further on how man is to be
understood in the question, or how the answer is to be represented.
Presumably the latter should refer to what it is to be a mai!J, as in the
parellel case of what it is to be a house ( 1041b6), but there is no
indication in Z. 17 of how this phrase is to be understood. In H. 2
he remains reticent about which differentia of matter marks out man
(or foot: 1042b31), and does not spell out which feature of man's
shape (morphe) or activity (energeia) is to play the role of arranged
thus in the house syllogism. The most promising candidate is being
two-footed, if the shape is to be specified (rn43b11, rn45a20). But
Aristotle is not explicit about this, and is once again proceeding at a
very abstract level (1044b16ff.). He is clearer on the specification of
what it is to be a man. This is unequivocally identified with soul of a
given kind (1043b2-4), which provides answers to the Z. i7 question
why is this body with this feature a man?
and its H. 2 rephraseci. version
Matter and Form 85
The answer to the first question would be that (the composite) man
belongs to the body because the composite is as it is because it has a
soul of a given kind, which in turn belongs to this body. If the
question is rephrased, the answer would be that being a biped belongs
to this body because being a biped is required if we are to have the
relevant type of soul, which belongs to this body. Being a biped is
necessary for the proper functioning and survival of the human soul.
On this construal, the distinctive type of soul serves as the teleological
cause (marked out by the middle term) which explains why the body
has given properties.
In the latter case a syllogism might be constructed along the follow-
ing lines:
Being a biped belongs to soul of a given kind
Soul of given kind belongs to body of type S
Being a biped belongs to body of type S.

The first premiss expresses the teleological claim: 'If there is a soul of
this type there must be bipedality.' There is no further explanation of
why the human soul requires this feature, and we have reached
explanatory bedrock. But what of the second premiss? In H. 2
Aristotle notes that in the case of house, the corresponding proposition
expresses a form-matter predication ('being a possessions-coverer be-
longs to these bricks'), where bricks are the matter with the relevant
potentiality (1043a15-16). Similarly, in the case of the man-syllogism,
the minor premiss would express a form-matter connection. 18 But it
is only in H. 6 that Aristotle seeks to specify the nature of these
connections more precisely (see next section).
On this conception, the unity of a composite substance depends on
there being one basic feature (specified as the middle term of. the
relevant syllogism: e.g. soul of a given kind) which belongs im-
mediately to the relevant matter (minor premiss). This feature should
have the following properties:
[a] it should be linked immediately to the relevant matter in such a
way as to explain why the matter is as it is;
[b] it should explain why the matter is possessed of other non-1-asic,
but necessary, properties (e.g. being a biped ... );
[c] the basic feature will constitute the essence of the composite
substance;
18 If it were not immediate, the process of investigation would not be complete.
In the Analytics Aristotle says little about the connection in the minor premiss
between (e.g.) fire being quenched and the clouds, but in Metaphysics H. 2 he
suggests that cases like these (e.g. 'calmness belongs to the sea') might be regarded
'lie ".ln"'ll"n-ru'C" tr.. l!r..,..,.,,../.,.,,...,,++,.... ... __ ,...,..1.;,... ...... ;,...._ ... /,....£ ..... ,.... • . - a - - L.\
86 David Charles
[d] its unity (as one feature) should be basic or self-evident. 19

In the Metaphysics the relevant structure of explanation is teleolo-
gical, unlike in the Analytics, where efficient causation predominates.
Further, it introduces new notions to carry through the project: matter
and form, potentiality and actuality. The form is the teleologically
basic feature, and this explains why the matter is as it is (in the com-
posite substance) and why it possesses non-basic, but teleologically
required, properties. 20 The relevant feature is required in both the
premisses of the man-syllogism, and both are immediate propositions.
This account is an explanatory one in two respects:
(i) the unity of the composite substance is explained by the rela-
tion between matter and form;
(ii) the identity of the form as the basic teleological cause is de-
termined without reference to the identity of the composite
substance.
The matter is present and is as it is for the sake of this form. The form
is as it is because it is the teleological basic feature wl)ich explains the
presence and nature of the matter. The identity of the form is partially
fixed by its teleological role independently of the identity of the
composite substance which it enforms. In this account, Aristotle's
reformulation in the early chapters of H of the explanatory questions
raised in Z. 17 is taken as evidence that he is following the explanatory
approach to the problem of unity. If he had been an advocate of a
non-explanatory approach, these chapters would have been an un-
necessary excursion. He should rather have argued directly that the
questions raised in Z. 17 were misleading or mistaken. However, his
19 The basic feature (or form) appears to be simple (cf. Met. 104.5"36ff.,
1041"9-II), as it is form without matter. This will be the primary object of
definition, as it gives the answer to the Analytics 'Why?' question by specifying
the essence of the composite (cf. 1041b6-9). However, just as in the Analytics
there are other definitions which contain more than the answer to the 'Why?'
question (e.g. 'thunder is noise in the clouds caused in a given way'), so in the
Metaphysics the universal composite may be defined with additional reference to
the matter as (e.g.) 'biped animal with soul of a given kind'. Attention to the
variety of types of definitions available in the Analytics makes it no surprise th;t1t
Aristotle can sometimes restrict definition to the form of composite substances
(e.g. Met. 1035°14-15), while elsewhere including their matter (Met. I036"28-30).
On this problem, see the papers of Michael Frede (1990: 113-29) and Michael
Ferejohn (this volume: 3 I0-17). The Analytics structures show how form of a
composite substance can be logically prior to matter, even though a full definition
of the composite turns out to contain both.
20 In these passages form is explicated in terms of actuality, and the latter
directly connected to goals: Met. 1050"21-3. Forms are, on this view, real entities
which possess intrinsically certain goals.
Matter and Form
attempts to articulate them in a way which makes them more readily
answerabl6tsuggest that he is developing a version of the explanatory
account.
IV. TWO APPROACHES TO H. 6
The explanatory account faces difficulties in the interpretation of H. 6

and parts of e. At very least, it leads to a radically different construal
of these chapters than the one favoured by the non-explanatory
account.
In H. 6 Aristotle takes up directly some of the issues addressed in
Z. 17, such as
what makes a composite one thing and not a heap?
and ends the chapter, it appears, by claiming:
it is similar to seek for the cause of one thing being something
and it being one thing (1045b20).
But the content of the chapter itself is elusive. Thus in 10453 23 ff.
Aristotle writes:
if-as we say-one [e.g. animal] is matter and the other shape [e.g. being
footed), and the one is potentiality, the other actuality, what is sought would
no longer seem to be an aporia (a problem). For the problem would be the
same even if the definition of cloak were rounded bronze. For this name
would be a sign for this definition. In this case what is sought is the cause of
the bronze and the rounded being one. But the problem has gone away
because one is the matter, the other is the shape. What is the cause of what is
potentially F being actually F in the case of things that come to be-apart
from what makes it so, the efficient cause? For nothing other is the cause of
what is potentially a sphere being actually a sphere, but this [namely the
cause] is what it is to be for each of them singly. 21
On the non-explanatory interpretation, he is proposing that if one
represents 'two-footed animal' in the form
matter/shape
potentiality I actuality
no further explanation of their unity is required. No genuine problem
remains: one has reached explanatory bedrock. Nothing more can be
21 On this reading, 'this' in 1045 3 33 refers to the cause mentioned in •32. On the
dissolutionist view 'this' would refer not to the cause but to the fact of what is
potentially a sphere being actually a sphere (namely to there being a unified
substance), which would give the essence of both actuality and potentiality as
abstractions defined by their contribution to a unified substance.
88 David Charles
said to explain the unity of a composite substance, precisely be-

cause these notions are abstractions from that of a unified composite
substance.
According to the explanatory interpretation, by contrast, Aristotle
is suggesting that if one views the problem of unity in this way, it is no
longer insuperable (an aporia), because one can see how to resolve it.
The question is now in an answerable form, so that one can proceed
to discover what is sought: what makes it the case that potentialityl-
actuality are paired in such a way as form a composite unity. To this
the answer is: because they have the same essence. The essence in
question is the actuality or shape, and this fixes both its own nature
(what is to be that actuality) and that of the matter which exists for
the sake of that actuality. The matter, or potentiality, is what it is
because it is the matter for this form or actuality.
The final lines of the chapter (1045h17-24) are also subject to these
conflicting lines of interpretation. They run as follows:
but-as has been said-the final matter and the shape are the same and one,
the one potentially the other actually [energeiai: dative], so that it is similar to
investigate
what is the cause of one thing being F
and
what is the cause of being one.
For each individual thing is something, and what is potentially and what is
actually are one in some sense. Therefore, there is no other cause [of it being
one/being F] except that which brings it about that there is something which
moves it from potentiality to actuality.
On the one hand, it seems that the explanatory question raised ,in
1045h18-20 is an important and sensible one: what is the cause of
unity? However, the final sentence clearly says that there is no other
cause of the unity of what is potentially F and what is actually F apart
from the efficient cause. And this might reasonably be taken to mean
that if one removes the efficient cause, there is no cause of unity .at
all; the unity of the composite substance is taken as basic (as in the
non-explanatory interpretation).
However, if the final sentence is interpreted in the latter way, it
contradicts the earlier passage ( 1045a30 ff.) in which Aristotle appears
to make the far more cautious claim that there is no cause of what is
potentially F being actually F except this: namely the one essence of
both. For this requires that there is a cause of the relevant unity of the
composite substance-the presence of a form, whose nature is fixed
by the relevant goal, which determines the nature of the matter. The
final sentences of H. 6 would be consistent with this earlier passage
(thus understood) if they were taken to mean that there is no cause of
Matter and Form
unity of substance distinct from the relevant actuality and potentiality,
because tNese encapsulate all the causal features required to explain
the unity of the composite substance. And this would be to understand
them in line with the explanatory approach.
This is a possible route to resolve the exegetical dilemma in H. 6.
But is it a plausible one? The argument in ro45b21-3 is very com-
pressed, but appears to run as follows.
P 1. Each thing is one thing (hen) of a given kind (ti).
P 2. The matter is the thing in potentiality, the form is the thing in
actuality.
I, 2 3. It is similar to seek the cause of one thing being something
and it being one thing. ·
4. 3 follows from I and 2 since one thing is something because
matter and form are joined as potentiality and actuality. But
when matter and form are joined in this way there is one
thing present.
But how to understand P. 2? In what way is the form the thing in
actuality? On one view, this would mean that the form is the com-
posite substance seen as an actuality, and the matter is the composite
substance seen as potentiality. On this understanding the notion of
one composite substance would be assumed as basic (as in the non-
explanatory approach). By contrast, according to the explanatory
interpretation, P. 2 would claim that the form is the thing in actuality
because the thing (e.g. man) is identified (in strict use of H. 2,
ro43a27-9) with the form (as the man in actuality). On this view, the
actuality is taken as basic and the matter is understood as potentially
that actuality, where the latter is identified with the thing itself (i.e.
the man) taken strictly. 22 But if the matter is potentially the actuality
itself, the natural question is: what is it for the matter to be potentially
this actuality? And that topic would be raised in H. 6, but not re-
solved until the following book.
Several considerations appear to favour the explanatory interpre-
tation of H. 6, although none is decisive. First (as already mentioned),
the questions raised in 1045b18-20 appear to be genuine ones. Indeed,
'what is the cause of the unity of the composite?' is the very question
which Aristotle had set himself in Z. 17 and H. 2-3. Second, there is
nothing to suggest that the theorists, whose views are discussed in
22 The main difficulty at this point is to determine whether the thing (hen ti) is
being identified with the form or with the compound (1043•27-9). The passage
under discussion ( 1045b20- 1) appears to allow both interpretations. I am indebted
to Michael Frede for advice on the interpretation of this sentence. See also
Theodore Scaltsas's discussion of this sentence (1985: 232-3).
90 David Charles
1045h7- 17, are asking a nonsensical or pseudo-question. Rather

they seem to be raising a genuine difficulty (namely that mooted in
1045a30-2) in a way which makes it aporetic, because (so formulated)
it is difficult to resolve. Once the question is correctly framed one can
find a feature which explains the relevant unity: the common essence
of the matter and form, which reveals the former as potentially F and
the latter as actually F. Indeed, from this vantage-point, the form i~
the teleologically basic feature which explains why the matter is as it
is. 23 Third, if one sees Aristotle as pursuing the explanatory project, it
is easier to understand why he proceeds to examine the concepts of
potentiality and actuality in Met. e. If he had favoured the non-
explanatory approach he should have said this.
There is no difficulty in seeing how
potentiality and actuality, or
matter and form
are unified to form a composite substance. For potentiality and
actuality are themselves abstracted from that unified substance.
Indeed, actuality is just that which together with potentiality
yields a substance of this type.
This would have been the end of the inquiry. If this had been his
approach, he should have fallen silent earlier.
If 'these considerations are accepted, Aristotle in H. 6 is employing
the intrinsic goal of the actuality to explain the unity of a composite
substance. This goal explains why the matter is as it is-because it is
organized for the sake of the goal. If matter and form are to be one,
they must both be represented in the basic proposition of a teleolo-
gical explanation which gives the essence of both, because both are
essentially directed to the goal implicit in the actuality. Indeed, to talk
of actuality and potentiality is precisely to introduce the ontuiogy
required to make these claims true. 24
H. 6 adds two further features to the explanatory picture. First, it
23 It is an interesting further issue whether the form should be regarded as a
property of a given type. This is discussed by Charlotte Witt ( 1989: chs. 4 and 5),
but lies outside the scope of the present essay.
24 This is not to say that matter (potentiality) and form (actuality) are them-
selves to be exhaustively defined in terms of their role in teleological explanation.

It requires only that they are the entities which make this form of teleological
explanation true. It should be noted that Aristotle sometimes identifies the goal
with the actuality (Met. 105oa2r-3), but elsewhere regards the actuality as that in
which the goal is present (Met. 1048b22-4). Sometimes the actuality is the goal
achieved, sometimes the achieving of the goal. 'Actuality' appears to be act/result
ambiguous (see Charles 1986: 135-9). In the present essay I have sometimes
preserved this ambiguity; .
Matter and Form 91
emphasi:z:es that the unity of the essence in the case of man (namely
the soutfis immediate (1043b2-3). It appears to be one of the simples
(ta hapla) specified in Z. 17 (1041b9-11) whose unity is a basic
feature, which cannot be established by investigation of the Analytics
model. But this is puzzling. Why cannot one question what makes the
human soul a unity, why it includes essentially certain features and
not others? Nor is it clear how a soul which consists in diverse
activities (e.g. walking, thinking, desiring, etc.) can be a simple unity
of this type. But these issues are pressing for Aristotle on any account,
and are not special problems for the explanatory approacb. 25
Second, H. 6 makes some progress in clarifying the type of entities
with which Aristotle is concerned. In H. 2 and 3 his talk has been at
the general level of man (1043b2), house (1043a31), animal (ro43a34),
and ice ( ro43a7 ff.). Indeed this is what one would expect if he is
guided by the Analytics conditions on explanation and definition,
which essentially involve universals. But in Z. 17 Aristotle bad written
of 'this body' (104rb6-7) and 'this matter' (1041b5), which suggests
that he is speaking of particular composite substances and not of such
substances 'taken universally'. In H. 6 he writes of man ( ro45a20),
which could be taken either in a particular or general way, but
adds that his account applies both to beings with perceptible matter
( 1045 3 36) which come into being ( ro45a3 I), and to those with intelli-
gible matter also ( ro45a34-5). Since the former are particular com-
posite substances (for only these come into being), the discussion
must at least address the question
why is this body two-footed/a man?
and give an answer in terms of the form of this particular composite.
But if intelligible matter is understood as the generic element in a
species, it should refer to animal (if 'man' is understood as 'biped
animal'). 26 And this can be seen to be like matter (cf. ro38a6) because
it is this particular matter (e.g. the matter of Socrates) seen in abstrac-
tion (universally) as the matter of an animal to which the predicate
'biped' applies. If this is correct, Aristotle is attempting to apply a
similar strategy in the case of both particular substances and kinds of
substances (namely particulars taken universally).
These remarks leave unaddressed an issue which has attracted con-
siderable scholarly attention in recent years: did Aristotle accept the
existence of both particular and general forms, or only one of these?
Several authors have argued that Aristotle believed in particular forms
25 I discuss this issue further elsewhere (1992a: 215ff.).

26 See Ross (1924: ii. 238).
92 David Charles
in addition to (or in place of) the general forms (or universals)

apparently required for proper knowledge and Analytics-style defini-
tion and explanation. 27 In this essay, I shall attempt to remain neutral
on this issue. My main line of argument is intended to apply indif-
ferently to particular and general forms alike.
However, acceptance of an explanatory interpretation has impli-
cations for this controversy. Since from this viewpoint form is seen as
prior to composite substance (whether individual or taken universally),
its proponent cannot accept that forms (whether individual or general)
are to be defined as the forms of composites (or the-matter-in-a-
composite); for this would be to treat the composite as explanatorily
prior to the form. Further, this interpretation appears (at least) prima
facie inconsistent with the view that particular (or general) forms
involve given matter in their essence (i.e. the answer to the funda-
mental 'why?' question); for if they did, the essence could not explain
the matter's possession of its (material) properties. In so far as this
essay provides support for the explanatory viewpoint, it limits the
types of forms, whether particular or general, that Aristotle can con-
sistently accept.
Acceptance of the explanatory viewpoint serves to locate several
controversial issues in a distinctive perspective. For its proponent, the
case for Aristotelian general forms (universals) is secure since they are
the object of scientific knowledge and the subject of lawlike connec-
tions. 28 Further, there is at least a strong prima-facie explanatory case
for accepting the existence of Aristotelian particular forms. 29 Indeed,
within this approach, the most pressing issue concerns not the existence
of particular forms, but rather (i) what 'particularizes' them and (ii)
whether, and in what ways, such forms are prior to general forms (ii:t
definition, explanation, being, etc.). 30 On one view, particular forms
are differentiated by the matter they enform (cf. Met. Z. 8, 1034a5-8).
If so, general forms would be explanatorily basic, and particular forms
the result of such universals being instantiated in given quantities of
27 e.g. the.recent books by Irwin (1988: 250-9), Frede and Patzig (1988), and
Witt (1989). Others who have taken this line include Albritton (1957= 699-708),
Sellars (1957: 688-99), Heinaman (1979: l-20), Lloyd (1981)', Frede (1985:
17-26), and Whiting (1986: 359-77).
28 The existence of general forms is explicitly defended in the Analytics (An.
Post. I. 24, 85b15-18). His formulation of explanatory claims invoke universals
(An. Post. II. 17, 99"21-4, 30-7), and apparently do not include singular terms
referring to particulars (see Patzig 1969: 4-8).
29 ln GA V, particular forms appear to be required to explain individual
genetic difference, as Cooper argues (1990 repr.: 79-84). Presumably, perceptual
forms are particulars (De An. II. 5, 417b22-5).
30 The latter issue is raised by Wedin (1991: 370-1) and Irwin (1988: 268-9).
Matter and Form 93
matter. A r:!tlical alternative (but one still within the explanatory
viewpoint) is to take as basic particular forms differentiated indepen-
dently of any connection with matter, and to regard general forms as
in some way derived from them. 31 Either alternative could accept
that (i) there are particular forms, (ii) Analytics-style definition and
explanation involve general forms, and (iii) knowledge requires the
existence of universals. For (i)-(iii) are neutral on the crucial issue of
whether particular or general forms are prior (and in what sense).
While the present essay maintains neutrality on this issue, its resolu-
tion remains of major importance within an explanatory approach.
V. TWO APPROACHES TO e. 6-8
These differing interpretations of H. 6 lead to differing views of e.

6-8. According to the non-explanatory interpreter of these chapters,
Aristotle is not seeking to elucidate the notion of the unity of a
composite substance. The notions of potentiality and actuality are to
be understood as the potentialities and actualities of one unified
substance. There is no attempt further to explain what these relevant
potentialities and actualities are. The relevant ones are just those
which stand as matter to substance (ousia) in one composite substance.
This dissolutionist strategy might seem to motivate Aristotle's remark
in ro48b7-ro, on this view (arguably) the most disappointing com-
ment in the whole of the Metaphysics:
All things are not said to exist actually in the same way, but only by analogy ...
for the relation is either that of process to potentiality or substance (ousia) to
some specific matter.
In H. 6 it seemed that Aristotle intended to employ talk of actuality
and potentiality to clarify the relation between matter and form in a
compound substance. But now it looks as if he is employing the latter
to illuminate the former. The relevant potentiality in a unified sub-
stance is one which stands to its actuality as matter stands to its
composite (e.g. the statue of Hermes: ro48a30-3, h2-5). The notion
of a unified composite substance is taken as basic. Matter is conceived
as the potential to be (or become) a unified composite. Form (if
present at all) must also be an abstraction derived from the notion of
the unified composite. No account of either matter or form is offered
31 This formulation is intended to cover those who regard Aristotelian general
forms as abstractions (Frede and Patzig), and those who view them in a more
realistic spirit (Irwin 1988: 262-3). This issue is discussed by Whiting ( 1991:
638 f. ).
94 David Charles
except in terms of their role in a unified substance. Aristotle's philo-

sophical ambitions are satisfied (it appears) by pointing to a set of
interdefined terms based on the prior notion of one unified composite
substance.
This proposal would be rejected by an explanatory interpreter of
these chapters, because it unduly restricts Aristotle's project. If this
were his goal, why did he struggle throughout e to clarify the relevant
type of actuality and potentiality? If he had followed this unambitious
strategy, he should have argued that any attempt to go beyond this
limited target was in principle mistaken. However, in e. 6~8 he
appears to suggest that the required potentialities and actualities are
of a distinctive type (namely matter of a certain sort and substance:
ousia), and to proceed to explain in what ways these are similar to,
and dissimilar from, the types of potentiality and actuality involved in
the case of processes. If he can do this without taking as basic the
notion of a unified composite substance, he is (in effect) developing
further his explanatory project in these chapters.
What analogies and disanalogies hold in Aristotle's view between
the potentialities and actualities involved in processes and substances?
In the case of processes, the potentiality is defined as the,potentiality
for this type of process: the potentiality to be made into a house. 32
This type of process is the one it is because of its goal. If processes
differ in goals, they will be numerically distinct even if they always co-
occur. 33 Since the process is defined by its goal (end-point), the
potentiality is essentially the potentiality to reach that end-point.
Thus, the end-point or goal defines the nature both of the process ~~~
move towards that end-point) and of the relevant potentiality. In a
favourite Aristotelian example, house-building is defined as the pro-
cess of house-construction by its goal (namely the house being made),
32 Phys. 201"16-18. I have defended this account elsewhere (1984: 19-27). For
a contrasting account, see Mary Louise Gill's discussion (1989: 186-94). My view
rests on construing the relevant capacities for change as essentially dynamic
capacities of substances. Others, by contrast, understand Aristotle to attempt
to account:for processes without taking such dynamic capacities as basic. Philoso-
phical discussion of this issue requires study of the adequacy of definitions
of change which do not rely on the presence of dynamic capacities (e.g. can
such accounts separate changes from a succession of distrinct states, or from
'Cambridge', relational, alterations which are not genuine changes?). These issues
raise exegetical problems concerning Aristotle's answers to Zeno's paradoxes of
change. In my view, Aristotle employed dynamic properties of substances to
answer Zeno's challenge, and this is why he accepted capacities for changes as
basic. But a full defence of this suggestion is a major, and separate, undertaking.
33 An example of this is, I believe, teaching and learning (Phys. 202b5-22),
which differ in goals (PJ,iys. 202•21-4). See my discussion of this (1984: II ff.).
Matter and Form 95
and the relevant potentiality of the builder is defined as the potential
to achieve this goal (namely to build a house).
If the case of matter and substance is analogous, the goal should
reveal the latter's nature (or essence), and the relevant potentiality
should be the potentiality to achieve that goal. Both actuality and
potentiality will be defined directly or indirectly in terms of the goal of
the actuality. The basis of the analogy is just this: in both cases the
relevant potentialities and actualities are what they are because they
are essentially goal-directed. Indeed, potentialities and actualities
would be introduced as precisely the entities required to make true
goal-directed accounts of this type.
However, while goal-directedness may supply the basic ground Of
the analogy between these two cases, there remain major points of
disanalogy, which emerge in subsequent discussion in 8. 6-8.
(r) The first pair of relata in the analogy (process, substance (ousia))
differ significantly (as Aristotle emphasizes in the remainder of 8. 6).
A process has as its goal its end-point. In the case of house-building,
the goal is the house being built. By contrast, in the case of substances
(as in that of activities) the goal is immanent (1048b21-2). The goal of
being a house is realized when there is a house. Once the goal (e.g.
being a belongings-protector) has been achieved a house can persist
for a long period, while in the case of processes, once the goal has
been realized there is no further process at all. So the goals are
of a different type in the two cases (as Aristotle hastens to explain:
1048a18-36). This is why processes essentially unfold or unwind
through time, while substances with their essences endure. In this
respect (in Aristotle's view) the latter are like activities and unlike
processes.
(2) The second pair of relata (capacity for change, matter) also
differ. In the case of persisting substances, the relevant capacity of the
matter is to be a house and not to become a house, to be a man and
not to become a man. In many cases, the capacity to become an F will
be lost when one is an F (e.g. a human); for once one is a fully-
ftedged human, one can no longer become one. In the process of
becoming an F one is not yet exercising one's capacity to be an F. For
if one were, one would no longer be in the process of becoming an F.
This is precisely how the capacities essentially involved in being and
becoming differ. 34 So if matter is defined as a type of potentiality, it
34 See Phys. 201b10-12, 28-9, 202"I. In these texts, the capacity to be a house
appears to be the capacity to be, in certain conditions, the composite, matter
arranged in a given way to protect belongings. I have defended this interpretation
elsewhere (1984: 19-21). For a contrasting view, see Michael Frede's discussion
96 David Charles
will be a different type of potentiality from the potentiality to become

an F (which belongs as such to a subject of change). Matter will rather
be essentially the potentiality to be an F.
(3) Under what conditions is matter possessed of this potentiality to
be, for example, a house? Or when is it (better put) matter of the
relevant type? (This issue, which is discussed in e. 7, was raised at the
end of H. 6.) In the case of becoming, something is potentially a house
(e.g. bricks/plinths) if it can be turned into a house (the composite)
without further addition or change in itself ( ro49aro- I I). These bricks
and plinths are the same before the building has begun, when it is
under way, and when the house is built. Qua what is potentially a
house they become (after a process which involves no alteration but
only rearrangement) the actual matter of a house. Thus, they possess
(at this stage pre-building) the potential both to become and to be a
house without further alteration (of the relevant type) in themselves. 35
But they are the matter of the house in virtue of their possession of
the latter potentiality, which will be present from the time when they
are capable of being made into a house until the house is finally
destroyed and the bricks broken (e.g. after many years of domestic
(this volume: 173 ff.). He suggests· that something retains the ability to be an F
precisely as long as it satisfies the conditions which made it initially appropriate
material to be turned into an F. This may well be true in cases like houses (which
Frede discusses), where the matter (i.e. bricks) remains relatively unchanged
throughout the process of building and after the house is built. But it does not
seem to apply to cases where the matter alters through a developmental process,
and is different at the end from the beginning (e.g. the matter of a man); for the
matter of the developed man has lost the ability to become a man. This is one
reason for resisting Frede's claim that the ability to be an F is the same as the
ability to become an F, 'seen in a different way' (this volume). This claim is
independent of Frede's central contention that the notion of potentiality to be an
F is different from the notion of power to change or be changed.
35 This seems clear in Aristotle's discussion of processes (Phys. 201•27- 05),
where bricks have the potential both to become a house and to be a house. In the
Metaphysics, bricks are used as examples of the matter (1049•9-11, see 1043a7-9)
of the house. So those things which are the matter of a house have at some point
the potential to become a house. (For this view of the role of bricks in processes,
see Charles (1984: 19-20).) However, this does not require Aristotle to identify
the potentiality to be an F with the potentiality to become an F, or to claim that it
is qua the matter of an F that bricks have the potentiality to become an F (even
though bricks could not possess the potential to be an F unless they also possessed
the potentiality to become an F). Indeed, Aristotle appears to focus in the
opening sections of e. 7 (1048b37-1049"r8) on the question of when something
first has the potential to be an F, and does not attempt the more ambitious project
of analysing this potential in terms of the potential to become an F. (Note the
temporal notions: 'when'. (48b36), 'anytime' (49•1), 'already' (49•2, 16).)
Matter and Form 97
use). By contrast, they are subjects of change in virtue of their posses-
sion of the former potentiality, which need only be present until the
house is built. The matter of the house, therefore, is what is present
unaltered (in the relevant respects) throughout the composing and the
composition of the house. On this account the house is composed
from this matter when it has been arranged in a certain way for a
certain purpose: protecting possessions. (Contrast the seed, which is
not yet the matter of a human being because it needs to be changed
before it can be the matter of a human: 1049a14-16.) On this account,
the matter cannot be strictly identical with the house (or its form)
since it has different persistence conditions.
(4) How are the relevant potentiality and actuality related in the
case of actually existing substance: e.g. a house? (This is discussed in
e. 8.) Since this actuality does not occur in another thing or in itself
as another thing, its relation to its potentiality must be different from
that between active and passive potentialities in cases of efficient
causal processes. The relevant potentiality is classified as a starting-
point of process or stability in itself as itself (1049b5-8). This is
because this actuality is what makes this potentiality what it is ( rn49b9-
II ), and so the potentiality's goal is to preserve this actuality over
time. In this way, the actuality is prior in account and being to the
potentiality (or nature), because the former defines the latter's goal
(and not vice versa). 36 Their interaction does not involve efficient
causation between distinct subjects of change. If it had done, actuality
and matter would have been two distinct entities (contrary to the
Unity Requirement) related in ways akin to that between the builder
and the bricks with which he builds. For subjects of change of this
latter type operate in another object or in themselves considered as
another object (e.g. a doctor healing himself: Phys. II. 8, 199b31,
Met. LI. 12, IOI9a17-18). 37
This chapter makes two further proposals about the nature of the
relevant relation. It makes it explicit that actualities (including sub-
stances in the strict sense of the forms of such substances, 105ob2-3)
are to be identified with goals ( 105d121 -4). Thus, in composite
substances, there is an intrinsic goal immanent in the substance which
makes the substance the one it is. Further, the intrinsic goal makes
36 This seems clear at Met. 10503 21-3, where Aristotle identifies the goal with
the actuality. See above, n. 24.
37 Mary Louise Gill offers a contrasting view of the role of efficient causation
(1989: r69ff.). She attributes an explanatory account of unity to Aristotle,
but in her view efficient and not teleological causation is the basic explanatory
ingredient.
David Charles
the potentiality the one it is ( 105oa8- I I): the capacity to realize an

actuality with goal G. Actuality and potentiality, thus understood,
constitute the ontology required to explain how matter and form can
be unified so as to share one teleological goal. For the goal makes the
actuality (the form) and the potentiality (the matter) what they are.
When actuality and potentiality share a goal in this way, the matter is
arranged as it is for the sake of a goal (e.g. safety of possessions, life
of a given kind), whose presence makes the composite substance (e.g.
house, man) the one it is. Indeed, what makes it the case that there
is one composite substance is that the relevant form, conceived as
actuality, and matter, conceived as potentiality, share one immanent
goal of this type. The relevant composite will persist for just as long as
actuality and potentiality are united in this way. Indeed, this is how
(on this view) the Unity and Persistence Requirements are fulfilled.
According to the explanatory interpreter, Aristotle in e. 6-8
separates in these four ways the type of potentiality and actuality
which is significant in considering composite substances. He isolates a
distinctive type of potentiality which characterizes matter (2 and 3), a
distinctive type of actuality (I), and a distinctive type of relation
between them (4). In doing this, he carries forward the explanatory
approach to the unity of composite substances initiated in z. 17 and
II. 6. The goal (according to the explanatory interpreter) which makes
the composite what it is can be specified (at least partially) indepen-
dently of the composite itself (in the way in which the goal of safety js
characterized independently of the house which achieves this goal).
The goal is what makes the actuality and potentiality the ones they
are. For actualities and potentialities are precisely the type of pheno-
mena whose nature and interconnection can be determined (in the
ways indicated) by such goals.
In this version of the explanatory account, the role of teleology is
central. If this is ignored; Aristotle can only be an explanatory theorist
if he adopts (what I shall call) a quick fix solution to the problem of
unity. In this actuality and/or potentiality are introduced independently
of the notion of a unified substance but without any further elucida-
tion, just because (it is claimed) one can see immediately that their
interconnection gives rise to a unified substance. However, this 'solu-
tion' fails to offer a genuine explanation of the unity of a substance. It
simply introduces two unanalysed terms (the correct relata), and says
that when the correct relata are correctly related there is a unified
substance. But this is no advance until one knows (at least) what the
correct relata are. Nor is much achieved merely by adding that the
correct relata in the case of composite substances are analogues of
potentiality and actuqfity in the case of processes. For even if these
Matter and Form 99
notions were immediately clear in the latter case, the basis of the
analogy would remain obscure. One would merely be told that the
correct relata in the case of composite substances are (in some un-
explained way) analogues of actuality and capacity in the case of
processes, and that they are correctly related to each other in a unified
substance in some (further unexplained) way analogous to that which
relates process-capacity and process-actualization. While this has the
form of an explanatory account (since at least some of the relata are
understood independently of the concept of a unified substance), it
lacks all content as an explanation. 38
In e. 6-8 the discussion appears to focus on composite substances
taken universally (e.g. man: 1049a1, house: 1049aII), and to discuss
actualities and potentialities considered as universals. But, as in H. 6,
Aristotle may intend to treat particular substances in a similar way. In
what follows I shall examine his proposal (as understood by the
explanatory interpreter) with special reference to the general case,
and merely indicate how it might apply to particulars. But this is to
remain steadfastly neutral on the issue of whether the general or
particular case is explanatorily prior.
VI. METAPHYSICS IN ARISTOTLE'S METAPHYSICS
I have argued in sections II- V that the explanatory viewpoint offers a

defensible interpretation of the passages we have considered, and that
taken as a whole it has several exegetical advantages. What does this
view, so understood, tell us of matter and form? In it the identity of
the goal, and so of form, is taken as basic, and the potentiality is the
one exercised in the actuality, in which the goal is immanent. But
what notion of matter is employed in this account? ls matter itself also
an independent starting-point? How are the goals themselves indivi-
duated? Does this account offer a genuine explanation (in any sense)
of the unity of a composite substance? What does this approach say of
the general project in which Aristotle is engaged?
Let us begin again with houses, and focus first on the role of matter.
Working from the top down, houses are made up of from planks and
bricks when ·they are arranged in a given way. If there are to be
houses, there must be planks and bricks arranged in a given way for
the sake of protection. It is the presence of this final goal that makes
these planks and bricks into a house. This is why houses can be
38 The 'quick fix' interpretation is not always clearly distinguished from the
dissolutionist, no-explanation, view. See above, n. 6.
IOO David Charles
defined as bricks and planks organized for a given teleological pur-
pose. Even if they had been, by accident, arranged in precisely this
way without that goal, the result would not be a house (but at best a
house look-alike). So it is the presence of a final cause which makes
the relevant planks and bricks into a house. Being a house, on this
view, cannot be reduced to being a mereological sum of bricks and
planks. The.re is more to its unity than that of the sum of its com-
ponents and their physical interrelations; for merely to specify these is
to ignore the goal whose attainment is required if there is to be a
house. And this will be so even if houses are composed from bricks
and stones, or the property of being a house 'supervenes' on given
properties of sets of bricks in given non-formally specified relations.
One introduces a higher level of ontology, houses, over and above
brick/plank collections (standing in given physical/spatial arrange-
ments) by invoking the final cause. Houses are the result of the
operation of the final cause as a principle which organizes the relevant
type of matter (in the ways indicated in the previous section).
Planks and bricks are introduced in a similar way. What makes a
piece of wood into a plank is that it is prepared for a specific purpose
(namely to be built with). One admits planks into one's ontology as
well as wood (or pieces of wood) when those pieces are worked into
the shape required for. building. The goals of being a house and of
house-building are internally connected, but not necessarily the same.
At each ascending stage in the hierarchy above the basic level
(level o),
o. Wood
I. Planks
2. House,
matter and final goal between them generate higher levels of entity,
and thus give rise to new composite substances on the basis of more
elementary ones grasped at the outset. (See Met. 1049ar5-25 for
comparable cases.) Thus, planks are to be represented as the result of
the operation of a given teleological cause on wood (as the matter).
And the house is composed from planks (and not, in the relevant
sense, wood), since wood has to undergo a further teleological transi-
tion to be made suitable to be turned into a house (without addition,
change, etc.). Thus, it is not wood-as such-which is the matter of
the house but wooden planks. This is not to say that there is no wood
in the completed house. It is only to say that the wood has to be
modified in certain ways before it can be the appropriate matter for a
house.
The defining cha:racteristics of levels 1 and 2 are derived from the
Matter and Form IOI
relevant goals at their level (respectively, building and being a protec-

tion). Wood is potentially a plank, as it is the material required if the
plank's goal (i.e. building) is to be achieved. But wood need not be
'alive' or 'pregnant' with plankhood or househood in any sense beyond
the relatively trivial one: it can be used to realize the goals which
define the higher-order entities: planks, houses. For wood itself can
be defined, as a stuff, independently of any connection with planks or
houses. In this way matter, as well as form, provides an independent
Aristotelian starting-point. In the generation of new levels, final
causation is basic because it is the principle which constructs one level
from the preceeding one. At each new level, the final cause deter-
mines the essence of the new composite substance, and fixes what it is
to be a house (e.g. to serve a given purpose). A house (the composite
substance, taken universally) will be planks and bricks arranged so as
to perform the purpose of protecting some goods. This particular
house will be these planks and these bricks arranged for this purpose.
The particular composite may indeed have its own particular form:
namely the form which belongs to this matter (and to no other matter).
And this may be so whether or not the nature of particular form is
derivative from that of the general form (or goal) which (on one view)
organizes this matter to make this particular composite. 39
Since Aristotle attempts to employ his principle of metaphysical
explanation quite generally, I shall conclude by considering the
problematic case of the human being. Let us begin with flesh, blood,
and bone (homoiomerous parts). Their material constituents can be
fully specified using (e.g.) heat, cold, and moisture (Meteor. LI. 12,
39ob2-10), and the elements (earth, fire, etc.) materially specified in a
way independent of sentience. They will constitute the basic level
(level o). Thus, as the first stage we can generate homoiomerous parts
as follows:
39 If one wished to represent these ideas somewhat more formally, and to take
the notion of general form as prior to that of particular form, one might proceed
as follows: begin with a = F(m; . .. m"), where a is the composite taken univer-
sally, and F a general teleological operation on matter of a given type. The
particular composite a* would be the result of the operation of F on m;* . .. mn *
(particular planks/bricks). The particular form which a* possesses would be the
form which is the result of the operation of general form F on m;* . .. m,," to
generate the composite a* with its distinctive form F*: the particular form which
belongs to m;* . .. m,, * and nothing else. If, on the other hand, one takes par-
ticular forms as explanatorily basic, one will begin with F*(m*; ... m"n) and
regard F(m; ... mn) as an abstraction derived from this case. And this will be so
whether F* is a particular in its own right or only so as a consequence of being the
one and only form which belongs to m;* . .. m",,. I am indebted at this point to
discussion with Kit Fine, who suggested this style of formalization to me.
I02 David Charles
material components (el earth, fire) serving goal G yield blood,
bone (cf. GA II. I, 734 31-4: with its reference to the account
(logos) of bone the (homoiomerous parts).
The second stage-for anhomoiomerous parts-is similar:
blood, flesh serving goal GI yield arms/hands.
Finally, we arrive at the highest level:
arms, hands, brain serving goal G2 yield the human being.
In this account, achieving goal G2 (e.g. a rational life of a given kind)
requires the successful realization of the lower-order goals (GI, G,
e.g. life involving movement or perception). At each distinct level in
the hierarchy new entities (such as blood, arms, human beings) are
introduced by continued reapplication of general principles of a
teleologically based process of construction, which operates pn matter
at the immediately preceeding level.
The human being is composed of, for example, arms and hands but
not of the basic material components (earth, fire, etc.), since the latter
require at least one additional teleological operation to be made into
matter (appropriate) for a human being. Thus it is not fire, as such,
that is present as the matter of a man, but fiery blood. This is noL,to
say that there can be no earth or fire (at all) in the completed man~''it
is only to note that it (like wood above) has to be modified in certain
ways before it can be the appropriate matter of a man. The priority of
form, and teleology, does not require that matter cease to be an
independent starting-point in Aristotle's metaphysics.
Against this background, there is no difficulty in specifying in G-
independent ways what the basic matter is at level o (wood, earth,
etc.). Nor is the possibility in any way excluded of specifying in G-
independent ways the types of combinations or mixtures of matter
present in human blood, flesh, or brains. But the crucial factor which
makes these new entities the ones they are, and fixes their essence, is
the teleological goal embedded in the metaphysical construction. And
the canonical way to specify the relevant principles of combination or
mixture must be in terms which invoke these goals. 40
40 The types of combination may differ from case to case: in one (e.g. parts of a
man) there may be a mixture (mixis) of components, in another a change in their
location or size (but no mixture): e.g. the case of a house. But in both the new
entity will be the result of the application of the goal to given matter (F(m;* . ..
m,.*) ), even though the mode of application differs. In the case of mixtures, the
basic elements will not be present-as in an atomist picture-when there is a
human being. (This point is emphasized by Mary Louise Gill) Even if mixtures
are to be specified canonically in goal-directed ways, this does not exclude the
possibility of there bemg ways of mixing basic elements so as to yield (e.g.) blood
Matter and Form 103
Aristotle's account of human beings, and their psychological states,
is a distinctive one. The teleological operations which introduce new
entities are not specifiable in terms of efficient causation, but require
reference to the goals at the relevant level. Nor are the new entities
themselves defined solely as 'whatever plays a given role in a scheme
of, for example, teleological explanation'. They are genuine entities in
their own right with their own essential properties and causal powers.
For Aristotle, desires and human beings are as ontologically respec-
table as tables and houses, or horns and horses. For these two reasons,
it would be importantly mistaken to align his account of desire or the
soul with any twentieth-century functionalist one. 41
From Aristotle's teleological perspective (as understood by the
explanatory interpreter), the form is explanatorily prior to the matter
of the composite, and for this and other reasons ontologically prior. 42
But in the order of construction, matter is the imput on which the
form operates to generate the new compound. So in the order of con-
struction, matter is compositionally prior to the compound, although
the form is explanatorily and ontologically prior to both compound and
which can be formulated in ways not involving such goals. See on this Charles
(r988: 30-6). Michael Frede makes a similar point with regard to house-building
(1992: ro1). Frank Lewis's discussion of mixtures, blood, and matter in this
volume (257-72) develops this possibility in an interesting and more detailed way.
41 For more detailed arguments against the functionalist interpretation of
Aristotle's psychology, see Charles ( 1984: 227-34). Code and Moravcsik ( 1992:
129-44) also reject this functionalist interpretation, but suggest, in extreme
reaction against it, that Aristotle's key notions of potentiality, actuality, matter,
form, taking on form without matter, etc. are all 'basic', 'primitive', and 'un-
defined' (in his theory) (1992: 137, r43). Their proposal, however, appears to
overlook the important connection between the final and formal cause (or nature)
of the relevant actualities of substances (cf. Met. ro41a27-32, ro44"36-b1,
rn50•21-3, etc.), which provides the basis for the explanatory viewpoint. The
latter approach attempts to illuminate Aristotle's key notions of form, actuality,
etc. without offering a reductive definition of them as (e.g.) 'whatever plays a
given functional or teleological role'. (For this latter distinction, see Grice 1975.)
Thus it strives to avoid the simple and stark dichotomy (either Aristotle was a
functionalist or else he treated forms as 'undefined primitives'), which underlies
much modern discussion of these topics.
42 Met. e. 8 presents a variety of arguments in favour of the ontological and
explanatory priority of actuality to potentiality. For this reason, Aristotle's view is
distinct from that of modern non-reductionist materialists who like Hilary Putnam
(1988: 73ff.) separate ontological from explanatory priority, and regard matter
as ontologically, but higher-order phenomena as explanatorily, prior. However,
Aristotle does regard matter as compositionally prior, 'the that 'from which', and
this is sufficient to render his views as 'materialist' on some weak but plausible
construals of 'materialism'. On this latter issue, see my paper (1992b: 265-96).
David Charles
matter. Given this separation between compositional and explanatory/

ontological priority, Aristotle can agree with materialist reductionists
that matter is compositionally prior, and with 'vitalists' that life (or
form) is explanatorily prior, without being forced to treat matter as
explanatorily prior (as in atomism) or conversely to regard it as 'alive
with sentience/form' in some wholly mysterious way (as in some
vitalist, pre-Cartesian, writings). Even sophisticated interpreters fail to
grasp the distinctive mid-position which Aristotle defines. In justified
reaction against attempts to depict him as a materialist reductionist
(for whom matter is explanatorily, compositionally, and ontologically
prior), they represent him as construing matter as in all respects
secondary, and so as advocating a vitalist conception. 43 But this is to
miss the philosophical significance of the viewpoint which Aristotle
advocates.
Aristotle's account is motivated in turn by his understanding of the
metaphysical project as set out in Met. e: to explain the nature of
higher-order composites (such as humans or houses) by using relevant
teleological goals to fix their nature. In this way, he explains what it is
to be an Fin a way separable from and prior to the account of how Fs
come to be or of the matter from which they are composed. From this
perspective, metaphysics can be conceived as the first philosophy, a
study prior to the a posteriori investigation of efficient causation or the
search for the physical bases for any proposed materialist reduction.
The problems with Aristotle's approach (so understood) are numer-
ous. Can one in fact individuate the relevant types of goal (immanent
in the actualities) independently of the composite substances they
enform? How far is it plausible to suggest that this basic actuality is a
simple unity which is the essence of an animal? What is the nature of
the teleological explanation required to underwrite this view in the
case of substances? These three questions are especially acute in
Aristotle's biological writings. How far can one individuate types of
animals solely in terms of their distinctive teleological goals and
means of achieving them? Does the form of the animal yield the
simple feature required by the Analytics model? Can Aristotle explain
43 See Burnyeat (1992: 25-6) and Kosman (1987: 388-91). In my view they are
correct to insist (against functionalist interpreters) that the principles of organiza-
tion (F) which govern the compound (F(m; ... mn)) cannot be fully understood in
terms of (e.g.) purely mechanical (non-teleological) properties of matter, but mis-
taken to conclude that m; ... m 11 cannot be fully specified in the latter terms.
The functionalist interpreter makes precisely the opposite error. Both overlook
Aristotle's mid-position: a plank is matter organized for a teleological goal (i.e.
house-building), where ,the latter cannot be fully captured in mechanical efficient
causal terms. '
Matter and Form 105
teleologically the existence of parts of an animal organized in a given
way? These questions raise acute difficulties for the teleological
approach to metaphysics to which the explanatory interpreter is
committed. 44 However, these issues, essentially involved in the meta-
physical construction we have been considering are (I conjecture) the
central ones for a proper understanding of Aristotle's metaphysical
project. 45
44 I discuss the second and third questions elsewhere ( 199w: 215-49; 1991:
101-28). On the latter issue, see also Charlotte Witt's discussion (this volume:
215 ff.). I am indebted to Nicholas Bailey for much discussion of issues concerning
teleology and metaphysical construction.
45 Earlier versions of parts of this essay were read at the Oxford Conference on
Aristotle's Metaphysics in 1989, at the Boston and Los Angeles Colloquia in
Classical Philosophy, and at Sendai University in l 991. I have gained from
discussion of these issues with Nicholas Bailey, James Bogen, Robert Bolton,
Myles Burnyeat, Kei Chiba, Kit Fine, Michael Frede, Montgomery Furth, Mary
Louise Gill, Allan Gotthelf, Wan-Ju Hyeon, Terence Irwin, Yasuo Iwata, Aryeh
Kosman, Frank Lewis, Martha Nussbaum, Theodore Scaltsas, Rowland Stout,
Pantazis Tselemanis, Sarah Waterlow, and Charlotte Witt.
5
Substantial Holism
THEODORE SCALTSAS
Substances are the most unified wholes in Aristotle's metaphysics,

being paradigmatically single but complex entities. My concern in this
paper is to understand the nature of their unity, and through it,
explicate an important aspect of the relation of the potential to the
actual, . which is fundamental in Aristotle's system. Although the
nature of this relation is illuminated by examples offered by Aristotle,
this is not sufficient, in view of the enormous explanatory weight that
is placed by Aristotle on this relation. An attempt must be made
to understand why it is that this relation can deliver such fundamen-
tal and rather spectacular-denigrated as 'magical' by its critics-
metaphysical results.
1 understand Aristotle's solution to be that the unity of a substance
is not achieved by relating its components to one another; rather,
unity is achieved by dissolving the distinctness of each of the sub-
stance's components. One of the threats to the unity of a substance is
the multitude of the material components that go into its creation, or
result from its destruction. The plurality of these components is prima
facie incompatible with the unity of the substance. But it is not jusC
the matter out of which the substance is created, or the matter into
which a substance is dissolved, that must be shown not to 'pluralize'
a substance. It is also, importantly, the many components of the
substance, which are synchronous with the substance, namely, its
constituent matter, its properties, its form, that pose just as much of a
threat, if not a more significant one, to the substance's oneness. ~.
concrete substance is a composite of matter and form, as Aristotle
repeatedly states. 1 The reality of matter and form in a substance is
established by separate existential arguments that Aristotle offers,
which therefore attest to the complex nature of a substance. 2 Thus,
the unity of a substance is put to the test, not only because of what
© Theodore Scaltsas 1994

1 e.g. Met. rn29a30-r, rn33°24-6, rn35"r9-21, rn37"7-9.
2 I thus agree with Frede (this volume: 187-8) that matter and form are real
for Aristotle, although they are not in actuality in a substance. Our concern will
108 Theodore Scaltsas
there was before the substance was composed, and what there will be
after it disintegrates, but also because of what there is while the
substance lasts.
The question that confronts one then is, how can the composite
substance, which is composed of matter, form, properties, be one,
rather than a plurality of many? Why doesn't this compositeness of
the substance render it into a sort of aggregate, or cluster of copresent
entities? It is this question that will occupy much of our attention in
this paper. I will argue that Aristotle's metaphysics offers the same
answer for the" unity of a substance, whether its components are taken
to pre-date it, post-date it, or to be synchronous with it. The same
answer also applies, whether the components are concrete or abstract
ones, since whether the substance is divided physically or by abstrac-
tion, the result is distinct components that are incompatible with the
oneness of the substance. Aristotle's solution is that all these distinct
components that are derived by dividing the substance either physically
os by abstraction are only homonymously components of the sub-
stance. They are incorporated into the substantial whole by being
reidentified, in accordance with the role they have in the whole (dic-
tated by the substantial form). 3 That the components that go to make
up a substance are identity-dependent on the form is expressed by
Aristotle in his claim that the· components are potentially what the
substance is. The potentiality of the components, whether they are
concrete bits of matter or abstract entities, is, for the actuality, deter-
mined by the substantial form. Thus there is an identity dependence
between the components and what the substantial whole is, since the
components' potentiality is determined by what the substantial whole
is. The resulting whole is therefore unified, not by internal relations
between distinct components, but by the identity dependence of the
(concrete or abstract) components on what the substantial whole is.
One of the conclusions of the present analysis will be that abstract-
ing parts from a whole is as much of a division of the whole as
physical division. Abstraction is not an ontologically neutral way of
describing a whole, but it dissects and pluralizes as much as physical
dispersal does. It is therefore not open to a metaphysician, and cer-
tainly not advocated by Aristotle, to account for the unity of a sub-
stance on the physical level, while treating it as a plurality of abstract
be to show, first, the difference between being real and being in actuality, and
second, why a multitude of real ~omponents does not result in the substance being
many.
3 In what follows, I shall speak of the dependence of the constituents on the
whole, meaning that they depend on what the whole is, not which whole it is.
Substantial Ho/ism
components. My claim will be that for Aristotle a substance is com-
plex, not because it is a conglomeration of distinct abstract com-
ponents like matter, form, or properties; a substance is complex
because such items can be separated out by abstraction, which is a
kind of division of the unified substance.
I. TYPES OF UNITY
The main difficulty in the unification problem is not for the many to
make up something single; rather, it is for the many to cease being
many. By that I mean that any number of entities can be unified into a
group just by including them in a group. Thus, this car, that tree, and
the house behind it are one, just by being classified as belonging to
this group of the three of them: they are one group of things. Clas-
sification into a group needs no more justification than listing the
items of the group. The ground for such unity is convention, and
the resulting oneness is entirely compatible with the plurality that the
things in the group constitute. This type of unity, is incidental to the
things that are unified. Diametrically opposed to it is substantial unity.
As we shall see, the grounds for substantial unity are metaphysical,
and the oneness of a substance is incompatible with the plurality of its
components.
In between the two extremes of unification, namely the conven-
tional and the metaphysical ones, are different types and levels of
unity whose grounds are physical, and weak metaphysical ones. Thus,
Aristotle distinguishes between being one by juxtaposition (as in the
case of the grains of sand in a sand-hill), being one by physical
connection (as in the case of a bundle of sticks), being one in the way
that an artefact is one, and finally being one in the way that a
substance is one. 4 Before turning our attention to the metaphysical
unity of organic substances, it should be observed that not any group-
ing of entities, grounded on metaphysical relations, achieves the unity
of substances. For example, two distinct substances can be one with
respect to their matter, if the matter of the first is the matter out of
which the second was also made. 5 But their oneness is compatible
with their being distinct and many. Similarly for substances which are
one in form, or simply qualitatively similar. T!:reir oneness is com-
patible with their being many. 6
4 e.g. Met. 1052"15-25.
5 Phys. 19ob24-5, Met. 1016"27-8.
6 Met. 1016b36- 1017"3. A more peculiarly Aristotelian and rather fanciful case
of unity is the unity of the white entity with the musical entity, when it happens
no Theodore Scaltsas
It is doubtful whether there is a difference of degree of unity in the

cases that result from things which share the same matter (at different
times), or the same property, or the same genus, or species. Are
things that have been made out of the same matter, e.g. by remoulding
the same quantity of brass, more of a unity than things which are the
same in colour, or in genus or in species? Things that share the same
species may be more similar than things that share a property or their
matter, but are they more of a unity? Generally, being one in these
senses is being similar with respect to an aspect or a component. But
this does not unite them into anything more than a group of different
substances. Their being one (in species, matter, quality, etc.) i<S com-
patible with their being a plurality of many substances.
So, the group of the car, tree, and house, is less of a single thing
than a sand-hill, or a bundle of sticks, and each of these is less of a
unity than a living organism. There is a sense in which even the sand-
hill or the bundle of sticks are aggregates of many, or wholes of
interrelated distinct things. The paradigmatic unity, namely, substan-
tial unity, is not compatible with being a plurality of many. The nature
of such a unity is the question with which Aristotle is concerned in the
central books of the Metaphysics. It is not sufficient for him to find a
respect in which the many components of a substance are one, if this
leaves the substance being a plurality of many. Rather, the com-
ponents of a substance are one in the sense of being a unified whole,
not a plurality of many. What is required is an account of a substance
which will show it to be complex without being plural, to be complex
and atomic. David Lewis has described the task accurately, although
he aims to avoid it7 in his system:. 'On the magical conception [of
composition], a structural universal 8 has no proper parts. It is this
conception on which "simple" must be distinguished from "atomic".
A structural universal is never simple; it involves other, simpler,
that one and the same substance is white and musical (Met. 1017a7-18). Here, the
unity involved is metaphysical, not physical. The white and the musical are not
grouped together because they overlap in the same space at a time. That would be
very similar to the relation of a sponge and the water that permeates its pores,
which is no more than a kind of juxtaposition, since it is possible to physically
separate the sponge from the water. But in the case of the white and the musical
we cannot separate the two entities physically from one another. Their relation is
not physical, but metaphysical: they are both instantiated in the same substance;
belonging to that substance as subject is what groups them together into a class.
7 In Scaltsas (1990), I have argued that David Lewis cannot avoid this kind of
unity even within his own ontology. (See in particular p. 595.)
8 For our present purposes, a structural universal can be thought of as a
substantial form.
Substantial Ho/ism III
universals ... But it is mereologically atomic. 9 The other universals

it involves are not present in it as parts. ' 10 Aristotle distinguishes
between three types of unity (regardless of whether the unified items
are concrete or abstract entities): aggregates, related wholes, and
substantial unities. In this paper I will concentrate on .the way that
unity is achieved in substantial wholes.
II. AGGREGATES, RELATED WHOLES,

AND SUBSTANCES
The least unified item is an aggregate of entities. An aggregate is

a collection of entities which stand unrelated to one another. The
aggregate can be described only by listing its members; there is nothing
that unites its members into a group of any kind. In that sense,
there is nothing that can provide any ground for the oneness of an
aggregate: it is a plurality, and only a plurality, of its members. To use
David Armstrong's example: 'The aggregate of all armies is identical
with the aggregate of all soldiers', 11 although an army is not a soldier.
By contrast, a group of interrelated entities does have a claim to
oneness, because it constitutes a related whole. Thus, the students in
a class are one in so far as they are interrelated and together they
constitute a single entity, the class. Here, it is possible to make
distinctions between interrelated wholes, depending on the type of
bond that relates them; e.g. if they are merel1' juxtaposed or unified
by glue or nails, or by being tied together. 1 There are borderline
cases between aggregates and related wholes, and borderline cases
between related wholes and single objects. Thus, we may waver as to
whether the objects on a beach are an aggregate or a related whole.
We may also waver as to whether the sticks in the fence are a related
whole, when they are in contact with one another, or just an object,
the fence. (Whether some things constitute a related whole or not
may be a matter of convention, or of functional efficacy, but may
also be a result of scientific considerations, e.g. in political science,
zoology, etc.)
9 i.e. it does not have distinct elements, in the way that (for Aristotle) the
elements of an aggregate are distinct. See the discussion of the aggregate argu-
ment below.
10 David Lewis (1968: 36), my emphasis.
11 Armstrong (1978: 30).
12 Met. 1052•19-20. Aristotle introduces further criteria of unity, e.g. depend-
ing on the degree to which a thing's movement is one and indivisible in place and
time.
II2 Theodore Scaltsas
The question that faces us finally is: if there is anything that has a
higher degree of unity than an interrelated whole, how is this unity
achieved? The answer to this question is given by Aristotle in his
aggregate argument, and in his analysis of the components of a sub-
stance in terms of the potential-actual distinction.
The aggregate argument is offered in Met. Z. 17. I have analysed
this argument in detail elsewhere, 13 so I will limit my discussion here
to the briefest mention of its premisses and its conclusion. Aristo!ie
gives the example of a syllable and investigates whether the syllable is
anything other than the letters that constitute it. 14 He claims that it is:
'As regards that which is compounded out of something so that the
whole is one-not like a heap, however, but like a syllable-the
syllable is not its elements, ba is not the same as b and a, nor is flesh
fire and earth' (1041b12-14). His argument is the following. Given a
whole like a syllable, when the elements constituting the whole are
dissolved, the whole does not exist any more, but the elements do.
Hence, the whole must be something over and above the totality of its
elements. If the difference between the whole and its elements is a
further element, then the same argument applies again: if we dissolve
the totality of the elements, the whole will be no more, but all the
elements will be. Hence, the whole cannot differ from its elements by
an element, since this still leaves the whole identical to a totality of
elements; but the elements can survive dissolution, qua aggregate,
while the whole cannot. (So, the hypothesis that a whole is over and
above its constituent elements by something present in the whole as
an element of it cannot explain why the whole is lost when dissolved,
even though all its elements survive.) Therefore, Aristotle concludes,
the whole must be over and above the totality of its elements by
something which is not present in the whole as a further element in
it, but by something that '1s the cause which makes this thing flesh and
that thing a syllable' (1041b26-7). And Aristotle continues:
this is the substance of each thing; for this is the primary cause of its being;
and since, ... as many [things] as are substances are formed naturally and by
nature, their substance would seem to be this nature, which is not an element
but a principle. (1041b27-31)
This is a complex conclusion to the argument, with a wealth of
n Scaltsas (1985). This is an existential argument for the substantial form in

a substance, which offers metaphysical reasons for positing substantial forms,
rather than merely linguistic ones which William Charlton mentions (this volume:
51-2).
14 In Scaltsas ( 1990) I showed the proximity between the aggregate argument
and Plato's discussion of the part-whole relation in the Theaetetus, 204-5.

Substantial Ho/ism I 13
metaphysical concepts being interconnected in it, whose significance

we need to analyse in what follows.
The minimal conclusion that we need to draw from this argument is
that the item by which the whole is over and above the totality of its
elements is of a different ontological type from these elements. This is
required for avoiding the regress that would ensue, as shown in the
argument. This result, significant as it is for the ontology of the
system, does not enlighten us as to what kind of entity that item will
be, other than that it will not be like the elements of the whole.
Aristotle does describe it as the primary cause of the being of the
whole, as the substance of the whole, as its nature, and its principle.
Although all· these descriptions are suggestive, pointing to a funda-
mental difference between the elements of the whole and that extra
item that makes them into a whole, much more needs to be explained
before it is appreciated precisely what this difference consists in.
Is the aggregate argument telling us anything more than that a
whole is over and above its elements by a relation, which is not an
element in the whole, but an item of a different ontological type? In
itself, this would be informative, in so far as it differentiates between
an aggregate of elements, and a related whole. Namely, an aggregate is
identical to the totality of its elements, whereas a related whole
consists of the elements plus an entity of a different status, which
relates the elements into a whole. But we know that this is not what
Aristotle is driving at here. The reason is that he identifies that extra
item with the substance of the resulting whole, and we know that for
Aristotle, substance is not relation. 15 Otherwise, the category of sub-
stance would reduce to the category of relation, which would make
nonsense of the whole of Aristotle's metaphysics. It is essential, there-
fore, that the cause of being of a substance, which unifies the substance
into a whole, not be construed as a relation; and that a substantial
whole be distinguished from a related whole.
I have argued elsewhere that it is possible to pay justice to the
Aristotelian distinction between substantial forms and relations, when
we realize that what is characteristic of Aristotelian relations is that
they do not alter the identity of their relata. 16 Thus, Callias' being
larger than Socrates, or the sun's moving around the earth, 17 are
relations which do not determine who, or what, Callias or the sun are.
This feature distinguishes relations from substantial forms, as we shall
see in what follows.
15 See the distinction in the Categories between the substance category and the
category of the relatives-relations, and Met. 10883 23,•30, bz, EN 1096•2r.
16 In Scaltsas (1990: 588-9).
17 Met. 1040•29-3r.
Il4 Theodore Scaltsas
The aggregate argument of z. 17 shows that the unification of an

aggregate of elements into a substantial whole requires an entity over
and above the elements which are unified. This naturally gives rise to
a further problem: what unifies the elements of the aggregate and this
new entity? More specifically, how is the regress of unifiers that
Aristotle warns us about in Met. Z. 17, rn41b22, to be avoided?
Certainly the unifier cannot be unified with the elements in the way
that the elements are unified with one another (by the unifier), for
then more unifiers would have to be posited to do that job, and
the regress would ensue. Aristotle realized this, and faced up to the
problem in Book Hof his Metaphysics. He did so by introducing the
notions of the potential and the actual to explain how it is possible for
the unifier to unite the elements of a substance into a whole, without
requiring further unifiers to bond the elements to their unifier. The
question facing us is, how is that possible?
Before embarking on an analysis of Aristotle's solution, I wish to
argue that although the aggregate argument in Z. 17 is presented in
terms of the unity of the concrete components of a substance, in fact,
it also applies to the problem of the unity of the abstract components
of a substance. The argument does not presuppose the concreteness of
the elements in the aggregate; rather, it is a general argument that
investigates the way that any aggregate of elements can be united into
a whole. Aristotle's solution applies to any part-whole relation,
whether the parts are concrete or abstract. Thus in H. 6 he introduces
the problem of the unity of wholes as follows:
To return to the difficulty which has been stated with respect to definitions
and numbers, what is the cause of each of them? In the case of all things
which have several parts and in which the whole is not, as it were, a mere
heap, but the totality is something besides the parts, there is a cause of
unity ... And a definition is a formula which is one not by being connected
together, ... but by dealing with one object. What then is it that makes man
one; why is he one and not many, for example, animal-biped, especially if
there are, as some say, an ideal animal and an ideal biped? Why are not
those Ideas the Ideal man, so that men would exist by participation not in
man, nor in one Idea, but in two, animal and biped? And in general man
would be not one but more than one thing, animal and biped. Clearly, then, if
people proceed thus in their usual manner of definition and speech, they
cannot explain and solve the difficulty. But if, as we say, one element is
matter and another form, and one is potentially and the other actually, the
question will no longer be thought a difficulty. (1045a7-22, my emphasis)
Plato's account of participation in Forms renders sensible substances
into bundles of properties. The partaker partakes of Forms, and the
resulting substance i~ an aggregate of instantiated Forms. After con-
Substantial Ho/ism IIS
sidering and dismissing the option that a whole is something over and
above its parts (by the cause of its unity, for Aristotle), Plato con-
cludes that a substance is identical to its parts. 18 It is not clear whether
the instances of Forms in a substance are universal, or they are
particular (i.e. tropes).1 9 But in either case, they render the substance
into a plurality of (abstract) parts, which presents as much of a
challenge to the unity of the substantial whole as the material parts of
the substance do. Thus in Z. 12 Aristotle puzzles about the unity of
the forms mentioned in the definition of a substance:
wherein consists the unity of that, the formula of which we call a definition, as
for instance in the case of man, two-footed animal; ... Why, then, is this one,
and not many, namely animal and two-footed? (103i'II-12) 20
And even if the genus shares in them [the differentiae], the same argu-
ment applies, since the differentiae present in man are many, for example,
endowed with feet, two-footed, featherless. Why are these one and not many?
(1037b21-3)
For Aristotle, the problem of the unity of a substance is seen as being
the same whether the constituents of a substance are abstract or
concrete entities.
Finally, Aristotle's concern for the threat that the abstract com-
ponents in a substance pose for the unity of a substance is further
verified by the scope he gives to the unity question in Z. 17. The
question ranges from the unity of a man and being musical (1041a10-
14), to the unity of a man and being an animal of such and such a nature
( 1041a20- r), the unity of the subject and what is predicated of it
(1041a23), of the sound and the clouds (when it thunders, 1041a24-5),
of the bricks and the house they make up, the bricks and being a house
(1041a25-7, b5-6), of a body and the state it is in (1041b7), and
generally, of the many parts and the whole they make up (1041b2). In
H. 3, while presenting the second aggregate argument, 21 he rejects
18 Theaetetus 205"8- IO. See my discussion of the Platonic account of the unity
of substances, Scaltsas (1990: 583-5).
19 Tropes are instances of properties which are particular-not universal; for
example, the particular wisdom or whiteness instantiated in Socrates.
20 It should not be thought that Aristotle is not concerned here with ontological
unity. That he is becomes clear from the fact that he immediately compares the
issue at hand with the unity of man and being white in a white man: 'For in the
case of "man" and "white" there is a plurality when one term does not belong to
the other, but a unity when it does belong and the subject, man, has a certain
attribute; for then a unity is produced and we have the white man' (103?1'14-18).
Here the unity is ontological (the subject, man, has the attribute, white), not
definitional.
21 For an analysis of the second aggregate argument in H. 3, see Scaltsas (198y
229-30).
mere copresence as an explanation of wholeness: 'The syllable is not
produced by the letters plus juxtaposition, nor is the house bricks plus
juxtaposition' (1043°5-6).
Let us then address the question of the unity of the abstract com-
ponents of a substance. For Aristotle this would include all items which
cannot be physically separated from the substance, but which can be
separated by description. 22 It does not follow that only properties
would be included in this aggregate. Any component of the substance
that is not physically separable from it, but is separable by abstraction
only, is abstract, including such items as the matter of the substance.
Not the matter in the sense in which the log is the matter from which
the box is constituted, 2.:i or in the sense in which water is the matter
into which vinegar disintegrates, 24 but matter in the sense in which
bread is the matter of a loaf. Bread cannot be physically separated from
the loaf (in the way that the log is separate from the box constructed
from the log, or the water from the wine); bread can be separated from
the loaf only by abstraction.
We therefore face the problem of the unity of a substance, even if the
aggregate is an aggregate of abstract components. In the case of the
matter and the form of the substance, the aggregate argument would be
applied as follows: Let us assume that a substance is the aggregate of its
matter and form. Now, the matter of a substance, for example, the
wood in a pine tree, can survive in another substance, for example, a
statue. The form of the pine tree, i.e. being a pine tree, also exists in
other pine trees. 25 But the original pine tree does not exist after we
make the statue out of it. Hence, it is possible for the matter and the
form of the pine tree to exist, without the tree existing, which means
that the pine tree is not the aggregate of its matter and form. 26
Furthermore, the tree is not matter plus form plus some further
element, since the same argument would apply again. What is required
22 Met. I042"29. Singling out the form by description divides the substance into
two abstract components, matter and form.
23 Met. 1049"23-4.
24 Met. 1045"1-2.
25 There is some affinity between the argument I am developing here and one
used by Armstrong (1991: 189-200). Armstrong's argument has to do with the
truth-maker for the proposition 'a is F', which requires more than the existence of
a and F.
26 As we shall see, the solution Aristotle offers is that a substance is not an
aggregate of matter and form. Rather, form is a principle, not an element of the
aggregate of substantial components, and the components that go into the make-
up of the substance are potentially what that principle stands for. If the form were
not itself a principle, then some such principle would be required for the unifica-
tion of the matter with the form in a substance.
Substantial Ho/ism II7
is that one of the items in the tree not be an element of an aggregate,

but a principle which unifies the rest into a substance-the tree.
But it might be objected that if there are individual forms, the
aggregate argument is rendered ineffective since the form of a sub-
stance cannot exist if the substance does not. To make the objection
even stronger, let us assume that, not only the substantial form, but a
substance's properties, too, can neither exist in other substances, nor
be dispersed. I will offer an argument, which I shall call the trope-
overlap argument, to show that a substance is over and above the
aggregate of its copresent, particular, properties; i.e. Aristotelian
substances are not bundles of tropes. Let us consider the aggregate of
Socrates' properties, and assume that all these properties are par-
ticular, i.e. tropes. Not being universal, the properties do not exist
in other substances, and therefore the argument presented in the
previous paragraph would not be forestalled. 27 Let us further include
among Socrates' properties second-order relations, namely relations
between first-order properties; for example, the relation of the nec-
essary copresence of colour and weight. It follows that the second-
order relational properties would not survive the dispersal of the first-
order properties (even if such a dispersal could be achieved). Hence,
the aggregate of Socrates' properties would not survive the dispersal
of its members. Therefore, the aggregate argument cannot be applied
to the aggregate of Socrates' tropes to show that that aggregate is
different from Socrates.
It is not dispersal that will show the difference between the aggregate
of Socrates' copresent, particular, properties and Socrates, but over-
lap. Within the context of Aristotelian physics, let us consider the
overlap of a sponge, which is submerged in water, and the water in
the location of the sponge. Within contemporary physics, we can
consider the overlap of Socrates and the nutrinos that shower through
him all the time. The properties of the water and of the sponge make
up an aggregate of copresent properties. Similarly with the properties
of the nutrinos and Socrates. The properties in each aggregate are not
necessarily copresent. But then, even the sponge's properties are not
all necessarily copresent with each other, since the sponge can survive
change. Consider then the aggregate of the properties of the sponge
and of the water. Why is that aggregate two things and not one?
Similarly with the aggregate of properties of Socrates and of the
nutrinos. Why are they not just one substance, but many? That
27 It was there assumed that the form of 'pine tree' of this tree also exists in
other pine trees.
we can separate the water from the sponge, or Socrates from the
nutrinos, cannot help us answer the question. It could be that the
initial aggregate is a single substance that divides into water and
sponge, or Socrates and nutrinos, very much like an orange divides
into slices, or an amoeba into two amoebas. The possibility of separa-
tion of the properties in an aggregate is compatible, equally, with
the initial existence of only one, or of more than one, substance.
Hence, the possibility of separation of the aggregate cannot help
us answer the question of whether the aggregate is one substance or
many.
If a substance is an aggregate of properties, then two copresent
substances will make up a single substance, since two aggregates make
up a single aggregate.' But there is no single substance that the sponge
and the water make up, or that Socrates and the nutrinos make up.
Hence, the sponge, the water, Socrates, the nutrinos, are not aggre-
gates of properties. More generally, a substance is not an aggregate of
copresent properties. A substance must differ from an aggregate by
something other than a further element, since this would still leave the
substance as an aggregate, only with more elements in it. A substance
must differ from an aggregate by something that will unify the sub-
stance in a way that an aggregate is not unified. The way a substance
is unified must explain why two substances do not make up a single
substance, while two aggregates make up a single aggregate.
The above (trope-overlap) argument shows that the distinction be-
tween an aggregate and a substantial whole remains even when the
constituents of a substance are taken to be the substance's particular
properties-tropes. It follows that the unification of a substantial
whole cannot consist in adding a further property (e.g. a particular
relational or a structural property) to the aggregate of the substance's
particular properties. That would simply augment the aggregate of
properties, for example, by a particular relational property, but not
unify it into a whole.
From the aggregate argument and the trope-overlap argument, it
follows that the abstract components of a substance cannot be unified
into a whole by a relation. Aristotle explicitly rejects such an account
of unification in Met. H. 6, when he rejects all types of relation such
as participation, communion, composition, connection posited by his
predecessors to unite such abstract entities as: whiteness and surface,
knowing and soul, health and soul, bronze and triangle, body and
soul. 28 He therefore owes us an account of the unity of the com-
ponents of a substance that does not render the unifier into a relation
Substantial Ho/ism I 19
between these components. But if not related, how are the com-
ponents of a substance unified into a whole? I will argue that according
to Aristotle, the components of a substance are unified into a whole
by losing their distinctness as they are incorporated into the whole. It
is not that they remain distinct and related to one another; rather,
they are unified by losing their boundaries, like a drop of water that
merges with the water in the glass at the cost of its distinctness. When
a component loses its boundaries by merging into the whole, the
component becomes identity-dependent on what the whole is, and
hence, not causally related to the whole (since a thing cannot be
causally related to itself). It is this identity dependence of the sub-
stantial parts on what the whole is that Aristotle aims to present by
introducing the potential-actual distinction in his final description of
the unity of a substantial whole, in H. 6. Briefly, what is characteristic
of the potential-actual relation is that x's potentiality is defined in
terms of y, where x is not y (either in the predicative or in the identity
sense of 'is'). When the potentiality is substantial, i.e. y is a substan-
tial form, 29 the actualization of the potentiality changes the identity of
the potential. What unites the parts of a substance into a whole is that
they realize their potentiality to be y, thereby losing their distinctness
and becoming (actually, not in potentiality) identity-dependent on the
form, y. 30
Ill. SUBSTANTIAL HOLISM
The problem Aristotle was facing is difficult. On the one hand, the
aggregate argument requires him to posit an extra item as a unifier of
all the diverse elements that constitute a substance. But Aristotle
29 e.g. of the bricks to become a house, rather than their potentiality to

become bleached by the sunlight.
311 It follows from this that I disagree with one of the views sketched out in
David Charles's contribution in the present volume, the 'non-explanatory view·.
according to which: 'if one represents animal two-footed in the form: matter:
shape:: potentiality: actuality. no further explanation of their unity is needed. This
is explanatory bedrock. Nothing more can be said to explain the unity of a
composite substance, because these notions are abstractions from that of a unified
composite substance' Charles (present volume: 87). But abstraction is not
ontologically neutral; rather, it interferes with the identity of the components of a
substance. It is therefore not clear why the extracted abstract entities would
explain anything about the unity of the components of a substance. Even saying
that they are one as the potential is one with the actual is in need of explanation.
Aristotle offers this explanation through his homonymy principle, as we shall see
below.
knew only too well what the consequence would be of positing a cause
of being for a substance which is other than the substance itself.
Whether the cause of being is a separate entity (like the Platonic
Forms) or a part of the concrete substance makes no difference; so
long as it is distinct from the concrete substance and it is possible to
relate the concrete substance to its cause of being by some kind of
causal relation, an infinite regress ensues. The reason is that if the
essence of something is a distinct entity, different from that thing,
then the essence will itself have a distinct essence, and so on ad
infinitum (1031b28-1032a4). So positing a substantial form (as a
unifier of the substance's components) as the aggregate argument
requires would threaten to open the gate to an infinite regress.
Furthermore, the substance would then be a related whole of distinct
components, namely the form and the matter~ This is so because the
aggregate argument requires the positing of the form in a substance,
and physical continuity in change requires a substratum surviving in
the substance. But then the substance would be a plurality (as related
wholes are), not a unified whole. Yet Aristotle wants to show that a
substance enjoys a far higher unity than that of a related whole of
distinct components. 31 So it seems that on the one hand he needs
different components to perform different functions in the substance,
but on the other, he does not want the plurality of these components
to undermine the unity of the substance.
The resolution of the dilemma is a measure of Aristotle's genius. It
rests on the introduction of the notion of potentiality, which allows for
something to be present without being present! The potentiality is
present although that which determines the nature of the potentiality,
namely the actuality, is not present. Hence, what is shared between
the potential and the actual cannot be a component they possess in
common.
The innovation that Aristotle is introducing here is to explain same-
ness between the potential and the actual in terms of two different
ways of being/, rather than in terms of a shared entity. 32 Then, along
with similarity, difference and, hence, change, can be explained
accordingly. Similarity is not the shared presence of a component, nor
is change the replacement of a component in a substance. Neither
sameness nor change requires a substance to be a cluster of distinct,
31 They are distinct because the form would be definable independently of the
matter, and the material substratum would have a nature and life-span indepen-
dently of the form it enmatters.
32 Sharing distinct components in common was the Platonic way of explaining
sameness, e.g. sharing a part of the Form with the Form.
Substantial Ho/ism 121
interrelated components. Rather, through the potential-actual rela-

tion, Aristotle is introducing identity dependence in place of copresence
of components: the components of a substance are identity-dependent
on what the whole is, and therefore cannot exist severed from the
whole. What a substance 'shares in common' with other things exists
only potentially in a substance, and can be derived only by dividing
the substance up (physically or by abstraction). 33 Thus a substance's
components are (only) potentially entities that can exist independently
of the whole, and conversely, entities existing independently of the
whole are (only) potentially the substantial whole.
Realizing that a particular substance depends on the substantial
form for what it is, without that substantial form being a distinct
constituent of that substance, is the key to understanding Aristotle's
account of the unity of a substance. In so far as the particular sub-
stance is dependent on the (abstract universal) substantial form for
what it is, the substance is posterior to the form, as asserted by
Aristotle at Met. 1029a30-2. In so far as the (enmattered) substan-
tial form is the particular substance itself, like the substance, the
(enmattered) form depends on the (abstract universal) form for what
it is, and hence is posterior to it. But in so far as the (abstract
universal) substantial form exists only as an abstraction from particular
substances, it is existentially dependent on the substances. 34 That it
can be abstracted from them does not mean that it exists buried within
them as a constituent. Abstraction divides what is unified in the way
that sprinkling divides the pool of water into distinct drops. Since the
abstracted form is not a distinct component in the particular sub-
stance, it does not threaten the internal unity of the substance.
Overall, Aristotle's position is that the substantial form is prior to the
particular substance with respect to identity dependence, but posterior
33 Although I agree with Sellars and Kosman, whose positions Lewis quotes
(this volume: 255-6) that matter and form are not individuals in the world, I do
not think that the matter and form of a substance are alternative descriptions of
the substance. Rather, they are entities that are derived by abstraction, where
abstraction involves the division of the substance into abstract entities.
34 Sally Haslanger describes my position as presenting the 'form to be "identity-
dependent" on the sensible substance' (Haslanger, this volume: 158 n. 39). In
fact, according to my position, the substance is identity-dependent on the abstract
(universal) form. Again according to my position, and unlike Haslanger's, there is
no such entity as the enmattered form which is different from, and a component
of, the concrete substance. The enmattered form is the substance itself. Therefore
neither construal of the form (abstract nor enmattered) on my position treats the
form as being identity-dependent on something more primary, the concrete sub-
stance. Hence, Haslanger's arguments against treating the form as dependent on
the substance do not apply to my position as she thinks they do.
I22 Theodore Scaltsas
to particular substances with respect to existential dependence, 35

without being an actual component in the substance, but being
derivable from it by abstraction. 36
Aristotle's rejection of the Platonic conception of a substance, 37 i.e.
as being composed of copresent distinct components, is explicit in
Aristotle's Metaphysics. In Z. 17 he shows that a substance is over and
above the totality of the distinct material parts into which it can be
divided, (1041b25-33); in Z. 16 he argues that the parts of animals are
not distinct components in the substantial organisms ( 104ob5- 15); in
Z. 13 he shows that the universals characterizing a substance cannot
be distinct components in the substance, on pain of an infinite regress
(1038b17-23); and in Z. 6 he shows that the substantial form of a
substance cannot be a distinct component in that substance, on pain of
an infinite regress, (1031b28-30). For Aristotle, the components of a
substance, whether concrete or abstract, emerge when we divide the
substance up, either physically or by abstraction. What this means is
that the entities that emerge from the division of the substance do not
exist in the substance but only potentially. Similarly for the com-
ponents that pre-date the substance and go into the make-up of the
substance by being incorporated into the substantial whole. They exist
before merging into the whole, but not in the whole. It is not that
these components vanish into thin air when incorporated into the
substance; rather, it is that they lose their boundaries and hence their
distinctness. To return to the drop of water example, the boundaries
that distinguish and individuate the drop of water are lost when the
drop becomes a non-distinct component of the water in the glass.
Similarly with the components that constitute a substance: 'all the
parts [of living things and the corresponding parts of the soul] must
exist only potentially, when they are one and continuous by nature-
not by force or even by growing together' (104ob14-15).
Further, just as we can divide the water into drops, which do not
:J:i The universal form is of course not existentially dependent on the particular
substance from which it is abstracted-at least in the case of the organic sub-
stances, the par excellence Aristotelian substances. Rather, it is existentially
dependent on the infinitely many members of the species.
On a related point, I do not understand why Charlotte Witt (this volume: 227,
n. 14) thinks that one cannot have an actual cause which does not exist at that
time. My grandfather is an actual cause of mine, although he does not exist. This
in fact resolves the puzzle that Witt addresses in her contribution, although Witt
does not embrace the position.
36 This position, which I am propounding in my paper, is not one of the ones
entertained/criticized in Sally Haslanger's paper in this volume .
.i 7 See Scaltsas (1990: 583-8).
exist as distinct components in the glass, but only potentially, thus we
can divide a substance up, either physically into concrete components,
or by abstraction into abstract components. The entities that emerge
from these divisions do not exist in the substance, any more than the
drops exist in the glass. According to Aristotle, dividing a substance
up into concrete components produces items that are not present in
the substance; he says: 'we shall define each part, if we define it well,
not without reference to its function .... [the parts of the body] cannot
even exist if severed from the whole; for it is not a finger in any state
that is the finger of a living thing, but the dead finger is a finger only
homonymously' (1035br6-25, my emphasis). The dead finger is not a
substantial component, but has disintegrated into matter (1035b21). It
is a finger in name only, not in the account that states what it is.
According to the homonymy principle, then, separation from a sub-
stantial whole involves the reidentification of the emerging com-
ponents. In conclusion, then, the substantial components are never
distinct; they exist only bound together seamlessly in the substance,
like the water drops in the water.
IV. POTENTIALITY ENTAILS HOMONYMY
What should be emphasized for our present purposes is that poten-

tiality entails homonymy. If x is potentially a y then x is homony-
mously a y. The reason is that although we can say that x is a y, for
example, these bricks are a house, 38 x is not actually a y, but only
potentially so. In that sense, an account of what x actually is would
not be an account of what a y is. Sox is only homonymously a y. The
potential here may be a log, or a lump of bronze, or it may be an
abstract entity like biological matter of low organization, or the
material substratum, for example, the wood in the log which is poten-
tially Hermes, or the body in a human being. (Aristotle does see a
parallel between the relation of the body to the human being it
constitutes, and the bricks to the house they constitute. ) 39 More
generally, anything that constitutes a substance (whether pieces of
concrete matter, bundles of properties, or clusters of opposites) is
38 Met. ro41b6.
39 'The question is why the matter is some individual thing, e.g. why are these
materials a house? Because that which was the essence of a house is present. And
why is this individual thing, or this body in this state, a man? Therefore what we
seek is the cause, i.e. the form, by reasons of which the matter is some definite
thing; and this is the substance of the thing' (Met. 1041"5-9).
homonymously the substance. For the substance to exist in actuality,

these constituents must merge by becoming reidentified in accordance
with the principle of the substantial form.
Suppose it is objected that there are always entities in an actual
substance which are only homonymously that substance; namely that
there is, actually existing in the substance, an aggregate of com-
ponents whose nature is independent of what the substance is. For
example, the first elements-earth, water, fire, and air. My claim is
that Aristotle would then conclude that there would be no substance.
What there would be is the aggregate of components (awpbr;, 104ob8-
IO), having (per impossibile) the substantial form as their accident. If
it were counter-claimed that both the aggregate of components and
the substance are actual and copresent, only at different levels of
internal structure, then these two actualities wduld be related to one
another by a relation other than fulfilment of potentiality. This is
because it cannot be the case that the lower level is actually one thing,
the upper level is actually another thing, and that the upper level is
also the actuality of the lower level. The lower-level thing cannot be
two actualities. So, adding levels of composition between what consti-
tutes and what it constitutes cannot allow for two distinct actualities-
namely what constitutes, and what it constitutes--to comprise one
and the same substance. Rather, two such actualities would have to be
related by the kind of relation that Aristotle dismissed when he
rejected participation, communion, composition, connection, and any
other kind of metaphysical bridge, between what constitutes and what
it constitutes. 40 He rejected them because the totality would be a
related whole of distinct actualities, not a substance.
We have already seen that the unity of a substance is threatened,
not only by distinct concrete components, but equally by distinct
abstract components. To individuate a component by abstracting it
from the substance is to divide the substance. This division has the
same effect as physical division: the entity that emerges-the abstract
entity-,.-is not present in the substance. The abstract entity that
emerges is not an actual component of the substance any more than a
severed arm is an actual arm. Separation from the substance destroys
the identity dependence on what the substantial whole is. Thus, the
abstracted entity emerges with an identity of its own, very much like
the severed arm, or the drop of water that is separated from the water
in the glass. Generally, any process by which we divide the substance
up into distinct components is a process of generating entities that are
independent of the form of the substance; hence, they are entities that
do not exist in the substance. 41
In the water drop example, we saw that the drop's merging with the
water in the glass involved the loss of the boundaries that separated it
into a distinct entity. The 'fate' of the drop was determined by the
nature of the entity it merged into. Similarly with any entity that
becomes a constituent of a substance. Its incorporation into the sub-
stance involves the loss of the 'boundaries' that separate it into a
distinct entity, and its reidentification in accordance with the role it
has in the whole; for example, inanimate matter becomes live flesh or
blood or bone when incorporated into an organism. Without this
reidentification, the incorporated components would remain distinct,
and the substantial whole would be identical to the aggregate of these
components. The case is the same with abstract entities. Socrates is
not an aggregate of distinct properties such as 'white colour', 'small
size', etc. It is not 'white colour' that exists in Socrates. Rather, the
incorporation of 'white colour' into Socrates results in white Socrates.
Properties, and other kinds of abstract entity, do not retain their
distinctness as components of a substance. Otherwise they would
simply be copresent, as elements in an aggregate. But, as we have
seen, substances are not aggregates. The colour white is not copresent
with Socrates; it loses its distinctness and merges into the whole,
resulting in a white whole rather than in Socrates plus the colour
white.
The substantial form dictates the principle of reidentification for the
components that merge into the substantial whole. So the whole,
along with the rest of the components, depend on the form for what
they are. The merging of the various components in terms of the
principle of the form does not alter what the form is, for then there
would have to be a further principle of reidentification for all of them,
and so on ad infinitum. Nor, on the other hand, could the form retain
its distinctness within the substantial whole, without undermining the
unity of the substance. Thus, although the unification of the com-
ponents in terms of the form preserves the principle the form stands
for, it does not preserve the distinctness of the form: the unification of
the various elements into a whole by the form gives the form number.
41 Frank Lewis and Mary Louise Gill develop a notion of concurrent matter
within the substance, which survives potentially in the substance. See Lewis for a
discussion of the contribution that the ingredients in potentiality make to the
substance (this volume: n. 52), and Mary Louise Gill, who attributes a conception
of lower concurrent matter to Aristotle, where matter survives only potentially,
as material properties in the substance.
The unified elements are the enmattered form, the particular sub-
stance, which differs from the universal, abstract form, not in what it
is, but in that it has number (or, concreteness). Having number is
nothing but bein~ an active unity of the sort that substantial forms can
'hold together'. 4 The distinction between the form in abstraction,
which is the principle of reidentification of the substantial components,
and the form actively unifying the whole is captured by Aristotle in his
distinction between being an actuality and being in actuality. 43 In
abstraction, the form is an actuality, qua principle of unification; but
when actively unifying the various substantial elements, the form is an
actuality in actuality, namely the concrete substance itself. 44
V. UNITY AND COMPLEXITY
The metaphysical position that has emerged from this analysis is that a
substance does not contain any distinct components. It makes no
difference whether the components would be concrete or abstract,
particular or universal, substrata or properties. So long as what they
are would not depend on what the whole they joined is, they would
divide that whole into elements and render it an aggregate. Hence,
integration into a substantial whole requires the reidentification of
the merging components. The identity dependence of the components
42 For a detailed analysis of the particularity of Aristotelian substances, and my

response to Kit Fine's puzzle (this volume), see Scaltsas (forthcoming, ch. 7).
43 In Scaltsas (198s: 227, or 1992a: 201-2) I show that Aristotle uses the term
ivipyf.la in the dative to talk of something being in actuality, while in the nomina-
tive the term can refer to an actuality which is not in actuality (e.g. a substantial
form which is not enmattered).
44 Sally Haslanger's account differs from mine in that she does not distinguish
between the form being an actuality from the form being in actuality. According
to her UHB principle of substantial unity, the form's being an actuality is its being
in actuality (present volume: 167, 168). Importantly, for Haslanger it is not the
composite substance, but the form, that is in actuality, i.e. being actively the
form; the composite substance only has the form. This imposes a degree of
distinctness of the form within the concrete substance which is incompatible with
the substance's unity. Further, it shifts subjecthood from the concrete substance to
the form of the substance, inviting the problems that follow from such a displace-
ment. (See Scaltsas (forthcoming: section 7.2).)
On my account, the abstract form is not an activity, but a principle of what a
concrete substance is. The form is an activity only when in actuality, i.e. when
enmattered. But the enmattered form is the substance itself. There is no distinc-
tion between the form in actuality and the concrete substance; hence, no ontolo-
gical gap between the subject that has the form and the subject that actively is the
form.
on what the whole is results in a seamlessly unified whole, rather than
a complex of interrelated but distinct elements.
lt follows that the only genuine parts of a substantial whole are the
ones that are dependent on what the whole is for what they are. The
reason is that, if the parts are identity-dependent on what the whole
is, the aggregate argument paradox does not apply, because a sub-
stance is shown not to be a (mereological) aggregate of parts. For the
aggregate argument to apply, it must be that the aggregate of the
parts survives the dispersal, while the whole does not. But if the parts
are identity-dependent on what the whole is, since the whole does not
survive the dispersal, nor will the parts, and hence, neither will the
aggregate of the parts.
Identity dependence on what the whole is also blocks the trope-
overlap argument. The reason is that for parts to constitute a substan-
tial whole, they must be more than merely copresent. They must be
identity-dependent on what the substantial whole is. Hence, mere
copresence of parts dogs not constitute a substance, although it does
constitute an aggregate of parts. That is why the overlap of substances
does not generate a new substance, while any two aggregates of tropes
make up a single aggregate of tropes.
That a substance is not composed of distinct elements does not
entail nominalism. Aristotle is a realist about universals, but that does
not require universals to be distinct components in substances. In fact,
he argues explicitly against this position in Met. Z. 13, where he shows
that universals do not exist as distinct entities even as components of
substance ( 1038b17-23). But universals can be derived, and defined
independently of the substances they are in, by abstraction. This is the
only kind of separateness that they can enjoy in Aristotle's ontology.
The fact that universals cannot exist either as physically separate
entities, or as distinct components of such entities, does not entail that
they are fictions of the mind. If that were the case, then unrealized
possibilities would also have to be existentially dependent on the
minds that conceive them, which is certainly not the case; unrealized
possibilities would have existed even if there were no minds to think
of them. Abstraction is not imagination. It is a way of singling out
entities which cannot exist as physical things. That they cannot exist as
physical things says something about their nature, not their reality.
Similarity between substances can be explained in a realist vein in
terms of universals, even though universals are not distinct elements
in the substances, nor physically separable; two substances are similar
if the same universal can be abstracted away from each of them. The
same holds for all abstract items of a substance, for example, sub-
stantial form, material substratum, qualities of any kind. None of
them exists as a distinct component in the substance, but similarity

between substances is explained by deriving by abstraction the same
substantial form, or material substratum, or quality, from each of the
substances.
The abstract components of a substance cannot exist in the sub-
stance, any more than what one does can coexist in the actual world
with what one might have done. The abstract components exist poten-
tially in the substance, in so far as they can be derived from it by
abstraction. In that sense, a substance is complex, not because it is
plural, consisting of distinct elements, but because it can be divided
up into distinct components (concrete or abstract ones). The com-
plexity of a substance allows Aristotle to explain similarity and change.
Its unity allows him to explain substancehood.
6
Parts, Compounds, and

