Documente Academic
Documente Profesional
Documente Cultură
in Aristotle
CAMBRIDGE
UNIVERSITY PRESS
for James
Contents
Acknowledgments
page ix
Introduction
I The Separation of Platonic Forms
II Referential Opacity in Aristotle
III A Theory of Substance
IV Substance and Aristotle's Epistemology
V The Separation of Substance
VI Substance and Teleology
1
5
21
40
63
83
100
Bibliography
Index
123
129
vn
Acknowledgments
Introduction
In Chapter I, a chapter that lays the groundwork for Aristotle's theory, I examine Aristotle's criticism of Plato for separating the Forms,
arguing that by 'separation' Aristotle has in mind numerical distinctness,
and I cite passages to show that he believes that the numerical distinctness of the Forms from sensible objects causes insoluble metaphysical
and epistemological problems. To see Aristotle's theory as a response
to Plato inevitably raises questions about the accuracy of Aristotle's
presentation of Plato's Theory. Even though in Chapter I and elsewhere
I do from time to time sketch, in a very broad way, various interpretations of Plato's own views, in a sense the question is irrelevant to my
project - if Aristotle's theory is a response to what he took Plato to be
saying, the impact on his own views will be the same regardless of his
skill as an interpreter. Nevertheless I must admit that upon reading
Plato and Aristotle, I find the view that Aristotle misunderstood or failed
to appreciate Plato's Theory to be largely false. Rather, I agree with
those3 who say that it is just inherently implausible that one of the finest
philosophers who ever lived should, after twenty years in Plato's company, have failed to grasp his views and the issues that underlie them.
But, as I have said, the cogency of my project does not depend on
agreement with this claim.
Having discussed Aristotle's criticisms of Plato in Chapter I, I turn
to Aristotle's own views in the subsequent chapters. If Aristotle is to
say, as his criticism of Plato makes it plausible that he should, that
substances are not numerically distinct from sensible objects, one might
reasonably expect that he holds them to be identical with sensible objects. But it has to be remembered that, like Plato, Aristotle wants
substances to be unchanging if they are to be epistemologically fundamental. One might suppose that if substances are forms and if forms
are universals this requirement could be met. But then again Aristotle
wants substances to be ontologically fundamental as well, a fact that
seems to argue for their being objects, not properties. In Chapter II I
discuss the problem of referential opacity in Aristotle, claiming that
Aristotle uses a distinction between numerical sameness and identity to
address many sorts of metaphysical problems, and in Chapter III I argue
that this distinction is the key to Aristotle's theory of substance. What
I hold is that substances are for Aristotle specimens of natural kinds,
where specimens, as particular forms lacking the accidents introduced
by matter, are numerically the same as sensible objects yet not identical
with them. While specimens of kinds are not eternal, within a kind they
are indistinguishable from one another, with the result that unlike sensible objects they are knowable.
3 One of them is Russell Dancy, to whom, as a result of a conversation in the summer
of 1988, this description of the stance and my confidence in its reasonableness are in
part owed.
INTRODUCTION
(i)
Yet even as I have been writing, others have also, and the two most
recent accounts of substance in Aristotle, Michael Loux's Primary 'Ousiai' (Cornell University Press, 1991) and Frank Lewis's Substance and
Predication in Aristotle (Cambridge University Press, 1991), came after
I had essentially completed this manuscript. As it happens, both Loux
and Lewis argue for forms as universals while my argument requires
them to be particulars, yet I have not attempted to provide an exhaustive
examination of all the texts that bear on this long-standing controversy.
Instead I have tried to consider a somewhat different cluster of issues
in such a way that they illuminate one another. For what I want to
contend is that, if read as criticism and revision of Plato in the way I
propose, Aristotle has a coherent view which, even if different from our
own, is nevertheless a philosophically challenging response to the experienced world.
I
The Separation of Platonic Forms
And Socrates gave the impulse to this theory, as we said before, by means
of his definitions, but he did not separate them from the particulars; and
in this he thought rightly, in not separating them. This is plain from the
results; for without the universal it is not possible to get knowledge, but
the separation is the cause of the objections that arise with regard to the
Ideas.
(Metaphysics XIII 9 1086b2-7)1
That Aristotle criticizes Plato for separating the Forms is a fact known
to every reader. However, what exactly it is that Aristotle wants to
criticize has, until recently, seldom been discussed explicitly and at
length, and indeed as exploration of the question has occurred, views
have differed considerably. It has been proposed by some interpreters
that when he criticizes Plato for separating the Forms, by 'separation'
Aristotle means their independent existence,2 that is, their capacity for
existing even if there were no sensible objects. But other interpreters
have held that he means their numerical distinctness from sensible objects,3 and some writers have thought that he means both of these.4
Thus despite the considerable importance Aristotle places on Plato's
separation of the Forms, there is disagreement about just what he is
objecting to. Moreover, besides our uncertainty about what Aristotle
meant, there is a further problem. For even as he criticizes Plato, Aristotle tells us that substances must be separate (Metaphysics VII 1
1028a34). Only after an account of Aristotle's theory of substance is
1 Unless otherwise noted, all quotations from Aristotle follow The Complete Works of
Aristotle: The Revised Oxford Translation, ed. Jonathan Barnes (Princeton, N.J.: Princeton
University Press, 1984). The page, column, and line numbers have where necessary
been corrected against the edition of the Greek text prepared by Immanuel Bekker
for the Berlin Academy, published in 1831. Emphasis follows Barnes except where
indicated.
2 Fine (1984); Hardie (1936), 73; Irwin (1977), 154.
3 Mabbott (1926). See also Morrison (1985), esp. 138-39 and 149-50.
4 Allen (1970), 131-32.
original good, for the destruction of that which is participated in involves also the
destruction of that which participates in the Idea, and is named from its participation
in it. But this is the relation of the first to the later, so that the Idea of good
is the good per se\ for this is also (they say) separable [choristen] from what
participates in it, like all other Ideas. (1217b 1-16; emphasis added)
Following Gail Fine, let us call the capacity to exist without the existence of some other thing the capacity for independent existence with
respect to that thing.6 Citing the passage just quoted, Fine argues that
when Aristotle talks of 'separation' he means a capacity for independent
existence7 and thus that in his criticism of Plato he (in the main, wrongly,
as she sees it) attributes to Plato the view that Forms are separate from
sensible objects in the sense that they are able to exist independently
of them. She says:
Aristotle is probably correct to say that at least some Forms, in some dialogues,
are separate. But he and others are incorrect to suggest that Plato, beginning
with the Phaedo, heralds separation as a new feature of Forms. On the contrary,
so far from this being the case, Plato never even says that Forms are separate;
it proves surprisingly difficult to uncover any commitment to separation; and
commitment to it emerges in unexpected ways and in unexpected cases.8
Fine understands Plato's Forms to be universals, and thus when she
denies that the independent existence of Forms is a key component of
Plato's Theory of Forms, what she is denying is that it is especially
important to Plato that the Forms be able to exist uninstantiated. But
whatever Forms are, Fine's conclusion about the role of separation is,
as she admits, very surprising. T. H. Irwin, for example, who agrees
that Forms are universals, thinks that even though Plato does not formulate it clearly and even though his arguments justify only the claim
that Forms are not defined in terms of sensible properties and are not
identical with such properties (which leaves open the possibility that
they might be identical with nonsensible properties of sensible objects),
he does in fact believe that he has established their capacity for independent existence.9
Since my concern is with the nature of separation, Irwin's conclusion
need be pursued only if it is plausible to hold that Forms are universals
6 Fine (1984), 35.
7 See Fine (1984), 33.
8 Fine (1984), 33-34. According to Fine the most likely candidates for separate Forms
are Forms of artefacts (as in the Cratylus and Republic X) since there would have been
a time when they existed which was prior to the work of human artisans, that is to say,
they would have existed in the absence of instances (76). If the account of creation in
the Timaeus is taken in such a way that, while Forms have always existed, there was a
time at which there were no sensibles, then, says Fine, many Forms are separate (79).
However, even in that case not all Forms would be so. For even before creation there
are some Forms that would have instances; traces of fire are found in the chaos, the
demiurge exemplifies justice and goodness, and so on (79).
9 See Irwin (1977), 154-55.
what is proved is that there exist some things that are not in flux. That
is to say, what has been proved is, at most, that there are things that
exist besides - in addition to - those that are in flux. Whether they
would exist even if those that are in flux did not do so is clearly another
question.
In short, I have no inclination to deny that Plato intended the Forms
to have a capacity for independent existence and indeed neither would
I deny that Aristotle thought that this was Plato's intention. But if a
capacity for independent existence is what is meant by 'separation', then
it would seem that Aristotle must attribute to Plato an argument for
the existence of Forms which is flagrantly invalid. In fact Fine offers a
way to avoid this conclusion. What she suggests is that Aristotle did not
take Plato to be arguing directly from flux to Forms capable of existing
independently but rather to be using the assumption that there is knowledge despite the existence of flux to establish only that there are Forms
which are universals, not that they are separate. Recalling that separation is taken by Fine to be a capacity for independent existence, why
then would Aristotle accuse Plato of separating the Forms? What Fine
proposes is that these universals or Forms are, in Aristotle's interpretation of Plato's ontology, the only available candidates for substances,
and substances, so Aristotle holds, are separate.12
Thus as Fine understands it, the argument for the separation of the
Forms - which, if Forms are universals and 'separation' means a capacity
for independent existence, is, as I have said, just an argument that
universals can exist uninstantiated - need not after all be invalid; it is
not so because the need for something which is not in flux is not by
itself supposed to be sufficient to establish that Forms are separate. Yet
even if the argument is not invalid, neither is it in Aristotle's view sound.
