Beruflich Dokumente
Kultur Dokumente
Explanation in
Aristotle's Metaphysics
Edited by
T. SCALTSAS, D. CHARLES,
and
M. L. GILL
©the several contributors and in this volume Oxford University Press 1994
Notes on Contributors ix
Introduction 1
THEODORE SCALTSAS, DAVID CHARLES,
AND MARY LOUISE GILL
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
Bibliography 355
Index Locorum 363
General Index 375
NOTES ON CONTRIBUTORS
A Puzzle Concerning
Matter and Form
KIT FINE
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
II. EQUIVOCATION
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
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.
A Puzzle Concerning Matter and Form 21
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
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
A Puzzle Concerning Matter and Form 25
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
V. COMPOSITION
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
A Puzzle Concerning Matter and Form 33
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
A Puzzle Concerning Matter and Form 35
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
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~\
A Puzzle Concerning Matter and Form 37
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
18 And hence the argument in Furth's book (1978: r8o, lines 15-21) from
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
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
Aristotle on Identity 47
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
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',
Aristotle on Identity 53
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
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
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
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-
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
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
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
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).
Individuals and 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
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
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
'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
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
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
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.
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
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
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-
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
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
(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
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
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
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
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
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
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
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).
II6 Theodore Scaltsas
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
27 It was there assumed that the form of 'pine tree' of this tree also exists in
other pine trees.
II8 Theodore Scaltsas
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
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
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
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
: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).
Substantial Ho/ism 123
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.
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).
124 Theodore Scaltsas
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.
126 Theodore Scaltsas
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
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
I. INTRODUCTION
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.
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
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.
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.
Parts, Compounds, and Substantial Unity 137
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
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
Parts, Compounds, and Substantial Unity 147
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
29 T. Irwin sometimes suggests this view; e.g. Irwin (1988: 251-2). It also
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
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,
Parts, Compounds, and Substantial Unity 153
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
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).
Parts, Compounds, and Substantial Unity 155
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.
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.
Parts, Compounds, and Substantial Unity 159
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
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::>
Parts, Compounds, and Substantial Unity 161
(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).
Parts, Compounds, and Substantial Unity
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-
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
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
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
Potentiality in Metaphysics 177
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
Potentiality in Metaphysics 179
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:
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.
Potentiality in Metaphysics 181
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
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
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
Potentiality in Metaphysics
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
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-
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
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
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.
The Activity of Being in Metaphysics 203
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
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-
The Activity of Being in Metaphysics 207
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
~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
The Activity of Being in Metaphysics 209
. 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'·-•·'<·•·'~
.. ..
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
The Activity of Being in Metaphysics 211
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~............................................. ······· ·················· ··· ..
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-
The Activity of Being in Metaphysics 213
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
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
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
The Priority of Actuality 219
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
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)
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
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
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
The Priority of Actuality 227
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
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
3 I owe this argument to Joseph Almog, though he is not responsible for this
application.
Essences, Powers, Generic Propositions 233
'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
234 Julius Moravcsik
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).
Essences, Powers, Generic Propositions 235
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.
5 For more on functionalism and Aristotle see Code and Moravcsik (1992).
Julius Moravcsik
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
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.
Essences, Powers, Generic Propositions 241
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
S. Levin, and T. Scaltsas; they are not responsible for the results.
IV
Matter and Form
II
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
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
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
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)
A Thing and its Matter 255
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
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
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.
A Thing and its Matter 259
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)
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)
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,
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.
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.
A Thing and its Matter
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
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.
A Thing and its Matter
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
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.
A Thing and its Matter 271
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.
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.).
A Thing and its Matter 273
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.
A Thing and its Matter 275
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
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
A Thing and its Matter 277
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
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
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
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.
Human Being and Individual Soul
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
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
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.
The Definition of Generated Composites 295
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
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).
The Definition of Generated Composites 297
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
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.,.
The Definition of Generated Composites 299
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-- ........... ,_
300 Michael Ferejohn
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.
The Definition of Generated Composites 301
(Burnyeat et al. 1979: 60, there referred to as 'b'), this cannot be the infamous
302 Michael Ferejohn
'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
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
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.
The Definition of Generated Composites
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.
306 Michael Ferejohn
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.
The Definition of Generated Composites
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.
310 Michael Ferejohn
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
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
The Definition of Generated Composites
Aristotle's Conception
of Metaphysics as a Science
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
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.
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 re- futation 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
Metaphysics as a Science
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
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
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
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.
12 What argument in particular does Aristotle forgo here to avoid begging the
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
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)
Metaphysics as a Science 337
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.
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
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
Metaphysics as a Science 343
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
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
Metaphysics as a Science 345
(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
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
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.
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
Philosophy, 64: 25-46.
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
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