Substantial Unity
SALLY HASLANGER
I. INTRODUCTION
In the Metaphysics, Aristotle articulates and defends two profound,

though difficult, theses: (i) sensible substances are composites of mat-
ter and form, and (ii) primary substance is substantial form. I am
concerned in this paper with two questions: first, in what sense is a
sensible substance composite: How are we to understand the relevant
notion of 'part', and the corresponding unity that results in the sensible
individual? Second, how are Aristotle's views about parts and unity
related to his views about priority: Will an understanding of the
composite nature of sensible substances lend insight into the priority
of form?
It is perhaps easiest to enter the tangle of issues motivating these
questions through the problem of unity. Sensible substances, it would
seem, are paradigms of unified wholes; they are each one thing. And
yet, the hylomorphic analysis tells us that sensible substances are
composed of matter and form. If matter and form are distinct, how do
sensible substances achieve their unity? What makes the two things
(matter and form) one? A prominent strategy in answering this
question is to qualify the sense in which matter and form are distinct-
from each other, or from the sensible substance. In the limiting case
one might deny that there is any (real) distinction between them at
all, thus simply eliminating the problem of unity. But short of this it
might seem plausible that the more ontologically dependent matter
and form are on the sensible substance, the easier it is to explain their
© Sally Haslanger 1994

I'd like to thank Alan Code, Terry Irwin, Mary Hannah Jones, Wolfgang Mann,
Charlotte Witt, and Stephen Yablo for valuable discussion on the issues raised in
this paper; special thanks to Mary Louise Gill for helpful editorial comments on
an earlier draft. I read a draft of this paper at the conference 'Aristotle's
Metaphysics', held at Oriel College, Oxford (July 1989), and benefited also from
comments by the participants.
130 Sally Has/anger
unity in the whole. Pursuing this line of thought, the problem of

substantial unity would appear to be solved once we explicate the
dependent status of matter and form.
On the face of it, this broad strategy is in tension with Aristotle's
view that the form of a sensible substance is prior to it (and to its
matter) in every relevant sense. Focusing on the idea that the form is
primary, it no longer looks promising to deflate the compositionality
of the sensible substance, for the form could not plausibly be viewed
as either identical to or dependent upon the composite whole or its
matter. Yet if we affirm that the matter, form, and sensible composite
are really distinct, then we are left again with the problem of substantial
unity. What does Aristotle mean in suggesting that matter and form
are one, 'not like a heap, but like a syllable' (Met. z. 17, rn41h12)?
In what follows I argue against the broad strategy of treating matter
and form as either identical to or dependent upon the sensible sub-
stance; although I grant that this strategy eases some of the tensions in
the hylomorphic analysis, to my mind it does so at too great a cost.
Drawing on Aristotle's broader views on part and whole, I argue
instead that matter and form are distinct proper parts of a sensible
substance, 1 and I offer a programmatic sketch of the unity of such
composites on which their unity is derivative and depends on the unity
of substantial form. It is a consequence of my view that sensible
substances are not paradigms of unity-at least in the sense that they
do not have the very highest degree of unity; sensible substances are
genuinely composite, but by virtue of their form they are highly
unified composites. Broadly speaking, my project is to sketch an
account of substantial unity which locates sensible substances within
Aristotle's hierarchical ordering of things that are, in particular by
offering an account that emphasizes the priority of form.
In discussing substantial unity, it is natural to want first a fairly clear
idea of what the matter and form are that combine in this unity; yet it
is notoriously difficult to sort out Aristotle's views on these com-
ponents. My aim here will be to explain the unity of a particular
sensible substance such as Socrates. 2 In inquiring into the nature of
1 I am assuming in line with standard terminology that x is a proper part of y iff
x is a part of y and x is not identical toy; x is an improper part of y iff xis a part
of y and x is identical to y. We may speak also of a 'proper' or 'genuine'
composite as being one composed of proper parts.
2 At different stages of the dialectic in the Metaphysics, Aristotle discusses
different candidates for the role of substance, so it is sometimes difficult to keep
straight exactly whose unity is under discussion. In some passages Aristotle is
clearly concerned about the unity of form or essence and the kind of definition
appropriate to form (Met. z. 12); at other points he is concerned about the unity
of sensible substances or composites, and the kind of definition that may be
Parts, Compounds, and Substantial Unity 131
this unity, I will assume, without accounting for, the unity of substantial
form. My own view is that the constituent form of a sensible composite
is the form common to the species rather than an individual form
peculiar to a particular substance. This is not the place to defend this
preference; so I have attempted to articulate the arguments and hypo-
theses under consideration in a way that is neutral on this point; at
least I intend to allow that in addition to the form common to a given
species, there are individual forms that function as the formal con-
stituents of concrete members of the species. So I shall use the term
'substantial form' for whatever functions as the formal component of
the substantial composite, be it individual or general. Because I do
not explicate the relevant notions of matter and form, there are many
puzzles I slide over without discussion ;3 however, my hope is that by
developing a schematic account of substantial unity we will gain a
better understanding of the constraints on both matter and form that
arise from their respective roles in a substantial composite.
II. INTERPRETATIVE BACKGROUND
Although in the central books of the Metaphysics Aristotle himself

consistently speaks of sensible substances as compounds or composites
of matter and form, commentators have hesitated to read this as
metaphysically revealing. Some commentators, for example, suggest
that we should not interpret Aristotle's use of compositional terms in
describing sensible substance to indicate that matter and form are, in
any 'ordinary' sense, parts of a compound. For example, J. L. Ackrill
comments:
'Constituent' is no doubt an unhappy word [to describe matter and form in
relation to sensible substance]: it is because matter and form are not, in the
ordinary sense, constituents that no question arises as to how they combine
into a unity. We might speak of the material 'aspect'. (Ackrill 1972-3)
appropriate to them (Met. z. 4, IO, II). But even further questions arise, since in
inquiring into the unity of sensible substances, one might be concerned with the
unity of a particular composite substance such as Socrates, or one might be
concerned with the unity of a species, i.e. the unity of the matter and form of a
composite substance 'taken generally' (Met. Z. IO, rn35h27-30 and Z. r r, rn37"6-
7). An account of substantial unity would benefit greatly from a more detailed
interpretation of this dialectic than I will offer here.
3 One important question I will largely ignore concerns the relation between
the constitutive or proximate matter of a substance and the matter from which the
composite is generated. For extended discussion of this issue, see Gill (r989),
Furth (1988), and Lewis (this volume).
132 Sally Has/anger
In line with Ackrill, many commentators have preferred to avoid

talk of matter and form as 'parts' or 'constituents' in favour of a
terminology of material and formal 'aspects' of a substance.
This resistance to regarding matter and form as 'parts' of a substance
is often supported by reference to Met. Z. 17, where Aristotle main-
tains that form (at least) is not an element of the composite substance:
But it would seem that this [namely the substantial form] is something, and
not an element, and that it is the cause which makes this thing flesh and that a
syllable ... And this is the substance of each thing (for this is the primary
cause of its being). (1041b25-7)4
But if we allow that there are parts of a substance which are not
elements, this text does not rule out treating both matter and form as
genuine parts. And this allowance should not be controversial, for in
the same passage Aristotle characterizes an element as 'that into
which a thing is divided and which is present in it as matter; for
example, a and b are the elements of the syllable' ( 1041°31-3). So at
least in this context, Aristotle is considering an element of something
as a material part of it. Assuming that there are other kinds of part
besides material parts, the claim that form is not an element of
sensible substance still permits form to be a part of a different kind.
We will return to consider what kind of part form might be, but we
may note here that Aristotle's views on part and whole are complex,
and he is working with more than one part-whole relation. As he
indicates in a number of places, both 'part' and 'one' (or 'unity') arc
used in several ways. 5 It may very well be that the sense in which
matter and form are parts of the composite is a sense that docs not
accord neatly with our 'ordinary' notion of part; however, our own
conception of part is sufficiently ill-formulated that we have more to
gain both in interpretative results, and in philosophical understanding
of the relation of part and whole, by following his suggestion that
matter and form are parts, than by beginning with a narrower regi-
mentation of the term.
There are other commentators, however, who propose that in the
final analysis of sensible substance there is no distinction between the
matter, the form, and the whole; thus, in genuine substances we
4 Here and following (unless otherwise indicated) I quote the translation by
W. D. Ross, revised by Barnes (1984), based on the Ross text (1924).

5 Concerning 'part', he lists various senses in Met. A. 25, and says explicitly at
Met. Z. IO, '"Part" is used in several senses' ( 10341>33). Concerning 'one' or
'unity', he lists a broad range of senses in Met. L1. 6, and Met. I. 1; and often he
states explicitly that there are different senses of 'one', e.g. Met. r. 2, 1004"22,
rno5"7; z. 4, 1030"8-12; /. l, 1052"15. Also, Phys. I. 2, 185h5ff.
should take matter and form to be identical. 6 On this view, if it is
meaningful to speak of matter and form as parts of a substance at all,
they are not proper parts but are at best improper parts. The most
important text to support this view is in Met. H. 6. 7 There, in the
process of discussing the unity of genus and differentia in a definition,
Aristotle explains definitional unity in light of the unity of matter and
form:
But, as has been said, the proximate matter and form are one and the same
thing, the one potentially and the other actually. (1045°17-19)
Interpretation of this passage is difficult, and is further complicated by
the role it plays in the account of definitional unity. Yet it serves as a
crucial text for explicating the kind of unity Aristotle envisaged be-
tween matter and form.
In considering the passage quoted from H. 6, it is important to note
that it prevents us from taking matter and form to be proper parts of
the substance only if we have reason to think that in order for X and
Y to be 'one and the same' they must be identical. Although there are
passages were Aristotle uses the terms 'one' or 'one and the same' to
characterize identicals, 8 there is a systematic use of these terms that
concerns the unity of distinct parts in a composite whole. As mentioned
above, Aristotle clearly allows that there are different ways for things
to be one: things may be one in form or essence; they may be one by
virtue of a common genus, or a common species, or by having the
same matter; they may be numerically one, accidentally one, or one
by nature (see Met. L1. 6). Unfortunately, when Aristotle speaks of
something being 'one', or of two things being 'one', he does not
always distinguish which sense of 'one' he intends. In a number of
passages it is a substantive interpretative question whether in saying
that X and Y are 'one', he means that X and Y are identical, or that X
and Y are together parts of a composite whole. Therefore, we should
hesitate to draw any conclusion from the passage in H. 6 before
6 See e.g. Kosman (1984: esp. 141-5; 198T 378). T. Irwin also sometimes
seems to endorse this claim. At Irwin ( 1988: 243) he argues for the conclusion
that 'the form, the formal compound, and the proximate matter are identical'
(my italics). However, a few pages later he also claims that, 'The descriptions
"matter", "form", and "compound" ... identify different features of the particular
substance' (my italics) (1988: 251). The former claim endorses the view that
proximate matter and form are identical; the latter claim seems to fit better with
the 'aspects' view mentioned above.
7 See also De An. II. 1, 412°4-9.
8 See e.g. Met. z. 14, ro39°1 and Met. z. 17, rn41•14-19, where being one
with oneself is at issue.
134 Sally Haslanger
exploring the different ways that matter and form might be 'one'. To
this let us now turn.
III. DIVISION AND COMPOSITION
There are a number of ways one might begin to classify different

senses in which something is a part of something. We may assume that
corresponding to different kinds of part there are different kinds of
whole or unity. If we consider the wholes, then one way to distinguish
different kinds of part is in terms of different principles of division
which, when applied to the whole, yield parts in accordance with that
principle. For example, the genus animal, if divided according to
species, yields as parts: horse, human being, monkey, etc.; if divided
according to individual, it yields as parts: Bucephalus, Socrates, etc.
Horse is a specific part of the genus animal; whereas Bucephalus is an
individual part of the genus animal (as well as an individual part of the
species horse). We may note that the genus animal has no (proper)
generic parts; if it is divided by genus we find that it is itself one
genus. 9 Roughly, if a composite is 'one' with respect to a principle
of division, then applying that principle to the composite yields no
parts. 10 ·
Alternatively, if we consider parts, we may distinguish different
principles of compounding or unification in virtue of which parts
qualify as members of a whole. For example, Bucephalus, Socrates,
Lassie, are unified in one genus because their genus is one, but they
are not unified in one species because their species are many. The
9 For a valuable discussion of principles of division and unity in connection
with Aristotle's notion of an individual, see Frede (1978: esp. 50-63).

HJ It remains obscure what 'applying a principle of division' amounts to. For
example, if there are some principles of division which do not divide something
because they are wholly inappropriate to that kind of thing, then they will yield
no parts; but in such a case it would be misleading to say that the composite is one
in that respect: e.g. a word may be divided according to vowels and consonants,
or letters (Met. L1. 6, rn16b18-24) but such a division would not apply at all to
e.g. animals. Even though division by letters yields no parts of an animal, surely
we should not say that animals are one letter. For this reason, the general
condition of having no parts with respect to a principle of division may serve as a
necessary condition on being one with respect to that principle, but it is not a
sufficient condition. Nevertheless, sometimes Aristotle seems to assert the con-
dition as a sufficient condition as well (e.g. Met. L1. 6, 1016b3-7). One option for
avoiding the difficulty would be to specify what it is for a principle of division to
properly apply to a composite, and then to restrict the condition to those principles
of division which properly apply to the composite in question.
parts of a log are unified in virtue of a natural continuity that results in
their having one movement (Met. A. 6, 10163 5-18). So, roughly, a set
of parts may be called 'one' with respect to a principle of unity just in
case applying the principle to the parts yields a kind of unified whole.
This suggests that we may gain insight into the relation between a
composite and its parts by considering, on one hand, the principles of
division by which the composite is divided into its parts, and, on the
other hand, by considering the principles of unity by which the parts
together constitute a whole. In exploring such principles we should be
aware that principles of division and unity form natural pairs: one
dividing the composite into parts, the other combining those parts into
the composite; in effect, one will function as the inverse of the other. 11
Thus, although to say that a composite is 'one', and that a number of
parts are 'one', we employ different senses of 'being one'; nevertheless
these senses reflect a correlation between principles of unity and
principles of division.
Since we are allowing for different kinds of part and unity, we must
take the unity of a composite to be relative to the principle of division
under consideration; likewise, we must take the unity of a group of
individuals to be relative to the principle of unity in question (Met. A.
15, 1021 3 9-13). (Recall that Socrates and Bucephalus are one in
genus but not one in species.) In general, however, a composite will
be 'one' with respect to the principle of division that divides things ac-
cording to the kind it is (Met. I. I, 1052b16-18; Met. A. 6, 10163 32-5).
As a result, in saying what something is we will indicate a principle
that does not divide it, and thus a principle by virtue of which it is
one. However, simply indicating that a composite is indivisible with
respect to a principle does not tell us how parts divided out by virtue
of other principles combine in that composite. Moreover, if a com-
posite is divisible in different ways, we may find ourselves formulating
principles of unity that combine different kinds of part. 12 In focusing
on the unity of matter and form as parts of a substance, our attention
11 My general observations about unity and division, here and throughout this
essay, owe much to K. Fine (unpub. paper). There Fine demonstrates how an
operator approach to part and whole can serve as the basis for a systematic
treatment of different kinds of part. This idea is useful in thinking about Aristotle's
views on unity, and applications of it have been suggested in F. A. Lewis (1985:
esp. sect. 1).
12 It is plausible to maintain that part-whole relations are transitive. If so, then
there will by hybrid part-whole relations which combine various sorts of part. For
example, if Socrates' flesh is a (material) part of Socrates, and if Socrates is an
(individual) part of the species Human Being, then Socrates' flesh will be a hybrid
(material cum individual) part of the species Human Being. See K. Fine (unpub.
paper: 2-3).
Sally Has/anger
will be primarily on principles of unity; however, this inquiry will lead

us to a better understanding of the idea that a substance is indivisible,
i.e. it is 'one', with respect to the principle indicating what it is.
IV. PRINCIPLES OF UNITY
In his discussion of unity in Metaphysics L1. 6, Aristotle outlines a

number of ways in which we can say that a group of things are one, or
better, are unified. The relevant contrast is between things that
together form a unified whole and things that, using Aristotle's term,
are (at best) a mere heap. For example, some things constitute a unity
because they are continuous and their movement is one; others because
they are accidents of a single subject; still others because their matter,
their form, or their genus is one. Considering the different kinds of
unity he says:
While most things, then, are called one from either doing or having or being
affected or being related to some other thing that is one, the things called one
in the primary way are those whose substance is one ... (Met. A. 6, rn16b7-9)
It is clear, both from his examples and comments, that Aristotle's
preferred way of formulating a principle of unity is to isolate an
individual such that the relation(s) of items to that individual constitute
the ground of their unity. 13 We might call such principles integration
principles, seeing the privileged individual as 'integrating' the parts
into a unified whole.
In considering integration principles, we should note that in some
cases the privileged item will be a member of the unity, and sometimes
13 I use the term 'individual' loosely here, allowing that the individual in
question may be particular or general. The only constraint we need add at this
point is that the privileged individual must itself be 'one', i.e. more than a mere
heap. Below I discuss difficulties which arise if the unity of the individual in
question itself depends upon the principle in which it serves.
It is also important to note that I have characterized an integration principle as
a way of formulating a principle of unity; I do so in order to allow that in some
cases an integration principle is equivalent to a principle which makes no reference
to a privileged individual, so this reference may be eliminable. Aristotle seems
also to allow that there are things which are unified merely b~ being connected,
continuous, or in contact (e.g. the Iliad). See rn16•5-15, 1030 8-12, rn45"7-13.
However, he suggests that such unities are inferior to integrated unities. It would
be a valuable project to inquire under what conditions reference to a privileged
individual is eliminable, and whether there is a substantive contrast between
integrated and connected unities; however, it is a project which goes beyond the
scope of this paper.
not. Thus, we might specify the unity of a family with reference to an

individual (say, parent) such that one is a part of the family just in
case one is that individual, or is the spouse of that individual, or is an
offspring of that individual. Similarly, one might specify the unity of a
train with reference to the locomotive that is itself part of the train.
Let us call principles of integration that rely upon a privileged indi-
vidual which is itself a member of the unity intrinsic integration prin-
ciples. In contrast, extrinsic integration principles rely upon an outside
individual. For example, one might specify the unity of a group of
siblings extrinsically by reference to the parents who are not them-
selves part of the (sibling-) unity.
At this point the idea of an integration principle is still very vague.
In order to use the general idea of integration to inquire further into
the basis for substantial unity, it would help to be clearer about how
such principles account for unity, and what sorts of constraints we
might place on a choice of privileged individual. To begin, it is
important to clarify what questions integration principles can help us
answer. Suppose we ask:
In virtue of what are matter and form unified in the composite
substance Socrates?
We may hope that our answer will enable us to answer each of the
following questions:
(a) Under what conditions are matter and form unified m a
composite substance?
(b) Under what conditions are matter and form unified m a
human substance?
(c) Under what conditions are matter and form unified m
Socrates?
Although each of these questions appears to raise the problem of
unity at different levels of generalization, we must be careful to separ-
ate the issue of unity from the issues of individuation and definition.
The issue of unity concerns the basis for considering a plurality of
things (namely, parts) one composite whole as opposed to a heap. In
considering the substantial unity of matter and form (as in (a)), we
want to know why we should treat them together as composing one
thing, a sensible substance: What are the conditions on arbitrary
matter m and form f, such that m and f combine in a sensible
composite? To go on to ask (as in (b)) how matter and form combine
to form a specifically human composite, the question shifts from the
issue of unity to focus on the kind of composite produced. Because
human beings do not differ from other kinds of (organic) sensible
substance in the kind of unity their matter and form exhibit, but in
Sally Has/anger
their characteristic form, we should expect that a full answer to (b)

will require a definition of human form. But in addressing simply the
issue of unity, we need not provide this definition.
If we go on further to consider the unity of matter and form in
Socrates (as in (c) ), the question raises difficult issues of individuation.
In order to determine what it is in virtue of which Socrates is distinct
from other members of his kind, we may seek those conditions on any
matter and form such that they result in the particular individual
Socrates (as opposed to, say, Coriscus). But a principle of unity may
specify the basis on which we can say that a plurality composes one
thing, without providing the basis for distinguishing that thing from
others of its kind. 14 To answer the question of individuation, we do
not characterize Socrates as exhibiting a special Socratic kind of unity.
Socrates and Coriscus are unified in the same way; rather, it is their
constituent matter (and depending on one's view, perhaps also their
individual form) which differ. The question of how to individuate
Socrates' matter and form may be pressing, but again, we need not
solve the problem of individuation in order to solve the problem of
unity. 15
Although an account of substantial unity will not enable us to define
human form, or to individuate the particular matter and form that
compose Socrates, we may still hope that an answer to (a) will provide
for a kind of answer to (b) and (c). In characterizing broad kinds of
unity by means of integration principles we should look to specify very
general conditions on the privileged individual and the parts' relations
to it. Then it is reasonable to account for the more specific unity of
particular composites or kinds of composites by demonstrating how
the general conditions are met in individual cases. For example, we
14 Aristotle's preference for characterizing unity in terms of integration princi-

ples can contribute to obscuring this point, since often by naming a particular
privileged individual which unifies a particular composite, we thereby individuate
the composite. (Think of the example of families.) However, it is clear that unity
and individuation can pull apart: e.g. Aristotle's comments on mental acts and the
objects of those acts provide a good example. Considering an act/object com-
posite, it will be individuated by the act, but unified by the object, since the
object is prior in account to the act. (I will discuss the importance of priority
below.) See e.g. Cat. VII; Met. I. ro53"3r-5, /. 6, w56"32-ro57''16, and A. 9; f.
Met. LJ. 15, 1021"27--"3.
15 Moreover, Aristotle (plausibly) thought that it is not possible to state general
conditions for the particular Socratic unity which apply to arbitrary matter and
form. Answers to (c) which do not take the individuation of Socrates' matter (and
form?) as given, will plausibly make reference to his particular causal history, e.g.
his mother's and father's contributions, and so will fail to be general. For good
discussions of this point, see White (1986) and Cohen (1984).
may broadly contrast generic unities from numerical unities by virtue

of the fact that generic unities rely upon a common genus being
predicated of the parts, whereas a numerical unity relies upon the
matter (or subject) being common to the parts. 16 We can then explain
the generic unity of Socrates and Bucephalus in the genus Animal by
virtue of the fact that Animal is a genus predicated of both Socrates
and Bucephalus. This explanation does not give us a basis for indi-
viduating Socrates and Bucephalus, though it does explain the basis
on which they are (generically) one. Likewise, if we can formulate a
general integration principle that captures the conditions on substantial
unity, then we can hope to explicate how it is that Socrates' matter
and form are unified in virtue of satisfying that principle.
Although so far I have mentioned no constraints on what might
count as the privileged individual that serves as a basis for an inte-
gration, it is clear that some individuals are more appropriate to that
role than others. At a number of points Aristotle suggests that the
unity of items to which such integration principles apply is derivative,
for their unity as a genuine whole depends on the antecedent unity of
the privileged individual and on the parts' relations to it (e.g. Met. r.
2). This is in keeping with the idea that unity comes in different
degrees: things that are integrated by an individual whose own unity is
derivative, will be unified only in a relatively extended sense of the
term.
We should beware, however, that it is possible to formulate inte-
gration principles that do not provide an acceptable basis for unity. To
see this it is important to introduce the idea of one thing's being prior
in account to another. 17 Roughly, something X is prior in account to
Y, just in case it is not possible to say what Y is, without making
reference to X, but it is possible to say what X is without making
16 On numerical unity, see Met. A. 6, ror6"r8-23, 1016"32-4; and Phys. I. 7.

19oa15. These passages indicate that we should be careful not to assume that x
and y are numerically identical iff x = y. For example, an accident is numerically
one with the subject in which it inheres, but the accident is not identical to the
subject. It may be helpful to think of numerical unity in terms of coincidence in
the actual world, or given just how things actually are-leaving aside modal
concerns (perhaps adding qualification with respect to time).
17 There are clearly different senses of priority at work in Aristotle (See e.g.
Met. A. n). Here, and in what follows, I am only addressing the issue of priority
in account, although for ease of exposition I will drop the qualification 'in
account' and speak simply of things being 'prior' and 'posterior'. For our purposes
it is particularly important to note that priority in account may not be equivalent
to priority in existence. It may be that two things are comparable with respect to
existence (neither can exist without the other), even if one or the other is prior in
account.
Sally Has/anger
reference to Y. This characterization is vague in several respects,
notably in so far as it remains an open question what is required to
'say what X is'. However, working with this rough idea of priority we
can begin to draw a contrast between different kinds of parts in a
composite (e.g. Met. Z. 10). A kind of part is prior to the composite
just in case it is not possible to say of the composite, what it is,
without making reference to that kind of part, but not vice versa. In
contrast, a kind of part is a posterior or derived part of a composite
just in case it is possible to say of the composite, what it is, without
making reference to such parts, and not vice versa. A composite is
prior to its posterior parts. Some parts of a composite may be neither
prior nor posterior. Let us call those kinds of part on which an
account of the composite depends (remaining neutral with respect to
whether an account of the parts also depends on it) the composite's
constitutive parts, ·and let us say that a composite is an aggregate of its
constitutive parts. 18
An example here will help. Consider to start, an abstract particular,
Socrates' pallor, that Socrates and only Socrates has. It is plausible to
say that Socrates' pallor is a part of Socrates, though a derived or
posterior part. In order to say what Socrates' pallor is, we must make
reference to Socrates, but we need not make reference to his pallor to
say what Socrates is. In contrast, consider the pair, Socrates and
Coriscus. Socreates and Coriscus are each constitutive parts of this
pair, since to say what the pair is we must make reference to them.
Because it is also the case that we need not make reference to the pair
in order to say of each of them, what they are, the pair consisting of
Socrates and Coriscus is an aggregate of the prior parts, Socrates and
Coriscus. Note, however, that Socrates himself is not an aggregate in
virtue of having his pallor as a part. There may be a related aggregate
consisting of Socrates plus Socrates' pallor of which Socrates' pallor is
a constitutive part; but this aggregate should not be confused with
Socrates, since Socrates' pallor is only a derived and not a constitutive
part of Socrates.
We can now see that principles of unity and division may be of
different sorts. Principles of division may divide a composite into
constitutive or posterior parts; likewise principles of unity may combine
constitutive or posterior parts into a composite. In seeking integration
18 Note that I am introducing the term 'aggregate' as a technical term to refer
to composites of constitutive parts. My use of the term differs from that of T.
Scaltsas (this volume: II r); Scaltsas defines an aggregate as 'a collection of
entities which stand unrelated to one another'. As I use the term, in general the
parts of an aggregate will be related to one another, and aggregates can very well
be genuine unities.
Parts, Compounds, and Substantial Unity
principles for a composite, it is important to remain sensitive to the

differences. In order to say of an aggregate, what it is, we must make
reference to the parts (or kinds of part) that constitute it; an aggregate
just is a unity of constitutive parts. Thus, principles that demonstrate
the unity of the constitutive parts will offer an account of what that
aggregate is. In contrast, principles of unity that combine posterior
parts, or that demonstrate the unity between a composite and a
posterior part of it, will not provide an account of the composite. 19
These observations indicate a basis for restricting the choice of
privileged individual in certain kinds of integration principles. If
an integration principle grounds the unity of constitutive parts in an
aggregate, then if the privileged individual in that principle is itself an
aggregate of that kind, its unity will depend on the principle in which
it serves. This will result in an unwanted circularity.
For example, suppose that we formulate a principle for specific
unity among individuals as follows:
(SpU) x and y are specifically unified iff there 1is a species s such
that xis ins, and y is ins.
For example, Socrates and Coriscus are specifically one because they
are in the species Human Being, which serves as the privileged indi-
vidual for their specific unity. But let us suppose for the moment that
what a species is, is just a special kind of aggregate of individuals. 20
Then, in order to explain what is it for something to be a species, we
will need to rely upon a principle that unifies its parts; after all, we are
supposing that what the species is, is a group of specifically unified
things. Which principle unifies the parts of a species? The principle
(SpU). Thus, allowing that there are no empty species and that nothing
is a member of two species, we can define the species aggregate using
the notion of specific unity:
(S) s is a species iff 3x(x is in s & 'v'y(y is in s iff x and y are
specifically unified)).
19 We should not conclude, however, that principles of unity provide proper

definitions of aggregates. It is plausible to read Aristotle as placing demands on
proper definitions which would prevent aggregates from having proper definitions
(see Met. Z. 4-6), at least those aggregates which have matter as a constitutive
part (1036"1-8, 1039b27-31, 1040"2-7, 1040"27-b4). This would allow, however,
that there are some composites with proper definitions, since not all composites
are aggregates.
20 It is very difficult to sort out Aristotle's views on species in the Metaphysics,
especially in relation to his views in the Categories. The suggestion that a species
is an aggregate of individuals is not intended as a likely interpretation, but is used
merely for illustration.
Sally Has/anger
But clearly this is a problem, for we rely on the notion of a species

to account for specific unity, but also rely on specific unity to provide
an account of a species; so given that accounts should not be circular
in this way, one or the other must be inadequate. One might avoid the
difficulty either by rejecting the principle of specific unity (SpU), or
by claiming that the individual members of a species are not consti-
tutive parts, i.e. that the species is not (strictly speaking) an aggregate
of individuals. This latter option is plausibly Aristotle's view; in order
to say, for example, what the species Horse is, we need not make
reference to individual horses, even if they are, in some sense, parts
of it.
However, it is important to note that in providing a principle of
unity that applies to posterior parts, we do not run into the same
difficulty. Posterior parts may be unified in virtue of their relations to
the composite ·of which they are parts, even if constitutive parts
cannot. In saying of a composite, what it is, we need not mention its
posterior parts; its unity is prior to the unity of those parts. For
example, we may account for the unity of Socrates' pallor and his ill-
health in the composite Socrates by virtue of their both being accidents
of him. In such a case Socrates serves both as the composite in which
the parts are unified and as the privileged individual that grounds the
unity. But because Socrates iii not an aggregate of his pallor and
ill-health, our account of what Socrates is need not mention these
accidents; they are posterior. This avoids the problem discussed above.
So, in specifying integration principles that ground the unity of a
composite, we should be sensitive to the kinds of part and kinds of
composite at issue. If we are concerned to account for the way in
which a group of parts constitutes a kind of composite, then we should
not use that very kind of composite as privileged individual in an
integration principle, on pain of circularity in our account. However,
if we are concerned only with the unity of derived or posterior parts in
a composite, then because the account of the composite will not
depend ,on this unity, we can avoid the circularity.
V. UNITY AND IDENTITY
We are now in a better position to compare unity and identity. It

clearly does not follow from the fact that A and B are unified, or are
one, that A and B are identical. Principles of unity do not, in general,
establish identity between the parts that compose the unity; nor do
they establish identity between an individual part and the whole that
contains it. Socrates . and Coriscus, although one in species, are not
identical with each other; although parent and child are unified in a
family, neither is identical with the family as a whole. Although there
may be some instances in which we might want to speak of a principle
unifying A and B, where A and B are identical, this is the exception
rather than the rule. If A and B are identical, then the principle of
their unity is simply identity. If desired, such unity could be stated in
terms of an integration principle: A and B are exactly one iff there is
something x, such that A= x and B = x.
Aristotle suggests in Z. 17 that in cases where a single object
is 'given', questions concerning its unity with itself can at best be
answered by repeating platitudes common to all cases, such as 'a thing
is itself' or 'each thing is inseparable from itself (1041a10-25). He is
notably dismissive of such questions, suggesting that the more signifi-
cant issue for inquiry is the unity of parts in a complex whole, for
example, the unity resulting when one thing is predicated of another
(1041 3 21-5). Thus, he proposes that if one (meaningfully) asks 'why
is X, Y?', we should not read the question as asking, 'why is X
identical to Y?'; rather, we should favour reading the X and Y as
distinct, typically one predicated of the other. Correspondingly, we
should seek answers to such 'why' questions not in platitudes about
identity, but in investigating what is the cause of the parts together
constituting the whole ( 1041a25-b2). The answer will typically cite the
form that is present in the matter (1041b6-7). If we fail to note that
our inquiry concerns the unity of parts in a whole, i.e. if we mistake
the issue of specifying the principles of unity with the issue of ex-
plaining the identity, we will be led astray:
The object of inquiry is most overlooked where one term is not expressly
predicated of another ... because we do not distinguish and do not say
definitely, 'why do these parts form this whole?' But we must distinguish the
elements before we begin to inquire; if not, it is not clear whether the inquiry
is significant or unmeaning. (1041a33-b3)
Although Aristotle's comments in Z. 17 are difficult, and cannot be
captured simply by contrasting identities and unities, his comments
offer us a guide to determining when we should deny an identity. This
will be important in the arguments that follow, for we shall want to
determine whether matter and form are unified in a substantial indi-
vidual by each being identical to it (and so to each other), or by virtue
of a principle that unifies distinct things in a proper composite. I
propose that we take the governing idea in Aristotle's contrast between
substantive and non-substantive questions to be that identity facts are
basic; roughly, there is no further fact in virtue of which an identity
holds. So, to put the suggestion boldly, if we can determine for X and
144 Sally Has/anger
Y, that there is no such basic relation that holds between them, i.e. if
their relationship is unavoidably mediated, then we should conclude
that X and Y are distinct (though possibly unified). Of course this is
not to say that only identities are basic-there may be a range of basic
or primitive facts-but at least identity is basic.
To see this point it is helpful to distinguish claims that are epistemo-
logically trivial from those that are metaphysically trivial. For example,
bracketing worries about vacuous singular terms, claims of the form: x
= x, are epistemologically trivial; they only offer information about x
that anyone understanding the statement will already know. In con-
trast, claims of the form: x = y are not epistemologically trivial; they
often provide new information. However, supposing that the identity
statements in question are genuine, i.e. that they are expressed in
proper logical form, it is reasonable to claim that there is a single kind
of metaphysical fact in virtue of which both sorts of identity statements
are true. 21 Assuming that a = b is true, it is in virtue of the same fact
that makes a = a true, namely the identity of the object in question
with itself. This identity fact, we might want to say, is metaphysically
basic; it is so fundamental that we cannot hope to explain it in light of
some further fact. 22 Following this we might see any genuine true
identity statement as metaphysically trivial because its truth rests on a
metaphysically basic identity fact.
In Z. 17, Aristotle suggests that in the context of metaphysical
inquiry we can assume that the facts and the objects we are trying to
understand are 'given'; in other words, we may assume that the
statements we assert are true, and turn to the metaphysical ground of
the truths in question. We want to know: what is it in virtue of which
they are true? If such a question concerns a statement that expresses a
metaphysically basic fact, then there will be no substantive meta-
physical reply; our answers will not call upon facts beyond the fact
expressed to ground the truth of the statement in question. 23 Assuming
21 e.g. on a broadly Fregean account, the fact in virtue of which a = a is true is
the same fact in virtue of which a = b is true, even though they express different
thoughts or propositions. On a broadly Russellian account, the two will express
the same proposition, but we must beware that many apparent identity statements
(e.g. those employing definite descriptions) are not expressed in proper logical
form, and so don't (strictly speaking) express simple identities.
22 Of course, it is difficult to define what counts as a 'basic' fact. I will not try to
do so here, nor will I attempt to explicate the 'in virtue of relation which I rely
upon. In addition to the condition I propose (that a basic fact not hold in virtue of
some further fact), one might also require that the fact be 'common to all things'
(Met. z. 17, 1041"19), or additionally that the fact in question play some special
epistemological role (consider Met. r. 4).
23 I am assuming that if in answer to the question, why is Socrates = Socrates,
we answer, 'each thing is identical to itself', the generalization does not provide
that we are not among sceptics who may question even the trivial
expression of such basic facts, 24 our best course in response is to
repeat the platitudes that somehow reveal the statement in question as
expressing a basic fact.
Aristotle explicitly endorses this response to questions of identity,
and contrasts it to the kind of substantive response appropriate when
considering questions of unity. Questions concerning unity require
different sorts of substantive answers depending on the kind of com-
posite in question, the kind of cause that is sought, etc. Facts about
unity, in contrast to facts about identity, are not metaphysically basic:
there are (typically) further facts that serve as the cause or the ground
of unity. For example, Socrates and Bucephalus are generically one in
virtue of each having Animal as their genus; the just and the artistic
are accidentally one in virtue of each being accidents of Socrates.
Here the parts in question are unified in virtue of more basic facts,
namely the relations that hold between an object and its genus, and
between a subject and its accidents, respectively. We should thus
expect that our inquiry into unity will lead us to explore the more
basic predicative relations in virtue of which items are jointly parts of
a whole. Although we may want to continue to count the identity of
something with itself as a limiting case of unity, the standard cases of
unity concern distinct parts which are somehow 'one', that is, they are
one in virtue of further predicative facts.
I mentioned above that we could take the fact that identity is basic
as a guide to determining what identities hold. Can we now make
something more of this suggestion? The broad strategy I suggested
was to determine for candidates X and Y, whether there is some basic
relation between X and Y; if not, then because identity is basic, X
must be distinct from Y. Although suggestive, this strategy is too
broad to be practical: how do we determine for given X and Y,
whether or not there is such a basic relation between them? Might it
not be the case that any two distinct things bear some basic relation to
each other? Drawing on Aristotle's contrast between substantive issues
of unity and trivial issues of identity, I propose that if X and Y are
one, but their unity is not basic, then X and Y are distinct. Another
way to put the proposal is this: if X is Y, but X's being Y is not basic,
then X and Y are distinct.
the 'ground' for the singular identity, i.e. the generalization is not a fact in virtue
of which the singular identity holds. Instead, the generalization serves to highlight
the structure of the fact in such a way that it becomes clear to us that it is basic.
24 This is to contrast the current inquiry to the discussion in Met. r. 4-8, which
('"{)Tif'Prnc;;:. ('\ltr ~r>rPnt~n1"i::-> r.f h..-.c-;.-. +..-u+J...,.. ;_....,. +t...,.,. 1.:,_L._ .-..C - - ..J; __ t - - - __ ,,_• - 1 •,•
Sally Has/anger
VI. MATTER, FORM, AND IDENTITY
I indicated above that there are two relatively common views con-
cerning the relationship between matter and form in a sensible com-
posite. The first view claims that matter and form are identical; the
second claims that matter and form are distinct 'aspects' or 'features'
of the sensible substance. Each of these proposals allows for different
interpretations, but it is important to keep them clearly distinguished.
On the view that matter and form are identical, the terms 'the matter'
and 'the form' refer to exactly one thing, although one may allow that
the referent has both material and formal properties. Casting the
point in terms of unity, one might say that the· concrete substance is
the privileged individual, and matter and form are unified by each
being identical to it, thus to each other. On the alternative view that
the matter and form are distinct aspects or features of a substance,
the terms do not co-refer to the concrete substance, they refer to the
distinct aspects or features that the substance has. I will discuss the
proposal that matter and form are identical in this section; I will
discuss the proposal that they are distinct aspects of sensible substance
in the next.
It is important to begin by clarifying the suggestion that matter and
form are identical. There is only one relation of identity. It is a
relation that everything necessarily bears to itself and to no other.
There has been considerable controversy over whether Aristotle ever
explicitly formulated 'our' notion of identity, particularly whether he
accepted Leibniz's Law as a condition on identity:
(LL) If x is identical to y, then whatever holds of x holds of y.
I think the evidence is overwhelming that Aristotle acknowledged this
principle, and relied on it in his argumentation. 25 He certainly accepted
that there is an intimate way of being 'one and the same, as you are
one and the same with yourself' (Met. Z. 14, 1039b1); and he also
accepted that it is not possible for something poth to hold true of an
individual and to fail to hold true of it (e.g. int. 7, 17b27-9). Given
this, it would be difficult to allow that although x and y are one and
the same, as you are one and the same as yourself, something holds
true of x and not of y. 26
25 On this issue, see e.g. Top. 152b25-8, Int. 17b26-31; cf. Soph. Ref. 179•37-
9, 169b3-6; Phys. 202b14-J6. For valuable discussions see White (1986; 1971);
F. A. Lewis (1982); and Miller (1973).
26 It may be useful here to consider the example of 'white man' and 'cloak'.
(See Met. z. 4, 1029h27 ff., and similarly, H. 6, 1045"25-9.) Aristotle clearly
intends that the word 'cl9ak' and the formula 'white man' should have the same
However, this controversy need not detain us here. Our interest in
determining whether Aristotle thought matter and form are identical
is surely an interest in whether he thought they are identical in
this standard sense; if he employs a different concept in explaining
their relationship, then our conclusion should be that they are not
identical.27 Moreover, our current focus is the hypothesis that matter
and form are identical. There may be various reasons to resist this
hypothesis; however, in evaluating it we may choose to grant, at least
for the sake of argument, that Aristotle was in a position to assert this
identity.
On the face of it, it is difficult to see I'tow one could defend the
suggestion that the matter and form of a sensible composite are
identical; in fact, Aristotle typically characterizes them in contrast to
one another. The substantial form, being primary, is one and the
same as its essence, the matter is not (1037a34-b7); the form is
properly definable, the matter is not (1036a28, 1036a2-9, 1037a27-30,
1039b20-31); the matter has material parts, the form does not
(1035a26-32); matter is an 'element' of composite substance, form is
not (1041b25-34); the form is predicated of the matter but the matter
is not predicated of the form (1043a5-6); the form is prior to the
matter, but (presumably) the form is not prior to itself (1035h12-22);
individual substance must consist of something whch is not the matter
or of matter, which is the form (1041h11-34, 1043brn-18). The list of
differences could grow very long. Even if one thinks that Aristotle did
not hold all of these theses, only one is needed to cause difficulty for
the claim that matter and form are identical.
There are two strategies for defending the suggestion that matter
and form are identical, in spite of these apparent differences. The first
strategy is to develop precise distinctions between different kinds of
matter and different kinds of form in order to allow that some kinds
of matter and form differ, while also claiming that there is (at least)
one kind of matter and (at least) one kind of form that are identical. 28
meaning, so that white man = cloak. If Aristotle denied LL, how would obser-
vations about the essence of cloak reflect on the essence of white man, and vice
versa?
27 Charles Kahn ( r98s: 328) suggests that Aristotle might have had a non-
Leibnizian concept of identity. I think, however, it would be wrong to consider
such a concept the concept of identity.
28 e.g. we could be sensitive to distinctions between the remote and the proxi-
mate matter of the substance; or the matter from which a substance is generated,
and the matter serving in the composite; or the species form and the individual
form; or the composite nature of artefacts in contrast to organic substances. I take
Kosman (1984; 1987) and Irwin (1988) both to incline towards this strategy.
148 Sally Has/anger
On this strategy, most of the claims listed above apply to kinds of

matter and form that are not identical; it is only a very special kind of
matter and form that are identical. I will postpone discussion of this
strategy until late in this section, though there I will draw on some of
the arguments that arise in the interim.
The second strategy is to claim that although the matter and the
form of a substance are identical, the terms 'matter' and 'form', and
phrases employing them, have different senses. 29 So terms such as 'the
matter of Socrates' and 'the form of Socrates' function as different
descriptions referring to one and the same thing. 30 It is difficult to say
how exactly this is supposed to solve the problem, unless we propose
that in applying the problematic theses to individuals, the relevant
matter and form terms occur opaquely, i.e. the substitution of co-
referential terms does not preserve truth-value because the terms
don't function purely referentially. Let's call this the 'descriptivist'
strategy.
An example will help demonstrate this approach. To simplify
matters, let us consider a singular statement concerning matter and
form:
(I) The form of Socrates is definable, and it is not the case that the
matter of Socrates is definable.
If it is true that:
(2) The form of Socrates = the matter of Socrates,
then we must explain why we cannot validly conclude the contradiction:
(3) The form of Socrates is definable, and it is not the case that the
form of Socrates is definable.
To handle this, the descriptivist might offer the following proposal. In
the identity statement (2), both singular terms refer to Socrates;
however, they do so by virtue of diffhent descriptive contents. 'The
form of Socrates' presents Socrates under the aspect of his formal
properties; 'the matter of Socrates' presents him under the aspect of
his material properties. In statement (I), we find the predicate 'is
definable'. Let us suppose that this predicate is sensitive to the de-
scriptive content of terms with which it occurs; so it is not the case
that the truth-value of claims in which 'is definable' occurs depends
29 T. Irwin sometimes suggests this view; e.g. Irwin (1988: 251-2). It also
appears in interpretations of Aristotle's philosophy of mind. See e.g. Nussbaum

( r~84: esp. 201-3).
- 0 It is difficult to state the proposal in generality, since the terms 'matter' and
'form' function in different grammatical roles. I will focus on the proposal as it
concerns matter and form terms functioning as definite descriptions.
only on the referent of the subject term. For example, we might

propose that the form of Socrates is definable only if Socrates, in
virtue of his formal properties, is definable; correspondingly, the matter
of Socrates is definable only if Socrates, in virtue of his material
properties, is definable. On this account, we might say that the de-
scriptive content of the singular term not only serves to pick out the
referent, but also serves to complete the pre~cate.
Since (continuing with the proposal) Socrates is definable in virtue
of his formal properties and not in virtue of his material properties, we
can explain why the claim ( 1), appearing to contrast form and matter,
and the identity (2) are true, even though the descriptions cannot be
validly substituted to reach the contradiction (3). The co-referentiality
of the descriptions in (2) is not sufficient to warrant their substitution
into statements whose truth-value also depends on the descriptive
content of the terms. Pursuing this line of thought, one could interpret
the generalization that form is definable though matter is not, to mean
that it is in virtue of its formal properties that something (namely a
substance) is definable, and not in virtue of its material properties.
More generally, the descriptivist strategy has us explain the appear-
ance that the matter and form of a substance do not share their
properties by challenging the idea that the problematic theses attribute
conflicting properties to matter and form. In these contexts either 'the
matter' and 'the form' do not have their ordinary referent, so we are
attributing conflicting properties to other (distinct) things; or the
properties attributed to the common referent of 'the matter' and 'the
form' although seemingly conflicting, are actually compatible. If this is
right, then one can continue to maintain that the matter and form are
identical: one could claim that in transparent contexts the terms have
the same referent, but the difficulty lies in properly interpreting the
linguistic contexts that give rise to the substitution failures.
Although it may be plausible that in the example just sketched, the
predicate 'is definable' generates an opaque context, I am not offering
it as a plausible interpretation of Aristotle. I don't think it is. The
point is to illustrate the kind and the extent of the project that would
be required to fulfil the hopes of this approach. The worry is that even
if one regards the example as promising, things will get progressively
more difficult in extending the general approach to cover all of the
theses that contrast matter and form, especially if one seeks a syste-
matic rather than an ad hoc account of the opacity. Of course, there
are many elegant devices one might borrow from contemporary dis-
cussions of opacity that could be employed to avoid the apparent
contradictions. This is not the place to argue against specific proposals;
instead I will begin by indicating why such a strategy does not look
Sally Has/anger
promising and will then move on to raise an objection drawing on

points that emerged in the previous section.
There are a number of reasons to be wary of the invitation to treat
as opaque those contexts that do not allow substitution of matter and
form terms, terms such as 'the matter of Socrates' and 'the form of
Socrates'. One significant basis for concern is that the substitution
failures can be localized to certain pairs of terms, in particular, to just
those pairs whose substitution would be licensed by the controversial
identities. For example, if one allows the identities
(4) The matter of Socrates = the body of Socrates
or
(5) The form of Socrates = the soul of Socrates,
substitutions based on these identities seem to preserve truth-value.
The difficulties arise only if one substitutes on the basis of identities
such as:
(6) The matter of Socrates = the form of Socrates,
(7) The body of Socrates = the soul of Socrates.
Of course, a context may be opaque only with respect to the substi-
tution of certain pairs of terms. But what is the explanation for the
fact that the context behaves differently for just those (allegedly) co-
referential terms occurring in (6) and (7)? Shouldn't we simply con-
clude that the terms are not, after all, co-referential, in short, that the
identity statements (6) and (7) are false? The problem becomes acute
if we acknowledge that there are independent grounds for questioning
the truth of the identities, since in order to maintain that the context
is opaque for just those substitutions, one must rely on the assumption
that the identities hold. This makes the whole project appear to be an
ad hoc effort to save a floundering hypothesis.
But aside from these broad doubts about the pri>spects of such an
account, it would be informative to consider whether there is anything
that would positively preclude attributing to Aristotle the view that
matter and form terms function as co-referential descriptions. A com-
plete discussion of this issue should include a discussion of Aristotle's
views on language and signification (e.g. in Met r. 4). However, we
may also consider the question more directly: if we can find compelling
arguments to think that matter and form are not identical, then this
will undermine the project of treating them as co-referential descrip-
tions. In the context of debate over the descriptivist strategy, the
challenge will be to find an argument that does not turn on finding
contexts in which substitution of matter and form terms fails; for this
complaint the descriptivists have a general strategy of response (treat
the contexts as non-transparent). In the argument that follows, I will

endeavour to show that concerning the matter, form, and composite
we have a basis for distinguishing them, i.e. that there are contexts in
which it is plausible to give the descriptive terms wide scope (aiming
to nullify the effects of their descriptive content), while affirming a
difference between the items referred to.
In considering whether matter and form are identical, I will assume
that if they are, then they are each identical to the composite of matter
and form. In effect, they are improper parts of such a composite. 31 I
will also assume that the concrete substance is the composite of matter
and form. So if the matter of Socrates and the form of Socrates arc
identical, then each is identical to the concrete substance Socrates. I
will argue that given Aristotle's remarks in Z. 17, we can conclude that
the matter of a substance is not identical to the concrete substance. In
short, although the matter and the composite 'are one', their unity is
not basic, but instead relies on the fact that the form is present in the
matter; hence, their unity should not be interpreted as identity. But if
matter is distinct from the composite of matter and form, then it
follows from the assumptions just stated that the matter and the form
are distinct.
Recall that in z. 17 Aristotle explicitly contrasts two sorts of answers
to questions of the form 'why is X, Y?': answers which display that
X's being Y is a basic fact, and answers which provide substantive
causes and principles of unity. His paradigm of non-trivial answers is
that which cites the form as explanation of why the matter is the
composite:
Since we must know the existence of the thing and it must be given, clearly
the [meaningful] question is why the matter is some definite thing; for ex-
ample, why are these materials a house? Because that which was the essence
of a house is present. And why is this, or this body in this state, a man?
Therefore, what we seek is the cause, i.e. the form, by reason of which the
matter is some definite thing. And this is the substance of the thing. (1041b5-
10, emphasis mine ) 32
So, consider the question: why is this body a man? Or:

(0) In virtue of what is this body a man?
31 Here I assume that matter and form are exhaustive parts of the sensible
composite of matter and form, i.e. whatever other parts the sensible composite
has will be included in one of these two. So if matter and form are identicaL
there's nothing left to distinguish that thing from the composite.
32 This is a slight revision of the Ross and Barnes translation; I have used
'some definite thing to translate rz in both occurrences; and have used simply 'this'
rather than 'this individual thing' for woi.
Sally Has/anger
In the context of the discussion we may take it as given that there is a

concrete substance of which the matter in question is a part, and that
the matter and the concrete substance are one; the question is in
virtue of what are they one? Thus we may interpret the question (Q)
as:
(Q1) Concerning some definite thing that is a man (and which we
may assume is one with this body), in virtue of what is it one
with this body?
In interpreting (Q) as (Qr), it makes clear that the issue in (Q) is not
why we include certain bodies in the species man, or why we predicate
the species of bodies (if we do at all); the issue is: what the relationship
is between something that is a man, and its body. Aristotle is using
the term '[a] man' to refer in an indefinite way to an individual,
thereby bracketing questions of individuation. 33 (In effect, we bracket
the question: why is this body this man, to consider what in general
makes particular bodies men.)
It is clear in providing an answer to (Q) that Aristotle does not
repeat platitudes concerning identity; this body is unified with some-
thing that is a man not because 'each thing is itself', but because the
body has the form. This invites the contrast between (Q1) and:
(Q2) Concerning some defini.te thing that is a man (and which we
may assume is one with this man), in virtue of what is it one
with this man?
By taking this as the question raised earlier in asking 'why is the man
a man?', it becomes clear why the answer 'each thing is itself' applies
(1041b16-18). 34 To (Q2), we can answer by repeating platitudes con-
33 Consider asking in English how we should treat the phrase 'a B' in statements
of the form 'A is a B'. We might consider including 'a B' as part of the predicate
'is-a-B', suggesting that the property of being a B is being predicated of A. Or we
might take the copula to express a relation between an object and a species: A is-
a B. However, an option plausible in some cases is that the locution 'a B'
functions to refer indefinitely to a particular B. (This use of indefinite descriptions
in English should be familiar, e.g. a man ran down the street, I went to the
movies with a man, etc.) I am proposing that in the examples in question, this last
option is the best way to capture Aristotle's point; so we should take the form of
'A is a B' to be (roughly): '3x(.Bx & a is x)' (allowing of course for different
senses of 'is'); or: '3x(Bx & a is unified with x). Although admittedly Greek lacks
the indefinite article, Aristotle's use of the indefinite pronoun (u) in raising the
initial question may alert us to a more general option for interpretation, applying
even to 'why is the man a man?' (1041"18). However, I offer this proposal on
philosophical grounds, granting authority to others on the Greek.
34 The interpretation I am proposing here contrasts with the interpretation
offered in Cohen (1984), and endorsed by F. A. Lewis (1985). On their view,
cerning identity. Concerning the man with which this man is one, it is
one with this man because they are identical: 'each thing is itself'.
Their unity is basic; there is no non-trivial way to explicate their
relationship. However, concerning the man with which this body is
one, its being one with this body requires a substantive explanation;
they are one in virtue of the body having the form. But if we suppose
that the man and the body are identical, how do we account for the
difference?
These observations indicate how to construct an argument for the
conclusion that, on Aristotle's view, the body and the individual man
are not identical. Let us grant that concerning the man and the body,
if they are identical, then they are such that their unity is basic. In
arguing against the descriptivist, it is important to put the point in this
way in order to give the descriptions wide scope; by doing so we can
expect that descriptive content of the terms is no longer interfering.
We now should ask, concerning that which is the body and that which
is the man (independently of these descriptions), is their unity basic?
Are we entitled to answer questions concerning their unity: 'each
thing is itself'? Relying on Aristotle's use in the quote above of the
simple indexical 'this' in reference to the matter, we should consider
how to answer the following:
(Q3) Concerning some definite thing that is a man, and concerning
this which is a body (and which we may assume is one with the
definite thing), in virtue of what are this and the definite thing
one? 35
Those who would claim that the body and the man are identical
should answer this question by offering a platitude that reveals that
the relation between the matter and the composite is basic, for ex-
ample, a platitude about identity 'common to all cases'. But this is
implausible as a reading of Aristotle's intentions. The text is clear that
this and the definite thing are one in virtue of this having the appro-
priate form; the explanation of their unity calls upon substantive
in considering 'why is the man a man?', Aristotle is considering a challenge to the

de dicto claim 'whatever is a man is a man', and finding the query pointless. This
interpretation makes little sense of his response that 'each thing is itself', and
further, it takes him to be confused about possible wide scope {de re) readings on
which the question may reasonably arise (namely of that which is the man, why is
he a man?). By taking the indefinite description 'a man' to refer indefinitely to a
particular man rather than taking it to function with the copula as a predicate, we
can avoid attributing to Aristotle such blunders.
35 In more modern terms we might put it: (Q3*) 3x(Mx & 3!y((By & Oxy) &
'?Oxy)).
154 Sally Has/anger
inquiry. 36 Given his lack of variables, there is little more Aristotle

could have done to indicate his intentions to speak of the objects,
however described.
In contrast, if we ask:
(Q4) Concerning some definite thing that is a man, and concerning
this which is a man (and which we may assume is one with the
definite thing), in virtue of what are this and the definite thing
one?
We do not answer (Q4) by saying that this has the form. Aristotle
explicitly contrasts the correct answers to questions such as (Q4) that
cite basic identity facts, with answers to questions such as (Q3) that do
not. But if the matter and the individual substance are identical, then
in both (Q3) and (Q4) the 'this' and the 'definite thing' refer to the
individual substance; so in both cases there should be no further fact
that explains why they are one; and in both cases we should answer
the question with the platitude that 'each thing is itself'. But this is not
what Aristotle says.
On the hypothesis I am criticizing, the fact that the matter and the
individual substance are one just is the fact that individual substance is
one with itself. But the argument just offered suggests that the latter
fact is basic and the former is not, for the matter and the individual
a
substance are one in virtue of further fact (about the form). So we
should conclude that in (Q3), this and the definite thing are not
identical; in other words, concerning the matter and the individual
substance, it is not the case that they are identical. It is because the
matter and the individual substance are not identical that the 'short
and easy way with the question' is unavailable.
In effect, the unity of the body and the man is derivative, resting on
the more basic fact that the body has the form. The body and the
composite are unified in virtue of the fact that there is a form which
the body has, such that the combination of the body plus the form is
identical to the individual man. The body is a proper part of the man,
occurring together with the form. Thus, the man is identical to the
man; the body is not; more generally, the individual sensible substance
is identical to the individual sensible substance, but the matter is not.
But if the matter is not identical to the composite, and if it is the
relation it bears to the form that makes the difference, then the
matter is not identical to the form. Substantial unity is not identity.
36 Even if we do think concerning that which is a man, and that which has a
body, that their unity can be explained by reference to trivial identity facts, the
body is a body and doesn't have a body (though the man does). So we've failed to
capture the point of (Q3).
Because this argument does not depend upon the failure of substi-
tution of matter and form terms, the descriptivist cannot complain
that the descriptive content of the terms is responsible for the appear-
ance that matter and form have different properties. The matter and
the composite, those things however described, are distinct; to deny
this is to miss the distinctive role of form as unifying cause. Moreover,
it would be misguided to resist by saying that the matter and the
composite are identical 'in a different way' from the way the composite
is with itself, and that this 'different way' is grounded in the form.
There is only one way to be identical, namely the way you are
identical with yourself.
Let me now consider briefly the alternative strategy mentioned at
the beginning of this section for defending the idea that matter and
form are identical, namely that there are different kinds of matter and
different kinds of form, and it is only one special kind of matter and
form which are identical. A thorough discussion of this strategy would
require examining a variety of different proposals concerning the
exact nature of the kind of matter and form which qualify for the
identity. Given limitations on the scope of this paper, my argument
against this strategy is more in the form of a challenge to such
interpretations than a proper rebuttal.
In short, the argument just given against the identity of matter and
form placed very few constraints on what could count as matter, form,
and composite. Thus I propose that it could apply generally to what-
ever might count as plausible candidates. The argument made only
two assumptions about the matter, form, and composite: (i) the form
(or the matter's having the form) is that in virtue of which the matter
is the composite, and (ii) this explanation offers a substantive ground-
ing of the fact that the matter and the composite are one; it does not
amount simply to repeating the fact in different terms. The texts
strongly support these two assumptions. Thus, I would propose that
any plausible account of the matter and form of an individual substance
must grant the minimal assumptions of the argument above in order
to do justice to Aristotle's treatment of matter and form, and his
views on identity and unity. But if this is right, we should reject any
account of matter and form that takes them to be identical.
In summary, I've argued that we should not accept the hypothesis
that the matter and form of a sensible composite are identical. Because
Aristotle consistently contrasts the matter and form, such an interpre-
tation is prima facie implausible, and we should resist attributing to
him such a view unless there are compelling reasons to do so. I've
argued that in order to square the proposed identity with the texts,
the descriptivist strategy would have us attribute to him a questionable
156 Sally Has/anger
and unwieldy semantics for virtually all of his central theses con-
cerning substance, and on any account of the identity we would be
forced to deny the distinctive role that form plays in offering a basis
for the unity of sensible substance. Thus, we should move on to
consider accounts of the unity of sensible substance that make sense
of the claim that matter and form, although distinct, are one.
VII. MATTER, FORM, AND UNITY

Let us now turn to consider more directly how matter and form are
unified in a sensible composite. If matter and form are distinct, in
virtue of what do they form a composite whole, namely the individual
substance? The text from H. 6 should provide an initial guide to our
answer: 'But, as has been said, the proximate matter and the form are
one and the same, the one potentially, the other actually' ( 1045h17- 19).
Given Aristotle's preference for specifying principles of unity by inte-
gration principles, it is plausible to see him as here implicitly indicating
a privileged individual to which matter and form bear relations, re-
lations that somehow concern potentiality and actuality. So we might
begin by asking what he intends to take as the privileged individual
that integrates matter and form in the sensible composite. We will
consider below how potentiality and actuality enter the picture.
In characterizing kinds of unity, Aristotle claims that 'the things
called one in the primary way are those whose substance is one ... '
(Met. L1. 6, 1016b8-9). With the expectation that substantial com-
posites exemplify a primary sort of unity, we should hope to specify
the unity of matter and form by virtue of their relations to one
substance. 37 But it is important to distinguish two different proposals
that would each take substance as the basis for substantial unity. As
Aristotle points out, sometimes in speaking of sensible substances we
are unclear whether we mean the form or the composite:
37 Aristotle says at several points that 'things whose substance is one and whose
essence is one are themselves also one' (Met. z. 13, 1038b13-14; also 1032•3,
104obI7_ 19). It should be clear from our discussion that this need not mean that
things whose substance or essence is identical are themselves identical. Although
things which are identical to the same essence are identical, there are many ways
one might be related to an essence, and many things might be related in the same
way (other than identity) to the same essence: e.g. Socrates and Coriscus may be
(formally) one by 'having' a common form (namely the species form); and the
matter of a substance and the composite substance may each 'have' a single form
(either individual or not), without being identical, because they 'have' the form in
different ways. So unless two things are identical to the same essence, sameness of
essence cannot be used to conclude identity. There has been much confusion on
this point. Cf. Irwin (1988: 243).
We must not forget that sometimes it is not clear whether a name means the
composite substance, or the actuality or form, for example, whether 'house'
is a sign for the composite thing, 'a covering consisting of bricks and stones
laid thus and thus', or for the actuality or form, 'a covering' ... (Met. H. 3,
1043"29-33)
Hence, two hypotheses stand out as reasonable. First, if one thinks of
matter and form along the lines of 'aspects' or 'features' of a sensible
substance, then one might take the composite substance itself as the
privileged individual that unifies its material and formal 'aspects'. On
this view, the matter and the form of Socrates are one in virtue of
each being 'aspects' of Socrates. (I intentionally leave the notion of
'aspect' vague here.) Alternatively, one might take the substantial
form as privileged individual. Although the form is itself part of the
unity whose composition is in question, it may serve as a privileged
individual in an intrinsic integration principle.
I will begin by considering the hypothesis that the composite sub-
stance itself serves to unify matter and form, and I will argue that this
proposal is inadequate. Since this is the view on which it is most
plausible to treat matter and form as 'aspects' of a substance, its
difficulties show that, at best, the 'aspects' terminology is seriously
misleading. In the next section, I will consider the alternative hypo-
thesis that the substantial form serves as the basis for substantial
unity. There I will offer a schematic principle of unity that explicates
the unity of matter and form in terms of the substantial form, and I
will argue that this better captures Aristotle's position. This principle
will also lend insight into the unity of the constitutive matter of a
sensible composite such as Socrates' body.
In order to evaluate the proposal that the individual substance
serves as the privileged individual to ground the unity of a sensible
composite, we should consider what kind of parts matter and form
are, and what relations the matter and form bear to the substance.
Both of these questions lead us into a dense tangle of interpretative
difficulties; without a thorough interpretation of Aristotle's views on
potentiality and actuality, proximate matter, and substantial form, it
will be impossible to offer a clear specification of all the options.
Granting the complexity of the issues and the difficulty of doing
justice to the many sophisticated interpretative projects, I propose the
following approach. I will argue that if one takes the concrete sub-
stance to ground the unity of matter and form, then one must deny
that the matter and form are both constitutive parts of it, in other
words, one cannot allow the account of a sensible substance to depend
on a reference to both matter and form. (This follows from our
discussion in section IV.) But if we allow two plausible assumptions,
Sally Has/anger
this can't be right. The first assumption is that an account of sensible

substance depends upon reference to form; the second is that the
relation of form to the composite is derived from a more basic relation
form bears to matter. Given these assumptions, if one proposes to
account for sensible substance by reference to the composite's relation
to form, then one must proceed to explicate this relation in terms of
form's relation to matter. This brings matter into the account of
sensible substance, so both form and matter must be constitutive
parts. If they are, then the substantial composite is an aggregate of
matter and form, and so (as I argued above) it cannot serve as the
privileged individual in the integration principle that accounts for its
own unity.
Can we flesh out this argument? Recalling the discussion in section
IV, we found that if an integration principle serves to unify a group of
constitutive parts in an aggregate, then the privileged individual cannot
itself be an aggregate of that kind, on pain of circularity. Suppose,
following the hypothesis under consideration, that matter and form
are substantially unified just in case there is a sensible substance to
which they each bear a special relation (yet to be specified). If to be a
sensible substance is just to be a unity of matter and form, then the
integration principle just indicated will fall prey to the objection that
an aggregate is being used to ground its own unity. So either the
sensible substance cannot serve as privileged individual in the account
of substantial unity-and the hypothesis is false-or reference to both
matter and form is not really required in an account of sensible
substance, i.e. they are not both constitutive parts. (The latter sugges-
tion that matter and form are not constitutive parts fits well with the
idea that they are 'aspects' of sensible substance. Plausibly an aspect
of something X depends on a reference to X to account for what it is,
but not vice versa; aspects are posterior, and so not constitutive parts
of a composite. 38 So to evaluate the hypothesis that the sensible
substance grounds the unity of matter and form, we must determine
whether an account of sensible substance makes reference to both
matter and form.
To determine whether both matter and form are required in an
account of sensible substance, it would help to include a study of
38 I grant that there may be other construals of the 'aspects' terminology. The
main point I am concerned to show is that an account which either (i) takes
matter or form to be posterior to the individual substance or (ii) takes the
individual substance to be the privileged individual in an account of substantial
unity, is mistaken. I believe that the argument I offer here also works against
accounts such as T. Scaltsas's (this volume) which take matter and form to be
'identity-dependent' on t,he sensible substance.
Aristotle's views on definition.'" This is a large task I will not under-
take here. For our purposes, we may begin work with two uncontro-
versial observations: first, Aristotle believed that in order to say of a
sensible substance, what it is, one must make reference to its form;
and second, he is less clear on what role, if any, we should grant
matter in an account of composite substance. So, granting that form is
a constitutive part of sensible substance, we should focus on whether
matter is also a constitutive part.
Aristotle clearly thought that if an account of something required
reference to matter, then the account could not provide a proper
definition. However, he also believed that composite substance was
not properly definable, at least not in the sense that a substantial form
is definable. It is plausible to think that this is at least partly because
an account of sensible substance must make reference to matter (see
Met. Z. 7; Z. 15). So if asked to say of a given sensible substance,
what it is, we seem to have two options. Either we can offer the
definition of its form, which is (presumably) a proper definition but
only in a derivative sense its definition (since the sensible substance
does not satisfy the definition in virtue of itself, but only in virtue of
its form); or we can offer a secondary sort of definition that includes
not only a reference to the form, but also to its appropriate kind of
matter (and although this may not count as a definition in the primary
sense, it will be the best we can do for something of this kind). Since
our interest is in the account that applies to the composite itself,
whether or not this is the strictest sort of definition, it is reasonable to
think that a reference to matter will be included in such an account.
Drawing on this rough picture of Aristotle's position, can we make
the argument more precise? Let us grant that a reference to form is
required in an account of a composite substance, and let us suppose
for the argument, that a reference to matter is not: matter fails to be a
constitutive part. If this is correct, then we should expect that to give
an account of the substance, we should characterize it as having a
certain (kind of) form. More generally, letting 'has*' indicate the
relation between the composite and the form, we might say:
(SS) c is a sensible substance iff there is a substantial form f, and c
has* f.
(Of course a substantive account of c will also offer a definition of the
form that c has*.) But there are compelling reasons to think that the
fact that a composite has* a form holds in virtue of the form's being
(metaphysically) predicated of its matter (see, for example, Lewis
39 I am drawing here on valuable discussion of Aristotle's views of definition in
Code (1985a; 1986a); and Lewis (1986).
160 Sally Has/anger
(1985); Modrak (1985); Code (1978; l985a)). Briefly, in accepting this

as Aristotle's view, we are in a position to provide a simpler account
of his views on predication; we gain insight into the discussion of
whether a form is a universal (and if so, in relation to what it is
universal); we gain the resources to explicate his account of substantial
generation and destruction; and we make good sense of the discussion
in Z. 17. Just as the relation between matter and the composite rests
on a more basic relation between matter and form; the relation be-
tween the form and the composite is likewise derivative. But if the
relation of composite to form depends on the relation between matter
and form, then we must further explicate the condition that 'c has* f'
in (SS). Letting 'has' indicate the relation of matter to form:
(SS+) c is a sensible substance iff there is a form f, and matter m,
and both m and fare parts of c, and m has f.
Since this brings matter into the account of sensible substance, we
must conclude that matter is a constitutive part. This contradicts our
initial assumption that matter is not a constitutive part. Although
there are several assumptions made in the course of the argument, I
submit that the weakest is the claim that matter is not a constitutive
part, and so we should reject it. 40
Because reference to both matter and form are needed in an account
of a sensible substance, then we cannot rely on sensible substance to
ground their unity. Therefore, we should reject the hypothesis that
the composite substance serves as the privileged individual in the
integration principle for matter and form. It follows that we should
not take matter and form to be unified in a particular such as Socrates
by virtue of their relations to the concrete individual Socrates, for
example, by each being 'features' or 'aspects' of Socrates. Socrates
cannot serve as the basis for the unity of his matter and form, since
the unity of Socrates depends in turn on their unity.
VIII. FORM AND INTEGRATION
If we explicate substantial unity using an integration principle, then

(once again) we are seeking that to which matter and form are each
40 Admittedly, there are points where Aristotle does characterize matter as
posterior to the composite substance, e.g. Met. z. ro, ro35b9-28. This raises
doubts about the argument just presented, since if the matter is posterior to the
composite, it is not a constitutive part. I would propose that although body parts
may be posterior to the composite substance, the body as a whole is not; and I
would suggest further that we consider reading passages in which Aristotle speaks
of matter as posterior as not primarily concerning the notion of priority in
~.H'·l"Annt h11t c.•nm.ath1nn m.nri::io 1il--.a nr1rvr1h1 in ,av1ct.anl"'i::>
related which serves as the ground of their unity. The passage in H. 6

suggests that the relations which serve as the basis for substantial
unity are being potentially x, and being actually x: 'But, as has been
said, the proximate matter and the form are one and the same, the
one potentially, the other actually' (1045b17-19). So we should ask:
What are the relevant relata for the relations being potentially and
being actually? What is it that matter is potentially and form is actually?
The best remaining candidate is the substantial form. So on the view
now before us form and matter are unified in a sensible composite in
virtue of the relation each bears to the form: matter is (in potentiality)
the form, form is (in actuality) the form.
Since Aristotle consistently speaks of form as prior to the composite,
the account of the form will not depend upon reference to the com-
posite, so we need not fear the circularity found above. Drawing on
the quote from H. 6, we might initially formulate the integration
principle for sensible substance as follows:
x and y are substantially unified iff
(i) y is a substantial form, &
(ii) x is potentially y, &
(iii) y is actually y.
Or:
(SU) x and y are substantially unified iff there is a z such that:
(i) z is a substantial ,form, &
(ii) x is potentially z, &
(iii) y is actually z.
The reasons one might prefer one of these formulations over the other
are fairly subtle. The second allows (but does not require) that there
is both a definable species form (z), as well as a distinct individual
form (y) that functions as the formal constituent of the composite.
For our purposes in offering a very schematic statement of the con-
dition this openness has some advantages, so let's work with the
second: (SU). 41
41 In considering (SU), there are several background concerns to keep in mind.