Rather what Aristotle holds, according to Fine, is that Plato errs in his
belief that Forms, being universals, can be substances.
In fact I would agree with Fine that Aristotle believes that universals
cannot be substances. Nevertheless I disagree with her understanding
of separation and in consequence with her construal of how, from the
fact of flux, Aristotle supposes Plato to have arrived at separate Forms.
Indeed the passage in Metaphysics XIII 9 from which the quotation at
the head of the chapter was taken and upon which Fine founds her
argument would not seem most naturally to support her interpretation.
Aristotle says:
as regards those who believe in the Ideas one might survey at the same time
their way of thinking and the difficulties into which they fall. For they at the
same time treat the Ideas as universal, and again as separable [choristas] and
individual. That this is not possible has been shown before. The reason why
12 Fine (1984), 51-53.
10
those who say substances are universal combined these two views in one, is that
they did not make them the same13 [autas] with sensible things. They thought
that the sensible particulars were in a state of flux and none of them remained,
but that the universal was apart [para] from these and different. And Socrates
gave the impulse to this theory, as we said before, by means of his definitions,
but he did not separate them from the particulars; and in this he thought rightly,
in not separating them. This is plain from the results; for without the universal
it is not possible to get knowledge, but the separation is the cause of the objections that arise with regard to the Ideas. His successors, treating it as necessary, if there are to be substances besides the sensible and transient substances,
that they be separable, had no others, but gave separate existence to these
universally predicated substances, so that it followed that universals and individuals were almost the same sort of thing. (1086a31-bll)
Passages such as this will be mined for another purpose in Chapter
V. At present, however, my point is that when Aristotle says that Forms
are both separate and universal, his diagnosis of Plato's error would
seem to be not, as Fine claims, that he made the Forms universals while,
if they are to be substances, they must be separate and therefore particular. Rather Aristotle's complaint is that Plato failed to make Forms
the same as sensible things (1086a36). To be sure, this claim could mean
that he failed to make them the same in kind - in other words, that he
failed to make them particulars - but since in fact Aristotle here also
says that Plato makes Forms both particular and universal, this would
be, to say the least, a peculiar complaint and, anyway, the manner in
which he continues suggests a different explanation. For if the reason
for postulating something not the same as sensibles is that the latter
are in flux, this reason tells neither for nor against the universality or
particularity of Forms - it counts only for the postulation of something
that does not have the property of being in flux. Thus their numerical
distinctness from things in flux would seem to be what is asserted.
II
Even though Plato's Forms do have the capacity to exist independently,
as I have argued, a capacity for independent existence would not seem
to be what Aristotle has in mind when he says that Plato's Forms are
separate. Rather, as I see it, the attribution of a capacity for independent
existence to the Forms is an assertion of the ontological priority they
have over the phenomena from which they are held to be separated.
But if this is so, it is necessary to establish what 'separation' does mean,
and I have already given a reason for favoring numerical sameness. In
Metaphysics I 6 987a32-blO when Aristotle gives his account of the
13 Barnes (1984) has 'identical'. My reason for preferring 'same' will become apparent
in Chapter II.
11
origin of Plato's Theory of Forms and addresses the question of separation, he says:
For, having in his youth first become familiar with Cratylus and with the Heraclitean doctrines (that all sensible things are ever in a state of flux and there
is no knowledge about them), these views [Plato] held even in later years. Socrates, however, was busying himself about ethical matters and neglecting the
world of nature as a whole but seeking the universal in these ethical matters,
and fixed thought for the first time on definitions; Plato accepted his teaching,
but held that the problem applied not to any sensible thing but to entities of
another kind - for this reason, that the common definition could not be a definition
of any sensible thing, as they were always changing. Things of this other sort, then, he
called Ideas, and sensible things, he said, were apart from these, and were all called
after these; for the multitude of things which have the same name as the Form
exist by participation in it. (emphasis added)
The same story seems to be told in Metaphysics XIII 4 1 0 7 8 b l 2 - 1 0 7 9 a 4
where Aristotle begins by saying:
The supporters of the ideal theory were led to it because they were persuaded
of the truth of the Heraclitean doctrine that all sensible things are ever passing
away, so that if knowledge or thought is to have an object, there must be some
other [heteras] and permanent entities, apart [para] from those which are sensible; for
there can be no knowledge of things which are in a state of flux. (1078bl2-17;
emphasis added)
After a discussion of Socrates, Aristotle then explains the difference
between his views and those of Plato:
For two things may be fairly ascribed by Socrates - inductive arguments and
universal definition, both of which are concerned with the starting-point of
science. But Socrates did not make the universals or the definitions exist apart
[chorista]; his successors, however, gave them separate existence, and this was
the kind of thing they called Ideas. (1078b27-32)
What are we to make of these passages? I have argued against Fine
that Eudemian Ethics I 8 does not show 'separation' to mean a capacity
for independent existence and that, because an argument from flux to
independent existence (unless taken in something like the way in which
Fine takes it) would be invalid, a capacity for independent existence is
not likely to have been what Aristotle meant. Of course, it is possible
that Plato's argument for the separation of the Forms, assuming that
Aristotle is right about the origins of the Theory, just is invalid. Nevertheless, as I have already said, the difficulty with the proposal that Plato
argues from the need for something that is not in flux to the need for
something that can exist without the existence of whatever is in flux is
that the two notions seem too obviously unconnected for fallacious
argument to occur. Besides, further and, I think, decisive evidence
12
Ill
In considering those passages in which Aristotle criticizes Plato for
separating the Forms, I have argued that, while Plato did take the Forms
to be capable of independent existence and while Aristotle thought that
he did so, such is not the basis of the latter's complaint. Rather, what
the argument from flux, the argument that Aristotle cites as the origin
of the Platonic Theory, can reasonably be supposed to show and what
Aristotle takes it to be supposed to show - is that Forms are numerically
distinct from sensibles. But if this is so, then when Aristotle objects to
13
14
15
qualification, - a procedure like that of the people who said there are gods,
but in human form. For they were positing nothing but eternal men, nor are
they making the Forms anything other than eternal sensible things, (cf. VII 16
1040b32-34)
This complaint is summarized in Nicomachean Ethics I 6 in a way that
makes the consequences clear:
And one might ask the question, what in the world they mean by 'a thing itselF,
if in man himself and in a particular man the account of man is one and the
same. For in so far as they are men, they will in no respect differ; and if this is
so, neither will there be a difference in so far as they are good. (1096a34-b3)
Thus Aristotle, beginning from their numerical distinctness, takes
Forms to be particulars and proposes that, if intelligible at all, Forms
are no more than eternal sensibles. Moreover in the Third Man Argument he seems to claim that the numerical distinctness of Forms will
lead to an infinite regress as well. Although the Third Man Argument
is no more than alluded to in Metaphysics I 9 (990bl7), there is a brief
discussion in the Sophistical Refutations. There Aristotle says: "Again,
there is the argument that there is a third man distinct from man and
from individual men. But 'man', and indeed every general predicate,
signifies not an individual, but some quality, or quantity or relation, or
something of that sort" (178b36-39). Even though the steps leading to
the troublesome third man are not spelled out, it seems evident that
Aristotle believes that for Plato 'man' signifies a particular and that this
is cited as the cause of a regress of distinct Forms. But since a lengthier
version of the Third Man Argument as given in his lost essay Peri Idem
is preserved in Alexander of Aphrodisias's commentary on the first book
of the Metaphysics, I will turn to the presentation of the Third Man
Argument in Alexander.
At 84, 22-85, 3 Alexander records the following objection to the
Theory of Forms:
If what is predicated truly of more than one thing is also [some] other thing
apart from the things of which it is predicated, being separated from them (for
this is what those who posit the Ideas think they are proving; for the reason
why, according to them, there is something, man-himself, is because 'man' is
predicated truly of particular men, who are more than one, and is other than
particular men) - but if this is so, there will be some third man. For if [the
'man'] predicated is other than those of whom it is predicated and subsists by
itself, and 'man' is predicated both of particular men and of the Idea, there
will be some third man apart from both particular men and from the Idea. And
in this way there will be still a fourth man, the one predicated of the third man
and of the Idea and of particular men, and similarly a fifth, and so on ad
infinitum.2X
21 Pages 121-22 of the Dooley translation of Alexander of Aphrodisias (1989). The
bracketed material is supplied by Dooley.
16
17
like Cleon or Socrates. Why does not one of the supporters of the Ideas produce
a definition of an Idea? It would become clear, if they tried, that what has now
been said is true. (1040a27-b4)
Aristotle's support for the claim that particulars are unknowable is
in part the alleged impossibility of counterexample: Just try to produce
a definition of Cleon or the sun or an Idea. But, of course, one reason
a definition of Cleon cannot be produced is that Cleon is transitory.
Earlier in the same chapter Aristotle has said:
there is neither definition nor demonstration of sensible individual substances,
because they have matter whose nature is such that they are capable both of
being and of not being; for which reason all the individual instances of them
are destructible. If then demonstration is of necessary truths and definition
involves knowledge, and if, just as knowledge cannot be sometimes knowledge
and sometimes ignorance, but the state which varies thus is opinion, so too
demonstration and definition cannot vary thus, but it is opinion that deals with
that which can be otherwise than as it is, clearly there can neither be definition
nor demonstration of sensible individuals. (1039b27-1040a2)
18
19
the Form of the Good on the grounds that craftsmen do not use it,
Aristotle then seems to suggest that their lack of attention to the Form
should undermine our confidence in its existence.