First, it is unclear what sort of variables (and correspondingly, what sort of
quantification) is appropriate here, e.g. should we be sensitive to a distinction
between first- and second-order variables and quantifiers to accommodate the
contrast between substantial form and the constitutive matter? Second, there may
be reason to hesitate in using variables which range over matter, for perhaps
matter is only potentially a 'this', not actually a 'this'. Third, it is unclear how
we should read the 'is' in the two occurrences 'is potentially' and 'is actually':
Should we isolate the 'is' from the accompanying adverb or not, and if so, should
we read it the same way in both occurrences? Keeping the~e caveats in mind (~nd
162 Sally Has/anger
(SU) plausibly accords with the idea that matter and form are
substantially unified just in case their substance is one, i.e. just in
case there is one substance (the substantial form) to which each is
appropriately related. The full story behind the unity of a sensible
composite must account for the unity of substantial form, sin~e the
unity of substantial form is assumed by the principle. It is likely that a
substantial form is not an aggregate; even if it is composite, its parts
are not constitutive parts (in the sense defined in section IV), so its
unity could be basic. Admittedly, the problem of the unity of form is
an important and difficult one for Aristotle, but we need not account
for its unity here, since we have no reason to dispute its qualifications
to serve as privileged individual.
Although the integration principle formulated in (SU) indicates a
general structure for substantial unity, it depends crucially on two
further relations: being potentially f, and being actually f. In some
sense, to understand what is required for the unity of sensible com-
posites we must understand what it is to stand in these relations.
So far I avoided discussion of Aristotle's views on potentiality and
actuality, concentrating on more schematic issues of unity. My prefer-
ence would be to offer the principle in this schematic form, leaving it
to others to interpret Aristotle's theory of actuality and potentiality. 42
But as it stands, (SU) doesn't work, and something more needs to be
said about actuality and potentiality in order to chart a way through
the difficulties. Let me emphasize, however, that my remarks are
intended to indicate how we might draw on an account of potentiality
and actuality to solve the difficulties; they are not aiming to provide
that account.
The problem is this: If we want to demonstrate how matter and
form are substantially unified to constitute an individual substance,
and if we place no further restrictions on the variables in (SU), then
the principle doesn't provide a sufficient condition on substantial
unity. Because for a given substantial form f there are different ways
to be actually f or potentially f, the condition as stated does not
sufficiently restrict the members of a substantial unity to be the proxi-
mate matter and the form of an individual substance. In short, the
condition will not rule out unwanted unities unless we specify further
the kind of potentiality and kind of actuality appropriate.
investigate whether these conditions capture the general idea behind Aristotle's
account of unity.
42 There are many useful discussions of these issues. One account I find es-
pecially helpful in connection with the general strategy I have been pursuing is
developed by Charlotte \\'.itt (1989: ch. 4).
For example, we can see that condition (ii) on the material consti-
tuent is too weak, by noting that there is at least a sense in which
Socrates' mother's katamenia prior to the presence of his father's
form has a potentiality for human form; if so, then her katamenia and
a substantial human form (either existing separately or as part of
another individual) would meet the condition for being a substantial
unity. But if we want the principle of substantial unity to apply only to
the matter and form of a concrete substance, this is unacceptable. Her
katamenia and human form (or his father's individual form) each
by themselves does not constitute an individual substance; until they
'combine', Socrates does not yet exist.
Similarly, condition (iii) raises difficulties. It is important to note
that Aristotle often treats the composite as the appropriate subject of
activities that realize the form (see, for example, De Anima I. 4
(408br-32), and II. 5). For example, it is Socrates (or a man) who
sees, or builds, or thinks, or lives a life. Yet if we allow that the
substantial composite meets the condition on the second constituent-if
it is the composite which is actually /-the resulting unity won't be the
unity of matter and form we are seeking. For example, if it is Socrates
who is actually alive, and is thus actively living a human life, then it
would appear that we could combine the matter (his body) and the
composite (the man) together as a new composite (body plus man);
the parts of this new composite would meet the condition (SU) for
being a substantial composite. This too would be unacceptable.
There are several ways to avoid these consequences. The most
attractive is to seek to restrict the kind of potentiality and actuality
involved in (SU) so that only the appropriate parts-the constitutive
matter and form-will meet the conditions. The clause of (SU) that
we might modify to constrain the material constituent of the composite
is the requirement that it is potentially the form. As I noted above,
not just any sort of potential will do here; but we can attempt to avoid
the unwanted cases by employing Aristotle's distinctions between
different degrees of actuality and potentiality. More specifically, we
might state the condition in (SU) so that the relevant potentiality for
the form must be a higher degree of (passive) potentiality for the form
than is found in the katamenia prior to the presence of the form. The
challenge then becomes how we are to specify the relevant level of
potentiality. There is a range of options worth exploring. But for the
purposes of this paper, let us assume that we can uncover a reasonable
Aristotelian account of the degrees of active and passive potentiality.
Granting this, we can insert a place-holder for this account in our
schematic unity condition by requiring that the material constituent of
a sensible composite have the highest degree of passive potentiality
Sally Has/anger
for the form in question. 43 On this view, the material constituent (i.e.
the proximate matter) of a human composite must have the highest
degree of passive potentiality for human soul. For ease of exposition,
let us say that something that has this highest degree of passive
potentiality with respect to a form f has a maximum passive potentiality
for f. If such a restriction on the appropriate kind of potentiality can
be explained and defended, it would have the consequence that the
potentiality requirement could only be met by the constitutive matter
of the composite. This would rule out the unwanted cases.
By specifying the constitutive matter in terms of its distinctive
potentiality for the form, we also gain a means for distinguishing the
unity of the matter from the unity of the composite. Briefly, the
composite is unified by virtue of an intrinsic integration principle, i.e.
a principle in which the privileged individual (the form) is a part of
the resulting unity. In contrast, the matter is unified by virtue of a
extrinsic integration principle; form is not a part of the matter, though·
it is the relation of material parts to the form that accounts for their
unity as a body. 44 To spell this out further it would be important to
explicate how it is that material parts (e.g. organs, flesh, etc.) are
related to the form of the composite, for on the hypothesis under
consideration it is in virtue of their relation to this form that they are
unified in a material whole (a body). But it is plausible that in giving
an account of such material parts one must mention their contribution
to the more complete potentiality for the form that is ultimately
exhibited by the body as a whole.
If human bodies are understood to be just those things that satisfy
the condition of having a maximum potentiality for human form, then
we can explain why a corpse is not (properly speaking) a body; at
death the body loses that potentiality for human form. Likewise, a
finger ceases to be (properly speaking) a finger if severed from the
whole, for then it ceases to contribute (in its fingerish way) to the
potentiality for human form. 45 So although the form is the basis
43 One way to develop such an account will be to focus on the contrast between
those passive potentialities which are realized if nothing hinders, and those which
are inadequate or incomplete unless (or until) there is a further change. Things
which have a passive potentiality have a potentiality to be acted upon, and they
have the highest degree of passive potentiality if they are being acted upon. See
e.g. Met. e. 7. For a valuable discussion of relevant issues see Gill (1991: esp.
246ff.).
44 This strategy of distinguishing the proximate matter and the concrete sub-
stance by employing the contrast between intrinsic and extrinsic integration

principles fits well with Frank Lewis's account in Lewis (this volume: sect. VI).
45 On this issue see, Met. Z. IO, 1035h10-14 and De. An. II. 1; also Ackrill
( 1972-3). Note also that this understanding of matter also explains why we need
for the unity of both the matter and the composite (by serving as
the privileged individual in the integration principle for each), the
principles will importantly differ; the principle of substantial unity
requires that the form be a constituent of the resulting composite, the
principle of material unity does not.
Allowing that this general strategy can rule out unwanted material
constituents, let us now turn to condition (iii), which requires that the
second constituent of the composite is actually the form (or is the
form, in actuality). Our problem was that this condition seems to be
met not only by forms, but also by composite substances such as
Socrates. To avoid the difficulty we might begin by drawing a dis-
tinction between actively being the form, and actively having or mani-
festing the form. Although in an important sense the composite realizes
the potential for the form-and so is what actually thinks or lives,
nevertheless, it is the form which (perhaps in a primary sense) actually
is the form-it is in actuality the activity of thinking or Jiving. 46 So,
for example, the composite realizes the potentiality for sight, and in
doing so actually sees. In seeing the composite manifests the activity
of sight, but this is not to say that the composite is actually being the
activity of sight. It is the form which actually is the activity of sight. 47
In a context where we are concerned with the realization of a potenti-
ality, it is appropriate to focus on the composite as the subject of the
activity; however in a context in which we are concerned with what
the potentiality is for, i.e. the activity that is to be manifested, it is
not worry that the replacement of material parts in a body over a lifetime results
in a series of bodies-in effect, a series of proximate matters-and so a series of
composite substances. The body endures so long as it retains the maximum
passive potentiality for human form, regardless of the replacement of individual
material parts; thus there is one composite of matter and form, i.e. one human
substance, throughout a human life. A similar point can be found in Locke's
Essay Concerning Human Understanding, ch. 27.
46 For valuable discussion of these issues, see Kosman (this volume). I am
sympathetic to Kosman's account of the relevant sense of cvipyeia in terms of
activity; see esp. sect. III of his paper.
47 It is a difficult question how to interpret the notion of being here; in keeping
with my efforts to remain neutral on the interpretation of the formal constituent
of the substantial composite, my intention is to leave the notion vague. We should
be careful to note, however, the notion of being may imply identity, though it
may not. If the constituent of the composite is an individual form, then it may be
that the particular activity of the composite is (in the relevant sense) the sub-
stantial form, though is not identical to it. Also, even if the relation in question
(of x being f) holds between the form and itself, this does not entail that the
relation is identity: there are many relations which hold between a thing and itself
which aren't identity.
166 Sally Haslanger
appropriate to focus on the form. With these considerations in mind,

it is plausible to say that it is only the constituent form which is being
the form, and so (in some strict sense?) is the form in actuality. Thus
to appropriately restrict the formal constituent we might revise the
relevant clause in (SU) to require that the second constituent in the
composite is actively being the form. The result of these changes will
be:
(SU + ) x and y are substantially unified iff there is a z such that:
(i) z is a substantial form, &
(ii) x has a maximum passive potentiality for z, &
(iii) y is actively being z.
This principle avoids some of the difficulties of (SU). However, the
condition it places on substantial unity still appears to be too weak if
we allow for the possibility that something may have a maximum
passive potentiality for a form, without that form being present in it; it
is plausible that there is only a genuine composite if the form is 'acting
on' the matter. There are two possible responses to this objection.
First, one might simply add to the principle the further requirement
that in some (undefined?) sense, the matter has the form. Although
this would avoid the problem of finding cases in which the form is not
'acting on' the matter, the worry is that the principle then fails to offer
an informative explication of substantial unity: by relying on an unde-
fined sense of the matter 'having' the form, we presuppose just what
we are seeking to understand. There is something right about this
complaint, in so far as our understanding of sensible composites rests
on whatever sense we can make of that relation. But at the same time,
the complaint asks for too much and grants too little. Admittedly, if
what we are seeking to understand is the relation of 'having' that
matter bears to form, then a revised (SU+) that employs that relation
will not be informative. However, it is not obvious that an explication
of this relation is the project at hand. The principle (SU+ ) is an
effort to explicate how it is that matter and form are 'one'. And as
such it indicates how the relation between matter and form is a kind
of unity resting on the prior unity of the form. On this account, a
sensible composite is one because it is not divisible by form; its
(constitutive) parts are indistinguishable with respect to the form that
accounts (albeit in different senses) for what they are.
But a seconi:-1 ~esponse is available that takes up the challenge that
we are seeking to understand the relation of matter 'having' the form.
The reply would be that we are to understand this relation of 'having'
as holding in virtue of more basic facts about potentiality and actuality;
about what matter and form are. It is in virtue of matter being
potentiality, and form being actuality, that they come together, one in
the other. On this view, potentiality and actuality are the more basic
notions in terms of which we are to understand the way in which
matter 'has' the form. In line with this proposal, one might refrain
from adding a further condition to (SU + ) which requires that the
matter in question 'has' the form, and instead develop a more sophisti-
cated account of the connection between passive potentiality and
actuality. For example, one might strive to show that something has a
maximum passive potentiality just in case there is a corresponding
active principle at work, the active principle being the form.
However, the true mystery lies in the transition from there being
two things (e.g. father's form and mother's matter) which are not yet
a composite individual, to there being one thing with its own active
principle. 48 What distinguishes that point at which it is correct to say
that there is one individual composed of two parts, rather than simply
two distinct things little better than a heap? At the level of generality
at which we are working, it is difficult to provide a more informative
account than to say that at the point of transition the two begin to
work together as a single unit realizing (actively and passively) the
same end. In some (admittedly still obscure and metaphorical) way,
the form's being an active principle and the matter's being a passive
principle is that in virtue of which the matter 'has' the form. (SU+)
gestures at this point of transition by demanding that the matter be in
its highest state of (passive) potential for the form, and demanding
that the form be actively doing its work-that it be actively being the
form; 49 (SU+) is an adequate formulation of substantial unity just in
case it is under these conditions the two are working together as a
unit, as a single composite individual. 50
48 For discussion of this issue, see Code (1987b).

49 My account strongly resembles Kosman's (this volume: sect. 111) in so far as
we both account for the unity of substance by relying on the unity of ability (or
potentiality) and activity in the fully realized substantial being. However, there
are places where Kosman speaks as if he might take his account to explicate not
merely a unity between matter and form, but an identity between them: 'ability
and realization are the same thing' (p. 205). I remain committed to the view that
the unity between matter and form is not an identity; even if 'ability and realization'
(in the relevant senses) are always present together, they are nevertheless distinct.
50 Someone might complain that I have not offered an adequate account of the
unity of sensible substances (even schematically), because I have not addressed
the threat of an infinite regress mentioned at the end of z. 17; does such a regress
emerge on all accounts that allow matter and form to be constitutive parts? Very
briefly, I see no threat of infinite regress if we allow that matter and form are the
sorts of things which are, so to speak, designed to unite. (Think of Frege's
ontology of saturated and unsaturated objects.) On my view, Aristotle's point in
168 Sally Has/anger
At this point, I will leave (SU+) in its schematic formulation,

recognizing the need for further elaboration; my hope is that it
specifies the general structure of an account of substantial unity and
indicates where some unresolved issues lie. Turning to an application
of the principle to cases, we can use (SU +) to specify what it is for
something to be a substantial composite:
(SC) c is a substantial composite iff 3x3y(x and y are parts of c &
x and y are substantially unified).
It is important to recall that by specifying conditions on substantial
unity, we do not thereby answer questions concerning the definition
or individuation of a sensible composite; thus, the principle of unity
leaves undetermined which matter and which form are parts of a
particular composite c, or how we are to define the form f that
grounds the unity. However, if we are given individuated matter and
form, we can see how the principle applies to the case.
For example, if we are looking for a principle that specifies in a
general way how human bodies and human form are related to con-
stitute a unity, we can get this through a rather straightforward appli-
cation of (SU+):
(UHB) x and y are substantially unified in a human being h iff there
is a z such that:
(o) x and y are parts of h, &
(i) z is the form of [a] human being (namely human soul) &
Note that this condition does not state conditions on 'what it is to be a
human being' or 'what it is to be a man' (that, of course, will be
provided by the definition of the form, and is in some sense assumed
by this condition); rather, it states for human bodies and human form
generally, that in virtue of which they constitute composite (human)
unities.
The question remains how we are to explicate the unity of the
matter and form in a particular composite such as Socrates. The
natural move is to simply apply (UHB) by instantiating it with respect
to Socrates' body and Socrates' form.
(US) x and y are substantially unified in Socrates iff there is a z
such that:
Z. 17 is that we should not view the form as just another 'element' in the
composite, as if it were just another material part. Given what forms are, they
play a fundamentally different role from material parts. Their presence in a
composite does not yield just a bigger heap; rather, it is what makes the elements
into something other than a heap: a unified whole.
(o) x and y are parts of Socrates, &

(i) z is the form of human being, &
Let me emphasize that (US) is simply an instantiation of (UHB) and
does not show that there is a special kind of unity appropriate to
Socrates. In short, the unity of Socrates is the unity of a human being,
which is more generally the unity of a sensible substance. We have
guaranteed the application of the condition to Socrates adding the
condition that the parts unified be his parts; the condition does not
give us a means for deciding which body/form pair is his. The principle
simply tells us, given the Socratic body/form pair, what it is in virtue
of which they are a (human) unity, rather than a mere heap. The
failure of the principle to offer criteria for distinguishing which body
and form belong to Socrates does not show it fails as a principle of
unity, for in doing the work of unity it is not also required to specify
conditions for individuation.
So what now can be said in response to the questions with which we
began: In what sense is a sensible substance composite? And can we
explain the unity of the sensible substance while doing justice to the
priority of form? On the account I have sketched, a sensible substance
is an aggregate of matter and form; that is to say, matter and form are
proper parts of the substance, and an account of the substance depends
on reference to both. But given this we cannot explain the unity of the
sensible substance by treating the matter and form as dependent or
posterior parts of a substance, unified by reference to the substance
itself. Instead a sensible substance is unified by virtue of an intrinsic
integration principle that places substantial form in the role of pri-
viledged individual; matter and form are unified by their respective
relations to the form that is itself a member of the unity. More
specifically, the matter is potentially and the form actually the form.
So, as indicated in (US), the unity of a particular composite such as
Socrates is grounded in the prior unity of the appropriate substantial
form, namely human (or Socratic) soul; it is through their respective
relations to this form that the human matter and form constitute one
human being.
Aristotle means to contrast this view with the more Platonic(?)
ideas that Being or Unity is the substance of things (see Met. B. 3-4,
esp. 1001a2-b25), or that unity is achieved through some sort of par-
ticipation in 'the One-itself' or 'Unity-itself' (see Met. Z. 16, ro4ob17-
22, and H. 6, 1045b2-8). Although every thing is, and is one, this
does not capture the complexity and variety in what it is to be, and to
be one. A particular sensible substance is, and is one, by being of a
170 Sally Has/anger
kind such that its constituent matter and form are related (either in
potentiality or in actuality) to the substantial form of that kind. The
unity of sensible composites is grounded in the more basic (or prior)
unity of substantial forms. This provides at least one sense in which
substantial form is prior to sensible substance: for any sensible sub-
stance there is a form which is the source of its being, and of its being
one.
III
The Potential and
the Actual
7
Aristotle's Notion of
Potentiality in Metaphysics e
MICHAEL FREDE
Aristotle thinks that we can not only make claims of the form 'A is F'
or 'A is an F' in the sense that A is Factually or that A is an actual F.
It also matters greatly to him that we can make claims of the form 'A
is F potentially' or 'A is a potential F'. For he thinks that the physical
world is characterized by the fact that there are such truths about it
which cannot be reduced to, or eliminated in favour of, claims to the
effect that something actually is something or other. To put the matter
differently, Aristotle thinks that there are truths of the form 'A
possibly is F', in some special sense of 'possibly', which cannot be
reduced to truths of the form 'A is actually G'. And given that he
assumes that an actual F is an actual being precisely in so far as it is an
actual F, he correspondingly assumes that there are not only actual
beings, for example, actual houses, actual human beings, actually
healthy things, but that there are also some kinds of possible being.
namely potential beings, for example, potential houses, potential
human beings, potentially healthy things. These potential beings, for
Aristotle, form a crucial part of the ontology of the sensible world.
Naturally one would like to know more precisely under which
conditions, according to Aristotle, something A can be said to be a
potential F, and which kind of possibility is supposed to be involved
when he talks about potential beings. And there is one text in which
Aristotle himself sets out to answer these questions systematically and
in some detail, namely in Met. e. So the obvious thing to do, if we
want to be clear about Aristotle's notion of potentiality, is to consider
this text in the Metaphysics. There are many more remarks in the
Aristotelian Corpus, even in the Metaphysics itself, for example, in
Met. A. 12, which are relevant to the question of Aristotle's notion of
potentiality, but in what follows I will disregard these to.focus on this
one passage in which Aristotle's professed goal is to develop and to
explicate a notion of potentiality.
It might be of some use, though, before we turn to this text, briefly
©Michael Frede 1994
174 Michael Frede
to consider why Aristotle at this point of the Metaphysics naturally

would turn to this topic. There are various reasons for this. The first
of them is trivial: Aristotle in the Metaphysics is concerned with the
question what it is to be a being. And since he thinks that there are
also, ineliminably, potential beings, he naturally thinks that a com-
plete answer to the question, what it is to be a being, will also have to
involve an answer to the question what it is to be a potential, as
opposed to an actual, being, and hence his theory requires some
elucidation of the notions of potentiality and actuality. He himself
says (e. 1, 1045°32-5): 'Since we talk of a being, on the one hand, in
so far as something is a substance or a quality or a quantity, but on
the other hand in accordance with potentiality and actuality ... let us
also make some determinations concerning potentiality and actuality.'
This, then, is an obvious reason for his turning to the topic of poten-
tiality at this point in the Metaphysics, having already dealt, at least in
a preliminary way, with substance and its non-substantial feat.ures
in z.
But there are further reasons why the question of potentiality would
naturally arise for Aristotle at this point in the Metaphysics which are
more intimately connected with the particular way Aristotle wants to
conceptualize reality in the Metaphysics. The first of these, at least in
general terms, is easy to specify. As is well known, Aristotle thinks
that there is no such thing as being, one general feature shared by all
things which are, and that hence there is also no single answer to the
question what it is to be a being. Rather, he thinks that there are basic
beings, namely substances, with their own .particular way of being;
and he thinks that everything else counts as a being only in so far as it
is related to these primary beings in the appropriate way, that, fo!fw
example, there are qualities only in so far as there are substances
which they qualify. So in Met. z he almost immediately begins to
focus on the question what it is to be a substance, what it is to be a
being in the particular way that substances are beings.
To get a handle on the answer to this questjon he beg_ins by con-
sidering the kind of substance we are familiar with, namely ordinary
physical objects. It is a consideration of these which is supposed to
give us at least the beginning of an answer to the question what it is to
be a being in the way that substances are beings. Now, as is also well
known, Aristotle thinks that these objects, sensible substances, need
to be analysed into form and matter. And this raises two problems: (i)
these notions themselves are not only technical, but also controversial;
so they need to be elucidated; and it is clear that Aristotle means to
explicate them in terms of the notions of potentiality and actuality. If
our ordinary object is .an F, then the matter, properly understood, is
Potentiality in Metaphysics 175
that which is potentially an F; the form, on the other hand, is the
actuality of the F-that about it which makes it an actual F, what is
actual about it as an F. This, clearly, is rather obscure and needs
explanation. Hence the need to turn to the notions of potentiality and
actuality.
(ii) Moreover, the assumption that a sensible substance is a com-
posite of matter and form gives rise to a further problem: a substance
is supposed to be a paradigm of something which is one thing, rather
than a heap or a transient compound of things. If it were a mere
composite of items which existed independently of and prior to it, it
could not itself be a basic, primary being. Hence Aristotle in the
course of Met. Z repeatedly insists that a substance cannot be a com-
posite of further items-to be more precise, a composite of further
actual items. The solution to the problem, as we see at the very end of
fl, is supposed to be this: matter and form are not two actual items
which combine to form a further actual item. There are not two items.
Only the form is actual; it is what is actual about the composite F as
such, whereas the matter is only a potential being, that which is an F
potentially. Needless to say, before we accept this resolution of the
problem, we want to learn a lot more about potentiality and actuality.
There is yet a further reason why at this point Aristotle turns to
potentiality and actuality; but this further reason is much more dif-
ficult to specify even in vague and general terms. As we already
noted, Aristotle in Met. Z and fl discusses only sensible substances. It
is obvious that a final answer to the question what it is to be a
substance will have to be one which also fits immaterial substances.
And Aristotle repeatedly in the course of Z indicates that the discus-
sion of sensible substances is only preliminary to a discussion of
immaterial substances. Since we never get the promised discussion of
immaterial substances, it is a matter of speculation precisely in which
way Aristotle thinks that we have ultimately understood what it is to
be a substance only if we have considered the substancehood of im-
material substances. But for various reasons it is tempting to assume
that this is supposed to be so because ultimately only immaterial
substances are primary beings, and hence substances in the narrowest
and strictest sense. If this is correct, there are various ways in which
Aristotle might have wanted to argue for this, of which I will only
mention one: We allow for potential beings, because even potential
beings have some kind of reality. But we also are inclined to say that
actual beings are more real, have more of a claim to be called 'beings',
than potential beings. Moreover, we might understand that somebody
wants to claim that primary, basic beings, those beings on which all
other beings depend for their being, should be most real, should have
Michael Frede
the strongest claim to be called 'beings'. Given this, one can see how
one line of argument may run. In the course of Met. Z Aristotle
argues that substantial forms are more strictly speaking substances
than the composites, the ordinary objects, they are the forms of. Now
the forms are actualities; they are what is actual about a composite F
as such. But though they are actualities, their being is tied to poten-
tiality in two closely related ways: (i) they need the appropriate
matter for their realization; (ii) they themselves, though actualities,
are constituted by a set of potentialities. For, to be a horse, for
example, is essentially to be able to do the kinds of things horses do
under the appropriate conditions. Moreover, given that forms depend
on matter for their realization, they also depend on something else
which brings it about, or which explains, that they get realized in a
certain matter. And this, Aristotle argues, ultimately presupposes that
there is a being which is a pure actuality, something neither tied to
matter, nor constituted by a set of abilities which it only exercises
given the appropriate circumstances. So there is a clear sense in which
the forms of sensible substances are not only less fully real than, but
also posterior to, and dependent on, a pure actuality, which for that
reason has more claim to be considered a primary being than they do.
But whichever the argument may be which Aristotle himself en-
visages, it almost certainly turns on the notions of actuality and poten-
tiality to exploit the fact that immaterial substances, immaterial forms,
are pure actualities, whereas the forms of sensible substances are tied
to potentiality. And so there would be a further important reason for
Aristotle to turn at this point of the Metaphysics to the notion of
actuality and potentiality.
Given the importance these notions have in Aristotle in general,
and given their crucial significance for the argument of the Metaphysics
in particular, we approach Met. e with rather high expectations.
Unfortunately these expectations seem to be disappointed. Atistotle's
discussion seems to be unclear, disorganized, and confusing. In fact,
some important commentators, like Bonitz and Ross, have accused
Aristotle of being himself confused. They have accused him of almost
immediately confusing again the very notions he sets out to distinguish,
namely the notion of potentiality and the notion of some kind of
active power, i.e. the ability something might have to produce a
change in something else.
Now one may oneself feel rather confused by the discussion in Met.
e and still not be prepared to believe this readily that Aristotle
himself should have been so confused about a basic distinction he
wants to make, as to proceed almost immediately and repeatedly to
confound what he meant to distinguish. And it also might give one
pause that, though Bonitz and Ross, each for his own part, seem to
have such a clear idea of what potentiality is supposed to be that they
are not even particularly concerned to elucidate the notion, their
respective ideas of potentiality turn out to be rather different. What is
worse, their characterization of potentiality hardly fits what Aristotle
himself has to say.
For Ross potentiality is primarily the ability of an object to pass
into a new state of or by itself. Now potentiality thus characterized
certainly covers some kinds of cases Aristotle is crucially interested in:
for example, the case of an organism which, given certain conditions,
passes into a new state in the sense that of itself it displays the
behaviour characteristic of its kind; thus it grows to a certain size of
itself, or is seeing something or thinking something, of itself. But it is
clear from Met. e. 7 that Aristotle wants the notion of potentiality
equally to apply to the case where the principle of change is external,
where there is an external agent, where a thing does not pass into a
new state of itself, but due to the agency of something else. For he
distinguishes two kinds of cases in which we can say that something A
is potentially F: (i) cases in which A is turned into an F by an internal
principle, and (ii) cases in which A is turned into an F by an external
principle. And as examples of the latter he adduces (a) somebody who
is potentially healthy in so far as he might be cured by the art of
medicine, and (b) something which is potentially a house in so far as it
is material in exactly the appropriate state to be turned into a building
by the art of building. But these are precisely cases in which what is
potential does not have the ability to pass into a new state of itself;
the patient requires a doctor to become healthy, and the building
material requires a builder. So Ross cannot be right to suppose that
potentiality is essentially the ability of something to pass into a new
state of itself.
Bonitz talks as if he identifies the notion of potentiality with a
notion of 'mere possibility', as he puts it. But surely this cannot be
right, either. For the main point of e. 7 is that it is not sufficient for
something A to be a potential F that there be the mere possibility that
A turn into, or be turned into, an F; if this were so, some earth, air,
fire, and water would be a potential human being; for there is the
mere possibility that, in the course of a long history, it be turned into
a human being; certainly there is nothing contradictory about the
assumption that it does. And yet Aristotle insists in this chapter that a
lot more is needed for something to qualify as a potential human
being. Thus the notion of potentiality must be a lot stronger than the
notion of a mere possibility. So Bonitz's notion of potentiality does
not seem to be adequate, either.
Michael Frede
Given this situation, one comes to suspect that, perhaps, part of the
sense of confusion which one feels when one reads Met. e, is due to
the fact that we tend to approach the text with the wrong expectations.
And part of the reason for this, I submit, is the term 'potentiality'
itself. It is a technical term, a term well-established by a long tradition
of Aristotelianism and Aristotelian scholarship, a term one instinc-
tively feels goes back all the way to Aristotle. All this encourages the
idea that there is this highly technical, more or less well-defined, more
or less clearly conceived, notion central to Aristotelian philosophy
which we expect Aristotle, here in Met. e, to lay out for us.
But we have to remind ourselves that the term 'potentiality' does
not go back to Aristotle, that it seems to come into common use only
very late, and that Aristotle himself, instead, uses the ordinary word
()6va1uc; and its cognates 156vaa0ai and 15vvar6v, words which even
ordinarily, but certainly in Aristotle, have a wide variety of uses or
senses, a situation which apparently very late scholastics try to rectify
by introducing potentialitas for l515vaµ1c; in one of its uses. And perhaps
it is a more realistic view of Met. e that Aristotle himself, now that he
feels forced to give a systematical, general account of the relevant
sense of l56vap1c; in which he wants to say that something is something
or other potentially, or that something is a potential being, does so in
a rather tentative and groping way, trying to distinguish and to clarify
the particular sense of ()1!va1uc; in which he wants to talk of potential
beings.
Let us look at the matter in some more detail. We can do so by
following Aristotle's own exposition in Met. e. 1, 1046a4-19. Aristotle
himself, of course, is quite aware that the word l56vaµ1c; has many uses.
He says so at e. 1, 10463 4-6. Some of these seem to hi:qi to be a
matter of sheer, accidental ambiguity or homonymy, and hence the~e
uses are of no systematical interest for a study of the notion he is
interested in (e. 1, 1046a6-7). An example of this for him is the use
of 156vaµu:; or 'power' in mathematics (1046 3 7-9). But, for the greater
part, Aristotle seems to think, these uses of (fovf!.µ1c; are systematically
related, related in such a way that their relation is of systematical
interest for a study of the corresponding notions. To be more precise,
he thinks that they form a family of uses with a certain internal
structure he also finds in other cases: they are related in such a way
that there is one basic use, and the other uses are all derivative from
this use and have to be explicated in terms of it (e. 1, 10463 9-10). He
identifies this use as the one in which ()(Jvaµzc; refers to the ability of
something to produce a change in something else, for example, the
ability of fire to heat things, or the ability of the doctor to cu;:e
patients (e. 1, 1046a10;-II). And he continues to explain how this use
of the term 15vvaµx; gives rise to other uses, such as the use in which
J15va;ur; refers to the ability something has to undergo a certain change
at the hands of the appropriate agent, for instance the ability of the
patient to be cured by a doctor.
Now it is tempting to assume that the relevant use of t56vaµu:; in
which we talk of potential beings is supposed to be a member of this
family of uses. And this, obviously, is one reason why Aristotle
considers the notion of potentiality in particular in relation to the
basic notion of <5vva1-m;. What we, then, have to see in order to
understand Met. e is how Aristotle distinguishes the basic notion of
()6vaµ1c; and the derived notion of potentiality, and precisely in which
way the notion of potentiality is supposed to be derived from the basic
notion of <fovaµ1c;. For it is precisely this basic notion of <51)vaµ1c; which
Aristotle is supposed to confuse with the notion of potentiality.
We might be better prepared to see whether Aristotle confuses
these notions or whether he actually manages to distinguish them, if
we take note of some simple distinctions. Aristotle here as elsewhere
is not particularly careful when it comes to indicating whether he is
talking about a word, or about a word in a certain use, or about what
is referred to by the word in a certain use, or about what is covered by
a word quite generally, whatever its use. Correspondingly, commen-
tators are often not particularly careful in this regard, either. Hence
they seem to think that when Aristotle distinguishes two or more uses
or senses of the term 01)vaµ1c; he also is distinguishing two or more
kinds of t51)vap1c;, two or more kinds of possibility, two or more kinds
of items in the ontology. Both Bonitz and Ross, for example, talk as if
there were active powers, on the one hand, and potentialities on
the other, as if these were two different kinds of possibility, which
Aristotle was then going to confuse.
Now I think that it is this assumption which is largely responsible
for the confusion. For it seems to me that what Aristotle is really
trying to do is this:
(i) he distinguishes various kinds of 15vvaµrc; and corresponding uses

of the term c51)vap1::;, including a basic kind and a corresponding basic
sense or use of the term 01)vap1c;;
(ii) in addition, he distinguishes a further use of the term which we
might mark by the word 'potentiality', namely the use of the term in
which we talk of something being a potential F and, hence, of a
potential being, a <5Vva1w ov;
(iii) but it is not the case that there is a further kind of J6vap1::;
which corresponds to this further use of the word c51!vap1::;; it is rather
the case that the term J1)va1nc; in this use ranges over various kinds of
180 Michael Frede
i5vvaµu:; we have already distinguished, among them the basic kind of
i5vvap11:;;
(iv) for example, the basic kind of i5vvaµ11:; itself, taken or under-
stood in a certain way, can be called a i5vvaµ1c; in the sense of poten-
tiality, with reference to which something can be said to be something
or other potentially, or can be said to be a potential being.
So it is not surprising, and by no means a sign of confusion, if
Aristotle, in talking about the basic kind of i5bvap1c;, at times talks as
if he were talking about potentiality. For potentiality is not a further,
distinct kind of i51)vaµ1c;, but just one or another of the kinds of i5vvaµ1c;
already distinguished, understood in a certain way.
Thus prepared, let us, at long last, having already considered e,
10463 4-19, turn to the remarks with which Aristotle himself intro-
duces his discussion of potentiality in Met. e. This is what he says (e.
I, 1045b34- 10463 4):
Let us also make determinations concerning t56vaµ1r; and actuality, and here,
first of all, concerning the kind of t56va111r; which is so-called most strictly
speaking, but which is not most useful for our present purposes. For poten-
tiality and actuality extend beyond the kind of potentiality and the kind of
actuality which are so-called solely with reference to some change. But,
having talked about this kind, we will in our discussions concerning actuality
also clarify the rest.
Aristotle here seems to be telling us that he is first going to~discuss
the basic kind of i5vvaµ1c;, to turn then to actuality and then, finally, in
the course of the discussion of actuality, to turn to potentiality. We
may wonder why he wants to postpone the discussion of potentiality
till we have come to actuality, but the programme itself seems to be
clear enough, and it also seems to be the one his discussion actually
follows. For chapters 1-5 deal with i5vvap1c; in the basic sense. In
chapter 6 Aristotle turns to actuality and, in the course of this discus-
sion, begins to talk about potentiality. In fact, chapter 7 is devoted to
the question of when something can be said to be something or other
potentially, and chapters 8-9 are devoted to a discussion of the
priority and greater worthiness of actuality over i5vvap1c;.
One sentence in the sketch of the programme quoted, though,
deserves more detailed consideration. It is the sentence 10463 1-2:
'For i5vvap1c; and actuality extend beyond the kind of i5vvap1c; and the
kind of actuality which are so-called solely in virtue of (or: with
reference to] some change.' Aristotle is making two claims, one con-
cerning potentiality and one concerning actuality. I want to proceed
by first considering the second claim and then try to construe the first
claim analogously.
Aristotle, then, is saying that actuality extends beyond the kind of

actuality which is solely so-called in virtue of some change. There are
two questions here: (i) what is the reference of 'actuality' in its first
occurrence? and (ii) what does Aristotle have in mind when he talks
about the kind of actuality which is so-called solely in virtue of some
change? Given the context, we should assume that the answer to the
first question is: actuality in the use or sense of 'actuality' in which we
talk of actual beings. As to the second question, an obscure remark in
e. 3, 1047"30- 0 2 may help. However obscure the remark may be,
Aristotle here clearly distinguishes two uses of the term 'actuality', a
narrower use in which it primarily refers to changes, and a further use
in which it seems to have the sense of 'actuality' here in question,
since he paraphrases 'actuality' in this use with evuJixaa, a tech-
nical Aristotelian term to mark actuality as opposed to potentiality
(1047a30-1). And Aristotle clearly says that the second use is arrived
at by extending the use of the term 'actuality' from a primary use in
connection with changes to other kinds or cases of actuality. So it is
tempting to think that it is precisely this which he has in mind in e. r,
1046a1-2: actuality, namely what is covered by the extended use of
the term, extends beyond what is covered by the term when it refers
primarily to changes.
To understand this better it may help to take into account that
Aristotle is actually using two terms for 'actuality': (i) evepyeza and (ii)
evreJixeza. The first word tends generally to be used by Aristotle for
doings, things something does, like pushing something, melting some-
thing, reading a book, cutting a tree, puzzling over something, seeing
something, and the like. Hence it is often also translated as 'activity'.
Another thing to keep in mind here is that there is a tradition in
Greek thought, which surfaces particularly clearly in Plato's Sophist or
later in the Stoics, according to which something which does some-
thing or other to something, or something which has something done
to it, must clearly be real, must have a certain degree of reality. What
is more, it is thought that the mere ability to affect other things, or to
be affected by them, presupposes some kind of reality. Thus Plato in
the Sophist offers to the Giants, as something they might agree on,
the suggestion that the ability to affect something or to be affected is
the mark of the real. And it must be against this kind of background
that Aristotle here, too, explains the extension of the narrow use of
tvepyeza, in which it primarily refers to changes, to cases of actuality
quite generally, relying on the close connection we even ordinarily see
between change and reality; he points out that one thing people quite
definitely do not want to attribute to things which do not exist is
change. Something which does not exist at all cannot be changing
r82 Michael Frede
something else or be undergoing a change at the hands of something

else (1047a32-b1). For, to claim that it does, would be to attribute
actuality to it.
Thus what Aristotle seems to have in mind ate. I, 1046a1~2, when
he talks about a kind of actuality which is so-called solely in virtue of
some change, seems to be precisely the kind of actuality in the sense
of the term iviipy1;w in which it primarily refers to changes. Looked at
just as changes brought about by something, as doings or sufferings,
we could call them evepyezai in the sense of 'activities'. But there is
another way of looking at them. Doings and sufferings are also a mark
of the real; indeed they can be regarded as themselves ways of being
real, more precisely of being actual. After all, for Aristotle, to be
heating something, or to be building something, is itself a way or a
form of being real, of being actual. And looked at in this way these
changes can be called tvepyc.1w in the sense of 'actualities' or, if we
want to avoid ambiguity, ivrel.Bxmm. But there are, of course, other
forms or ways of being actual, other kinds of actuality. And so the use
of the term 'actuality' in the sense in which we talk of actual beings
extends beyond the use of the term in which it refers primarily to
doings or changes.
This, then, is what Aristotle seems to be saying in the sentence in
e. I, 1046a1-2 under consideration. But what he has in mind b'ecomes
still clearer if we look at the corresponding section in the actual
discussion of actuality, namely e. 6. In particular, it becomes clearer
that 'actuality' in the sense which we are trying to elucidate does not
refer to a distinct, special kind of actuality which Aristotle is trying to
distinguish from the actuality of changes, or rather of things changing;
it is not itself a further kind of actuality. It is rather the case that the
use of the term we are interested in is a further use which covers both
changes and other kinds of actuality. Aristotle in e. 6 is very con-
cerned to distinguish various kinds of actuality, in particular two kinds
of doings, a distinction the entire second part of the chapter is devoted
to. It is a distinction between changes and other kinds of doing-
doings like seeing, intellectually grasping, or the like-which are not
changes. So, if we follow the intuition that anything which is actually
doing something or other must as such be real and, what is more,
actual, it is evident that the term 'actuality' in its relevant use extends'
beyond mere changes to items like seeing, intellectually grasping,
living, and the like. What is more, it is clear from this chapter that
there are yet other forms or kinds of actuality in the relevant sense:
for instance, substantial forms (cf. 1048b9). For, clearly, being a house
or being a human being is a way of being actual as much as, if not
more so than, building a house.
Potentiality in Metaphysics
But let us note that the idea in e. 6 is not that there is a kind
of actuality which is change and another kind of actuality which is
actuality in the sense we are trying to clarify. It is rather the case that
there is kind of actuality which is change, and that there are other
kinds or forms of actuality which equally deserve to be so-called by
extension. But, as Aristotle emphatically insists (cf. e. 6, 1048a35-b9),
there is no single thing which we would be referring to when we talk
of different kinds of items as actualities in the sense we are after.
There is no single thing which all the items called 'actuality' in the
relevant sense have in common and which we could capture and
specify by a definition. They are merely so-called by analogy. When
we say that something is an actual house, and when we say that
something is actually building a house, or is an actual builder, we have
one and the same use of ' ... is an actual ... ' or ' ... is actually ... '.
But what corresponds to it is not some single, elusive kind of actuality
or reality to be found in both cases. Rather, what corresponds to this
single use are different kinds of actuality; the form of a house, which
is the actuality referred to in the first case, is a completely different
kind of actuality from the actuality referred to in the second case,
which is a change. Nevertheless, the force of ' ... is actually ... ' or
' ... is an actual .. .' in both cases is the same; there is in both cases
the same unified use of the word 'actual' or 'actually'. In both cases it
is claimed that a potentiality is realized and that, given this realization,
the object in question has achieved a certain degree of reality, namely,
actuality. And the basis for this uniform use is the fact that the form
of the house is the analogue of the act of building a house in that it is
related to the matter of a house and to a house in the way in which
the act of building is related to the ability to build and to the builder.
It is this analogy which allows for a new, uniform use of the term
'actuality' across different, but analogous kinds of actuality. What
gives an actual house its actuality is the form, just as the act of
building gives the actual builder his actuality.
It seems to me that there is no way to read e. 6 in such a way that
Aristotle would be distinguishing different forms or kinds of actuality
and singling one of them out as the kind of actuality which we are
concerned with when we talk about actual beings. It seems clear,
rather, that he wants to say that there is no single kind of actuality in
virtue of which all actual beings are actual, but that there is a use of
'actuality' which extends analogously to the various kinds of actuality.
Against this background we can now turn to the other part of the
claim in e. 1, 1046a I -2: <5vva;uc;; extends beyond the kind of <>vvap1:;
which is solely so-called with reference to some change. Again, <51)vaw:;
in its first occurrence here should refer to <5vva;w; in the sense in which
Michael Frede
we speak of potential beings. And the kind of <>6vaµH; which is solely

so-called with reference to some change should be none .other than
i56vaµ1c; in its basic sense of an active power, given the context, and
given the language of e. 6, 1048a26, where Aristotle, too, talks of the
kind of <>vva1-uc; which is so-called with reference to change and by this
seems to refer to i515vap1c; in the primary sense. So Aristotle would be
claiming that i56va1-uc; in the relevant sense extends beyond <>15va1-uc; in
the basic sense. The claim, hence, is not that besides i56va1-uc; in its
basic sense there is another kind of <fova1-uc;, namely, potentiality. The
claim is rather that <56vaµ1c; in the sense of 'potentiality' covers <51!va1uc;
in the basic sense, but extends to other kinds of '56vap1c;. The text,
neither here nor in e. 6-9, warrants the assumption that potentiality
is itself a further kind of i51iva1-11c;, and the analogy with actuality
suggests the contrary.
Let us look at this in more detail. We should already now be able to
see how i515vaµ1c; in the sense of potentiality could be thought to cover
i51!vaµ1c; in its basic sense. For we already noted earlier that Aristotle
himself relies on a more widespread notion that to attribute a Joing to
something is to attribute a degree of reality to it, namely, some
actuality; moreover, we noted that there was the further view that the
mere ability to produce a change, to affect things, conferred some
degree of reality on what has this ability: things which do not exist at
all cannot cure patients; what can cure patients must have some
degree of reality. And so it is not surprising that Aristotle, too, should
think that a basic i56vaµ1c; can be thought to confer some minimal
degree of reality on what has it, indeed that having it is a way of being
real, namely to be a potential, rather than an actual being. But there
are other kinds of i56vaµ1c;, and they, too, can be called i5vvaµ1c; in this
further use or sense, since they, too, confer some degree of reality on
what has them in an analogous fashion. Thus there is the kind of
bvvaµu; which consists in the ability to undergo a change, and, as we
already saw, Aristotle and others quite plausibly think that this kind
of 06vaµ1c;, too, confers a certain degree of reality on what has such an
ability. Something which did not exist at all could not be able to
undergo a change.
But this matter may become clearer if we look again at the part of
Aristotle's actual discussion which corresponds to the programmatic
sta·tement in e. 1, 1046a 1 -2 which we have been considering, namely
e. 6, 1048a27-30. Aristotle says:
For making the appropriate distinctions [sc. concerning actuality] the poten-
tial, too, will at the same time become clear, namely that we do not just call
'potential' that which has it in its nature to change something else or to be
changed by something else, either unqualifiedly or in a certain way, but that
we also call things 'potential' quite differently.
We should note that he is not saying that the potential is a further

kind of ovvarbv besides the ones listed. Rather he is saying that
something qualifies as potential for quite different reasons, either
because it has an active power or because it has a passive ovvaµ1c;, or in
an altogether different way. He does not say that only those things
count as potential which are potential in this different way which we
now have to elucidate. Aristotle's remark about the potential amounts
to the claim that the term bvvaµrc; in the sense of 'potentiality' covers
not only the basic. kind of ovvaµzc;, its passive counterpart, and, of
course, particular modes of the two, but also a quite different kind of
t56vaµ1c; or different kinds of 06vaµ1c;. And what he has in mind is
readily apparent, if we heed his remark that this will be clear once we
make the appropriate distinctions concerning actuality. Now the major
distinction made in chapter 6 is the one already alluded to, that
between two kinds of doings, namely changes and things like seeing.
Now the ovvaµzc; which corresponds to the actual building of a house is
the active ability to build a house. But there is also an ability which
corresponds to actually seeing, namely the ability to see. And this is
neither the basic kind of active <5vvaµrc;, nor its passive counterpart,
nor a mode of either. For seeing is not a change. And yet this ability,
too, has a corresponding actuality and analogously confers a degree of
reality on what has this ability. So it, too, can be called a <~vvaµ1c; in the
sense of 'potentiality'.
If, moreover, we take into account that Aristotle in e. 6 insists that
there is no definition for either actuality or bvvaµrc; in the sense we are
interested in, that instead these terms in the relevant use apply only
analogously, the following picture emerges:
There are different kinds of ovvap1c; and corresponding kinds of
actuality, including the basic kind of <lVva1nc; and the corresponding
actuality which is a change. But these different kinds of t51)va1nc; can be
taken and understood in a certain way: They can be understood as
each in an analogous way conferring a certain degree of reality on its
bearer, which in each case is to be understood by contrast with the
corresponding kind or form of actuality. And in so far as we consider
a kind of 01)vaµ1c; in this way we consider it as a potentiality. Corre-
spondingly with actuality. There are different kinds of actuality each
of which in an analogous way confers a degree of reality upon its
bearer, which in each case has to be understood in contrast to the
corresponding <56vaµ1c;. And in so far as an actuality is regarded in this
way it is called an 'actuality' in the relevant sense.
186 Michael Frede
It now becomes clear also why Aristotle wants to consider poten-

tiality only once he has begun with a consideration of actuality. In the
case of actuality it is abundantly clear that there are other forms
or kinds of actuality than change. Indeed, if we press the matter,
according to Aristotle, change barely qualifies as an actuality. It is
rather an incomplete or imperfect actuality. But once this is clear it is
also clear that potentiality extends beyond <56vaµ1r;, in the basic sense,
indeed beyond any kind of <56vaµ1r;, which is merely a <56vaµ1r;, for
change in something else or by something else. And once this is clear,
it is also evident that there must be two quite different uses or senses
of the term <56vaµ1r;,, or, to put the matter differently, that the notion Off
potentiality is different from the notion of a power to change some-
thing else or the broader notion of a power to change or be changed
by something else. But this notion of potentiality is not the notion of a
further kind of <561wur;,. Rather, it is a notion of these very kinds of
(56vaµ1r;, we have already distinguished taken in a certain way, namely
as corresponding analogously to an appropriate kind of actuatity and
as, thus contrasted, conferring a degree of reality on their bearers.
It seems that if we understand Aristotle's remarks about potentiality
in this way, it is easier to make sense of some passages which other-
wise are difficult, if not impossible, to explain. It should, for example,
strike us as remarkable that Aristotle in Met. L1. 12, where he tries to
list the more important senses of <56vaµ1r;,, does list various kinds of
06vaµ1r;,, including the basic one, but does not have a separate entry for
potentiality. This would be extremely strange if Aristotle thought of
potentiality as a further kind of 06vaµ1r;,, distinct from the ones already
listed, but is easy to explain if we assume that potentiality just is
certain kinds of J6iwur;, taken in a certain way.
And there are other passages, it seems to me, which only make
sense if we do not assume that potentiality for Aristotle is a further
kind of 06vap1:;. Thus, I think, we can make sense of chapters 8 and 9
of Met. e only if we assume that potentiality is just a certain ()r!va1-a:;,
already distinguished, taken in a certain way. For in e. 8 and 9
Aristotle proposes to compare actuality and o6vaµz:; in various regards,
for example, as to priority. Now by t56va1-ur;, here, given the position of
these chapters in the book, and given the cases these chapters discuss,
Aristotle must mean 'potentiality'. But let us consider the way he
introduces the discussion. He says (e. 8, 1049b5-rn):
It is obvious that actuality is prior to Juvaf11:;. And by <)1!va111;· I do not mean
just the one defined which is called a principle of change in something else or
in itself qua something else, but quite generally any principle of change or
rest. For a nature, too, belongs to the same family as a 01!vuf1ts [sc. in the
sense defined as a principle of change in something else]. For it, too, is a
• I 'I , • , I ~ ' -. • '""' '
So Aristotle, in specifying the scope of what he wants to talk about

when he says that actuality is prior to potentiality, does not specify a
certain further kind of 06va1-m;, but refers to kinds of o6vaµu; we are
already familiar with, 06vaµ1; in the basic sense, and other kinds of
06vaµ1; which are so-called derivatively, because they, too, are prin-
ciples of change, for example, natures. They all can be regarded as
potentialities, if they are taken in a certain way, namely as related
analogously to the appropriate kind of actuality. Unless we understand
Aristotle in this way, the discussion in e. 8-9 will not be about
potentiality, as it should be, given the structure and composition of
the book.
So, to sum up our discussion so far: There is a basic sense of o6vapu;
in which the active power to produce a change in something else is
called a o6vapu;. There are derivative uses, in particular the use in
which something which has the capacity to undergo a change can be
called a 06vaµ1;, but also the use in which a nature, or the use in
which, for example, the faculty to perceive, which is constitutive of a
nature, can be called a or5vaµz;. To these correspond different kinds of
o6vaµu;, different kinds of items in the ontology. But they can all be
taken analogously to each other in a certain way, and thus taken, they
are called 'potentialities'.
If we approach the text with this understanding, Aristotle no longer
seems to confuse the very notions he is trying to distinguish, and his
discussion seems to gain in order and coherence. But there are still
many details which remain obscure, and it is to some of these that I
will turn now, hoping in this way to clarify further the question of the
precise relation between the notion of a 06vaµ1; in the basic sense of
the term and the notion of potentiality.
In e. 7 Aristotle turns to the question of when we can say that
something is an F potentially. A problem he is particularly concerned
with and tries to address here is the following: Somebody might think
that if we have some material which in the course of a long history
could turn into, or could be turned into, an object of the kind F, then
this material constitutes the matter of an F and thus a potential F,
something which is potentially an F. This is what Aristotle wants to
deny. The conditions on potentiality have to be much more stringent,
if the notion is to be helpful, that is, is to have the explanatory power
Aristotle wants it to have. The question is where we draw the line
short of only allowing for Fs which are already actual: at what point
can we begin to talk of a potential F?
Aristotle distinguishes two kinds of cases, (i) cases in which some-
thing turns, or is turned, into an F because there is an external agent
who has the ability to produce this kind of change in something else,
.1 • • ... .. • 'I ........ •
188 Michael Frede
which something turns into an F because it itself has the appropriate

ability to change itself in this way, as when, for example, plants in the
course of their natural development grow to a certain size. It is not an
external agent which makes them grow to this size, it is rather an
internal principle which accounts for their growth to a certain size
given the appropriate external conditions.
Aristotle begins his consideration of the first kind of case with the
following remarks: 'Similarly it is not the case that just anything could
be cured by the art of medicine or by chance, but there is somethiri~
for which it is possible to be cured, and it is this which is potentially
healthy' (1045a3-5). I think that the possibility of spontaneous cures
does not affect the issue, and so, for the sake of simplicity, I will
proceed as if Aristotle claimed that the only way for a patient to be
cured is to be treated by a doctor.
Given our earlier discussion, it is worth noting in passing""that the
potentiality in this case does not seem to be anything else but the
passive capacity of the patient to undergo a certain kind of change
brought about by a doctor, understood, though, as conferring a certain
reality on the patient as a potentially healthy being. But setting this
aside, let us see what Aristotle may have in mind when he says that
not just anything can be cured by medicine. I think that there are four
kinds of cases we need to distinguish:
(i) Of course, for something to be cured by medicine it has to be in

the first place the kind of thing which can be ill or healthy. Stones
cannot be cured, but human beings, for instance, can. For they are
the kind of thing which can be healthy. But this sense of 'can' in 'can
be healthy' obviously is much weaker than the sense of 'can' which we
need if we want to capture the force of 'A is potentially healthy'. We
see this immediately if we look at a second and a third kind of case.
(ii) There are living beings which, as such, can be healthy; but the
particular specimens we are dealing with happen to be ill in such a
way that there is no possibility that they could ever be healthy again.
They are irreparably or perhaps even fatally ill. Obviously they are
not potentially healthy, either, though at least they are the kind of
thing which can be healthy.
(iii) There are living things which are ill in such a way that no art of
medicine, however advanced it may be, could cure them, given the
state they are in. But there might be a state which they could be in, a
state which is causally attainable by them, such that, if they were in
that state, they could be cured. If might be the case, for instance, that
they still have to grow to a certain size or weight to be able to undergo
a certain operation. So they are not ill in such a way that they could
never again he healthy. They have not irretrievably lost their health.
They can be healthy again. But this sense of 'can', too, is still too
weak to capture the force of potentiality. For they are not yet poten-
tially healthy, since, given the state they are actually in, they cannot
be cured by the art of medicine.
(iv) Finally, there are living things which are ill, but ill in such a
way that there is nothing about them which would prevent them from
being cured by the art of medicine. And it is only these things which
are potentially healthy.
This distinction of kinds of case and its relevance may become
even clearer if we consider the other example Aristotle addresses, that
of a house. This is what he has to say about that case (ro49a8-II): 'In
a similar way a house is potentially. If nothing in this, i.e. in the
matter, prevents it from becoming a house, and if there is nothing
which has to have been added or been taken away or to have changed.
then this is a house potentially.' It will again simplify matters, but not
unduly, if we suppose that a house is entirely made of bricks, nothing
but bricks, and that these bricks can be put together in the appropriate
way so as to form a house. When, then, do we have a potential
house? In this case, again, we may distinguish four situations:
(i) We need the right kind of material; not just any kind of material
will do to build a house. It must be the kind of material which can
constitute a house, for instance, bricks.
(ii) Suppose, then, that we do have bricks. This is not enough. For
the bricks may be irremediably flawed. They may have been baked in
such a way that they collapse under the kind of pressure they have to
be able to withstand when they form a house. So such bricks are not a
potential house, either. As Aristotle puts it, there is something about
them which prevents them from becoming a house, just as if they
were not the right kind of stuff in the first place.
(iii) The bricks may not be irremediably flawed, but they still might
not be in the right kind of state to be turned into a house by the art of
building. They might, for example, still be unbaked. So there can be a
history in the course of which they come to form a house. But, in the
state they now still happen to be in, no art of building can turn them
into a house. They still need to be baked. But this is not a matter of
the art of building, nor part of the process of building a house, but a
matter of some other art. It is this sort of thing Aristotle seems to
have in mind when he says that a given material counts as a potential
house only if it does not require further change before it is turned into
a house.
(iv) Finally, we have the case in which we have not only bricks, but
Michael Frede
bricks in exactly the state they need to be in to be turned into a house

by the art of building. And it is only in this case that the bricks count
as a potential house.
We may again note in passing that here, too, the poten.tiality does
not seem to be anything in addition to the passive bvvaµu; of the bricks
to be turned into a house, taken as being constitutive of a potential
house and thus conferring a certain reality on them.
But what also emerges from a consideration of these different kinds
of cases is (i) that what counts as a potential F here, in the case of
some material with a passive capacity to undergo a change, is deter-
mined by what single process or change there is which could turn a
potential F into an actual F, and that (ii) this in turn depeu,.ds on
which active principles of change there are which would govern and
explain such a change or process. To put the matter crudely, there
being potential houses depends on there being such a process as the
building of a house, and this in turn depends on there being such a
thing as the art of building. Thus, at least in this kind of case, we
can see how a passive power to be turned into a house and hence
potentiality presupposes the basic kind of ovvaµ1c;' the active power to
bring about a change in something else; that is, we can see why
Aristotle attributes such prominence-indeed primacy-to the active
power to produce a change in something else.
So for something to be a potential F it is not good enough that in a
long series of changes it finally can end up as an F. There must rather
be a specifiable single change or process which would turn it into an F.
And it thus itself has to be such that this process or change could turn
it into an F. Hence a tree in the woods will not count as a potential
table, nor even the part of it which actually in the end turns up as
a table. There is a process of making tables, and this process does
not begin with trees in the wood, but with the appropriate pieces of
wood.
Thus the potentiality is a potentiality relat~ve not to some indefinite
possible series of changes which finally terminates in an F. It is
relative rather to a specifiable single change or process. And this, of
course, is reflected by the fact that the passive capacity, which itself is
the potentiality, is the capacity to undergo a certain change brought
about by something which can bring about this change.
But this, too, in turn deserves to be looked at more closely. Let us
assume that we have the bricks which could be used to build a house if
they are moved about in the appropriate way till they form a house.
Obviously they could not be turned into a house unless they could be
moved about in pr~cisely this way. But let us now consider these
bricks and think of all the different ways they could be moved about
and all the different structures they could form as a result of being
moved around in different ways. They, indeed, have the passive
capacity to be moved in all these ways, so as, in the end, to form one
of these structures. Certainly this capacity is not the capacity to be
built into a house. It is true that it is part of this capacity to be able to
be moved about in exactly the way in which they would be moved
about if they were to be turned into a house. But even to think of
them as having this particular ability is not yet to think of them as
having the capacity to be turned into a house. For they might actually
be moved about in this way without this being a case of house-
building. It might be by sheer accident that they happen to be moved
in this way. What makes this being moved in this way a case of house-
building is not their being moved in this way, but the fact that their
being moved in this way constitutes an exercise of the art of building.
And this does not mean that in the case of house-building something
more, or something else, happens that does not happen when the
bricks get moved precisely in this way only by chance. It only means
that what happens has to be explained in a certain way, namely in
terms of the art of building.
So, the suggestion is, there is such a thing as the process of building
a house, only because there is such a thing as the art of building, the
active power to turn something into a house. And it is only because
there is the art of building and hence the process of building that we
can identify the ability of the bricks to be moved about in a certain
way as the passive ability to be turned into a house. It might be in this
way, then, that the passive ability to be turned into a house by
something else is not the coequal counterpart of the active ability of
something to turn something else into a house. It may be in this way
that the active ability is supposed to be more basic than its passive
counter-part, a claim which at first seems counter-intuitive, since one
is inclined to think that the active Jvvaµu; and the passive 01!\!(xpu;
presuppose each other. Indeed, Aristotle himself, towards the end of
e. 1, rn46a19-29, goes to considerable lengths to explain that in a
way the active 01!vaµ[(; and the passive 01!vaµ1i; are one <5vvaµ1i;.
But if we assume that this potentiality for a house is none other
than the passive capacity of the material to be turned into a house by
an exercise of the art of building, a further problem arises. Something,
some material, is a potential house if it is in such a state that it can be
turned into a house by the art of building, that is, if it has a certain
passive ovvaµ1i;. But, one will point out, Aristotle is committed to the
view that an actual house can be analysed into matter and form, and
hence, to the view that an actual. house is constituted by a potential
192 Michael Frede
house and the actuality of a house. But if the potentiality to be a

house is none other than the passive <5vvaµ1r; to be turned into a house,
and the material has already been turned into a house, how can it be
said to be still a potential house? The answer is easy: the matter, of
course, can no longer be turned into this house, but it retains the
ability to be turned into a house, another house. Though this is clear
enough, it, too, deserves to be considered in some more detail. Even
when turned into a house, it is crucial for the continued existence of
the house that the material retain its ability to be a house. And it
seems that it retains this ability precisely as long as it satisfies the
conditions which made it the appropriate material to be turned i~to a
house in the first place. Thus, if the bricks begin to crumble, they
might well be tending toward a state which would have disqualified
them from being a potential house even before they were turned into
a house, exactly to the extent that this crumbling now threatens the
continued existence of the house. Similarly, if we have an actually
healthy person, what underlies the health-the person independently
of being healthy-remains potentially healthy even after having been
cured by a doctor, namely in so far as he continues to be in a state
such that, if he were to be ill, he could still be cured by a doctor. This
is a specifiable condition which a person could lose. For, as we have
seen, a person could not only fall ill in such a way as to be in the
present condition incurable, but even in such a way as to be irreparably
or fatally ill. In either case he would no longer be potentially healthy.
But a healthy person retains this potentiality; in fact, it is a pre-
condition for his health. To fall ill is to lose one's health, but not
necessarily one's ability or potentiality to be healthy.
Perhaps we can now say more precisely how the use of <51!vaµ1r; in
the relevant sense of potentiality presupposes the primary use of
<>vvaµ1r;. It does so somewhat differently from the way in which, for
example, the use of <5vvaµ1r; in the sense of a passive power does. For it
is dependent on the primary use in a twofold way: (i) ·to understand
this use we have to see that <5vva1ur; in this use covers kinds of M>va1ur;
which are called M;vaµ1r; in a way to be explained by the primary use;
(ii) we have to understand the primary use because it itself naturally
gives rise to the further use of <5vvaµ1r; in the sense of potentiality-a
use, though, in which the term by natural extension applies to other
kinds of <56vaµ1r;, as well. And this should go a long way towards
explaining why an Aristotelian discussion devoted to potentiality
devotes the first five chapters, and thus the major part of the discus-
sion, to the basic kind of <5vvaµ1r;-a fact which originally, when we did
not see the strong connection between <56vaµ1r; in the basic sense and
<56vaµ1r;, in the sense of potentiality may have contributed significantly
to our sense of disappointment to hear so little about potentiality itself

in a treatise which was supposed to be devoted to the topic.
Thus it seems, to return to our point of departure, that the sense of
confusion with which Met. e tends to leave us is, indeed, to a con-
siderable extent due to the fact that we tend to approach the text with
well-entrenched preconceptions, in particular the preconception that
potentiality is a certain kind of item is the ontology, a distinct, some-
what mysterious, kind of possibility. Even if we can free ourselves
from these traditional preconceptions associated with the term 'poten-
tiality' and let ourselves be guided by Aristotle's own words, the text
still does not become as transparent and lucid as we might wish. There
still remain a great number of difficulties. But these now no longer
seem to reflect Aristotle's own confusion, but rather the difficulties he
encounters in trying to develop his position on the general question of
what it is to be a potential being.
8
The Activity of Being in

AR YEH KOSMAN
I
This essay is about being active and active being in Aristotle's onto-
logy; that is, it is about the role in his ontology of the concept
represented by the Greek term evepyeia (as well as its correspondent
term bvvaµ1i;;). In speaking of Aristotle's ontology, I mean to be speak-
ing of the enterprise represented primarily in the collection of treatises
we know as the Metaphysics. Aristotle has no term that is the exact
equivalent of ontology or for that matter of metaphysics; he calls the
enterprise carried out in the treatise we have come to call the Meta-
physics 'first philosophy'. But it is clear that this is ontology, for he
describes it as a science (a theoretical discipline) concerned with the
nature of being.
The question concerning being is for Aristotle a question of great
importance and profundity, but its origins are quotidian. In con-
sidering it, he wants to know about the nature of that most universal
and fundamental feature of the universe that we invoke when we
remark upon the fact that the window is open or closed, that we are
seated and are in this room, that Abramowitz is a human being or is
pale, that life is short, art is long, and fate is fickle. It is the being
involved in these and all similar predicative facts (which means. of
course, everything) whose structure and nature Aristotle endeavours
to understand.
The question of the Metaphysics, therefore, concerns being as such,
that is, in general: what Aristotle calls being qua being. For Aristotle
this question is only secondarily concerned with the extensional ques-
tion of what sorts of things there are, or even what sorts of things are
fundamental. Its primary concern is with the nature of the being of
those things, and specifically with the nature of the being that charac-
terizes fundamental things, that mode of being that all call substance-
© Aryeh Kosman r994

Aryeh Kosman
oi)aia-and that is said to be the explanatory principle of being in

general.
On Aristotle's view that mode of being is the being exhibited in
~91)le.thing.:S.J2.~il11LttlH!LJLi,¥, rather than, say, where it is or 'what
colour it is or how large it is. This mode of being is best understood,
he argues, in terms of the notion of evipys1a; near the end of the long
and intricate argument that consitutes books z, H, and e of the
Metaphysics, he makes clear that the concept of cvipysia is central to
his developed understanding of substantial being and therefore, given
the claims of the Metaphysics, to his understanding of being in general.
It has long been standard to Anglicize tlvipyt:ia as actuality and its
correspondent term <5viww:; as potentiality. This choice, as I will sug-
gest, mistakenly makes notions of otherness, change, and becoming
central, rather than what are for Aristotle the ontologically primary
notiO!!LQL§~lf:i.9J~J!lityJ:1.ncl being. In these pages, I want to sketch a
theory of evip_r~~.il.§...£!.<;!ivity and J6!J.'::!P.15: .~S..-~IJ_i!!!Y...S?r .JJ.9wer, and
suggest why the notion of activity is key to Aristotle's ontology, and
why it behoves us to read that OQtology as one for which a central and
governing concept is that of the activity of being.
II
Aristotle's discussion of evipyf;la is found largely in book e of the
Metaphysics. It is a frustrating discussion; the remarks immediately
preceding it lead us to hope that a proper understanding bf the nature
of activity will solve the problem of the unity of substance, a problem
that has emerged with particular clarity in the last several chapters of
book Z, although it has governed its developing argument throughout.
The issue of substance's unity arises directly out of the inquiry in
book Z of the Metaphysics into the nature of substance. Aristotle
attempts there to answer the question, What is substance? by reference
to the several candidates that recommend themselves on the basis of
our most obvious philosophical intuitions concerning that mode of
being. But these attempts all fail, and for importantly related reasons.
In chapter 3 of that book, Aristotle considers the consequences of
our earliest recognition that substance is connected to the notion of
being a subject, that it is, as we learn from the discussion in the
Categories, not a mode of being said of a subject or present in a
subject, but the subject itself of which accidental being is predicated
and in which modes of dependent being are present. Chapter 3 records
the attempt to build upon this fact that substance is a mode of being
of the ontological subject an account of substance as essentially and
merely subject. But this attempt is rapidly shown to be deficient in its
The Activity of Being in Metaphysics 197
failure to account for the determinate being of substance; for on this

account, substance becomes, as it was later to be understood by such
philosophers as Locke, substratum, that is, matter: the indeterminate
something, I know not what, that underlies being rather than the
determinate mode of being that Aristotle wishes to make central to
his ontology.
This deficiency, as it is later to emerge, is dialectical and temporary;
the refusal to accept an account of substance as substratum or indeed
as matter is linked to an immature understanding of substratum and
matter, and it is importantly qualified before the end of book e. But it
is an important deficiency, for it constitutes a failure to recognize what
is, on Aristotle's account throughout the Metaphysics and already in
the Categories, a central ontological fact: the fact that determinate
being is the condition of the openness to further determination that
alone allows an entity to serve as logical or ontic subject. An instance
of substance is, as Aristotle puts the point differently time and again
in his discussion, both a this and a what, and its being a what is a
condition of the possibility of its being a this.
If we think in these terms and pose the problem concerning sub-
stratum in relation to the necessity of things being what they are, it
will become clear that the move in chapter 4 to a discussion of
essence, of an entity's being what it is, is not simply a move to the
next item on an arbitrary list; it is the next step in an argument whose
dialectical progression leads naturally from haecceity to quiddity, from
thisness to whatness. The attempt in chapters 4 to 6 to give an account
of substance strictly in terms of quiddity, however, proves unsatis-
factory precisely because of this dialectical relationship, precisely be-
cause of its failure to account sufficiently for the individuality of
substance.
The relation between substance as substratum and substance as
essence, which may be thought of as the dialectical poles of Aristotle's
argument, is more generally the relation between subject and pre-
dicate in the structure of being, the relation between something that is
a being, and the being that that something is. This relation in turn
parallels the better known relation central to Aristotle's discussion,
that between matter and form. Matter and form are often thought of
in the context of change and becoming: the matter of some entity, say
a cloak, is what the cloak is made out of, while the form is that which
makes what the cloak is made out of into, specifically, a cloak. But
there is a more general account of matter and form that is appropriate
to the larger and prior context of being; the matter of a cloak is that
which is the cloak, the form that by virtue of which that which is the
cloak is a cloak. These are the terms in which the discussion of the
Aryeh Kosman
varieties of form in book H of the Metaphysics is cast. The matter·-Of a

lintel, for instance, is the oak beam that is (as we might say, busy
being) a lintel. But its form, that which makes it a lintel rather than,
say, a threshold, is the position in which the beam is placed, above the
door frame rather than, as with a threshold, below it.
Hyperthyrology is not, of course, the study of substance. For lintels
are importantly not substances, and neither are cloaks nor brass
spheres. Aristotle repeatedly turns to these entities not as examples of
substances, but because they are entities in which it is easier to
abstract matter and form and examine them as separate principles,
precisely, that is, because these entities are not substances, not entities
for which matter and form are the same thing. Their study merely
makes possible the formulation of important issues concerning sub-
stance in terms of matter and form.
The attempted accounts of substance as matter and form, which are
thus essentially accounts of substance ?ls subject and being, surrender
to the more obvious fact that substance is the compound of matter
and form. But at the end of book Z (Met. Z. 17), in one of the most
elegant arguments of that book, Aristotle sets out to show that matter
and form cannot be elements in the being of substance, and the
relation between them cannot be one of the joining of elements. We
are therefore left with the question: In what sense is substance a
compound?
More critically, we are left to ask how it could be a compound, to
inquire, that is, not simply into the nature of substance's unity, but
into its very possibility. For the question of this unity is not merely
a technical problem; it is at the heart of the inquiry into being.
Specifically it is at the heart of the attempt to rescue being from
collapsing into either of the dialectical poles I have suggested are
found in Aristotle's argument. These poles are not, as I may have
seemed to suggest, merely developments of our intuitions concerning
substance; they are central characters in versions of what I take to be
the important competing theory of being that Aristotle addresses.
Central to such a theory is the claim that every entity in the world
of nature is an instance of a relation between some subject and what it
is. In the version we might call, with appropriate winks and nudges,
Platonist, no such entity, and a fortiori no such subject, is then in fact
a being. Properly speaking, being belongs only to the transcendent
what to which such an entity stands in relation, and the (so-called)
being of that entity, its being, as it were, that what, is only an instance
of such a relation, a relation that we may recognize in the important
but problematic Platonic concept of participation.
Aristotle's earliest response to this view is characteristically gen-
erous; substantial being is granted to a variety of candidates. Not

every entity, but the right sort of entity, is substance in the primary
sense; what it is also is substance, but in a secondary sense. More
generally, it is the relation between them that is being, the predicative
being signalled precisely by our predicating n Kara r1v6r:;, this of that:
by our saying, in other words, 'this is that'.
But in the more mature ontological project of the Metaphysics,
Aristotle undertakes, in a way that allows him to go beyond what is in
fact his gentle early criticism of Platonism, a more radical critique of
the theory of relational being. While this theory, he argues, gives an
accurate account of accidental being (where the categories provide the
types of predicative being and substance the subject), in substance
itself there is no relation of being to an independent subject. It is a
human being who is seated, who is fair-complexioned, or who is here
in the library; in each of these cases some being is predicated of that
human being as of a subject. But human being is not in turn predi-
cated of some other subject, for that which is human is nothing other
than the human being who is being human, and it is this human
being's being human that is the primary being, the substance, that we
mark as 6 nr:; av()pwnor:;, some particular human being, some parti-
cular instance, that is (though no particular particular instance) of
being human.
We may phrase this fact in terms of matter and form; the oak beam
that is a lintel is an entity that can be removed from its position, and it
is this very oak beam that, in the proper position, will be a lintel. This
does not allow us to say without qualification that a lintel is an oak
beam; it is oaken, made out of oak, but not exactly oak. Nevertheless,
that which is a lintel has a being independent of its being a lintel. The
matter of a lintel, in other words, is itself something other than a
lintel, and it is this something else that is the matter relative to the
positional form that constitutes it as a lintel.
But there is nothing other than a human being, that is, there is no
other thing, that is the matter of a human being; for the flesh and
bones of a human being are themselves human in a way that no oak
beam is in itself hyperthyral. Instances of substance, we might say, are
form all the way down. Or, we might alternatively say, they are
matter through and through; for to see the respect in which matter is
that which is the being that constitutes the entity it is, is to see the
sense in which Socrates is the matter of the walker, for it is Socrates
who is walking. More accurately, we need to say that if substance is
per se being, there can be in it no separation of matter and form, as
there is between an oak beam and its position, no separation that is,
between that which is the substance and that which the substance is.
200 Aryeh Kosman
If we are to sustain the notion that substance is per se in this sense,

and consequently sustain the analysis that avoids the separational
pitfalls of Platonism, we will need to be able to give a proper account
of how it is that in substance, matter and form, subject and being, are
one. This is the problem of explaining (the very possibility of) the
unity of substance. It is, I suggest, an important problem, and at the
heart of Aristotle's enterprise.
Aristotle's strategy is to appeal to a kind of relationship where the
unity of the relata is clear-the relationship between an ability and its
exercise, that is, between an instance of ability and the activity that is
its realization-to explain the more problematic unity of matter and
form that has developed as central in the argument of book z.
Clearly then, if we go on defining and speakiiig in this manner, it will not be
possible to explain and solve this difficulty; but if, as we say, there is on the
one hand matter and on the other form, the one ovvaµr.1 and the other
t.vcpyeiq., what we are after will no longer seem to be perplexing. (Met. H.
6, 1045"20-6. Cf. also 1045b15-24)
Unfortunately, however, we find little that is explicit about this
strategy in book e. Instead Aristotle embarks upon a discussion, first
of ability and then of activity, as though the question of substance had
been answered, or at least momentarily forgotten. And when, well
into book e, he turns specifically to a discussion of activity, things do
not get better; even those most accustomed to the frustrations that
await the reader of Aristotle may feel some vexation. At the begin-
ning of book e he tells us that
we will first discuss ovvaµ/(; in its strictest sense, even though it is not the most
useful sense for what we are now after; for ovvaµ1r:; and tvi:nr.ia are spoken of
in more senses than those with respect to motion alone. But when we have
talked about that sense, we will then, in the course of discussing i:vtpyeia,
explain the others as well. (Met. e. I, 1045b35-1046a3)
The opening of chapter 6 looks as if it is finally going to fulfil this
promise:
Now that we have talked about ovvaµ1<; with respect to motion, let us say
something about svencia, explaining both what it is and some of its features.
For the ovvarov will at the same time be made clear by our discussion, since
we do not call ovvarov only that which naturally moves something else or is
moved by something else, either without qualification or in some special
sense, but also use the word in another sense, which is the purpose of the
inquiry in which we have discussed these earlier senses as well. (Met. e. 6,
1048"25-30)
But the remarks that immediately follow do not in fact give us a
definition of the important kind of ability by means of an explanation

of activity. The defining seems to go in the opposite direction;
evep}'eta is a state of affairs obtaining not in the way we mean when we say [it
does] ovvaµe1. (Met. e. 6, 1048a30-2)
Nor do things improve as we go along, for Aristotle immediately
denies the possibility of any unequivocal definition of activity, claiming
that
evep)'eta is not spoken of in every case in the same sense, but by analogy: as
this is to this or in relation to this, so that is to that or in relation to that. In
one sense it is as motion in relation to an ability, in another as substance in
relation to some matter. (Met. e. 6, 1048b6-8)
I think we should not allow our exasperation with Aristotle's ellipti-
cal and laconic diction to blunt recognition of the importance of this
hermetic description of activity. Aristotle's strategy here is analogous
to that of negative theology; it is by contrast to the articulate tale told
about Jvvaµ1:;, as an ability for change that we are to understand the
less clear but more important sense of <51!vaµ1:;, and its correspondent
sense of evepyeia, the sense that informs Aristotle's developed theory
of the activity of being. It is this more important sense of the r:vepyua
of a <5vvaµ1:;, as the exercise of an ability, as the complete expression of
an ability, that explains, on Aristotle's view, the genuine unity of
matter and form in the being of substance.
The introduction of the contrast between activity as the principle of
su9st-?-nce and activity as the principle of motion is thus not merely
h~riS,)ic, not simply a pedagogic ladder that once climbed should be
discarded. The contrast must be kept sight of if we are properly
to understand the nature of being. The sentence that inscribes this
contrast-in one sense it is as motion in relation to an ability, in
another as substance in relation to some matter-may serve, I think,
as an epitome of Aristotle's argument in the central books of the
Metaphysics; it suggests that the argument of those books is directed
toward revealing activity, properly understood, as the key to the
elucidation of the nature of substance-being (and therefore of being in
general).
III
Let me sketch what that proper understanding is, beginning with
activity as a principle of motion. Aristotle holds that a motion is a
certain kind of realization, the realization of an entity in so far as it is
able to be something other than it is. The well-known definition of
202 Aryeh Kosman
motion in Physics III is stated in precisely these terms. Aristotle,

however, makes clear-although commentators have had difficulty
hearing this-tl1~_!_~~j9eL!_!<?Ll!!~~n -~YJ~~E~!'l:!!<?P... !Jie p[QJ;fIS.S. by
w_!!_~~~ -~~-~~- ':I:~. ct_~il!!Y . i_s,. rt::.<ll.~~_9_; t11.<lL.P!2-C.~s,s,js, _ fil~£i§.~Jyth.e . . enti ty
th_a,L!.).J::!.§ill!le.means. .to..he...denning, .llQLth~.£9.!1£~I>!Jv_!~Dl1§ ..9.Lwllich
he !1-!~<lllS. . .t.2..Q~fiQ~jJ. Nm •.he . .xeminds. J1s... .ca.nhc:LpQssibly..me.an the
prQQY-fL<?.Uh~.!.E~~li.?'.~.tiPJl.~ .. the..physical.huilding,.foLe.x.ampk..~ich
re~l:!I!S,.1!2.!!!._..!h.~---<l.C:L<?L!?µilgjp.g. T~I~~l~~tjgIJ. th2Lhe. wisb~s to
define ~-l.!l_QJ.!2.!L1Q.J2ejs..JnJact..an.. .acti vity, the_~c:.!!Yit..Y..QDLP.Utenti al
enJ.i!.Y.e.rior .!2.jt§_r.~aj!~.2.!.i2!1:h~.Y!n&..!2em~runpleted.
This activity is the entity's most fully exercising its ability, its most
actively being able to be what it is able to be qua, that is, being able to
be it, while not yet actually being it. This is what Aristotle means by
saying in the Physics that a motion is, to use the English we are most
accustomed to, t~ realizatiO!L~Q2.l~.!lili!L.~iD.!L9..~.!Ll?.Oli:'._~~a/.
(Phys. III. 2, 201a11)7Becoming, in other words, is for Aristotle the
full exercise of an entity's ability to be otherwise, its most actively
being able to be other than it is.
It ~~t?E~-~~.t::.!?.'1.~.~~.<lfcJ.S,. .t2...S..l:1.PP_<?,S,.~.!.h~!..!h.<? . .c:211c:.~e!~-QL!?.\!.YQP.1;Ju1d
tvipy~!?:...... (l!>.~~.!Y..<l!.1:~-~~.!.i.Yl!Y=~~!tY.,~Jroi:g__ !.h~.!.9f. £hJmgs:_QI ..mQlion.
This is a tempting supposition. A ovvaµu; may be seen as an ability
that an entity has to undergo some transformation in which it becomes
other than it is; an tvf.pycia will then be the result of such transfor-
mation relative to a prior <56vaµu;. Here's how one reader of Aristotle
puts it:
the form that supersedes another in a process of change is called the actuality
(energeia) of the previous ability (dunamis). Thus matter and form regarded
as factors in a process of change become potentiality and actuality (or ability
and act). (O'Connor 1964a: 51)
But this is a mistaken view of the order of conceptual priority
between motion and act. The conce.Q!._Qf..motionj§_g~fi..Q~Q~YA.t:i§totle
in terms of a prior concept of activity that does not depend definition-
alfy-~..Q.Q:ii~=tiiai=o.fmo!~;··-5u-ch····actlvlty.1s_f_eailzecf···wlth.re.specito
its correspondent ability not, as Aristotle puts it in the De Anima,
by a change from one characteristic to another, but by the active
exercise of a characteristic that was present but dormant (De An. II.
5, 4r7a32-b2). These prior concepts of ability and activity allow
Aristotle to give an account of motion in terms of activity: an entity is
in motion when its ability to be other than it is, an ability inactive
1 Similar versions of the definition are at 201•29, 201 "5, 202"7, at Phys. VIIL I,
251•9 and at Met. K. r, 1065"16, and 1065"23.
when the entity is at rest, is actively exercised, when, that is, the
entity is most fully being able to be different.
But although a motion is in this sense an activity, Aristotle contrasts
motions with such activities as seeing, which he describes as activities
in a stricter and more literal sense. In such activities, there is no
difference between the acting out of an ability qua ability and the
acting out of that ability simpliciter. The realization of being able to
see is seeing. There is no process by which that seeing is realized that
can be said to be the realization of the ability to see qua ability, while
the seeing itself is the full realization of that ability. Nor is there an
end other than seeing that seeing is in its nature designed to bring
about, which would make seeing the realization of the ability qua
ability, while that end would be the full realization of the ability to
see. It is for this reason-indeed it is this fact-that, as Aristotle puts
it, to see is to have seen. But to move is not to have moved; in the
case of motion, there is always a further end-having moved-that is
other than the motion and for the sake of which alone the motion
exists. 2
It is because the exercise of an ability qua ability and the realization
of that ability as the end other than itself to which it is directed are
distinct in the case of motions, that a motion cannot both sustain itself
and achieve its end. In order fully to be itself, a motion must cease
to be. Because of this auto-subversive character, a motion is, so to
speak, on a suicide mission. A motion is fully realizable only
posthumously; while alive, it has not yet fully achieved its being. For
the achievement of that being lies in an entity other than itself, in the
full realization that it is destined to bring about only at the expense of
its being.
The ability, therefore, that a motion may be said to be, is consumed
in the course of the realization to which that ability is ultimately
directed. Like time, the ability that is at once exercised and realized in
a motion gets used up in the course of the motion. This is only to say
again that a motion is the transformation of an entity by the exchange
of one characteristic for another; and since the first characteristic is
the occasion of the very ability in question, that ability is lost in the
transformation.
But in the case of an activity like seeing, since the ability's exercise
and the realization toward which it is directed are one and the same
2 What is a motion in one context may be thought of in another context as an

activity carried out for its own sake; one may build to learn or to practise
building. But building is none the less a concept dependent on what is being built,
that is, on the building which is the goal of building.
204 Aryeh Kosman
thing, the activity, far from replacing and using up its ability, is the
full manifestation of that ability. Indeed in many cases in which
Aristotle is interested, such as the development of a virtue or the
acquisition of a skill, it strengthens and develops that ability (EN
II. I, II03a34-b2).
We now can see what Aristotle has.. in mind in the remark from the
De Anima I quoted a moment ago. In motions, entities exchange one
state for another and different state, and in so doing surrender an
ability in order that an appropriate end may be realized. But in the
realization of an activity, while it is true that an ability has in a sense
been replaced with its correspondent realization, no change has taken
place, and as a consequence no ability has been destroyed; the ability
has in fact been brought into full being.
All of this is brought sharply into focus in the sentence that follows
the one I quoted from the De Anima:
Neither is 'being acted on' used in a single sense; in one sense it means the
destruction of something by its contrary, while in another sense it means
rather the preservation of a being in the state of ability [bvvaw:1 ovrod by
an actual being [evreJ.cxr.iq ovw;] that is like it in the way ability is like
actuality. (De An. II. 5, 417b2-5)
The J1!va1ac; introduced in the opening chapters of book e is, as

Aristotle says, that more common concept of the ability an entity
has to change, its ability to undergo motion (or to cause motion in
another) understood most generally as the principle of the possibility
of entities to be other than they are. Such a principle presupposes the
concept of a change in the being of an entity, and of a process of
change in which the ability to be different in a specific way, as
manifest in that very process, is consumed in the entity's changing.
But the notion of b61 aµ1c; to which the discussion in book e points,
1
the notion that Aristotle describes as more important to the purposes

of that book, is concerned not with change, but with what for Aristotle
is a prior and in some sense deeper structural feature of being: its
ability, paraphrasing the locution we have read in the De Anima, to
appear now as ro exe1v µr( evepydv Jt, being but inactive, and now as
u) t.vepye'iv, being active (De An. II. 5, 417a35-b2). This feature of
being, its structural propensity to be now latent, now active, is the
central concept in Aristotle's theory of <56vap1c; and t.vtpyeia in relation
to being.
Since this sort of ability and its realization are the same being, and
since the realization, as we have seen, is in fact the awr17pia-the
preservation (some might say the salvation; Hegel might say the
aufhebung)-of its ability, the relationship between ability and realiz-
ation is preserved even in the fullest moment of realization. It may

be necessary for heuristic purposes to introduce the relationship by
reference to the distinction between a sighted person with eyes open
and that person with eyes shut, or between a person who speaks
French when silent and that person when actually speaking French,
but the person whose eyes are open still has the ability for sight, and
the person who is speaking French is still able to speak French;
precisely the point I have been arguing is that at_!!Je lll()IJ1~nt of
realization abilities o,fthissort are most fully active.
In the realization of the ability that an entity has in this sense, then,
in an entity's actively exercising its being, the ability is not consumed;
it is precisely preserved and made manifest, is called forth into the full
and active exercise of its being, SQJliM.Tt;l:.ll.i:Z<:ttion clges not replace
ability~hJJ1i$ . l:.l.Qility,.a.nd isJh~ . <?.<::<::l:tsion .fQr Jh.~ . f4ll~st. 9,x1dJ11ost"real
expression of th(! <i.'2iliJy JhaUtjs.
It·· is· this sense in which i56vaf.l1c; and evipyeza explain the unity of
substance. Aristotle proposes a solution which is deceptively simple:
th~_!:_t!__ i~_,_h~-~<lYS~.matt~Land..fo.rm, anc!Jl1e 011e is relatec). to ability
andtb~_gt}1er to activity: ro_pev i5vvqf.ll') .I<?.~.f. evepye_iq (Met. H. 6,
1045a24). How are we to understand this? Well, in one sense, as we
have seen, matter is that out of which an entity is made or comes to
be, that which is transformed into an entity. In this sense, matter is
related to ability in the sense we have seen to be applicable to motion
and change. But it is not in this sense that the matter of substance is
said by Aristotle to be 'able'; this is the force of his claim that the
sense of i56vaJ.lzc; and evipyc:za important for the understanding of sub-
stance is not that sense relative to motion, but that other, 'more
important' sense he invokes.
In this sense, ability and realization are the same thing, present
together in that full exercise of being that is nothing other than the
activity of the one entity that both are. And it is in this sense that
matter is the ability of which form, as the being· which that matter is,
is notlifflg"otlier thal1 the full realization. jt is the sense in which
sub.Stan-ce"is'activiiy", a'sense ma<le'"C'iearTi1"''6'ook A' where the mode of
substance that is paradigmatic of the being of substance in general is
revealed as an entity whose very 01)afo is 1:vtpyua, that is, whose
essential nature is activity (Met. A. 6, 1071h20).
IV
A proper understanding of the connection between substance and
activity makes clear the sense in which Aristotle's ontology is finally a
theory of being in its full verbal activeness, a theory, that is, of the
206 Aryeh Kosman
activity of things being what they are. The substance that Aristotle
sets out to explain is a feature of entities in so far as they are busy at
work in this activity of being. To identify substance as tvtpyeza is
simply to make this claim about the ontological centrality of activity.
I think that there is much of interest to say about the wider onto-
logical implications of this centrality. Here, however, I wish to pose a
somewhat more limited question about the relationship between sub-
stance and activity, a question to which I think there is an easy
answer, but an interesting one. Consider Aristotle's identification of
substance as activity in relation to another feature of his ontology, the
fact that in the Metaphysics animals are said to be paradigm instances
of substance or, as we should more correctly describe Aristotle's view,
that animal being is said to be the paradigm mode of sensible sub-
stantial being. This fact means that substance should be said, in some
important sense, to be soul; for the soul is understood by Aristotle to
be the substantial form and essence of animal life. Since it is also the
case that the soul is said to be a kind of realization-the realization,
as Aristotle puts it, of a natural body potentially alive-this fact might
in turn seem to reinforce our sense of Aristotle's identification of
substance and tvtpyeza.
But does it in fact? For although Aristotle's account of soul in the
De Anima is indeed in terms of notions of realization and potentiality,
he actually defines soul not as a kind of tvtpyeza, but as a kind of
frreJ.txeza, using a neologism, once Anglicized as entelecliy, that sig-
nifies realization in the sense of a developed state of being and whose
i~ntification with activity seems to beg the question.
As long as we are talking about activity, however, the concepts of
ilvtpyaa and tvrdtxeza are functionally equivalent. Indeed, in such
cases, tvreJ..txeia should be read as a gloss upon and extended sense of
evtpyeza, signifying that mode of activity that is complete-tvrd~(
that unlike the incomplete activity of motion-tvtpyeza are),~(-iS its
own end. To suppose otherwise is to suppose that the only interesting
contrast to the kinetic is the static; but Aristot:).e's project is precisely
to reveal as a more apposite contrast that between the kinetic and the
perfectly energetic.
All of this points to the aporia that I now wish to address. What is
stressed in the early chapters of book II of the De Anima is that
/;vreAtxeza is said in two senses, and the introduction of soul as evreJ.txeia
is in the context of the distinction between these senses, a distinction
at work everywhere in Aristotle's thinking but particularly critical to
his analysis of lflvx~-soul-and the structures of animate life in
general: the distinction between a first and second grade of realization
(what we have come ,to call first and second actuality)-that is, be-
tween an achieved dispositional ability for a certain activity and the

realization of such an ability in the exercise of that activity. Jhe
q ~i 01.LL~ant Jo aQ.<;!rJ~§.~j.s_simply, this.:.ls.S,uhsta.ru:eJir.s.t,,. . ~,secgnd
gE_(l.2.C::..E~~.!rn.tiQD? Is substantial being, in other words, to be thought
of on the model of an ability or on the model of the active exercise of
that ability?
It's not difficult to see how arguments on both sides of this question
might go. On the one hand, the invocational text at the end of
Metaphysics H identifies activity with form. But if the form of para-
digm substances is soul, and if the soul is a mode of first-grade
realization, an active structure of ability, it becomes attractive to think
that substance is most centrally to be understood in the same way, as
a kind of ability for further activity.
The analysis of activity in book e, however, seems to point us in
the other direction. For it reads the relationship between ovvaµI(; and
tvtPYcia ' which is meanTfoexP..........
........................................................... "'"tJ.•.,•.•-JY... Q.I~ma
lain··n:e·unl· .......Hei.aiidfofiii
...... ____. . ."'················' as
a -reT~_betwee!l.fu:.~!~...~!!Q...J?.~£<:?!!~:_g!(l.~~ .E~(ll~~~!L?1:'1:!. that is,
between_ tvtnf!.<! ..f!.S.JLfm.:m 0L.4il!a,u.1'..{which.is.. the.x~.alizat.ion9f a
pri6.tabii}ty)_.·_ an<:t.evtpr.c11J . ii;1Jhe Jull . . s.ense . . of .actiYity, .ivtpycia . . as . the
exeiciSe of sucl1an ability. We should therefore read our master
text~'as ovaia in relation to some matter'-to mean 'as some mode
of second-ornde
___ .. _ _ _ _
realization-activitv-in relation to somethfr1g.
°'~~~·-"'~"''~'~'''''•'"''>'>.'''"'"'~''"''"'~''>''•'"~.•>'""~~ ""'>""''-'""""'"""'':..:.J. ...~.,,,_,,~--"'"'.,._,,...,,~M~'-~-~,. ........ ,H-WM""""'""""'-~'"~'~~.--,,.V,.m.o>-'-••••
••..
an~!~gous _t,~-~!.~!~.£E~~~~e.~!~~~~o~~~~i. On this view, substantial

being-oMia-is an activity m exercise; what Aristotle clearly has in
mind is, specifically, the activity of an entity's being what it is. Which
of these two answers should we endorse as the correct response to our
original question?
v
I want to suggest that neither of these answers is correct, not because
there is a better answer, but because the question I have posed is itself
fundamentally misconceived. For it is misle~ding to ask whet]ler th.~
activity in question, the activ_ity of <!_1l_entity's being what it is, is fi~s~
orsecond-grade realization; with respect to activities of this sort,
therelsnosuchalsffiicFon.-111e-a:a1vrtyoT5e1ilga·lliiiiiaii-b"elng~· tor
example;ISIUiry-reahzea by any individual human being at all, simply
by virtue of its being human. And in general, the activity of any
substance, as the substance's being what it is, is fully realized in that
being and is so in the way that thinking, for example, is realized not
simply in the ability to think, but in the active exercise of thinking that
Aristotle calls fJcwpia.
In this sense, to be sure, substantial being may be said to be more
208 Aryeh Kosman
like a second-grade realization than like a first, more, in other words,

like the activity of seeing than like being able to see . .B.J1Lthis_way_of
sneaking is misleadiniY. For the abilitv to see is still nresent in ~
~i~~~~~-t~it1~tft:JRl:'":~j·~~~:.1~:1~f~~-~:-!~i;~~ii~~~~f~~~g~i-~~
saw, are the most complete expressions of their capacities; seeing just
is being able to see most fully active. So it would be mistaken to
characteril".~_qr)_qJg as . <1... 11129~.9L~.e£Qnd:giade..reaIIzition.:Ifi:1iai-1edus
to-~iipp~-~~-~~~t it \Vas_ not at t~~-~-~,m~ti111_~.?~E~!-~_gE~<:l~.!5:~!.i.?'..~!ion.
What I have"'jlisfsaio'Ts·; fo be sure, true both about activities in
general and about the specific class of activities that I am suggesting
Aristotle identifies with an entity's being what it is. But there is an
important difference between the general case and this specific class of
activities that is the source of a further concern. One of the obvious
facts about capacities and their exercise in general is that an entity
may exhibit an ability without actually realizing that ability-µ'7 evepye'iv
& , as Aristotle puts it (De An. II. 5, 417b 1). It is part of, the C:?:illI?!:l:ign
a~E~!~.M~.&~Ei.<1!!.~...!.~.Jr.isi~!"'"--~-s_ ~:Vi~~2.!.!~=i~gi!!irIY.. ~E>~:i1 !.~?! . . .an ..
~~-~!.!!Y-11!~.Y.J2.~I911R . tQ,,(l,,, . .§,\!l?iec:!. . ~X.\;'..Il . . .\:Y.h.~.Il-the.s.ubj,e_cL.is..11QLilC.tiYf:ly
engage? in J.~-~.,~~~rs.i§~.QfJll<:tL~\?.iJity.
~i(.wiili..I~~-~r.d..1Q.Jh~...<1c:JiYil.Y...Jh.~n. . &.ri~!Qll~jd_~ntifi~-~ . . \:Y.i.!l:L§!! b-
g(IQ£~~J;t,.Me_gariau...w.o.uld ...be.... correct. tQL.!h~.[~ _ is."UQlhi.ng . . . thaL.>Ve
sh9gld ...describe.. as . hav:ing the ....ability ... to .. be... a human.being that is
nqt,,.CJ:.<:;tiY~ly,J2~ing,.SQ, Here, therefore, lapsing into a condition of
inactivity is a more radical lapse than any mere closing of the eyes or
failure to contemplate actively the Pythagorean theorem. [ndeed,
~~iiii~:--·1~e;~~~s.!ii~~~~~~~~·@i~f¥y9~:l~~i~~~~ i~h~h~;::eo~fs~:~~:
between having sight and actively exercising that sight, such that a
dead human being might still be said to be human, in the way that
someone with his eyes closed might still be said to see; as Aristotle
Pl1!~-~1,:...P.. -.~!:.(ldOX:~.l!Y_j_J-1:!!__with ~-<?_d reason, a de.':':~--~l1m~!1-..~~!..~ ·not
~ hl;1_1,'!la!?-J~~ng Int. 1, ~~). -
There is a sense, revealed in Aristotle's embryology, in which the
m~118-~§.IP<lY.~-~.-~~-~.!~.!?.~.P?_~~-t.H!.i!!!Y~l1tJ1.Q!....~~tjy~!y_a,__~l11P<l_f!!?.~ing.
For on Aristotle's theory of generation, the menses is the matter out
of which the embryo, and therefore mediately the animal itself, is
generated (GA I. 19, 727b31). -~-~!__t.~ ..-~l1EP()~~.~~~!.!.~~~-f~<::_!Js a
coJilll~r .to the argument I've _ just ~ive11 . is to .. c()nfuse . matter for the
ge11~xa.iiiiii~~=6I.~]i:iiiiiaii:Ii~fiig:~\Y1!Ji~th:af Jii~iimat,,e,jpatter . •. tnans a
huma1t.being.s.Jiring...b.ody.. Alexander (1891: 580) makes fflis'very
mistake in his explanation of our master text, and in doing so he
subverts the force of Aristotle's argument by thinking of ability and
activity on exactly the wrong model, on the model of becoming rather