In the two passages just cited, Aristotle uses the separation of Forms
as grounds for doubting that knowledge of Forms, even should there
be such, would yield knowledge of the sensible world. But in Metaphysics
I 9 he also mounts a general attack on the acquisition of knowledge of
the elements of things, an attack that seems to be aimed at Forms:
And how could we learn the elements of all things? Evidently we cannot start
by knowing something before. For as he who is learning geometry, though he
may know other things before, knows none of the things with which the science
deals and about which he is to learn, so is it in all other cases. Therefore if
there is a science of all things, as some maintain, he who is learning this will
know nothing before. Yet all learning is by means of premises which are (either
all or some of them) known before, - whether the learning be by demonstration
or by definitions; for the elements of the definition must be known before and
be familiar; and learning by induction proceeds similarly. But again, if the science
is innate, it is wonderful that we are unaware of our possession of the greatest of sciences.
Again, how is one to know what all things are made of, and how is this to be
made evident? This also affords a difficulty; for there might be a conflict of
opinion, as there is about certain syllables; some say za is made out of s and d
and a, while others say it is a distinct sound and none of those that are familiar.
Further, how could we know the objects of sense without having the sense in
question? Yet we should, if the elements of which all things consist, as complex
sounds consist of their proper elements, are the same. (992b24-993al0; emphasis added; boldface indicates Barnes's emphasis)
The problems attendant to the Theory of Recollection will be considered in Chapter IV. But it is already apparent that in arguments such
as these Aristotle is contending, in effect, that in addition to all the
metaphysical problems of the Theory, the Forms cannot accomplish
what he believes to be Plato's aim, namely, to explain how there can
be knowledge of anything in a world that seems to be full of flux. Yet
even if Plato's Theory is thought by Aristotle to fail, the questions it
was intended to answer remain, and it is from within their presuppositions that Aristotle too has to wrestle with problems of universality,
particularity, ontological priority, and knowability. As he says in Metaphysics III:
If [the first principles] are universal, they will not be substances; for everything
that is common indicates not a 'this' but a 'such', but substance is a 'this'. And if we can actually posit the common predicate as a single 'this', Socrates
will be several animals - himself and man and animal, if each of these indicates
a 'this' and a single thing. - If, then, the principles are universals, these results
follow; if they are not universals but of the nature of individuals, they will not
be knowable; for the knowledge of anything is universal. (1003a715)
20
Thus it is that Aristotle also will try to say what substances or forms are
and how they have the universality and particularity he claims is necessary. Likewise he will try to explain what the relation between substances or forms and sensibles is, and how, by encountering sensible
objects, one knows substances. The examination of his views in the
chapters to come will test my contention that what separates Plato and
Aristotle is - separation.
II
Referential Opacity in Aristotle
21
22
23
is accidentally the same as the seated man, but Socrates, and not the
seated man, is essentially the same as his essence or form; the builder
is a cause of the house, and the musician is not (even if the builder is
a musician); the man, but not the pale one, becomes cultured. This list
- to which could be added '. . . is a substance', '. . . is per se [roughly,
'. . . is necessarily'] F', '. . . is a species o f . . . ' , ' . . . is the genus of. . .',
and so on, as well as \ . . moves', '. . . builds', * . . . sculpts', and all
other action verbs - includes not only the modal and propositional
attitude contexts that many philosophers besides Aristotle have thought
to be opaque, but also contexts that might seem to be paradigm cases
of the transparent and contexts involving Aristotle's technical vocabulary.
So far I have explained what is meant by 'referential opacity' and I
have described why, where identity is concerned, failures of substitutivity would seem to be problematic. But in fact whether Aristotle had
such a notion as identity is disputed. In the Topics, he himself expresses
what appears to be the principle of the indiscernibility of identicals or,
as it is often called, Leibniz's Law:
Moreover, examine them in the light of their accidents or of the things of which
they are accidents; for any accident belonging to the one must belong also to
the other, and if the one belongs to anything as an accident, so must the other
also. If in any of these respects there is a discrepancy, clearly they are not the
same. (152a33-37)
Or again:
Speaking generally, one ought to be on the look-out for any discrepancy anywhere in any sort of predicate of each term, and in the things of which they
are predicated. For all that is predicated of the one should be predicated also
of the other, and of whatever the one is a predicate, the other should be a
predicate as well. (152b25-29)
These passages and others from the Topics (for example, 132a2728) are usually taken as good evidence that, at least at the time of that
treatise, Aristotle held a standard view of identity.6 But there is also
reason to have qualms. At De Interpretatione 1121a 7-14, for example,
Aristotle says:
6 In his definition of 'numerical sameness' in Topics I 7, Aristotle distinguishes three
senses of the expression, namely, sameness where there is an alternative name or
definition ("its most literal and primary use"), second, sameness rendered by a nonessential property peculiar to that kind of thing, and third, sameness drawn from
accident, as in describing Socrates as "the creature who is sitting" (103a23-31). While
these senses may seem to foreshadow later distinctions, 152a33-37 and 152b25-29,
just quoted, make it clear that, with regard to substitutivity, no use is made of them
in the Topics. Moreover Aristotle's example of the first and most literal sense is at least
peculiar since the example (doublet and cloak) is not, as it stands, an example of
numerical sameness.
24
Of things predicated, and things they get predicated of, those which are said
accidentally, either of the same thing or of one another, will not be one. For
example, a man is white and musical, but * white' and * musical' are not one,
because they are both accidental to the same thing. And even if it is true to say
that the white is musical, * musical white' will still not be one thing; for it is
accidentally that the musical is white, and so 'white musical' will not be one.
Moreover in Sophistical Refutations 24 Aristotle clearly says that numerical sameness is not sufficient for substitutivity, a claim that raises
serious problems of interpretation. Nicholas White, for example, holds
that Aristotle's grasp of identity weakens7; Gareth Matthews argues that,
on the contrary, the restrictions Aristotle introduces show that he is
moving toward an understanding of the notion.8 What is clear is that
Aristotle thinks or comes to think that there are two varieties of numerical
sameness, essential sameness and accidental sameness {Metaphysics V 9
10l7b27-1018a9), such that substitutivity is said to hold in the first case
in certain contexts where it fails in the second. The problem therefore
is this: If numerical sameness just is identity, as elucidated by Leibniz's
Law, what can the distinction between essential and accidental sameness
come to and how can it be used to explain certain failures of substitutivity?
On the other hand, if either accidental sameness or essential sameness
(or both) is other than identity, what is their relation to identity?
II
One strategy for dealing with failures of substitutivity is that of Frege.
When we are presented with a belief context or a modal context in
which substitutivity seems to fail (for example, 'it is necessary that the
Morning Star is the Morning Star' is true but 'it is necessary that the
Morning Star is the Evening Star' is false), Frege's response is to say
that, in such sentences, 'the Morning Star' does not, despite appearances, refer to the Morning Star and 'the Evening Star' does not, despite
appearances, refer to the Evening Star. That is, in modal and belief
contexts, Frege claims, 'the Morning Star' refers not to the ordinary
referent of the expression 'the Morning Star' (namely, Venus) but rather
to the sense of that expression.9 Since 'the Morning Star' in 'it is necessary that the Morning Star is the Morning Star' does not, on Frege's
view, refer to the Morning Star but to its sense and since the sense of
'the Morning Star' is different from the sense of 'the Evening Star', it
is hardly surprising that substituting one for the other fails to preserve
7 White (1971), 177. I will not argue against White's view that, after the Topics, Aristotle
confuses 'X and Y are the same' with 'X and Y are one', where the latter is taken to
mean that both X and Kmake up a single entity (see 187). However both Miller (1973)
and Matthews (1982), 230-35, do so effectively.
8 Matthews (1982), 233.
9 See Frege (1952), 67.
25
26
27
28
and patiency in Physics III, Aristotle himself seems to offer an explanation. Here he says:
Nor is it necessary that the teacher should learn, even if to act [to teach] and
to be acted on [to learn] are one and the same, provided they are not the same
in respect of the account which states their essence (as raiment and dress), but
are the same in the sense in which the road from Thebes to Athens and the
road from Athens to Thebes are the same. . . . For it is not things which are in
any way the same that have all their attributes the same, but only those to be
which is the same. (202bl0-16; cf. 202al8-20)
However, as has already been said with regard to Dancy's presentation
of it, the problem with Aristotle's explanation of referential opacity,
namely, that only to things the same in definition do all the same properties belong, is that it seems to be nothing more than a restatement
of the difficulty; that is, there are cases in which a is apparently identical
with b where substitutivity fails. In what way is the road from Athens
to Thebes the same as the road from Thebes to Athens? In the next
section I propose an answer to this question.