than of being. 3
Th_ese facts show how wrong it would be to think of an entity's
being . what~it'iS-as a first-gra~e- r~~H?~H9ri. Bl!! .t~~y sfo)~ld, ..also
suggesr~t() . ·us:tne·infe1idt)i'§L!hiol<.ing.of .such.being in terms of first-
andsecorid-~-rade'it;aliZati()n at . an. For where ability and exercise are
alwafS'aiiaofnecessity together, i'he distinction itself becomes otiose.
Substantial beings are simply actively what they are.
. i• .•. VI .
i-·Q };j;'>.,.. i <_)!_'-~··''!..\ .~-~~;.~ ·<;_,J;,;,:1 /\,"-:7~"1,·. ,~'\---<1._.:
There is another sense in which we may want to say that a human
befng1s·ariiy--poteritfally..liiiman~·-'ffiis.Ts--ffiesense·1n····:wnicli..someone
may-fie·~·nuiiiari''aiicr··ya 'nor be realizing those activities that are
distinctively human; indeed, a person may not even have actualized
the rude potentialities for developed forms of disposition charac-
teristic of being human, may not yet, for example, have learned to
~~~i~~~B~~~~~~~~ifq~~s,~-~ii:~~iii~~iii~~~i~:~n-~~~~!£g[t:Jt~~v~~r~
simple 1tiia"fii1ffre~l1izea''act'lvirr'""'"'""'''''·''''''''·'•••··-·········--···························-····
Jl'ere~·me·n·;-·15·an'0Tfiei"aspect'of how we are to think of substantial
being with respect to first- and second-grade realization. The activity
of a substance's being what it is, most paradigmatically being an
animal of such and such a specific sort, is at the same time the
animal's being some complex of capacities or powers for the activities
(and for further capacities) characteristic of that animal. Thes~_.,Sg
pacl!!Y~...bmv.s:.Y~I,.D~~Q.J:!Qt ~e realizt;~ for the. anii:nal t?. b_e actively
what it is. For
'""'"''°'"'~"""""""~'\'-
the activiti~S'"th.at'm'ak'e'·ti;:;a"s\ilisiance'i:iiiiiii~ are
--~.,'-"""""~~-"~"m;- ·~·--''"-,.__.. "·-~-·""'-'~'''""'·'"'"'""·'·-~ ,-,,,~'°''"·'·-, ~~·-'-'-''' -•'-''·'>""<''-~--,0'·-•·'<·•·'~
.. ..
di~tiQ£tJrQrn .. the . activity ...of . . the ..substance.:s... being.that....nature.~.····e.ven

thQ_l!glLAL~totl~.m.odd§.lh.e..lf!tt.~LmQ.rn..s.ub.tk.notion.uponthe..former
mQ!~_ob~iQ.!l_LQUe. This is why there is no contradiction between a
discussion in book 4..Qf.!!t.~. !iJ.e,_fqpfiy§iq~JMft . 4.'.7).!QL1~35.:::-b9J~.here
bo~J.1.....'!c.!?il.,!YJ:!!lQ_~ctLvity... ar.e ..said. J.o. . .characterize ...substance and are
thus posterior to it, and our master text in which substance is identified
as activity.
The fact that the activity of a substance's being what it is may
consist of its exhibiting certain capacities rests upon features charac-
3 Alexander may have been influenced in his understanding by Aristotle's
description, in the context of a very different sort of argument at Met. H. 4,
1044"35, of the 1wwµffv1a as 'dv6pcimov ri, airia dx;; i!2r/. But the rri.igmder-
Stfl!1c:li_r:i._g_~~E()f111_ecte~~jth a more. general misunderstandin~L~~Q,iSfi!iiJo...which
Aris.tutle.b.as. compkteJLtffe::~ri,aJy.~i~.~9r~i!liStaoc~-:1'yJfieJj~inning of e and turns
at that point to a related but independent discussion o(poteiitialrty-<frid actuality.
210 Aryeh Kosman
teristic of such principles of being as ability and activity. I want to

illustrate these features in a roundabout way by recalling a short
argument at the beginning of Metaphysics Z. 6.
VII
In the case of entities that are said to be what they are in the mode of
accidental unity, Aristotle there argues, an entity is different from
what it is to be what that entity is. A white human being, for example,
which is an instance of a human being who happens to be white, is
different from what it is to be a white human being. For if, Aristotle
argues, a white human being were the same as what it is to be a white
human being, then it would follow that what it is to be a human being
would be the same as what it is to be a white human being. For since a
white human being, say Abromowitz, is the same individual as the
human being in question, then if that individual is thought to be
identical with what it is to be whatever being she is said to be, all
those beings would be identical (Met. Z . ..6..-.LQJ.r.19-~ZJ).
A~tle here i~vo~eey tl!!J!nity_pf .flJLi!lillYill..'!.<!L~mi1LlY!tj£QJ!ru:ler
d!ff~i.:enJ descriQ!iO!J,§..~.~hiJ?itL<!!!f~reu~,.1?£k!~~.A.ll!:LiU:g~.Jhat.iUYe
W£!'£._~ntily~tn. be. . the,same~as..a~.ru:.411 ..oLits.J;lifferept
b~!~_g~ 1_!hm.!h~ diftsrs.:nl.lleings..~m.tl.d..he.J:.~Y£S?!Lt9.~D.ne..id__entical
Q~ing. The argument shows, in other words, that the distinction be-
tween the unity of an individual entity and the plurality of its different
beings would be undermined if we were to assume the identity of an
individual and its being in each case.
Aristotle's argument here relies on a simple fact central to his
ontology, ~act ~ha~.. ~.O,X..~~.!!.Sl.~..!':!~~_y.~sl,~':1:1..~.Q.mY. !!!.<lYJ~~-i'!!!2~!!£ed
~~~fnt~~~·¥f;~·y~f~;fih;~}~i-i~7~~~·~t~=~~!~~o~
being will be distorted; we might even say hyperbolically that being
itself will be lost, for a theory that loses sight of this fact will be
unable to account for the ability we have to jdentify and distinguish
among the different predicational beings of a single individual entity.
But if an entity is always identical with each of its beings, then if the
entity is one, the beings must be one, or if the beings are many, there
will be many entities. In either case we will be unable to maintain the
distinction between and compossibility of the unity and singularity
of an individual entity and the diversity and plurality of its several
beings. But in fact, that which is in the library and is pale and is short
and is thinking and is a philosopher is in each case none other than
the single individual entity Abromowitz.
This inversion of t!J.e more conventional form of the one and the
many is central to the thinking of both Plato and Aristotle, and I take
it to be equally applicable to activities. When a builder is building, the
activity in which she is engaged may be thought of either as a motion
or as an activity. If I walk from Anaheim to Azusa, that might be a
journey to a goal and thus a motion, but it might also be a case simply
of strolling, a mean but authentic instance of activity. In either case it
will be motions that make up the activity of walking, so that even
when it is thought of as an activity, it will consist of various motions.
This is equally a~-r:!~1:!!J.Y~!!!1~~inJbe.mosLcentraLactivities'°f an
animal's life, the.activiti~~1baL,eJu:ts,!il!tte..the.essentiaLexercise.4)fjts
na~?re;.. T0r···1ne~-~~.-~-2E~i~L2iJnQ1i.Qn§.,..J?gJ1tJn .Q~!C1.!L. .~~~.48!.2~C1Ily;
see1ng,-ToYe·xa~pJ~.!j~Lmade ...up ofJ:lne.nnd.another.monon.
rlftfie"same·w~ 1 I wisQ.JQ..EW.e.....th.e actb.'i!.YJh<.il.iL~L~!!Q~.!!.!lf~ 's
being~WfiaflTls can_cor.sist2L9JJ1~.rJ:t.c.tiritLes_thi!lJ!re"~!t~-.£~~..~l1~:
st~~e i~.z...~!1d.!h£!L!!t;;!Y..~L3-:!1.Y.mom~mJ;ie_ma.nifm~o_gly~jE:.!~£!2!!!1 of
abili.!i~d...QQ~erLfoLtho~~LiU~tirities. This is in brief the rather
simple point of this essay. There is undoubtedly an interesting story to
be told about why activities constitute the privileged modes of being
that make up the content of substantial natures. But in addition, ac-
tivity provides for Aristotle the model in terms of which to think of
the most abstract mode of substantial being, that being by which
substances are whatever it is they are. It is a mode of activity. A.l!.c!.
sin~:-~~..i.~. . !~-~~!~111 t.'?!.8.a.~i~~-~--':1!1~.~-!!.~_gr_a.:!~~E~~!!E.<l!~~~~t!~E~h_a t is
~~~~lI~~fi~~::~~¥6i~1We!i~l1W1~y~~~vii~·~1fit~1Y*~e~:~~ :th6~t ~citYi~ ~~
'in-soilis:. . £:~~~s~~i§i!l?ifan£~I~?:~a·c:eLtaliiJnatier':·· · · ····
All of this by itself tells us little about the details of Aristotle's
theory of substance, or about the larger elements of his metaphysical
picture. I suggested earlier that the model of activity solves the prob-
lem of the unity of matter and form and allows Aristotle to develop a
subtle and powerful alternative to platonist ontologies and to his own
earlier ontology as well. But here I've meant only to suggest how we
are to understand one aspect of Aristotle's identification of substance
as activity, and to highlight the quite general use that he makes of the
notion of activity in developing his theory of substance. H.J~ .. not
simpb'_Jhat _c!i!f~-~-Q!.J.!£!iYH!~~-!!1~~.~.J:!Q..~,Qd determine the m1ture of
:!m~~isri0iP-:~J.f~e~fs~~%~i·°-~~L~~r~~lFE1~~~f~s~~~iW~~tf~-a~~~-~;~!
......... .. . . . . . . . .........&... ,.............12......."·-···-----Y~............................................. ······· ·················· ··· ..
4 As long as we don't see this, we will be correct to be bothered by the worries

about motion-like features of tvepyeia raised by Ackrill (1965). Ackrill generates
these worries by pointing to entities which are first described as modes of ivtpyw1
and then making trouble by redescribing these evtpye1w as motions.
212 Aryeh Kosman
actively. be the s~!?§.!~n!!~.L ~!!.!.!!i~L!.h~Y.. __ <,lf~· If we are to under-

sfandArisfoiie;s identification of substance as activity, we need to be
diligent in not confusing activity proper and motion, even when that
confusion is the subtle one of confusing substance wffii"Tfie···crevel-
op.m~rit~9La::-suhstaniiaCuaiure~·B~t·-I've ···11ere··cfaimed' Hiarwe ··must
moreover r~sist .. th!n~iE&..Qfthi.L'!.~!!Yi!Y.. '1..t.f'2!!!!~!.ly_.~.9~ili~J!!.wtiL.the
po~,:E,s_~!!.~L<;,tS.!iYJti.Ys..... th.~! . . .~.<l.!S~....!!ILlh~.... 9~t~.rmiJ1at~.~mt~.~9L~JlY
gfy~~Jzl!a11£~,i,..,.s.'.Y~Jl ...Wh~nJb,tK.§t,,,il£!i~i!.~~-s•..~E.~...!h~,Y!fY.JJl..2S!.5:!§. . ,,<;?f
act!Y!lY_.i!Ll~.E!!!.s... 2t.~hi1;b . Ads1Ptl~J}nally,.J111g~rs~.<lnQ.s..th~ ..b~i11g of
subs.ta.nee.
Perceiving, reproducing, thinking-engaging in the variety of modes
of activLty that make up animal life-these are all in their way acti-
vities. But it is not these activities to which Aristotle is referring when
he identifies substance and activity, for they characterize substance in
the mode of both ability and activity. There is a~acti~ity of bei~g that
is not formally equivalent to any of the.st! .c:lt!1~r,iliin!g:"J?:2~~~s. ~!Jd
activities that constitute the specific nature of particular individual
suosfiinces;"Tfis''tliis"acfivity''or··a···tning's•.beliig''·wh·ar·1r···1s···1nat··15its
sUbst;Il'ce being, its ovaia. At one extreme, this activity of being
is most paradigmatically manifest in that being that is nothing but
activity, the being of divine substance. But it is, at the other extreme,
figured by those very activities that make up the natures of specific
substances.
VIII
There is a final chapter to this story, to which I alluded at the
beginning of this essay and which I will here only outline; it is the
chapter that returns us to the original question of the Metaphysics, to
which the question concerning substance is subordinate: the question
concerning being as such.
Earlier I claimed that understanding the connection between sub-
stance and activity might help us to see the respect in which Aristotle's
ontology is a theory of being as activity. To identify substance as the
activity of entities busy at work in the activity of being what they are
is, given the paradigmatic status of substance, to make a claim about
the ontological centrality of activity to a theory of being in general.
This much larger story about being as activity would do much to show
how specious are depictions of Aristotelianism that characterize an
ontology of substance as an ontology of inert objects, and that con-
trast it to theories that portray a more active and dynamic view of
being, or how misleading is our own unconscious assimilation of
Aristotle's ontology to an ontology of things. Such depictions over-
look precisely the centrality that Aristotle affords to activity in his

developed theory of substance, and therefore in his developed theory
~~!~~~g ~~~s:~!f£:l~offi: ~5~J~r~c~~~!_i~~1m~~ ~;~~bs~~iihc~~{J~{~ ~)~-

activitY:C:c?m¢~J~»figurebeingin general. ······· ..............
We may experience some dissatisfaction with this act of figuring as a
mode of metaphysical explanation. But such dissatisfaction would be,
it seems to me, a mark of that Philistinism that Aristotle ridicules
early in the Metaphysics (Met. r. 4, rno6a6), the ignorance that does
not know the sort of explanation it is reasonable to expect in first
philosophy. F:2~.. Arist()t!e's pr:ic:tice rrierely. ~xhibits . t~e ~ecessary
transgressi()~ of synecdoche 1 th::i,t perpte~ing>fr§p~ in whiclj'a'Yhole is
a
figured by p~_r(':"hose being :is apart d,epemts up911its 1l()Jge,i11g the
\.V.hole,,!h~tit.i~.ca!l~~HP~!:l !() fig~f_e. Metaphysics, which by its very
project of explanation is doomed to traffic in global synecdoche,
calling upon this or that part of being to figure being as a whole,
cannot be free of this transgression.
So if Aristotle ~pp_e,.<.tl~ !2 <l._~pe,ci~~-1!!.().<:!~ gf_ b,eing~ . . activi.ty_, in
exp.@Iqtfig:::J:i:~frig._is a. whole., . thisJP&rnly ma,~ks the fact that evtp1·w1
join~.. -~~-~h-A.rist9.leli!;ln .· IJ9tions.. of ...~rL.'!.s dJ;;~~-~ricl. N,,z·:· JO,frri:-~nd
matter-not,!QH§Jhathave.heentrnmla,Jed ff()I,II_their . loc:<1J Q()01ains to
do·=~9iI'Inthe.high and airy courts oLfirsLphilosophy. Such trans-
lation is an inescapable feature of the project of edification in the
Metaphysics, which of necessity looks to a particular feature of being
to figure and thus explain being in general.
Any governing trope, however, in terms of which we choose to
figure being in general will of necessity be transgressive. But it will
equally carry not merely rhetorical but conceptual force as well;
'the tendency, the thrust, points to something', as Wittgenstein once
remarked about language that overflows the facts it tries to express.
We have too long read the .ruling metaphors of Aristotle's meta-
Pli'sics····in
. .Y .......... ""'"'ihe·11·11tof
·' """'""'g.... ,..•...., •.... "·a:·.noder··overnea··b·.····· notions 0rc:ha:n···· e ·and
........,.." .....,. , ............K............., ...............Y.....,........... .. ,..................,.S..,,.... .
be,,c9Il}~gg_x(:l_~h~r .!h,l:l!l:,l?t!?~}gg",gf, the.. . !lbiHtY. . ~2. ~e ()tbe.r .r<;t.!he.r .~hl:ln of
th.: . . ~-~ilgy f2L.§.~Jf~.~.~t?J:~§,§i2lJ•.. .RI . t.be. . 9Qntrc,t~t . . Qet\Ve.e.11 . . . th.e. . possible
and th_ereal ratber tb::t11Jll.at9etween, c,t\:>iHt,y(:l~d its .~x~rci_~e. My
argument iri this essay has been that we need to listen more carefully
to Aristotle's choice of the ruling metaphor of evtpysia. For it seems to
me that Aristotle could have done much worse than to choose activity
as that particular feature of being with which to figure being in
general.
9
The Priority of Actuality

in Aristotle
CHARLOTTE WITT
Many puzzles surround Aristotle's distinction between potential and

actual being (Met. A, 1017a35-b9). How are we to understand his
typical illustrations of the distinction-a capacity and its exercise
(like sight and seeing), and an undeveloped substance and a mature
substance (like unripe and ripe corn)? How are the two kinds of
examples like one another? And how do the two additional important
examples of potential and actual being-matter and form, perishable
and eternal substances-fit in? One of the most interesting and central
puzzles, it seems to me, is what Aristotle means when he says that
actual being is prior to potential being, and why he thinks it is so.
The priority of actual being could be a basic Aristotelian assump-
tion, or it could be a matter of definition, or it could be self-evidently
true. It is none of these. That it is not simply assumed by Aristotle is
clear from the fact that he gives reasons for it in Met. e. 8. And
Aristotle's arguments, together with the fact that he never defines
actuality, rule out the idea that the priority of actuality is secured by
definition.
The third alternative, that it is self-evident that actuality is prior to
potentiality, might appear the most attractive position. If what is
actual is what exists, and what is potential merely possible, then our
intuitions might grant priority to the existent in relation to the merely
possible. 1 This explanation for the priority of actuality is not available
to Aristotle, however, because what is actual cannot be defined as
what exists, and contrasted with the potential, understood as what is
merely possible. We can see this from Aristotle's examples. As I
mentioned above, Aristotle typically illustrates the ontological use of
the distinction between potentiality and actuality by contrasting a
© Charlotte Witt 1994

1 Not even the priority of the existent in relation to the possible goes uncon-
tested, however: 'This makes manifest a peculiar circumstance, which is relevant
throughout the whole of philosophy, namely that within the ontological sphere
the possible is higher than everything actual' (Heidegger 1982: 308).
216 Charlotte Witt
capacity and its exercise or by contrasting an immature with a mature

substance. Clearly, both the capacity to know and the activity of
knowing can exist in a person at the same moment, and so the
capacity to know is not a mere possibility. And similarly, an immature
substance like a child, a potential human being, exists; being in poten-
tiality does not mean being merely possible. And, since Aristotle's
examples do not suggest any other obvious grounds for assigning
priority to the actual, it is far from self-evident. It is a puzzle.
I. THE PUZZLE CONCERNING PRIORITY

IN SUBSTANCE
In Met. e. 8 Aristotle discusses three kinds of priority relation be-

tween actuality and potentiality; actuality is prior in time, in definition,
and in being. It is fairly easy to understand why Aristotle holds
that actuality is prior to potentiality in time. For he believes that
there must be an actual entity at the origin of any causal order, and,
indeed, at the origin of the changing world as a whole. Just as an
actual human being generates an embryo so, too, the pure actuality of
the Prime Mover is the origin of all change in the sublunar realm. It is
also fairly easy to see why Aristotle might think that actuality is prior
to potentiality in definition. Consider the case of capacities and their
active realizations. A capacity like sight is defined in relation to a
given activity, seeing. Or consider the case of potential and actual
human beings. We define a foetus as a potential human being, in
terms of its eventual form and realization, and not the other way
around. A human being is not defined as an ageing foetus. The puzzle
I wish to explore in this paper arises with the third kind of priority
attributed to actuality-priority in being (ovaiq.).
What does Aristotle mean when he says that actuality is prior to
potentiality in being? He does not provide a definition of priority in
being in our text; elsewhere, however, he defines priority in being as
follows:
A is prior in being to B if A can exist without B but B cannot
exist without A. 2
2 My definition of priority in being is a paraphrase from Met. LI. rr, ror9"1- 14.
It is priority Kara rpvatv mi ovuiav ('in nature and being'). Aristotle's discussion of
the existence of mathematical objects in book M. 2, ro77"36-h4 also connects the
idea of priority in being with ontological independence. In this text Aristotle
contrasts priority in definition and priority in being. An attribute, for example,
may be prior in definition to a compound entity (white may be prior in definition
The Priority of Actuality 217
Let us call this kind of priority 'ontological priority'. The priority

in being of actuality means that actualities can exist independently
of potentialities but potentialities cannot exist independently of the
existence of actualities. Actualities are ontologically independent and
potentialities are ontologically dependent upon them. 3
It is easy to confuse ontological priority with causal priority. It is a
confusion because it is one thing to claim that an entity cannot come
to exist and/or continue to exist without the causal agency of another
entity, and a very different thing to hold that one entity cannot exist
without another entity. For example, in the Categories Aristotle holds
that attributes cannot exist without the existence of primary, individual
substances, and there is no mention of any causal relationship be-
tween the two; it is a relationship of ontological, not causal, de-
pendence. Now, as we have seen, in the course of his explanation of
the temporal priority of actuality, Aristotle does attribute causal
priority to actuality: 'For from the potential the actual is always
produced by an actual thing, for example, man by man, musician by
musician; there is always a first mover, and the mover already exists
actually' (1049b24-6). But Aristotle distinguishes clearly between
temporal priority and priority in being, and he discusses them sep-
arately in our chapter. Hence, it is both a mistaken interpretation and
a philosophical confusion to use Aristotle's discussion of the temporal
and causal priority of actuality to interpret its ontological priority.
I have suggested that we interpret the priority in being of actuality
to mean that actualities can exist independently of potentialities but
to white man), but it is not prior in being because it cannot exist separately from
the compound.
3 I am aware of only one other passage that explicitly discusses priority relations
and that conceivably could be used to interpret what Aristotle means by priority
in being. It occurs in Cat. 12, 14°10-13: 'For of things that reciprocate as to
implication of existence, that which is in some way the cause of the other's
existence might reasonably be called prior by nature.' If we use this text to
explain priority in being of actuality, Aristotle would be claiming that potentialities
and actualities mutually imply the existence of each other, and that actualities are
the cause of existence of potentialities. We cannot, however, use this text to
explain the priority in being of actuality because it conflates priority in being and
priority in time instead of providing an independent explanation of priority in
being. Aristotle's explanation of the priority in time of actualities in relation to
potentialities points precisely to the fact that a generation of a human being
requires an existing mature human being, an actuality. The mutual implication of
existence is also a part of priority in time, since an existing mature human being is
preceded in time by a child. And so on. It is also worth noting that Aristotle's
example in the Categories of this kind of priority is peculiar (he claims that facts
are prior in nature to the true sentences that they 'cause'), and has no plausible
relationship to the examples discussed in our text.
218 Charlotte Witt
potentialities cannot exist independently of actualities. Unfortunately,

when we look more closely at Aristotle's examples of priority in
being, and his explanations of them, it is at first sight implausible to
interpret the priority of actuality as ontological priority. Aristotle
says:
But it [actuality] is also prior in being; first, because the things that are
posterior in becoming are prior in form and in being, for example, man is
prior to boy and human being to seed; for the one already has its form, and
the other has not. Secondly, because everything that comes to be moves
towards a principle, that is, an end. For that for the sake of which a thing is,
is its principle, and the becoming is for the sake of the end; and the actuality
is the end and it is for the sake of this that the potentiality is acquired. For
animals do not see in order that they have sight, but they have sight that they
may see. (105oa4-u)
Here, Aristotle explains the priority in being of actuality in relation to
potentiality not in terms of ontological priority, but rather in terms of
a teleological relationship between the two. And there is no obvious
connection between the teleological relationship which unites poten-
tiality to actuality, and the idea that potentialities are ontologically
dependent upon actualities. Further, when we consider Aristotle's
examples, it is not easy to see how ontological priority applies to
them. What sense can we make of the idea that a man can exist
without the existence of a boy, but a boy cannot exist without the
existence of a man? And, what sense can we make of the parallel
notion that the activity of seeing can exist without the capacity for
sight, but that the capacity for sight cannot exist without seeing?
Aristotle's reasons for attributing priority in being to actuality are
equally difficult to square with the concept of ontological indepen-
dence. Activities and fully realized substances are prior in being to
capacities and immature substances for two related reasons:
(I) the actualities have their form and the potentialities do not;
(2) everything that comes to be moves towfirds a principle or end.
The first point, that capacities and undeveloped entities lack their
form while activities and mature entities have them is very peculiar
indeed. Why would Aristotle hold that a boy lacks its form, pre-
sumably the human form? 4 Surely all members of the human species
4 It is worth noting Aristotle's precise vocabulary. He contrasts a nal; (boy)
with a mature male human being (dw/p) rather than with the term frequently used
to denote the human species (&.vOpwnoc;) (1050"5-7). Does the use of avifp
make it wrong to call a boy potentially human as I do in this paper, on the
grounds that the form the boy lacks is the mature male form rather than the
have that form. Second, even if it be granted, as Aristotle says,

that everything that comes to be moves towards an end, this claim
provides no obvious basis for discriminating between potentialities
and actualities. It would seem to apply equally well to boys (potential
beings) and men (actual beings), because both boys and men come to
be for an end according to Aristotle's teleological vision of generation.
The two points read together do suggest a difference between the way
in which a boy exists for the sake of his end and the way in which a
man does. If we take the end in question to be the form, then we
might understand Aristotle to be saying that the boy exists towards
that end, since he lacks his form, but the man does not because the
form is already fully realized in him. Even so, it is very puzzling that
Aristotle would say the boy Jacks his form.
As we have seen, Aristotle's idea that actuality is prior in being to
potentiality is deeply puzzling on two grounds. Neither the most likely
theaning of the priority (ontological priority) nor Aristotle's reason
for attributing it to actuality (his teleological justification) seems to fit
or be true of his examples.
One response to these puzzles is to reject the interpretation of
priority in being as ontological priority. Given Aristotle's teleological
language, for example, it might make sense to understand priority in
being as explanatory priority. On this view he would be saying that
potentialities exist for the sake of those actualities that explain their
existence. 5 For example, the activity of seeing explains the existence
of the capacity of sight in a human being. It is the relevant end or goal
of a natural process that explains the existence of the stages of that
process; the sequence that constitutes the process does not explain the
existence of the end or goal.
There are two important considerations against the idea that priority
human form? I think not, b.ecause the logic of Aristotle's point is clear. He treats
the boy and the seed in a parallel fashion; each exists potentially and each lacks
its form. It is possible, but unlikely, that the forms in question are different.
Further, even if Aristotle were saying that the boy lacks the form of the mature
male (rather than a male human being) the main point of the examples is still the
same, namely that immature entities lack the form they will later realize.
5 T. H. Irwin gives a very clear statement of the explanatory priority view as
follows: 'He [Aristotle] claims that actuality is prior in substance to potentiality
because potentiality is for the sake of actuality; we have sight for the sake of
seeing and not the other way round, and similarly "the matter is potential because
it might reach the form" (ro5oar5), not the other way round. The traits and
activities that explain teleologically the structure and composition of some body
are the actuality for which the body has the potentiality, and they are the form for
which the body is the matter' Irwin (r988: 237).
220 Charlotte Witt
in being simply means explanatory priority. In the first place, there

is no textual basis for this interpretation. Where Aristotle defines
priority in being he describes it as ontological not explanatory priority.
Further, where he explicitly describes the priority of explanations of
natural changes in terms of ends, forms, and definitions over material
explanations, he connects explanatory priority with priority in time
and definition:
For coming to be is for the sake of being, not being for the sake of coming to
be. Hence Empedocles was wrong in saying that many attributes belong to
animals because it happens so in their coming to be, for instance that their
backbone is such because it happened to get broken by bending. He failed to
recognize, first, that the seed previously constituted must already possess this
sort of capability, and secondly that its producer was prior not only in
definition but in time; for it is the man that generates a man, and therefore it
is because that man is such that this man's coming to be happens so. (PA
640"18-26)
Aristotle clearly does attribute explanatory priority to forms and

ends-to actualities. But in this passage explanatory priority is linked
to the priority in time and definition of the actuality, the producer;
there is no mention of priority in being. And this is what we would
expect given our earlier discussion of the connection between
Aristotle's understanding of causation and the priority in time and
definition of actuality. Because Aristotle defines priority in being in
terms of ontological priority elsewhere, and because the explanatory
priority of actuality can be clearly linked to the other two sorts of
priority that Aristotle attributes to actuality, it would be a mistake to
hold that what Aristotle means by priority in being is priority in
explanation.
II. POTENTIALITY AND ACTUALITY

AS WAYS OF BEING
Before I give my solution to the puzzles surrounding the ontological

priority of actuality, I should discuss an important assumption I have
been making concerning Aristotle's distinction between actuality and
potentiality. Earlier in this paper, I referred to the 'ontological use' of
the distinction between potentiality and actuality, and throughout I
have assumed that Aristotle typically illustrates the distinction using
examples of two sorts, both capacities and their realizations, and
immature substances and mature substances. On my view both kinds
of examples are instances of the relationship that holds between
'matter and some substance' in contrast with the sense of potentiality

and actuality associated with motion (Met. 1048b6-9).
The idea that Aristotle singles out a sense of potentiality and
actuality relevant to the inquiry into substance has been argued in an
original and important paper by Aryeh Kosman (1984). In his paper
Kosman presents clear and-to my mind-decisive textual evidence
that Aristotle distinguishes between the meaning of potentiality and
actuality as it is used to analyse motion, and its meaning in his inquiry
into substance. At the opening of Met. e, for example, Aristotle
says: 'First let us explain potentiality in the strictest sense, which
is, however, not the most useful for our present purpose. For poten-
tiality and actuality extend further than the mere sphere of motion'
( I045b35-I046 3 4). 6
What do the terms 'potentiality' and 'actuality' mean in their onto-
logical use? The answer to this question is very difficult; it depends to
a large extent on how one interprets Aristotle's examples, and, to
begin with, what one thinks his examples are. 7 As we have seen in
Aristotle's arguments for the priority of actuality he mentions both
examples of the capacity and exercise sort, and examples of immature
and mature substances. 8 There is clear evidence that Aristotle does
not consider activities like seeing and thinking to be motions (1048b18-
34). Motions have an external telos, and perish when it is reached;
activities have an internal telos that is realized at every moment. A
potentiality for change or motion and its corresponding actualization
differs from a potentiality for an activity. In the former case, the
potentiality is used up as it is actualized, in the latter case potentiality
and actuality coexist. Given this difference, it makes sense for Aris-
0 Michael Frede argues that Aristotle should not be interpreted as singling out
a further sense of ovva;w; (potentiality) useful for the inquiry into substance.
Instead, the sense of ovvaµ1;; appropriate for the investigation of substance is just
the range of senses Aristotle has already distinguished understood in a different
way, namely in terms of their respective degrees of reality. For the details of
Frede's interpretation see Frede (this volume).
7 In a recent paper, Charlton (1991) argues that in its ontological use the
distinction between potentiality and aCtuality refers to the relationship of exempli-
fication between universals and particulars. As I point out in my commentary,
however, I do not think that Aristotle's typical examples of the distinction support
Charlton's interesting thesis (Witt 1991). Charlton's thesis has its textual base in
certain texts where Aristotle makes a connection between matter, universality and
potentiality, and contrasts them with form, the individual and actuality. If one
simply considered these examples, one might define the ontological use of the
distinction between potentiality and actuality as Charlton does.
8 Both sorts of examples are also mentioned at Met . .d. 7, w17"35- 0 9 and 6. 6,
rn48"25- 09, the two other central texts in which an ontological use of potentiality
and actuality appear.
222 Charlotte Witt
totle to distinguish between two related meanings of the distinction

between actuality and potentiality.
But what about the other kind of example-boy and man, unripe
and ripe com? Aren't growing and ripening motions? It is important
to see that Aristotle likens these examples to the examples of activities
and not to motions. It is clear from Met. A. 7 that Aristotle thinks of
the different ways of being a substance like com or a human being as
parallel to the different ways of being a knower; one can be a knower
or a human being in capacity or as exercising the capacity (ro1i'5).
Aristotle is describing an ontological relationship; he is specifying
different ways of being corn or a human being rather than describing
the motions of ripening or growing. The upshot of including immature
and mature substances as examples of the ontological use of poten-
tiality and actuality is that we Il}!!.S.L~!IEUl.JLQJJLAtistoklian ontolQgyJ,o
i!l_~hi,9.~~.Ld.~sLthaLin.a.ddiilim. tQ.kiruis..filJ1eing, su b§lfill..s.e, quality,
~n8...~h~.J£~.L-~~--S.L~~Q-~:Ya,}~Luf~.hs.tall£~.i~£!nd
t.h~J:.£.st. For our immediate purposes, however, it is enough to keep
in mind that with the ontological use of potentiality and actuality,
Aristotle describes different ways of being F where F can refer to
either a capacity or a substance. In these terms our puzzle is why
Aristotle might think that certain ways of being F, namely being
potentially F, are ontologically dependent upon other ways of being
F, namely being actually F.
III. THE SOLUTION OF THE PUZZLE
The clue to a resolution of the puzzle lies in Aristotle's reasons

for holding that actuality is ontologically prior to potentiality: ( 1)
actualities have their form and potentialities do not; (2) everything
that comes to be moves towards a principle or end. My thesis is that
these reasons, properly understood of course, do indeed explain the
ontologi.cal priority of actuality.
The rough idea is that if we view a child as an entity en route
towards its end, as being for the sake of its end, then two conditions
obtain: (a)the c:hildd()t!.~ n()tp_ave its end; (b}none..the.less..the_end
pgt_§L~xist. Indeed, the existence of the child, since it exists for the
sake of its end, is dependent upon the existence of the end. And what
is that end? For Aristotle, the m(lture human ~~i.11gLthe entity having
theform, i§Jhe encl ofJbe ~biW~ thatfor the.sake.oL\Nhich the child
exists,:thatwhich musLexis.Lif the. child exists. The reverse depen-
dency does not hold. The adult has its form and does not exist for the
sake of or towards SO¥Jething else, as least not something else with its
form. 9 The existence of the adult does not depend upon the existence
of the child.
How does thjs line of thought apply to the example of a capacity
and its exercise? Aristotle thinks of the activity as the end or goal of
the capacity. Sight is the capacity to see, to engage in that activity;
sight is directed towards that end, which must exist if sight is to exist.
If the activity did not exist then neither would the capacity that is
directed toward it. ln~Q!!!rast-'-th~-~-ti~!!yj~JlQl.Jli.rn.~t~Q!.9_~1!!.<t~Jhe
cap~city; !Lf~Aristotle tells us later in our text
that eternal beings are pure actualities that exist independently of any
correlative potentiality ( 105ob6-8).
What is common to Aristotle's treatment of both cases is the idea
that the potentialities exist for the sake of the actualities. The idea is
that if an entity, state, capacity, or what have you is directed towards
an end, then the existence of that entity or what have you is de-
pendent upon the existence of the end. Of course we need not don
Aristotle's teleological glasses, we need not view a child as for the
sake of a mature human being or sight as for the sake of seeing. But
Aristotle holds that in so far as the child and the capacity are among
the class of things that come to be, they are also among the class of
things that exist for the sake of their ends (Met. 105oa7- IO). b11_d
hi~.i~LII!..J_~~~~l_.<t_i:i_y___t_e_!~glogi<.:_<t] _ eptj!y _ or.l:\t;:tiyjty _ _ js exis.tentially
depericte...l!Ll1:P.<:mJ!1~.~xist~m:eotlts end.
Does the teleological relationship I have just described between
potentialities and actualities explain or justify the ontological depen-
dence of potentialities upon actualities? At first sight, my proposed
explanation appears extremely implausible. In order to spell out
exactly what seems implausible about it, let us agree with Aristotle for
the moment that a child exists for the sake of a fully developed actual
human being. Consider a young child, Sally. One might think that
what Aristotle means is that she exists for the sake of herself as an
adult, for the sake of the grown-up Sally. But Sally the adult does not
now exist, and sadly, may never exist. Since the young Sally can exist
even though adult Sally does not exist, young Sally cannot be existen-
tially dependent upon the mature Sally. Indeed, it is a basic feature of
teleological judgements that an entity can be said to exist or to act for
the sake of an end that does not exist and might never exist or be
realized. Hence teleological judgements appear peculiarly ill-suited to
9 Aristotle does posit a teleological relationship between eternal and perishable
substances, but he does not hold that eternal substances have the forms of the
perishable substances. Rather, the natural substances are directed towards the
eternal because they desire to be eternal-to have a nature other than the one
they do have (De An. 415•22).
224 Charlotte Witt
provide the rationale for Aristotle's claim that potentialities are on-
tologically dependent upon actualities.
Fortunately, however, the statement that the young Sally exists for
the sake of herself as an adult, what I have been calling the mature
Sally, is just one way of filling in the teleological blank, and it is not
the most perspicuous formulation of Aristotle's position. What is
important for Aristotle is not a particular telos-Sally as an adult-but
rather the idea that the child exists for the sake of being a mature
person. Sally's telos is the type or species which she will realize, and
not the token or individual she will become. 10 On this view, the end
or actuality in question is the species. And, if this is right, then we can
make some sense of the ontological dependence of the potential on
the actual. For the human species, unlike the adult Sally, exists now,
and so it is possible that Sally's existence might be dependent upon it.
And we can understand the relationship between a capacity, like
sight, and its actuality, seeing, in analogous fashion. Aristotle's poinC
is that the existence of a capacity is ontologically dependent upon the
existence of a certain type of activity, not that a capacity is dependent
upon a particular realization of it. 11
Thus far I have only explained how it is possible for Aristotle to
ground the ontological priority of the actual in the teleological re-
lationship between actuality and potentiality. The next question is
why Aristotle finds it at all plausible to hold that the existence of a
potentiality, for example a developing substance like the youthful
Sally, depends upon the existence of the human species. The idea is
that Sally is a human being only potentially; Sally exists for the sake
of an end, in her case being an actual human being. What does this
mean? In the first place, it means that Sally's form or essence, to
whatever extent she has realized it, is the human form or essence. But
it is present in her only potentially. Perhaps she has not yet fully
IO In a discussion of three paradoxes of becoming in Met. z-e, G. E. L. Owen

interprets Aristotle as holding that in the statement 'a seed becomes a tree' the
phrase 'a tree' does not refer to a particular individual, but to the form or nature
(Owen 1978). Owen's analysis of change is obviously parallel to the analysis I am
offering of what is referred to as the goal in teleological judgements. This is not
surprising since becomings, whether artificial or natural, are all teleological pro-
cesses for Aristotle.
11 Does it make sense to hold that a capacity-the capacity to do geometry for
example-is ontologically dependent upon a corresponding type of activity? What
sense does the idea of a type of geometrical activity make? Do we have to add to
Aristotelian species a population of activity-types? Not really, for Aristotelian
activities are all activities of agents, and so what needs to exist in order for there
to be existing activities as the ends for capacities are active agents. An individual's
capacity to do geometry is ontologically dependent upon the existence of geome-
+-: . . . -~ • .......... +; ••..:..... - ...... ..-...-1 ............. +--.-..11 •• ..-1,....: ___ ..... .-. ..... -.-+--.
developed the ability to know, or-think of a baby-to move or

nourish herself. But these are among the basic capacities that con-
stitute what it is to be human. So how should we answer the question:
What is Sally? We have three options: We can stick to our guns and
say she is not a human being (but rather a very mature foetus), or we
can fiddle with the properties that constitute being human in order to
include Sally, or we can do what Aristotle does which is to say that
she is potentially human. But, to say that she is potentially human is
to say that she exists towards or for the sake of the end of being
human, and that end, Aristotle thinks, must exist. The actual existence
of that type of being, being human, is necessary for Sally's exist~11ce
not_~~ca_use Sally has to .!?~5-Q_!!!~_S.QQ!ethffiilruh~.futur.e_b.ut.b.ecause
she ha~.tQJ2.e...somethiag,,rigb.Ln.ow.
By conceiving of actualities or ends as species or types we have
found a way of making intelligible and perhaps even persuasive
Aristotle's claim that potentialities are ontologically dependent upon
actualities. And we have done so in a way that makes sense of his
examples and makes use of his arguments. But the account thus far is
not the whole story. To see why not, we need to consider two im-
portant features of Aristotle's thought that are relevant to the topics
we have been discussing. The first point is terminological; Aristotle
does not apply the label 'actuality' to the species or type as he should
according to the proposed interpretation. If the species or type were
the full realization or actualization of a given kind of being we would
expect Aristotle to refer to it routinely using his vocabulary of ac-
tuality. Generally,. how~yer, h~. ~(!sery~~ Jh?Ll!<Imi.nQ1Qgy_JQI.~.ither
the individual or. the form. 12 The reason for Aristotle's terminology is
that Aiisfotelian.speClesthemselves are existentially dependent upon
the existence of their individual members. 13 In other words, Aristotle
holdst~at ifthere were no. incJ.iyifiuaLh1,1m~ii].?ciiigs:,.tfit!~species or
type would n~L~~ist. In· the light of these considerations we need to
extend our original interpretation. We need to reconsider once again
what is implied by the fact that young Sally's telos is to be human, and
upon what her existence ultimately depends.
I have argued that Aristotle holds that the existence of a potential
human being like Sally is dependent upon the actual existence of her
kind of being, being human. This interpretation was intended to
12 For a typical passage connecting the term 'actuality' with the individual, see
Met. M. IO, 1087a10-25. For a list of texts exemplifying the connection between
actuality and form, see Bonitz 215a16-20.
13 Aristotle's Categories contains the most explicit statement of the dependency
of the species on its individual members (1"20-b9; 2a34-b6). Aristotle maintains
this position in the Metaphysics where he states that universals do not exist
• ~ • . . / h "
226 Charlotte Witt
capture the Aristotelian intuition that the goal or end that is relevant
to what Sally is is not her particular (possibly unrealized) future self,
but rather a certain kind or form of being. But I have just pointed out
that the type or species itself would not exist, and so would not exist
as an end for Sally, unless there were actual individuals of that type
or species. Ultimately, then, the existence of Sally, or any other
potential human being, is dependent upon the existence of individual,
actual human beings. This development of the species interpretation
preserves the idea that Sally's telos is the realization of a certain kind
of being, and it provides an existing telos for her. Moreover, it is
supported by the first of Aristotle's reasons for the ontological priority
of the actual, namely the idea that the potential being lacks its form
and the actual being possesses it. In Aristotle's words, 'Man is prior to
boy and human being to seed; for the one already has its form and the
other has not.' Pretty clearly Aristotle's examples of actualities are
individuals and not species. Species do not have forms, individuals
do. What is relevant abouJJheindividualsin at~leological context,
how~y~i~,!~.t!J~il" .fm;ms, and. .lh.aJ f<l~t js. prnp.edy c.apiur~·~L~Yi~ying
that Sally .~x:isJs. fQL!h~.~sak.~..9.Lbeing..a.matur.~. (a£!J!S!D.h!!!!1~JLbJ~~.ing;
ratherJhan..s~ing.1hat.fall~fuLt~s.ake...aLherselfJ!§,,.iUL:.td.µlt.
So, Aristotle thinks that if there were no actually existing human
beings, then a child like Sally· would not exist, at least not as a
potential human being. The reverse dependency does not hold. For
we can imagine an Aristotelian universe in which mature human
beings exist and have long life expectancies, but are unable or un-
willing to reproduce. The fact that no children exist, that there are no
potential humans, is not life-threatening to the adults. What they are
is actual since they have fully realized their forms. Actual human
beings are not existentially dependent upon potential human beings.
Hence actuality is ontologically prior to potentiality.
It might be helpful at this point to pause and consider the following
scenario that might appear to be an unfortunate consequence of the
position l have attributed to Aristotle. Suppose there were only one
adult and one child in the cosmos, one actual and one potential
human being. Now suppose the adult dies. On the view sketched
above, the child immediately ceases to be potentially human. This
imaginary example might seem so implausible as to cast doubt on the
correctness of my interpretation of priority in being. How could the
fate of one substance alter the nature of another?
To see how Aristotle might cope with this sort of example it is
important to locate its exact source in the position I attribute to him.
It seems to me that the difficulty originates in Aristotle's view that the
existence of the specie,s, a kind or type of being, depends upon the
existence of its individual members. In other words, the difficulty
arises not because Aristotle thinks that the existence of something
potentially human requires the existence of its end or goal-the exist-
ence of a certain kind or type of being; ratheL.the...difficulty.arises
because Aristotle thinks that the latter only exists if there are actual
·ma.ra
hl!ii!~_i_i_._l?_~~rigs.' trA:fisfotfeliera···e1iher speeies·coma-exisfquite
independently of whether or not it is instantiated or that a species
exists if it ever has been instantiated at one time or other, then the
imagined scenario could be defused. 14
Since I do not think that Aristotle would embrace either of these
possible responses to the example, however, we must find some other
Aristotelian response to it. I think he would reject the scenario out-
right on the grounds that any genuine species always has members or
is eternal (GA II. I, 731b24-732au). How exactly to understand the
eternality of Aristotelian species-the status, nature, and implications
of the position-is a large topic that cannot be broached here. Since it
is clear that Aristotle does hold that the species always have members,
he does have grounds to reject the troublesome scenario.
Aristotle's puzzling text that asserts the priority in being of the
actual turns out to yield a thumbnail sketch of his teleological vision
of the existence and careers of perishable substances. I would like
to end by mentioning two philosophically interesting features of
Aristotle's vision. F!rn!.,_.withlh~.JlQtjQJlJh<:itth~t;(!J:l.It! Wl1Ys Qfj:)~igg in
addition .1-2 ~!nc!s..9LJi.~ing, Aristotle has proposed a solution to the
oniOlogiCal difficulty surrounding his biological substances; entities
like ourselves that realize their forms in time. The problem here is not
one of specifying the form correctly by getting the right list of pro-
perties; rather, the problem is that Aristotelian form, however you
specify it, undergoes a process of realization. It is not here in a
14 Dory Scaltsas has proposed an alternative solution to the problem of furnish-

ing Sally with an end or telos. According to Scaltsas, all one needs is a referent
for Sally's end or actuality, but there is no need to insist that the end in question
is actually existing. So, one could specify Sally's end as being human even if all
human beings had ceased to exist. I believe that this is a possible position for
Aristotle, but not in fact his view. For Aristotle, as I say in the text, the eternality
of the species effectively blocks the counter-intuitive example of Sally's essence
changing because of the elimination of all actual human beings. I take it that the
eternality of species is not merely an empirical claim for Aristotle, but it is a
central principle of his view of nature. My second reason for rejecting Scaltsas's
suggestion is that I find it very implausible that Aristotle would consider some-
thing that doesn't exist to be an actuality. While 'actuality' does not simply mean
'existence' (potentialities exist too) surely existence is a minimum condition for an
entity being an Aristotelian actuality. Potentiality and actuality are relative terms,
like matter and form; if a potentiality exists so must its correlative actuality.
228 Charlotte Witt
child-except potentially. Aristotle's vocabulary of potentiality and
actuality-of ways of being-allows him to say that the child is human
even though her form is not yet actual.
The second philosophically interesting issue concerns Aristotle's
conception of teleology. If the child exists for the sake of an end,
Aristotle thinks, then the end must exist. In the case of the child the
end in question is being human, a certain kind of existence that we
can specify in terms of a range of human capacities and activities. The
jchild exists for the sake of becoming fully human; she is en route
!towards that end. Is Aristotle right in thinking that in order for a
teleological judgement of this kind to be meaningful, the end must
exist? Could a child or any other individual undergoing biological
development towards an end, be going towards a non-existent end? If
there were no end, in what sense would she be existing for the sake of
an end? And, if Aristotle is right in holding that the end must exist,
what does it mean to say that the end, a kind or type of being, exists?
On the view I have attributed to Aristotle, it means that individual
human beings exist. But perhaps Aristotle is mistaken on this point,
and different kinds of being, different species, can exist quite in-
dependently of whether or not there are any individual cases of them.
As is so often the case with Aristotle, a puzzle of interpretation has
become transformed into a series of philosophical puzzles; there are
questions concerning the correct understanding of teleology, the on-
tological status of species and their role in teleological judgements,
and the precise contours of our ontological categories. And, as is also
so often the case with papers on Aristotle, I will leave these philoso-
phical questions open for discussion. 15
15 I would like to thank the participants at the Oxford Aristotle Conference.
the UNH Colloquium on the History of Philosophy, and the Los Angeles Ancient
Philosophy Colloquium for their helpful questions and comments. I am grateful to
Mary Louise Gill and Theodore Scaltsas for a set of written comments, and to
Mark Okrent for his insights into the nature of teleological judgements.
IO
Essences, Powers, and Generic Propositions

JULIUS MORA VCSIK
What are essences? Aristotle and the student of the modern natural
sciences are both concerned with this question. Some recent analytic
philosophy assumes that the essence of a natural kind K will be a
collection of truths that are necessarily and uniquely of it and its
members. If so, essences should be expressed by sentences of the
form: 'All and only As are Bs' in which 'A' names the natural kind
under scrutiny.
This essay shows that such a conception of essence is inadequate. It
neither captures Aristotle's notion of the essential, nor that of the
modern biologist. For the essence of a natural kind includes con-
stituency, structure, and function. Thus an adequate analysis has to
account for all three of these ingredients. Functioning is, however, not
an all-or-nothing affair. Things can function better or worse. Thus an
entity can more fully embody its essence, or less so. But how could
mere generalizations capture this phenomenon? Let us consider the
following examples:
( r) Beavers build dams.
(2) Humans have eyesight.
It is in the nature of beavers to build dams, and having eyesight is a
part of the human essence. But our two sample sentences do not
translate into universal generalizations. Sentence (I) does not entail
'all beavers build dams', for there are disabled beavers or beavers
living under abnormal conditions; these creatures do not build dams.
Likewise, there are blind people and others living under circum-
stances under which they cannot see.
These few examples show already that some essential aspects of
natural essences cannot be expressed adequately by lawlike gener-
alizations. Sentences like (1) and (2) have a different semantics.
Linguists have called such sentences 'generics'. This essay looks at
some aspects of their semantics, and then relates these to Aristotle's
notions of potentiality and essence. This will help us in understanding
the dynamic aspects of Aristole's conception of nature.
©Julius Moravcsik 1994
230 Julius Moravcsik
I. GENERICS AND THE NATURAL
We shall consider sentences of subject-predicate form with bare

plural subjects and suitable predicates. 1 The notion of suitability
employed here will remain-to the annoyance of some readers-
intuitive, to be illuminated by the examples that follow.
Let us now compare (1) and (2) with:
(3) Germans drink beer.
(4) Cougars cross this range.
Though similar in syntactic form to ( 1) and ( 2), these sentences do not
describe parts of essences, but rather what happens usually or fre-
quently, but not always. Such sentences are called 'habituals'. We
have, however, only our knowledge of nature and semantic intuition
to guide us when we separate (1) and (2) from (3) and (4). The
following test helps. Sentences of this form that preserve meaning
even if we add modifiers like 'usually' or 'frequently' are habituals.
Applying this to (3) and (4) we obtain:
(3a) Germans usually drink beer.
(4a) Cougars cross this range frequently.
The additions preserve meaning. If, however, we do the same with (1)
and (2) we obtain: ·
( ia) Beavers usually build dams.
(w) Humans usually have eyesight.
These sentences do not express what (1) and (2) do, for while these
latter ones say something about the natures of beavers and humans,
the former do not. The claims about nature could be true even if (rn)
and (2a) were false. For example, environmental damage could render
most beavers deformed and most humans blind. Under such circum-
stances ( rn) and ( 2a) would be false, but ( 1) and ( 2) would still be
true.
Mere consideration of the meaning of a predicate will not tell us
whether it will figure in specifying essence. In
(5) Cheetahs run fast
we see a verb of activity with a modifier as the predicate. The predi-
cate can be used to describe members of various species on various
occasions. But in (5) this predicate contributes to describing the es-
sence of cheetahs. They are one of the fastest species of carnivores,
and their amazing running ability is functionally vital, for it enables
1 For linguists on generics see Carlson (1982) and Kratzer (1988).

Essences, Powers, Generic Propositions 231
them to be fine hunters even when facing competition from larger

predators.
The difference between essence-expressing generics and mere habi-
tuals may not be vital for syntax. It is, however, importaHt for meta-
physics and biology. In the cases considered we see natural potentia-
lities described. The habituals describe merely statistically significant
frequencies. Potentialities and powers, as we shall see, are basic
constituents of biological essences, both for Aristotle and for modern
biology.
In passing, let us consider the similarity between the essence-
specifying sentences we saw and sentences of the form:
(6) The whale is a mammal.
(7) The lion is tawny.
These sentences are semantically similar to ( 1) and ( 2), though they
have a definite article followed by a name for a kind as subject rather
than bare plurals. Since (6) and (7) are about species, one might
assume that the same is true of (1) and (2). But maybe essences have
different aspects. (6) says so'hlething that does not admit exceptions,
while (7) could be true only most of the time. Thus to say
(8) Whales are mammals
is to say something true but misleading, for it can be read as admitting
exceptions while (6) does not admit such a reading.
Can we subsume generics under some type that has a standard
logical form? Let us start with:
(9) All beavers build dams.
Reasons have been given already why this is not equivalent to or
entailed by ( 1 ). The latter says, roughly, that it is in the nature of
healthy beavers to build dams under normal circumstances. Dam-
building is a unique and essential power of beavers but not all of them
have it. This contrast is puzzling for the modern semanticist. For
the two sentences have important similarities; both are lawlike and
support counterfactuals. 2 ( 1) says something that applies to actual
beavers and ones that might have been, or will be, born, but unlike
(9) it allows exceptions. Therefore, it cannot be equivalent to, or be
entailing:
(10) Necessarily all beavers build dams.
Yet there is some modal element in ( 1). This helps to explain why we
run into trouble when we try weaker versions like:
2 For an explication of lawlikeness see Goodman (1955).

(II) Most beavers build dams.

( 12) With respect to any beaver it is highly probable that this animal
builds dams.
These sentences contain claims of high-frequency occurrences, but
these coincide with the manifestations of natural potential only under
optimal circumstances.
Perhaps we should construe 'potential' as a semantically primitive
modifier. We then generate sentences like:
(13) All beavers are necessarily potential dam-builders.
But (13) claims too much. Not all beavers have this potentiality, even
though it is a part of their essence. Deformed beavers possess only a
part of the essences of their species. We need for a correct version of
( 13) the modifiers 'healthy' and 'normal contexts' in the subject and
predicate slots.
Potentiality or power is more than mere habit or disposition, but
less than necessary structure or necessary aspect of species-specific
agency. As we saw already, not all potentialities are linked to gen-
erics; some-such as being a mammal-admit of no exception when
applied to a species.
Since we deal with biology and Aristotelian metaphysics, we have
been considering biological examples. In these cases the generics seem
to have existential import; there must exist something corresponding
to the subject expression. 3 A species is not just a property. It is a class
of actual and derivatively possible entities, causally connected, and
with a common causal origin. If tigers died out and later a very similar
species developed without any causal links to the extinct tiger species,
we would not call it a continuation or resurgence of the same species,
tiger, but rather a newly emerging species with striking qualitative
similarities to the earlier tiger species. Thus in these cases we have
both existential import and lawlikeness.
One might try to avoid having to acknowledge the unique semantic
force of essence-revealing generics, and try. to account for the
phenomena by adding pragmatics to the semantic analysis. Just as
Grice attempted to analyse conjunction, for example, by positing the
minimal semantic content familiar from truth-functional logic, and
adding conversational impliciturs, so one might try the same with
generics. But though one could accept the logician's analysis of con-
junction and then add the conversational implicitur that orders con-
juncts when time is involved in order to explain the difference between
3 I owe this argument to Joseph Almog, though he is not responsible for this
application.
'they got married and had a baby' and 'they had a baby and got
married', such strategy does not work in the case of generics. One
cannot fix universal generalization as the semantic minimum and then
add the rest of the force of generics as pragmatic factors, for a
universal generalization like (9) says too much rather than too little.
In order to start with a minimal content we would have to start with
existential sentences like:
(14) Some beavers build dams.
But this is now too weak. It does not have the lawlike force of
generics, it does not extend actual cases, and it says nothing about
essences. For spiders weaving webs and beavers building dams are
among key facts for defining the respective species. This is no more a
matter of pragmatics than matters requiring causal links, counter-
factuals, etc.
Leaving pragmatics we might look to scientific methodology for
help. Both in the natural and in the social sciences we often do not
study what actually takes place, since that is likely to be the result of
too many factors. Thus we study certain phenomena under idealiz-
ation. For example, we outline how gases behave under ideal con-
ditions, or we do the same with Economic Man, and then add the
intervening variables to describe what actually occurs. Could we
characterize the study of biological potentialities as the study of
phenomena under idealization?
One logical difference surfaces right away. For statements about
the Economic Man, or gases under ideal conditions, do not carry
existential import, while generics do. This difference is a clue to
deeper dissimilarities. Studying something under idealizations is to
project away from the actual into the realm of the theoretical that is
abstracted from experience. But studying a biological power of a
species is not to take something abstracted from actual occurrences
and changed into a different kind of entity and then theorize about it.
Rather, it is to take the actually occurring, and see in it order and
structure that goes beyond what just happens to be the case. For
example, thinking of pure water as H 20 is more like thinking of
something as idealized. But the healthy beaver is not like H 20;
rather, he is more like pure water in the sense of clean potable water.
From the point of view of abstraction from experience the healthy
beaver is more like clean potable water, rather than pure H 20.
These remarks show the futility of reducing potentiality to other
notions more familiar from modern logic and semantics. Let us take
this notion as an irreducible primitive, and investigate its properties.
The examples given above suggest that generics have a species as
their subject. What is a species? We need not think that these must be
either universals or particulars. Other entities such as sets or types
and tokens too do not fit that dichotomy. Perhaps the same is true of
species. 4
We can rule out universals and particulars quickly. A pure Platonist
might think that generics are about the attribute of beaverhood. But,
as we saw, this clashes with the intuition that these sentences have
existential import, and the condition of necessary causal link between
members of a species.
On the other hand, to think of a species as just the set or collection
of actual particulars that happen to be its members in the actual world
is also unrewarding. For the essence-revealing generics will be true
not only of the specimens that happen to be the actual ones, but also
of those that might have come into being and are related by what
could have been natural causal ties to the actual ones.
Neither Platonism nor nominalism seems to have the right answer.
Roughly, a species is a set of elements, some of which are actual and
have a common causal origin and are linked via other causal links,
while the non-actual ones are projectible from the actual cases. The
elements have also some unique necessary properties in common.
These properties include principles of individuation. When the species
is the designatum of a mass term like gold, there is a higher term with
principles of individuation, like 'metal', that defines the kind in ques-
tion. The members of the species partake of activities essential to
their development or maintenance. These activities allow better or
worse realizations, and have potentialities underlying them. Biology
studies these underlying potentialities. These can be exercised better
or worse; hence the link to generics.
Not all generics characterize powers. Consider:
(15) Humans have two legs.
The healthy, normal human has two legs, but not all humans do. But
even possession of legs is linked to the realization of potentialities
needed for functioning. Legs make locomotion possible for humans.
Those who suffer loss of limb would have problems surviving in a
world without technological aids.
Some functioning is needed for minimal persistence. The underlying
potentialities like those for breathing, or blood circulation, must be
exceptionless in their realization. This is not true of other poten-
tialities that bring us flourishing beyond the minimum.
The frequent mention of functioning in the previous pages might
suggest to some a possible link between the exercise of potentialities
4 See Mayr (1987).
and the position called functionalism in current philosophy of mind. 5
Functionalism aims at defining various mental capacities in terms
of what the exercise of these accomplishes. There is, then, some
similarity between functionalism and a delineation of essence partly in
terms of potentiality. Both deal with properties entailing dispositions.
But they deal with potentialities in different ways. Functionalism is
non-committal concerning the nature and ontology of the processes of
realization and the status of the agents. An Aristotelian approach to
powers and potentialities is not neutral on these matters. It is realist
with regard to powers, and has definite ontological commitments with
regard to the nature and status of the agents involved. Within the
Aristotelian system, if we encounter entities with different maturation
schedules and tnaintenance schemata, then we have two distinct
species, even if in terms of accomplishment the two processes lead to
similar results. No such stance is adopted by functionalism.
In summary, generics have their own nature. Some of them ascribe
potentialities to species. These potentialities are parts of essences. In
the next section we shall see how Aristotle captures the force of
generics, and how within his framework potentialities are seen as
constituents of essences. This view will exhibit similarities to that
taken by biological methodology.
II. GENERICS AND ARISTOTLE'S METAPHYSICS
Aristotle uses a locution which has been translated as 'usually' or 'for

the most part'. We shall see that in important contexts what Aristotle
has in mind is best rendered with the word 'naturally', and that the
locutions in question carry the force of generics. In Met. E. 2 Aristotle
contrasts both what is always and what is 'naturally' with the acci-
dental ( 10273 8- II). It will not do to translate the crucial phrase here
as 'for the most part', for the accidental need not be always just a rare
occurrence. Furthermore, as lines 20- 1 show, Aristotle wants so
contrast the objects of scientific study with entities that Jack this
status. Both that which is always and the 'natural' are on the scientific
side, while the accidental is on the other side. Since Aristotelian
science is not probabilistic, we can assume that he is contrasting
the eternal and natural with the accidental that is not a subject of
scientific study, and would not be such even if it occurred in some
contexts often.
Phys. II. 7 provides further evidence for this rendering. Here
5 For more on functionalism and Aristotle see Code and Moravcsik (1992).
Julius Moravcsik
Aristotle discusses adequate answers to 'why-questions'. In 198b5-6

he talks about what Ross and others have called efficient cause, that
is, showing that from one thing another will result, either unexcep-
tionally or 'naturally'. Since this discussion takes place within the
framework of articulating the ingredients of the nature or essence of
entities and the 'cause' in question is a part of this, the passage
establishes a clear link between the 'natural' and essence; a link that
could not exist if he had 'for the most part' in mind.
Finally, in Met. Z. II Aristotle tells us that one cannot leave matter
out of the study of the nature of things (rn36b22-3), and then (lines
28-31) goes on: 'An animal is something perceptible, and it is not
possible to define it without reference to movement.' But while some
of these movements or processes go on all the time, others will take
place naturally but not always. Again, here the rendering 'for the
most part' would make no sense. We define in the Aristotelian frame-
work natures, not frequencies.
To be sure, the distinction just drawn is not completely explicit in
Aristotle. This is most likely to be due to his teleological optimism.
Nature, for Aristotle, works in functional ways, and in his view what
prevents the realization of potentialities is likely to be the exception.
Most of what one might call motions today are seen by Aristotle as
the realization of potentialities. Potentiality too is for Aristotle an
irreducible basic concept. In Met. 8. 7 he describes it as: 'In the cases
in which the source of becoming is in the very thing which comes to
be, a thing is potentially all those things which it will be of itself if
nothing external hinders it.' Some entities have a built-in schedule
of development and maintenance, and the entities keep realizing
their natural potentialities in accordance with this schedule if nothing
hinders them. For example, beavers realize their dam-building poten-
tials and humans their potential to see. Potentiality thus contrasts with
actuality. Aristotle relies on this contrast when giving various deline-
ations and examples of potentiality. 6 Today one might assume that
actuality is the key notion, to be taken for granted and then used to
shed light on potentiality. But for Aristotle this is not so. The actual is
not just what happens to be the case, but-optimally-the realization
of potentialities. Thus actuality in this framework is just as teleological
as potentiality. 7 In the text cited above (rn49a13-14) the key notion
is not something happening to be the case, but some power being
realized.
6 For relations of potentiality to possibility see Witt (this volume: 215).

7 Witt (this volume: 219) discusses explanatory priority between actuality and
potentiality. According to my interpretation that relation is mutual.
In Met. e. 6, 1048a35-7 it is explicitly recognized that potentiality is
not explained by reducing it to other notions, but by examples and
analogies. This shows that for Aristotle potentiality is not the same as
possibility or frequency, but is a fundamental metaphysical concept.
Phys. I. 7 presents matter also as a basic non-defined concept. It is
described as what underlies various changes, and in passages con-
cerning various explanatory factors it is referred to as what makes up
basic constituency. Every kind has its own peculiar matter. Thus this
notion differs radically from the Cartesian notion of matter.
Met. H. 2, 1042b9- 1 r shows how matter and potentiality are related.
Here matter which .'is, potentially' is contrasted with the actual sub-
stantial elements in sensible things. These lines can be read in a
number of ways. On one view, matter and its potentiality are two
distinct entities; matter is actual, and it has potentialities to change in
various ways. According to another reading, Aristotle says only that
'matter is potentially this or that'. On this reading, matter exists
actually but has its nature always partly in terms of potentialities. Is,
then, matter an entity that possesses the potentialities as an entity
possesses its attributes, or is potentiality one of the constituents of
matter?
A third, and to my mind most plausible, reading attributes to
Aristotle the identification of matter and potentiality. This fits best
Aristotelian biology, the centre of his metaphysics of substance. An
animal is a series of potentiality realizations, determined by an innate
schedule. Once we listed the powers and the conditions of realization,
no additional entity needs to be mentioned; there is no matter over
and above these elements. But in the history of an animal at any given
time already partially informed matter is realized in further stages.
Thus at any given time there is already a substance partially actu-
alized, in a stage, with other stages still to be realized. 8
This view, as said before, contrasts sharply with the Cartesian view
of matter as extension. 9 It differs from the Cartesian view not only in
adding to its framework something, namely potentiality, but also by
not taking concrete bits of stuff as the basic parts of matter. Aristotle
has a much more dynamic view of the reality of sensibles. The world
contains processes of potentiality realizations, some constant, some
natural but admitting exceptions. Hence there are also movements
preventing the orderly realization of potentials. The schedule deter-
8 Frank Lewis (this volume: 250-1) shows that for Aristotle matter is essential
to a compound if it is capable of becoming it. On my hypothesis identifying
matter with potentiality, this follows naturally.
9 Code (1987a).
Julius Moravcsik
mmmg development and maintenance is not-for Aristotle-a

separate entity. It is neither just an attribute of substances, nor a
particular that is a part of substances. Within this view there is no
separate entity that is the schedule; rather there is a sequence
of potentiality realizations that proceeds 'humanly', or 'in beaver-
fashion', etc. This view fits Michael Woods's observation that 'Socrates
is a human' is not strictly a matter of relating humanity with its
instances. 10
This dynamic interpretation gives the 'naturally' a prominent posi-
tion. It designates what is for Aristotle not a marginal but an essential
ingredient in nature. Matter as potentiality can then be seen as realized
better or worse, more or less. With this in mind, let us tum to Met. Z.
8, 1034a6-8, where we are told that matter individuates. For example,
Callias and Socrates are the same in form but different in matter. This
doctrine can be read in different ways. Individuation for Aristotle
takes place within kinds. Thus first we locate the entities under con-
sideration within their species. Our example locates two entities with-
in the species human. Today many philosophers would say that what
makes the two humans distinct is that they have different properties.
But Aristotle cannot give this response. For on that account the
distinguishing properties will tum out to be accidental and not necess-
ary aspects of the entities in question. The necessary essential pro-
perties of humans are, however, that which we share. Could we say
that what makes the two entities different is that they are made up of
different collections of stuff? But is then the criterion for persistence
also difference in stuff? This is absurd, for by this criterion we would
,be always different humans since we as human organisms continuously
gain and lose material parts.
The identification of matter and potentiality solves the puzzle.
1
Callias and Socrates are different because they are constituted by
distinct sequences of potentiality realizations. There is one such series
going on over here, and another one over there. One is Callias, the
other S~crates. These sequences of potential~ty realizations are not
accidents of the two humans but necessary constituents. The 'material
parts' in the modem sense change all the time, according to the
potential to be realized and the schedule for realization. Both of these
differ also from species to species.11
The differences between interpretations of individuation can be
seen sharply when we consider maintenance and persistence. Accord-
10 Woods (this volume: 289).

11 More on entitative explanatory structures in Moravcsik (1991).
ing to the 'stuff' interpretation, Aristotle has to hold the view that
things persist only if they keep most or all of their material parts. It
does not take, however, knowledge of modern biology but just plain
common sense, ancient or modern, to realize that as we grow older
we do not keep the same parts of blood or flesh in our bodies. This
fact causes no problems for the dynamic view. For within that con-
ception persistence of Aristotelian matter consists of the right flow of
changes and thus of potentiality realizations. These leaq us from
infant stage to adolescence, and to the adult and old age stages. The
amount and kinds oi stuff that move along these lines is governed by
principles that differ from genus to genus.
Given this view, puzzles like the Ship of Theseus are not problems
for Aristotle. A ship has its parts replaced gradually, and out of the
discarded parts a new ship is built. According to the interpretation
proposed for Aristotle, the ship remains the same ship in spite of-
indeed on account of-all the rebuilding. It is a functional unit, and
its life-span consists of a series of potentiality realizaTwns... What
happens to the discarded stuff is of no concern to Aristotle in the case
of the ship any more than in the analogous case in which humans
discard some of their material stuff. The situation would not change
conceptually even if miraculously science could construct a new human
from the discarded stuff.
Humans are individuated within their species, then, on the basis
of different potentiality-realizing sequences. Humans persist because
these sequences will continue if nothing intervenes. Human per-
sistence requires pr~s,:esses of maintenance. If all came to a standstill,
there could be no nfafntenance and no persistence for humans or
other living things. The situation is obviously different for unchanging
abstract entities. But for living things persistence is maintenance,
and that can take place in better or worse ways. This is where the
potentialities that are realized naturally but not always enter. Main-
tenance varies depending on how healthy or even outstanding a
specimen is. Some will survive even in crippled form. This is the
primary place for the 'natural but not always' to take place. A beaver
that cannot build dams can survive in the company of other beavers
that can, and a blind human can survive in a caring community.
But the human who sees well and the beaver that builds dams realize
more of the natural potential of their respective species; and this
point is independent of the varying frequency of the healthy or the
disabled.
A potentiality is specified when we know its source, the sequential
structure that its realization must follow, and the final state that it
,..~,__-- ...... .
Julius Moravcsik
must attain and maintain. These elements account for what Aristotle
regards as the nature of an entity. 12
Let us now consider a sample of Aristotelian texts in order to
show that both the essence-revealing and the habitual is expressed by
Aristotle with the help of the locution mentioned.
In Phys. 198b34 Aristotle is talking about the normal way in which
teeth and other natural parts develop. Having these specific parts is
included in what is our nature. Aristotle's point can be summarized as
'teeth come into being (naturally) in such-and-such a way'. The same
point is made in GC 333b4-6.
There are many examples of the construction in question in Historia
Animalium. This is hardly surprising since this work deals mostly with
material that we would today assign rather to the naturalist than to
the biologist. Needless to say, the biologist-naturalist distinction is
absent from the Aristotelian framework. Various passages do show,
however, that discussions of the normal and healthy will crop up
systematically in Aristotelian biology. Let us look at a few passages
from the work cited.
In HA 56ob19-20 the normal maturation schedule of the common
hen is described. Further, in 5683 11-12 we find specifications for the
natural age at which certain kinds of fish are to conceive. Egg matur-
ation and ability to conceive are not accidental qualities. These are
fundamental aspects of certain kinds of animal life, since they are
parts of the biological regeneration process for some species. Thus the
normal healthy schedule and age specification indicate ways in which
some of the natural potential that is partly constitutive of animal
essence becomes realized.
In HA 587b30- 1 a natural correlation between flow of milk and
lack of menstruation is described. In this case too we deal with
phenomena linked to essential functioning. Book IX. 7 describes
conjugal fidelity within certain species.
In each of these cases the special Greek construction used by
Aristotle can be translated into modern English by the use of generic
sentences.
So far we have considered examples involving what we would de-
scribe as natural functioning constitutive of biological nature. As
noted earlier, generics can be used also to describe 'habituals', that is,
mere frequency or regularity. Thus it is not surprising that we find
Aristotle using the construction mentioned to express habituals. For
12 Lewis (this volume: 260) ponders whether potentiality exists before the
animal. On my view potentiality must exist before a given realization, but at the
same time potentialities exist always partly realized. To block danger of regress
we have the Prime Mover.
example, in HA VIII. 3, 592b28-9 some birds are said to feed usually
on certain kinds of grub. This is a matter of mere habit. If the birds
were transported into another habitat, they would avail themselves of
different kinds of food with the same nourishment value. Cases like
this support the hypothesis that the Aristotelian usages cover the same
dual range as the English generics.
Having looked at biological texts, we turn to passages in which
Aristotle discusses methodological matters and invokes the notion of
the natural as sketched in this essay.
In EN I. 3, 1094b11-27 and I. 7, 1098a25-30 Aristotle compares
approaches to different disciplines, and in the course of doing this
he states that we cannot expect the same kind of exactness in the
accounts-logoi-of all disciplines. As an example he cites politics,
and then turns to a discussion of human goods. In both of these
cases Aristotle thinks that we cannot expect complete exactness, but
only statements concerning what is 'natural' or 'for the most part'
(1094b21), both in the premisses and in the conclusions. The reason
given is that there is much variety in the things studied by 'political
science'-better rendered as 'the science of communal affairs', and
also in matters of human goods. Aristotle cites two examples to
illustrate his point. Both wealth and courage are goods, but in certain
contexts people can be harmed by these. Thus in such matters we can
aim only at what is naturally the case. In section 3 of the same book
he mentions as a contrast geometry, a discipline in which we can state
what is always the case.
Why is it that in the case of the goods one cannot state what
is always the case? Reflection on the examples should show that
Aristotle is not thinking of statistical probabilities or mere frequency
and regularity. Being courageous is one of our potentialities. Its
manifestation in the right way requires appropriate contexts. The
same consideration applies to wealth. Having wealth is good for
people provided that they acquire it in the appropriate contexts and
know how to use it. Hence the truth of the following: courageous
actions are good for the agent; wealth-constituting things are good for
the agent. We cannot say that these items are always good for the
agent, for the circumstances may not be felicitous or the agent may
not have an appropriate character structure.
One interpretation of these texts has them contrast mathematics
and geometry, where one can obtain what is always true, with ethics,
for example, where we can talk only about the usual and the regu-
larities that admit exceptions. 13 According to the other interpretation,
13 This seems to be the position taken in Mignucci (1981).
Julius Moravcsik
the contrast is between studying abstract relations holding between

unchanging elements at all times, and studying potentialities that
constitute human flourishing. Like any potentiality, these require
sound agency and appropriate circumstances, neither of which can be
guaranteed in every case. The appropriateness of circumstances is not
a matter of frequency, nor is the matter of having a sound agent a
matter of what the majority is like, as we know only too well. In view
of these considerations, the second interpretation recommends itself.
The texts just surveyed show why we should render the Aristotelian
locution as 'natural' in the cases in which natural potentiality rather
than mere habit is at issue, and they also show how the account of
potentiality sketched fits their claims.
III. THE PLACE OF POTENTIALITY AND

GENERICS IN ARISTOTLE'S PHILOSOPHY
In this concluding section we shall raise a few points concerning

this interpretation of generics, the 'natural', and potentiality and its
relation to other parts of Aristotle's thought and to some modern
biological methodological questions. First-L~EY)_s there [l,Q}pportant
role __f~._!~gf!!£!i~~_iti_Ari§~Q!1<,:~~LQ.i§cu~ion~)"._l(Q:g1it~E E~i~_2(lin_g?
This omission is hardly specific to these fundamental notions. Others
like individuation, the difference between mass terms and terms with
principles of individuation, the actual and the potential, etc. do not
figure in these discussions either. The absence of all of these key
metaphysical notions is due __!Q__1h~tas:LtPJlL!h~§~ g_Q_DQL<:!ff~,c:,Lthe
mosLgeneraLasp.ec.t£.QLs~llogisti.s:_re!!SQ!!JJJ£. We can 'reason across'
the mass-count distinction, or different principles of individuation.
The generics, or the 'natural', affect only some of the sciences. They
are useful in biology and psychology, but not in mathematics or
geometry. Furthermore, what prevents the realization of powers or
potentialities from time to time are what Aristotle would consider
accidents. But Aristotelian science does not deal with the accidental.
'Hence Aristotle can view reasoning about the natural as universal
\generalizations from a syllogistic point of view that abstracts from the
accidental.
Generics provide Aristotle also with a way to escape Platonism.
Activities that are parts of the essences of biological species are at
times polluted by other factors. Shall we then specify this part of
nature as ideal activity that is only partly realized in this world?
Philosophers are left with at least three options. The Platonist would
posit the Form of the (lCtivity, and say that the various actual activities
partake of this Form in various ways. As an alternative we· could

consider talking about the activities under idealizations. This ideal
activity is then linked to actual activities by the introduction of a string
of variables.
Aristotle wants to avoid the first way, and lacks the technical
apparatus for the second. But in the case of biological potentialities he
can fall back on the 'natural', which we express with generics. Some
essential activity takes place always, and others in the 'natural' or
healthy cases. This move helps to keep essences on the level of
potentialities realized, and not raised to Platonic heights.
We have seen the importance of generics for Aristotle's meta-
physics, and his correct use of these in connection with potentiality.
We saw also that he uses the same phrase to describe the habitual. He
does not comment explicitly on the difference between the two uses.
How far was he consciously aware of the dichotomy? We see that he
uses the notion of the general, and that he builds on it a part of his
conception of potentiality. If generics and their equivalent express
only frequencies and regularities, then they cannot help in formulating
accounts of essences. Thus we see that Aristotle needs the distinc-
tion, employs the phrases in the right contexts, and he relies on the
'natural' at the right points in his theorizing. The available evidence
does not allow us to go further than that.
Aristotelian metaphysics is not a metaphysics of concrete things as
ultimate elements of reality, but 1!..!!1.tl<wh)"'sLes-,·ef--µo.w.erand.pQten-
t@Ji!y. We saw also, however, that potentiality as matter individuates.
Does this mean that Aristotle studies primarily the particular? How
does this fit the Stagirite's view that science concerns the general? Our
answer has to be that Aristotle wants to study both the general and
the particular. The potentiality that a beaver has is general in the
sense that the species has it primarily. It is not abstracted from
individual cases. At the same time we saw that biological generics
have both existential import and lawlike projectibility. In this way
Aristotle can combine the insight that particulars are necessary for
powers, and that power manifestations occur as distinct individual
processes involving the development of substances, with the insistence
that potentialities are what we share and not what gives us individu-
ally different natures.
Let us at this point also show how Aristotle's account of the natural
helps the methodology of modern biology and medicine. For the
potentialities Aristotle regards as central and not occurring without
exceptions are also central to modern biology. Yet they do not fit the
strait-jacket of the analytic-synthetic dichotomy. Biology needs a way
of talking about things that are essential to a species, have existential
import, and cannot be expressed strictly with universal generalizations.

Aristotle's way of talking about the natural, and its equivalent in a use
of modem generics provides an alternative. It helps us to talk about
what is essential even if it is not fully realized. This also helps medicine
where at times the question is not, 'living or dead?' but, 'how much
life, how many capacities left, how much vital functioning left if we do
this or that?'
For modem semantics Aristotle provides not so much an alternative
1 but rather a challenge. His treatment of the study of nature shows the
!) need for an irreducible notion of potentiality and for ways of talking
!! about the 'natural' that is lawlike, projects beyond the actual and has
Uexistential import, and admits exceptions. We need these locutions
and to understand their meaning. If modem symbolic logic and the
semantics that philosophers built on it is to accommodate all the ways
in which we can express fruitfully the insights of science and other
rational enterprises, then these disciplines will have to come to terms
with the Stagirite's insights. 14
14 I am indebted for helpful suggestions to N. Cartwright, A. Code, J. Driscoll,
S. Levin, and T. Scaltsas; they are not responsible for the results.
IV
Matter and Form
II
Aristotle on the Relation between

a Thing and its Matter
FRANK A. LEWIS
I. OPENING REMARKS ABOUT

ARISTOTLE'S ONTOLOGY
Conspicuous entries in the early ontology of Aristotle's Categories and

Topics include individual substances, for example a, and their ac-
cidents, for example <f>. If some accounts are right, his ontology also
makes room for accidental compounds of individual substances with
their various accidents, for example a + <f>: in English, Socrates
seated, musical Coriscus, and the like. 1 The accident, ¢, is metaphysi-
cally predicable of a, and the accidental compound, a + </>, exists if
and only if in fact <P is metaphysically predicated of a. The fact that cf>
is metaphysically predicated of a is what grounds the truth of the
corresponding linguistic predication 'a is cp'. 2 Where <Pis (metaphysi-
cally) predicated of a, finally, Aristotle will also say that </> is an
accident of a.
With the introduction of the dual notions of form and matter in
©Frank A. Lewis 1994

1 The notation 'x + y' here is meant to express the notion of compounding,
first brought to my attention by Kit Fine, which I take to be primitive in Aristotle's
theory: other associations which this same notation may have in other contexts
should be disregarded. There is more on Accidental Compound Theory (ACT) in
Lewis (1982) and (1991); cf. Matthews (1982). In ACT, a phrase of the form
'Socrates seated' for Aristotle will not inevitably require the Russellian reading,
so that it refers to Socrates, who is the one and only thing that is seated, but may
pick out instead the accidental compound of Socrates with the accident, being
seated. Socrates + seated is not identical with its parent substance, Socrates, for
the two have different persistence conditions: only Socrates, but not Socrates +
seated, will survive the change when Socrates stands.
2 In my jargon, metaphysical predication is a relation between items in the
ontology: between a metaphysical subject, Socrates (say), and-not a predicate (a
linguistic item), but-a predicable, man (say) or pallor, without quotation-marks.
This is in contrast to linguistic predication, where what is predicated is a linguistic
item-a grammatical predicate. By 'predication' here I will mean metaphysical
predication, unless I expressly say otherwise.
Frank A. Lewis
Phys. I, the ontology is enlarged to include cases of matter, for

example m, and forms, for example !fl, where Aristotle thinks of the
individual substance as itself in some sense a compound of these two,
a r6& tv r0& (a 'this-in-this'), or ro eK rovrwv ('what is out of these'
( = matter and form)), or as a avvo},ov ('composite'), or a avvo).11 ovaia
('composite substance'). Thus, the individual substance, a, is identical
with the compound, m + !fl· In such a compound, further, the form,
If/, is (metaphysically) predicated of the matter, m (and m is subject to
!fl), in a way that, as Aristotle tells it in Phys. I, parallels that in which
an accident is (metaphysically) predicated of a substance (and the
substance is subject to that accident). I shall also say that If/ supervenes
on m. 3 It will be convenient too to have a word for the relation
between If/ and a itself, that is, m + If/: we will say that If/ is constitutive
ofm +!fl.
II. COMPLICATIONS: THE TOP-DOWN VIEW OF

MATTER, AND THE APPEAL TO HOMONYMY
In Phys. I, Aristotle talks as though the two cases of the (metaphysical)

predication relation run pretty much parallel to each other: at least as
far as the account of becoming is concerned, he feels that he can
safely generalize about both, before coming down to specifics about
either (I. 7, 189b30-2, cf. 19oa13-14). In fact, however, the two kinds
of case differ in fundamental ways. One obvious point of difference is
the multi-step nature of the model featuring form and matter. What in
Phys. I Aristotle identifies as the matter of an individual substance-
the bronze, for example, that is the proximate matter of the statue-is
itself a compound of some further form and some further matter:
and this further matter itself, no doubt, can be analysed in turn by
reference to still other layers of form and matter, ending only when
we reach prime matter. 4 This complication by itself makes for a
3 I use the word 'supervenes' here as a term of art for one half of the relation
of metaphysical predication (the other half is the accident-of relation); so my use
of the term has nothing to do with its use by Kim, Teller, and others in the
contemporary philosophy of mind. As a gesture towards keeping things straight, I
use the noun 'supervention', as opposed to their 'supervenience'.
4 Some, e.g. C. Williams (1982: 2u-19), have doubted that the notion of
prime matter is coherent; others, e.g. Charlton (1970: 129-45), that the notion is
present in Aristotle; I believe that the sceptics are wrong on both counts, but the
point is peripheral to my present concerns, and I shall not try to argue it here.
More recent sceptics, with an ingenious alternative account of the transformation
of the elements without prime matter, include Furth (1988) and Gill (1989);
Furth's account is criticized in Scaltsas (1989: 83).
A Thing and its Matter 249
difference with the other half of Aristotle's system of (metaphysical)

predication, which contains only a single layer, in which various ac-
cidents are predicated of their parent substances.
Aristotle offers quite different illustrations of the multi-step model.
In the biology, for example, he suggests that in a living animal, the
simple bodies are the matter for the uniform parts, the uniform parts
the matter for the non-uniform parts, and these finally the matter for
the whole animal. At the same time, in what appears to be a very
different and, I suspect, more central kind of case (see section VIII
below), he also describes how the male parent, acting as a craftsman
and using heat and motions in the pneuma as his tools or instruments,
imposes successive layers of form or soul-nutritive, sensitive, loco-
motive soul-on the material that comes initially from the female.
Still other complications have to do with how the theory of matter
and form meshes with the account of change, specifically, substantial
change, where an individual substance comes to be or passes away.
Aristotle's analysis here takes its start from the account of accidental
change. But not any old kind of accidental change will do: in fact, we
need a}),oiwmr;, change in sensible quality, as opposed to (say) change
in place, in particular, that variety of a}).oiwmc; that also involves
acting and being acted on by: in this last set of cases, a substance, a,
becomes actually </> from having been potentially </>, by the agency of
something (else) which itself is actually </> ( GC I. 7 )-a thing gets
heated up (say), as opposed to turning pale. 5 Here, finally, we have
the appropriate model for substantial change. In the account of sub-
stantial change too, in the standard case at any rate, we must be able
to insert reference to an efficient cause that appropriately involves the
form that will be the constitutive form of the product: in cases of
natural coming-to-be, the efficient cause will be an actual member of
the same kind as the offspring, and will have the form of the offspring
as its own constitutive form ('it takes a man to beget a man'), but in
cases of human craftsmanship, the constitutive form of the product is
present only in the craftsman's mind.
In other respects, the theory of form and matter and the account
of substantial change get worked out in ways that simply have no
counterpart in the account of accidental change. In substantial change,
significantly, the form of the product is implicated, not just with the
5 In cases of accidental change that exemplify ai).oiwm~ but not acting and
being acted on by, e.g. when Socrates becomes pale, there is no efficient cause
that is actually pale to transmit its pallor to Socrates, and the accident, pallor.
pre-exists only in the sense that before turning pale, Socrates is potentially pale.
cf. Ross (1924) on Met. Z. 8, 1033b5-6.
250 Frank A. Lewis
efficient cause, but also with the matter. In Met. H. 5, Aristotle

contrasts a 'normal' product, for example, wine, and the 'abnormal'
product, vinegar, to which the normal one can give way. The same
matter is present both in the normal product and in its abnormal
counterpart: Mwp (water) is the matter both of wine and of vinegar.
But it is the matter of the one KafJ' e.;1v Kai Karau) Bl6or;;, 'in accordance
with a disposition and in accordance with the form', and of the other
Kara arip1Jaiv Kai <f>Oopav r1v n:apa <f>vaiv, 'in accordance with a privation
and a destruction that is contrary to (its) nature'. Again, the water is
potentially wine, and also potentially vinegar, but the two potentia-
lities differ in the same way (1045a1-2): the potentiality the matter
has for the abnormal product, vinegar, is only a degenerate one.
Properly, matter is matter for, and has the potentiality for, the normal
product. 6
The addition of the notion of a power or potentiality (6vvaµ1r;;) to the
account of matter is crucial. In Met. e. 7, Aristotle argues that,
supposing the matter of a thing to be properly its proximate matter,
the matter is potentially the thing in again some suitably immediate
sense. For example, something is the matter of a house, only if it has
the appropriate passive 6vvap1:; for being made into a house once it
comes into contact with an agent with the corresponding active l5i!va1u:;
and once the other realization conditions obtain ( 1049a5- I 2); or it
must actually constitute a house, as a result of the interplay of such
6vvaw;1r;;. The upshot is a heavily top-down view of matter. According
to Aristotle's settled view, matter is always the matter suitable for
some end: 7 thus, the matter of a house includes rafters, beams, floor-
boards, and the like, but strictly, not just wood or timber. 8 And
the matter of a man is apparently flesh and bone (Aristotle's stock
account), or better, a body with the appropriate organs. 9
Aristotle's point is not just that something is the matter of a given
kind of thing, only if it can come to constitute (if it does not already
constitute) a thing of that kind-that bricks (say) are the proximate
matter. of a house, only if they can be built, or actually have been
built, into a house. Nor is he holding merely that it is a necessary truth
that bricks are the matter of a house, only if they are capable of being
built (if they have not actually been built) into a house. More radically,
his point is that it is essential to the bricks that they should be the
0 See also Phys. I. '9, 192"16-25.

7 Phys. II. 2, 194"27-1>15, PA I. I, 6391>26-640"1, 642"12-14.
H m1o1, n).fr601, sv).a, Met. H. 2, 1043"15, PA I. 5, 645"34.
9 E.g. PA I. 1, 642"12-14, De An. IL 1, 412"28-413"3. For discussion of the
complications involving .the various parts of the living animal, see sect. V below.
A Thing and its Matter
proximate matter of or for a house, and hence essential to the bricks
that they-should be capable of being built (or should actually have
been built) into a house. 10
The tie between matter and the passive power for being made into a
thing of a given kind suggests that something is the matter of a thing
only in so far as there exists a process for actually obtaining a thing of
that kind from that matter.11 This is a very different conception from
the view at work in Met. Z. 3, where matter is thought of the way a
metaphysician might think of it, as the result of mentally 'stripping
away' predicables to reveal the underlying subject. In Z. 3, there is no
suggestion that this 'thinking away' somehow traces in reverse the
actual processes by which the finished thing was obtained from its
matter, and the concept of matter seems to work entirely in abstrac-
tion from any actual or possible physical processes by which the thing
did or could come into existence.
The top-down view of matter is taken to get further support from
Aristotle's notion of homonymy. Very roughly, a thing is homony-
mously a so-and-so, according to Aristotle, when it is not able to
behave in the way typical of members of its purported kind. For
example:
A corpse has the same shape and form (n/v avrijv rou (Jxi/paror;; pop<fliJv), but yet
it is not a man. Again, a hand constituted in any old way, for example, a
bronze or a wooden one, cannot be a hand except homonymously, like the
doctor in a painting. For it will not be able to perform its own work, just as
flutes of stone cannot (perform) their work, or the doctor in a painting.
Similarly to these cases, none of the parts of a corpse is still of such a sort, I
mean for example an eye or a hand. (PA I. I, 64ob34-641 3 5)
A severed hand, for example, or the hand of a corpse, once had the
ability to perform as hands should but has now lost that ability. 12
We have seen that matter essentially has the potentiality for some
standard product, best of all, for a natural substance. Homonymy
gives the other side of the story: it has to do with a falling away
from the standard product, where a homonymous man (say) or a
homonymous hand is the lapsed state of a substance or its parts. 13
These two sides of the story-the passive power in the matter for the
product, or the homonymous product which has fallen away from
10 Cf. Z. 15. 1039b9-30: matter has the nature such that it is capable of being
and of not being; that is, it has the constitutive power, which may or may not be
realized, for being made to constitute a so-and-so.
11 Here I have been influenced by a remark in conversation by Michael Frede.
12 Here I suppress some complications in the account of homonymy noted in
Lewis (r99r: 251-2).
Frank A. Lewis
its proper state-both contribute to the strongly teleological cast of
Aristotle's philosophy of nature.
But the notion of homonymy has also been taken to suggest that in
certain cases, the tie between matter and thing. is even closer than
anything suggested so far. As Ackrill explains in an influential paper,
in the case of living things:
The matter is itself 'already' necessarily living. For the body is this head,
these arms, etc. (or this flesh, these bones, etc.), but there was no such thing
as this head before birth and there will not be a head, properly speaking,
after death. In short-and I am of course only summarizing Aristotle-the
material in this case is not capable of existing except as the material of an
animal, as matter so in-formed. The body we are told to pick out as the
material 'constituent' of the animal depends for its very identity on its being
alive, in-formed by psuche. (Ackrill 1972-3: 126, his emphasis)
I am inclined to think that the view Ackrill suggests of matter is
correct, for the set of cases he discusses here, and given a certain
conception of matter. 14 For example, imagine an animal whose con-
stitutive form or nature is the capacity for perception of a given
particular sort, that is, it is (a variety of) sensitive soul. Aristotle's
definition of soul in the De Anima apparently commits him to the
view that the matter of sensitive soul is a body with the appropriately
functioning sense-organs. 15 On this view, the sense-organs that help
13 Cf. PA I. 1, 641a18.
14 The concession is a guarded one. I suppose that what Aristotle identifies as
the proximate matter of an animal, namely its living body, is essentially endowed
with the kind of life typical of that animal. But this is not to say that, in general,
matter is essentially alive, or even that on each of the different possible con-
ceptions of the proximate matter of an animal, the proximate matter essentially
has the kind of life typical of the animal itself (on the 'different conceptions'
hinted at here, see sect. VIII below, esp. n. 57). For other reservations about the
account in Ackrill, see Lewis (1991: 252-3).
15 The definitions in De An. IL 1 have a special logical character that Aristotle
comments on by implication later in De An. II. 3. A definition of soul, he says,

should not be a common formula, which applies in the same way to all the
different varieties of soul, but is proper or peculiar to none (oii<J1:v(I~ ... u?1v ovrwv
/Oto:; Myo; oMt rwra '[() oiKciov h'Ui TO awpov el,>o:;, II. 3, 414h26-7). Instead,
the definition he gives in II. r is more properly a definition-schema, and applies
differently to each of the different kinds of soul. But since the different kinds of
soul are arranged in a natural sequence, the way in which the general schema for
definition is understood varies systematically according to the place the kind of
soul under discussion occupies in that sequence. A given kind of soul, by the
lights of Aristotle's definition-schema, will be the relevant first actuality, paired off
with a natural body that potentially has the relevant kind of life, that is, a body
characterized by the relevant set of organs. Accordingly, sensitive soul, to take a
particular kind of soul, is the first actuality of a body whose characteristic organs
are sense-organs.
constitute the matter of the animal must themselves, in their own
right, have the capacities typical of sensitive soul. 16 If so, two con-
clusions apparently follow: sensitive soul is essential to the sense-
organs, so that the form is essential to the matter on which it super-
venes; and what appears to come to the same thing, the sense-organs,
and hence the matter of the animal, cannot exist independently of the
animal itself. In general, then, in at least a certain range of cases-
perhaps artefacts are excluded-the form that supervenes on a given
matter is also essential to it, and matter is not only essentially matter
for a given product, but also in an important set of cases it cannot
exist independently of the product.
It is important to see that these conclusions go well beyond the
points urged above, that matter is teleologically driven, and that it is
essential to bricks (say), which are the matter for a house, that they
be capable of being built (if they are not already built) into a house.
These claims by themselves come nowhere close to the conclusions
drawn from the Appeal to Homonymy and backed up by the De
Anima, that the form or soul that supervenes on a given matter is
essential to it, so that the matter of an animal cannot exist without
that form or soul.
The idea that in certain cases a form is essential to the matter on