Ill
We have examined several proposals for understanding referentially
opaque contexts in Aristotle, including Lewis's and Matthew's suggestion that in such cases the referents are in fact not the same. In a passage
of the sort noticed by Lewis and Matthews, Aristotle says:
For to be a man is not the same as to be unmusical. One part survives, the
other does not: what is not an opposite survives (for the man survives), but notmusical or unmusical does not survive, nor does the compound of the two,
namely the unmusical man. (Physics I 7 190al7-21; cf. Prior Analytics I 47b2934, and On Generation and Corruption I 319b25-31)
Aristotle seems to want to say that the unmusical human being's ceasing
to exist is something distinct from a human being's ceasing to be unmusical; the unmusical (that is, this case of it) does not survive and the
unmusical human being does not survive. On the other hand, that the
accidental is "obviously akin to not-being" (Metaphysics VI 2 1026b21)
and "practically a mere name" (1026bl3-14) is said to be shown by
the fact that, although unmusical Coriscus ceases to exist, there is no
process by which he does so (1026b22-24). Thus what seems to be
needed is a view that is intermediate between saying, on the one hand,
that accidental unities are entities numerically distinct from the sensible
objects that have the accidents in question and saying, on the other
hand, that they just are those sensible objects, inappropriately described. When Coriscus comes to be musical, Aristotle seems to think
there comes to exist not just an instance of the quality musicality but
29
30
(where the hyphens indicate that now 'qua' is part of a referring expression), and because 'qua' when used referentially is a filter of properties,
the result again is generality. That is, just as it is true that rectangles
have four right angles regardless of the color and age of the tables, in
other words, that for purposes of geometry two tables are, assuming
their size to be irrelevant or the same, indistinguishable, so it is as well
in other cases. Since 'Socrates-qua-musician' abstracts from all the properties of Socrates except musicality and 'Callias-qua-musician' does the
same, then even though they are, in virtue of being numerically the
same as two different sensible objects, two specimens, Socrates-quamusician and Callias-qua-musician are qualitatively indistinguishable.
In short, on the account which I am proposing, Socrates and
Socrates-qua-builder are numerically the same. They are not, however,
identical. Because not all the properties of a are true of a-qua-0, in a
move which is of crucial importance for his theory of substance, there
is a distinction to be made between numerical sameness and identity.
Postponing until later chapters all questions regarding the status of
kinds and the naturalness of any given kind, I want to call x-qua-0 a
specimen of the kind 0, and what I am suggesting is that what Aristotle
calls "accidental unities" are in fact specimens of kinds. That is to say,
what I take Aristotle to hold is that Callias is numerically the same as
the musician (and also musical Callias). Nevertheless Callias is not identical with either of these, and it is just because 'the musician' and 'musical Callias' refer not to Callias but in both cases to a specimen of the
kind musician, where a specimen of a kind has only the properties essential to members of that kind, that failures of substitutivity occur.
I have contended that 'the builder', 'Socrates seated', and other such
expressions for accidental unities are not, for Aristotle, definite descriptions that fix the reference to someone (namely, Socrates) who is
short, snub-nosed, a philosopher, and also a builder and seated. On the
other hand, I have also claimed that to deny that 'the seated one' refers
to Socrates does not have to mean that the seated one is, as Lewis
thinks, an entity numerically distinct from Socrates, even though Socrates and the seated one (where Socrates is seated) differ in properties
and so are nonidentical. To be sure, the seated one (that is to say,
Socrates-qua-one-seated) is identical with what it is to be one seated,
and this is what provides the generality required for explanation. But
from the fact that Socrates-qua-one-seated is identical with what it is
to be one seated, it of course cannot be concluded that Socrates and
what it is to be one seated are so. Likewise it does not follow that the
definition of the kind, even though it is a kind to which the sensible
object can be said to belong, is necessarily the definition of the sensible
object. If Socrates, for example, belongs to the kind (it is to be remembered that I am not committed to supposing that it is a natural
31
32
33
the musical is a man, just as we say the musical builds, because the builder
happens to be musical or the musical happens to be a builder; for here 'one
thing is another' means 'one is an accident of another*. So in the cases we have
mentioned; for when we say the man is musical and the musical is a man, or
the white is musical or the musical is white, the last two mean that both attributes
34
proaching human beings.24 But this problem does not affect the general
structure of the solution I propose. For even if contrary to Aristotle's
usual view, there were some way to know Coriscus as Coriscus and even
if the kind were not approaching thing but something so narrow that the
only approaching thing which could be a member of that kind would
be Coriscus (perhaps something like the kind thing now within sight and
approaching the viewer from the only direction from which he can now be
approached),25 still, as long as being a member of this kind is not Coriscus's essence, to know a specimen of this kind is not to know that
Coriscus is the same as this specimen.
In short, despite the peculiarity of the paradox, how my account of
accidental unities would deal with knowledge contexts is clear. Coriscus
is a specimen of the kind approaching thing, and we can know what it
is to be a thing approaching. But since the definition of what it is to
be approaching is not the definition of Coriscus, to know Coriscus is
not to know that it is he who is approaching.
35
36
Thus it would seem that Irwin too attributes to Aristotle a nonstandard concept of identity;30 Sandra Peterson may do so as well. What
Peterson argues is that Aristotle's handling of opaque contexts is close
to that of Carnap, in whose system a distinction is made between identity
of extension and identity of intension.31 For Carnap substitution in
modal contexts requires sameness in intension, that is, sameness in all
state descriptions, while substitution in propositional attitude contexts
requires both intensional equivalence and intensional isomorphism. 32
Aristotle does not make the latter distinction. Nonetheless he does claim
that essential sameness (that is, the sort of numerical sameness in which
there is sameness in definition) is required for substitutivity to hold in
modal and propositional attitude (knowledge) contexts, and that, Peterson believes, is enough to justify the comparison.
However, if my interpretation is correct, the views of Aristotle and
Carnap are not especially close. In the first place, while Carnap's individual concept Socrates - the referent of modal contexts in which
there is seeming reference to Socrates - includes that which distinguishes Socrates from Callias in all state descriptions, on Aristotle's
account, as we have seen, Socrates-qua-human being and Callias-quahuman being turn out to be qualitatively indistinguishable. Moreover,
where Carnap would consider 'the fluteplayer' and 'the housebuilder'
(in a given state description) as extensionally equivalent definite descriptions, Aristotle would, I have argued, take them to refer to specimens of different kinds. Finally even though whatever is true of Socrates
is true of that which is numerically the same as a specimen of the kind
human being or a specimen of the kind builder, to refer to Socrates-quahuman being or to the builder is not to refer to Socrates but to a
specimen of a given kind, where, as I have said, what is characteristic
of specimens is that they (unlike the sensible objects with which they
are numerically the same) have only the defining characteristics of the
kind. But if this is so, then the numerical sameness that holds between
Socrates and Socrates-qua-</> is not identity, and Aristotle's distinction
30 Hartman (1977) may be yet another example. He says: "As his use of qua suggests,
Aristotle distinguishes entities according to their descriptions; thus an entity under a
certain description is not fully identical to what would normally be considered the
same entity under a different description. In effect, Aristotle makes mutual logical
dependency a requirement for true identity . . . " (73). To distinguish between true
or full identity, on the one hand, and accidental identity (still said not to involve
numerical distinctness) (75), on the other, would seem to revise our understanding
of identity.
31 Peterson (1969), 145-50. Hussey (1983), 69-70, argues that Aristotle's treatment of
referential opacity anticipates parts of Frege's, but Hussey's account is still like Peterson's. Hussey says: "Aristotle's point is that . . . 'Leibniz's Law' need not be true
unless there is sameness 'in being' or 'in definition', on top of ordinary identity" (69).
32 See Carnap (1947), 46-59, 9-11, 100.
37
38
39
the same thing, are obviously answered in the same way . . ." (1032a68). Yet if this is his answer to the question, it is an answer which proves
elusive. For that the question is to be answered in the same way does
not, unfortunately, tell us what the answer is; some interpreters have
thought that Aristotle means us to conclude that Socrates and to be
Socrates are the same,33 and others that they are not.34 But since Socrates has many accidental properties, if (as I take VII 6 to assert) he is
the same as his essence, then in this case, which is certainly a case of
essential sameness, there will nevertheless be a failure of substitutivity.
A defense of the claim that, though not identical, Socrates and his
essence are the same - and with it an interpretation of substance in
Metaphysics VII-VIII - is the subject of the next chapter.
33 See, for example, Hartman (1977), 57-87, esp. 63; Sellars (1967), 115; Rorty (1973),
402; Irwin (1988), 218.
34 See Furth (1985), 114.
Ill
A Theory of Substance
40
THEORY OF SUBSTANCE
41
42
for which reason it might be suggested that kinds just are classes or
sets. But even if this were so, not all classes would be kinds. For a class
is commonly defined as any collection of objects, 3 while at the very least,
not any arbitrary collection could be a kind; the members of a kind
need be alike with respect to certain nontrivial properties. Some classes,
of course, do meet this condition. All members of the class of white
things are white; all members of the class of human beings are rational.
Yet I would argue that this fact disguises a fundamental difference. For
whereas the identity of any class or set is solely a function of its membership, the identity of a kind, by contrast, would not seem to depend
upon its extension; 4 that for Aristotle kinds could have had other members seems beyond doubt. But if this is true, then even if, for example,
the class of human beings includes all past, present, and future human
beings, it would not be the same as the kind human being. More generally,
for any kind, it seems that there could have been other members of
that same kind.5
I have said that if kinds are not arbitrary collections there must be
some grounds that determine which objects are members of a given
kind. In fact what would seem to determine membership in a kind is
the presence in any object of certain properties; Socrates, for example,
is a member of the kind human being in virtue of his being rational or
two-footed. But - and this is crucial - even though both the kind and
the properties that determine membership in it are counted by Aristotle
as said of a subject, a kind is not identical with the properties essential
for membership in that kind. Laboring without the advantage of an
indefinite article (which is lacking in Greek), in the Categories, Aristotle
explains how kinds are both like and unlike properties. He says:
Every substance seems to signify a certain 'this' [tode ti\. As regards the primary
substances, it is indisputably true that each of them signifies a certain 'this'; for
the thing revealed is individual and numerically one. But as regards the secondary substances, though it appears from the form of the name - when one
speaks of man or animal - that a secondary substance likewise signifies a certain
'this', this is not really true; rather, it signifies a certain qualification \poion ti\
- for the subject is not, as the primary substance is, one, but man and animal
are said of many things. However, it does not signify simply a certain qualification [poion ti], as white does. White signifies nothing but a qualification [poiori],
whereas the species and the genus mark off the qualification of substance they signify substance of a certain qualification (3b 10-21).