which it supervenes, lies behind a number of often puzzling, if not
outright sceptical, conclusions. One consequence some have argued
for 17 is the idea that Aristotelian matter is not (in our sense) purely
material: that the proximate matter of an animal is necessarily living
matter, while there is no level of matter below its living flesh and the
like that escapes the animal's form or soul-no matter, that is, that is
not necessarily living. The case for this view of (biological) matter
and its negative consequences for the functionalist interpretation of
16 I take it that Aristotle's view that the matter of sensitive soul must be a body
with the appropriately functioning sense-organs cannot be the triviality that the
matter in question includes the appropriately functioning organs in virtue of its
being matter to sensitive soul; instead, I take it, it is a pre-condition of being
matter for sensitive soul, to be explained independently of being matter to
sensitive soul, that the matter be alive in the appropriate way.
17 Other apparent consequences include the suggestion by Ackrill that if a
given form is essential to the matter on which it supervenes, then it supervenes
essentially on that matter-that is, the relation of supervention itself is essential.
Again, Ackrill appears to suppose that if the form is essential to the matter, then
it will not be possible to pick out or identify the matter in a way that is logically
independent of the form, so that (on Ackrill's view) it is hard to make sense of
the idea that the form and the matter together make up a compound in some non-
standard but still recognizable meaning of 'compound'. These views are discussed
in greater detail in Lewis (1991: 254 and n. 18).
254 Frank A. Lewis
Aristotle's philosophy of mind in particular are spelled out in a tren-

chant paper by Burnyeat. 18 .
Finally, the Appeal to Homonymy has also breathed new life into
an old account of the nature of the distinction between a thing and its
matter. According to the Appeal to Homonymy, as we have seen, in
the case of animals at least, matter exists, only by virtue of actually
constituting an animal of the appropriate kind. Homonymy has also
been taken to suggest that the tie between a thing and its matter is
even stronger. According to the doctrine of homonymy, apparently,
the body that is the proximate matter of an animal is its living body;
and surely the body in this sense just is-that is, is identical with-the
animal itself . 19 Accordingly, the animal and its proximate matter are
outright identical.
This is a bold view, but it receives support from what is at first sight
a very different source, for the identity of a thing with its matter is
also a component in a proposed solution to Aristotle's problem of the
Unity of Substance. This is our topic in the next section.
III. THE UNITY OF SUBSTANCE
For Aristotle, questions about the unity of substance centre around

his analysis of the individual substance as a compound of form and
matter, in which the substantial 'last' form is metaphysically predi-
cated of or, more particularly, supervenes on the proximate matter.
Given this analysis, what makes it that the result of (metaphysically)
predicating a (substantial 'last') form of the relevant proximate matter
is a unity? Aristotle's answer to this question is given in Met. H. 6,
and on the dominant reading of the chapter, the answer is a highly
surprising one. In the closing lines of the chapter Aristotle is usually
taken as saying that the proximate matter of a thing and its form are
identicat. 20 If a thing's matter and form are truly identical, then the
question of how together they make up a uniW simply does not arise.
18 Burnyeat ( 1992).
19 I take this to be the argument at work in Kosman (1987: 379).
20 For a sampling see Peck ( 1953: xii, xlvii, 1II (note on I. 20, 729•34) ), Sellars
(1976: u8), Rorty (r973: 394), Hartman (1977: 99 n. II ('somehow identical'-
apparently in a way that stops short of 'full identity', pp. 75-80) ), Kosman ( 1987:
378), Balme (1987 repr.: 295), Gill ( 1989: 8 ('somehow identical')). What Aristotle
actually says is this: 'the proximate matter (~ eaxar17 v.ic17) and the form are the
same and one, the one potentially, and the other actually' (1045br7-r9). The
reference to potentiality and actuality should be understood in light of the earlier
claim in. the chapter: 'the one is matter, and the other form, and the one (is)
Now by 'form' in this solution to the unity question, Aristotle is
most reasonably supposed to mean individual form (that is, a form
that is unique to the particular individual substance it typifies). If,
further, seemingly following a suggestion in Met. z. 6, we imagine
that the individual form of a thing and the thing itself are also identical,
then we get a three-way identity between the form of a thing, its
matter, and the thing itself.
But if these three are identical, we are urgently in need of an
account of how they can possibly differ from one another, as surely
they must. A classic statement of the kind of view required is in
Sellars (1967). Sellars asks about the sense in which a compound
material substance is a composite or 'whole' which (on Sellars's view)
has an individual form and an individual piece of matter as its 'parts':
his answer is that
the individual matter and form of an individual substance are not two in-
dividuals but one. The individual form of this shoe is the shoe itself; the
individual matter of this shoe is also the shoe itself, and there can scarcely be
a real distinction between the shoe and itself. What, then, is the difference
between individual form and matter of this shoe if they are the same thing?
The answer should, by now, be obvious. The individual form of this shoe is
the shoe qua
potentially (a sphere), and the other (is) actually (a sphere)' (1045"23-4),

where the supplement, '(a sphere)', is suggested by the explanations at 1045"31-
3. The claim that x and y (here, matter and form) are one and the same, can be
understood in a variety of ways: (i) x is the same as, or even identical with, y
(different suggestions have been made about the nature of the sameness relation
in question, which on some accounts is identity, on others, something less than
this, for example, 'accidental identity'); the relation cannot be symmetric, how-
ever, if we are to make sense of Aristotle's immediate qualification, 'the one
potentially, the other actually'; (ii) unusefully here, each of x and y is one, or is
one and self-identical; (iii) unusefully again, since it is the problem of unity that
Aristotle is here trying to explain, x and y together make up a unity; (iv) there is
something, z, such that x and y both are z, for example, Aristotle and Speusippus
are one and the same thing, namely philosophers; or even (v) x and y are
genidentical: they are different development stages of one and the same thing (so
Schwegler ad foe. )-but it is unclear how this answers questions about the
synchronic unity of matter and form. The reading (i) gives the identity thesis for a
thing's proximate matter and its form mentioned in the main text. My own view is
that (iv) is the right reading here. Aristotle is repeating the point from rn45a23-4
discussed above, that there is one and the same thing, in fact, one and the same
kind, k, such that the matter is potentially a k (it potentially constitutes a k), and
the form is actually a k. I shall reserve further defence of this reading, and of the
interpretation that goes with it, for a different occasion (see Lewis (1992)), but it
is worth going on record here that as far as the sheer translation of the Greek is
concerned, the identity thesis in (i) is by no means assured.
Frank A. Lewis
(piece of some appropriate material or other-in this case leather) serving

the purpose of protecting and embellishing the feet.
The individual matter of this shoe is the shoe qua
piece of leather (so worked as to serve some purpose or other-in this case
to protect and embellish the feet).
Thus, the 'parts' involved are not incomplete individuals in the real order,
but the importantly different parts of the formula
(piece of leather) (serving to protect and embellish the feet)
projected on the individual thing of which they are true. (Sellars 1967: 118,
his emphasis)
This point of view is echoed in Kosman (1987):
The term 'matter' refers to entities taken a certain way; matter is a principle
of being or nature, and not a category of entity in nature. The world does not
divide up into instances of matter and of form; matter and form are ways of
thinking about the entities of the world relative to an explanatory or de-
scriptive context. (Kosman 1987: 362, his italics; cf. 390)21
On the 'projectivist' or 'constructivist' line of thought offered in these
passages, the distinction between a thing and its matter is an inten-
tional one: it reflects different ways of thinking about one and the
same thing, but not any difference between individuals, complete or
otherwise, 'in the real order'. 22
We now have on the table two sources of support for the identity
thesis for the relation between a thing and its matter: the problem of
the Unity of Substance, which I will say nothing further about here; 23
and the concept of the animal's matter as its living body, based by
Ackrill on the Appeal to Homonymy and supported by Aristotle's
definition of soul in the De Anima. I shall argue that we can safely
grant the concept of matter for living things described, without buying
into the projectivist or intentional view, or supposing that a thing and
its matter are ever identical.
21 Kosman elaborates his views in his contribution to this volume: see in
particular his rejection of a theory of 'relational being' and of any 'separation of
matter and form' (his sect. II). In contrast to a lintel (say), where the matter 'is
itself something other than a lintel', Kosman argues that 'there is nothing other
than a human being, that is, there is no other thing, that is the matter of a human
being' (his emphasis), on the grounds that the flesh and bones of a human being
'are themselves human': reasoning of this kind is discussed in my sect. VI.
22 For criticism of the constructivist viewpoint in general, see Devitt (1991),
and for criticisms of the view in Aristotle, in addition to sect. IV below in the
main text, Lewis (1992) and esp. Charles (this volume).
23 An account of Aristotle's treatment of the unity of form and matter in the
compound material substance in Met. H. 6, without the identity of form and
proximate matter, and without any of the other identities envisioned by Sellars, is
given in Lewis (1992).
IV. BLOOD AS THE MATTER OF AN ANIMAL
The counter to the intentional view of the distinction between a thing

and its matter that comes most quickly to mind is that it makes
Aristotle's project no longer a part of ontology. This consequence is
cheerfully admitted by Sellars, but it puts an uncomfortable distance
between Aristotle and what is usually thought to be his proper subject-
matter. There is also textual evidence against the view in Phys. I,
where Aristotle introduces his notion of matter as subject to form by
analogy with the relation between an individual substance, Socrates
(say), and his accidents, musicality and the rest. On the majority
reading at least, 24 Aristotle here thinks of matter as a subject that
persists through the possession of, in turn, the lack or privation and
subsequently the form-for example, the bronze that is transformed
from a lump into a statue-much as Socrates can exist both while he
is unmusical and also after becoming musical. 25 Given the assumption
of a persisting matter, it is easily argued that a thing and its matter are
not identical. If a single matter is both the 'before' matter of a thing
and its concurrent matter, then evidently, the thing's matter exists
when the thing does not, hence, matter and thing cannot be identical.
But, it may be objected, perhaps Aristotle changes his mind about
the assumption that there is a persisting matter, or the assumption is
suited better to artefacts than to examples in the biological world,
where Aristotle's best candidates for substances are found. If we are
to rebut the identity thesis, not just for the case of the bronze statue
and its matter, but also for animals and their matter, we must find a
case in the biological world where one and the same matter both can
be made to first constitute the animal-is its 'before' matter-and is
also the concurrent matter of the animal.
Blood appears to give us the example we need. Blood plays a part
both in the generation of an animal, and in nourishing the animal
once it is formed. Hence, it has good credentials for being both the
concurrent and the 'before' matter of the animal. 26
24 See e.g. Code (1976) in contrast to Charlton (1970) and Barrington Jones
( 1974).
25 Similarly in Met. z. 7, Aristotle argues that coming-to-be is impossible if no
part of the product pre-exists, and that matter is both a part of the product and
also what undergoes the coming-to-be, as required ( 1032b31-1033•1, cf. 1032"19-
22). See also Met. A. 3, 983b6-18.
26 The place of blood as both matter 'for' and matter 'of' the living animal has
already been noted in Freeland (1987). Ackrill (1972-3: 128-33) had already
asked whether an answer to the Appeal to Homonymy might involve recourse
to matter of a lower level than the non-uniform parts. It is worth emphasizing
Frank A. Lewis
Blood, on Aristotle's account, is the last stage of nutriment in

blooded animals, worked up by the animal's internal heat. It exists for
the sake of nourishing the animal (PA II. 3, 65oa32-b19), and hence is
a good candidate for its concurrent matter. Aristotle asserts outright
in a number of places that 'blood is the matter of the whole body'. 27
And in a striking analogy, he suggests that:
[T]he system of blood-vessels in the body may be compared to those water-
courses which are constructed in gardens: they start from one source, or
spring, and branch off into numerous channels, and then into still more, and
so on progressively, so as to carry a supply to every part of the garden ...
These things are done because water is the material out of which the plants in
the garden grow ... In the same way, Nature has provided for the irrigation
of the whole body with blood, because blood is by nature the matter of it all
(navrbi; v2,, nt</>vKe roiJro) ... the blood (or its counterpart) is, potentially,
the body (that is, flesh-or its counterpart). (PA III. 5, 668a14-27, following
Peck's translation)
Meanwhile, blood also supplies the residue in the female which is the
'before' matter out of which the animal is first constructed. 28 Again,
Aristotle uses analogies to describe how the animal is first formed:
Beginning at the heart, the blood-vessels extend all over the body. They may
be compared to the skeleton models which are traced out on the walls of
that even if the argument for the non-identity of a thing and its matter applies
to animals and their matter as well as to the statue and the bronze, blood will
not be the matter of the animal in exactly the way in which the bronze stands
to the statue. In particular, blood is worked up into the various bodily parts
by a variety of 1uc,11;, where the matter exists only potentially in the product.
On µi~1i;, see the next to last paragraph of the present section, and esp. sect. VIII
below.
27 PA II. 4, 651 3 14, cf. III. 5, 668a4, GA I. 19, 726b5-6.
28 Blood produces various useful residues in the mature animal that are relevant
to the generation of an offspring. The residues produced will depend on the
amount of heat the animal has available to concoct the blood. The female is
weaker and has less heat (Aristotle discusses the connection between heat and the
degree of perfection of an animal at GA II. l, 732b32-733b23), hence the residue
in her case will be the rnra1i~via or menstrual fluid which is the nutriment or
matter out of which the Jvvaµ1r; or power in the male semen will be able to put
together an offspring (GA I. 19, 726b30-727•2, 727b14-728•33, cf. Met. H. 4,
1044"34-5). But only the male has sufficient heat to concoct the blood further so
that it becomes semen: the female is as it were an infertile male, and is unable to
concoct semen out of the blood, owing to the coldness of her nature (GA I. 19,
728•17-21). The semen contributed by the male, however, which is again a useful
residue from the bloodlike nourishment, contains the right heat and motions to
fashion an offspring out of the materials the female supplies (GA I. 21, II. r,
734b19-735a29). But Aristotle is emphatic that the male contribution is not part
of the matter of the offs?,ring.
buildings, since the parts are situated around the blood-vessels, because they
are formed out of them. The formation of the uniform parts is effected by the
agency of cooling and heat; some things are 'set' and solidified by the cold
and some by the hot ... (a8) As the nourishment oozes through the blood-
vessels and the passages in the several parts (just as water does when it stands
in unbaked earthenware), flesh, or its counterpart, is formed: it is the cold
which 'sets' the flesh, and that is why fire dissolves it. As the nourishment
wells up, the excessively earthy stuff in it, which contains but little fluidity
and heat, becomes cooled while the fluid is evaporating together with the hot
substance, and is formed into parts that are hard and earthy in appearance,
e.g. nails, horns, hoofs and bills ... (a18) The sinews and bones are formed,
as the fluidity solidifies, by the agency of the internal heat; hence bones (like
earthenware) cannot be dissolved by fire; they have been baked as it were in
an oven by the heat present at their formation. (GA II. 6, 743"2-21, Peck's
translation; cf. 744°11-745"18)
Finally, at GA II. 4, 74oh34-7, defending the identification of genera-

tive soul and nutritive soul, Aristotle points out that the matter on
which each is brought to bear is one and the same:
Similarly, just as in animals and plants themselves at a later stage, the or1vap1:;
('power') of the nutritive soul makes growth out of the nourishment, using
heat and cold as its instruments ... so in the same way at the very outset. it
sets the natural object that is coming to be. For, the matter by which the
object grows is the same as that out of which it was originally set, so that the
i5vvaµ1; which makes the object too is the same. (GA II. 4, 74ob30-6, my
emphasis)
It is perhaps barely possible that Aristotle here has in mind two

distinct matters, the 'before' and concurrent matters respectively, and
is committed only to holding that they are the same in kind-they are
both alike blood. But his conclusion about generative and nutritive
soul is that they are straightforwardly identical, and he gives no hint
that in saying that the 'before' and concurrent matters too are the
same, he has something weaker in mind. The natural reading of the
passage, in line with Aristotle's views in Phys. I, is that a single
accumulation of blood is involved in the composition of the animal
and is its single matter, from the very first stages in the animal's
formation (where the matter is constituted by the Karaµ~vw), and
taking in the entire life of the animal thereafter. If this is his view, the
matter of the animal exists when the animal itself does not, and even
within the biological world, a thing and its matter are not identical for
every choice of matter and thing.
But there are objections to be considered. First, blood is not plau-
sibly the proximate matter of the animal; so we have yet to show that
260 Frank A. Lewis
the animal and its proximate (concurrent) matter are not identical. 29
The short answer to this is that if the animal is not identical with its
more remote matter, blood, then not even its proximate matter is
identical with the animal. For, suppose the animal and its proximate
matter are identical, that is (where pm(a) is the proximate matter of
an individual substance a),
a= pm(a).
Putting 'pm(a)' for 'a', we have
pm( a) = pm(pm(a) ).
And in general, we can prove that
a= pm(a) = pm(pm(a)) =pm( ... (a) ... ),
and so on, all the way down to the blood. 30 But we already know that
the animal is not identical with the blood. Accordingly, the animal is
not identical with even its proximate (concurrent) matter.
That would seem to be the end of the issue, were it not for a second
and more troubling objection, that blood is not in fact the concurrent
matter of the animal at all. At most, someone might argue, the
passages quoted about blood and nourishment show that blood is the
daily matter out of which the animal grows-as it were, its daily N,
01!-but not that it is ever the concurrent matter of the animal. The
reason is simple. Aristotle tells us that the animal has the non-uniform
parts as its matter, and the matter of these in turn is the various
uniform parts of the animal, flesh, bone, and the like; but (the ob-
jection goes) there is no lower concurrent matter of the animal than
these to be had. Flesh, for example, is a uniform part, hence, it is
flesh 'all the way down' (better, perhaps, 'all the way through'). 31
However closely we look, Aristotle's anti-atomist view of the process
of mixture (pi~1;) by which flesh is produced guarantees that in prin-
ciple there is only flesh there to be discerned. So there is no concur-
rent matter for the animal lower than its various uniform parts-flesh,
bone, and the rest-including the blood now coursing in the animal's
veins; but the blood and, below blood, the other materials-earth,
29 I owe this objection to James Bogen.

30 This streamlined version of the argument that appeared in an earlier draft is
due to Kit Fine. The argument assumes that each of 'pm(a)', 'pm(pm(a))', and
the rest, is uniquely referring: that at each stage from the animal on down, a
single conception of the matter of the thing is at work, hence at each stage, there
is a single proximate matter of that stage. If, however, as I suggest in sect. VIII
below, Aristotle has more than one conception of the matter of a thing, the
argument will not be valid without some disambiguation to ensure that the same
matter-of relation is in force throughout, cf. n. 50.
31 I am indebted here to a remark in conversation with Myles Burnyeat.
air, fire, and water-out of which the uniform parts have been made-
these are only the animal's 'before' matter, and not its concurrent matter
at all.
These difficulties make this a good point at which to do some stock-
taking. I have cited the example of blood to argue that even among
living things, there is a case of (concurrent) matter that is not identical
with the animal itself; from this, as we have seen, it follows that the
animal and its proximate (concurrent) matter too are not the same.
But this result must still be squared with what we seem to know about
the proximate matter of an animal. As we saw in section II above, a
number of considerations point to the conclusion that in certain cases,
a form is essential to the proximate matter on which it supervenes.
But if the form of the animal is essential to the proximate matter on
which it supervenes, as well as to the animal itself, how can Aristotle
avoid the conclusion that in the case of living things at least, proxi-
mate matter and thing are identical?
At the same time, by thinking of blood as the concurrent matter of
the animal, we have also raised an new question about how far down
an animal's matter goes. A prime candidate for an animal's proximate
matter is its body, uniform and non-uniform parts and all, and it is
notable that these various parts are at the same time spatially deter-
mined parts of the animal. Like the animal itself, its spatially deter-
mined parts must be alive to be properly members of their respective
kinds-they must be living flesh and living bone, or living hands and
living eyes. So the question arises, is there a concurrent matter in the
case of a living thing below the level of its spatially determined parts,
that is not necessarily living in the way that the spatially determined
parts are? To repeat: Is there a lower level of matter, all the way
below the uniform parts, where we find stuffs that are not spatially
determined parts of the animal, and are not essentially alive, but
which are still parts of the animal's concurrent (non-proximate)
matter? I shall suggest that the answer to these questions is 'yes';
working out the details will take us below blood, which I have argued
gives one case of the concurrent matter of the animal, to the lower-
level matters-earth, air, fire, water-that lie beyond the reach of the
substantial form or soul of the animal altogether. 32
32 To this point, we have been interested in the question: Once blood is mixed
to make flesh and the rest, can we properly describe the blood as the concurrent
matter of the resulting compound? But if blood is in fact the concurrent matter of
the animal, and the blood itself has a concurrent matter, then this further matter
will also be the animal's concurrent matter (hence, we must take the 'the' in 'the
concurrent matter of x' with a grain of salt, since a thing can have many
concurrent matters). On this showing, the different question: Is there a concur-
Frank A. Lewis
These together give the two issues I shall address in the remainder
of the paper. A useful preliminary to both sets of questions will again
be the study of blood: this will be our topic in section V that follows
next.
V. THE METAPHYSICAL STATUS OF BLOOD AND

OF OTHER PARTS OF THE ANIMAL
(i). Heat as an external principle relative to the blood

At PA II. 2-4 Aristotle sets in place the theoretical framework within
which blood can have the physiological role he has earmarked for it.
An important first step in his study is to come to terms with the
conflicting views of his predecessors about the varying degrees of heat
of different animals, and this in turn requires some distinctions about
hot and cold, and later solid and fluid, which are the principles of the
physical elements (II. 2, 648b9-10, cf. II. 1, 6463 14-21). Of the
distinctions Aristotle draws, perhaps the most interesting is his idea
that a thing's heat can be derived from within it or from without: the
heat is either foreign (d)),6rpw~)-boiling water, for example-or it is
intrinsic or per se-for example, a flame.
Aristotle introduces this distinction in order to rationalize the dif-
ferent senses he finds in which one thing, A, can be hotter than
another, B: A heats something C more than B heats C, or A melts or
burns C more than B does, A produces a more violent or painful
sensation than does B, or A cools down more slowly than B, or it
warms up more quickly. Given these different meanings of hotter
than, one thing will not be hotter than another in all the available
ways at once. The divergences, Aristotle suggests, are explained by
the foreign/intrinsic distinction: for example, a flame (which is hot per
se) burns you more than does boiling water (which is hot per accidens),
but the boiling water causes the stronger sensation to the touch.
The salient feature of blood is that its heat is externally derived. Its
heat comes not from its matter, namely (predominantly) earth and
water, both of which are cold, 33 but from its efficient cause. Blood
rent matter of the animal, below the spatially determined parts, that is not itself
necessarily living? plausibly is answered by pointing not to blood, but to the
concurrent matter of blood, namely (predominantly) earth and water (cf. n. 33
below).
33 Meteor. IV. II, 389b9- 18, IV. IO, 389a19-22, adds air as a constituent in
blood, while GC II. 8 insists that all uniform parts are composed of all four
elements.
must be concocted out of the animal's food by the animal's nutritive

soul, which 'makes' growth out of the nourishment using heat and
cold as its instruments (GA II. 4, 74ob29-741a3).
The fact that blood's heat is external to it is reflected in uncertainty
in how we are to talk about blood and its heat. In the Parts of
Animals Aristotle notes that:
In one way, blood is hot, for just as if we were to signify boiling water with a
single term, blood is spoken of like that; but in respect of the substratum and
what it is that is blood (o non: ov aiµa foriv), it is not hot. And in itself
in a way it is hot, and in a way not; heat will be included in its definition, just
as pallor is included in the definition of a pale man; but in so far as (it is hot)
due to external influence (Kara naOo~), blood is not in itself hot. (II. 3,
649b22-8)34
As this and other passages show, people disagree over whether

blood is hot or not, even over whether in itself blood is hot or not.
There are good reasons for either answer: the body is blood's natural
or proper context, and there blood is hot; but the heat is not its
own.
Even if the blood's heat comes from its efficient cause, not from its
matter, why must its heat remain external to it? In the case of human
sexual reproduction, for example, the efficient cause (the male parent)
successfully transmits his form so that ultimately it becomes a prin-
ciple internal to the offspring as well. On the story Aristotle tells, the
male is a member of the same kind as his offspring, and the heat and
motions present in the male semen that go proxy for the form work on
the blood-like residue or Karaµftvia from the female to 'set' the foeta-
tion in the way that rennet works on milk to set it (GA II. 4,
739b20-30). Here, the form or soul of a human being is involved in
the efficient cause by way of the male parent, and the form governs
the process by which the foetation is first formed in the uterus 'from
without': it is an external principle of the change. But, notably in
Phys. II. 1, Aristotle distinguishes a natural object from an artefact on
14 'Blood in itself is not hot': at 649°n-14, on the subject of ice, Aristotle

goes further: ice is solid 'accidentally' (rnra m1µ/fe/J1JK0:;), that is, due to external
influence, but 'in itself' ice is fluid. On blood, see further at Meteor. IV. 11,
389°9- IS: 'So long as blood, semen, marrow, rennet and the like keep their
proper nature they are warm, but once they perish and lose their proper nature,
(they are) no longer (warm) ; for what is left is the matter, which is earth or
water. So there are two views about them, and some regard them as cold, some as
hot, seeing that as long as they are in the natural organism (ev r~ </>i!o'ei, cf. PA
II. 3, 649°29-30), they are hot, but when they are separated, they solidify.' (Lee's
translation, with corrections.)
Frank A. Lewis
the grounds that the former has a nature or internal principle of

behaviour, while the principle of behaviour of an artefact remains
external to it. Accordingly, the offspring must internalize its principle
of behaviour, and this it does with the formation of the heart, from
which the subsequent 'putting in order' (<5zaK6aµ17mc;) of the animal's
body is derived; there comes a time when the foetation 'must manage
for itself, just like a son who has set up a house of his own indepen-
dently of his father' (GA II. 4, 74oa7-8, Peck's translation). 35 Nothing
of this sort happens, however, in the case of blood. Again, blood's
heat comes from its efficient cause, not its matter, but in this case the
heat does not become an internal principle. Even when the animal is
fully formed and viable on its own, it needs a heart to supply the heat
necessary to concoct the nourishment into blood. On this account, the
heart is the 'principle of the blood' (apxi/ rov a'iµawr;, PA III. 4,
666a24), and is the source (apx1/) and cause of its heat and fluidity (III.
5, 667b22-31).
(ii). Heat as a form-analogue of the blood

These results raise a striking question. If the heat of the blood is
external to it in the way Aristotle says, on what grounds do we treat
blood as a natural object? The· difficulty has to do with questions
about blood and form. Heat is essential to the blood, and part of its
definition; and the heart is the apx1/ of the blood because it supplies
the needed heat and fluidity. But the heat is never an internal prin-
ciple. So heat is part of the form of blood, if anything is; but at
the same time, it is not an internal principle or nature, contrary to
Aristotle's assertion in the Metaphysics that the form of a thing and its
nature strictly so-called are one and the same (LI. 4, rnr4b35-rnr5a11).
Without an internal principle, moreover, given the discussion in Phys.
II. 1, blood is more like an artefact than a natural object.
The proper solution seems to be this. To be blood in the full sense,
blood must have heat from the appropriate source. But heat is not
part of the form of blood. Rather, it is the form-analogue in blood,
and similarly, whatever material parts underlie blood's heat are only a
matter-analogue for blood. So blood cannot lose its heat and still be
properly blood; but since the heat is only a form-analogue and not a
form proper, it is not necessary for the heat to be an internal principle.
35 On the question of when the foetation acquires an internal principle of its

own, cf. also Met. e. 7, 1049•13-18, on the contrast between cases where the
principle (apx~) is in the thing itself that has it, and where some outside principle
is still needed (eKcfvo r5e frtpa; dPX~; &fra1, •16- 17).
But blood is still more like an artefact than a natural object, since its
behaviour as blood is not governed by an internal principle. Unlike
that one-time artefact, the human offspring (say), blood never acquires
an internal principle, and is more artefact than natural object for all of
its existence.
The characterization of heat as the form-analogue of blood is taken
over from the discussion of pseudo-substances-primarily artefacts,
but there are other examples as well-in Met. H. 2. Aristotle is
notoriously willing to use artefacts as examples in setting out his
analysis of individual substances into form and matter. But the
examples contradict his official view, and in H. 2 he admits that such
items as honeywater, bundles or books or boxes, thresholds and
lintels, or breakfast and supper, or broadening the range of cases,
winds, or even ice (mentioned also in the discussion of blood at PA
IL 3, 649b11-14), are not substances at all. 36 H. 4, 1044b3-20 adds
eclipses and sleep to the list, and most interesting of all, H. 2, rn42b31
adds hands and feet. But while these are not themselves substances,
none the less, 'the analogous thing holds in each of them; just as in
substances what is predicated of the matter is the actuality [that is, the
form] itself, so in the other definitions, (it is what resembles this)
most' (Met. H. 2, rn43a4-7). He goes on to suggest that in defining a
threshold or a house, we should mention the appropriate materials
(wood or stone, or bricks and timbers) suitably configured, or that ice
should be defined as water solidified or condensed in such and such a
way. Even with solidification (avvbva(oµevov, rn43a4, cf. PA II. 3,
649a15-19), however, ice is not a substance. 37 I am suggesting that
blood too belongs on the list of things cited in Met H. 2 as pseudo-
substances.
36 Cf. H. 3, ro43b21-3.
37 Aristotle writes: 'It is clear from these that if substance is the cause of each
thing's being (airia roD Elm1 crnaro1'), we must inquire among these what is the
cause of each of these thing's being (rel a'inov wD elvw ro1!rnJv c1<<wrov, •3·-4).
Now substance is none of these (ovoev roiirwv, "4 [= ice, etc., cf. wvrwv eKaarov,
"3-4]), not even coupled together (mll'<foa~6,111:1·ov, •4 [sc. with the appropriate
differentiae, e.g. solidification]), but none the less the analogous thing holds in
each of them, (ro43"2-5) (the continuation is quoted immediately above in the
main text). Here, I take it that 01W:v ro1)rw1.• ( 1043"4) has ro1!rwv iiKaarov ("3-4)
as its antecedent, referring to ice and the like, not rovro1:; ("3) with Ross, referring
to the various 61mf>opai that are the a/nm· roD dvw of a thing. I also take
avw5va~6111:1·ov to govern an implicit reference to the relevant differentiae, contrary
to Ross's view that the appropriate supplement is: when coupled with matter. The
question is not obviously settled by the Greek by itself. although this is obscured
by translations going back at least as far as Schwegler's, which proceed without
notice as if the supplement Ross suggests were actually present in Aristotle's text.
266 Frank A. Lewis
(iii). Extension to the other parts
My next step is to suppose that the results obtained for blood hold for
the run of uniform and non-uniform parts of an animal as well. Life to
the hand or eye, for example, is like heat to blood: in its natural,
proper context, the hand or eye is alive, as blood is hot-but like the
heat of the blood, the hand's life or the eye's is externally driven, in a
way that is determined by the form of the whole animal. The form or
soul of the animal is an internal principle governing the animal's
typical behaviour; the same form or soul also governs the behaviour
of the animal's hand or eye, but it is an external principle relative to
them. 38 The living hand or eye of the animal, then, is only a pseudo-
substance, and has the form or soul of the animal as its constitutive
form-analogue. In line with this, as we have seen, the hand and the
foot find themselves along with ice and the rest on the list on which I
have argued blood also belongs in Met. H. 2.
I have suggested that like blood, the uniform and non-uniform parts
of an animal in general have no internal principle of behaviour. We
will have grounds for doubting this claim, if it can be shown that the
case of blood is sufficiently anomalous. Three arguments attempt to
distance the case of blood from that of the other bodily parts. (i) In
contrast to the animal's flesh, bone, or its various non-uniform parts,
blood is not strictly a part of the animal at all. 39 Against this, how-
ever, we have blood expressly included in the standard lists of bodily
parts at PA II. 2, 647brn-16 and HA III. 2, 511b2-4. (ii) Whatever he
may think about blood, at Phys. II. 1, 192b9-14 (cf. Met. Z. 2,
1028b8- 13) Aristotle outright asserts that the parts of an animal in
general do have an internal principle of behaviour. But at the begin-
ning of Met. Z. 16 at 104oh5-16, he takes it all back: even the fact
that they have a principle of change originating from a point within
the joints will not give the parts of an animal a principle of the
appropriate sort, and will not make them substances. 40 (iii) A further
reason for thinking that unlike blood some parts of an animal at least
do have an internal principle has to do with the relation between the
form or soul of the whole animal and the animal's various parts. I
·' 8 For the view that the uniform and non-uniform parts of an animal have no
internal principle of change of their own, cf. also Waterlow (1982: 88).
39 If the text can be trusted, Aristotle explicitly denies that blood is a part of
the animal at PA II. IO, 656°21. Aristotle's descriptions of blood as situated in the
blood-vessels 'as in a container' point in the same direction (PA II. 3, 650"34,
0 7-8; the same metaphor appears also at PA III. 4, 666"18).
4° Cf. Waterlow (1982: 48-9, 53, 88-9). Notice also the </>opn at Met. Z. 2,
rcu8°10, and E's </>1!ai;1 <5i: </><itu;v civw at Phys. II. I, 192°9, in place of </>1)a1:1 ptv in
the other manuscripts and printed by Ross.
have supposed that the animal's soul is an internal principle relative to

the animal itself, but that the various parts have that same form or
soul as an external principle of their behaviour. This picture is con-
tradicted if instead the form or soul of the animal is distributed over
the various relevant parts, so that, for example, the eye has sight as its
form proper, where sight is a principle that is internal to the eye. 41 But
the evidence on the point seems equivocal. 42 At De An. II. 1, 412b18-
413a3, for example, Aristotle is running a three-way comparison
between artefacts, the parts of an animal, and animals themselves:
given this parallel, sight is 'the substance in accordance with the
definition' (ovuia ... ~ Kara rov 261 ov) of the eye; but the tone of the
1
passage is counterfactual in nature, and Met. H. 2 has already taught

us to treat comparisons of the kind under way here with caution.
Again, if the form or soul of the animal is just the sum of the forms of
its relevant parts, how is the animal itself a unity, over and above the
collection of its parts?43
I have argued that neither blood nor any of the uniform and non-
uniform parts of an animal has an internal principle of behaviour, but
only a constitutive form-analogue. We are now in a position to move
up a level, and ask about the form of the whole living animal-the
form that supervenes on the matter to produce the animal itself. The
study of blood has suggested a sense in which the form of the animal
is present as an external principle of behaviour to the different parts
that make up the animal's matter-what quota of form, then, can be
left over to supervene on the matter to produce the animal itself? Can
this plausibly be the same form that is present to the animal's parts?
But can it plausibly be anything less? This brings us to the first of the
two main issues raised at the end of section IV above. If the form is
the same in the two cases-both for the animal and for its matter-can
we escape the conclusion that the animal and its proximate matter
themselves are one and the same? This will be the topic of section VI,
which follows next.
41 I am indebted here to a comment by Charles Young.
42 The question at issue is whether the different 'parts' of the soul (whichever
these may be) are Xluptarov ramµ. De An. II. 2, 413hq-15, cf. xwp1arov 1\Ctra
µiydJm;, III. 4, 429aII-I2, xwptarov µ£yi;Oet, III. 9, 432a20. Plato takes a distributive
view of soul in the Timaeus (Tim. 69d-72d), and arguments in the first book of
the De An. assume a distributive view of soul (De An. I. 4, 408a10-13, I. 5,
4rrb7- 14; but the passages are polemical, and plausibly, the distributive view is
adopted here only because it is a feature of the positions under attack). Aristotle's
final view appears to be that the different 'parts' of soul are not 'locally distinct',
see De An. I. 5, 411hr5-30, II. 2, 4r3°u-414"3, esp. 0 27-32, with Hicks (I90T
327, n. on 413°29).
43 De An. I. 5, 4rrb5-14, cf. Waterlow (1982: ch. 2, esp. p. 90).
268 Frank A. Lewis
VI. THE FORM OF THE LIVING ANIMAL,

AND THE ALLEGED IDENTITY OF THE
ANIMAL WITH ITS PROXIMATE MATTER
Aristotle sometimes suggests that the non-uniform parts are the

matter for the whole living animal. But a given non-uniform part is
not only a part of the matter of the animal, but also a spatially
determined part of the animal as well. If, then, the animal's matter is
just the sum of its living non-uniform parts, it seems as if the dif-
ference between the whole animal and its matter must be just shape,
or a mode of spatial determination. And this gives too 'thin', that is,
too nearly 'Democritean', a concept of form for the whole animal,
appropriate more to artefacts than to living substances. In the case of
artefacts, it can seem plausible that their constitutive form-that is,
the form that supervenes on the proximate matter-is nothing more
than their shape or mode of spatial determination: the familiar
examples of the bronze sphere or statue come to mind here. 44 But
Aristotle rejects this conception of form for the bona fide Aristotelian
substances, namely the living creatures of the biological world, where,
he argues, form is far more than just shape. 45
44 Met. fl. 6, rn45"25-33, Z. 3, rn29"3-5. The thresholds and lintels and other
examples from fl. 2 are also relevant, though here it is fair to notice Aristotle's
warning that 'a purpose may exist as well' in such a case (ro43"8-9).
45 Aristotle's argument for this point comes in the course of a criticism of
Democritus in the celebrated chapter that begins the PA. Aristotle is arguing
against the view he finds in his Pre-Socratic predecessors, that a complete expla-
nation of nature can run in terms of just matter, or (slightly better) matter in
combination with the glimmerings of an efficient cause, Empedocles' Love and
Strife, for example, or Mind, or the like, leaving off the necessary reference to
the appropriate form or nature or power (c>Vvaµ1:;): 'But if man and the animals
and their parts exist by nature, we must say of each part-of flesh, bone, blood,
and all the uniform parts, and similarly of the non-uniform parts such as face,
hand, foot-in virtue of what, and in respect of what sort of power (Kara noiav
,fovap1v), each of them is such as it is. For it is not enough to say what it is made
of, for example, of fire or of earth. If we were speaking of a bed or some such
thing, we should be trying to define its form rather than its matter, for example,
the metal or the wood ... we should have to speak also of its shape (uxqpa) and of
what sort of thing it is in respect of its form. For its nature in respect of
conformation (~ ... Karn u)v µoprf>~v ¢1!m:;) is more important than its material
nature' (PA I. I, 640°17-29, following Balme). Despite this, in the passage
immediately following Aristotle cites the case of Democritus to suggest that the
concept of form too may not come so very easily (the continuation is partially
quoted in sect. II above). Aristotle's own words, uxq1w and pop¢~, in the lines
above and later at [,34, like the artefact example, suggest a 'Democritean' concept
of form as just 'shape'. But there are familiar cases, which Aristotle collects under
the label 'homonymy', where the shape of a hand, for example, is evidently
A Thing and its Malter
Alternatively, if the matter of the animal is its living body, sense-
organs and all, as the De Anima suggests, then it seems that there is
no difference between the animal and its matter. If so, then ap-
parently the conception of the form that is constitutive of the thing-
that is, that supervenes on its proximate matter and is distinctive of
the whole animal as opposed to its matter-is not just thin, but has
shrunk to the vanishing-point.
Contrary to these suggestions, I shall try to make sense of the view
that the proximate matter of the animal is its living body, sense-organs
and all, and that the form that supervenes on the proximate matter is
the full-fledged soul or nature of the animal. This gives a concept of
the constitutive form of the animal sufficiently rich to escape Aristotle's
anti-Democritean strictures. But how are we to overcome the re-
dundancy that apparently has this view in its grip? We have supposed
that the form of the animal is already present essentially to the body
and its sense-organs. If now the same form also supervenes on the
body and its organs-what can adding the form for a second time
contribute to the body that is not already present to it? In particular,
if the body by itself already has all the form the animal could ever
need, doesn't this tend to show that the animal is identical with its
living body? And since the living body is the proximate matter of the
animal, does this not lead in turn to the Sellars-Kosman conclusion,
that (in the biological world at any rate) thing and proximate matter
are identical? 46
These objections fasten on the idea that the animal's form is es-
sential also to the animal's body. If the form is essential both to the
animal and to its body, how do the animal and its body differ? A
possible response lies in the idea that although the same form is
essential to both, it is essential to them in different ways. For example,
the body and its sense-organs are essentially capable of sensation. But
sensitive soul is essential to the body and its organs in a special way:
they are essentially capable of sensation, because sensitive soul is their
constitutive form-analogue. That is, sensitive soul is essential to the
present, but no real hand exists. Aristotle concludes that Democritus' account is
oversimplified (l.iav ... ani.wc; i;'ip17ra1, 641"6). Democritus sees no difference
between natural objects and artefacts, a wooden hand (say) made by a craftsman;
with a real hand, however, its form must include more than just its shape.
Accordingly, artefacts can misrepresent the true nature of form and matter in two
different ways: they not only fail to have an internal principle of behaviour; but
also typically (although not exclusively) they employ too 'thin'-that is, too
nearly 'Democritean'-a concept of form.
46 Seen. 21 above.
270 Frank A. Lewis
body and its sense-organs, but at the same time, it is an external

principle of their behaviour. Exactly this combination occurs in the
case of blood and its heat: as we have already seen, Aristotle explains
in the case of blood that 'in itself in a way it is hot, and in a way not;
heat will be included in its definition, just as pallor is included in the
definition of a pale man; but in so far as (it is hot) due to external
influence (Kara nafJor;), blood is not in itself hot' (PA II. 3, 649b25-8).
I am supposing that the relation of soul to the body and its sense-
organs is the same as that of heat to blood-it is essential to them, but
at the same time, it is an external principle of their behaviour.
Now an external principle of this sort is first introduced under the
influence of the form or soul of the male parent, which is an external
principle of behaviour relative to the offspring and its parts, but an
internal principle relative to the male parent itself. Once the creature
is fully formed, however, and independent of the male parent, it has a
form or soul that is an internal principle of its own behaviour; but the
form remains an external principle relative to the body and its sense-
organs. Accordingly, once the animal is fully formed and can 'manage
on its own', the form of the animal performs in two different roles: it
is the constitutive form-analogue of the animal's body and its various
organs, but it is also the constitutive form simpliciter of the whole
living animal. ·
On this story, sensitive soul is essential to the animal, and it is
essential also to the animal's body and its sense-organs. Even so, the
animal and its body differ, because the substantial form that is an
internal principle of behaviour relative to the animal and is its con-
stitutive form proper, is only an external principle relative to the body
and its organs and is only its constitutive form-analogue. Similarly, it
is not redundant that sensitive soul should supervene on the body.
The body by itself has no internal principle of behaviour: the form or
soul that governs its behaviour is external to it. But this form must be
internal to something-if not the male parent,i then the animal itself.
The forrh, then, is an external principle of behaviour relative to the
body, only because it is internal to the animal itself. And its being
internal to the animal means that it is the constitutive form of the
animal and at the same time supervenes on the animal's proximate
matter. 47
I will mention just one difficulty with this proposal. I suggested
47 With the conclusions of this paragraph, compare the suggestion by Haslanger
(this volume: sect. VII) that the substantial form is an intrinsic integration
principle with respect to the compound material substance, but an extrinsic in-
tegration principle with respect to its matter.
earlier that, like blood, each of the uniform and non-uniform parts
of an animal has no internal principle of behaviour, but only a con-
stitutive form-analogue. It is a far cry from this to the conclusion that
the living body itself, which is organized out of all the living uniform
and non-uniform parts, also has no internal principle of its behaviour-
that for it too, the form of the animal is an external principle of its
behaviour and, hence, not its constitutive form proper but only a
constitutive form-analogue.
Worse still, there is a textual reason for thinking that the form or
soul of an animal is an internal principle of behaviour relative to the
animal's body. In the account of soul in De An. II. I, Aristotle says
that the body that is the matter of the animal is a natural ( </>vazKi>V}
body-that is, a body with an internal principle of behaviour (De An.
II. I, 412 3 20, cf. br6-17). This, he says, is why an animal's body
cannot properly be compared to an axe (II. I, 412bn-17). 48 By way
of mitigating the difficulty, notice that Aristotle does not say what
internal principle the animal's body is required to have. I have
claimed that where the body is a body capable of sensation (say),
sensitive soul is not an internal principle of the body's behaviour. But
this is consistent with that body's having some other form or soul as an
internal principle of behaviour, namely the varieties of soul respon-
sible for nourishment and growth (II. r, 4r2ar3-r5). So it is open to
us to argue that the living body, that is, the body that is essentially
capable of sensation, has no internal principle for its behaviour as a
body of that sort, namely capable of sensation.
I have tried to defend the idea that its soul is 'a principle internal to
the animal, but not to its (living) body. This line of thought, if it can
be sustained, may pay dividends for Aristotle's psychology. On this
view, for example, it will be reasonable to say that the animal sees
with its eyes, but not (strictly) that its eyes see. And in general, it
might be argued (although I shall not pursue the topic here), the unity
of consciousness of the animal is due to the fact that it, the animal,
and not its different sense-organs, is the centre of awareness. 49
Our main subject in this section has been the possibility of a dis-
48 The details of the proposed comparison itself are obscure. The example of
the axe is introduced to help explain the requirement in the definition of soul that
soul is the substance in accordance with the logos of a body of a certain kind,
namely a natural body. On this showing, we might expect the axe to be compared
to the animal itself, not its body; but the comparison is eventually rejected
because the axe 'is not a <f>va1Kov awµa', with its own internal principle of be-
haviour, which makes it seem that the counterpart of the axe is properly the
animal's body, not the animal itself.
49 Here I am indebted to discussion with David Keyt.
Frank A. Lewis
tinction between the proximate matter of the animal and the animal
itself. Independently of questions of how successfully its proximate
matter can be pried apart from the animal itself, however, it is also
worth asking whether there is some lower-level matter that can be
separated from it. 50 Even if the proximate matter were so bound up
with the form of the animal that we cannot tell the animal and its
proximate matter apart, how far down does the form reach? Is there a
concurrent matter of the animal, at a level below the body and the
uniform and non-uniform parts that are also its spatially determined
parts, that is independent of the form of the animal? This is the topic
of the next section.
VII. IS THERE A CONCURRENT MATTER OF

THE ANIMAL BELOW THE ANIMAL'S
SPATIALLY DETERMINED PARTS?
SOME REMARKS ABOUT MIXTURE
According to Aristotle, the uniform parts of an animal have for their
matter the simple bodies or so-called elements, earth, air, fire, and
water. Like the uniform and non-uniform parts above them, the
simple bodies are functionally defined-each has its distinctive
epyov or work-but unlike the higher-level parts, the simple bodies
have a distinctive 'work' that is not spelled out in terms of their
contribution to the living animal. So it appears that they can at once
make up the living animal-be its concurrent matter-even while
they themselves exist independently of the animal's form or soul.
Aristotle's study of blood helps reinforce this idea. 51 Given that
heat is not an internal principle of behaviour or nature for the blood,
and equally solidification is not part of the nature of ice, it is clearly
appropriate to ask what there is to blood apart from its heat, or to ice
apart from its solidity. What is it that is ice so long as the correct
solidification is present, but all along is something else in its own right
besides? What constitutes ice is in its own right water. Similarly, the
material basis of blood constitutes blood in the presence of the ap-
propriate heat, but all along is something else in its own right-what
50 If we are forced to say that the proximate matter and thing cannot be
separated, but attempt to make out a notion of lower matter that is independent
of the form of the animal, how are we to view the argument in sect. IV above,
that if thing and proximate matter are identical, then the identity goes all the way
down, so that even the lowest matter and the thing are identical? Presumably, at
some point there will be a shift in the notion of proximate matter, so that at some
level the identities begin to fail, cf. n. 30 above.
51 This point has already been noted in Charles (1985, unpub.).
Aristotle identifies as a quantity of water and earth in the appropriate
proportions.
Plausibly, the distinction between blood and the material basis for
blood can be replicated for all the uniform parts of the animal. As we
have already seen, in its natural, proper context, flesh or bone (say) is
alive, as blood is hot-but like the heat of the blood, the life of flesh
or bone is externally driven in a way that is determined by the form of
the whole animal, which is an external principle relative to them. In
addition to the living flesh or bone, then, doing (or at least capable of
doing) its proper work within the living animal, and which has the
form or soul of the animal as its constitutive form-analogue, there
exists also the material basis for the flesh or bone, which comprises all
the correct material parts, but which can exist independently of the
animal's form or soul.
If this is right, then there is room for a notion of the concurrent
matter for the animal below the animal's spatially determined parts.
One familiar objection to this suggestion will be reserved for the
beginning of section VIII below. An objection of a rather different
sort involves Aristotle's technical notion of µi¢ff; ('mixture') at work at
the lower orders of matter. Aristotle holds that an animal's uniform
parts, flesh and bone and the rest, are produced as the result of µi¢u:;,
which he sees as an alternative to both the Empedoclean and the
atomist view of the construction of matter. Aristotle emphasizes that
11il;1; is not the division of things-to-be-mixed into parts, which are
then placed side by side to make the finished product ('mixture' in our
or Democritus' sense, perhaps, or what in his own theory of matter
Aristotle labels (Juveun; ('composition')), but the 'union' or 'uniting'
(iivwm;, GC I. IO, 328b22) of two or more stuffs to become some
new stuff, which has its own new properties, and which is wholly
uniform (I. IO, 328a4,a8-12). Each of the original 'contributing' stuffs
exists altered in the product of the µi?;z;, and exists potentially there,
but each can in principle be recovered from the product, and returned
to its original state (I. IO, 327b24-31, cf. II. 7, 334a31-5).
Where a thing is the result of µi¢u;, if its initial ingredients are
present at all, they cannot be present as spatially determined parts of
the thing, since on Aristotle's account of µi(,z;, they exist only poten-
tially, but not actually, in the mixture. Three competing conclusions
about the matter of the mixture present themselves. First, if the initial
ingredients of flesh (say) are not present as spatially determined parts
of the flesh, and instead the mixture is flesh 'all the way through', then
possibly flesh and the result of mixture in general can have no concur-
rent matter at all. I am inclined to distrust this conclusion. For one
thing, it imposes considerable restrictions on Aristotle's account of the
274 Frank A. Lewis
sequence of matters, from the animal's non-uniform parts to the
uniform parts to the simple bodies, earth, air, fire, and water, which
on this view can be intended only as an analysis of the matter from
which an animal is constructed, and not as a look at the different
layers of matter that exist concurrently within the animal. More im-
portant, the conclusion also sets too much store by the notion that in
every case, something is part of the (concurrent) matter of a thing,
only if it is a spatially determined part of the thing. Whatever its merits
elsewhere, this condition manifestly fails where the thing is the result
of µi<;zr;. A second and, I think, preferable view is that flesh in fact has
a concurrent matter, namely so much potential earth, air, fire, and
water, each of which is a part of the (concurrent) matter of the flesh
but (since each is only potential earth, air, and the rest) not one of its
spatially determined parts. The initial ingredients cannot be physically
isolated within the product in the way that a good Greek atomist
might want. But they contribute their distinctive properties and
powers to the mixture. 52 And they can be recovered from it, so that
they exist in actuality once more. This is not matter in the way that an
animal's non-uniform parts comprise its matter, or that bricks are the
matter of the wall (cf. GC II. 7, 334a27-b2). But why should we take
the bricks in the wall as the universal paradigm for matter? The
theory of mixture, rather, supplies an obvious set of cases in which
Aristotle turns his back on the connection between a thing's matter
and its spatially determined parts.
I have argued that the result of µi¢1c; has a concurrent matter that
must be distinguished from the spatially determined parts of the thing.
This proposal is essentially weakened by yet a third suggestion, that
the concurrent matter of the mixture exists only potentially but not
actually there. 53 It is worth taking precautions here against the fol-
52 <Td>(nw )'<lp ff ovvaµ1c; avrcvv (327b30-r ). Thus, the fact that the initial in-
gredients of a mixture-earth, air, fire, and water-retain their tendency to go to
their natural ,place, helps explain the corruptibility of objects in the sublunary
world, GC II. 10; for similar reasons, soul is needed to hold the body together,
Oe An. II. 4, 4163 6-9, cf. I. 5, 4rrb7-9. Aristotle also discusses the way in which
the elements are present in the result of pil;u; in GC II. 8. His attempt to reduce
ail the properties and powers of bodies to the properties, hot, cold, moist, and
dty, typifying the four simple bodies that are the basic initial ingredients of any
mixture (GC II. 2, cf. Meteor. IV. 4-11 and Mourelatos 1984: 8, 13), again
presupposes that the initial ingredients make some continuing contribution to the
c9mpound. This same assumption lies behind the view in Meteor. IV that when
t~e mixing is done, we can reason back from various properties of flesh, bone,
and the like, to the proportion among the initial ingredients from which they are
mixed.
53 Gill (1989) and Scaltsas.,(this volume), see nn. 54 and 55 below.
lowing argument. For Aristotle, as we have seen, the four elements,
earth, air, fire, and water, that are the initial ingredients of any
mixture exist only potentially in the mixture. If, then, earth, air, fire,
and water are the concurrent matter of the mixture, doesn't it follow
that its concurrent matter exists only potentially in the mixture, and is
not actually present there? 54 But I doubt that we should use the
definite description 'the concurrent matter of this mixture' rigidly, to
conclude that that item, the matter, exists only potentially in the
mixture; instead, the most we are entitled to conclude is (giving the
description wide scope) that there is something, namely so much
earth, air, fire, and water, that serves as the concurrent matter of the
mixture, and that it, the quantity of earth, air, and the rest, exists
potentially there. That is, the concurrent actual matter of the mixture
is potential earth, air, fire and water. 55
54 Gill ( 1989: i47), for example, moves without comment from the point that
the elements, earth, water, air, fire, are potentially present in the mixture to the
claim that the matter is potentially present there.
55 Textual support for the view that a thing has no actual concurrent matter has
been found at Phys. I. 9, 192a25-9: Aristotle says, speaking of the matter, 'As
that in which, it does in itself pass away (for what passes away, the lack, is in it).'
By this, he may mean to say of the matter, here picked out by its being what
(once) possessed the lack, that it is destroyed when the lack is gone and the
product comes to be, so that there is no actual concurrent matter; this is how he is
understood by Gill (1989), who writes that in generation, 'the preexisting matter
is actually destroyed but preserved in potentiality' (Gill 1989: 148, my italics; cf.
pp. 8, 241); in the strict sense, according to Gill, only the properties of the
pre-existing matter survive (p. 241 ). But 'actually destroyed' is not something
Aristotle expressly concedes in his discussion of µit,1:;, where in the argument
against the sceptic at GC I. 10, 327a34- 06, he works hard to distinguish µif,i:; from
cases where one or more of the elements is destroyed. In the Phys. passage,
Aristotle may equally well be talking of the compound of the matter with the
privation, which ceases to exist when the thing comes into existence; in the formal
mode, he may be challenging the propriety of the description, 'that in which the
lack resides', to pick out the matter, once the lack ceases to exist (cf. Ross 1936:
498). It does not follow, of course, that there is not also a single continuing matter
that once was the pre-existing matter of the thing and is now its concurrent
(actual) matter.
Finally, according to Scaltsas (this volume), Aristotle's solution to the problem
of the Unity of Substance requires that again an individual substance have no
actual concurrent matter; he suggests that the material components of a thing
'lose their distinctness' and become 'identity-dependent' on the substantial form
when they are incorporated into the whole, and that the distinctness of matter
and form from the substance emerges only when we 'divide the substance by
abstraction'. I find this view uncomfortably close to the 'constructivist' or 'pro-
jectivist' views of Sellars and Kosman cited and rejected in sect. III and also
repudiated by Scaltsas, n. 33, the more so because Scaltsas does not clearly
separate his position from the view that in the absence of-or even despite?--
abstraction the matter and the form are identical with the substance.
Frank A. Lewis
VIII. POSTSCRIPT: THE DIFFERENT CONCEPTIONS

OF MATTER
I have been defending the idea that there can be a concurrent actual
matter of an animal that exists below the level of the animal's spatially
determined parts. A second line of objection to this view, predicted
towards the beginning of section VII above, is independent of the
theory of µi~l!;; it involves instead doubts about the utility of the
elements as the animal's matter. Ackrill argues that while the elements
can exist before and after the animal does, so that the form of the
animal is evidently not essential to them, they constitute too low a
level of matter to do any good. They are not potentially the thing in
the immediate sense Aristotle apparently requires in Met. e. 7. 56
This complaint has to do with the connection between the concept
of matter and notions of change or process which I argued earlier are
a prime component in Aristotle's theory of matter. But it is not
usually noticed that in identifying the matter of the animal as its non-
uniform parts, and the matter of these in turn as the uniform parts,
Aristotle has already left notions of change or process far behind. But
he has clearly done so. No one, I think, would seriously imagine that
when an animal comes to be, first to be formed are the uniform
parts-flesh, bone, and the like~and that out of these, next, the non-
uniform parts-hands, eyes, and so forth-get put together, the
animal itself finally emerging as a triumphant compilation of these
last. Aristotle's picture is rather that of the artist who sketches in the
whole drawing in outline first, and then goes back to fill out the details
later (GA II. 6, 743 11 20-5). Ackrill is right, then, that the elements
are not potentially the animal in any immediate sense, but a similar
point goes for the face, hands, and the rest; these too are not poten-
tially the animal either, at least not in the sense considered so far, that
they are capable of being made to constitute (if they do not already
constitute) the living animal. For a notion of matter that does involve
change and process, we must look elsewhere, specifically, to the
embryology, where (I would argue) the successive varieties of soul-
nutritive, locomotive, sensitive-that are introduced at different stages
in the development of the foetation create different levels of matter,
ready to be transformed by the arrival of the next level of form into
an item of a yet higher level, until finally the finished animal is
produced. These different embryological stages present an alternative
notion of matter that restores the connection with change or process-
and one that is free of the troubles we have been considering. Matter
at each stage can be essentially what it is, independently of the form or
soul that supervenes on it. On this conception of an animal's matter,
then, there is no redundancy of supervenient form, and no trace of
the troubling view that the proximate matter and the animal itself arc
identical. 57
Perhaps, then, the conclusion we should settle for is this. Variant
conceptions of matter undoubtedly exist, as divorced from consider-
ations of change as the metaphysician's conception in Met. z. 3,
already mentioned in section II above: the sequence, non-uniform-
uniform-simple, in the physiology, or the idea in the psychology that
the body with sense-organs, essentially capable of sensation, is as
matter to sensitive soul as form, or that the body with a heart and
blood-vessels, essentially capable of nourishment and growth, is as
matter to nutritive soul. It is a topic for another day to ask what
purpose these variant conceptions serve, or how they are related to
what I take to be the standard conception, defined around the notion
of change. The variant conceptions of matter involve their own dis-
tinctive views about how matter is related to the form that supervenes
on it or to the animal it constitutes. But these are very different
notions of matter, and whatever difficulties they may encounter leave
the standard conception of matter untouched. 58
57 Ackrill (1972-3: 131-2) draws attention to the different kinds of soul

involved at the different stages of the embryological development of a human
being (say), and notes that (had the development of the creature been suitably
frustrated) 'the human body might have remained at the merely vegetable or
merely animal stage', or 'might have failed to develop beyond the animal stage'.
But, he argues, there is no real comfort here: the relation of sensitive (animal-)
soul to the body of the human foetation at its animal-stage, for example, is open
to the old problem, that the body at this stage is necessarily living in precisely the
way conferred by sensitive soul. My suggestion is that we should look instead at
the relation between sensitive soul and what developed into the animal-stage of the
foetation, namely, the foetation in its still earlier, plant-like stage. This is the
notion of matter associated with change, and as Ackrill himself notes, had things
gone differently, it might have stayed merely vegetative: that is, its relation to the
form or soul that supervenes on it is accidental. For more discussion, especially of
the interesting BODY-body distinction employed in Cohen (1987), see Lewis
(1991): 257 n. 23.
58 Sect. II of this paper is largely an abbreviation of some pages from Lewis
(1991); I am grateful to Cambridge University Press for permission to reuse this
material here. I owe thanks also to audiences at the University of California,
Irvine, the Aristotle Metaphysics Conference at Oriel College, Oxford, the Pre-
dication Conference at the University of California Humanities Research Institute,
and the Conference on Aristotle's Thought at the University of California, San
Diego, for their tolerant and helpful discussion of earlier drafts of this paper. An
earlier versioff of the paper benefited from comments by James Bogen, Kit Fine,
and Mary Louise Gill; I am grateful also to David Charles for detailed comments
on the penultimate version. Despite all the help I have received, I shall be
<nrnri<Prl if <>nvonP <>orPP~ with :tll T h::ive said here.
12
The Essence of a Human Being and

the Individual Soul in Metaphysics Z and H
MICHAEL WOODS
In this paper, I raise, at a fairly general level, some questions about

the hylomorphic conception of the relation of the soul to the body; in
particular, as it appears in Metaphysics z and H. I was stimulated
to approach the issues in this way by Bernard Williams's article
'Hylomorphism' ( 1986). I do not attempt to respond to this article in
detail, and I shall in fact be occupied with certain aspects only of
Williams's arguments.
Williams argued that hylomorphism 'earns its reputation as every-
body's moderate metaphysics of mind' by wobbling between two
options: on the first of these, it is in fact a version of materialism-a
version that, in essentials, treats human beings as physical bodies in a
certain state, though it preserves certain features of the ordinary
conceptual scheme by allowing that, when Callias dies, something has
gone out of existence. 1 An uncompromising version of materialism
would simply treat Callias' life as a phase in the life-history of the
continuing material substance that is Callias' body. On this alternative,
there is no room for using the term 'soul' as a count noun, applicable
to individuals; nor could Callias' soul be reasonably thought of as an
entity enjoying the status of a continuing substance-even as one
dependent on a physical body of a certain sort.
The alternative view, Williams claims, manages to secure for an
individual soul the status of a genuine individual, but only at the cost
either of lapsing into incoherence, or adopting a notion of soul that
treats it as a type, which could, in principle, be multiply embodied.
This is something that hylomorphism does not want, and indeed is
©Michael Woods 1994

1 This needs '! qualification, or rather a caveat, since Aristotle wishes to leave
open the possibility that some parts of the soul, at least, might survive death in
the ordinary sense. In fact, it is, I think, clear that this is not thought by Aristotle
to require any modification of the claim that, when Callias dies, Callias the human
being ceases to exist.
280 Michael Woods
precisely what hylomorphism ought to provide us with a way of

avoiding.
It is certainly true that many recent writers have recommended
Aristotle's as a moderate view, and a serious competitor to currently
held views of the relation between body and mind. For example
Jonathan Barnes (1971-2: 114), arguing that Aristotle held an attri-
bute theory of mind, regards it as 'at least as good a buy as anything
else currently on the philosophical market'. Richard Sorabji ( i974:
64) claims that Aristotle's view resists classification according to
current pigeon-holes. Howard Robinson ( 1983) and Christopher
Shields ( 1988), while arguing that Aristotle was a dualist, are careful
to distinguish his dualism from Descartes's. Most recently, T. H.
Irwin (1988: sections 155, 156) has argued that Aristotle was neither a
dualist nor a reductive materialist, and adopted a position that was
neutral between what Irwin calls quasi-dualism and compositional
materialism.
My concern in this paper is with the more general question whether
the hylomorphic model provides room for a position on the relation
between mind and body (or soul and body) that is significantly dif-
ferent from materialism, as it is now understood.
I think that it is possible to see, in outline, how one might respond
to Williams; but what I am particularly concerned with is a problem
posed for Aristotle's view by the position taken by one who accepted
the first horn of the dilemma: the problem that the form-matter
model does not have the resources to yield distinct particular sub-
stances of the various sorts that one ought to be able to distinguish. 2
Thus, given that Socrates' body is a human body only so long as
Socrates is alive (that is, so long as Socrates exists), it becomes hard
to distinguish Socrates from his body; and it becomes hard, too, to
find a place for Socrates' soul, if that also is to be distinguished from
Socrates, and to have the status of a substance and a particular,
and not merely that of a property-the property of being alive, for
example.
What I want to do is to sketch an interpretation of the application
of the matter-form model which is not, I hope, actually incoherent,
and argue that it can be supported from a reading of certain passages
in Metaphysics Z and H and the De Anima. I shall not inquire whether
Aristotle is appropriately classified as a Dualist or a Materialist. I am
not primarily concerned to show that this conception of the soul is one
that we could believe in, so I am not offering a reply to Myles
2 This apparent problem with the hylomorphic model was noted by Edwin
Hartman (1977: 98): 'Aristotle does not need both the concept of person and that
,...,J:,\.. .... - ......- 1.. ....... .-l •• '
Human Being and Individual Soul 281
Burnyeat's arguments ( 1992) that the Aristotelian conception of the

mind is now, in a fundamental way, one that we cannot make sense
of; my purpose is to find an intelligible and coherent application of the
hylomorphic model that will provide for the phenomena that the
model was intended to explain in the first place. But I also want to
examine whether we can preserve the various things that Aristotle
himself was committed to. So I will begin by saying what I take those
to be. I shall be assuming that a satisfying interpretation ought to
make room for the following elements of Aristotle's position.
First, it must be such that there is some point left in examples like
that of the bronze sphere. I take it that part of the reason that this
seemed to Aristotle to be a helpful model for understanding the
relation of body to soul is that we do not find it puzzling that a piece
of bronze and its shape should contribute to forming an individual
concrete thing. 3 Whether one regards that as an untypically simple
instance of an object constituted by matter into a certain form, or
rather a case that is meant to help us to grasp the application of the
model to human beings by analogy, the application to the case that
concerns us ought at least to preserve the features of the original
case that make the union of matter and form in the composite
unproblematic.
Moreover, the conceptual resources provided by the form-matter
model must be deployed in such a way that they can be applied
outside the case of living animals, yet at the same time room is
made within that model for what is distinctive of an individual living
organism like a human being. Aristotle applies the hylomorphic
model to kinds of stuff designated by mass terms, regarding a type of
matter of a more developed, more structured, kind as resulting from
the informing of matter of a more basic kind. Thus, he is ready
to think of blood, flesh, iron, and lead, and in general, what is
commonly described as proximate matter as themselves decompos-able
into matter and form. He also applies it to artefacts, and, more
generally, kinds of thing that do not count as substances by the most
exigent standards, such as are invoked at Met. Z, ro4ob5-8. Likewise,
he applies it to the parts of living animals, though these do not count,
strictly ,4 as substances. When I say that the distinction between
3 Cf. De An. 412"6-8: 'Hence we should not ask whether the soul and body
are one, any more than whether the wax and the impression are one, or in general
whether the matter of each thing and that of which it is the matter are one' (tr.
Hamlyn 1968).
4 Cf. z ro28b8 f., where the parts of animals are said to be generally agreed to
be substances. It may be thought that this was no more than an l:w5o¢ov that
Aristotle set out at the beginping of the discussion without necessarily wishing to
Michael Woods
animals and other things that he applies the model to must be capable
of being made using the resources of the model itself, I mean that it
must be possible to characterize in terms of the model itself what
marks off substances in the fullest sense-living things and, in parti-
cular, human beings-from those things that are substances only in an
extended sense.
It may be thought that this requirement is too strong, and it has
become a commonplace to say that form, in the sense in which it is
the correlative to matter, must not be thought of narrowly as an
object's physical shape or structure. All the same, at De An. II,
4r2b6-7, in a passage quoted in n. 3, Aristotle appeals, as we have
seen, to the transparent unity of a piece of wax and its shape to throw
light on the unity of body and soul. The exact interpretation of that
passage is highly controversial; 5 but it serves to illustrate the way in
which Aristotle is content to refer to the familiar example of stuff
.wrought into a certain shape when arguing for the hylomorphic con-
ception of the soul. We may also compare passages in which the
bronze statue is invoked.
As I mentioned at the beginning, when outlining Williams's attack
on the hylomorphic conception, it would seem to be a desideratum
that room should be made for a distinction between the individual
soul and the individual human being; that, at any rate, seems to be
among the common beliefs that Aristotle starts from and is required
by what Aristotle actually says. 6 What this means is that we must be
able to distinguish two non-synonymous sortal terms, 'human being'
and 'soul' which are true of different things.
Further, it is part of Aristotle's position that there is a relation of
priority between the individual human being and the individual soul.
The soul is more basic than the individual human being, which in fact
presupposes it; likewise, the notion of a human being has to be
explained in terms of that of a human soul, and not vice versa. This is
surely an essential part of any position that conceives of the individual
human body as a composite of form and matter, with the soul as the
form, and this is made very clear at H. 3, 1043a36f., where Aristotle
says that the predicate 'animal' is ambiguous, applying both to the
soul and to a soul-in-a-body (which I take it means, to a human being,
as opposed to a human soul). Likewise, at 1043b3, Aristotle seems to
allow as a possibility that 'human being' should apply, in different
senses, both to the soul alone and to the compound. 7 (This is taken to
5 See Irwin (1988: 285).

6 See e.g. Met. Z, 1037"9-ro.
7 This passagce is discussed by Irwin (1992: sect. 132).
Human Being and Individual Soul
be parallel to the ambiguity that also occurs with 'house'.) The con-
nection of these two applications of 'animal' is said to be an example
of 'focal meaning':
For soul is the substance and actuality of a certain body: and 'animal' might
be applied to both (form and compound) ; it will be spoken of with reference
to one thing, not in one account. (trans. Irwin i992)
A satisfying interpretation of Aristotle's use of the hylomorphic
model must make the individual soul enough of an independent in-
dividual for the question seriously to arise whether it is appropriate to
treat the individual soul as the subject of thoughts and perceptions.
On the whole, Aristotle holds it is better not to say that the soul
thinks, but rather that the human being does; but in the Metaphysics,
Aristotle's view is less consistent, and in Z. I I, rn37a7 f., he toys with
the suggestion that the name 'Socrates' has a double reference (omov
[u17µaivs1]), and that Socrates is also his soul (in addition to being the
individual human being). This suggestion, presumably, is along the
same lines as the suggestion, already mentioned, that 'animal' and
'human being' apply in distinct senses to the concrete human being
and to the soul. Now if the individual soul is taken to be an attribute
or a property, it will hardly be intelligible that Aristotle should have
so much as considered this possibility. There are, of course, good
reasons for insisting that the soul of an animal is not any sort of
property-not, for example, the property of being alive, or the
instance of the property in the individual animal. Moreover, Aristotle
is quite explicit in his rejection of the view that the soul is some sort of
attribute in the De Anima and in his rejection of the conception of the
soul as a harmonia. I shall return to this point later.
Finally, there are two very important doctrines, one the converse of
the other, that Aristotle accepted about human beings, and indeed
any other kind of living organism.
There is first the fact that there is very little latitude over the kind
of material which a .form can be embedded in. Aristotle holds that the
human soul must occur in a human body with human flesh, bones, etc.
In short, the form must go to constitute a recognizable human being,
with a human body and the sort of matter that is required for that.
The next point is connected with that, but also needs to be dis-
tinguished from it. It is also true that a human body only counts as
such when it has the form: what the form structures is not available to
be structured antecedently to the structuring. Likewise, a dead finger
is only homonymously a finger, 8 and spilled blood is only homony-
Michael Woods
mously blood. 9 We must not think of the human body as something

that can be now animated and now not.
Although these two last points are connected, and they both re-
present ways in which the application of hylomorphism to living things
has features that distinguish it sharply from its application to cases like
that of the bronze sphere, they need to be separated. Even if the
human form could be realized in different kinds of materials, that
would not affect the point that a human body counts as such only
when living, and when capable of the various human activities. Simi-
larly, the doctrine that a dead hand is only homonymously a hand
would still stand even if living hands could vary much more in respect
of their material constitution than they can in fact.
If we now consider the case of artefacts in general, and not simply
untypically simple cases like that of the bronze sphere, it seems that
these last two of Aristotle's doctrines about human beings are distinc-
tive of the application of the hylomorphic model to living creatures.
For, in the first place, in the case of many artefacts, there is consider-
able latitude in the sort of material constituents that may have the
form. Aristotle himself mentions the constraints on what would count
as a saw, 10 but it is clear that, with another of his examples, the
example of the house, there may be a multiplicity of material com-
ponents that may go to make up a structure serving the functions of a
house as these are defined in Metaphysics H. 11 Further, it is hard to
think that Aristotle would have felt much temptation to deny that,
for example, a brick not integrated into a larger structure was only
homonymously a brick. 12
Ackrill (1972) assumes that the same view should be taken about an
instrument or tool as about a bodily organ. But, so far as I know,
Aristotle nowhere commits himself to the homonymy thesis about an
item such as a brick. Nor does such a view follow from the view,
which he certainly would have accepted, that only a brick (or a saw or
a door-lintel) that is capable of discharging its function can be non-
homonymously so-called. 13
If we think of the claim that a dead human body is only homony-
mously so-called as both a generalization and consequence of the
claim about a severed hand, it is clear that there is no room for
9 See Haslanger (this volume: 164 f.) for further discussion of such examples.
10 See Phys. II, 200"11-15.
11 Met. 1043a31-3.
12 At Cat. 7''5-11 he says that in the case of a rudder, there must always be
something of which it is the rudder, but that is hardly the same thing.
13 As Ackrill notes, it is not in general easy to decide when such an object does
satisfy that condition.
anything parallel to the human body (or indeed the 'body' of any
other living organism), as conceived by Aristotle in the case of a
house. A randomly arranged, and possibly scattered, set of bricks,
etc., is not even homonymously a house; and when they are appro-
priately arranged, we have the compound of form and matter and not
something on which the form supervenes.
These two features of the human case just mentioned: that the form
identified with a human soul should not occur except in a material
body of a human kind, and that the human body should not exist
independently of the form, are what led Williams to question whether
there is room, in Aristotle's application of the hylomorphic model for
a distinction between two, or indeed three, individuals: the human
body, the individual human soul, and the composite of the two.
Finally, we need an interpretation that caters for the identity of the
form with the essence of a human being, while preserving the dif-
ferences between human being and human soul already alluded to.
That these are Aristotelian doctrines is not entirely uncontroversial.
I think that we can find a coherent view fairly readily if we are
prepared no longer to insist on some of these desiderata. Alternative
views that have been considered can be distinguished according to
which of these requirements they are prepared, sometimes under
pressure, not to ascribe to Aristotle. Indeed, it is barely an exag-
geration to say that most of the recent interpreters of Aristotle's view
of the relation between body and soul have, in effect, sought to
show that some of the requirements were not after all insisted on by
Aristotle.
Thus Barnes (1971-2: 106-7) holds that the hylomorphic model is
a hindrance rather than a help to understanding Aristotle's theory; he
also holds that Aristotle espoused an attribute theory of mind. 14 But,
in doing so, he effectively discounts Aristotle's statements about a
soul being a substance, and the priority of the soul to its components,
along with the distinctively hylomorphic character of the theory. As
we have seen, the possibility of making room within the hylomorphic
picture for what is distinctive about living organisms seems to depend
upon the fact that in their case the form supervenes upon matter
that is apt for receiving the form only when it does receive it, and,
in addition, allows little scope for variation; but this very feature
threatens to make the bronze statue example unilluminating.
Alternatively, if we are prepared not to insist on the priority of a
soul to a human being, one might be tempted to see the relation
14 ( 1971 -2: I 13): 'Aristotle thus emerges as the holder of an attribute theory of
mind; and that is, I suggest, his greatest contribution to mental philosophy.'
286 Michael Woods
between soul and human being as analogous to that between human

being and person: being a human being is not the same thing as being
a person, even if in fact the only persons there are are human beings.
But that, regarded as an interpretation of Aristotle, gets the priority
the wrong way round, since it is in virtue of being human beings with
the capacities that they do have that persons qualify as such; whereas,
as we saw earlier, a full explanation of what it is to be a human being
must bring in the human soul, since the explanation that Aristotle
favours of what it is to be a human being clearly incorporates the
notion of the relevant kind of form.
In fact, the proposed comparison between Aristotle's contrast
between human being and human soul and the contrast made by John
Locke and other philosophers subsequently between human being and
person seems to have difficulty in accommodating the denial that the
human form could be realized in matter in a number of alternative
ways; since, to allow that being a human being is not the same thing
as being a person, and therewith the view that some modern writers
have adopted, that Socrates the human being is not to be identified
(strictly) with Socrates the person does seem to depend upon at least
making sense of the possibility that there should be persons that are
not human beings; but this conflicts with the Aristotelian doctrine just
mentioned. A number of modern writers 15 have argued that there are
good reasons for regarding human being but not person as a natural
kind term; but it would hardly be Aristotle's view that soul is not a
natural kind term. In other respects, the human being/person analogy
seems promising, since having distinguished what are arguably two
distinct things, Socrates the human being and Socrates the person,
someone might hold that 'Socrates' has a dual reference, as suggested
in Met. Z.
Whatever may be thought of the bronze statue, one may reasonably
think that any application of the hylomorphic model would require us
to see at least some analogy between an individual human being and a
house, with the soul corresponding to the internal organization of
bricks and mortar that enables it to serve its function; but, as we have
seen, this would seem to involve the possibility of the very same
material parts occurring both in a structured and an unstructured
state; but although, at a certain level, Aristotle can identify the same
matter in both the living and the dead human being, this is not true of
the body, upon which the form directly supervenes.
What I want to do now is to sketch a position that meets these
desiderata, and can be supported by certain passages in Z and H. I
15 e;g. David Wiggins (1980: 170-3).

shall have time to examine only a few passages, and indeed the main
reason for thinking that this position is worth further development is
that it meets these desiderata.
I begin with the problem that, if we are to make anything at all of
the bronze statue analogy, it would seem that we must be able to
make sense of the same form in different sorts of matter; it would not
be enough that the form could be found in different parcels of matter,
if those could not be distinguished qualitatively. From that, all that
one could derive would be the notion of an undifferentiated property
which they all share-the property of being a bronze statue, for
example. In order for the form/matter contrast to have any purchase
at all, we need to find a duality between structure and what is struc-
tured. In order to disentangle the structure from the matter, I need to
be ready to envisage the possibility of different things sharing a struc-
ture but differing in the kind of matter involved.
Now there is, indeed, no difficulty in disentangling the common
human form from the qualitatively different human bodies that they
structure. These qualitative differences between human beings do not
conflict with Aristotle's denial of variable realization, since the dif-
ferences in characteristics are compatible with the possession of the
standard elements of a human body.
But it is notorious that, if all we have in the way of a form is the
structure common to human beings in general, we do not seem to
have the resources to distinguish Socrates' soul from Callias'. It is at
this point, of course, that particular or individual forms are invoked.
But I think the question needs to be asked what sort of particular
form the hylomorphic model provides us with the means of introducing.
We begin by noting that the form of a human being has commonly
been taken to be a universal or property that has instances, because,
in the sphere case, we have no difficulty in seeing the same shape
occurring in different parcels of matter. However, it seems to me that
it is misguided to think of this shape or structure as something of
which the particular material thing of that shape is an instance, or
again as a property of that thing. The spherical piece of bronze is not
an instance of the spherical shape, nor is that spherical shape one of
its properties. Of course, in virtue of the fact that it has that shape
there is a property that it has-the property, in fact, of having that
shape. If we believe in particular properties, we can then say that
there is a particular property that it has-but that will be a particular
instance of the property of being rectangular, not an instance of the
shape of a rectangle. A shape is simply not the sort of thing that can
be said to have instances.
So I want to say, quite generally, that the structure of something is
288 Michael Woods
not one of its properties, though having a certain structure is. So this
point is not something that depends on the special features of living or
organic substances. These observations apply equally to the case of a
house. If the question is asked whether the structure of something is a
universal or a particular, the answer must, I think, wait until we have
refined further our notion of what it is to be a universal. We should
encounter a similar uncertainty if the question is asked whether
numbers are universals or particulars. If we ask what should count as
an instance of a certain structure, the best we can do is to cite specific,
more determinate, ways in which that structure might be realized, and
it seems that those, like the structure we began with, would be
multiply realizable.
So much, I would claim, can be defended in advance of the attempt
to understand Aristotle. But Aristotle does, of course, address the
question how we can make room for a difference between matter and
form, given that we do not find the human form occurring in different
kinds of matter, in Z. 11. He begins by saying that it is hard to
disentangle what belongs to the form, in view of the fact that we do
not find it separated from a particular sort of body in the case of
human beings ( 1036a3 I - b7). None the less, he says, even if it happened
that all circles were bronze, that would not mean that a distinction
would not need to be drawn. (The point seems to be that we might
easily overlook a possibility; all the same, that possibility is there, and
that is enough to show the need to distinguish the circle as a form.)
I think it is reasonably clear, from the passage immediately fol-
lowing, that Aristotle thinks that this is the wrong way of trying to
grasp the human form, because, as he says at 1036b23, 'some things
are perhaps "this in this", or these things in such a state'. Since it is
intended to show something about the form of a human being, it
must be making the suggestion that, although only form is properly
mentioned in a definition, a proper understanding of what the human
form is will involve the recognition that we have the specific structure
that counts as the human one only when we have a structure that is
dependent upon a particular body, developing in the way appropriate
to the kind.
A human form is essentially a form that animates a continuing and
developing human body. Part of the reason for this is that the form is
the structure that is the source of the various distinctively human
capacities. Among those we should want to include the forms of
perception that are possible to a creature with human sense-organs,
human desires, beliefs, and intentions, all of which presuppose the
possession of a human body. Moreover, the type of human body is
one that is capable of persisting and developing in the characteristic
human way; but that, in turn, presupposes the human form, to provide
a basis for reidentifying the human body.
That the form only should be mentioned in a definition is true, but
it is still misguided to attempt to 'remove' the matter altogether,
because the presence of a human body is partly constitutive of that of
the form, even though we need to appeal to the notion of the human
life in order to reidentify the same body through replacement of
matter.
This may be what Aristotle had in mind when he said that certain
things-and here he clearly has in mind certain forms, notably the
human form-are 'this in this'. If the position I have sketched is along
the right lines, a relation of reciprocal dependence holds between the
human body and the human form. The human body depends on the
human form for its status as a human body, and for its reidentifiability;
but the human form, in another way, depends upon the human body
for its discernibility in that human body.
I think that this symmetrical dependence is perfectly coherent, and
serves to provide a rationale for two of the characteristic features of
Aristotle's hylomorphism mentioned earlier: the absence of latitude in
the type of matter in which the form is embedded, and the fact that
the soul is the form of a body that does not exist antecedently to its
possession of the form. And the symmetrical dependence can be
grounded in some recognizable elements of Aristotle's philosophy of
mind.
It should be clear that, if such a view is anything like Aristotle's,
there still remains the possibility of distinguishing a human body from
the continuing structure that constitutes it as such. The explanation
precisely invoked the idea of the human form as something that is
partly constitutive of the fact that we have a continuing human body;
and the very same form persists through the various vicissitudes in the
life of the continuing human body.
But will this version of hylomorphism make room for the human
soul as a substance; as something that is a genuine individual and not
merely an attribute, and something that is not a type of which there
may be many instances? In my sketch of this position I have spoken of
the human form as that which is recognized throughout the life of the
human body.
I have already argued that it is not correct to regard the human
form as something that has instances, and I have also argued that the
human body, with its distinctive characteristics, plays an essential part
in the identification of the form as a human form. The form is
essentially the form discernible in a human body that is a persisting
thing, developing in a way characteristic of human beings, and surviv-
Michael Woods
ing through replacement of matter. That being so, it makes good

sense to think of the form-the human form-as having a certain
career in a particular human body. The human form is something that
persists through time in its occurrence in a particular body. Thus, the
human form, which is the same in the case of Socrates as it is in the
case of Callias, is something that has a career and a life-history in its
occurrence in a particular body. When we speak of Socrates' soul, as
distinct from Callias', we are thinking of the human form in its
occurrence in a particular body. Socrates' soul is something that can
be recognized through the various changes that befall Socrates in his
life; but what persists throughout is the human form, the very same
item as persists through Callias' life. But there is a sense in which
Socrates' soul is unique to him, because in speaking of Socrates' soul,
we are thinking of the human form as an object persisting through
time animating Socrates' body.
13
The Definition of Generated Composites

in Aristotle's Metaphysics
MICHAEL FEREJOHN
Consider where matters stand at the beginning of Metaphysics Z. 4 in

the search for the identity of primary substance which occupies most
of Z. 1 In the previous chapter Aristotle had narrowed the field
of leading candidates for primary substance to the familiar triad of
matter, form, and the matter/form composite (i.e. the concrete par-
ticular), and had then gone on to eliminate matter from further
serious consideration on the ground that it cannot be what primary
substance must be, namely something which is 'separate' (xwpunov)
and a 'certain this' (r6& u) 2 (1029a27-8). Now since Z. I and 3 both
indicate Aristotle's belief that an important clue to the nature of
primary substance is contained in the observation that there is a close
conceptual connection between substance (or more accurately, the
substance of a thing) and the 'what is it' ( ri i<rn) or 'essence' ( ro ri 1}v
elvw) of that thing, a question that naturally comes to the fore at the
beginning of Z. 4 is whether, and in precisely what sense, each of the
remaining two candidates (form and composite) 3 can be said to have
©Michael Ferejohn 1994

1 Virtually every recent commentator on Met. z agrees that the bulk of the
material in this difficult book is organized in some way or other around a putative
search for the identity of substance (ovaia) in the primary, or unqualified,
Aristotelian sense of that term. Unfortunately, there is very little beyond this on
which one encounters general assent. In particular, there seems to be no consensus
about the nature or the location of the eventual outcome of this search. Indeed, it
is not even generally agreed that the search is intended to have a definite outcome
and is not just an 'aporematic' exercise, as is suggested in Owen (1978) and Code
( 1985b). Worse still, there seems to be very little agreement about the basic
ground rules and substantive assumptions under which the search is conducted.
2 This verdict is undermined by the fact that Aristotle seems to say at z. 8.
rn33b2r-2 that form (his preferred candidate for primary substance) also is not a
rb& rz, leaving it unexplained why this violation is not vicious.
3 z. 3 also contains what could be taken as an argument to rule out the
matter/form composite as primary substance (1029•30-4 together with 1029"5-7),
which would seem to make the rest of Z superfluous as a search for the identity of
primary substance since it would leave but one candidate (form) in the running.
Michael Ferejohn
(or to be) 4 an essence. But since it is also Aristotle's settled view that
a definition (6pzaµ6r:;) is a formula (},6yor:;) which signifies a thing's
essence, this is tantamount to the question of whether (and in what
sense) each of the form and the composite satisfies what I will refer to
as the Definability of primary substance.
To address this question, Aristotle opens the main discussion of
Z. 4 at 1029b12- 14 with an announcement that he will begin his
investigation into the connections between essence (and definition)
and substance by treating the issue 'logically' (}.oyzKwr:;). When this is
set beside other methodological passages, most importantly De An.
I. I, 403a25, where Aristotle uses this adverb or its associated adjec-
tive, it can plausibly be taken to suggest that Aristotle thinks there is
an alternative mode of treatment available, one which he usually
describes as 'physical' (rpvmK6r:;). 5 Moreover, the De Anima passage
also suggests that the best sort of treatment will somehow incorporate
results from both of these kinds. In what follows I propose to pursue
these suggestions by arguing for the following three theses.
(TI) To begin with, the 'logical' treatment of the Definability of
substance announced at 1029b12-14, and which occupies most of Z. 4
and 5, amounts to a survey of certain peculiar 'formal' features of the
formulae of various sorts of things, and of certain 'logical' puzzles
arising out of those features which seem to preclude the Definability
of the sorts of things in question. Looking far ahead for a moment,
the idea is that since certain plausible candidates for substancehood
seem to share some of these features, Aristotle will be able to salvage
their candidacy only if he can show how these logical difficulties can
be dissolved or bypassed in their cases. On this interpretation, the
'logical' puzzles of Z. 4 and 5 form a sort of proving ground for the
theory of primary substance developed by Aristotle in later chapters.
(T2) One major reason for classifying the investigation in these
chapters as 'logical' (as opposed to 'physical') is that it treats various
For this reason, it is much better to understand d<pertov at a32 as conveying what is
essentially a conjecture that the matter/form composite 'should be set aside' (i.e.
not initially considered) as the projected target of the inquiry on the grounds that
it appears to be more 'intelligible to us' (i.e. more familiar) than the form,
whereas Aristotelian investigations typically make a 'passage to what is more
intelligible' ( ul µcraftaivw si~ ro yvwpzp<iJrcpov) from what is more familiar ( J029b3).
4 z. 6 is a study of what things can be understood as not just having essences,
but also as being (identical with) those essences. Excellent treatments of this topic
and its relation to the Third Man Argument are to be found in Owen (1965b),
Code (1985a), and F. Lewis (1991).
5 See also Phys. IL 2, 194"10-15, GCI. 2, 316"5-14, Met. E. l, 1025"19-10263 7,
Z. l l, 1037arn-20.
The Definition of Generated Composites 293
proposed objects of definition from the entirely 'static' perspective
typical of the Organon, exclusively as ovra-'things that are',
without taking into account that many of these bvra are also
y1yvopeva-'things that come to be'. 6 Thus, the 'logical' discussion in Z.
4 and 5 stands in marked contrast to 'physical' investigations found in
such works as the Physics and Generation and Corruption, where
many of the same entities are conceived of 'transtemporally' as the
subjects of change of various sorts, and therefore as amenable to a
'deeper' kind of metaphysical analysis involving such notions as per-
sistent substratum, replacement of mutually opposed qualifications,
actuality and potentiality, and most importantly, matter and form.
Now even though neither the word lfJV<JZKo;; nor its adverb is actually
used in the chapters following Z. 4 and 5 to signal a transition from a
'logical' to a 'physical' mode of treatment, I shall argue that it is none
the less possible to view Z. 7 as making just such a transition by intro-
ducing terms and concepts characteristic of 'physical' investigations.
(T3) Equipped with the results of this 'physical' investigation,
Aristotle is then prepared, beginning in z. IO- 12, to produce a
'combined' account which incorporates the results of both earlier
treatments.
Thus, to give a quick preliminary overview, my primary aim here is
to show how what is added by Z. 7 is useful (and even necessary)
to advance Aristotle's investigation of Definability in z. IO and I 1
beyond where matters are left at the close of Z. 5. More specifically, I
will argue that certain conclusions developed within the 'physical'
treatment of y1yvo11eva in Z. 7 are instrumental in Aristotle's sub-
sequent efforts in z. IO and II to show that the 'logical' puzzles of Z.
4 and 5 do not entirely vitiate the Definability of sensible matter/form
composites.
In making a case for (T3), I will thus be defending to some extent
the integrity of the structure of z as it has been transmitted. I should
not, however, be taken as assuming or arguing that all of its chapters
were written or delivered in one continuous narrative sequence; that
would do an injustice to Aristotle's obvious organizational and com-
municational abilities. 7 On the other hand, I do want to claim on the
6 This expression is used here in the widest possible sense to cover all types of
change, including cases where a substance 'comes-to-be-F' (for any non-substantial
substituend for 'F'), as well as the coming-to-be simpliciter of items in any
category.
7 Thus, I do not understand my position here to be at odds with the remark in
Furth (1988) that 'Z. 7-9 make up a separate unit not originally continuous with
Z. l-6' (114-15), nor with the observation in Burnyeat et al. (1979) (also known
as the 'London Group') that these chapters 'were written for another purpose'
294 Michael Ferejohn
basis of the interpretation given here that we need not think of this
part of Z at least as just a loose-leaf compilation of fragmentary
discussions unified by little more than the fact thay they all concern
the same general topic (primary substance). On the contrary, if my
general view here is correct, one can see in the present positioning of
these chapters (however it came about) an orderly progression of
argumentation directed at the eventual emergence and defence of