3 See Massey (1970), 355. See also Mates (1965), 29.
4 To be sure Quine (1970), 8, thinks otherwise. He says: "There is no call to reckon
kinds as intensional. Kinds can be seen as sets, determined by their members. It is just
that [because not all classes or sets are such that their members share certain properties]
not all sets are kinds."
5 See Wolterstorff (1970), 258.
THEORY OF SUBSTANCE
43
When Aristotle says that man or animal is poion ti but not simply poion
ti (or not simply poion, according to some manuscripts) (3b 18), the
grounds for the distinction are as follows: since its name is applicable
to many things, a species is not a particular, yet because of the relation
between species and their members, species are nevertheless not properties. Even if it is too much to claim that any ontology involving objects
and properties must be committed to the existence also of collections,
the reasonableness of Aristotle's distinguishing kinds from essential
properties and including both in his ontology cannot be denied.
To this point I have said that, since like a class a kind can have
members, a kind would seem to be some sort of collection. Moreover
as it is a collection whose membership is determined by the presence
of certain properties, there could have been other members of the same
kind.6 Finally, because a kind is a collection, it is easy to see that most
properties of its members will not be characteristic of the kind. Of
course, given that Aristotle distinguishes between properties and kinds,
it will follow also that, while for Aristotle there are many human beings,
each with the property rationality, there would be, on his view, no
property man.1 Rather, as we have seen, 'man' ('human being') is the
name for a kind,8 and a kind is not a predicable universal but a collection.
II
We have examined the distinction between kinds and properties which
underlies the ontology of the Categories. Within this ontology, as we
have seen, Aristotle identifies that which is neither said of nor present
in anything else as primary substance. That is to say, primary substances,
6 Compare Mondadori (1978), 37, who accommodates the shifting membership of natural
kinds into possible world semantics as follows: "the view according to which natural
kind terms (rigidly) designate the corresponding kind (or species, or substance) can be
represented in modal semantics by assigning as extension to a given natural kind term
/ the corresponding kind (or species, or substance), and as intension a constant function
taking each possible world into a kind (a species, a substance). The latter, in turn, can
be represented by means of a function taking each possible world into a set of objects
(the set of tigers, or lemons, or samples of gold, as the case may be)." As a result of
this two-step definition, Mondadori can say both that natural kind terms are rigid
designators and that the class of tigers in one possible world is different from that in
another (see 38).
7 To be sure, Aristotle in De Interpretation 7 offers "man" as his example of a universal
(that which is of such a nature as to be predicated of many subjects), but the distinction
between properties and kinds is not made (or needed) in that context.
8 Pellegrin has marshaled evidence to show that in the biological works and elsewhere,
genos and eidos can function at very different levels of generality. If this variability
affects Aristotle's metaphysical views, it seems to me to complicate but not undermine
the account of substance which I want to give. That is to say, I will argue that substances
are specimens of natural kinds; if natural kinds can for Aristotle equally well be set at
various levels of generality, specimens of those kinds will vary in level accordingly. See
Pellegrin (1986), esp. ch. 2.
44
THEORY OF SUBSTANCE
45
Of course one's view of the relation between the Categories and the
Metaphysics depends in large part on what one takes Aristotle's theory
of substance in the Metaphysics to be. What I myself want to claim is
that, apart from the addition of the matter of which objects are composed, there is a certain continuity between Aristotle's ontology in the
Categories and the Metaphysics. However, since in the Metaphysics 'sub-
46
THEORY OF SUBSTANCE
47
every detail at that moment.21 Irwin, on the other hand, concedes that
particular forms are only indirectly objects of knowledge but argues
that this is acceptable. He says, "Aristotle cannot . . . allow that the
particular is in itself an object of definition and scientific knowledge;
but he can still fairly insist that scientific knowledge and definition apply
to a particular form."22 "On this view," he continues, "particular forms
partly satisfy the requirements for being objects of knowledge. Not
being universals, they are not the primary objects of knowledge; but
they are among the objects that scientific knowledge applies to. We
might say they are what science is about, though not what it is o/]"23
That Balme's proposed ontology is not Aristotle's seems clear. As for
Irwin's proposal, the difficulty is that it leaves Aristotle with the dilemma
he identifies in Metaphysics III 6: universals are knowable but are not
substances while particulars are not knowable (1003a5-17). To put the
point another way, if particulars are primary ontologically and universals
are most fundamental epistemologically, there is nothing that meets the
requirements for substance.
I have argued that if Socrates' form includes accidental features such
as Socrates' snub-nosedness, Socrates' form will not be knowable. On
the other hand, if Socrates' form does not include, for example, being
snub-nosed,24 then it is open to Aristotle to distinguish what is true of
Socrates and what is true of Socrates' form in such a way that the latter
is knowable. However, if he is to avoid Platonism, Aristotle must at the
same time deny (and if Socrates and to be Socrates are said to be the
same, he does deny) that Socrates and to be Socrates are numerically
distinct entities. I will now argue that the notion of a specimen of a
kind, introduced in the preceding chapter as a solution to the general
problem of referential opacity, shows how this can be accomplished.
Ill
Earlier I proposed to use the expression 'Socrates-qua-musician' in such
a way that by abstracting from all properties of Socrates except musicality it designates a specimen of the kind musician. Though my use of
4
Socrates-qua-</>' to designate a specimen of a kind is an interpreter's
term of art, it is one clearly derived from Aristotle. Indeed there are
many occurrences of 'qua' in Aristotle; some typical examples are: It is
not qua water or qua air that water or air is transparent (De Anima II
21 See Balme (1984), 5.
22 Irwin (1988), 263.
23 Irwin (1988), 263-64.
24 For a view of particular forms such that Socrates and Callias do not have qualitatively
different forms, see Frede (1987a), 63-71 and (1987b), 78. Sellars (1967), 112, says
that the particular form of this bronze sphere is the bronze sphere itself simply qua
sphere; see also Sellars (1957), 698. Hartman (1977) holds that the particular form
of Socrates is Socrates qua man. See 57-87.
48
THEORY OF SUBSTANCE
49
50
accidental unities like that of Socrates and musical; for these are the same only
by accident. (1037a32-1037b7)
Of course Aristotle does not mean that curvature exists without matter.
The point is rather that, since its matter is not part of what it is, between
curvature, that is to say, the geometrical shape, and its essence, there
will be identity. But like a snub nose, on the other hand, Socrates is a
whole that includes matter. Therefore although Socrates is essentially
the same as what it is to be a human being and accidentally the same
as the musical, in neither case is there identity.
I have defended the thesis that, as a result of the restrictive function
of 'qua', it is possible to refer not only, for example, to Socrates, but
also, by using the expression 'Socrates-qua-human being', to a specimen
of the kind human being. I have argued too that, despite the differences
between any two members of the same kind, since specimens of a kind
include no accidental properties there is no difference at all in the
qualities had by specimens of the same kind. But of course it is just this
indistinguishability that makes specimens of kinds, unlike the sensible
objects with which they are numerically the same, knowable. For at least
in the Metaphysics,26 the goals of science and metaphysics are limited to
the necessary and the universal. The result is that even if for Aristotle
the ideal candidates for knowledge would be the Platonic Forms, understood as eternally existing, necessarily unchanging patterns whose properties are all essential, specimens of kinds can be an acceptable approximation of the Forms without their separation.
I have argued that Socrates-qua-musician is knowable for Aristotle
because there is, in some sense, a definition for what it is to be musical;
of course, to say that x-qua-0 is a knowable object is not necessarily to
say that x-qua-0 is first in knowledge. Yet even though Aristotle would
say that qualities (and hence specimens of kinds derived from qualities)
have an essence and a definition in a secondary sense (VII 4), I have
as yet given no criteria that would establish that musician is a kind derived
from the properties of human beings, rather than conversely. To do so
would make it apparent that the requirement for epistemological priority is not one that is independent; to be not just knowable but first in
knowability will turn out to be, in other words, a characteristic not of
specimens of just any kind. Only specimens of natural kinds can be
specimens that are first in knowledge because it is only they that are
ontologically fundamental.
26 In fact Aristotle did realize that certain features not common to all members of a
species can be genetically transmitted. For a clear discussion of how this fact affects
the notion of form and the limits of knowability in the biological works, see Sharpies
(1985), esp. 120-22.
THEORY OF SUBSTANCE
51
IV
Although Aristotle says often enough that knowledge is belief about
things that are universal and necessary (e.g., Nicomachean Ethics VI 6
1140b31-32), in Section III I argued that for Aristotle substances specimens of natural kinds - are not universals but nevertheless are
knowable just because they, as it were, mimic the universal. However
if this is so, the inconsistency of Metaphysics VII-VIII can be resolved.
That in VII-VIII Aristotle asserts that substance is form, that form is
universal, and that no universal is substance, there is no doubt. 27 Some
philosophers have believed that the inconsistency cannot be overcome,
and others, taking Aristotle's approach to be aporetic, have supposed
that he might not be committed to all three claims. But most have sought
to avoid these conclusions either by reinterpreting one of the three
statements or by declaring one of the three constituent terms to be
ambiguous.
While every permutation has probably been tried, 28 recent solutions
have tended either toward particular forms or toward taking some universals to be substance. My view, that forms are specimens of natural
kinds, has evident affinity to the view that there are particular forms, 29
although also to a view according to which substances are neither particulars nor universals.30 An alternative takes form (substance) to be a
universal, perhaps something like a formal cause, 31 thereby denying
Aristotle's apparent claim in VII 13 that no universal is substance.
What Aristotle says in VII 13 is, in part, as follows:
For it seems impossible that any universal term should be the name of a substance. For primary substance is that kind of substance which is peculiar to an
individual, which does not belong to anything else; but the universal is common,
since that is called universal which naturally belongs to more than one thing.
Of which individual then will this be the substance? Either of all or of none.
But it cannot be the substance of all; and if it is to be the substance of one,
this one will be the others also; for things whose substance is one and whose
essence is one are themselves also one. (1038b815)
He then goes on to claim that substance means that which is not predicable of a subject but that the universal is predicable of some subject
27 For a list of relevant passages, see Lesher (1971), 169. Lesher does not distinguish
between passages that say only that the substance of a thing is form and those that
52
always (1038bl5-16), and later concludes, "If, then, we view the matter
from these standpoints, it is plain that no universal attribute is a substance, and this is plain also from the fact that no common predicate
indicates a 'this', but rather a 'such' " (1038b34-1039a2), a conclusion
he repeats in VII 16 at 1040b23-24. 32
It might appear that VII 13 is decisive evidence in favor of a view,
such as mine, which denies that forms are universal. Nevertheless John
Driscoll, Alan Code, and both Loux and Lewis, among others, have
argued that the VII 13 claim is limited to universals that are predicated
of particulars while substances or forms are universals that are predicated of matter,33 in which case substance can be universal after all. As
Driscoll says, "Since the formal cause or essence taken without any
material substrate is definable, it cannot be a particular (cf. Z, 15,
1039b20-22, 1040a5-7), and since it cannot be a particular, it must
be a universal in the broad sense that every nonparticular is a universal."34 Though the arguments of Loux's and Lewis's recent books have
more complexity than I can here do justice to, I will indicate some
points where there is disagreement.
What Loux argues is that in the Categories primary substances are
the familiar concrete particulars that belong to substance-kinds; a certain horse and a human being such as Socrates are his examples. As
Loux says:
The fact that a basic subject falls under its species cannot rest on or be grounded
in some prior instantiation of the said-of or the present-in relation; and, as I
have said, these two relations exhaust the tools Aristotle has at his disposal in
the Categories for ontological analysis or reduction. But, then, the result of the
interaction of [the basic-subject criterion for ousiahood and the doctrine of
essentialism] is that a basic subject's falling under or belonging to its lowestlevel substance-kind is a primitive, bedrock fact about the world, a fact not
susceptible of further analysis. For want of a better name, let us call this result
the Unanalyzability Thesis.35
15) gets most of his attention. His interpretation of it is discussed later in this chapter.
33 See Driscoll (1981), 150-52; Code (1986), 413; Loux (1979), 23; and Lewis (1985),
66-67. This way of reading VII 13 is derived from suggestions in earlier papers:
Woods (1967), 216, and Albritton (1957), 705.
34 Driscoll (1981), 151.
35 Loux (1991), 35.
36 Loux (1991), 159.
THEORY OF SUBSTANCE
53
Were it not a necessary fact about a thing that what it is to be that thing and
the thing itself are one and the same, it could not play the kind of explanatory
role that the Aristotle of Z and H37 wants to reserve for form. Form is to provide
us with a final answer to the question why things are what they are, and only
if a thing is necessarily the same with what it is to be that thing do we have the
requisite guarantee that there is no explanation in terms of something else for
that thing's being what it is, for its being the kind of thing it is. Only if a thing
is its own essence can it be its own ousia. For a thing of this sort, the question,
Why is it that kind of thing? has a ready answer that precludes the possibility
of any further questions. It is the kind of thing it is because it is one and the
same as being that kind of thing; and it could not be otherwise.38
54
THEORY OF SUBSTANCE
55
the only two ready alternatives I know of for explaining Aristotle's express
assertion that forms are subjects.50
In short it seems to me difficult to see how if form is universal, form
is a this in any way that allows it to be ontologically fundamental. Moreover it may be that there is also another problem. For in Chapter V it
will be noted that some interpreters have argued that matter in Aristotle
is such that its having the form it has is essential to it and even that, as
matter, matter has no properties other than its potential for form.
Although both Loux and Lewis respond by suggesting that Aristotle
may have had more than one way of thinking about matter,51 insofar
as such views of matter have plausibility, Loux's and Lewis's claim that
form is predicated accidentally of matter is in jeopardy.
Having considered Loux's and Lewis's defense of the claim that~form
is universal, let us return to the problem with which this section began,
namely, that of the apparent inconsistency of Metaphysics VII-VIII. If
one reads VII 13 in such a way that forms are universals predicated of
matter, it is clear how inconsistency has been avoided; substance is form,
form is universal, and some universals, namely, those that are predicated
of matter, are substances after all. But if, as I argue, substances are
specimens of natural kinds, the inconsistency is to be resolved rather
by reinterpretation of the claim that form is universal.
Of course, because substance is said to be form, and because there
is, or so I have conceded, a sense of 'form', namely, the differentia, in
which form is a universal, the triad could also be made consistent by
maintaining that substance is formj while form2 is universal. Nevertheless, as I see it, Aristotle intends his argument to be more cohesive than
a solution that appeals to equivocation would produce. For Aristotle's
point in saying that form is universal is that substances must be knowable, and I have argued that, even though Socrates, the sensible object
with which Socrates-qua-human being is numerically the same, is not
knowable, Socrates-qua-human being is knowable. That is to say, on my
proposal, it is, strictly speaking, false that form (in the appropriate sense
of the term) is a universal. Likewise it is false that the form - a given
specimen of a kind - does not come into being and pass away, though,
of course, it is true that the species does not do so. What is true (and
what is intended by the claim that form is universal) is that particular
forms have in virtue of their indistinguishability the epistemic virtues
of universals.
V
I have been arguing that in the Metaphysics Aristotle takes substances
to be specimens of natural kinds that are numerically the same as but
50 Lewis (1991), 303-4.
51 See Loux (1991), 180-83, and Lewis (1991), 250-58.
56
not identical with sensible objects, and I have claimed that specimens
of kinds are, as for Aristotle they need to be, knowable. Indeed it seems
to me an interesting fact that even in the Categories, where there are
no epistemological issues, Aristotle offers as examples of primary substances neither (as we might have expected) 'Bucephalus' and 'Socrates'
nor even 'this horse' and 'this human being' but the very general ho tis
hippos and ho tis anthropos (lb4-5) - an individual horse and an individual
human being.
To be sure, most interpreters of the Categories, including, as we have
just seen, Loux, take Aristotle to intend by primary substances ordinary
sensible objects such as Socrates, objects which in the Metaphysics are
considered to be composites of matter and form and not primary substances. Yet I am far from sure that the Categories is so straightforward.
Of course I would not claim that at the time of the Categories Aristotle
had already worked out the implications of distinguishing a human being
as such from Callias, Socrates, and so on; clearly if he had done so he
would have found it necessary to distinguish the subject of which something is said from the subject of which some things are said and in which
others, namely, accidents, are present. Rather, the point is just that
even in the Categories the differences among individual horses or individual human beings seem not to be of importance.
But if in the Categories the differences between one human being and
another are taken to be less interesting than the similarities and in the
Metaphysics it cannot be otherwise as there substances are supposed to
be knowable, what I want to conclude from this fact is, as I have said,
that Aristotle differs from Plato chiefly in rejecting the numerical distinctness of Forms from sensible objects. Whether for Aristotle a thing
and its essence or form are numerically distinct from one another is,
as I indicated at the end of the last chapter, the question of Metaphysics
VII 6; indeed it is well known that G. E. L. Owen once argued that in
VII 6 Aristotle rejects what is usually called the Nonidentity Assumption
in cases of strong (essential) predication, thereby providing an answer
to the Third Man Argument.
In fact there does seem to be a connection between the Third Man
Argument (or more generally, the Theory of Forms) and VII 6; specifically, if 'separation' means, as I argued in Chapter I, numerical
distinctness, a denial of numerical distinctness would just be the rejection of one of its premisses. According to Owen, however, the rejection
of the Nonidentity Assumption leads Aristotle to "embarrassment":
[for Aristotle] if we take any primary subject of discourse and say just what it
is, we must be producing a statement of identity, an equation which defines the
subject. And this in turn helps to persuade him that the primary subjects of
THEORY OF SUBSTANCE
57
58
THEORY OF SUBSTANCE
59
60
THEORY OF SUBSTANCE
61
he says: "of any character or property, x, that a particular has, the reality
is ho estin x, which it could not be if it were had by anything and which
therefore must be independent or 'separate' from all manifestations of
itself as a property."69 The second is that, once Forms are so described,
what is striking is the degree to which, but for Plato's inclusion of Forms
for properties, Plato's Forms so understood would indeed be similar to
Aristotle's substances, except as regards separation. Finally, the third is
to suggest that it is at least arguable that Aristotle is right about the
consequences of separation, that is to say, that Cherniss's assessment
notwithstanding,70 Aristotle's criticisms are very much on target after all.