Aristotle's considered view about the nature and identity of primary
substance. 8 However, what I have to say below admittedly concerns
only a part of this progression, for I do not believe that the Meta-
physics' final position on the Definability question is yet in sight by the
endofz. 1r.
I
One possible approach to the topic of Definability would be to give a
full and precise formulation of this condition, and then to test various
candidates by seeing how well each of them fits that formulation. It
appears, however, that Aristotle works in the opposite direction in z.
4 and 5. His strategy there is instead to identify various kinds of items
which seem clearly to violate Definability, and to determine in each
case the exact nature of the violation, in this way developing a pro-
gressively richer understanding of what the condition entails. Fur-
thermore, it is assumed throughout that a genuine definition must
constitute a 'unity' in some exceptionally strong sense of that term. At
the upper end of this ranking is the limiting case of primary substance
itself, which is ex hypothesi perfectly Definable, and which is iden-
tified provisionally in Z. 49 as the 'species (pl.) of a genus' (103oa12).
(54). My main contention here is that whether chs. 7-9 were inserted in their
present position by Aristotle himself, as the London Group conjectures, or were
placed there by a later editor, the crucial fact is that the doctrines developed in
chs. IO- II depend essentially on insights developed in the preceding chapters. I
do therefore take my conclusions here to conflict with the assertion in' Frede and
Patzig (1988) that the arguments of z. 7-9 intrude upon the rest of Z (i. 24-5).
8 However, what I have to say below admittedly concerns only a part of this
progression, for I do not believe that the Metaphysics' final position on the
Definability question is yet in sight by the end of z. 11. See n. 36 below.
9 By 'provisionally' here I mean for the purposes of the discussion of Definability.
It does emerge in Z. 12 that the definition of a species by its genus and proper
differentiae does constitute a unity in the sense required. This, however, does not
guarantee that species will satisfy Aristotle's other requirements for primary
substance, e.g. the ro& rz requirement of z. 1 and 3. See Owen (1978) and Code
( 1985b) for a defence of the view that Aristotle thinks nothing can possibly satisfy
both conditions simultaneously.
And at the opposite end, again as a limiting case, Aristotle places his
stock example of the Iliad, rehearsing his standard view that even
though this case satisfies the very minimal requirement that a defini-
tion is a formula that 'means the same as the name' ( 103oa IO), it
nevertheless is clearly not an instance of Definability because the
formula in question can be counted a 'unity' only in a very marginal
sense due to the 'holding together' (rep avw:xsi) of its elements in a
continuous sequence. 10
Aristotle's project in z. 4 and 5 is to situate between these ex-
tremes two sorts of cases whose shortcomings he evidently thinks are
especially instructive. The first of these is what he calls 'compounds'
(a6v0sra) of substances and non-substances, such as the peculiar entity
signified by 'white man'. Briefly, Aristotle argues at 10303 2-7 in Z. 4
that a thing of this sort is not an 'essentially (single) particular thing'
(onsp r6& ri), and so cannot satisfy Definability, because it involves
'one thing being predicated of another' (aJ),o Kar' a}),ov liy1Faz).
However, immediately prior to giving this argument, Aristotle also
entertains (but doesn't adjudicate) another apparent objection to the
Definability of white man. At 1029b29- 103ou2 he considers the sug-
gestion that this compound might fail to have (or be) an essence
because it is not even something 'said per se' (Ka€J' w!u) },tyc:rw), and
that this failure is in turn due to what he terms the fallacy of '[defining]
from addition' (1!K npoaetawx;), which is described at 1029b32-3 as
that of attempting to define some simple expression 'A' and ending up
instead defining some larger expression 'AB' of which it is but a part.
Unfortunately, Aristotle never actually says how this fallacy is sup-
posed to be relevant to the case of the compound white man, and it is
not at all obvious how to manipulate that example so that it looks
even remotely as if it involves this sort of mistake. None the less, this
inconclusive discussion evidently provides a useful transition into Z. 5,
for that chapter opens with the question of whether another sort of
case, what are there called 'coupled things' (avv&61)(iaµc:va), might fail
to satisfy Definability because they too involve the 'from addition'
fallacy. The example Aristotle uses to illustrate this sort of case,
snubness (m116n7r;), is by far the most prominent in 'logical' discussions
of definition. The frequent recurrence of this example does not just
reflect some fetish on Aristotle's part. Its homeliness should not
obscure the fact that snubness is after all a very carefully chosen
stand-in for the natural biological substances which are Aristotle's real
°
1 Cf. An. Post. II. 7, 921'32 with II. 9, 931'36, and also Met. If. 6, 1045"14.
Oddly, the latter two passages replace avvi:xi:i with avl'iifopq> ('binding together"),
which is contrasted with avvexci in Z. 4 (at ro3obro).
Michael Ferejohn
objects of interest throughout his discussion of Definability in Z. This

is particularly clear from Met. E. 1, where he explicitly groups snub-
ness together with such 'physical things' (qmmKa) as 'nose, eye, face,
flesh, bone, and animal generally, and leaf, root, bark, and plant
generally' ( 1025b36- 1026a3), and then intimates strongly that because
of their affinity, a study of the former should prove useful in under-
standing the latter. 11 ·
What makes this even a prima-facie case of the 'from addition'
fallacy is the fact that snubness is what Aristotle calls a 'per se
affection': an attribute which cannot be defined without somehow
bringing in its proper subject (in this case, nose). 12 This might be
taken to mean that if one tries to define snubness by itself, one ends
up defining the compound snub nose instead, a predicament which
more or less fits the description of the 'from addition' fallacy given at
1029°31-3. To a modern reader, this might just be seen as a tip-off
that 'snub' can only be defined contextually, by giving definitional
11 One might well ask why Aristotle resorts to this proxy-example instead of
dealing directly with biological substances. The answer, I think, is that directly
discussing these paradigmatic substances would encumber him with a problem he
regards as tangential at this juncture: that of saying exactly what counts as the
form of a biological organism. The doctrine of De An. II. r, 4r2bro-4r3"ro (and
cf. Met. Z. II, ro37"26-9), that the form of a living thing is its 'soul' ('!fvxff),
simply moves the problem back one step to that of determining what constitutes
the soul of a biological organism of a given kind. Is it, for instance, (i) the
distinctive cluster of capacities for certain modes of activities which characterize
that kind of organism, or (ii) the complex structural organization of the material
parts of the organism which ground those capacities, or perhaps (iii) some
'hylomorphic' combination of both of these? As undeniably important as this
problem is to Aristotle's overall philosophical system, it is understandable that he
$hould be reluctant to broach it in the Metaphysics, where it is not crucial to his
immediate purposes. What makes snubness so convenient as a substitute for
biological organisms is that it is possible in its case simply to identify the form of
the concrete particular (the snub nose) with its shape (concavity). Of course, this
convenience is also shared by Aristotle's other pet examples of matter/form
composites, the bronze statue and the bronze sphere. But the difficulty with these
artefactual composites is that they lack the crucial inseparability of form from a
specific sort of matter so characteristic of biological organisms (Met. z. r r,
1036°3-4). Hence, snubness is a particularly good example because it allows
Aristotle to exploit the convenient (if not quite accurate) shape/form equation
without worrying about disanalogies between artefacts and natural things.
12 This doctrine of per se affections appears to arise out of a confluence of the
purely 'logical' doctrine in An. Post. I. 4 that there is a certain class of 'per se'
attributes whose defining formulae must mention their proper subjects (cf. An.
Post. I. 4, 73"38- 03 with Met. Z. 5, 1030°24), with the more metaphysical
doctrine of Cat. ro that certain types of accidents (e.g. blindness) by their very
nature must of necessity inhere in the proper sort of 'recipient' (c5cKrtKb~) (in the
case of blindness, animals that by nature are sighted).
equivalents of larger expressions containing them. As we shall see,
however, in Met. Z. IO and 11 Aristotle himself ends up drawing an
entirely different moral from this observation.
II
At 103ob18-28 Aristotle reasons that because 'coupled things' such as
snubness bear some sort of necessary connection to their proper
subjects (since snubness, unlike concavity, is necessarily an attribute
of noses) they at least possess more unity than do m1v0na like the
white man of Z. 4, and are therefore more likely to satisfy Definability.
But then, in the immediately following passage (103ob28-1031a1), he
sets out the most formidable 'logical puzzle' in Z. 4 and 5 in the form
of an argument which appears to show that snubness cannot after all
be defined. What follows is a literal translation of that argument in
which the statements it involves have merely been numbered and
separated from each other, and from inferential connectives.
But there is another puzzle (dnopia) about ['coupled things'].
If (I) snub nose and concave nose are the same,
then (2) snub and concave will be the same.
But if (3) this is not the case,
[then] because
(4) it is impossible to speak of snub without the thing of which it is a
per se affection [namely the nose]
(for snub is concavity in a nose),
either (5) to say 'snub nose' is impossible,
or (6) it will be saying the same thing twice: 'concave nose nose'
(for 'snub nose' will be 'concave nose nose').
Therefore
(7) it is absurd that essence should belong to such things,
But if (8) this is not so,
(9) there is an infinite regress
(for in 'snub nose nose' there will be yet another 'nose').
Although this argument has already been subjected to much scholary
attention, I shall not pause to evaluate the relative merits of the
various previous interpretations of it that have been put forward, nor
shall I try to say exactly how these earlier interpretations agree or
disagree with my own. Instead, I will simply sketch one way of
interpreting the argument for which I want to claim a significant
virtue: it not only makes sense of each of the statements and infer-
ences displayed above, but also makes clear (as most other versions
do not) how the line of reasoning in (7)-(9) can be understood as
building upon that developed in (1)-(6). In other words, this inter-
Michael Ferejohn
pretation makes clear how the infinite regress mentioned in (9) is

supposed to be generated by principles and assumptions introduced
earlier in the passage.
As I understand it, the ultimate premisses of the argument are
given by statements (r), (3), and (4). More specifically, it is assumed
from the beginning in ( 1) and throughout the argument that the
expressions 'snub nose' and 'concave nose' are strongly equivalent,
where this technical expression (abbreviated below as '=df') denotes
a connection between the two expressions strong enough to license
the same sorts of substitutions that would be allowable between a
term and its definition. Thus, (1) can be paraphrased as
(1') 'snub nose' = df 'concave nose'.
Then, (2 ) raises the possibility that someone might propose to
'cancel out', as it were, the common parts of these two phrases by
appeal to some such principle as the following,
(*) if AB = df AC, then B = df C,
and thus take the remainders, 'concave' and 'snub', also to be strongly
equivalent. (3) then summarily dismisses this proposal (for a reason
not given here, but which could be taken as obvious: the terms 'snub'
and 'concave' are not even co-extensional), 13 and (5) and (6) then
delineate what seem to be the remaining possibilities concerning the
Definability of 'snub'. This further delineation, however, is controlled
throughout by (4), which expresses the essentially negative point that
whether or not 'snub' or (snubness) turns out to be definable at all,
since it is a per se affection it cannot be defined unless its definition
makes mention of nose. In the light of this assumption, then, (5) and
(6) respectively set out two unattractive alternatives concerning the
seemingly innocent expression 'snub nose'. On the proposal considered
in (5), nose is not mentioned, and so, by (4), 'snub' is not definable,
and it turns out that 'snub' is not really a discrete expression with an
independent meaning. In this case it is concluded that one is not really
13 It might be objected that Aristotle already commits himself to (*) in z. 4 at
ro29h22. where he says, 'if being a white surface and being a smooth surface are
the same, then being white and being smooth are one and the same'. However,
even though this passage contains dative + infinitive forms which are usually used
to denote essential being, it is more charitable (given the obvious fact that not all
smooth things are white) to take this not as asserting definitional equivalence
between 'white' and 'smooth', but only to mean that the whiteness of white
surfaces is tantamount to their smoothness (on some unspecified physical theory).
Incidentally, understanding a parallel point in the case of snubness and concavity
would explain why Aristotle says at ro3ohr8- 19 that concavity, as well as snub-
ness, is a per se affection of nose. For on a perfectly natural understanding, the
snubness of a nose just is .its concavity, though this certainly would not support
lfhpr~l int~rc11h(.·t-it11tlu1t" hii.tnu:•a.n tho1,- .... o::ioc-""'"",..t;...,'3 ....,,....,..,,... .0.,.
saying anything at all about a nose in calling it 'snub', so that the
expression 'snub nose' is 'impossible'. This of course is the aporia
towards which the argument points. On the other hand, (6) is gen-
erated by the contrary supposition that 'snub' is defined, and further-
more (in keeping with (4)), that it is defined as 'concave nose'. That is
to say, a crucial but suppressed step in the argument is from (4) and
the definability of 'snub' to
(4 *) 'snub' = df 'concave nose'
whose right side does after all mention nose, as (4) requires. In this
case we can then immediately derive (6), since 'snub nose' becomes
'concave nose nose' by the replacement of 'snub' in it with its hypo-
thesized definiens, 'concave nose'.
As I said, the chief virtue of this interpretation is that it allows us to
understand the reasoning of (7)-(9) as a compressed summary and
elaboration of that given in ( r )-( 6). In particular, (7) recalls the
negative alternative considered in (5), that nose is not mentioned, and
its consequence (by (4)) that 'snub' cannot be defined at all. In the
words of (7), the consequence is that snub 'has no essence'. On the
other hand, (8) and (9), like (6), reduce to absurdity the positive al-
ternative on which 'snub' is defined, and moreover (again, in keeping
with (4)) is defined in a way which makes mention of nose in the way
sanctioned by (4 *) as 'concave nose'. From this point, the infinite
regress in (9) is generated straightforwardly by a sequence of alter-
nating substitutions on the expression 'snub nose' authorized respec-
tively by
(4 *) 'snub' = df 'concave nose',
on the one hand, and the standing assumption
( 1 1 ) 'snub nose' = df 'concave nose',
on the other. For in the first place, from 'snub nose' and (4 *) it is
possible to derive 'concave nose nose'. But ( 1 ') then licenses the
replacement within this of 'snub nose' for 'concave nose', which
produces 'snub nose nose'. At this point, however, it is possible to
reapply (4*) to 'snub nose nose', which then generates 'concave nose
nose nose'. And since a reapplication of (I') then allows us to replace
'concave nose' in this with 'snub nose', we now have 'snub nose nose
nose' (which captures what is meant is (9) by the explanatory remark
that there will be yet another 'nose' in 'snub nose nose'). But since
this pattern of alternating substitutions can be continued ad infinitum,
a consequence of the positive alternative considered in (6) and (8)-(9)
is that the seemingly innocent expression 'snub nose' actually turns
out to be equivalent to the expression formed by 'snub' followed by
• r- _• .. 1 ~.t..----.t..~---- _J: '·----' J"""I! ____ A-!-'--•1-'~ ----111-- ........... ,_
intolerance of infinite regresses, it is reasonable to interpret the argu-

ment as pointing aporetically to the conclusion that (5) and (7) are
true-that snubness is not definable at all.
Before going on, it is crucial to notice how pivotal (4 *) is in this
reasoning. In the most general terms, the argument is that because
(4) is true, we are forced to conclude that if 'snub' were definable at
all, it would have to be defined as in (4*), as 'concave nose', in which
case the bad consequences of (6) and (9) ensue. But nowhere in this
argument does Aristotle either defend or question the assumption that
(4 *) is in fact the only remaining possibility consistent with (4). That
is, Z. 5 doesn't even acknowledge the possibility that there may be
some other way of defining 'snub' which also makes mention of nose
(as (4) requires), but does not fall victim to the logical puzzle just
surveyed. Indeed, I shall argue in section IV that with the aid of
insights developed during his 'physical' investigation of generated
composites in Z. 7, Aristotle eventually uncovers just such a possibility
in Z. IO and 11.
III
In contrast to the puzzle-orientated methodology in Z. 4 and 5,
Aristotle sometimes employs a style of philosophical investigation
which calls to mind a remark of Ryle's in 'Systematically Misleading
Expressions'.
There are many expressions which occur in non-philosophical discourse
which, though they are perfectly clearly understood by those who use them
and those who hear or read them, are nevertheless couched in grammatical or
syntactical forms which are in a demonstrable way improper to the states of
affairs they record (or the alleged states of affairs which they profess to
record). Such expressions can be reformulated and for philosophy ... must
be reformulated into expressions of which the syntactical form is proper to
the facts recorded (or the alleged facts alleged to be recorded). (emphasis
added) 14
To employ a fashionable term of art, we might think of the method
recommended here as the attempt to develop a reflective equilibrium
between one's linguistic intuitions and philosophical theories in the
following manner. On the one hand, one's non-inferential attitudes
about the queerness of certain descriptions are supposed to provide
symptoms of metaphysical difficulties as well as clues to their ultimate
solutions. On the other hand, the correct metaphysical accounts, once
produced, would be expected to provide enlightening philosophical
14 Ryle [1931-2] 1971: 41-2.
explanations of why certain expressions, while technically correct,

none the less had sounded odd to the discerning ear. I now want to
show how Aristotle employs this sort of methodology in Z. 7 to
resolve an issue which was left obscured in Z. 5, namely that of
determining the precise manner in which nose should be mentioned in
the definition of snubness (as (4) requires but doesn't explain).
As befits an Aristotelian 'physical' inquiry, the announced subject
of Z. 7 is 'things which come to be' (y1yv6µsva), and Aristotle opens
the chapter naturally enough by recalling elements of an earlier
'physical' treatment of the same topic in Phys. I. 7. He begins at
1032a12- 13 by distinguishing three types of 'coming-to-be' (yevsm;) 15 -
'natural' ( qJIJ(J"[;/), 'technical' ( rexvn), and 'automatic' ( ano ravroµaum)
-and claiming that in all three types it is possible to distinguish (i)
that by which (6<p' mS) the y1yv6µsvov comes to be, (ii) that from which
(e¢ m3) it comes to be, and (iii) what ( ri) the y1yv6µsvov comes to be, 16
henceforth simply 'the product' (1032a12-15). He then goes on to
specify what these are for each of the three t~pes of coming-to-be
(ro32a15-b31). Next, after declaring at 1032 31-1033a1 that the
possibility of any coming-to-be whatever requires that there be some
matter which pre-exists before the coming-to-be, and which is in some
sense or other a 'part' of the thing which comes to be, he poses for the
first time what will eventually become the pivotal question of z. IO
and II, whether this matter belongs (that is, should be mentioned) in
the formula of what is generated.
Immediately after posing the question, at 1033a3-4 Aristotle floats
an answer that he evidently regards as both 'logical' and unsatisfactory.
He observes that in the case of bronze circle, for example, one might
consider the geometrical shape as a genus (of circles of various mate-
rials), and different material constitutions (in this case, being com-
posed of bronze) as differentiae which distinguish various 'subspecies'
within this genus. I classify this answer as 'logical' because in treating
the material constitution of a thing simply as an occurrent charac-
teristic which can be used to distinguish it from other things, it ignores
altogether the part played by matter in the generation of the com-
posite. 17 The evidence that Aristotle regards this answer as unsatis-
15 As 1032a14-15 makes clear, this is meant to include cases of a substance

coming-to-be-such-and-such (i.e. taking on some non-substantial attribute), and
also the coming-to-be-simpliciter of both substances and non-substances.
16 See preceding note. Where the change in question is a case of coming-to-be
simpliciter, this will be the y1yvoµevov itself; where it is a case of coming-to-be-such-
and-such, it will be the attribute gained during the change.
17 Contrary to a suggestion entertained and rejected by the London Group
(Burnyeat et al. 1979: 60, there referred to as 'b'), this cannot be the infamous
factory is that it becomes clear by the end of z. 11 that in the crucial

cases of naturally generated (biological) substances, a given form can
be realized in only one kind of matter, which makes material consti-
tution useless as a means of making finer subdivisions among things
with the same form. In any event, it is clear enough that Aristotle
rejects this 'logical' answer, 18 for at 1033a6-8 he disambiguates the
'genus as matter' doctrine, because it has things the other way around: here
matter (or material constitution) is aligned with the differentia, and form (or
shape) with the genus. Incidentally, this curious notion of material constitution
functioning as a differentiae of mathematical kinds also makes a brief appearance
at An. Post. I. 5, 74•35-b5 (a more congenial setting, since 'diachronic' consider-
ations are conspicuously absent from the Organon). In Ferejohn (forthcoming) I
argue that Aristotle's reasons for treating bronze isosceles triangle as a further
subspecies of isosceles triangle in the Analytics passage ultimately have to do with
a peculiarity in his example, and not with any deep-seated metaphysical views
about the nature of matter (a subject which evidently had not yet caught his
attention).
18 This view is opposed to that of Gill ( 1989), who takes what I am calling the
),o)'IK()(; and unsatisfactory answer at 1033"3-4 to be the official position of z. 7,

and interprets the entire remainder of the chapter (1033"5-23) as dismissing an
apparent objection to this position. The objection, on Gill's interpretation, is that
in standard usage the composite is not actually called by the name 'bronze' but by
its derivative adjective, 'brazen', so that the matter is not strictly mentioned in the
formula of the composite. According to Gill, Aristotle's response is that this
pattern of usage is a linguistic 'illusion' brought about by the understandable
tendency (which I discuss below) to treat the matter as that from which the
composite comes to be, and the fallacious inference from this that the matter
doesn't survive the generation of the composite (in which case the name of the
matter would not properly apply to the product).
Gill's interpretation not only puts z. 7 in direct conflict with Met. e. 7,
1049•18-24, a passage where (as she herself acknowledges (pp. 123, 151-5))
Aristotle is clearly defending the very same linguistic practice she alleges he is
attacking here, it also demands an implausible construal of the language of z. 7
itself. The question at 1033"1-2 is whether the matter of a composite should be
mentioned in its ..!o)'o;, and Aristotle's immediate response is that we say what
composites are 'in both ways' (aµrportpwr;). Clearly, ~he most natural way of
understanding this is to take aµrportpwr:; to point backward to the options presented
by the question itself, and this would lead us to expect the subsequent lines to
specify one way of describing a composite which does mention its matter, and
another way which does not. On the interpretation I propose, that is exactly what
occurs: 1033"3-4 describe a ).oy1Kor:; (and less accurate) manner of describing the
composite which does make mention of matter (as a material 'differentia'),
whereas the rest of the chapter presents a rpvaiKbr:; (and more accurate) alternative
manner of speaking in which the name of the matter is omitted in favour of
an adjectival form. By contrast, Gill's interpretation requires her to construe
a;irportpw~ as pointing forward, and only to the unsatisfactory proposal at "3-4.
On her view, then, the two ways denoted by aµrporepwr:; are 'mentioning both the
matter ... and the form' (p. 122). Two considerations weigh heavily against this.
original question at 1033a1 in a way that virtually demands a 'physical'

answer: he now asks whether the 'that from which as matter' (e¢ 06 la;
vll.17~) the product comes to be should be mentioned in the formula
of that product.
It is in response to this reformulated question that Aristotle brings
his 'Rylean' methodology to bear on the question of the Definability
of matter/form composites. Recalling his earlier conceptual distinction
(at 1032a13-16, and cf. Phys. I. 7, 19oa5-9) between what a y1yvb1t1::vov
comes to be (and what consequently can be 'said' (ll.ey1::rat) of it), and
that from which it comes to be, he first turns his attention to altera-
tion, a type of change he evidently regards as least problematic,
taking as his example a change in which the non-substantial attribute
healthiness replaces its privative opposite, unhealthiness, in some
persisting substantial subject (~ particular human patient). In such a
case, Aristotle insists, even th6ugh both the underlying subject (the
man) and the privation ('the unhealthy') may correctly said to become
healthy, so that both of the following descriptions are unexceptionable,
(ro) The man becomes healthy,
(u) The unhealthy (thing) becomes healthy,
only the privation can be identified as that from which the healthy
thing is generated. This is to say that
(12) The healthy (thing) comes to be from the man
is at best 'systematically misleading', because a proper 'that from
which' is typically replaced by its opposite, and (12) consequently
fosters the false impression that the patient fails to survive the success-
ful cure. 19 Furthermore, Aristotle observes, because there is such a
clear separation in this case between the 'that from which' (i.e. the
privation) which does not survive the change ('the unhealthy'), and
the product ('the healthy'), there is not the slightest temptation to use
the name of the former to describe the latter by calling the product
(i.e. the healthy thing) 'unhealthy'.
Notice that what makes this 'clear separation' between the 'that
from which' and the product so obvious in alterations is that there is
always an easily identifiable non-persistent privation available to serve
To begin with, the issue of mentioning the form (indeed, the distinction between
matter and form) never arises earlier in the chapter (which would make a refer-
ence to it by aµrporepwr; at "3-4 highly unlikely). More importantly, what is
presented at "3-4 on Gill's view is not two ways of defining a composite (as the
adverbial form would lead us to expect), but a single way which mentions two
items (matter and form).
19 So far, this faithfully echoes Phys. I. 7, 190•5-9.
Michael Ferejohn
as 'that from which' the change proceeds. By contrast, Aristotle now

implies, this typically does not obtain where the change in question is
the generation of a substance. For example, in the building of a brick
house or the making of a bronze statue, 20 he holds that even though
there are in fact certain (low-level) forms (or perhaps 'quasi-forms')
that get replaced by those of house and statue (perhaps heap and
amorphous lump, respectively), he describes these 'privative forms' at
rn33a14 as 'obscure' (M17A.oi;) and 'nameless'(av<i>vu1wi;), by which
he no doubt means that they are not easily identified, and so are easily
overlooked.
Because of the relative obscurity of the privation in substantial
change as compared with alteration, Aristotle reasons that since there
must be a 'that from which' for every change, there is a natural
inclination to allow the matter (which after all, like the privation, does
become the yiyv6µevov) also to fill this apparently vacant role. In that
way one is seduced into thinking and talking about the peculiar 'entity'
introduced at rn33a6 by the curious phrase 'that from which as matter'
(El¢ ob wi; i!).17i;), and into tolerating such descriptions as
( 13) The statue comes to be from bronze,
(14) The house comes to be from stones,
even though (as Aristotle points out at rn33a20-3) these are strictly
speaking inaccurate (i.e. 'systematically misleading') because a genu-
ine 'that from which' is always replaced during a change, whereas mat-
ter is the substratum that persists throughout a substantial change. 21
But where, it might be asked, is the actual harm in this fiction of
allowing matter to take on this additional role in substantial change?
Aristotle's answer, I believe, is that it can easily lead into a deeply
confused idea of the very nature of a material substratum, namely
that it is something that possesses its own individuative nature which
is independent of what forms happen to be imposed on it.
20 These artefacts are of course not paradigmatic Aristotelian substances, and
may in fact not be substances at all in the final analysis. Here as elsewhere
Aristotle resorts to artefactual examples to make points about natural substances
in order to avoid potentially distracting questions concerning the form of biological
organisms (see n. II above). Furthermore, the task of identifying the privative
quasi-forms that are replaced in the natural generation of a biological organism
would be even more difficult than it is for artificial generations.
21 Phys. I. 7, 190•24-6. Code (1976) argues convincingly against the view in
Jones (1974) that matter perishes during a substantial change, but it is also
somewhat misleading to say, as Code does, 'the same matter' survives such a
change. That gives rise to what I will argue presently is the un-Aristotelian idea
that matter has its own independent conditions for identity through time.
Notice that the 'opposites' that get replaced in substantial changes

are always substantial forms (or at least quasi-substantial 'privative
forms' such as lump and heap), and that these would nowadays be
classified as sorta/ concepts. From this it follows that the proper 'that
from which' in a substantial change is always a particular thing (in
Aristotelian terms, a r6& u) falling under the relevant sortal. Thus, in
the case of ( 13), the proper description of the change in question
would be something like
(15) The statue comes to be from a certain lump (of bronze).
However, as noted abbve, Aristotle regards such proper descriptions
as generally unavailable in practice because the privative forms such
changes involve are generally 'obscure and nameless'. 22 His worry,
then, is that this dearth of linguistic resources makes it all too easy to
think that a sentence like ( 13) is itself a proper description of the
change in queston. But since a genuine 'that from which' is always a
r6Jc u, one can then be led to misconstrue the expression 'bronze' in
( 13) as if it made reference to a definite individual thing. On this
misconstrual, the material substratum would then, as it were, illicitly
borrow the appearance of being a r6& u from its assumed role as the
'that from which' of substantial generation.
On the diagnosis offered in Z. 7, even though this way of thinking
is mistaken, it does not in itself lead inexorably into objectionable
metaphysics. We could after all just take (13) as a piece of convenient
(if inaccurate) shorthand by construing the expression 'bronze' in it as
implicitly containing a hidden sortal-as actually referring to some-
thing like the bronze lump of sentence (15). 23 On the other hand,
Aristotle evidently believes that a thoroughly incoherent conception
of matter emerges if one both mistakes (13) for a proper description,
and at the same time continues to think (correctly) of the bronze as
22 In fact, the situation is even worse than my use of the generic privative sorta)
'lump' here (and Aristotle's use of 'heap' (uwpo<;-) at Met. z. 16, ro4ob9) makes it
seem. For on a natural understanding of Aristotle's usual practice in the Meta-
physics of equating specific form with shape, it would seem there is not really just
one 'privative' form of 'lump', but a different such form for every determinate
shape which is not interesting or common enough to deserve a name, and the
number of these is of course infinite. But since virtually all of these specific
'privative' forms really are obscure and nameless, limitations on linguistic re-
sources will in general make the particular proper descriptions of substantial
change inexpressible, and a sentence like (15) should probably be construed as
general and schematic: (15') The statue comes to be out of a bronze something or
other, where the appropriate (determinate, but 'obscure and nameless') sorta)
must be left unspecified.
23 Or perhaps, in light of the preceding note, to a bronze something or other.
something which persists throughout the generation of the statue. For

this latter thought clearly entails that whatever turns out to be the r6&
n referred to by 'bronze' in (13), it cannot be the bronze lump, since
obviously the lump is destroyed during the generation of the statue.
From this position, one is finally led into the profound confusion of
thinking that the change in question involves a third individual thing-
the bronze itself-something which is distinct both from the bronze
statue that is generated and from the bronze lump that is destroyed.
In this way the bronze itself is conceived of as if it were a definite
individual which persists through the change, at one time assuming
the (privative) form of lump, and later taking on the form of statue.
Of course, to think in this way requires the possibility of making
absolutely form-independent references to the matter of a composite
as if it had its own independent principles of individuation. But this
Aristotle firmly believes to be impossible, as is evidenced both by his
statement in Z. IO that 'matter by itself (KaO' avr1?i') is unknowable
(ayvwow<;)' (1036a8-9), and by his later assertion in Z. I I that the
definition of a concrete substance cannot include (strict) reference to
matter, since it is 'indefinite' (a6p1arov) (1037a27-9).
In keeping with the description of the 'Rylean' methodology given
at the beginning of this section, we should expect Aristotle at some
point to argue that the truth of the correct metaphysical position on
this issue (namely, his own position that the matter of a substantial
change is not an individual which persists through the change) is
reflected in, and conversely explains, our considered judgement about
the relative queerness and perspicuity of various alternative ways of
describing substantial change and its products. This is exactly what he
does at Met. Z. 7, 1033a17-20. Aristotle had argued earlier at 1033a7-
8 that in the case of alteration there is not the slightest temptation to
call the product (the healthy thing) by the name of the privative 'that
from which' (the unhealthy), because the latter is easily identifiable,
and is therefore easily seen to be replaced by its opposite during the
change. By contrast, inasmuch as matter is the persistent substratum
in a substantial change, he recognizes that here is some rationale for
applying some matter-signifying expression to the product, and this
µiight appear to justify calling the statue 'bronze'.
Ultimately, however, Aristotle argues that this last inference is
fallacious. The problem, as he sees it, is that if matter is also permitted
to fill in as the 'that from which' for the 'obscure and nameless'
privation, then since the proper 'that from which' in a substantial
Change is invariably a TOOe Tl, the USe Of the noun form 'bronze' for
this purpose can easily create the false impression that one is making a
form-independent reference to the bronze itself as if it were a free-
standing individual in its own right. 24 This false impression would
in turn then engender understandable confusion as to whether the
expression 'bronze statue' should be taken to signify (i) an individual
which is a statue and has the property of being composed of bronze,
(ii) an individual which is the bronze itself and has the property of
being statue-shaped, .or perhaps (iii) a 'compound' entity somehow
composed of both <:>tAhese.
As we saw above, of these only (i) is regarded by Aristotle as
philosophically defensible, and he consequently thinks that a fully
perspicuous grammar should have a way of applying some bronze-
signifying expression to the statue which excludes both (ii) and (iii).
Accordingly, he claims that a philosophically informed usage would
not actually call the product of a substantial change by the name of
the matter, but would instead employ an adjective derived from the
name to express the idea that the product is '(made) of the matter'.
So, for example, he recommends at 1033a17-20 that we say 'not that
the statue is wood, but that it is wooden, nor that it is bronze but
brazen', and so forth. 25 His point is that even though an expression
24 Notice that on the present interpretation, the parallelism expressed by

wancp at 1033a17-20 is weaker than a perfect analogy. Certainly, Aristotle's
main point here is that it is never correct to call the product of any change by the
name of 'that from which' it comes to be, and that this holds as much for sub-
stantial changes as it does non-substantial changes. However, the reasons behind
the conclusion are very much different in the two cases. In a non-substantial
change it is because the 'that from which' does not even overlap temporally with
the product. In a substantial change it is because matter (which has been permitted
to assume the role of the 'that from which', is not a ro& n, and therefore should
not be designated by a noun-form.
25 Whereas I take the shift to adjectival forms to be defended in this passage as
a philosophically sophisticated counter-measure to the inclination to think of the
matter as a definite individual, Gill ( 1989) thinks that Aristotle is attacking it here
as a misguided linguistic reaction on the part of ordinary speech to a perceived
analogy between the two sorts of change: 'Just as speakers do not call the
resulting healthy thing by the name of that from which it came to be (they do not
call the healthy thing "sick"), so they do not call the statue by the name of that
from which it came to be (they do not call the statue 'bronze'). In the case of the
healthy thing they call it 'man', and in the case of the statue they modify the name
of bronze and call the statue "bronzen"' (p. 124). In addition to the fact men-
tioned above (in n. 18) that this intetpretation sets Z. 7 in direct opposition to an
evidently parallel discussion in e. 7 where Aristotle is clearly defending this very
same linguistic practice, the above analogy, if pursued rigorously, would actually
produce a result exactly opposite to the one Gill desires. For as the first clause of
her final quoted sentence indicates, Aristotle's full view (both here and in the
Physics) is not just that the product of an alteration (a) should not be called by
the name of the 'that from which', but also that it (b) should instead be called by
the name of the persistent substratum (namely the man). However, if we take
Michael Ferejohn
such as 'the bronze statue' is not technically incorrect, it is none the

less 'systematically misleading' because it appears to contain a free-
standing reference to matter, and should therefore be replaced in a
perspicuous grammar by 'the brazen statue', which does not even give
the appearance of mentioning bronze as if it were an individual in its
own right.
This point, which I will call the paronymy thesis, is registered by
Aristotle in the context of a discussion not about definitions, but about
descriptions of substantial changes and their products. Still, if it is kept
in mind that I033 3 3-20 is an immediate response to the question (at
I033a2) of whether the matter (of generated things) should be men-
tioned in their J.6yo1, we have ample reason to suspect that the
paronymy thesis is intended to have important application to the
Definability question. In the next section I shall endeavour to sub-
stantiate this suspicion. For on the view I am advocating, when
Aristotle returns to the issue of the Definability of matter/form
composites in Z. IO and I I, he does so with the object of combining
his earlier 'logical' and 'physical' investigations by showing how the
paronymy thesis developed in Z. 7 opens the possibility of allowing
the definition of a composite somehow to contain mention of its
matter, but in a way that avoids the regress argument of Z. 5.
IV
Aristotle opens z. IO at I034b20-4 by reinstating a more general form
of a question which had actually been raised briefly and dropped at
I033 3 I-2 in Z. 7= given that a definition is a complex formula, and
given that (in the usual case) the thing defined is also a complex of
parts, should the formulae (or names) which signify the parts of the
thing be included in the formula of the whole thing? Characteristically,
Aristotle's answer (at I034b24-7) is both yes and no-that in some
cases the parts should be mentioned in the formula of the whole,
while in other cases they should not. However, as the question is
originally treated at the beginning of Z. IO, it is not confined to
the serious candidates for primary substance. For the two sides of
into account the fact that the persistent substratum of a substantial change is the
matter-a fact unaffected by the decision to allow the matter also to assume the
role of the 'that from which'-the strict analogue of (b) is that we should call the
statue by the name of the matter (i.e. that we should call it 'bronze'). The catch,
of course, is that this is so by virtue of the matter's proper role as substratum, not
its assumed role as a 'that from which'. But the crucial point is that the analogy
would give us no reason for shifting to the adjectival form.
Aristotle's answer are illustrated at 1034b25-7 by a pair of examples

that do not seem even remotely substance-like. He argues that the
formula of a circle should not include that of the circle-segments
which make it up, whereas the formula of a syllable should include
those of the letters which it contains. The best way to understand
these examples (and others added later to reinforce them) is as some-
thing like 'control cases' which Aristotle expects the correct account
to respect and explain. His central project in Z. 10, then, is to make
clear what it is about these two superficially similar cases which
accounts for the fact that they turn up on opposite sides of the
question at 1034b20-4, and then to apply the principle of distinction
developed for these cases to the more difficult and more important
cases of things with some plausible claim to substancehood.
The principle by which Aristotle ultimately distinguishes these two
control cases centrally involves a relation he describes at 1034b31-2 as
the 'priority of independent existence' ( np cfvat avcu a)),~hm•
np<'mpov), and which I will refer to as ontological priority. Briefly, the
position developed at 1034b28- 1035 3 14 is that in cases like that of the
syllable, where the part can exist without the whole but not vice
versa-that is, where the part is ontologically prior to the whole-the
part is also logically prior, which is to say that its name or formula is
contained in the formula of the whole. By contrast, in cases such as
the circle and its segments, where the whole can exist without the part
but not vice versa, he maintains that the part should properly be
thought of as a 'merely material' part which should not be mentioned
in the formula of the whole. 26
The positive side of the distinction proposed here is relatively
straightforward and unproblematic, since it is not hard to understand
and endorse the claim that the letters of a syllable are ontologically
prior to the syllable itself. On the other hand, it is not so easy to
comprehend the contrasting point on the other side of the distinction,
that a circle is ontologically prior to its segments. For one can cer-
tainly imagine a disorganized collection of lines of equal uniform
curva~ure that have not yet been composed into a circle. Fortunately,
some illumination on this issue can be obtained from Aristotle's dis-
cussion of the divisibility of geometrical magnitudes in Phys. VIII. 8.
At 263 3 24-b9 he argues that as long as a line remains not actually
divided (that is, remains uncut), its segments can be said to exist only
26 Cf. Cat. 12, 14•30-5, and Met. L1. II, 1018b38-1039"15. There is admittedly
some difficulty in understanding this distinction as exhaustive, since it fails to
recognize the possibility of intermediate cases where part and whole are ontologi-
cally interdependent.
potentially, and that upon its actual division, since the line is no
longer a continuous magnitude-no longer a single line-the products
of the division are no longer line-segments, but rather (undivided)
lines themselves. To put it another way, the point is that line-segments
exist only as potential divisions of actual lines. This rationale can be
transferred straightforwardly not only to the case of the circle and its
segments, but also to Aristotle's other mathematical examples of
'merely material' parts in z. rn. 27
Having distinguished his two control cases, Aristotle turns at 1034b34
to the central business of making out this distinction between 'logical'
and 'merely material' parts in the very special case of substance.
However, inasmuch as there are still two outstanding candidates for
substance (namely form and composite), and since Z. 7 has already
characterized matter as a 'part' of the composite, he now actually
confronts two more refined queries: (i) should the matter of a matter/
form composite be mentioned in the formula of its substance if the
substance is identified with its form, and (ii) should the matter
be mentioned in the formula of the substance, if the substance is
identified with the composite itself?
Aristotle's answer to the first of these questions is simple and firm.
He insists throughout both Z. IO (1035a1-7; 1035a26-b2) and Z. II
(1037a24-5) that matter should not be mentioned in the formula of
the form of the composite for the pre-emptive reason that form
considered by itself is 'without matter' (avsu /J).fl~) (1035a29).
That is, because a form doesn't even have material parts, the question
of whether any of its parts are 'merely material' never arises. By
contrast, he implies at 1035a4-7 that since a matter/form composite
does have material parts, 28 it is reasonable and important to ask
whether those parts are 'merely material', or whether they should
instead be mentioned in its formula. 29
27 Inn. 30 below, I argue that the same rationale can also be transferred to the
single anatomical example (of the finger as a merely material part of the human)
that Aristotle inserts among these mathematical examples at w34"28-33.
28 See also w35"24. Here again Aristotle works with the heuristic examples of
snubness and the bronze statue to make a point which is ultimately supposed to
apply to natural composites.
29 Where I interpret z. IO as developing and employing a distinction between
logically prior and logically posterior (or merely material) parts, Frede (1990)
construes the project of the chapter more narrowly as an attempt to determine
which things are parts of the form, and which are 'parts only of the composite' (p.
II 8). Since he also interprets Aristotle as holding throughout z. 10 and II that
the composite has no formula of its own (but cf. 1034°32, and also 1037•24-7
together with n. 31 below), for Frede the question of whether the formula of the
composite should mention its material parts is a non-starter. On my view, the
The Definition of Generated Composites 3I I
But of course, just to say that matter should not be mentioned in

the formula of a substance unless 'substance' is taken in the sense of
'composite', is not yet to give an affirmative answer to question (ii)
above. In fact, in the immediately following passage ( 1035a8- I 2)
Aristotle draws on insights developed in Z. 7 to construct an argument
that matter cannot after all be mentioned in the formula of the
composite either. At 1035a8-9 he abruptly makes a point which
had not yet come up in z. IO, that either the form or the com-
posite, but never the matter, can be identified with an individual thing
(emowv), and then claims that this is the explanation of the dif-
ference between the examples of the syllable and the circle. As I
understand this compressed passage, it invokes the principle intro-
duced at 1034b28-1035a14 that the only parts which should be men-
tioned in the formula of the whole are those that are ontologically
prior to the whole. However, in this passage I believe Aristotle is also
relying on the insight that if x is to be ontologically prior to y, it must
be capable of existing as an individual in the absence of y. But since
he had already concluded in z. 7 that matter by itself cannot exist as
an individual under any circumstances, the implication here is that
matter can never be ontologically prior to the composite, and therefore
should never be mentioned in its formula.
On this interpretation of the passage, Aristotle's 'abrupt point' at
rn35a8-9 actually fits quite well with what he wants to say about
his two control cases. For, in the first place, there seems to be no
difficulty at all about the individuation of letters occurring outside of
syllables. On the other hand, the passage from Phys. VIII. 8 discussed
earlier (263a24-b9) suggests that the potentially existing segments of
an undivided line fail to satisfy the standard conditions of linear
identity, namely the possession of definite starting and stopping
points, whereas the actually existing divisions of an actually divided
question is on the table, though it ultimately receives a negative disposition.

On Frede's interpretation, Aristotle's ultimate reason for distinguishing the two
'control cases' is presumably just that the syllable is a form (whose parts are
therefore 'logical' parts), whereas the circle is a composite (whose parts arc
consequently material). But if~ cmUaP1/ is understood throughout Z. IO as denoting
the form of the syllable, Aristotle's explanation at I035aIO- 1 I, that the formula
of the syllable should mention its letters because they are parts of the formula of
the form, becomes vacuous, since explanans and explanandum are now identical.
More generally, on Frede's view it is hard to see how the priority of 'independent
existence' contributes at all to the overall argument of z. IO. Yet it clearly does.
as for instance when Aristotle argues at I035b22-5 that the finger is logically
posterior to the animal because it cannot exist as a finger except as part of the
living organism.
line, though individual, are not individual line-segments but individual

lines. 30
What is more, it is now understandable that Aristotle goes on to say
at rn35a14 that circle-segments are 'closer to the form' (f.yyvrtpw rov
dJovs) than such materials as the bronze of the statue. He had
already claimed at rn35a7-9 that form, unlike matter, is correctly
spoken of as an individual, and there is a very natural explanation of
why line and circle segments seem more individual-like than a 'mere
stuff' such as bronze. For all the segments of a line might be said to be
situated on that line, though of course only potentially (where this
expresses a counterfactual thought: if the segments were to exist, they
would exist where some of the line actually does exist). For this
reason, Aristotle believes it is very easy {though ultimately mistaken)
to 'visualize' their potential termini, and in that way to conceive of the
segments as all somehow actually existing within the line and waiting
to be divided out. As such, however, they would be thought of (again,
mistakenly) as a group of very small individual lines, each with its own
distinctive starting and stopping points. On the other hand, since
30 Incidentally, this interpretation of 1035a8-12 also helps to explain the Jone

and seemingly incongruent anatomical example Aristotle includes among his
mathematical examples of 'merely material' parts earlier in Z. IO at 1034b28-34.
At first sight, it is hard to understand Aristotle's claim at 1034h23 that the whole
body is ontologically prior to the finger as meaning that a detached finger cannot
exist as an individual. However, the opening passage of Z. 16 (104ob5-8) indicates
that the analysis he has in mind for this case is parallel to the one reconstructed
above for the mathematical examples on the basis of Phys. VIII. 8. On this
understanding, the parallel point is that so long as the finger remains undetached
it is literally a finger, but only potentially an individual thing, whereas once
detached it is actually an individual thing, but is no longer literally a finger (but
merely, as Aristotle puts it at rn35b25, a finger-homonym). Gill (1989: 127-33)
accepts this point for the 'functional' (non-uniform) parts of animals, but claims
that passages like 1035•19-20 (where Aristotle speaks of the man being 'destroyed'
(rpOdperw) into flesh and sinew and bones) show that he rejects its application to
the 'lower constituent' matter of organisms, which includes both the uniform
animal parts (flesh, bone, etc.) and the four elements. However, in many places
(e.~ Meteor. IV. 12, 39oa10-13, PA I. 1, 641•20, GC 321b29-32, and GA II. 1,
734 24-7) Aristotle applies the point explicitly to the uniform parts of animals.
Furthermore, the claim at the beginning of Z. 16, that no parts of animals are
substances, is not restricted to the non-uniform parts (as it should be on Gill's
interpretation). Hence, on her view he mistakenly (or deceptively) claims to have
shown this for all animal parts even though he does not actually think it holds for
some. In view of this counter-evidence, it is preferable to interpret Aristotle at
1035•19-20 and like passages as following the linguistic practice which he himself
documents in other places by applying terms like 'flesh' and 'bone' homonymously,
especially since alternative ways of describing the matter into which a living
organism is destroyed do not come readily to mind.
there are almost no limits (excepting those entailed by sheer size or

mass) to the possible appearances the bronze of a given statue might
present as a result of its taking on alternative forms (where these will
again include the plethora of 'nameless and obscure' quasi-forms
mentioned in Z. 7), it is impossible even to visualize that matter as an
individual possessing no particular form whatever. This, I suggest,
conveys the force of Aristotle's remark at 1036a8-9 that matter in its
own right (KaO' a!lrqv) is 'unknowable' (ayvwaro<;;).
In view of the argument at 1035ag_ 12, it might now seem that
Aristotle's answer to question (ii) above is a simple negative, and that
his final position is that matter cannot be mentioned in the formula of
either the form or the composite. But here again Aristotle answers
what seems to be a yes-or-no question both ways. For near the end of
Z. 1 I, in a passage which recapitulates the positions developed in Z.
IO, at 1037a24-6 he reiterates his conclusion that matter is not men-
tioned in the formula of the substance qua form, but then he claims at
10373 26-7 also to have established that in the case of the composite,
there is one sense in which matter is mentioned in the formula, and
another sense in which it is not. 31
The sense in which the composite does not have a formula 'with
matter' (µmJ. rfj<;; iJ),17<;;) ( 1037a27) has already been sufficiently do-
31 Frede (1990) translates rn37"26-7 as follows, 'Of the composite substance,

there, in a way, is a definition, and, in a way, there is not' (p. II6), and then takes
the following sentence (•27-9) to explain this by specifying that the composite has
a definition only in the most radically attenuated (indeed, almost farcical) sense
that it possesses the form, and the form has a definition. Finding in this summary
passage the view that the composite can have no formula other than that of the
form, and therefore none which mentions the matter, Frede argues ingeniously
(but, I think, unsuccessfully; see n. 32 below) that other passages from z. IO and
11 (most importantly, rn36b26-30) which seem to contradict this view can
plausibly be interpreted otherwise. Notice, however, that on Frede's translation
of 1037a26-7 (the passage on which so much of his interpretation rests), it does
not fit well with its immediately preceding context. Aristotle says at rn37"22-4
that the general question he has been treating is why in some cases the parts of
the thing defined should be mentioned in the formula of its essence, and in other
cases not. The question is not what things have definitions, and in what sense they
have them. Further, at •24-5 he claims in a µiv clause to have shown that on one
hand the formula of the substance qua form will not mention material parts
(again, because the form has no material parts in the first place). So when he
turns to the case of substance qua composite in the second half of this µtv ... of
construction, he should say whether the material parts should be mentioned in its
formula, not digress into the separate question of whether the composite even has
a definition. Thus, to preserve the integrity of this parallel construction, I believe
that the crucial clauses at 1037"26-7 should be interpreted to mean that in one
sense the composite does have a formula which mentions the matter, and in
another sense it does not have a formula of this sort.
cumented. In giving as his reason for this that matter is 'indeterminate'

(abp1arov), Aristotle is simply rehearsing his reasoning in Z. IO that
since matter cannot legitimately be thought about as if it were a free-
standing individual, it is therefore improper to attempt a direct re-
ference to the matter by the inclusion of its name in the formula of the
composite. Again, I have argued that this position can be traced
ultimately back to Z. 7, where Aristotle was seen to argue that a
composite cannot strictly speaking be called by the name of the
'matter out of which' it is generated for essentially the same reason
given in Z. IO: the matter is not an 'individual thing', and so cannot be
ontologically prior to the composite.
But simply to leave matters there would be to ignore the positive
thesis in Z. 7 that the composite can be called by an adjective
paronymous to the name of the matter, as well as the negative point
left so prominent in Z. 5 that one cannot define a composite such as
snubness (nor, on the authority of I025b32, can one define a natural
composite substance) without somehow making mention of its matter.
My principal contention here is that the other sense Aristotle alludes
to at I037a26, the one in which he says there is a formula of the
composite which somehow (mii;;) mentions its matter, is one that
arises out of both of these views.
Notice, to begin with, that the main thrust of z. I 1 is to soften the
distinction, which had been drawn so sharply in Z. IO, between sub-
stance construed as the composite thing (which necessarily contains
matter), and substance construed as form (which can't contain matter).
By working through a progression of cases, Aristotle presses the point
that even though it is always possible to separate matter and form in
concept, none the less as one moves away from examples of artefacts
(where a given form can generally be instilled in many different kinds
of matter), and towards the all-important examples of biological sub-
stances, one finds more and more that the forms in question, while
still of course separable in concept from their material instantiations,
begin to display strong essential dependencies upon particular kinds of
material constitution. For example, he points at 1036b3-5 to the
obvious intimate connection between the form of human (i.e. the
soul) and the flesh, blood, etc., which make up composite human
specimens. On the basis of this observation, he then declares at
1036b22-6 that to 'do away with matter completely' (arpaipcfv riJv
i!l.17v) when trying to define sensible composites is to be 'over-
reactive' (m;pil:pyov) and 'leads away from the truth' (am::iyn ano roi3
a),17eoD;;). For, he continues, the fact is that some sensible composites
(indeed, all natural sensible composites) just are (and should be
defined as) instances of 'this in that' (r6J' sv u{)Je), that is to say, a
certain form in a certain kind of matter. 32
Returning now to 10373 27 with this in mind, it becomes significant
that what Aristotle actually says there is not (as some have understood
him) 33 that there is a defining formula of primary substance (which is
certainly true, but not directly relevant here), but rather that there is
a formula 'according to' or 'with respect to' (Kara) the primary sub-
stance. To make this remark continuous with its immediately preceding
context, which is clearly concerned with the question of whether
matter should be mentioned in the formula of the composite thing, 34 it
should be taken to mean (a) that there is such a thing as a formula 'of
the composite according to its (primary) substance', and (b) that this
32 This passage along with Aristotle's elaboration and defence of it at 1036"28-

33 constitute the most formidable obstacle to the contention in Frede (I 990) that
'Z. 10 and 11 in general clearly takes the line that in the definition of substance
there is no place for the mention of material parts or matter' (p. 118). Conse-
quently, Frede is obliged to reject the most natural interpretation of b28-30,
where Aristotle says that it is not possible to define an animal 'without its parts
being disposed in some manner' (aw:v rwv txomvv nw;), and appears to mean
by this that the definition of the composite animal should include mention of the
disposition of its material parts. Against this, Frede suggests that the most
Aristotle commits himself to in this passage is that the definition in question
should imply (but not state) that the thing defined is the sort of thing that has
material parts of a certain kind. In support of this suggestion, Frede points out
that b28-30 also states that the animal cannot be defined 'without motion' (avw
K1v~m;w;), and Frede claims that 'it is rather doubtful whether Aristotle thinks that
the definition of an animal also has to make reference to motion or change' (p.
121). However, this last claim is not so obvious if one takes into account that a
good Aristotelian definition of a certain kind of animal (or for that matter. the
genus animal itself) will presumably make reference to a distinctive set of cap-
acities (e.g. perception and self-movement), and that these will in general be
capacities to be moved or changed in certain ways.
Frede himself evidently has some qualms about his proposed interpretation of
1036b28-30, since he also offers the secondary line of defence that its implausi-
bility can be offset somewhat by the fact that virtually the rest of z. IO and 11
concurs with the summary passage at 1037"26-30 in holding that matter should
not be mentioned in the definition of the composite. Thus he writes, 'even if the
suggested interpretation of 1036b28-30 should be unacceptable ... the remarks in
ro36b28-30, rather than the remarks in the summary, should be treated as
anomalous and requiring special explanation' (p. 122). Since I argued inn. 31 that
even Frede's leading passage, the summary at 10373 26-30, allows for a manner in
which some reference to matter should be included in the definition of the
composite, I do not find this attempt to motivate his departure from the most
natural interpretation of 1036b28-30 compelling.
33 Frede 1990: rr6-17.
34 See n. 3 r above.
special sort of formula does somehow mention its matter. The ex-
planation given for this at 1037a29-30 is that the primary substance of
a thing is the 'indwelling form' (ro el<5or:; ro evov) which combines with
matter to make up the composite. This is meant to suggest that it is
possible to give a formula of a composite substance partly by means of
giving the defining formula of its indwelling form (i.e. its primary
substance). But this cannot simply mean that the two formulae are
identical. That would not just soften, but utterly obliterate, the distinc-
tion between form and composite. It is much more plausible, there-
fore, to understand Aristotle here as looking back to his observation
at 1036b23 that the composite should be conceived of as a 'certain
[form] in a certain [matter]' (r6'5' ev up&).
On this interpretation, then, to give a formula of a composite
'according to its primary substance' is to define its form as indwelling,
or what comes to the same thing, to define it as enmattered. Thus, for
example, when at 103?1129 he cites the formula of the soul (the form of
human) as the appropriate one for the composite human, he is not
simply saying that the form and composite have the same formula, but
rather that if the formula of soul is some phrase 'XYZ', then the
formula of human would be 'XYZ in flesh, blood, etc.' Similarly, if
the formula of concavity is 'PQR', then that of the composite snubness
would be something like 'PQR in a nose'. Not coincidentally, this sort
of formula of matter/form composites is explicitly referred to as
'enmattered formulae' (evvl.01 }.oyo1) in De An. I. I (403"25).
If we now examine more closely how references to matter occur in
such 'enmattered' formulae, it is possible to see the final position of
Z. 10 and I 1 as incorporating the earlier discussions of Definability in
Z. 5 and 7. For by contrasting these evvhn lcbyo1 with the 'formulae
with matter' ruled out at 1037a28, Aristotle is in effect invoking the
paronymy thesis of Z. 7 by asserting that even though it is not strictly
speaking possible to use the name of the matter in the formula of a
composite, it is none the less possible to make an indirect allusion to
the matter in that part of the composite's livvl.or:; Xoyor:; which sig-
nifies the enmatterment of the form. In other words, we are now to
think of such expressions as 'in-flesh, blood, etc.', 'in-bronze', and 'in-
a-nose' as fused prepositional phrases in which apparent references to
matter are only apparent. 35 However, when construed in this way,
35 It might seem that the view I am locating at the end of z. r r is already
present at Z. 5, 103ob3r-2, where Aristotle remarks that snubness is 'concavity in
a nose' (Ko1).br11r; iv fnvi), \and perhaps also at Soph. El. 31, 181 1135-182"7). How-
ever, the context of ro30'3r-2 clearly indicates that the statement in question is
not a positive technical suggestion for defining snubness, but rather a parenthetical
remark couched in the vernacular and inserted for the sole purpose of supporting
these 'enmatterment' phrases turn out to be virtually equivalent in

function and meaning to paronymous adjectives such as 'brazen' and
'wooden', whose use was rationalized in Z. 7, since there is really very
little difference between describing a statue as a certain shape in-
bronzed and describing it as a brazen instance of that shape. 36
We can now finally consider what becomes of the 'logical' puzzle in
Z. 5 concerning the composite snubness on this proposal. Recall that
what really powered that argument was the idea that the essentially
negative assumption (4), that snubness cannot be defined without
somehow mentioning nose, seemed to require that it be defined (as in
(4*)) in a way that explicitly includes the name 'nose'. However, we
have just seen that in Z. 11 Aristotle holds that there also cannot be a
formula of a composite 'with matter', which effectively rules out the
hypothesized definition of 'snub' in Z. 4 as 'concave nose'. But if there
is no formula of a composite with matter and also none without
matter, it might seem to follow that composites cannot have any
definitory formulae whatsoever. My final suggestion is that Aristotle's
recognition in Z. I I of the possibility of i:vul.01 },6yo1 of composites
enables him to escape this apparent dilemma. For even though such
formulae clearly are not with matter, as (4 *) would have it (because
they don't include the name of the matter), they are also not quite
Aristotle's essentially negative point (at rn30°30- I) that it is impossible to define

snubness without somehow mentioning nose. What is more, on the view I am
defending, since Aristotle's survey of the philosophical problems incurred by
directly mentioning matter in the definition of the composite does not take place
until z. 7, there is no reason to interpret him in z. 4 as assuming the sophisticated
response to these problems (according to which an 'enmatterment' phrase such as
'in-a-nose' does not really mention nose) developed in z. 1 I.
36 I do not mean to suggest that Z. r I settles the question of whether hvi.01
i.111•01 satisfy the very strong unity condition required for full-fledged definitions.
Evidently the initial move in Aristotle's treatment of this issue occurs at the
beginning of Z. 17, when he says that it is necessary to make a 'fresh start' in the
investigation into the nature of primary substance (rn41°6-7), and then imme-
diately introduces a thought which, curiously enough, had not come to light in the
earlier chapters: that 'substance is a sort of "principle" (apx~) or "cause" (airia)'
(rn4o"rn). Then, at 1041°7-9 he invokes this thesis by characterizing the primary
substance, or form, of a sensible composite as the cause by virtue of which its
matter constitutes not just an 'aggregate' (awpb.;), but a definite and unitary thing.
This in turn allows him to argue in Met. H. 6 (rn45•23-33) that because a
substantial change is a process in which some matter which has the potential to
take on a certain substantial form is transformed into an actual individual instance
of that form, the form and matter of a sensible composite are unified in the sense
that the former is actually the same (individual) thing that the latter is only
potentially. For a very different interpretation of Aristotle's strategy in H. 6, see
David Charles's contribution to this volume.
Michael Ferejohn
without matter because of the indirect allusion to the matter accom-

plished by the fused prepositional phrases (or equivalent paronymous
adjectives) which signify that the form in question is being defined as
an enmattered form. Notice, however, that on this way of satisfying
(4), to play on the language of the argument itself, there will not even
be one 'nose' in snubness, but only an 'in-nosed' (or, equivalently, a
'nosen'). But that way leads into no regress.
v
Principles of
14
Aristotle's Conception
of Metaphysics as a Science
ROBERT BOLTON
I. ARISTOTLE'S CONCEPTIONS OF SCIENCE
At the beginning of his exposition of his own positive metaphysical

doctrine in Metaphysics r. I, Aristotle announces that the subject is a
science, the 'distinct science which investigates being qua being and
the things which pertain to this in itself' (1003a21). This claim, that
metaphysics is a science (brnn~µr1), is one which may initially strike a
modern reader as peculiar, and implausible, in various ways. To begin
with, it may be hard for us to believe that the body of metaphysical
truth has anything like the explanatory structure which our paradigms
of science tend to achieve. That is, it may be hard to think of
metaphysics as capable of discovering some small set of basic principles
or laws which are absolutely fundamental to what is, from which all
the other significant metaphysical truths can be derived and, more
importantly, explained. Secondly, it may be hard for us to think of
metaphysics as having anything like the kind of potential for epistemic
success which we expect of a typical science. For instance, metaphysics
does not seem to have or to be able to collect a body of highly
warranted empirical or other information generally known to re-
searchers which they can use to test; and to confirm or reject, alterna-
tive proposals concerning the fundamental metaphysical truths. In fact,
metaphysical truths, or at least some of them, such as the principle of
non-contradiction, are so general, and so ubiquitous in their appli-
cation, that they seem to be presupposed by any body of information
that might purportedly be used to test and confirm them.
This account of some features of some typical sciences, features
which metaphysics may well seem to us to lack, is, of course, one
which many philosophers would now wish to reject. The idea that a
science requires and aims to find objectively fundamental brute
principles which uniquely explain why all of the other truths of that
science hold involves a realist conception of theoretical scientific truth
© Robert Bolton 1994
322 Robert Bolton
and of explanation which some would now reject. Some would also
question the idea that we need or even can possess a testing procedure
which relies on a relatively theory-neutral body of empirical or other
data which we can and should use to evaluate competing theories in a
given science. However, these are not features of a genuine science to
which Aristotle would obviously object. It is, in fact, Aristotle himself
who is primarily responsible for the development and the recurrent
influence of the conception of science on which these features are
crucial. In An. Post. I. 2, for instance, he describes at length the
requirement that genuine scientific knowledge must fit into the kind
of real explanatory structure outlined above; and in Prior Analytics I.
30, for instance, he elaborates an empirical version of the discovery
and confirmation procedure just described and requires it 'for any art
or science whatever' (46a17-27 at a22). 1
The difficulty in seeing metaphysics as a science, as science is de-
scribed in Aristotle's own Analytics, has not been lost on recent
interpreters of Aristotle's Metaphysics. It is now standard to hold that
in the Metaphysics, reasonably enough from our own contemporary
perspective, Aristotle develops a new conception of science. On this
conception it is not required that a science have the kind of systematic
explanatory structure described in the Analytics. Moreover, in his new
science, Aristotle uses his dialectical method, or an extension or
development of it, to reach definitively confirmed results without re-
ference to any body of, even relatively, theory-neutral empirical or
other data. This, roughly at least, is the result to which perhaps the
dominant line of research in recent years on Aristotle's conception of
metaphysics as a science has led us at this point. 2
1 This account of Aristotle's picture of science in the Analytics is developed and

defended against alternatives in R. Bolton (1987).
2 The most fully developed version of this view is found in T. H. Irwin ( r 988).
There is a well-entrenched school of interpretation according to which the

Analytics is not ever relevant, in the way suggested above, to understanding
Aristotle's actual procedures of inquiry and confirmation in any of his scientific or
philosophical works. Rather, his method of inquiry and confirmation is always
predominantly the method of dialectic described in the Topics. It has also been
typical for those who follow this approach to add that the Analytics is relevant for
something if not for inquiry or confirmation, in a genuinely scientific subject such
as physics. Its use is to show how to systematically understand or present results
after inquiry and confirmation. However, even this limited relevance of the
Analytics-in either of the two versions sketched-has seemed to most inter-
preters in this tradition to be annulled when we come to the Metaphysics, since
Aristotle there abandons the requirement that a science be expressible in the kind
of systematic structure described in the Analytics, even if he retains a version of,
or a successor to, his dialectical procedure for confirmation. I here ignore these
dif~erences in the tr~atment of Aris~otle's Il1ethodology prior to the Metaphysics
Metaphysics as a Science
The main question which faces us now concerns the adequacy of

this result. Does Aristotle, in Met. r, show us that he has given up the
requirement of systematic structure for a science developed in the
Analytics; and does he show us that a special dialectical procedure for
justification, or some extension thereof, which is non-empirical and
holistic in important ways, is the only procedure we need to use for
confirmation, or for understanding, in the newly conceived science? It
will be economical to concentrate here on the latter question since a
proper treatment of it will take us a long way toward dealing with the
former. 3 Discussion of this question has frequently concentrated on
Aristotle's treatment of the so-called common axioms, particularly the
principle of non-contradiction, in Met. r. This is natural enough given
that the common axioms constitute one of the types of basic first
principles marked off in the Analytics (e.g. in An. Post. I. rn) and
Aristotle's full treatment of the principle of non-contradiction should
show us whether he is taking a different approach to the justifi-
cation or the understanding of principles than the one he earlier
recommended. Aristotle's discussion of this principle has also seemed
to many to show how genuine a priori knowledge of this and other
truths of metaphysics is possible. Recent interest in the question
whether anything is knowable a priori also gives Aristotle's discussion
a special timely relevance.
II. ARISTOTLE'S DEMONSTRATION BY ELENCHUS
The crucial part of Aristotle's discussion of the principle of non-

contradiction comes in r. 4 where he proposes to 'demonstrate by
elenchus', that is to demonstrate by refutation, that those who deny
the principle are mistaken (1006arr-12). The precise form which this
elenctic demonstration takes is, however, not easy to determine and
there is a wide variety of interpretations of its content and significance
in the literature. One of these goes counter to the prevailing trend.
That is the view that Aristotle's intended proof is not a dialectical
argument, but in fact a demonstration in much the sense in which a
demonstration is understood in the Analytics. It is not, of course, a
demonstration of the principle of non-contradiction itself. If anything
is clear in Aristotle's discussion it is that there can be no strict
demonstration of the principle itself ( rno6a5- II). Rather, it is a
demonstration of some truth about the principle. Just as in r. 3
Aristotle attempts to demonstrate a truth about the principle, namely
that it is 'the firmest principle of all' ( rno5b22 ff.; cf. 10063 3-5), so in
"
Robert Bolton
r. 4 he attempts to demonstrate another truth about it, namely that it

is what we may call a super first principle, a principle presupposed not
only by certain demonstrable truths but also by other indemonstrable
principles and, indeed, by all coherent thought and speech. A main
argument in support of this reading is that some of the premisses in
the elenctic proof introduce Aristotle's own special semantical and
metaphysical doctrines which one could not expect an opponent to
grant in a dialectical discussion. A second argument is that if we
construe Aristotle's aims in this way then he will not violate his own
claims that proof is always from what is more intelligible or better
known and that nothing is better known than the principle of non-
contradiction. Aristotle and his audience believe that everyone knows
this principle without proof; so he could hardly be trying to prove it to
anyone. 4
There are, however, two facts which rule out this reading of
Aristotle's aims. First, Aristotle's actual main conclusion at the end of
his elenctic demonstration is as follows:
It is not possible that it is at the same time true to say that the same thing is a
man and not a man. (1006h33-4)
So what Aristotle actually proves and concludes here is the principle
itself, or, more specifically, a chance substitution instance of the
principle. He does not ever demonstrate in this text a special truth
about the principle, namely that it is a super first principle which
everyone believes. It may be, and we will want to explore whether it
is somehow the moral or lesson of what he does prove that the
principle is a super first principle. But this is not anything which is
anywhere proved or demonstrated.
The second, and ultimately more important, reason why Aristotle's
actual proof cannot at all be intended to conform to the model for
explanatory proof or demonstration in the Analytics is because he
makes it a clear requirement of his proof procedure that the person
responsible for the premisses is his opponent, who must accept what is
required for the proof in response to questions (1006a15-I8, 25-6).
Aristotle makes it quite clear elsewhere that the aim of those engaged
in a proof procedure which includes this requirement is not demon-
stration (Soph. El. II, I72a15-21, An. Post. I. II, 77a33-4). More-
over, Aristotle does not accept assumptions which make genuine
proof of this principle, or instances of it, either impossible or un-
necessary. Certainly, for Aristotle, nothing is more intelligible by
4 This type of view is developed in A. Code (1986b), e.g. at pp. 346 and
354 ff. See also Code (1987a).
Metaphysics as a Science 325
nature than this principle. But Aristotle does not believe that the only
genuine proofs which yield knowledge are proofs from what is more
intelligible by nature, i.e. explanatory demonstrations. There are also
proofs from what is more intelligible to us, such as inductive proofs
and dialectical proofs, which may yield knowledge of facts even if
they do not explain those facts. (See An. Post. I. I, 71a1-9, I. 13;
and further below.) It is a serious mistake to think that the only thing
that counts as a 'proper proof' for Aristotle is an explanatory proof
or demonstration from what is more intelligible by nature. Thus,
when Aristotle claims that the principle of non-contradiction is non-
hypothetical and thus known by anyone who knows anything he need
only mean, given the context, that everyone who has any scient;ji.c
knowledge knows it (1005b8-17, cf. An. Post. I. 2, 72a16-17). Else-
where, in fact, he explicitly claims that all scientific first principles,
since they are not innate, must be learned from previous knowledge
(An. Post. II. 19, 99b26-30). So he clearly does not think that
anyone who knows anything, in any way, knows the principle of non-
contradiction.
It is much more common to suppose that what Aristotle intends to
give us here is not any kind of demonstration but rather a dialectical
proof. That has seemed reasonable in view of the fact that it is a
standard feature of dialectical argument that premisses are conceded
by an answerer in response to questions. But fairly general agreement
on this point has not been matched by extensive agreement on further
details. One reason for this may be that while there is much agreement
that Aristotle's demonstration by elenchus is dialectical, typically very
little attention is paid to what that precisely involves. In specific, there
are few references in the literature on Met. r to the details of
Aristotle's discussion of the rules of dialectical argument in Topics
VIII or, more significantly, to the details of his elaborate discussion of
the rules for dialectical elenchus, or refutation, in the Sophistical
Refutations. Since Aristotle says his argument is an elenchus ( 1006a I 8)
and the Sophistical Refutations gives us his theory of elenchus, particu-
larly dialectical elenchus, it would seem that a useful new strategy to
follow to achieve a better understanding of Aristotle's aims and results
in Met. r would be to investigate what the bearing on this text is of
his own explicit discussiop of elenctic dialectical arguments.
III. DIALECTICAL RULES IN METAPHYSICS r. 4
As an introduction to this line of inquiry it is worth drawing attention

to one further feature of Aristotle's elenctic refutation, beyond the
Robert Bolton
fact that he insists that its premisses are his opponent's concessions in
response to questions. There is another clear indication that Aristotle
is explicitly following the rules of some special form of question and
answer discussion in generating his refutation. At an early stage in the
main argument he insists on a certain type of answer from his respon-
dent as essential for people 'to t5ialJ:rw0az with each other' ( 1006°8).
Later he insists again that if his respondent does not answer in a
certain way '01) ()ta}i,ycraz' (1007a20). Aristotle begins Soph. El. 2 with
the comment that: 'There are four kinds of argument used in ref>
t5ial~tyea0ai.' It is clear from his discussion there that this term, ro
<5taJJ:yw0az, designates rule-governed argumentation based on question
and answer discussion. He goes on to describe the rules of argument
for each of the four kinds. We must ask, then, which of the four kinds
of argument in ref> t5ia/J;yw0az Aristotle is using an his elenctic
refutation.
The four types of argument which exhaust ro <>ialJ:ywOw are
Didactic, Dialectical, Peirastic, and Eristic (165a38-9). Didactic argu-
ments are, according to Aristotle's description, strict demonstrations
(165b1-3, 8-9). Even didactic demonstrative arguments, it should be
noted, require the consent and agreement of an answerer to premisses
introduced by a questioner. So it is not enough, as is often supposed,
to guarantee that the elenctic demonstration is a dialectical argument
and not a genuine demonstration, that its premisses are conceded by
Aristotle's opponent. What does guarantee this, however, is the
stronger requirement that the premisses of a didactic argument are
'taken on trust' by the answerer (b3). That is, the responsibility for the
correctness and appropriateness of the premisses in didactic argument
belongs to the questioner, who is trusted on this matter by the an-
swerer. This is precisely not the case in the elenctic demonstration, so
that demonstration cannot employ didactic, genuinely demonstrative,
argument. Obviously enough, we cannot suppose that Aristotle intends
his elenctic demonstration to be an eristic argument. So we are left
with just two options, either the elenctic demonstration is a dialectical
argument or it is a peirastic argument. Since Aristotle makes it clear,
in Soph. El. 8 and II (I69°25, I7Ib4-5) that peirastic argument is a
special type of dialectical argument, the elenctic refutation will be a
dialectical argument, either one which simply follows the generic rules
for dialectic or one which also follows the more restrictive rules of
peirastic. In the Sophistical Refutations, however, Aristotle assigns the
job of skilled dialectical elenchus to peirastic (8, I69°20-9). This by
itself should lead us to expect that the elenctic refutation, as an
elenchus, will be a peirastic argument.
Fortunately, we do not need to look further at the details in the
Sophistical Refutations to see whether the elenctic demonstration in

Met. r. 4 is a dialectical or, more specifically, a peirastic argument.
Aristotle gives us a direct answer to this question in r. 2, rno4b25-6.
There he says: 'Dialectic is nezpaauKif concerning those things on
which philosophy is yvwp1ar1K17.' Ross translated this passage as fol-
lows: 'Dialectic is merely critical where philosophy claims to know.'
This translation has been followed, in more or less equivalent
variants, by much of the later tradition. 5 In this translation the text
has served as a primary basis for large-scale results in the interpretation
of Aristotle's methodology in the Metaphysics. Owen, for example,
took this passage to show that Aristotle had abandoned his former
confidence, expressed in Top. I. 2, in the ability of dialectic to estab-
lish principles in science or to lead one to know them, and had
'demoted' dialectic to a lesser department of its old province. 6 In
another variant in this tradition the passage has been held to show
that before the new methodology of the Metaphysics Aristotle never
regarded dialectic as capable of more than mere testing or criticism
or, thus, as capable of leading one to knowledge of the truth of any
results. Thus, since Aristotle later goes on to use the elenctic refutation
to establish results dialectically, so it is argued, he must be introducing
a new form of dialectic, not earlier envisaged, a form which is not
'merely critical'. 7
However, in view of the material in the Sophistical Refutations it is
difficult to support the type of translation on which these interpre-
tations are based. The term ne1paauK1 (or 'the art of testing') does not
introduce a descriptive feature of all dialectic, or all dialectic used in
scientific inquiry-that it merely tests or criticizes or probes but does
not establish anything. As we have just seen, m;1paar1K17 is a technical
term uniformly used elsewhere to describe a special type of dialectical
argument. In Soph. El. 11, Aristotle uses the very same language as
he uses in Met. r. 2. He says: '1 <5taJ,f,K!IKft nezpaarzK1 earz (m;pi (JJV)'
(171b9). He means by this not that dialectic can merely test or probe
on certain matters but rather that dialectic uses its special peirastic
form or capacity on certain matters (171b4-5). This then is what he
means in Met. r. 2: When it comes to those matters which (first)
5 Tredennick translated the passage: 'Dialectic treats as an exercise what

philosophy tries to understand.' Kirwan has translated it: 'Dialectic probes where
philosophy seeks understanding.' Irwin has translated it: 'Dialectic tests in the
area where philosophy comes to know.' These translations are found, respectively,
in the Oxford translation, the Loeb Library, the Clarendon Aristotle Series, and
in Irwin ( 1988: 174; cf. vii).
6 Owen 1986: 216.
7 Irwin 1988: 174-5.
Robert Bolton
philosophy deals with, dialectic should use its special peirastic form or
capacity. This does not in the least require that when it deals with
philosophical subjects dialectic merely probes or tests or criticizes but
does not establish or lead one to know anything. Aristotle has no such
view of peirastic in the Sophistical Refutations. For reasons which we
will go into shortly, he believes that peirastic is a method of 'true
refutation' which can be used to objectively refute specific claims and,
thereby, to objectively prove certain claims (9, 17oa22-6, 36-9,
bg_ II). So there is no need to supersede dialectic as peirastic, in the
Met., in order to develop a procedure which is more than merely
probing or critical. 8
It is also a mistake to claim that the doctrine that dialectic should
specifically use its pcirastic capacity in dealing with scientific matters
(such as first philosophy deals with) is a new doctrine which demotes
dialectic or changes Aristotle's view of it. In the Sophistical Refutations
Aristotle .. specifically introduces peirastic as a form of dialectic de-
signed to deal precisely with the subjects that the scientist deals with,
without requiring the special scientific capacity which the scientist has
(9, 17oa36-bll). The claim in Met. r. 2-that dialectic uses its
peirastic capacity to deal with those matters which philosophy uses its
special scientific capacity (its yvwpumK~) to deal with-is in fact an
exact and concise summary of the doctrine of the Sophistical Refutations
on this point. This does not demote dialectic from the sphere assigned
to it in Top. I. 2. There Aristotle gives as his reason why dialectic
plays an essential role in the search for scientific principles, the fact
that dialectic is s¢1rraauK~-an art of examination (101b3). He does
not explain how dialectic does its testing on specifically scientific
matters or what kind of success it may expect until the Sophistical
Refutations, where he assigns this special function to the art of pcirastic
testing. In any case, it is clear that Aristotle actually tells us in
Met. r. 2 that the form of dialectic which is to be used on the subjects
8 Irwin's proposal is that Aristotle first, in r. 2, condemns dialectic as capable

merely of testing or criticism on scientific topics and then in r. 4, without
warning, introduces a new form of dialectic which is more than merely critical
which then becomes his dominant philosophical method in all later work, without
ever producing any identification of this new method, and expecting us to see that
the supposed general condemnation of dialectic as a tool for science in r. 2 was
meant only to apply to earlier forms of dialectic. But Aristotle does not say that
dialectic used to be only peirastic, he says that what he now calls dialectic is, i.e. is
now, peirastic, when it concerns the topics of (first) philosophy. Since Aristotle
clearly thinks that this procedure, as used in Met. r. 4, establishes its conclusion,
this is itself sufficient indication that the procedure has, for him, this power. The
similar use of the same procedure in Phy. I is discussed in Bolton (1991).
which first philosophy or metaphysics deals with is peirastic. What

should this lead us to expect about the elenctic demonstration?
IV. THE RULES OF PEIRASTIC ARGUMENT
There are two special requirements introduced in Soph. El. 2, which

must be satisfied by the premisses of peirastic arguments, over and
above the general requirement imposed on every dialectical argument
that its premisses must be evoo¢a or noted beliefs ( 165b3 ff.). The
first of these is that peirastic premisses must be things actually believed
by the answerer or opponent (165b4-5). They are not simply pre-
misses which the opponent must concede, because they are hcfo¢a
of an appropriate sort, as is required in any dialectical argument
(Top. VIII. 5-6). Nor can they be accepted on trust from the ques-
tioner (165b1). They must be the actual standing convictions of
the answerer or opponent. This, of course, fits appropriately with
Aristotle's insistence in the elenctic demonstration in r. 4 that the
opponent alone is fully responsible for the premisses of the argu-
ment ( 1006a 17- r 8). The second, and more striking, special require-
ment for peirastic premisses is that they are 'things which must be
known (E:i<5tvai) by anyone who pretends to have scientific knowledge
(imwrf/µr1)' (r65b5-6). Given this, the premisses of peirastic argu-
ments, unlike the premisses of dialectical arguments generally, must
be true since what is known, or knowable, by certain people must be
true. This is part of the reason why peirastic proofs do genuinely
establish their conclusions and genuinely refute the views of opponents
which are contradictory to these conclusions.
Aristotle later expands on this requirement in Soph. El. I I. There
he makes it clear that by the one who pretends to have knowledge
he means the typical opponent whose position is being tested in a
peirastic argument (171b5-6). So the premisses must be things which
are known by Aristotle's opponent. This is further confirmed by
Aristotle's characterization of peirastic premisses as things which
'everyone, even ordinary people ... know themselves, no less than
[those who profess brmn/µIJJ' (172a30-4). This permits 'everyone ... to
fudge those who profess knowledge' on every subject 'since there are
many of these [commonly known] things in every area' (172"30-7). It
has often been claimed that there are severe limitations on the ability
of dialectic to genuinely establish things since dialectic is restricted to
reasoning based merely on certain accepted beliefs or iivoo¢a which
may well be false. As we have already noted, some have argued that
this limitation required Aristotle, in Met. r, to invent a successor to
330 Robert Bolton
dialectic or a new stronger form of dialectic capable of arguing from

true premisses and of truly establishing things. But, as we can now
see, there was no need to invent a form of dialectic with this capacity;
it already existed in peirastic. We can now see why peirastic has
this capacity. Peirastic argues from premisses which are known by
the opponent and also commonly known and not merely commonly
believed. Thus, conclusions deduced from these premisses are also,
thereby, known. 9 How the premisses get to be known is, of course, a
crucial question. For present purposes that question will be best
explored with the actual premisses of the elenctic demonstration in
view. But Aristotle's account of peirastic as described thus far already
imposes significant constraints on our interpretation of the elenctic
demon~tration. Its premisses must be things which it is plausible
to suppose that the opponent, and everyone, knows. It will not
even be enough if they are, either implicitly or explicitly, believed
by everyone. 10
9 Some would suggest that 165°5-6 only requires that the questioner select and
the answerer agree to, as premisses in peirastic argument, things which the
answerer (believes and) would know if he had br1arf/p17. But this would require the
questioner and answerer to know what actually counts as brtariJJlll in some area in
order to select proper premisses and Aristotle explicitly says that this is not
required for the proper practice of peirastic. In any case, r72"30-7 requires actual
(non-scientific) common knowledge of the premisses. This is discussed in further
detail in Bolton (1990, 1991, 1993).
10 This constraint fits well with another important feature of Aristotle's discus-
sion in Met. r. 4. We have noted that Aristotle describes his elenctic demon-
stration as an instance of HJ <>wi.i:yfo(1w, rule-governed argumentation based on
question and answer discussion. There are many indications of this in the text
beyond those already noted. (See esp. 1006"19, 24, 1007'18.) But these indications
abruptly cease at 1007''20. After that point Aristotle goes on to offer many further
objections to those who say that it is possible 'to predicate contradictories at the
same time' (wo7°r8). He attempts to show that this claim, in various forms, has
various unacceptable consequences. But he makes no claim that these various
unacceptable consequences are things for which his opponent actually must take
full responsibility in the way required in peirastic argument. Aristotle says, for
instance, that it is a consequence of his opponents' claim that 'they do away with
substance and essence' ( wo7''20- r ). They say that something can be a man and
not a man. But being a man is in fact the substance of what is a man and that
implies that what is a man cannot (either at the same time or ever), while being
that very thing, be something other than a man, which it would be if it was (ever)
not a man. As various writers have pointed out, it is quite implausible to suppose
that Aristotle's opponent here must be someone who accepts the aspects of
Aristotle's essentialism or the particular essentialist claims to which this argument
appeals. A fortiori, it would be implausible to suppose that these reflections of
Aristotle's essentialism are known by Aristotle's opponent, and everyone else, or
that Aristotle could think that they are. But this is what is required in peirastic
Metaphysics as a Science 33r
V. FURTHER REFLECTIONS OF
PEIRASTIC PROCEDURE
We can also use the Sophistical Refutations to settle another important

preliminary point. There is much variation and uncertainty in the
literature over just what claim or claims of his opponent Aristotle is
trying to refute in his elenctic demonstration. In Soph. El. 5 Aristotle
gives us his official definition of an elenchus. It is worth quoting now
for the help it gives us on this and further matters:
An elenchus is [a syllogistic proof of] the contradictory of one and the same
thing [as the opponent holds], not in name merely [by virtue of an ambiguity]
but in fact, and not using synonymous words but the same words [as in the
opponent's claim], where [the contradiction follows] of necessity from the
things which are granted [by the opponent] with the point at issue not being
included, (and follows] in the very same respect, relation, manner, and tense
(as in the opponent's claim]. (167''23-7)
This makes it clear that the claim which Aristotle's opponent makes
is precisely the thing contradicted by Aristotle's conclusion. Aristotle's
main conclusion is: It is not possible for the same thing at the same
time to be a man and not to be a man (1006b33-4). So the opponent's
position will be precisely the claim that just this, expressed in just
these words, is possible. Aristotle's elenchus will start, then, from the
assumption that just this claim is to be subjected to examination. At
rno6b20-2 Aristotle says himself that the dispute with his opponent is
precisely over the question 'whether it is possible for the same thing at
the same time to be and not to be a man'. This rules out, for example,
the proposal offered by many interpreters, that the opponent's thesis
employs the name 'man' in the subject position and is of the form
'man is F and not F'. The name 'man' only occurs as a predicate in the
opponent's thesis and the opponent's thesis, in fact, is not a subject-
predicate proposition. Its form is 'it is possible for the same thing at
the same time to be F and not F'. 11
argument. So we must be careful not to import into our discussion of the elenctic
demonstration, features of Aristotle's later argument where he gives no indication
that he is still engaged in peirastic argument.
11 This means that when Aristotle goes on to insist that his opponent signify
something by his thesis, this cannot be a requirement which has to do with what
the subject term of his thesis designates, as many have supposed. It also means
that the requirements which Aristotle introduces in the De lnterpretatione for
certain strict subject-predicate propositions will not be directly relevant, as some
have also supposed, to understanding his discussion of his opponent's position.
See e.g. J. Lear (r980: 104ff.).
332 Robert Bolton
VI. THE REQUIREMENT rt ;.tyB1v
The first requirement which Aristotle mentions in beginning to con-

struct his elenctic demonstration is that:
The opponent shall simply Tl Jim; if he ).tvn nothing then it is absurd to

look for a h11·0.;- with someone who exovra }.or·ov about nothing, in so far
as he does not f:xc1 [J.ayov]. For such a person, as such, is already like a
vegetable ... The starting-point in all such cases is ... the request [for the
opponent] that he should affirm at least that he rrr;µaiw;1v n, both for himself
and for another, since this is necessary if in fact he },tyo1 Tt. Otherwise there
can..be no h:lyo;; for such a person, neither with himself nor with another.
(I006aI2-24)
Interpreters standardly suppose that Aristotle's first request as

questioner here, that his opponent J.tm u, is the request that the
opponent 'say something' or 'speak about something' (e.g. some real
object of scientific study) which is designated by the subject term of
his thesis. He dismisses the opponent who will not do this in the
proper way, which includes 'signifying something', as someone who
has no capacity for rational thought or discourse (}coyo;). However,
these translations do not fit easily with Aristotle's descriptions of
dialectical and peirastic argument. The request 'say something' or
'speak about some real kind of thing' is not a proper dialectical or
peirastic question to begin discussion with (Int. II, 2ob22 ff.). The first
requirement for an opponent or answerer in a dialectical or peirastic
discussion (after taking up a position, which Aristotle's opponent has
already done here) is 'to either grant or not grant what has been
asked' (Top. VIII. 5, 159b2-4). That is, the answerer must }J;ye1 u in
the sense that he must 'affirm something' by saying definitely yes or
definitely no to questions and in that way IJyr;1 u.
That this dialectical requirement is what Aristotle has in mind in the
elenctic demonstration seems clear from a passage later in r. 4 where
he describes the position of an answerer who will only say 'yes and no'
or 'not yes and not no' in response to questions but never simply 'yes'
or simply 'no'. Aristotle says about the person who takes such a
position that ov(}f;v U1 s1 ( 1008a3 I), using the very same language he
1
used earlier; and, just as in the earlier case, he goes on to characterize

the person who in this way does not definitely say 'yes' or 'no' to
anything (µr;(}i;v 6no}.aµfiaw;1) is no different from 'the vegetables'
(roo8b10-r2). In this case, however, the person who o!lOh· }Jye1 is not
the person who says nothing; he may well say many things of the form
'X is F and not F'. Nor is he the person who does not refer to
anything real. If he says 'man is vile and not vile', but refuses to
definitely say 'yes' or 'no' in answer to questions, he still refers to a
real kind. What he does not do is to definitely concede anything, and
that is why he oveev )iyr,1. Aristotle mentions this here, presumably,
because his opponent, as a Heraclitean who believes in the unity of
opposites (rno5b25), might well be one who is tempted to answer 'yes
and no' to questions.
If we understand what Aristotle means by the requirement that the
answerer Jiyr,1 ri, namely that he answer questions with a definite 'yes'
or 'no', how does the person who refuses to do this thereby fail to
exuv J.6yov or, as Aristotle later puts it, to 61w11tvm J.6yov ( 1006a26).
It does not seem plausible to suppose, as is usually claimed, that such
a person has, by this refusal, abandoned the capacity for reasoning or
for all rational thought and discourse. For instance, a person could
engage well enough in demonstrative reasoning, and express this in
discourse since this does not require one to answer questions at all. So
what does Aristotle mean by the failure to exr,zv or 6noµevr,1v ).6yov.
Again, Aristotle uses just this sort of language in his discussions of
dialectical procedure as a semi-technical description for the role of the
answerer. The answerer's role is often summarily described as 6ntxw·
).6yov, 'to submit to argument'. This the answerer does by saying
definitely 'yes' or 'no', in the proper manner, in answer to questions,
and, thereby, allowing the questioner to construct a J.6yoc;, that is an
argument, to test the answerer's claim (Top. VIII. 5, r59a38ff.). So
the opponent who does not excl\! or 6noµtvuv J.6yov, who does not
co-operate by answering in the way minimally required to allow an
argument against his position to be constructed, is the opponent who
simply will not submit to question and answer examination. This would
perhaps be reprehensible for Aristotle's opponent in Met. r, and
make him like a vegetable, but this opponent need not, thereby, have
abandoned all reason and coherent thought.
This is an indication for us already of what Aristotle's general
strategy is in his elentic demonstration. He wants to refute his op-
ponent's claim that it is possible for the same thing to be and not be a
man on the simple condition that his opponent will submit to the
minimal requirements for dialectical and, in particular, peirastic argu-
ment. His opponent, by his denial of the principle of non-contradiction
on which all argument is based, may 'subvert argument but [by an-
swering properly] he submits to [peirastic] argument' (1006 3 26). Since
peirastic is, for Aristotle, the minimal adequate procedure to use for
genuine refutation, where refutation by demonstration is impossible,
this means that his strategy is to refute the opponent, and thereby
establish the principle of non-contradiction, on the basis of the
minimal procedures which anyone must follow who is prepared to
334 Robert Bolton
open himself to genuine refutation. (See Soph. El. 9.) This does not
mean, however, that the argument itself is a merely hypothetical, or
transcendental, argument. The argument should be a peirastic argu-
ment and, thus, only use premisses which are independently known by
everyone, and in particular by the opponent. As such, it will be a
non-hypothetical argument.
.., VII. THE REQUIREMENT NOT TO BEG

THE QUESTION
There is further indication that this is Aristotle's general strategy in

the first moves which he makes to develop his requirement that the
opponent must 'affirm something'. The opponent has taken up the
position that it is possible for the same thing at the same time to be
and not be a man. Our first question, Aristotle says, cannot be 'the
request that our opponent shall affirm (}ciyezv) that something is or that
something is not. For this right away some [opponent) could take to
beg the question' ( 1006a20- 1). 12 Concern about begging the question
is, of course, another matter which is quite prominent in the Sophistical
Refutations. It is one of the classical fallacies discussed there in
Aristotle's account of the theory of elenchus. One of the main ways in
12 What argument in particular does Aristotle forgo here to avoid begging the
question? His thought is illuminated by the parallel discussion later in r. 4. There

he considers the following strategy for objecting to the claim that e.g. the same
thing is a man and not a man. One may, he says, invoke the true principle of
detachment that 'X is F and G implies X is F'; and infer thereby that the thing in
question is a man. Then one could invoke the true principle that 'if Xis F then it
is not the case that X is not-F' and conclude that nothing is both a man and not a
man. This argument, Aristotle says, 'they perhaps could say begs the point at
issue' (rno8°1-2). Why could an opponent claim, with some justification, that this
line of argument begs the question? The argument does not actually use the
principle that 'nothing is a man and not a man' or 'nothing is both F and not-F'.
But it does use the principle that 'if something is F then it is not the case that it is
not-F', and the opponent could reasonably argue that this is insufficiently different
from the point to be proved to be used as a premiss without begging the question.
In Top. VIII. 13 Aristotle counts it as a case of begging the question in dialectic
if something immediately convertible with the point to be proved is used to prove
it (163"rn-13; cf. An. Pr. II. 16). This seems to be the strategy which Aristotle
declines to follow in the elenctic demonstration. His opponent has taken the
position that 'it is possible for the same thing to be a man and not to be a man'.
But Aristotle will not proceed to in effect invoke the rule of detachment and ask
that he answer 'yes' or 'no' to the conjuncts and follow the strategy later described
because the opponent can claim that he has not been refuted if the question