Of course to say that Plato's Theory can, not implausibly, be taken
as straddling the distinction between particular and universal is not to
say that this is how Aristotle took it, anymore than to say that this is
how Aristotle understood the Theory shows Plato's intent. Nevertheless
that such an interpretation of Plato is not implausible and that it is
remarkably compatible with Aristotle's allegations, discussed in Chapter
I, remains to the point. Indeed in Metaphysics VII 14 there is what I
would claim is an acute analysis (at the level of the genus) of the problem
with types, if these are understood as real and yet, as Plato supposes,
numerically distinct from their tokens:
For if the Forms exist and animal is present in man and horse, it is either one
and the same in number, or different. . . . If there is a man-in-himself who is a
'this' and exists apart, the parts of which he consists, e.g. animal and two-footed,
must indicate a 'this' and be things existing apart and substances; therefore
animal too must be of this sort. Now if animal, which is in the horse and in
man, is one and the same, as you are one and the same with yourself, how will
the one in things that exist apart be one, and how will this animal escape being
divided even from itself? Further, if it is to share in two-footed and manyfooted, an impossible conclusion follows; for contrary attributes will belong at
the same time to it although it is one and a this. (1039a26-b4)
In short, as I understand Aristotle, the crux of his proposal, worked
out in his own theory of substances as a response to the Theory of
Forms, is just this: that the straddling of universal and particular which
is characteristic of Plato's Forms, and indeed which is necessary if what
is ontologically fundamental is also to be knowable, can be made unproblematic only if Plato's separation of Forms from sensible objects
is rejected. However, if Platonic separation is rejected, so Aristotle
argues, forms, though particular in virtue of their numerical sameness
with different sensible objects, have on account of their indistinguishability the knowability characteristic of universals.
In this chapter I have argued for understanding Aristotle's substances
as specimens of natural kinds and have claimed that, so understood,
69 Cherniss (1957), 261. See also 260.
62
IV
Substance and Aristotle's Epistemology
63
64
In the Meno Plato is faced with the paradox of discovery; how, the
dramatic character Socrates asks, paraphrasing a complaint raised by
Meno, can one try to discover either what he knows or what he does
not know? One would not seek what he knows, for since he knows there
is no need of inquiry, nor what he does not know, for in that case he
does not even know what he is to look for (80d). Plato's answer is to
argue that we seek what we in some sense know. Plato illustrates his
proposal, the theory of recollection, by eliciting from an uneducated
slave knowledge of geometry. Challenged to find a way to double the
area of a square, the slave twice proposes procedures that give the wrong
result, is led by Socrates to see that these are wrong, and then, having
conceded that he does not know, is able to follow Socrates' explanation,
worded as an interrogation, of how it is to be done. This knowledge,
the knowledge the slave comes to have, is said to result not from teaching
but from questioning (85d).
Plato's claim that Socrates does not teach the slave can look tendentious, but perhaps it is not so, given that it is intended to cover a priori
and not empirical knowledge; if Socrates supplies the solution to the
problem, he nevertheless expects the slave to see why what he proposes
concerning squares is true.2 To be sure, the example also suggests that
the recovery of knowledge may require a guide, and even at that what
is recollected may at first have a "dreamlike" quality (85c). Nevertheless,
Socrates says, "if the same questions are put to him on many occasions
and in different ways, you can see that in the end he will have a knowledge on the subject as accurate as anybody's" (85c-d) - he will be able
to produce the proof himself and to see why it is to be done in this
way. His true opinions, Socrates tells us, will then have been turned
into knowledge (86a). In fact it seems that not only will he have knowledge but he will then know that he knows.
With the doctrine of recollection Plato has produced a solution to
the paradox of discovery - he has claimed that we in fact know what
we do not now seem to know - but it is a solution beset by a variety of
difficulties. On any theory of knowledge there is need to explain acquisition and verification, yet that the objects of knowledge are for Plato
not now readily available for inspection causes special problems. In the
Meno it is said that the soul, since it is immortal and has seen all things
both here and in the other world, has learned everything that is (81c);
in the Phaedrus Socrates speculates that our souls were then like charioteers and their horses that rushed after the gods through the heavens
and saw the Forms along the way (246a-248b). But it is interesting that
in the Phaedrus Plato himself worries that we might have been unable
to attend properly to our experiences (248b), a state of affairs that
2 See Moravcsik (1971), esp. 65, and Vlastos (1965).
65
"Once the final truth is recognized," Gulley adds, "it cannot 'run away'.
We know that we have found the truth."6
Problems about the acquisition of knowledge and its verification
arise, as I have said, on any account of knowledge; they are hardly
unique to Plato. But the point is that because for Plato the objects that
we are said to know are not now experienced and because, indeed, we
for the most part are not even aware of having knowledge of them, the
3 Guthrie (1978), 113, claims that in the Theaetetus Plato cannot be thinking of the Theory
of Recollection because at 197e he says that the aviary is empty at birth. I take this to
be a feature of the story which is without philosophical significance for determining
Plato's overall concerns.
5 Gulley (1954), 195.
4 White (1976), 51-52.
6 Gulley (1954), 195.
66
67
68
recognition. Rather he seems to believe that an account of human knowing can be given which has an immediacy - and also a credibility lacking in Plato's. In the sections that follow, Aristotle's account will
be examined.
II
The goal of the Theory of Recollection is to explain how learning occurs.
Plato chiefly proposes that, starting from perceptual cues, we regain
access to knowledge already within us. In Posterior Analytics II 19 Aristotle himself attempts to explain how, beginning from the perception
of sensible particulars, scientific knowledge - that is to say, knowledge
of universals resting on indemonstrable first principles - can be attained.
That his target is the Meno is plain from the outset. In the first chapter
of Book I at 71a 24-30, he has said:
Before the induction, or before getting a deduction, you should perhaps be
said to understand in a way - but in another way not. For if you did not know
if it is simpliciter, how did you know that it has two right angles simpliciter? But
it is clear that you understand it in this sense - that you understand it universally
- but you do not understand it simpliciter. (Otherwise the puzzle in the Meno
will result; for you will learn either nothing or what you know.)
Although the goal is said to be knowledge of "first principles,"
whether by first principles Aristotle means primitive concepts or indemonstrable propositions, or whether he vacillates between them, is
unclear. But, as several interpreters have recently pointed out, since a
first principle is for Aristotle a definition and a definition is the articulation of a concept, it really makes no difference.11 Having argued that
knowledge of principles is not, on the one hand, innate (Posterior Analytics II 19 99b26-7) and that, on the other, it is impossible for such
knowledge to come about if we are ignorant of it and have no capacity
for it at all (99b30-32), Aristotle concludes that we must have some
capacity from which to derive our knowledge of the principles (99b3234). Of this capacity, he says:
And this evidently belongs to all animals; for they have a connate discriminatory
capacity, which is called perception. And if perception is present in them, in
some animals retention of the percept comes about, but in others it does not
come about. Now for those in which it does not come about, there is no knowledge outside perceiving (either none at all, or none with regard to that of which
there is no retention); but for some perceivers, it is possible to grasp it in their
minds. And when many such things come about, then a difference comes about,
so that some come to have an account from the retention of such things, and
others do not. (99b34-100a3)
11 See Modrak (1987), 161-64, Kahn (1981), 387-97, and Barnes (1975), 249, 259-60.
69
70
71
72
73
Both Barnes and Kahn cast doubt on the idea that for Aristotle we
can know what a human being is by perceiving Callias and others as
human beings. Aristotle's psychology does not, they believe, allow Callias
to be perceived in such a highly conceptualized way. But in that case
how do we come to know what a human being is? Kahn's suggestion is
that for Aristotle it is the active intellect that permits us to learn from
experience, by acting on us so as to actualize our potential intellect.20
Indeed, according to Kahn, far from being a version of empiricism,
Aristotle's alternative to Plato is "superrationalism." As he says,
As far as the active intellect is concerned, the rationalist doctrines of innate
ideas and infallible intuition convey too weak a picture of its complete and
unwavering grasp of all noetic truth. As far as we are concerned, however, the
rationalist model simply does not fit, since our potential intellect is not stocked
at birth with noetic principles and does not acquire them by any act of direct
intuition that we can perform. One may, if one wishes, speak of the active
intellect as continuously intuiting the forms and essences of the natural world.
But we can enjoy such an intuition only to the extent that we succeed in realizing
its activity in our own thought and knowledge.21
Despite the fact that, as Kahn understands Aristotle, we have no innate
ideas, "if the intellect can grasp the noetic forms and essences in the
phantasms provided by experience, that is because it already knows or
is these very forms before experience." 22
Thus Kahn's view of Aristotle can, I think, not unfairly be called
Platonic - the forms can be known from experience because the forms
were known before experience. Nevertheless if Kahn is right about the
role of the active intellect and if he is right to suppose the active intellect
to be something apart from us,23 how is it that the knowledge had by
the active intellect comes to be ours as well? What Kahn offers on
Aristotle's behalf falls very short of being an explanation. He says:
And this process of learning and exercising science, although it has a metaphysical cause and even a metaphysical guarantee in the super-rationalism of
the active intellect, must be achieved in our own experience by the ordinary
processes of induction and hard work: there is no epistemic button we can push
in order to tune in on the infallible contemplation of noetic forms by the active
intellect.24
74
what is the connection between that and the engagement of the active
intellect on our behalf? Whether the forms are buried within our memories or whether they are known by an active intellect that is distinct
from us, the problem of access is the same. Nevertheless it must be
remembered that the problem the involvement of the active intellect is
intended to solve arose because Kahn, in rejecting the idea that we can
perceive Callias as a human being, takes Aristotle to be committed to
a perceptual content of only proper and common sensibles. But some
interpreters believe Aristotle intends a more generous perceptual content than that which Kahn allows. One such proposal will be the subject
of the next section.