is begged.
which an elenchus can misfire-and turn out to be only a sophistical
elenchus-is if the question is begged. This is already explicit in the
definition of elenchus to which we have referred. Since Aristotle
wants to demonstrate his result by elenchus, one of the minimal
conditions for success is that he not beg the question. One might want
to argue, as some have, that Aristotle's elenctic proof of the principle
of non-contradiction must beg the question since no proof of this most
fundamental principle could fail to do this. This is acceptable, some
would further argue, so long as Aristotle's opponent accepts the pre-
misses of the proof since, in that case, he begs the question against him-
self. But this is not acceptable on Aristotle's theory of elenchus. In
any peirastic elenchus the opponent accepts the premisses, but this does
not thereby make begging the question acceptable in peirastic elenchus.
By the very definition of elenchus which Aristotle gives, no elenchus
is genuine which begs the question (167a25-6; cf. 181a16-19).13
As Aristotle says in Soph. El. 2T
If it is clear [that some concession would beg the question] it must not be
granted, not even if it is ey<5o¢ov and expresses the truth. If this should be
[at first] not noticed then since such arguments [those that beg the question]
are bad arguments, one should charge the error [in not noticing] back to the
questioner on the ground that he has not followed the rules of discussion (ml
<>1e1J.eyµi:vov). For an elenchus does not beg the question. (181a16-19)
VIII. THE REQUIREMENT TO SIGNIFY ONE THING
The primary move which Aristotle next makes to further develop the
strategy which he will follow, to avoid begging the question, is to
request that the opponent concede that he signifies something distinct
which involves more specifically, as Aristotle later indicates, that he
signifies one thing. Aristotle adds that this is a requirement for .A6yo;
(roo6a21-5, b11-13).
What is involved in the opponent's concession that his claim signifies
something distinct or signifies one thing? This is another requirement
which has been understood in many different ways in the literature. It
is also, again, a requirement discussed at length in the account of
elenchus in the Sophistical Refutations, particularly in chapter 4. Just
13 Irwin, p. 185, seems to interpret Aristotle as saying, at 10o6a15-18, that

begging the question is acceptable since 'the respondent is responsible for such a
thing' (818). But the 'such a thing' which the opponent is responsible for is not
begging the question but lxm },oyov, i.e. for sustaining the argument by agreeing
to its premisses (3 14).
Robert Bolton
as begging the question must be avoided to avoid fallacy in refutation

so the failure to signify something distinct or to signify one thing leads
to fallacy, what we often call the fallacy of ambiguity or equivocation.
For Aristotle this covers a whole bundle of types of fallacy concerned
with language including those based on the simple ambiguity of words,
on the. ambiguity of words in combination, on syntactic sentential
ambiguity, on the ambiguity of inflections, of accents and other things
( 16911 24 ff.). All of these involve cases of multiple signification (<Jf/JWiw;1
nh;iw, 166a16), or ambiguity (aµ</J1Po},fa), or equivocation (bµwvvµia,
166a22), or the failure to signify one thing (tv <Jf/µaiw:1, 166a2-3) or to
signify the same thing by some expression ( ra1Jro <Jf/µaivet, 166a25).
This is directly reflected in the definition of elenchus quoted earlier,
where it is specified that the contradictory of the opponent's position
must be deduced 'not in name merely (i.e. by virtue of an equivo-
cation] but in fact'. In general, as Aristotle later sums things up in
Soph. El. 6:
It is standard [in a refutation] for everything [sentences, words, phrases,
syntactic forms, inflections, accent patterns, etc.] to signify some distinct
thing (rb& Tl <J17µaivc1). (168 3 25-6)
Where this is not the case fallacies arise. So the requirement that
the opponent signify something or, more explicitly, signify some dis-
tinct thing, introduces another of the conditions that the opponent
must be willing to accept to be open to genuine refutation of his
position. If he doesn't accept this he can still engage in thought and
reasoning; in particular he can still engage in fallacious reasoning. But
if he does this, or is not willing to accept requirements which are
essential to rule out this kind of reasoning against his position, then
he does not permit genuine examination of his position, since that
requires that one follow rules of discussion which do not leave the
door open to fallacious refutation.
The requirement for an elenchus to signify one thing and, thus, not
to employ terms equivocally, so as to permit refutation of an opponent
not merely in name but in fact is directly reflected in Aristotle's
discussion in r. 4. He restates the import of the requirement that the
term 'man' signify one thing in the following way:
It will not be possible for the same thing to be and not to be [a man] except
by virtue of an equivocation (bµwvvµia), just as if what we describe as a man
others were to describe as not a man. But this is not the problem in question,
whether it is possible for the same thing at the same time to be and not to be
a man in name, but rather in fact. (rno6b18-22)
Here Aristotle uses just the language which, as we have seen, he uses
in the discussion in the Sophistical Refutations of the requirement that
names, etc., in an elenchus must signify one thing.
A number of recent writers have, however, taken a quite different
view of Aristotle's requirement that his opponent agree that he sig-
nifies some distinct thing. They have claimed that, given Aristotle's
theory of signification, this is meant to require the opponent to agree
that he is talking about some real kind of thing with a real essence
which is a proper object of scientific study. There are a number of
questions which one might raise for this view. For one thing, if this is
what Aristotle requires then he is not using a strategy which can
establish that the principle of non-contradiction covers claims about
things which he and his opponent might agree do not belong to real
kinds, such things as artefacts (e.g. coats), accidental beings (e.g.
musical grammarians), and things that are not (e.g. goatstags). But
Aristotle is explicitly trying here to deal with the principle of non-
contradiction as it applies to being qua being, that is to whatever is
in any of the ways in which things are ( rno5b9- I I). One of the ways in
which things are, is as accidental beings (E. 2). Another of the ways in
which things are is as things which are not (I'. 2, 1003b9-10). Aristotle
makes it quite clear that the study of being qua being will cover these
things because these are among the things that are in one of the ways
in which things are. This means that the principle of non-contradiction
needs to be established as a principle that covers artefacts and non-
entities as well as things which are as real objects of scientific inquiry.
Another difficulty for this proposal is that in the course of his
discussion in r. 4 itself Aristotle applies the requirement to signify
something distinct not only to the term 'man' but also to the terms
'not to be', 'not-man', 'being one', and 'coat' (1006a29-30, b13-15,
22-7). It is doubtful whether the term 'being one' signifies some real
kind with a real essence for Aristotle (r. 2, 1003b22ff.), and it is quite
clear that the term 'coat' and the terms 'not to be' and 'not-man', do
not signify any real kind with any real essence (see Z. 4). Aristotle
also supposes that the situation where the names which the opponent
uses do not signify one thing, or even any definite range of things
which can be distinguished, is one where the opponent does not even
think anything (1006a34-bII). This implication reflects the doctrine of
the De lnterpretatione according to which the primary signification of
a name is a thought ( 16"3-9). Aristotle also introduces the notion that
expressions 'signify some distinct thing' or 'signify one thing' in the De
lnterpretatione (16a17, b20, 18a17, 19b9). There are, however, many
names, sentences, etc., mentioned in the De Interpretatione which
Robert Bolton
signify something and signify one thing by signifying one thought, for
instance 'goatstag', 'Kallippus', and 'pirate-boat', which do not signify
any real kind (16 3 17, 19-26). The failure to signify some real kind
does nq,t involve the failure to think anything on Aristotle's view. So
when he speaks here of a failure to signify something which does
involve a failure to think, the type of signification which he has in
mind cannot be the signification of real kinds. It must rather be the
signification of what is thought, which may or may not be accompanied
by the signification of real kinds. 14
But the most serious objection to this account of signification is that
in the Sophistical Refutations where Aristotle explicitly describes how
the requirement to 'signify one thing' operates in an elenchus produced
through TiJ Jw),eyw()ai he does not mean this. This requirement applies
there to any elenchus, including any peirastic elenchus, whether the
elenchus aims to deal with or refute claims about real kinds or not.
Surely what Aristotle means by the requirement to signify something
in his elenchus in Metaphysics r is explained by his discussion of what
that requirement means in an elenchus. The result of not signifying
one thing, Aristotle explicitly says in r. 4, is that one rules out To
<>mUyr.aeai (cf. roo6b7-9). If so then whenever one does not rule out
TO Jzal.Byr.aem on any topic at all one must signify one thing. If so, then
the requirement will not preclude the possibility that Aristotle's proof
procedure will be applicable to any sample instance of the denial of
the principle of non-contradiction, even where the terms do not
designate kinds with real essences, as it should be given the range of
applicability of the principle as a principle concerning being qua
being. Of course it may be that some further feature of Aristotle's
elenctic demonstration rules out the general applicability of his pro-
f:edure. Some have argued that there is just such a feature so we will
11eed to explore the further premisses in the argument to see whether
this is so.
14 Aristotle does suppose in the De lnterpretatione that a single term equivalent

to 'man-and-horse' does not signify anything ( 18a19-26) and that 'not-man' sig-
µifies one thing only 'in a way' (19b8-9), evidently because there is no actual thing
or no clearly demarcated class denoted by these terms. This reflects Aristotle's
doctrine that there are two modes of signification for words. Words signify
thoughts, primarily, but also things (16"3-8). Here he denies that the expressions
in question are significant in the second mode, just as he says that expressions
such as 'runs' and 'is' do not by themselves signify any thing (npc1ypa) though they
do still signify thoughts (16b19-25). But signification in the second mode is not
restricted to real kinds any more than signification in the first mode is. Aristotle
expresses no reservation as to whether 'pirate-boat' or 'Philo' signify in the second
mode ( 16"19-29, 32 ff.), when these examples are discussed in the same context
as 'not-man' ("29-32). For further discussion of these matters see Bolton (1984).
IX. ARISTOTLE'S CONSTRUCTION OF

HIS ELENCHUS
We are in a position now to set down the initial premiss of Aristotle's

elenchus. This is what the opponent agrees to in conceding that he
signifies something (1006a21-4). As we have noted, it has been stan-
dard to understand this as the concession that the subject term in the
opponent's claim (taken to be 'man') signifies something. But the
opponent's claim has no subject term. The only item introduced into
the discussion so far is the opponent's whole thesis (1006a1-2; cf.
b20-2). It is required, as we have seen from the Sophistical Refutations,
that the whole claim, along with all its parts, signify something dis-
tinct. Since no part is particularly in focus at this point, it is most
natural to take the concession by the opponent to be that his whole
claim signifies something distinct. That is, the first actual premiss in
the elenchus is this:
(I) 'It is possible for the same thing at the same time to be a man
and not to be a man' signifies something distinct (1006ar8-25).
Aristotle's next move, after securing his opponent's agreement that
his sample denial of the principle of non-contradiction signifies some
distinct thing is introduced as follows:
First, then, it is clear that this, at least, is true, that the name 'to be', or 'not
to be' signifies a particular thing. ( 1006a28-30)
Here Aristotle simply brings out one of the things involved in his
opponent's concession that his thesis signifies something definite. As
we have seen from the Sophistical Refutations, the requirement that
the opponent's concessions in an elenchus each signify something
involves the necessity that every signifying part of what he concedes
signifies some definite thing. This includes, in the case at hand, the
expressions 'to be' and 'not to be'. 15 So the opponent concedes a
second premiss.
(2) The expressions 'to be' and 'not to be' each signify a particular
thing ( 1006a28-30).
15 I omit full discussion of 1006"26-31. The text is uncertain and the material
introduced is not needed for the main argument. It reflects the doctrine of the
Sophistical Refutations that every signifying part of what is conceded in an
elenchus must u1Jµaive1 rb& r1 (168•25-6); and thus it simply prepares the way for
the more important claim that 'man' signifies one thing. As we shall see more
clearly below, in a dialectical argument premisses are often introduced for orna-
ment or other purposes which do not turn out to figure directly in the later
argument (Top. VIII. r, r57''1-13, cf. VIII. 6, 160"1-2).
340 Robert Bolton
The main point of this move, by Aristotle, is to focus attention on

the parts of his opponent's claim and the need for each of them to
signify something in order to prepare the way for the next concession,
which is introduced as follows:
If the name 'man' signifies one thing, let this be two-footed animal. I mean
this by signifying one thing: Assuming man is this [namely, two-footed animal],
if anything should be a man, being for a man will be this. (1006"31-4)
Here Aristotle extracts two further concessions from his opponent

which constitute the third and fourth premisses of his elenchus:
(3) The name 'man', as used in (1), signifies one thing which we
will take to be two-footed animal (1006a31-2).
(4) So, if anything should be a man, being for a man will be this,
namely two-footed animal (1006"32-4).
The third premiss merely requires the opponent to agree that just as
his whole thesis, and its parts 'to be' and 'not to be' signify something
distinct, as he agreed in premisses (1) and (2), equally the term 'man'
signifies one thing, that is one and the same thing in every occurrence,
including its occurrence in 'not a man'. This, again, reflects the
minimal requirement for an elenctic premiss set out in the Sophistical
Refutations. Aristotle further specifies what the one thing signified by
the term 'man' in each occurrence is, namely two-footed animal. It
does not matter for his purposes here exactly what the term is under-
stood to signify, he merely picks something which is plausible enough
as an ordinary definition (l,oyoc;, 1006113) of what the term signifies.
This does not mean that it is merely arbitrary what the term signifies,
however, since it does matter that the term not signify the same thing
as the term 'not-man'. So Aristotle goes on immediately to counter or
pre-empt this move by his opponent (100611 13-28). This is easy enough
to do because, as he points out, the dialectical problem (ro anopovwvov)
which his opponent raised in his thesis was not whether the expres-
sions 'man' and 'not-man' could be truly applied to the same thing, for
instance when the two are conventionally used to signify the same
thing. That is obviously possible and uncontroversial and thus raises
no dialectical problem. What the opponent meant to defend was the
controversial and problematic claim that the two terms 'man' and
'not-man' can be truly applied to the same thing when they signify two
quite different things (100611 20-8 with 1007a1-7). However, Aristotle
cannot, and does not, assume here, as a part of his elenchus that the
terms 'man' and 'not-man' must signify incompatible things. That
would open him up to the charge of begging the question, since that
is, in effect, just what he is supposed to prove against his opponent.
So he is careful not to request any such concession. He merely requests
that the opponent concede that 'man' and 'not-man', as used in his
own thesis, signify different things. That begs no questions and it helps
to guarantee that the term 'man' is not used equivocally in the argu-
ment, which it presumably would be if 'man' and 'not-man' signified
the same thing. So the opponent has to concede:
(5) The terms 'man' and 'not-man' signify different things and,
thus, so do the expressions 'being for a man' and 'not being for
a man' (which indicate what the two terms signify) (1006h13-28).
Like the second concession, this premiss plays no direct role in the
main argument, for reasons which we will explore shortiy, but serves
only to prepare the way for various steps in it.
The maiti argument continues by introducing an implication of
premiss (4) using different language. The opponent has conceded that
'man' signifies the one thing, two-footed animal, and that this means
that if anything should be a man, two-footed animal will be what being
for a man is. Aristotle recalls this and reformulates this using overtly
now in place of the subjunctive ('should be ... will be' in (4)) the
sentential necessity operator with the indicative.
(6) It is necessary that if anything is a man, it is a two-footed animal
(1006b28-30).
Aristotle justifies this by reference to the earlier concession in (3)
which is explicated in (4) (1006h30). His argument then moves swiftly
to its conclusion.
(7) 'Being necessary' (in (6)) signifies that it is impossible for the
thing not to be ( 1006b3 I -2).
(8) Therefore, it is not possible if anything is a man that the same
thing is at that time not a two-footed animal (1006b31-2).
(9) Therefore, it is not possible that the same thing at the same
time is a man and is not a man (1006b33-4). 16
16 This swift conclusion has surprised Aristotle's interpreters. Why does he wait
so long to do this? Why does he not introduce the material in premiss (4) in the
form in which it is expressed in premiss (6) and move right to his conclusion,
avoiding the special language in premiss (4) and the lengthy diversion between (4)
and (6)? Once again, Aristotle's discussion of dialectical practice suggests the
answers to these questions. He says in Top. VIII. I that it is important to stay
away from the precise language of the conclusion as long as possible and, rather,
to obtain concessions using other language which can then be converted directly
at the end of the argument into premisses which use the precise language of the
conclusion (155°29-32; cf. l56"27ff.). For in this case, Aristotle says: 'The an-
swerer will not foresee the basis on which the final conclusion will be reached since
the course of the previous reasonings has not been clearly articulated' (156"18- 19).
342 Robert Bolton
'X. DOES ARISTOTLE APPEAL TO HIS

ESSENTIALISM?
The main question of interpretation which faces us at this juncture

concerns Aristotle's use, in premisses (4) and (5), of the phrases
'being for a man' (ro Bfvm avepciJ1w?) and 'not being for a man' (ro
µ~ 1.dvm avepcvmµ) which indicate what the terms 'man' and 'not-
man' signify. Many recent interpreters have seen these phrases as
introducing main features of Aristotle's metaphysical essentialism.
What his opponent is taken to concede in premisses (3) and (4) is that
what the term 'man' applies to is essentially a two-footed animal. That
is, the thing the term 'man' applies to cannot fail to be a two-footed
animal so long as it exists.
So understood, the opponent's concession is, for instance, incom-
patible with views commonly accepted in antiquity about transmi-
gration and the afterlife. What is now a two-footed animal, on these
views, can become a four-footed animal such as a wolf, or can become
a demigod rather than an animal. There are indications that Heraclitus,
Aristotle's presumed opponent here, himself held such views and it is
hard to see how Aristotle could expect just any opponent of the
principle of non-contradiction to make concessions which immediately
rule them out. In any case, and more generally, it is hard to see how
any essentialist commitment like this could figure in a premiss in a
peirastic argument. Peirastic premisses must be things which the op-
ponent not only believes but knows and which everybody knows and
is ready to grant, even people without special scientific or philosophi-
cal training. The very idea that there are essential properties, which
specify what a given individual must be to be that very individual,
even if this can be philosophically defended, as Aristotle himself
thinks, is not something commonly known or readily granted by
everyone. 17
To further contribute to this goal, Aristotle says: 'You should prolong the discus-
sion and introduce points which are not of use for the argument ... For when
there are many things available it is unclear [until the end] on which sort of thing
the falsity [of the answerer's position] depends' (157•1-3).
17 Cf. Code (r986b: 346). Irwin (1988: 182), tries to avoid this criticism in the
following way: 'We need not commit ourselves to any strong doctrine of essential
properties here; we need not say that there is one property which must persist in
all change, or that the property can be described in purely qualitative terms;
perhaps some primitive ostention or unanalysable temporal continuity is all that is
needed.' There are two difficulties with this. ( 1) It does not describe Aristotle's
actual argument; he picks one very ordinary qualitatively describable property
which he expects his opponent to have no trouble with. (2) The further one goes
Fortunately, there is no need to read Aristotle as introducing such
essentialist claims. To begin with, as we have just seen, the premisses
which use the special language, for example 'being for a man', are not
the premisses on which the main argument depends. Aristotle wishes
to avoid the specific terminology of his main argument, particularly
the terms 'necessary', 'possible', and 'impossible'-as long as he can,
and he uses the special language because the concessions made in that
language can be easily converted into the concessions he needs even
though they might well seem to an opponent to be concerned with
other matters. The main concession which he needs is: (6) It is
necessary that if anything is a man it is a two-footed animal. This
claim straightforwardly employs 'necessary' as a propositional operator
and does not itself require any essentialist commitments. However, it
could still be claimed that the special language itself involves such
commitments. But we know from other passages in the Metaphysics
that it need not. Aristotle is quite happy to say that being for an F
applies to some F where this does not imply that being for an F
belongs to that thing essentially. For example, being for a pale thing,
or being for a large thing, belongs to pale things, or large things, but it
need not belong to them essentially (Z. 6, 1031b22-8). Similarly,
being for a pale man belongs to pale men, but it not only does
not belong to them essentially, it belongs to nothing essentially
(1031a19-28). So Aristotle is happy to use phrases of the form 'being
for an F' where this does not introduce any essential property of what
is an F. Nevertheless, Aristotle thinks that this sort of language, even
where it is not used to make essentialist claims, still supports claims
that certain propositions are necessary. In Parts of Animals II. 3
(649b21 ff.) Aristotle says that something can be (in) 'the being for F'
by virtue of what 'we signify by the name F' without this specifying
the real subject which some F is. Nevertheless, by virtue of what
'we signify', this being for F belongs to what is F intrinsically. He
uses blood and pale man as examples. A similar case is introduced in
the present context. A wine, Aristotle says in Met. r. 5, may taste
sweet at one time and not at another. But, given what being sweet is,
'necessarily whatever is going to be sweet is of this sort' (rnrnb25-6).
So whatever being sweet may be, necessarily if anything is sweet it is
of this sort, and this is true quite independently of whether the thing
in the directions Irwin suggests (e.g. so that the essential property is some ordered
set of temporally successive properties or some non-qualitative demonstratively
identified property such as, say, 'being this'), the more one gets, once again, into
elaborate philosophical doctrines which not just any opponent can be expected to
accept or even understand.
344 Robert Bolton
which is sweet is so essentially. So all that Aristotle's opponent need
be conceding in agreeing that the name 'man' signifies two-footed
animal is that being a man for anything which is a man would always
be being a two-footed animal. This is all that premiss (4) need involve
and this is something quite innocuous which any ordinary opponent
could be expected to grant and, indeed, to know if that opponent used
the term 'man' in the ordinary way which Aristotle takes for granted
here. 18
XI. THE MAIN FEATURES OF ARISTOTLE'S

ELENCHUS
We can now focus on the elenctic demonstration as a whole and

consider its character.
(r) 'It is possible for the same thing at the same time to be a man
and not to be a man' signifies something distinct (1006a18-25).
(2) The expressions 'to be' and 'not to be' in (1), each signify a
particular thing ( 1006a28-30).
(3) The name 'man' as used in (r), signifies one thing which we will
take to be: two-footed animal (1006a31-2).
(4) So, if anything should be a man, being for a man will be this,
namely two-footed animal (1006a32-4).
(5) The terms 'man' and 'not-man' signify different things and,
thus, so do the expressions 'being for a man' and 'not being for
a man' (which indicate what the two terms signify) (1006°13-28).
(6) It is necessary that if anything is a man, it is a two-footed animal
(1006°28-30).
18 Of course, Aristotle does himself believe that being for a man is the essence
of whatever it belongs to, because he believes that man is a substance and this just
means to him that being for a man is the essence of whatever it belongs to. This is
a point which he introduces himself later in r. 4 when he goes on to argue that
those who deny. the principle of non-contradiction do away with substance
(rno7 11 2off.). He relies there on the claim that what 'man' signifies, e.g. two-
footed animal, is in fact the substance and thus the essence of whatever is a man
(1007a25-7). But this is a new claim, one which is not introduced in the elenctic
demonstration. It is introduced in a new argument designed not to establish the
principle itself on the basis of what the opponent will accept but to show that the
denial of the principle is incompatible with certain other things, such as certain
views about substance and essence, which Aristotle thinks he can defend but
which he does not ever say his opponent must accept or know. As we have seen,
the indications that Aristotle is engaged in peirastic dialectical argument, where
the premisses are accepted by and are the responsibility of his opponent, stop
nrior to thi~ :irir11mPnt
(7) 'Being necessary' (in (6)) signifies that it is impossible for the
thing not to be (1006b31-2).
(8) Therefore, it is not possible, if anything is a man, that the same
thing is at that time not a two-footed animal (1006b31-2).
(9) Therefore, it is not possible that the same thing at the same
time is a man and is not a man (1006h33-4).
The main thing to note about this argument is that it conforms to all
the requirements in the Sophistical Refutations for an elenchus. To
begin with, the premisses are agreed to by the opponent, all of the
signifying elements signify one thing (i.e. there is no kind of equivo-
cation), and, most importantly, the argument does not beg the ques-
tion. It uses neither the principle of non-contradiction nor anything
directly convertible into it as a premiss.
It will be useful to contrast this account of the argument with
various other accounts offered recently in the literature to see how
well, on those other accounts, Aristotle's argument conforms to these
requirements for an elenchus. Here is one account of the crucial point
in Aristotle's proof by refutation:
An assertion divides up the world: to assert that anything is the case one must
exclude other possibilities. This exclusion is just what fails to occur in the
absence of the law of non-contradiction ... One cannot assert P and then
directly proceed to assert - P: one does not succeed in making a second
assertion but only in cancelling the first assertion. This is the ultimate reason
why an opponent of the law of non-contradiction cannot say anything. 19
Whatever the merits of this argument may be on its own, as an
argument against Aristotle's opponent it clearly begs the question. It
assumes that P and -P cancel each other, that they are incompatible.
But this is just what Aristotle needs to prove, since it is just what the
opponent denies. Since Aristotle definitely wants a non-question-
begging refutation, this cannot be Aristotle's argument unless he is
terribly confused. It is not sufficient to respond to this that Aristotle is
not interested here in an argument which convinces his opponent, but
only an argument which makes clear to us what the opponent's mistake
is, and the above account of the argument does this. Aristotle points
out himself that his argument will not convince certain people, for
instance those who believe that the only convincing argument is a
strict demonstration (1011a3-16). But he is not so much interested in
an argument which convinces his opponent as in an argument which
refutes his opponent, by the proper standards for a refutation. An
argument which begs the question does not do this.
19 Lear 1980: 112; Irwin (1988: 181), similarly, makes it a part of Aristotle's
~- - . - -- - -
Robert Bolton
Another common assessment of the main move in the argument

goes as follows: Aristotle's opponent acknowledges that his subject is
essentially a two-footed animal. It follows directly from this that a
subject that is not a two-footed animal is not the same subject. If the
opponent insists that it is the same subject because he is not denying
that the subject is a two-footed animal, he is only saying that this
subject is also not a two-footed animal, this is Aristotle's reply:
[The opponent] can hardly maintain this with the standard conception of
negation which we need to use to understand [the opponent's] thesis as he
intends it to be understood in the first place (in asserting that man is not a
[two-footed animal]). 20
If Aristotle is relying on 'the standard conception of negation', that
is on the notion that predicates such as 'man' and 'not-man' are
incompatible, then he quite blatantly begs the question. That is just
what he is supposed to prove. But Aristotle simply does not accept an
argument which begs the question as a refutation. This shows that on
the essentialist reading the argument is a non sequitur unless it begs
the question. In addition, it is worth noting that if Aristotle is relying
on the 'standard conception of negation', and needs to rely on it, then
all of the essentialist machinery which he is presumed to need is
beside the point. If Aristotle can assume straight away that predicates
such as 'man' and 'not-man' are incompatible, by the standard con-
ception of negation, then he can easily refute his opponent whether
man is an essential property or not
By contrast, then, with these accounts of the elenctic demonstration,
the account offered above strictly conforms to Aristotle's require-
ments for an elenchus mentioned above. Does it also conform to
Aristotle's other requirements for the premisses of a peirastic argu-
ment? For this, the premisses must be things the opponent actually
believes and knows, and also matters of common knowledge, and it is
plausible enough that they are. The opponent's thesis, that the same
thing can both be a man and not a man, is certainly not a matter of
the opponent's knowledge or of common knowledge; but that this
claim is significant, plausibly enough, is. That the term 'man' signifies
some one real essence may not be a matter of the opponent's belief or
knowledge, or of common knowledge; but, that the term 'man' is in
fact used non-ambiguously, in utterances like the opponent's thesis
and concessions, is. That anything which is a man is essentially a two-
footed animal (on the understanding that two-footed animal is what
'man' signifies) is hardly a matter of any opponent's belief or knowl-
20 Irwin 1988: 182 n. 9.

edge; but that (on this understanding again), if anything is a man then
necessarily (while it is a man) it is a two-footed animal, plausibly
enough, is. Aristotle has often been criticized for using in his proof
controversial essentialist premisses; on this account of his proof these
criticisms are misdirected. Finally, it is, plausibly enough, a matter of
common belief and knowledge that if something is necessarily the case
then it is not possible for it not to be the case. These premisses ( ( 1),
(3), (4), and (7)) are the main premisses of Aristotle's proof.
XII. ARE ARISTOTLE'S PREMISSES

PRESUPPOSITIONS OF COHERENT THOUGHT?
If these premisses are matters of common knowledge, are they also

something more? It has been standard, as we have noted, to suppose
that in his proof here Aristotle draws on very special premisses,
premisses which no one could give up without abandoning necessary
conditions for significant thought or discourse. That has been taken to
make the argument a kind of transcendental a priori proof. However,
if one were to give up premiss ( 1) and hold, for instance, that denials
of the principle of non-contradiction are not significant, that might
have very little impact on one's intellectual life. Or, to deny premiss
(3) and suppose that 'man' does not signify one thing, as used in
various typical utterances, might be wrong but its acceptance is not
obviously a presupposition of all coherent thought or discourse. After
all it may well be that we do unawares use some names equivocally
which turn out on inspection not to signify one thing. Again, to deny
premiss (7) and to give up the equivalence of the terms 'necessary'
and 'not possibly not' or to deny premiss (6) by supposing, say, that
all uses of the term 'necessary' are incoherent, would require one to
give up a certain amount of modal talk but not to give up all intelligent
thought and discourse. So these actual premisses of Aristotle's proof,
even if they are true and commonly believed and known, do not seem
at all to have this special character. Nor does Aristotle ever claim that
they do.
The one premiss by reflection on which Aristotle is led to make
claims which tend in this direction is premiss (3) 'man' signifies one
thing. Aristotle does not make the implausible claim that this premiss,
the one actually used in the argument, is one no one could intelligibly
give up, though he certainly believes it is true and a matter of common
belief and knowledge that in typical claims, including the opponent's
thesis and concessions, the term is used to signify one thing. What he
says is that if the term 'man', or any other term, does not signify one
Robert Bolton
thing (or at least a finitely distinguishable set of things) then it cannot

be used to express any thought in any utterance where it is used or,
thus, to express any claim used in question and answer elenctic ex-
amination or other reasoning ( 1006a34-bII). This passage is obviously
interesting and important, and it is easy enough to see why it has
so attracted the attention of interpreters that, in the absence of
any comparison with Aristotle's theory of elenchus in the Sophistical
Refutations, it could be taken to reveal the heart of Aristotle's
elenchus. To understand this passage properly, however, we need to
keep several things in mind. First, the claim that one of the premisses
(3) or part of it-that 'man' signifies one thing-is a requirement for
the opponent's thesis to express a discussable thought, is not extended
by Aristotle to cover all the other premisses. He does not say that it is
a requirement for thought or discourse, or for thought or discourse
about certain subjects, that semantic truths (as in (3) ), generate
necessary truths (as in (6) ), though he takes it that an ordinary
opponent will agree to this implication. Nor does he say that it is a
requirement for thought that in (7), the term 'necessary' signifies not
possibly not, or that it signifies anything at all; though, again, it pretty
obviously does ordinarily signify this. So the claim which Aristotle
makes about (3) is not one which he extends, or which could easily be
extended, to all the premisses of this argument. More importantly,
even if it were possible to extend this claim or some similar claim to
cover all of the premisses, that would not make any such claim a part
of the argument. The claim that it is necessary to signify one thing by
what one says if one is to express a thought is not a premiss of Aristotle's
elenchus since it is not anything the opponent takes responsibility for.
This point deserves further elaboration.
There has been a general tendency in the interpretation of Aristotle's
elenchus to suppose that what he is attempting to do is to isolate what
his opponent 'must presuppose' in order to coherently state his position
and to refute him, that is, to construct the elenchus, using these
necessary presuppositions as premisses. 21 But this badly misconstrues
Aristotle's theory of elenchus. An elenchus is based not on the necess-
ary presuppositions of an opponent's thesis, or of its intelligibility or
the like, but only on what is explicitly granted, and actually believed
and known, by the opponent. Thus when Aristotle does reflect on a
presupposition of one of his opponent's concessions and claims that it
must be true if the opponent is to express a thought by his thesis, this
does not mean that, nor does Aristotle indicate that, this presuppo-
sition is a part of or a premiss for the elenchus.
21 Irwin (1988: 181) typifies this now standard approach.

So Aristotle is not introducing here a special form of dialectical
argument where the premisses are necessary presuppositions of what
people say, or of the intelligibility or discussability of what they say,
rather than what they explicitly grant and take responsibility for. This
is clear from the beginning of the argument (1006a18, 25-6). When
Aristotle says there that the starting-point of his elenchus is that his
opponent must at least signify something, by his thesis, if he is to
affirm anything (1006a20-2), he means, as he goes on to point out,
that the claim that he signifies something by his thesis is something
that the opponent is asked to grant ( 1006a24). 22 It enters the argument
only on that basis. A dialectical peirastic elenchus establishes its
conclusion only on the basis of what the opponent grants, believes,
and knows and any additional presuppositions might well satisfy none
of these requirements. The opponent in Aristotle's elenchus here
believes and knows the premisses, and so do the rest of us. That
enables the argument, as a proof from premisses which are known on
the basis on which he and we know them, to genuinely establish its
conclusion. But neither the opponent, nor the rest of us, need know
any additional presuppositions, so they are not part of the epistemic
basis or evidence for what is established by the elenchus itself.
Of course, such additional presuppositions may still be facts, and
Aristotle may think that they show something important. Provided we
distinguish what is established by the elenchus from what is established
by reflection on the elenchus we can ask what it is that Aristotle
thinks is brought out by the latter. Perhaps there is a moral or deeper
lesson to Aristotle's argument, distinct from what the elenchus estab-
lishes, which indicates in a different way that conformity to the prin-
ciple of non-contradiction is a requirement for any significant thought
or discourse. Here is one account, from the literature, of how such a
deeper lesson is brought out by Aristotle in what the writer properly
calls Aristotle's meta-elenctic argument:
The meta-elenctic argument is a reflection on the elenctic argument and
establishes the following claim: if the opponent of the PNC is to signify
something then he will be committed to the PNC governing whatever he has
signified in that it would be impossible for the contradictory of what he has
signified to be true if what he has signified is true ... Reflection on the
elenctic argument ... is designed to show that the mere possibility of signifi-
cant thought and discourse requires adherence to the principle ... 23
22 This indicates that rr) d¢10Dv ... iiym (roo6a19) governs m71mive1v r1 ("21).
'The starting-point ... is the request that he should affirm ... that he signifies
something [by his thesis], both for himself and another, since this is necessary
[that he should agree to this] if he is to affirm [i.e. agree to] anything.'
23 Code 1987a: 145.
350 Robert Bolton
One part of this account seems in a way correct. Aristotle does try
to show, and to show by the elenchus, that 'if the opponent of the PNC
is to signify something [by his thesis], then [given other premisses] he
will be committed to the PNC'. But Aristotle does not try to show
that if the opponent signifies something 'he*will be committed to the
PNC governing whatever he has signified in the sense that it would be
impossible for the contradictory of what he has signified to be true if
what he has signified is true'. One of the things that the opponent has
signified, by his very thesis, is the denial of the principle of non-
contradiction. But Aristotle does not try to show that if this denial is
true then it is impossible for the contradictory of this denial (i.e. the
principle of non-contradiction itself) to be true. If anything he shows
that if this denial which is the opponent's position is true, as the
opponent claims, and thus significant, as the opponent concedes, then
it is necessary for its contradictory, namely the principle itself, to be
true. 24 The upshot of this point can be put more simply. Aristotle
cannot be trying to show that whatever is significant must conform to
the principle of non-contradiction because one of the things that he
takes to be ~ifnificant is the very denial of the principle of non-
con tradicti on. 2
This shows clearly that Aristotle cannot be trying to establish that
conformity or adherence to the principle of non-contradiction is a
necessary condition for any significant thought or speech. Those who
have interpreted Aristotle in this way often go on to criticize him
for making this claim. 26 They introduce quotations, from Hegel and
Lenin, for instance, in which various counter-instances to the principle
of non-contradiction are asserted. They then say: 'These claims may
24 Of course, if the principle of non-contradiction is (proved) true then if the

denial of the principle is true it will be impossible for the denial of the denial (i.e.
the principle) also to be true. This may create some trouble for Aristotle's
opponent if he wants to affirm the principle as well as deny it. However, this is
not the trouble Aristotle creates for the opponent in his elenchus. The trouble
created there does not arise from the fact that the opponent cannot affirm the
principle (if he significantly denies it) but that he must affirm it if he significantly
denies it. For the purposes of this elenchus, the opponent whose position is simply
the denial of the principle (in a sample instance) takes no position on whether he
also wants to affirm the principle (in this sample instance).
25 Code further claims that if 'the mere possibility of significant thought and
discourse requires adherence to the principle' then 'if one did not believe that it
governed his utterances ... he would not be making significant statements' (p.
145). But this is a non sequitur. It does not follow from the supposition that one's
utterances must conform to the principle to be significant, that one must believe
they conform for them to be significant.
26 See Dancy (1975: 36, 142).
be absurd but they are significant. So Aristotle is wrong.' But Aristotle
does not say that these claims are not significant, he agrees that they
are significant and uses that very fact as a part of his proof that such
claims are false. So that is all that he is trying to show, not that if you
say things that don't conform to the principle of non-contradiction you
have no coherent thought, but rather that given the coherent thought
you do have, plus other matters of common knowledge, you are
wrong.
XIII. ELENCHUS AND SCIENTIFIC JUSTIFICATION
Aristotle's proof, then, is a peirastic proof as he led us to expect it

would be. It argues only from what is commonly known and accepted,
not, however, from premisses no one of which could rationally be
given up. Does Aristotle believe, nevertheless, that the proof which
he offers gives us by itself the proper basis or justification for genuine
scientific knowledge of the principle of non-contradiction? If he does
then his procedure for justification is quite different from the one laid
out in the Analytics. But can he believe that this argument provides
the scientifically adequate justification for the principle? He does not
say that it does, and there are many indications in Met. r that he
would not regard this argument as providing the appropriate scientific
justification. To begin with, in r. 2, as we have seen, Aristotle con-
trasts the peirastic power which dialectic uses on scientific subjects
with the scientific, the gnoristic, capacity of philosophy (1004h25-6).
The whole point of peirastic as Aristotle makes clear in the Sophistical
Refutations is to provide a procedure by which we can genuinely know
and establish things, by refutation, without having to know or establish
them scientifically ( 11, 172a21 ff.). So since the elenctic demonstration
is a peirastic argument it will not yield scientific knowledge. This is
reflected in various ways in r. 3-4. In r. 3 where he introduces his
discussion of the common axioms, Aristotle makes it clear that he still
thinks of such principles in the same way he does in the Analytics. He
describes them as 'the most firm principles of the subject' ( 1005h9- IO).
He explains this as meaning two things. First, these principles are not
hypothetical. They must be known (scientifically) in order to know
anything else, and thus they are not (scientifically) known only hypo-
thetically, that is, only if other things are known from which they are
(scientifically) deduced (1005h14-16). Secondly, they are the best
known things in the subject in the sense that they are already known
(scientifically) before other things are (b16- 17). Aristotle elaborates on
this in r. 4 when he points out that the principle of non-contradiction
in particular is more an indemonstrable basis for the scientific demon-
stration of other things than anything else ( 1006a10- 11 ). These re-
marks suggest, first, that Aristotle still thinks of a science, and of
metaphysics in particular, as having, or being expressible in, the same
kind of demonstrative explanatory structure as in the Analytics. They
also ascribe to the principles of such a science the same kind of
epistemological or cognitive status which is ascribed to such principles
in the Analytics, where the axioms are taken to be things one must
know to know anything (I. 2, 72a14- 18); to be non-hypothetical (I. 3,
72b13-25); to be known before and better than anything depen-
dent on them (I. 2, 71b26-35); and to be epistemically unshakeable
(72a37-b14). Given these requirements, from r. 3 and the Analytics,
it seems clear that Aristotle's proof by elenchus of the principle of
non-contradiction cannot by itself provide the justification or cogni-
tive basis required for scientific knowledge of the principle of non-
contradiction. This justification must establish the principle as the best
known thing of all and as something known but not known because
other things are known. The justification provided by the proof by
elenchus does depend on other things being known, and these things
are better known to us but scientifically or by nature less well kno'Yn
than the principle must be, according to Aristotle's own statement, so
they cannot by themselves yield by deduction the kind of knowledge
required for scientific knowledge of the principle.
This point will become clearer if we consider further the epistemic
status of the particular premisses of the elenchus. We have noted that
the premisses of an elenchus are matters of common knowledge. How
do they get to be matters of common knowledge? In particular how
does the crucial part of premiss (3), that the name 'man' signifies one
thing, get to be known by anyone or everyone? This question seems,
at least at a simple level, easy enough to answer. It is a claim about an
actual word of our language and whether an actual word of our
language (or Aristotle's language) does signify one thing is, as we
would put it, an empirical question. We know from experience that
the name 'man' signifies one thing. This simple matter of fact is,
however, a crucial premiss in Aristotle's proof. So what we know by
deduction from it cannot be more well known than it. Yet the principle
of non-contradiction when scientifically known must be better known
than it and better known than every other premiss of the deduction as
well. So this deduction cannot by itself give us scientific knowledge of
the principle.
Of course, it is important to remember that Aristotle's actual de-
duction only proves a particular sample instance of the principle of non-
contradiction, not the principle in full generality. Aristotle certainly
expects us to see, however, that this proof could be repeated for any
other expression which signifies one thing. This gives us, then, the
basis for a generalization to the principle in unrestricted form. But
this procedure for reaching the principle is inductive, it relies ultimately
for its force on our knowledge of certain particular matters of fact
known from experience which are required to provide the basis for
the inductive generalization.
If our knowledge of the principle of non-contradiction is reached in
this way, however, then it is reached in just the way the Analytics tells
us all scientific principles are reached. According to the Analytics, as
we have seen, we start from items of experience and we proceed from
them to general principles which cover them. Moreover, according to
the Analytics, the procedure which we use to reach general principles
is induction. Since we start this process from items of experience,
Aristotle says: 'It is necessary for us to come to know the primary
things [in a science] by induction' (An. Post. II. 19, 10ob4).
It is, of course, problematic how it is for Aristotle that by induction
we come to know things which are better known, in some appropriate
way, than the data on which the induction is based. But that is not a
problem we need to solve here. It is clear enough both in the Analytics
and in the Metaphysics that Aristotle believes we can somehow do this
and this gives rise to no special problem for the Metaphysics. In
Met. A. 1-2, it should be noted, Aristotle presents the very same
picture of how we acquire scientific knowledge as he does in the
crucial chapters of the Analytics to which we have referred. We start
from experience and by some sort of process of generalization from
this we come to know all scientific principles (981"5-7 with a24-7).
We know that Aristotle has the very same picture of science in mind
as in the Analytics, in Met. A, because he refers us there to the
Ethics for further details on what he means by scientific knowledge
(A. I, 981°25 ff.), and in the Ethics (VI. 3, 7-8) the account he offers
is the same as in the Analytics. In fact in the Ethics he refers us to the
Analytics for further details (1139°31-3). 27 In Met. A. 1-2 Aristotle
quite explicitly says that metaphysics, which he calls there aorpia,
with its own especially general principles, is a science in just this
sense (982"1-6). (This by itself indicates that Aristotle does not have
a new view of science in the Metaphysics.) This is confirmed in Ethics
VI. 8, where Aristotle also explicitly says that in metaphysics (aorpia)
27 The reason why in Met. A. T he refers to the Ethics rather than to the
Analytics is possibly because the Ethics discusses the full range of related cognitive
states which he has been discussing in the Metaphysics, including aorpia and ri:xv11
as well as l'nian/1117, and the Analytics does not. See 98 I h25-7.
354 Robert Bolton
just as in natural science 'the principles are reached from experience'
(l!~ r!1mB1p1a~, 1142a19; cf. An. Post. II. 19, 100"6-9). 28
If in the Metaphysics, both A and r, this is still the way all scientific
principles are reached, by induction from items of experience, then
one significant conclusion can be drawn. Our knowledge of the prin-
ciple of non-contradiction is not, for Aristotle, a priori. Or, it is no
more a priori than our knowledge of any other basic scientific principle
such as the principle of natural science that a lunar eclipse is a
blockage by the earth of the moon's light source. (See An. Post. II.
8.) We come to know all scientific principles, by induction, in just the
same way. It will be difficult, no doubt, for us to give up the feeling
that there is the material for some non-question begging a priori proof
of the principle of non-contradiction somewhere in Met. r. It will
perhaps be even more difficult for us to give up the feeling that such a
proof is somehow possible. But this is not, it seems, a feeling which
we share with Aristotle.
28 In Ethics VI. 7 (1141a12-20, "33-b3) Aristotle understands aocpia as a uni-

versal science as well as a science of the divine, just as in Met. A. 2 (982a8-10,
21-5; 983"5-II).
BIBLIOGRAPHY
AcKRlLL, J. L. (r963), Aristotle's Categories and De Interpretatione (Oxford).

- - (1965), 'Aristotle's Distinction between energeia and kinesis', in
Bambrough (1965), 121-41.
- - (1972-3), 'Aristotle's Definition of Psyche', Proceedings of the Aristo-
telian Society, 73: 119-33, repr. in Barnes, Schofield, and Sorabji (1979),
4: 65-75.
ALBRlITON, R. (1957), 'Forms of Particular Substances in Aristotle's Meta-
physics', Journal of Philosophy, 54: 699-708.
ALEXANDER (1891), Alexandri Aphrodesiensis in Aristotelis Metaphysica Com-
mentaria ( Commentaria in Aristotelem Graeca), i, ed. M. Hayduck (Berlin).
ANSCOMBE, G. E. M. (1953), 'The Principle of Individuation', Proceedings
of the Aristotelian Society, suppl. vol. 2T 83-96, repr. in Barnes, Schofield,
and Sorabji (1979), 3: 88-95.
ANTON, J., and PREUS, A. (1992) (eds.), Aristotle's Ontology (New York).
ARMSTRONG, D. M. (1978), Universals and Scientific Realism: Nominalism
and Realism, i (Cambridge).
- - (1991), 'Classes Are States of Affairs', Mind, 100: 189-200.
AscLEPlUS (1883), In Aristotelis Metaphysicorum Libros A-Z Commentaria
(Commentaria in Aristotelem Graeca), ii, ed. M. Hayduck (Berlin).
AUSTIN, J. L. (1961), Philosophical Papers (Oxford).
BALME, D. (1972), Aristotle's De Partibus Animalium I and De Generatione
Animalium I (Oxford).
- - (1980), 'Aristotle's Biology Was Not Essentialist', Archiv fur Geschichte
der Philosophic, 62: 1- r 2, repr. in Gotthelf and Lennox ( l 987), 291 -302.
BAMBROUGH, R. (1965) (ed.), New Essays on Plato and Aristotle (London).
BARNES, J. (1971-2), 'Aristotle's Concept of Mind', Proceedings of the
Aristotelian Society, 73: 101-14, repr. in Barnes (1979), 4: 32-4r.
- - (1984) (ed.), The Complete Works of Aristotle, The Revised Oxford
Translation, 2 vols. (Princeton, NJ).
- - SCHOFIELD, M., and SoRAB.JI, R. (1979) (eds.), Articles on Aristotle, 4
vols. (London).
BENNEIT, J. (1971), Locke, Berkeley, Hume (Oxford).
BERTI, E. (1981) (ed.), Aristotle on Science, Proceedings of the Eighth
Symposium Aristotelicum (Padua).
BLACK, MAX (1952), 'The Identity of Indiscemibles', Mind, 61, repr. in Loux
(1976), 250-62.
BOGEN, J., and McGUIRE, E. J. (1985) (eds.), How Things Are (Dordrecht).
BOLTON, R. (1976), 'Essentialism and Semantic Theory in Aristotle: Posterior
Analytics II. 7-10', Philosophical Review, 514-44.
- - (1984), 'Aristotle on the Signification of Names' in the Greek Philoso-
phical Society (1984), 153-62.
- - (1987), 'Definition and Scientific Method in Aristotle's Posterior Analy-
tics and Generation of Animals', in Gotthelf and Lennox (1987), 120-66.
Bibliography
Bolton, R. (1990), 'The Epistemological Basis of Aristotelian Dialectic', in

Devereux and Pellegrin (1990), 185-236.
- - (1991), 'Aristotle's Method in Natural Science: Physics I', in Judson
(1991), 1-30.
- - (1993a), 'The Problem of Dialectical Reasoning in Aristotle', Ancient
Philosophy.
- - (r993b), 'Aristotle's Account of the Socratic Elenchus', Oxford Studies
in Ancient Philosophy.
- - (forthcoming), 'Science and the Science of Substance in Aristotle'.
BoNITZ, H. (1868), Aristotelis Metaphysica (Bonn).
BuRNYEAT, M. F. (1992), 'Is an Aristotelian Philosophy of Mind Still Cred-
ible?', in Nussbaum and Rorty (1992), 15-26.
- - et al. (1979), Notes on Book Zeta of Aristotle's Metaphysics (Oxford).
CARLSON, G. N. (1982), 'Generic Terms and Generic Sentences', Journal of
Philosophical Logic, 145-8r.
CHARLES, D. (1984), Aristotle's Philosophy of Action (London).
- - (1985), 'Aristotle on Materialism and the Explanation of Behaviour',
unpub. paper.
- - (1986), 'Aristotle: Ontology and Moral Reasoning', Oxford Studies in
Ancient Philosophy, 4: II9-44.
- - (1988), 'Aristotle on Hypothetical Necessity and Irreducibility', Pacific
Philosophical Quarterly, 69: 1-53.
- - (1990), 'Aristotle on Meaning, Natural Kinds and Natural History', in
Devereux and Pellegrin (1990), 145-67.
- - (1991), 'Aristotle on Teleological Causation', in Judson (1991), 101-28.
- - (1992a), 'Aristotle on Substance, Essence and Biological Kinds', in
Cleary and Shartin (1992), 215-49.
- - (1992b), 'Supervenience, Composition, and Physicalism', in Charles and
Lennon (1992), 265-96.
- - and LENNON, K. (1992) (eds.), Reduction, Explanation and Realism
(Oxford).
CHARLTON, W. (1970), Aristotle's Physics: Books I and II (Oxford).
- - (1972), 'Aristotle and the Principle of Individuation', Phronesis, 17:
239-49.
- - ( 1983), 'Prime Matter: A Rejoinder', Phronesis, 28: 197-2n.
- - (1991), 'Aristotle on the Uses of Actuality', in Cleary and Shartin
(1991), 1-22.
CHERNISS, H. (1944), Aristotle's Criticism of Plato and the Academy (New
York).
CLEARY, J. J., and SHARTIN, D. (1991, 1992) (eds.), Proceedings of the Boston
Area Colloquium in Ancient Philosophy, 5, 6 (Lanham, Md.).
CooE; A. (1976), 'The Persistence of Aristotelian Matter', Philosophical
Studies, 29: 357-67.
- - (1978), 'No Universal is Substance', Paideia, 65-74.
- - (1985a), 'On the Origins of Some Aristotelian Theses About Predica-
tion', in Bogen and McGuire (1985), 101-3r.
Bibliography 357
- - (1985b), 'The Aporematic Approach to Primary Being in Metaphysics
Z', Canadian Journal of Philosophy, suppl. vol. 10: 1-20.
- - ( 1986a), 'Aristotle: Essence and Accident', in Grandy and Warner
(1986), 411-39.
- - (1986b), 'Aristotle's Investigation of a Basic Logical Principle', Canadian
Journal of Philosophy, 341-57.
- - (1987a) 'Metaphysics and Logic', in Matthen (1987), 127-49.
- - (1987b) 'Soul as Efficient Cause in Aristotle's Embryology', Philoso-
phical Topics, rs: 51-9.
- - (1993), 'Essentialism and the Life Sciences in Aristotle', unpub. paper.
- - and MoRAVCSIK, J. (1992), 'Explaining Various Forms of Living', in
Nussbaum and Rorty (1992), 129-45.
COHEN, S. MARC (1984), 'Aristotle and Individuation', in Pelletier and King-
Farlow (1984), 41-65.
- - (1987), 'The Credibility of Aristotle's Philosophy of Mind', in Matthen
(1987), 103-25.
COLSON, D. D. (1983), 'Aristotle's Doctrine of Universalia in Rebus',
Apeiron, IT 113-24.
COOPER, J. M. (1988), 'Metaphysics in Aristotle's Embryology', Proceedings
of the Cambridge Philological Society, 214: 14-41, repr. in Devereux and
Pellegrin (1990), 55-84.
DANCY, R. (1975), Sense and Contradiction (Dordrecht).
DEVEREUX, D., and PELLEGRIN, P. (1990) (eds.), Biologie, logique, et meta-
physique chez Aristote (Paris).
DEVITI, M. (1991), Realism and Truth (Oxford).
DRISCOLL, J. (1981), 'EIAH in Aristotle's Earlier and Later Theories of Sub-
stance', in O'Meara (1981), 129-59.
EDWARDS, J. ([1754] 1957), Freedom of the Will, ed. P. Ramsey (New Haven,
Conn.).
FEREJOHN, M. (forthcoming), 'Selecting Immediate Premises for Aristotelian
Demonstration'.
FINE, K. (1983), 'Aristotle on Substance', unpub. paper.
- - 'Some Remarks on the Theory of Part-Whole', unpub. paper.
FREDE, M. (1978), 'lndividuen bei Aristoteles', Antike und Abend/and, 24:
16-39, trans. and repr. in Frede (1987), 'Individuals in Aristotle', 49-71.
- - (1985), 'Substance in Aristotle's Metaphysics', in Gotthelf (1985), 17-
26, repr. in Frede (1987), 72-80.
- - (1987), Essays in Ancient Philosophy (Minneapolis).
- - (1990), 'The Definition of Sensible Substances in Met. Z', in Devereux
and Pellegrin ( 1990 ), II3-29.
- - (1992), 'On Aristotle's Conception of the Soul', in Nussbaum and Rorty
(1992).
- - and PATZIG, G. (1988), Aristoteles: Metaphysik Z, 2 vols. (Munich).
FREELAND, C. (1987), 'Aristotle on Bodies, Matter, and Potentiality', in
Gotthelf and Lennox (1987), 392-407.
FREUDENTHAL, G. (1994), Aristotle's Physiology and Chemistry: The Theory
Bibliography
of Soul, Vital Heat, and Pneuma and its Foundation in Aristotle's Early
Theology (Oxford).
FURTH, M. ( 1978), 'Transtemporal Stability in Aristotelian Substances',
Journal of Philosophy, 7S: 624-46.
- - (1988), Substance, Form and Psyche: An Aristotelian Metaphysics
(Cambridge).
GEWIDMET, P. (1985) (ed.), Aristoteles Werk und Wirkung (Berlin).
GILL, M. L. (1984), 'Aristotle on the Individuation of Changes', Ancient
Philosophy, 4: 9-22. _
- - (1989), Aristotle on Substance: The Paradox of Unity (Princeton, NJ).
- - (1990), Review of Frede and Patzig (1988), Journal of the History of
Philosophy, 28: 602-5.
- - (1991), 'Aristotle on Self-Motion', in Judson (1991), 243-65.
- - (1993), 'Comment on A. Code: Essentialism and the Life Sciences in
Aristotle', unpub. paper.
- - (forthcoming), 'Aristotelian Compounds', Special Issue, Pacific Philo-
sophical Quarterly, ed. Robert Bolton and Frank A. Lewis.
GOODMAN, N. (1955) Fact, Fiction, and Forecast (Cambridge).
GOTTHELF, A. (1985) (ed.), Aristotle on Nature and Living Things
(Pittsburgh).
- - and LENNOX, J. (1987) (eds.), Philosophical Issues in Aristotle's Biology
(Cambridge).
GRAESER, A. (1987) (ed.), Mathematik und Metaphysik, Proceedings of the
Tenth Symposium Aristotelicum (Bonn).
GRANDY, R., and WARNER, R. (1986) (eds.), Philosophical Grounds of
Rationality: Intentions, Categories, Ends (Oxford).
GREEK PHILOSOPHICAL SOCIETY (1984) (ed.), Language and Reality in Greek
Philosophy (Athens).
GRICE, H. P. (1975), 'Method in Philosophical Psychology', Proceedings of
the American Philosophical Association, 48: 23-53.
HAMLYN, D. W. (1968), Aristotle's De Anima: Books II, III (Oxford).
HARTMAN, E. (1977), Substance, Body, and Soul: Aristotelian Investigations
(Princeton, NJ).
HEIDEGGER, M. (1982), Basic Problems of Phenomenology, trans. A
Hofstadter (Bloomington, Ind.). •
HEINAMAN, R. (1979), 'Aristotle's Tenth Aporia', Archiv filr Geschichte der
Philosophie, 57: 1-20.
HICKS, R. D. (1907), Aristotle: De Anima (Cambriflge).
HUME, D. (1739), A Treatise of Human Nature, ed. L. A. Selby-Bigge
(Oxford, 1888).
IRWIN, T. (1988), Aristotle's First Principles (Oxford).
JoNES, B. (1974), 'Aristotle's Introduction of Matter', Philosophical Review,
83: 474-500.
JUDSON, L. (1991) (ed.), Aristotle's Physics: A Collection of Essays (Oxford).
KAHN, C. (1985), 'On the Intended Interpretation of Aristotle's Metaphysics',
in Gewidmet (1985).
KANT, I. (1781, 1787), Kritik der reinen Vernunft (Hamburg, 1957).
Bibliography 359
KING, H. R. ( I956), 'Aristotle without Prima Materia', Journal of the History
of Ideas, IT 370-89.
KIRWAN, C. (1971), Aristotle's Metaphysics, Books r, ~' and E (Oxford).
KOSMAN, L. A. (1984), 'Substance, Being, and Energeia: The Argument of
Metaphysics Theta', Oxford Studies in Ancient Philosophy, 2: I2I-49.
- - (I 987), 'Animals and Other Beings in Aristotle', in Gotthelf and Lennox
( I987), 360-91.
KRATZER, A. (1988), 'Stage-level and Individual-level Predicates', University
of Massachusetts Linguistics Manuscripts.
LEAR, J. (1987), 'Active Episteme', in Graeser (1987), 149-74.
- - (1980), Aristotle and Logical Theory (Cambridge).
LEE, E. N., MoURELATOS, A. P. D., and RORTY, R. M. (1973) (eds.), Exegesis
and Argument: Studies in Greek Philosophy Presented to Gregory Vlastos,
Phronesis, suppl. vol. l (Assen).
LESHER, J. H. (I971), 'Aristotle on Form, Substance and Universals: A
Dilemma', Phronesis, 16: I69-78.
LEWIS, D. (1968), 'Against Structural Universals', Australasian Journal of
LEWIS, F. A. (I982), 'Accidental Sameness in Aristotle', Philosophical
Studies, 42: 1-36.
- - (1985), 'Form and Predication in Aristotle's Metaphysics', in Bogen and
McGuire (1985), 59-83.
- - (I986), 'What is Aristotle's Theory of Essence?', Canadian Journal of
Philosophy, suppl. vol. rn: 89- I3I.
- - (1991), Substance and Predication in Aristotle (Cambridge).
- - (I992), 'Aristotle on the Unity of Substance', unpub. paper.
LLOYD, A. C. (1970), 'Aristotle's Principle of Individuation', Mind, 79, 5r9-
29.
- - (1981), Form and Universal in Aristotle (Liverpool).
LOCKE, J. (1690), Essay Concerning Human Understanding.
Loux, M. (1976) (ed.), Universals and Particulars: Readings in Ontology
(Notre Dame, Ind.).
- - (1991), Primary Ousia: An Essay of Aristotle's Metaphysics Z and H
(Ithaca, NY).
McTAGGART, J. (1927), The Nature of Existence, 2 vols. (Cambridge).
MATIHEN, M. (1987) (ed.), Aristotle Today (Edmonton, Alberta).
MATTHEWS, G. B. (I982), 'Accidental Unities', in Schofield and Nussbaum
(1982), 223-40.
MAYR, E. (1987), 'The Ontological Status of Species: Scientific Progress and
Philosophical Terminology', Biology and Philosophy, 145-66.
MIGNUCCI, M. (I981), 'Hos Epi to Palu', in Berti (1981).
MILLER, F. D. (1973), 'Did Aristotle Have the Concept of Identity?', Philo-
sophical Review, 82: 483-90.
MODRAK, D. K. (1979), 'Forms, Types, and Tokens in Aristotle's Meta-
physics', Journal of the History of Philosophy, I7: 371-81.
- - (I985), 'Forms and Compounds', in Bogen and McGuire (I985), 85-roo.
MooRE, G. E. (I903), Principia Ethica (Cambridge).
Bibliography
MoRAVCSIK, J. (I991), 'What Makes Reality Intelligible?', in Judson (1991),

31-47.
MosER, P. K. (1983), 'Two Notions of Substance in Metaphysics Z', Apeiron,
IT 103-12.
MouRELATOS, A. (1984), 'Aristotle's Rationalist Account of Qualitative
Interaction', Phronesis, 29: 1-16.
NEWTON, I. (1686), Principia Mathematica, trans. A. Motte (Berkeley, Calif.,
1934).
NussBAUM, M. (1984), 'Aristotelian Dualism: Reply to.Howard Robinson',
Oxford Studies in Ancient Philosophy, 2: 197-207.
- - and RoRTY, A. (1992) (eds.), Essays on Aristotle's De Anima (Oxford).
O'CONNOR, D. J. (1964a), 'Aristotle', in O'Connor (1964b), 36-61.
- - (1964b) (ed.), A Critical History of Western Philosophy (New York).
O'MEARA, D. J. (1981) (ed.), Studies in Aristotle (Washington, DC).
OWEN, G. E. L. (1965a), 'Inherence', Phronesis, 10: 1-9.
- - ( 1965b ), 'The Platonism of Aristotle', Proceedings of the British
Academy, 50: 125-50.
- - ( 1978), 'Particular and General', Proceedings of the Aristotelian Society,
78: 1-21, and in Owen (1986), 279-94.
- - (1986), Logic, Science and Dialectic (London).
OWENS, J. (1978), The Doctrine of Being in the Aristotelian Metaphysics, 3rd
edn. (Toronto, Ontario).
PATZIG, G. (1969), Aristotle's Theory of Syllogism (Dordrecht).
PECK, A. L. (1953), Aristotle: Generation of Animals (Cambridge, Mass.).
PELLETIER, F. J., and KING-FARLOW, J. (1984) (eds.), New Essays on Aristotle,
Canadian Journal of Philosophy, suppl. vol. IO.
PUTNAM, H. (1988), Representation and Reality (Boston).
ROBINSON, H. (1983), 'Aristotelian Dualism', Oxford Studies in Ancient
Philosophy, l: 123-44.
RoRTY, R. (1973), 'Genus as Matter: A Reading of Metaphysics Z-H', in
Lee, Mourelatos, and Rorty (1973), 393-420.
Ross, W. D. (1924), Aristotle: Metaphysics, 2 vols. (Oxford).
- - (1936), Aristotle: Physics (Oxford).
RYLE, G. (1931-2), 'Systematically Misleading Expressions', in Ryle (1971),
2: 39-62.
- - (1971), Collected Papers, 2 vols. (London).
ScALTSAS, T. (1985), 'Substratum, Subject, and Substance', Ancient Philo-
sophy, 5: 215-40, rev. and repr. in Anton and Preus (1992), 179-211.
- - (1989), Review of Furth (1988), Philosophical Books, 30: 82-5.
- - (1990), 'Is a Whole Identical to its Parts?', Mind, 99: 583-98.
- - (1992a), 'A Necessary Falsehood in the Third Man Argument', Phrone-
sis, 3T 216-32.
- - (1992b), Review of Loux (1991), Philosophical Books, 33: 139-42.
- - (r993a), 'Aristotle's "Second Man" Argument', Phronesis, 38: 117-36;
repr. in Scaltsas (1993b).
- - (1993b) (ed.), Aristotelian Realism (Athens).
- - (forthcoming), Substances and Universals in Aristotle's Metaphysics
(Ithaca, NY).
Bibliography
SCHOFIELD, M., and NUSSBAUM, M. (1982) (eds.), Language and Logos:

Studies in Ancient Greek Philosophy Presented to G. E. L. Owen
(Cambridge).
ScHWEGLER, A. (1847), Die Metaphysik des Aristoteles (Tubingen).
SELLARS, W. (1957), 'Substance and Form in Aristotle', Journal of Philo-
sophy, 54: 688-99.
- - ( 1967), 'Aristotle's Metaphysics: An Introduction', in Philosophical
Perspectives (Springfield, III.), 73-124.
SHARVY, R. (1983), 'Aristotle on Mixtures', Journal of Philosophy, So:
439-57.
SHIELDS, C. (1988), 'Soul and Body in Aristotle', Oxford Studies in Ancient
SMITH, J. A. (1921), 'TOL1E Tl in Aristotle', Classical Review, 35: 19.
SORABJI, R. (1974), 'Body and Soul in Aristotle', Philosophy, 49: 63-89,
repr. in Barnes, Schofield, and Sorabji (1979), 4: 42-64.
WATERLOW, S. (1982), Nature, Change, and Agency in Aristotle's Physics
(Oxford).
·WEDIN, M. V. (1991), 'PARTisanship in Metaphysics Z', Ancient Philosophy,
II: 361-85.
WHITE, N. (1971), 'Aristotle on Sameness and Oneness', Philosophical
Review, 80: 177-9.
- - (1986), 'Identity, Modal Individuation, and Matter in Aristotle', Mid-
west Studies in Philosophy, II: 475-94.
WHITING, J. ( 1986), 'Form and Individuation in Aristotle', History of Philo-
sophy Quarterly, 3: 359-77.
- - (1991), 'Metasubstance', Philosophical Review, roo: 607-39.
WIGGINS, D. (1980), Sameness and Substance (Oxford).
WILLIAMS, B. (1986), 'Hylomorphism', in Woods (1986), 189-99.
WILLIAMS, C. F. J. (1982), Aristotle's De Generatione et Corruptione
(Oxford).
WITT, C. (1989), Substance and Essence in Aristotle: An Interpretation of
Metaphysics VII-IX (Ithaca, NY).
- - (1991), 'Commentary on Charlton', in Cleary and Shartin (1991), 23-7.
Wooos, M. J. (1986) (ed.), Oxford Studies in Ancient Philosophy, 4. A
Festschrift for J. L. Ackrill (Oxford).
ZELLER, E. (1897), Aristotle and the Earlier Peripatetics, trans. by B. F. C.
Costelloe and J. H. Muirhead, 2 vols. (London).
INDEX LOCO RUM
ALEXANDER OF APHRODISIAS Prior Analytics
In Aristotelis Metaphysica Commentaria I. 30 322
580 208 46a17-27 322
468 22 322
ARISTOTLE II. 16 334n.
Categories 55, 63-4, Posterior Analytics
70, I. I 71a1-9 325
196, I. 2 322
217, 71"26-35 352
247 72"14-18 352
2 1"20-h9 225 n. 72a16-17 325
1"6-9 55 72a37-"14 352
3 196 I. 3 325
5 2a34-"6 225n. 72"13-25 352
2"1-6 55 I. 4 296n.
3"I0-24 52 73a38-"3 296n.
3"15-2r 52 l.5 74"35-"5 3020.
7 138n. I. IO 323
7''5- II 284n. I. II 77•33-4 324
IO 296n. I. 24 85"15-18 92n.
12 14"30-5 309n. II. 7 92h32 2950.
14hI0-13 217n. II. 8 93a16-20 8rn.
93a29-36 81 n.
De Interpretatione 331 n., 93"30-6 81 n.
337 93"I0-14 81 n.
l 16"3-9 337 IL IO 93"36 295n.
168 3-8 338n. 94"4-6 8rn.
16"17 337,338 IL I I 94"21-2 8rn.
2 16"19-26 338 98b35-8 8rn.
16"19-29 338n. II. 17 99a21-4 92n.
16"29-32 338n. 99"30-"7 8rn.
16"32 338n. 99"30-7 92n.
3 l6h19-25 338n. II. 19 99"26-30 325
16h20 337 100"6-9 354
7 17"26-31 146n.
17h27-9 146 Topics 247,
8 18"17 337 322n.
18"19-26 338n. I. 1 IOOa20-l 33on.
IO 19h8-9 338n. 100"4 353
19°9 337 I. 2 327-8
II 20"22ff. 332 101"3 328
21"23 208 I. 7 44-5, 57,
64
Analytics 86, 91, 103a9-IO 44
104, 103"23-39 44
322. I. 8 I06"5-6 329
351-3 VII. l 152"25-8 146n.
364 Index Locorum
VIII 325 Physics 63, 293,
VIII. I 155b29-32 341 n. 307n.
156"18-19 341 n. I 248,257,
156"27 341 n. 259
157•1-13 339n. I. 2 185b5 l32n.
157•1-3 342n. I. 7 237,301
VIII. 5-6 329 r89b30-2 248
VIII. 5 159•38ff. 333 190•5-9 303
r59b2-4 332 r90•13-r4 248
VIII. 6 160"1-2 339n. 190•15 l39n.
VIII. 13 334n. 190•24-6 304n.
163"10-13 334n. l9obr5 44
190"24-5 109n.
Sophistical Refutations 325, 328, l9ob24 44
345, 191•10-14 79
348 I. 9 192"16-25 25on.
2 326-7, 192"25-9 275n.
329 l92b9 266n.
165 3 38-9 326 II. l 263-4
r65b1-3 326,329 l92b9-14 266
l65b3 326,329 II. 2 l94"27-b15 25on.
l65b4-5 329 l94bro-15 292n.
l65b5-6 33on. II. 7 235-6
r65b8-9 326 198°5-6 235
4 166"2-3 336 II. 8 l98b34 240
166"16 336. l99b3l 97
166"22 336 II. 9 20011 1I 202
166"25 336 III 202
5 326 III. I 201 3 1I-15 284n.
167'123-7 331 201"16-18 94n.
6 168"25-6 335-6, 201"27-"5 96n.
339n. 201"29 202n.
7 169b3-6 146n. 201"5 202 n~
8 326 20Ibl0- 12 95n.
169b20-9 326 201b28-9 95n.
16gb24 336 202"1 95n.
169b25 326 202"7 202n.
9 334 IIl.3 202 3 21-4 94n.
170"22-6 328 202b5-22 94n.
170"36-bII 328 202b14-16 146n.
170"36-9 328 III.6 207•19 48
17d'8-11 328 IV.4 212"5-6 62
II l71b4-5 326,327 IV. II 219"1-2 62
171b5-6 329 V.4 65
17rb9 327 227°23-4 66
172"15-21 324 227"29-32 66
172"21 351 227"30-3 66
172"30-7 329, 228•3-19 66
33on. 228"3-7 66
172"30-4 329 VIII. I 251"9 202n.
24 179•37-9 ,146n. VIII. 8 309,
15l.r 3 1h-1n ;.,,....e
""'
Index Locorum 365
De Cae/o 412b4-9 r33n.
I. 9 18 412b6-8 28I-2
412bI0-413aIO 296n.
De Generatione et 412hu-17 271
corruptione 293 412b16-17 271
I. 2 316a5-14 292n. 412h18-413a3 267
I. 3 318h32-3 67n. II. 2 413hu-414a3 267n.
I. 5 321h25 26,39 413b14- 15 267n.
321b29-32 312n. 413b27-32 267n.
I. 7 249 413b29 267n.
I. IO 327a34-b6 275n. II. 3 252n.
327b24-31 273 414b26-7 252n.
327b30-1 274n. II. 4 415a22 223n.
328"4 273 4163 6-9 274n.
328118-12 273 II. 5 163
328h22 273 417•32-4r7h2 202
II. 2 274n. 417•35-417b2 204
II. 6 333b4-6 240 417br 208
TI. 7 3343 27-b2 274 4I7b2-5 204
334a3r-5 273 417b22-5 92n.
II. 8 262 n., III.4 429 3 II-I2 267n.
274n. III.9 432a20 267n.
II. IO 274n.
Sense and Sensibilia
Meteorologica 44Ihr5-30 267n.
4
IV 274n.
IV.4 274n. Historia Animalium 240
IV. IO 38911 19-22 262n. III. 2 5I lb2-4 266
IV. II 389b9-18 262n. VI.2 56ob19-20 240
389b9-I5 263n. VI. 13 5683 II-I2 240
IV. I2 389h29-39oh1 29 VII. II 587b30-1 240
390"IO-I3 312n. VIII. 3 592b28-9 241
39ob2-IO IOI IX. 7 240
DeAnima 202,207, De Partibus Anima/ium
253, I. I 268n.
269, 639b26-640•1 25on.
280, 640•18-26 220
283 64oh17-29 268n.
I. I 403a25 292,316 64ob34-64r15 251
l.4 408 3 10-13 267n. 641"6 269n.
408b1-32 163 641"18 252n.
I. 5 4Ilb5-I4 267n. 641"20 3I2Il.
4IIb7-14 267n. I. 2 642•12-14 25on.
4IIb7-9 274n. I. 5 645 3 34 25on.
4IlbI5-30 267n. II. 2-4 262
II. I 164 n., II. 2 646"14-21 262
252 n., 647bI0-16 266
271 648h9-IO 262
412"7 48 649"15-19 265
4I2"8 50 II. 3 649bII-l4 263 n.,
AT?aT'l-TI;; ?"'7T
,.,/'.,,.
366 Index Locorum
649h25-8 270 B. 3-4 169
649h29:...-30 263 n. B.4 1001"2-"25 169
650"32-hr9 258 r. 323
650"34 266n. r. I 321
65oh7-8 266n. 1003"21 321
II. 4 651"14 258n. r.2 50, 139,
II. IO 656h21 266n. 328
III.4 666"18 266n. 1003h9-ro 337
666"24 264 1003h22 337
III. 5 667h22-31 264 roo4"22 132n.
668"4 258n. 1004h25-6 327,351
668"14-27 258 roo5"7 132 n.
De Generatione Animalium r. 3-4 351
726h5-6 r.3 1005°8-17 325
I. 19 258n.
1005h9-11 337,351
726"30-727"2 258n.
1005h14-16 351
727h14-728"33 258n.
1005°16-17 351
727"31 208
1005h22 323
I. 20 728"17-21 258n.
1005h25 333
729"34 254n.
I. 21 258n. r. 4 150,323
II. I
r. 4-8 145 n.
731"24-732"1 I 227
732h32-733h23 258n. r. 4 1006"1-2 329
1006"3-5 323
734h19-735"29 258n.
1006"5-11 323
734 °24-735"26 29n. 1006"6 213
734h24-7 312 n.
1006"10-11 352
734°31-4 IOI
1006"11-12
II. 4 323
739°20-30 263 1006"12-24 332
740"7-8 264 1006"14 335n.
740°29-741"3 263 1006"15- 18 324,
740°30-6 259 335n.
74oh34-7 259 1006"17-18
II. 6 743"2-21 258-9 329
10015"18-25 339, 344
743°20-5 276 1006"18 335n.,
744hl I -745"18 259
v 92n. 349
roo6"r9 33on.,
Metaphysics 86, 174- 349n.
5, 195, 1006"20-2 349
199, l006"20-1 334
206, 1006"21-5 335
213, 1006"21-4 339
283, 1006"24 33on.,
322 349
A. 1-2 353 1006"25-6 324,349
A. I 981"5-7 353 1006"26 333
981"25 353 1006"28-31 339n.
982"1-6 353 1006"28-30 339,344
982"8-ro 354 1006"31-4 340
982"21-5 354 1006"31-2 340,343
983"5-11 354 1006"32-4 340,344
983h6-18 257n. l006"34-h1 I 337,348
a.2 00'1. 3 T-h6 6? n rno6°R '2 ?h
Index Locorum 367
1006hn-13 335 l0l6h36 64
roo6h13-28 340-1, l0l6h36-I017"3 109n.
344 1017•3-4 61,64
1006"13-15 337 t:..7 222
roo6h18-22 336 1017•7-18 II On.
1006h20-8 340 1017"35- 1017"9 215
1006h20-2 331, 339 1017"35-h9 209,
1006"22-7 337 221 n.
1006h28-30 341,344 1017h5 222
~
1006h30 341 t:.. 8 1017h24-6 67n.
1006h31-2 341,345 t:..9 43-4
1006h33-4 324, 331, 10183 6 43,45
341, 1018"8-9 45
345 10183 10 43
10073 1-7 340 t:.. II l39n.
rno7 3 8 33on. 1018h38-1019"15 309n.
10073 20 326, 1019"1-14 216n.
344n. t:.. 12 173, 186
1007•25-7 344n. 1019"17-18 97
roo7h18 33on. t:.. 15 1021"9-13 135
10083 31 332 1021"27-h3 138n.
loo8h1-2 334 t:.. 25 l32n.
1008h10-12 332 1023h18-22 37
r.5 l0l0h25-6 343 E. I 1025h19-10263 7 292n.
r.6 lOIIa3-16 345 1025"32 314
t:...4 1014h30-1 25 1025"36- 1026"3 296
l0l4h35-1015•1 I 264 E. 2 337
t:.. 6 43-5, 50, 1027"8-11 235
132 n., !0273 20-1 235
133 z 71, 174-
1015"16-34 44 6, 196,
1015"16- 19 25 200,
10r5h36-1016"17 64 286,
ro15h36-1016"6 61 291,
1016"5-15 l36n. 293
1016"17-21 45 Z,H 33, 63n.,
1016"18-23 l38n. 224 n.,
1016"27-8 109n. 280,
1016"28 45 286
1016"32-5 135, Z,H,and 196
138n. 0
ro16"3-7 134n. Z. I 291,
1016b7-9 136, 156 294n.
I016h8-II 64 z. 1-6 293n.
1016°1 l-16 43 Z. I 1028"13-31 49
1016°u-13 43 Z.2 1028h8 281 n.
1016h18-24 134n. 1028h8-13 266
1016"31-5 6o 1028"10 266n.,
1016°31-3 22, 42-3, 28xn.
48 Z.3 196, 251.
1016"32-4 l39n. 277,
rorn""l<;:-6 A"l 20T.
368 Index Locorum
294n. 300-1,
I029"3-5 268n. 302n.,
I029a5-7 50, 291 n. 305,
I029"26-30 71 307n.,
I029"27-30 48, 67n. 308,
I029"27-8 49,291 3IO,
I029"30-4 291 n. 314,
I029"30-2 121 316
I029"30-1 I07n. Z.7-9 294n.
I029•32 292n. Z.7 I032 3 12-15 301
Z.4 131 n., I032"13-16 303
197, I032"14-15 301 n.
291-5, I032"15-b31 301
297, I032"19-22 257n.
337 I032h12-13 301
Z.4-6 141 n., 1032b3 I- I033•1 257n.,
197 301
Z.4 I029b3 292n. I033•1-2 302n.,
I029b12-14 292 308
I029b22 298n. I033"1 303
I029b27 146n. I033•2 308
1029b29- I0303 2 295 I033"3-20 308
I029b3r-3 296 I033•3-4 301,
I029b32-3 295 302n.
I030"2-7 295 I033 3 5-23 302n.
I03o"IO 295 I033 3 6-8 302
I030"12 294 I033"6 304
I03ob8-12 132n., I033"7'!:.8 306
136n. 1033"14 304
I03obIO 295n. I033"17-20 3o6-7
Z.5 292-7, I033"19-22 257n.
300-1, I033"20-3 304
308, Z.8 61-2, 64,
314, 70-1
316 I033b5-6 249n.
I03d'18-28 297 I033b21-2 291 n.
I03ob18-19 298n. I033b24-6 107n.
I03ob24 296n. I0343 5-8 22,45,
I03ob28- I031"1 297-8 48,6o,
I03oh30-r 3r7n. 92
I03ob31-2 316n. I034"6-8 238
Z.6 122,210, Z. IO 131 n.,
255, 140,
292n. 308,
I031"19-27 2IO, 343 3IO,
I031h22-7 343 314,
I031b28-I032"4 120 316
I031b28-30 122 Z. IO-II 294n.,
I032"3 156n. 297,
Z.7 159, 300-1,
257n., 308-IO
'20'1. Z. IO-T2
"'"'
Index Locorum 369
Z. IO 1034h20-4 308-9 1036h22-3 236
1034h23 312n. 1036h23 288,316
1034h24-7 308 1036h26-30 313 n.
1034h25-7 309 1036h28-33 315 n.
1034h28- 1035•14 309, 311 1036h28-31 86,236
\034h28-34 312 n. rn36h28-30 86n.,
1034h28-33 310n. 315 n.
1034h31-2 309 rn37•5-7 78
1034h32 310n. rn37"6-7 131 n.
rn34h33 132n. rn37•7-rn 283
rn34h34 3IO rn37•7-9 rn7n.
1035"1-7 3JO rn37•9-rn 282n.
rn35"4-7 3rn rn37"10-20 292n.
rn35"7-9 312 rn37•22-4 313 n.
rn35"8- 12 3 l I, rn37''24-7 3rnn.
312 n., rn37''24-6 313
313 rn37''24-5 310,
rn35"8-9 311-12 313 n.
rn35"rn- r r 164 n., rn37"26-30 315 n.
31 In. rn37"26-9 296n.
rn35"14 312 rn37•26-7 313
rn35•19-21 107n. rn37"26 314
rn35"19-20 312 n. rn37"27-30 147
rn35"24 3rnn. rn37•27-9 306,
rn35"26-32 147 313 n.
rn35"26h-2 rn37"27 313,315
rn35"29 3IO rn37"28 316
rn35h9-28 16on. rn37"29-30 316
rn35hrn-14 164 n. rn37•29 316
rn35h12-22 147 rn37"34-h7 147
rn35h14- 15 86n. z. 12 130 n.,
ro35h16-25 123 29411.
rn35h21 123 J037hII-12 115
ro35h22-5 31 In. rn37h14-18 n5n.
rn35h25 283, rn37h21-3 115
312 n. rn37h26-9 296n.
rn35h27-30 131n.,78 rn38"6 91
rn36"1-8 141 n. z. 13 122, 127
ro36"2-9 147 ro38°13-14 156n.
rn36"8-9 306,313 rn38°17-23 122, 127
Z. II 131n., z. 14 rn39h1 133 n.,
235, 146
288, rn39°9-30 251 n.
294, z. 15 6o-1,159
314, rn39"20-31 147
316, rn39"27-31 141 n.
317 n. rn40"2-7 141 n.
rn36"28 147 rn4o"rn 317n.
rn36"31-h7 288 rn40"27- 04 141 n.
rn36h3-5 314 rn40"29-31 113 n.
rn36h3-4 296n. z. 16 56, 122,
rn36"22-6 314 312 n.
370 Index Locorurn
1040°5-16 266 rn41025-33 122

rn40"5-15 122 rn41°25-7 132
rn40°5-8 281, 1041°26-7 II2
312 n. rn41b27-41 Il2
rn4ob8-rn 56, 124 1041°31-3 132
rn40°9 305n. H 175, 198,
rn40°17-22 169 207,
rn40°17-19 156n. 284
rn40"25-7 225n. H. I rn42"27-8 48
Z. 17 22, 56, rn42"28-9 67n.
84, 86, rn42"29 116
I 14, H.2 84-5,
122, 265-6,
143-4, 267
160, I042°9-1 l 237
167n., 1042°31 84,265
198, rn43"2-5 265n.
317n. 1043"3-4 265 n.
1041"10-25 143 rn43"3 265n.
1041"10-14 II5 1043"4-7 25, 265
rn41"14-19 133 n. 1043"4 265 n.
rn41"18 152 n. rn43"5-6 147
1041"19 l44n. rn43"7-9 96n.
1041"20-1 II5 rn43"8-9 268n.
rn41"21-5 143 1043"15 85, 25on.
1041"23 rr5 1043~'25-6 85n.
1041"24-5 115 1043"27-9 89
1041"25- 0 2 143 H.3 II5
rn41"25-7 II5 rn43"31-3 284n.
I041"26 34 1043"36-7 282-3
w41"27-32 rn3 n. 1043°2-4 84,91
rn41"33-"3 143 1043°3 282
ro4r 0 2 115 104305-6 38, l 16
1041°4-9 56 1043b10-18 147
1041°5-10 151 1043h21-3 265 n.
ro41°5-9 r23 1043"11 84
1041°5-6 91, r15 H.4 rn44"34-5 258n.
ro41°6-9 86n. 1044"35 209n.
ro41°6-7 84, 91, ro44"36-"1 103n.
143, I044h3-20 265
317n. 1044°16 84
ro41°6 r23 H. 5 250
1041h7-9 317n. I045"1-2 II6, 250
ro41"7 rr5 1045"3-6 27n.
ro41"9-II 86,91 w45"3-5 188
1041°1r-34 147 H.6 22, 63n.,
1041°12- 14 l 12 71, 85,
I04lhI2 130 87-93,
ro41h16- 18 152 96,
rn41°22 114 II8-
ro41h25-34 147 19,
Index Locorum 371
133, 1046"1-12 180

161' 104611 1-2 181-4
254, 1046"4-19 178
., 317n . 1046"4-6 178
1045"7-22 rr4 1046"6-7 178
1045a7-13 136n. 1046•9-10 178
1045"7-10 33 1046"10-II 178
1045"14 75n., 1046"19-29 191
295n. 1046"31-3 284n.
1045"20-6 200 0.3 1047"30- 104702 181
1045"20 84,91 1047"30-1 181
1045"23-33 87, 317 n. 1047"32- 104701 182
1045"23-4 75n., 0.5 1048"18-36 95
255n. 0.6-9 184
1045"24 205 0. 6-8 93-9
1045"25-33 268n. 0.6 95, 182-
1045"25-9 r46n. 3, 185,
1045"29 75n. 2II
ro45"30-3 75n.,88, 1048"25-€0 200
90 1048"25- 9 221 n.
1045"31-3 91, 255n. 104811 27-30 184-5
1045"32 87n. 1048"30-2 201
1045"33 87n. 1048"35- 10481>9 183
1045"34-5 91 1048"35-7 237
1045"36 86,91 104806-9 221
1045°2-8 169 1048h6-8 201
1045h7-24 34,88 1048°7-10 93
1045°7-16 90, 1048"9 182
II8n., 1048h18-34 221
124n. 1048b21-2 95
1045"15-24 200 1048°22-4 9on.
1045°17-23 75 n., 88 0.7 75n.,96,
1045b17-19 133 n., 177,
156, 187,
161' 236,
254n. 250,
1045b18-20 88-9 276,
l045b20-1 89 307n.
1045"21-3 89 1048b36 96n.
0 63n.,71, 1048b37- 1049"18 96n.,
104, 276n.
173, 1049"1 96n., 99
193, 1049"2 96n.
196-7, 1049"5-12 250
200, 1049"8-1 I 189
204, 1049"9-1 I 97
207, 1049"10-11 96
209n. 1049"II 99
1045h32-5 174 1049"13- 18 264n.
1045b34-1046"6 180 1049"13- 14 236
1045°35- 1046"3 221 1049"14-16 75n., 97
372 Index Locorum
1049"15-25 100 1065023 202n.

1049a16-17 264n. /\. 205
1049"16 96n. /\.. 3 l070aII- 12 67n.
1049"18-24 302n. 1070"13- 18 67n.
1049"23-4 II6 /\.. 5 68,70
1049"35 67n. l07la20-2 69
0.8 75n.,97, 1071"27-9 68
215-16 /\.. 6 1071°20 205
e. 8-9 186-7 /\.. 9 l38n.
1049°5-10 186-7 M.2 1077"36-04 216n.
1049°5-8 97 M.lO 1087a10-25 225n.
1049°7-lO 75n. n. N. l 1088"23 ll3 n.
1049°9-11 97 1088"30 ll3 n.
1049024-6 217 1088°2 l 13 n.
105oa4-11 218
2r8n. Ethica Nicomachea
1050"5-7
1050"7-lO 223 I. 3 241
1050"8- I I 98 1094°11-27 241
1050"15 219n. 1094021 241
105oa21-3 86n., 90, I. 6 1096"21 n3n.
I. 7 1098a25-30 241
97, 103
1050°6-8 223 II. l l 103"34- l 103°2 204
I. I 43,45, VI. 3 353
50, n39°31-3 353
132n. VI. 7-8 353
1052"15-25 109n. VI. 7 l 1'41-l" 12-20 354
1052"15 132 n. l 141"33-03 354
1052"19-20 61 n., VI. 8 II42"19 354
I II n.
PLATO
1052"22-5 43
105201-3 46
1052"16- 18 Parmenides
135
l052bl6 132" 51
45
1053"31-?, l38n. Sophist 181
I. 3 1054"32- '2 45 Theaetetus
1054°16-17 22,45
204-5 II2n.
1054b23-30 44
I. 6 1056b32- 1057''16 l38n. 205"8-10 II5n.
I. 9 45 Timaeus
10s8b1-II 6on. 69d-75d 267n.
K.9 1065b16 202 n.
GENERAL INDEX
Numbers in italics signify cross-references within this volume
ability, see power analogy 25, 34-5, 38, 40, 60, 78, 84,
abstraction 50, 52, 76, 78, 88, 91, 93, 93-5, 98-9, 183-7, 201, 21 I, 213,
I08, 114-16, II8, l'.'.I-4, 126-8, 237, 257-8, 265, 281, 286-7, 307 n.
198, 233, 242, 275 n., 314; see also analytic-synthetic dichotomy 243
division animal 206, 209, 237, 249, 254, 296; see
accidents 145, 235, 238, 242, 247, also biology
263 n., 296 n.; accidental Anscombe, G. E. M. 39 n.
compounds 247, 337; Accidental aporia (anopia) 88, 206, 291 n.,
Compound Theory 247 n.; see also 297-300, 340
compound (of form and matter) Aquinas, T. 47-8
account, see definition argument 325, 330 n., 333, 335;
Ackrill, J. 56 n., 284 n., 13 l -2, 164 n., dialectic 196, 323-30, 332-3, 340,
2II n., 252-3 n., 256, 257n., 276, 344 n.; didactic 326; elenchus 323
277n., 284 et passim, definition of 331; eristic
acting and being acted upon by, see 326; hypothetical 334; peirastic
affecting and being affected 326-30,332-5,346,349,351;
activity (i!vtpycia, 1lvre}.txcw) 84, 163, regress l 13-14, 120, 122, 125, 260,
165-7, 181-2, 195 et passim, 297-300; rules of 326, 329, 332-3,
221-4, 228, 249, 296n.;seealso 336; Third Man Argument 292 n.;
form; potentiality; power, active transcendental 334, 347; see also
actuality 28-9, 75 et passim, I07, 112, aggregate, argument; belief; trope-
I 14, II9-26, 128, 133, 156-7, overlap argument
161 -9, 173 et passim, 215 et Armstrong, D. 111 n., II6 n.
passim, 236, 242, 249, 254n., 265, artefacts 20- I, 24-5, 31, 147 n., 257,
274, 293, 3IO,317 n.; defined by 263-8, 281, 296 n., 304, 314, 337
goal 97,219,221-2;degreesof atomism 49, I04; Aristotle's
163-5; as ivrcJ.txaa 181, 204-7; anti-atomist view of mixture 26o,
first actuality 206-9, 252 n.; kinds 274; Atomic view of substance
of 181-5, 201-7, 221; relative 1IO-11; Atomist view of matter
227 n.; see also activity; power, 273; see also Democritus;
active Empedocles
affecting and being affected 181, 186, Austin, J. L. 48
249 axe 271
aggregate I08, II0-18, 124-5, 140-2, axiom 323, 351
158, 161, 169, 317 n.; argument
112-15, u7, 119-20, 127; see also Bailey,N. I05n.
heap Balme, D. 67n., 254n., 268n.
Albritton, R. 13, 68 n., 92 n. Barnes, J. 132 n., 15 l n., 280, 285 n.
Alexander of Aphrodisias 208, 209 n. becoming 95-6, 197, 202, 209, 213,
Allaire, E. 71 n. 218, 224n., 248, 257n., 293, 301
Almog, J. 232 n. beg the question, see fallacy
alteration (aJ.).ofwm;) 65, 94 n., 303-4, being 50, 95-6, 174, 195, 197-200,
306; see also change, in sensible 202,204-5,209-13,220,227,339,
quality 344 n.; being of 195, 343; qua
374 General Index
being (cont.): change 27, 30, 50, 65-6, 70, 94n., 97,
being 337-8; relational 199; see 120, 128, l76etpassim, 197,
also actuality; potentiality 201-4,213,216,220,239,276-7,
belief (hcfo~a) 329-30, 343, 346, 293, 303-4, 315; accidental
35on. 249-50; motive 6r, 65, 200-4,
Bennett, J. 41 206, 211-12, 221-2, 236, 258n.; in
biology 104, 227-8, 229, 231-5, 237, sensible quality (a).}.oiwm.;) 249;
240, 242-4, 249, 257, 295-6n.; substantial 26, 249-50, 304,
Aristotle and modern 243-4; 306-7; see also principle, of change
idealization of 233 Charles, D. 41 n., 42 n., 49n., 57n.,
Black, M. 58-9n., 61-2 81 n., 82 n., 83 n., 90 n., 91 n.,
blood 239, 257 et passim, 281, 284, 314, 94 n., 95 n., 96n., 119n., 256n.,
316, 343; as pseudo-substance 265; 272 n., 277 n., 317n.
see also menstrual fluid (Ka-ra11~via) Charlton, W. 22, 6on., 61 n., 62 n.,
body of an animal 29, 60, 101, I 18, 123, 65 n., 81 n., l 12 n., 22 l n., 248 n.,
150, 151-4, 164, 168, 208, 249-50, 257n.
256-9, 270-2; in contrast to BODY Cherniss, H. 68n.
of an animal 277 n.; organic 250, Chiba, K. 82n., 105n.
252-3, 258-9; see also matter circle 309- 12
Bogen, J. 105 n., 26on., 277n. classification 108-9
Bolton, R. 81 n., 105 n., 322 n., 328 n., cloak 57, 87, 146 n., 197-8, 337
330 n., 338 n. Code, A. 56 n., 67 n., ro2 n., 129 n.,
Bonitz, H. 176-7, 179, 225 n. l59n., l6on., l67n., 235 n., 237n.,
boy 218-19,222,226 244n., 257n., 291 n., 292 n., 294n.,
brick 34, 38, 53, 85, 96-7, 99-100, 304 n., 324n., 342 n., 349, 350 n.
n5-16, 119, 123, 157, 189-92, Cohen, S. M. 22, 42-3, 59n., 6on.,
250,251,253,274,284-6,304 65n., l38n., l52n.,277n.
bronze lI8, 123,248,281-6,301, Colson, D. 51
· 302 n., 305-6; cube 37; hand 25 l; coming-to-be, see generation
sphere 51, 87, 198, 255 n., 268, component, see parts
287-8, 296n., 301; statue 25, composite (uvvo}.ov) 63, 76 et passim,
65-6, 248, 257-8n., 268, 285-7, 107, 129-36, 140-2, 145, 151,
296 n., 304-8, 310 n., 313, 317 154-8, 163, 168-70, 175-6, 248,
Bucephalus 135, 139 255, 273, 291-3, 296n., 302n.,
Burnyeat, M. 48, 78 n., 80 n., 104 n., 310-II, 315; composition
105 n., 254, 260 n,, 280- l, 293 n., (11vvllc111.;) 16- 19, 22, 30-5, 134,
301 n. 219n.,273
compound (of form and matter)
Callias 13- 19, 26, 33, 42, 45, 51, (uvvllcra) 15-16, 20, 30-1, l 12,
59-61,67,69, 113, 138,238,279, 175, 198, i16 n., 247 n., 248, 253 n.,
287 254-5, 261 n., 275 n., 295; see also
Carlson, G. N. · 23on. composite
Cartwright, N. 244 n. constitutive-of relation 38, 143, 157,
qategory 65, n3, 199, 228, 256 237-8, 248-9, 251 n., 254, 276
c;ause 34, 68-9, 88-9, n3, n9, 143, constructivism, see projectivism
216- 17, 232, 234, 317 n.; efficient contradiction 330 n., 331, 345 n.; see
cause 81-3, 86-8, 97, 103-4, 236, also Principle, of
249-50, 262-4, 268n.; final 83, Non-Contradiction
100-1, 218-20, see also teleology; Cooper,JohnM. 67n.,92n.
first 216; formal 46, 50, 81; Coppock, J. 58n., 71 n.
material Sr, 2II, 303-5 Coriscus 25, 140-2, 156 n.
General Index 375
count nouns, see sortal terms enmattered H1yo1 308- II , 3 r 3- I4,
craftsman 249 316-17
Entrapment 21-4, 38-40
Dancy, R. 3son. essence ( rb ri 1/v clvai) 20, 39, 49,
definability 17. 147-9, 161, 292 et 79,81,85,86n.,88,90, 120, 147,
passim, 308, 316 151, 156n., 196-7, 206, 224,
definition 33, 44-5, 49, 57, 80, 82 n., 229-36,238,240,242-4,250,253,
86-7, 91, 93, rr5, 123, 130 n., 141, 291 -2, 297, 299, 330 n., 337-8,
159,t68, 183,200,216,220,264, 342-4, 346-7
292-5, 298-9, 291 et passim, 308, ethics 241
3r5, 340; contextual 296-7; from existential import 232-4, 243-4
addition 295-6; of composite 159,
310-17; of form 49, n5, 159, fallacy 334, 336; ambiguity 336; beg
310-17; of matter 148-9, 159; the question 334-5, 340-r, 345;
schema, see soul (definition of in equivocation 16- 19, 336, 341, 347:
De Anima) 252 n.; see also unity of non sequitur 346; see also
definition principle, of Non-contradiction
Democritus: 268-9, 273; see also female parent 249
atomism; form Ferejohn, M. 86n.
demonstration 322-6, 333, 335, 345, Fine, K. 47n., l26n., r35n., 24711.,
35 r-2; see also argument, didactic 26on., 277n.
Descartes, R. 280 finger 123, 164,283,310n.,3r2n.
Devitt 256n. fission 23- 5
dialectic, see argument flesh 2r,27,39,45,49,53,60-1,77-8,
difference 63 n., 70, 92-3; qualitative rnr, 112, r25, r32, l35n., 199, 239,
60-1. 69 250, 256 n., 259-61, 266, 273-4,
differentia 75 n., 84, 133, 294 n .. 301, 276,281,296,312n.,314,316
301-2 n. focal meaning 178, r82-4
dilemma 58-9, 61, 280 foetation 263-4, 276-7 n.
divisibility 55-6; into letters 34, 35, form 15 et passim, 49-50, 66-8, 86,
38, 40, rr2, u6, 134 n., 309; of II2n., 113, II6, I2I, 125, 129-31,
syllables 35, II2, II6, 130, 132, r47 et passim, 174-5, 197-200,
309, 3II-l2 2 r3, 280 et passim, 215, 2r9-20,
division 82n., I08-9, 120-4, 126, 128. 224 n., 225-8, 247-8, 264, 268,
132, 310; principle of 134-6 291, 293, 296 n., 302 n.
doctor 87, 178-9, 188-9, 184, l92ff. asactuality 86n.,88-90, 157, l6r,
Driscoll, J. 68 n., 244 n. 169-70, 174-6, 218, 222,
LJvvaµu;, see potentiality 254-255 n.
dualism 280; quasi-dualism 280 constitutive 33-4, 49, 56, 62-4, 76,
107-8, 125, IOI, 143, 199, 248-9,
eclipse 354 252, 254, 268-7r,273, 277
Edwards, J. 58n. in contrast to form-analogue 264-7,
element 16, 22-30, 38-9, 56, 61, 62 n., 269-71
68, IOI-2, III n., rr2-13, 1I6-17, 'Democritean' concept of 268-9
124-8, 132, 143, 147, 198, 248n., as end 219, 222-4
259-63, 268n., 272-6, 31rn. formal 'aspect' 132, 146-8, 157-8
embryology: 67, 208, 216, 276-7; view human 138, 218- 19, 316
of matter suggested by Aristotle's individual form 19-21, 30, 51, 66-7,
embryology 208, 276 II7, 147n., 161, 226, 255, 287, 289
Empedocles 220, 268n., 273 indivisible 45, 56
end, see teleology internal/external principle of
General Index
form (cont.): Gill,M.L. l3,56n.,57n.,62n.,63n.,

behaviour 86n., 97, 167, 186-9, 66n., 68n., 71n.,94n., 102 n.,
221, 264-7, 270-3 105n., 125n., 129n., 131n.,164n.,
and matter, see matter; unity 228n., 248 n., 254n., 274 n., 275 n.,
obscure and nameless, see privation, 277 n., 302 n., 307 n., 312 n.
privative forms goal-directed, see teleology
particular 52, 63 n., 65 n., 68-9, 91, goatstag 337
IOI Goodman 231 n.
Platonic 120, 242 Gotthelf, A. 105 n.
as principle n6 grammar 55, 337
privative, see privation Grice, P. 103 n., 232
proper 69, 130 growth 26, 39
as property 9on.
relative 227 n. habitual proposition 230- l, 240-3
of species 147n., 156n., 161 Hamlyn, 281 n.
and substance, see constitutive-of hand 28-9, 77, 251, 265-6, 268n., 276,
relation 284
substantial form 56, 69, 97-8, Hartman, E. 57n., 67n., 68n., 69n.,
120-1, 124, 128, 129-34, 147, 159, 254 n., 280 n.
161, 176, 182, 254, 261, 305; of Haslanger,S. 57n., 121n., 122n.,
living animal 206, 268; see also 126n., 27on., 284 n.
priority heap (awpbc;;) 33, 37-8, 56, 87, II2,
supervenient 253, 257, 261, 267-8, 129-30,'136-7, 167, 168 n., 169,
277 175, 304-5
as rbt'ic n 67, 70 heat 56n., 249, 258-9, 263-5, 272;
universal 17, 19-21, 50, 131, 160 'foreign' (<UJ.6rpwr;;) and intrinsic
see also constitutive-of relation; 262-3; as 'form-analogue' 264
matter; priority; pseudo-substances Hegel, G. W. F. 204, 350
Frede, M. 20, 56n., 67n., 68n., 69n., Heidegger, M. 215n.
86n., 89 n., 92 n., 93 n., 95 n., Heinaman, R. 92 n.
96n., 103n., 105n., 107n., 134n., Heraclitus 333, 342
22l n., 251 n., 294 n., 310n., 313 n., Hicks, 267n.
315n. holism, substantial 107 et passim; see
Freeland, C. 257 n. also parts; wholes
Frege, G. 41, 46, 78n., 144n., 167n. homonymy 28-9, 31, 39-40, 108,
Freudenthal, 56n. 123-4, 251, 253-4, 256, 257 n.,
function 229-30, 236, 239, 25 I, 256, 268 n., 283-4; see also Paronymy
268n., 272 Thesis
functionalism 103, 235, 253-4 horse 55, 57, 176
Furth, M. 13, 38, 61n.,62n., 105 n., house 24-5, 31, 34, 38, 56, 81, 83-5,
131n.,248n., 293n. 94-9, JOO- l, 104, IIO, II5- 16,
fusion 23-5. 123, 157, 182-3, 185, 189-92,
250-1, 265, 284-6, 304
generation 64, 66, 160, 219, 249, 257, Hume, D. 41,49
275 n., 293, 301, 304 Hyeon, W.-J. 105 n.
genus 42-4,55,75n.,91, 110, 133, hylomorphism 14, 20 n., 28, 36n., 37,
134-5, 139, 145, 239, 294n., 129-30, 279 et passim, 296 n.;
301-2 n., 315; as matter 159-60, hylomorphic model 20 n., 37
289, 301, 302 n., 337-8; see also
enmattered ).byo1; 'being' as 174; ice 263 n., 265-6, 272
generic proposition 229 et passim identity 14, 17-18,21-2,27,30,32,41
geometry 241 -2, 309- IO et passim, 108, I 13, 119, 120- r,
General Index 377
124-8, 1.33, 142-56, 210, 254et Identity 14-15, 17, 146, l47n.
passim, 292 n.; basic 42, 50, Lenin, V. I. 350
143-4, 151; dependence rn8, Lennox, J. 71 n.
I 19-28, 158 n.; of form with matter Lesher, J. 5 1, 69 n.
45-6, 76-7, 97, 146-9, 154, 254 et letters, see divisibility
passim; of form with thing 45, Levin, S. 244 n.
49-50, 150, 285; of matter with Lewis,D. II0-11n.
thing 45, 146-7, 150, 154, 253-4, Lewis, F. Bon., 103 n., 105 n., 121 n.,
257 et passim; numerical 28, 32, 125n., lJin., 135n., l46n., 152n.,
43-5, 57, 67, 70; qualitative; l59n., 164n., 237n., 24on., 247n.,
simple 143; specific 58; 251 n., 252 n., 253 n., 255 n., 256n.,
undermined 50 277 n., 292 n.
Iliad 136 n., 295 line 309-12
~magination 127 lintel 198-9, 198, 256 n., 268 n.
mdividual (sKwnov) 55 et passim, Lloyd, A. C. 49, 67n., 68n., 92n.
136, 152, 168-9, 197' 210, 221, Locke, J. 49, 165n., 197, 286
224-7,242,283,311,312; logical vs. physical inquiry 292-3, 301-
privileged, see unity, substantial, 3, 308, 317
integration principles of London Group, see Burnyeat, M.
individuation 47, 57 et passim, 69, Loux, M. 58n., 59n., 67n., 68n., 69n.
138-9, 152, 168-9, 234, 238-9,
242-3, 306 McTaggart, J. McT. E. 47
induction 353-4 male parent 249, 263, 270
inherence 55-6, 139n., 196, 296n. man 55, 295, 324,330 n., 331, 336-7,
intelligibility 292 n., 324-5, 347-51 340-7
Irwin,T. 67n.,68n.,92n.,93n., Mann, W. 129 n.
105 n., 133 n., l47n., 148n., 156n., mass terms 234, 242, 281
219 n., 280, 282-3, 322 n., 327 n., materialism 1'03, 280; compositional
328 n., 335 n., 342 n., 345, 346n., 280; reductive 79, 104, 280
348n. mathematics 178, 216 n., 241-2, 302 n.,
Iwata, Y. 105 n. 312 n.
matter 13 et passim, 48, 59-60, 107-9,
Jones, Barrington 257n. u6, 120-1n.,123, 129-34, 147 et
Jones, M. H. 129n., 304n. passim, 174-5, 187, 189, 191-2,
197-201, 208, 2n, 213, 219n.,
Kahn, C. 147 n. 236-7, 247 et passim, 291, 293,
Kant, I. 46-8 302n.,313
Keyt, D. 271 n. 'before' matter (cc; oli) 64, 147 n.,
Kim,J. 248n. 257-61' 301, 304
King, H. 62 n. biological 253
Kirwan, C. 42, 327 concurrent matter 257-61, 272-6
knowledge (1:mcrrifJU1) 329-30, 334-5, construction of, see mixture
342, 346-7, 351-3 continuityof 43,61-5,70,95n.
Kosman, A. 78n.,8on., 104n., 105n., different conceptions of 131 n.,
l2ln., l33n., l47n.,165n., 167n., l47n., 155
221, 254 n., 256, 269, 275 n. in essence 92, 253 n., 301
Kratzer, A. 23on. as u1wKeipcvov 237, 257, 304-8
indeterminate 314
Lawrence, G. 13 individuation and 59-63, 238
Lear, J. 68 n., 33rn., 345 n. intelligible 91
Lee, H. 263 n. material 'aspect' 131, 146, 148-9,
Leibniz. G. T7-r8: T.eihni7'~ T.~w nf r.;:'7-R
General Index
matter (cont.): 264, 268n.

multi-step nature of model for 16, necessity 81 n., 229, 250, 341, 344, 348
l47n., 164, 248-9 negation 340
not Cartesian 237--,8 Newton, I. 46-7
not rb& rt 48, 68, 305 nominalism 127
persistence of 23-7, 238-9, 257, Normore, C. 58 n.
273-4, 302 n., 304-6 number 33, 41, 125
as potential 76, 88-90, 93, 96-8, Nussbaum, M. C. I05n., 148n.
161, 169-70, 174-5, 189-92,
219 n., 221, 237-9, 242, 250, O'Connor, D. J. 202
254n., 255, 273-7 Okrent, M. 228 n.
prime matter 27-8, 47, 197, 248 oneness, see unity
proper matter 69, 85, 130, 237, 250, opposites 123, 303-4, 333
255, 302 n. Owen, G. E. L. 56n., 224n., 291 n.,
proximate matter 16, 22, 28, 3 l-2, 292 n., 294 n., 327
35, 36n., 64, 77-8, 131n.,133, Owens, J. 68 n.
147 n., 148-9, 162, 164 n., 208, >;f-
248, 250-4, 259-61, 267-72, pale 140, 247, 249, 343-4

281-3 et passim Parmenides 63
relation of form to l 4-17, 63, 75 et Paronymy Thesis 308
passim, 154-5, 191-2, 197-9, 248, participation 114-15, u8, 120, 124,
252-3, 276-7, 286 198
relative 227 n. particular 56, 138, 221n.,234, 243
remote 147n., 260 parts 26, 33-4, 36-40, 53, 56, IOI, I04,
seat of activity 2u !07, !IO-II, 114, 120, 122-3, 127
top-down view of 250 et passim, 213, 238-9, 255 et
see also form; identity, of matter with passim, 308, 3IOn., 339-40; of
thing; parts; unity, material animals 122-3, 261-8, 312n.;
Matthews, G. B. 247n. improper 151; logical 3rn, 313;
Mayr, E. 234n. material 3IO; non-uniform parts
medicine 243-4 !02, 249, 257 n., 260, 266, 268,
Megarian philosophers 208 271-4, 276-7, 312 n.; posterior
menstrual fluid (Kcm1p~vm) 163, 208, 142; prior 309-II; proper 130;
258, 263 simple 277; spatially determined
methodology 327-8 parts of a thing 261, 261-2 n., 268,
Mignucci, M. 241 n. 274, 277, 281; uniform parts IOI,
Migration, see matter, persistence of 249, 26o-1, 261-2 n., 266, 271 -4,
Miller, F. D. 146n. 276,312n.
mixture (pii;1;) 258n., 273-6 Patzig,G. 67n.,68n.,92n.,93n.,
modality 15, l39n.,231,341,347 294n.
Modrak, D. 69n., 16on. Peck, A. 254 n., 258-9, 264
Moore, G. E~ 47 person 286
Moravcsik, J.M. E. I03 n., 235 n., philosophy, first 213, 353
238n. place 46-7, 62, 274n.
more and less 238 Plato(Platonism) 42,44,51, 112n.,
Morning Star and Evening Star 43, 65 n4-15, 120, 169, 198-200, 211,
Moser, P. 51 234, 242-3, 267 n.; on substance
motion, see change 114, 120, 211, 234
Mourelatos, A. 274 n. pneuma ( nvcv1w) 56 n., 249; heat and
motions in 56 n., 249
natural kind, see genus politics 240
nature 97, II3, 229, 238, 240, 242-4, possibility 18, 173, 177, 179, 191,
General Index 379
215-16, 236n., 237, 343 216, 217n., 220
posterior 64, 121-4, 139n., 140-2, privation 226, 257, 275 n., 303-4;
158, 169,209,218,311 privative (nameless and obscure)
potentiality (<>vvaµ1,) 28-9, 68, 75 et forms 304-6, 313
passim, HY], II2, u4, n9-26, process 94-6, 190, 202-3, 219, 251.
125 n., 128, 133, 156, 161-9, 173 et 276
passim, 195 et passim, 215 et product 25, 202, 251, 257n., 273, 301,
passim, 229, 231-2, 235-7, 239, 303-4, 306
242-3, 250-1, 254n., 274, 293, projectivism Son., 256-7, 275n.
310,317n.;degreesof 163-5, property 44, 49, 57, 85-90, rn7-8.
188-9; entails homonymy 123; not 116-18, 123,225,227,232,274,
Ll1!vaµ1, 178 et passim, 196; 342 n.; see also soul
ontological dependence on proprium, see property
actuality 222-6; primitive 233, pseudo-substances 265-6; see also
236; relative 227 n.; uses of form, in contrast to also
179-80, 185-7, 192 form-analogue
power 56n., 173 et passim, 190, 196, psychology 242,271
202-9,211,213,223-4,231,234, Putnam, H. rn3 n.
242-3, 250, 258 n., 268 n., 274; Pythagorean theorem 208
active powers 176, 179, 185-7,
190-1, 195 et passim, 250; passive quality 17,174,293,342n.
powers 250- 1; see also potentiality quantity 174
predication 48, 50-3, 55-7, 85, 1I9, realism 127, 321
139, 143, 145, 152 n., 160, 195-7, reduction 79, rn3 n., rn4, 173, 237; see
199, 210, 295; metaphysical 144, also materialism, reductive
159, 247-8, 254; linguistic 247 n. refutation, see argument, elenchus
prime matter, see matter, prime relation 49, 78, 98-9, rn7-8, l 13-14,
prime mover 216- 17 I 17-18, 124, 143-6, 152 n., 161,
principle 65, 112-13, 124-5, 135-8, 227n.
218, 256, 3 f7 n., 321-3, 327; Robinson, H. 280
internal/external principle of Rorty, R. 254 n.
change 177, 186-9, 221, 236, Ross,W.D. 91n.,132n.,151n.,
263-7, 273; of Non-contradiction T?6-7, 179, 265 n., 266 n., 275 n ..
321 et passim; of Sensible 327
Substance 159-60; of Specific rudder 284 n.
Unity 141 -3, 166; of Substantial Russell, B. A. W. 144n., 247n.
Composition 168; of the Unity of a Ryle, G. 300, 303, 306
Human Being 168; of the Unity of
Socrates 168-9; see also heat, sameness, see identity
'foreign' and intrinsic; unity, saw 284
substantial, integration principles Scaltsas, T. 57 n., 8011., 89 n., rn5 n ..
of II2n., 1I3n., n5n., 122n., 126n.,
priority 50, 79, 92-3, rn3-4, 121-2, r58 n., 227 n., 228 n., 244 n., 248 n.,
129-33, 138-40, 148-50,, 161' 274 n., 275 n.
166-70, 175, 186-7, 202, 204, 207, Schwegler, A. 255 n., 265 n.
215-18,226;inaccount 79,86,97, science 244, 321-3, 325, 327-9, 332,
139-40, 158, 202, 216, 219-20, 337, 351 -4; in relation to peirastic
236n., 282, 309; of form 147, 169, argument 351
220; ontological dependence 55, seed 97, 218, 219n., 220, 224n., 226,
64, 97, 121, 175, 216-20, 222, 258n., 263
225 n., 309-12 n., 314; of soul to seeing 203-5, 208, 2II,215-19, 223,
human being 206, 216, 28<;; in time 229, 296n.
General Index
Sellars, W. 67 n., 80 n., 92 n., 121 n., structural properties l 18;
254n., 255-7, 269, 275 n. structural universals l 10
semantics 244 subject, see substance, substratum
semen, see seed substance (ovaia) 15, 20, 25, 30, 49,
sense organs 267-72 55-6, 75 et passim, 107 et passim,
set 234 195-200, 207-8, 211, 221-2, 243,
sexual reproduction 212, 246, 263 291-2, 310-18, 33on., 344n.
shape 84, 87, 268, 282, 287, 296n., as activity 195 et passim
305n. animal as paradigm 206, 227
Shields, C. 280 biological 295-6
signification 331 et passim; empirical as bundle n4, 120
question 352 endurance of 95, 98, 223, 234
simple bodies, see elements immaterial 175 '
Smith, J. A. 68n. individual substances 126, 197, 247
snubness (aiµoT'l,;) 295-301, 3 IO n., particular 79, 199
314,316-17 per selKaO' w!Ti1 199-200, 295-300
Socrates 13-19,26,33,42,44-5,51, perishable 223 n.
57,59-61,66,67,69,76-7,80, primary 292, 294, 316
II3, 117-18, 130, 135, 137-9, sensible 175
140-2, 156-7,238,246,248-9, separate 49, 127, 291
257,280,283,286-7 as subject 126 n., 196-200
Sorabji, R. 280 substratum 30, 120, 123, 127-8, 197,
sorta! terms 49-50, 242, 279, 282, 305 263, 293, 296, 307 n.; see also
soul (lf!vx~) 84, 85, rr8, 150, 206-7, 249 matter, as i!1wKdµevov
et passim, 279 et passim; attribute Ti CaTI 291
theory 280; definition of soul in De . rbiie Tt 48-9, 67-70, 15 l n., 291,
Anima 252, 256; distributive view 294n.
of soul 267; generative soul 259; see also activity; essence; form,
as harmonia 122, 283; human soul substantial; inherence;
218-19; individual, particular soul pseudo-substances; unity,
30, 282-3; locomotive 249, 276; substantial
nutritive soul 249, 259, 263, 271, supervenience 44-6, roo, 248, 253-4,
276; sensitive soul 249, 252-3, 268-70, 285-6; see also
269,276 predication
space 58,62 sweet 343-4
species 15, 17-18,28n.,30,32,44,55, syllogism 82 n., 84-5, 242, 331
57-8, 64, 65 n., 67-8, 91, IIO, 122, synonymy 33 r
131, 134-5, 141-3, l56n., 2IO-ll,
218, 224-8, 230-5, 238-40, 243,
294 n., 301; defined 141; teleology 81, 84-6, 90, 94-105,
dependence on Individuals 225, 218-19, 222-8, 236, 250, 252-3;
228, 239; eternality of 227; see also immanent 105
form; set; token Teller 248 n.
Speusippus 255 n. Theseus,,;;hip of 20-1, 24, 239
sphere, see bronze, sphere thing 31-2, 247 et passim; see also parts
statue, see bronze, statue thunder 81-2
Stoics 181 time 32-4, 46-7, 59, 62, 66, 70- I, 75,
Stout, R. ro5 n. 216-17
stripping away (<'uf>aipwff;) 251 token 56,224-6, 234
structure 34-5, 40, 204, 219 n., 229, Tredennick 327 n.
232-3, 239, 287-8, 296 n.; tropes II5, n7-18, 213; trope-overlap
General Index
argument 117- 18, 127 universal 56, 78, 91, 93, 97, 131 n.,
Tselemanis, P. rn5 n. 221 n., 229, 234; taken as 68, 78,
type, see toker> 99
'usually' or 'for the most part' 235,
unity (or 'being one') 22, 33-5, 41 et 240-1
passim, 76 et passim, rn7-8, 129 et
passim, 200, 267, 294n. vegetable 332-3
accidental 25, 56, 109, n5, 133, 136, vinegar 250
145, 2!0 vitalism 79, I04
by analogy 60, 65 n.
artificial rn9 water 250
basic 153-4 Waterlow, S. 105n., 266n., 267n.
of consciousness 271 Wedin, M. 92 n.
by physical connection 109 white 65, u5, II8, 298n.; man l46n.,
of definition n9-26, 294-5 295, 297
explanatory vs. reductive models White, N. l38n., l46n.
79-94 Whiting, J. 43, 49n., 6on., 65n., 68n.,
formal 42-3, II5, l3on., 133, 7on., 92 n., 93 n.
168-70 wholes l IO- 13, II5- 16, u9, 124-5,
generic 42-3, 57-60, 64, 65 n., l 15, 127 et passim, 213, 255, 269,
133, 136, 139, 145 308-9; mereological 36-40, 100,
individual 2!0 Ill, 267, 308- l l; see also holism,
by juxtaposition 109, r l 5 substantial; parts
material 133, 136, 164 Wiggins, D. 286n.
in movement 135, 136 Williams, B. 279 ff., 285
numerical 34, 41-4, 55-6, 61, 64, Williams, C. 248n.
70-1, 139 wine 250, 343
by physical connection 109 Witt,C. 67n.,9on.,92n.,105n.,
in relation to time 76 et passim 122 n., l29n., 162 n., 221 n., 236n.
specific, see species Wittgenstein, L. 213
substantial 37, 48, 56, 109 et passim, Woods, M. 238 n.
lJ4, II8-21, 126, r29etpassim,
196, 205, 254 et passim, 275 n.; Yablo, S. l29n.
integration principles of 136-45, Young, C. 267n.
157-8, 160-5, 169-70
of wholes 114-15, u9, 127 Zeller, E. 68
see also aggregate; wholes Zeno 94

Unity, Identity, and Explanation in Aristotle's Metaphysics

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

Unity, Identity, and Explanation in Aristotle's Metaphysics

Hochgeladen von

Copyright:

Verfügbare Formate

Unity, Identity, and

CLARENDON PRESS · OXFORD

Oxford is a trade mark of Oxford University Press

Published in the United States

All right.s reserved. No part of this publication may be reproduced,

British Library Cataloguing in Publication Data

Library of Congress Cataloging-in-Publication Data

Typeset by Best-set Typesetter Ltd, Hong Kong

I. The Identity of Composite Substance 11

I. A Puzzle Concerning Matter and Form 13

II. The Unity of Composite Substance 73

III. The Potential and the Actual 171

7. Aristotle's Notion of Potentiality in Metaphysics e 173

IV. Matter and Form 245

V. Principles of Aristotle's Metaphysics 319

ROBERT BOLTON is Professor of Philosophy at Rutgers University, and

Prevailing movements. in contemporary metaphysics have fuelled

stood as active being. Kosman considers Aristotle's distinction in De

shows that in the case of organisms the form cannot be instantiated in

Montgomery Furth has written, 1 'given a suitable pair of individua,ls

©Kit Fine 1994

will be equally good grounds for rejecting the other's premisses. We

There is an obvious challenge to the logic of the argument, which

4 In a rigorous formulation of the argument we could substitute a notion of

III. COMMON FORM

The common form assumption states that it is necessary that cospecific

believed in universal forms, there would be no difficulty in supposing

Let us consider whether the puzzle might be solved by rejecting

strengthening, which we may call (Material) Individuation, says that

6 There is also a principle of individuation corresponding to Strong Entrap-

We come now to the last and most cogent putative instance of

12 In Met. H 5, 1045"3-6, Aristotle writes: 'If from a corpse is produced an

to give up the traditional interpretation, especially in its relevantly

13 One might adopt a radical position according to which elemental or other

interpretation of substantial change. But even without a substratum

I now wish to consider the question of whe.ther the puzzle can be

VI. THE MEREOLOGICAL PUZZLE

Let us return to those solutions to our puzzle which require that we

Metaphysics H. 3, 1043b5, 'the syllable does not consist of the letters

Variation to Migration can be resisted.

19 Above, p. 26. The passage has been used in support of Variation by

Anscombe (r953: 83) and by Hartman (I97T 59-60).

requiring what might seem to be an arbitrary distinction between the

are liable to assume that he is talking about our numerical identity.

But is Aristotle here saying what it is for a substance identified in

That is not because there is something spooky, what Aristotle

What is an Aristotelian individual, and what makes the individual

©Mary Louise Gill r994

7 I thank Paul Coppock for posing the dilemma to me.

9 Cf. Black (1952) in Loux (1976: 256).

11 This interpretation is proposed by Charlton (1972: 244-5), who finds support

Whiting (1986: 360).

13 Cf. Met. I. I, 1052a19-21.

This line of questioning either continues indefinitely, yielding at each

15 See Met. a. 2, 9943 1-h6.

pomtmg out that their bodies do not compose a single continuous

By treating matter within a composite as a subject in which sub-

tinuity. 19 As in Z. 8, matter appears to individuate in the weak sense,

(1986: 365-9). Aristotle's treatment of the individuation of changes in Phys. V. 4

indivisible form, the entire track (called by Aristotle the £v (~)

22 I have discus~d Phys. V. 4 in more detail in Gill (1984: 14-21).

on account of their matter but are the same in their indivisible

"13-18; cf. z. 3, w29"27-30: see also GC I. 3. 318°32-3.