IV
According to Kahn's account of the movement from perception to
knowledge, incidental objects of perception fall outside the scope of
perception. Modrak, however, uses Aristotle's remark that though one
perceives the particular, perception is of the universal to suggest a way
in which universals might be grasped through perception after all. She
says:
[At 100al5-b3 Aristotle makes] the crucial point that there is a sense in which
universals are grasped through perception. The universals in question are those
that determine the features of the sensible particular; these include features
perceived kata sumbebekos as well as those perceived kath * hauta. Indeed, in the
case of substances, the single most important universal, namely, the substance
sortal that the particular falls under, is perceived kata sumbebekos. What is incidental at one cognitive level becomes essential at another (cf. 89a33-37). The
essential features of a concrete particular are general characteristics. Hence the
perceptible individual is not only the object of perception but is also the vehicle
for the apprehension of the universal.25
She explains how for Aristotle this is possible:
The sensible particular is a token of a type, and we apprehend the type in virtue
of apprehending the particular. In Aristotle's metaphysics, the substance type
is ontologically prior to the token; the essential characteristics of an individual
human being are determined by the substance type, the species anthropos. We
perceive a particular man (i.e., a token of a certain type or species), but the
perception is of man (i.e., the type is the ultimate determinant of the content
of the perception). Many of the distinctive perceptible features of a particular
object are type-dependent. For instance, the difference between a cat and a
dog is more easily recognized through perception than the difference between
one cat and another, (cf. Physics 184a24-bl4)26
One might conjecture that Barnes and Kahn would find Modrak's
view problematic; Kahn, it will be recalled, doubted that we could per25 Modrak (1987), 168.
75
be directly affected by properties that are not sensible in the strict sense
(all the italics are mine) would seem to be either vacuously true - incidental perception is not nonincidental perception - or tendentious.
That is to say, Kahn has said nothing to show that incidental perception
of the sort prescribed in Barnes's definition is not available to Aristotle.
On the contrary, such a view would seem to be required by what Kahn
himself says of Aristotle's understanding of even animal awareness:
"[sense perception] permits a dog to recognize its master and distinguish its master from the master's horse, without knowing that its master
27 Others who have attempted to find middle ground include Ross, Sorabji, and Cashdollar. Ross (1949), 678, says of this passage: "we perceive an individual thing, but
what we perceive in it is a set of qualities each of which can belong to other individual
things." Sorabji (1992), 197, argues that for Aristotle some perceiving is propositional
(we perceive that something is the case), in other words that "perceiving that" need
not be described as an inference of reason. As he puts it, "Coincidental does not
mean inferential." Cashdollar (1973), 156, 161, argues that incidental perception,
even though predicative, is nevertheless a case of perception.
76
is a man or that the horse is a horse."28 Nor has Kahn shown that (for
suitable perceivers) from such a beginning the universal cannot be attained.
Thus I want to agree with Modrak that Aristotle uses incidental perception to explain how perceptible particulars can be the vehicle for
knowledge of the universal and that it is reasonable for him to do so.
Specifically, what Modrak argues is that it is the capacity for representation which allows us to reach knowledge of the universal from sensory
awareness of particulars. When the content of a perception is retained
in a phantasm or memory image, the image, being an image of Callias,
will include some of his accidental features. Nevertheless once images
of other human beings and other images of Callias are retained as well,
then (in the human soul) it is possible for even the essence to be derived.
This progression need not, Modrak contends, involve a superimposition
of the images into a confused general image; that, she maintains, is a
view that is neither philosophically plausible nor true to Aristotle. 29
Rather, the proposal is that any of these particular images can be reused
to represent the others. She says:
if Callias is perceived as tall, thin and balding, the phantasma of Callias employed
in thinking about the essence of human beings would include (some of) these
idiosyncratic characteristics, but the thought would ignore them. To grasp an
intelligible form is to reinterpret the content of an appropriate phantasma.
Nevertheless, the phantasma is a necessary component of the thought. Just as
in the extramental world essences inhere in matter, essences-in-thought inhere
in the phantasmata that serve as their material substrata.30
We have seen that on Modrak's view the gap between the phantasm,
which is still particular, and knowledge of the universal is closed by the
capacity for representation. That is to say, it is representation that takes
us from the phantasm to thought (432al2-14). In this it seems to me
that she is right. I will not pretend that no philosophical problems arise
in understanding how representation is supposed to work, nor will I
try to determine whether Aristotle wants to appeal to a general image. 31
But what needs to be emphasized, I think, is the connection between
28 Kahn (1981), 402. Aristotle's motive for considering incidental perception to be a
case of perception seems in fact to be animal intelligence; animals, he wants to hold,
lack reason and yet they respond appropriately. At Metaphysics I 980b25-27, a passage
not unlike Posterior Analytics II 19, Aristotle claims: "The animals other than man live
by appearances and memories, and have but little of connected experience." Of this
passage Ross (1949), 677, says: "I.e. experience is a stage in which there has appeared
ability to interpret the present in the light of the past, but an ability which cannot
account for itself; when it accounts for itself it becomes art." Sorabji (1992), 195209, has an excellent account of the problem posed for Aristotle and other ancient
philosophers by animal intelligence.
29 Modrak (1987), 166.
30 Modrak (1987), 169.
31 D. Frede (1992), 291, seems, unlike Modrak, to believe that Aristotle does make use
of a general image and that his doing so is not unsatisfactory.
77
Aristotle's metaphysics and his epistemology. I will argue next that the
fit between Aristotle's epistemology and his metaphysics, on the interpretation of the theory of substance which I have offered, should be
taken as a reason for thinking that interpretation to be correct.
78
but do not have the accidental properties of the latter. However, to say
that Aristotle proposes that substances are, as it were, unseparated types
and thus that he must explain how there can be knowledge of these
would be to hold the same view. Moreover, the result of the process
that Modrak has described, namely, the reuse of a phantasma to arrive
at knowledge of the universal, or more precisely, at knowledge of an
unseparated type, is just knowledge of a specimen of a kind. Thus
Aristotle's epistemology, as Modrak describes it, would match his metaphysics as I understand it to be.
In Aristotle's account of the movement from perception to knowledge, it has often been noticed that he does not address skepticism of
the sort that concerns us - or even Plato. As his starting point, for
example, Aristotle claims that in perception the perceiving and the
object perceived are the same (431b20-432al); having postulated that
in perception the form of the object is taken in without the matter in
something like the way a piece of wax takes on the imprint of a signet
ring (424al7-24), Aristotle does not then concern himself with whether
there might not really be a surface to be seen or whether the white
human-shaped surface might nevertheless not be a human being.
In other words, Aristotle never doubts that we have knowledge. But
if he is addressing the problems introduced by Plato's Theory of Forms,
the otherwise puzzling limits of his epistemological interests make sense.
To start from experience and assume it to be experience of something
outside oneself is to do no more than Plato did in granting us awareness
of the phenomenal world. The difference between Plato and Aristotle
arises rather in the relation between what is in flux and what is unchanging. If, contrary to Plato, what is perceived is numerically the
same as what is to be known, then, Aristotle believes, the world is such
that what is perceived is also, in a way, what is both knowable and known.
That is to say, if what is perceived is numerically the same as what is to
be known, then Plato's peculiar problem of retrieval on the basis of
vague and ambiguous phenomenal cues vanishes, making inquiry - rational investigation from a base of experience - possible. As T. EngbergPederson explains Aristotle's theory, ''[One can attend to] particular
cases with the consequence that insight into some universal point is
acquired."33
Thus even if we may go wrong on a particular occasion, even if nous
is not presumed to guarantee the truth of whatever is grasped,34 by his
claim that substances are numerically the same as sensible objects,
Aristotle believes that he has eliminated reasons of the sort that confront
Plato for doubting that we can and do know that which is ontologically
fundamental. To put the point another way, in not considering Cartesian
33 Engberg-Pederson (1979), 305.
34 See Engberg-Pederson (1979), 308, and Lesher (1973), esp. 58-65.
79
grounds for doubt, it can fairly be said that Aristotle just assumes that
there is an intelligible world and that we can have knowledge of it.
Nevertheless what needs also to be said is that as a consequence of these
assumptions, Aristotle takes it to be a criterion for a successful metaphysical theory that it propose a world where such knowledge can plausibly be obtained and that by his rejection of separation he believes that
this is just what he has done.
VI
I have agreed with Modrak that Aristotle's account of the acquisition
of knowledge depends on the reuse of a phantasm in such a way that
any human being can represent any other. Of course, there are many
features of Aristotle's epistemology and philosophy of mind which I
have not considered. Nevertheless I believe that I have said enough to
show that his epistemological view is encouragingly parallel to the metaphysical account of substance given in earlier chapters. For there I had
argued that, as a result of the restrictive function of 'qua', it is possible
to refer not only to Socrates but also, by using the expression 'Socratesqua-human being', to a specimen or paradigm of the kind human being.
I had argued too that since accidental properties are eliminated there
is no qualitative difference among specimens of the same kind, with the
result that for Aristotle substances are knowable despite not being universals. Now I have argued in this chapter that, in the extraction of the
knowable from perception, what are known are in fact specimens of
kinds.
But if the metaphysical grounding of Aristotle's epistemology is as I
have argued it to be, then contrary to what is sometimes thought, Posterior Analytics II 19 and the parallel account of perceiving and thinking
in the De Anima are consistent with what Aristotle says in Metaphysics
XIII 10. In that chapter Aristotle describes the limits of knowability,
saying:
The statement that all knowledge is universal, so that the principles of things
must also be universal and not separate substances, presents indeed, of all the
points we have mentioned, the greatest difficulty, but yet the statement is in a
sense true, although in a sense it is not. For knowledge, like knowing, is spoken
of in two ways - as potential and as actual. The potentiality, being, as matter,
universal and indefinite, deals with the universal and indefinite; but the actuality,
being definite, deals with a definite object - being a 'this' [tode ti], it deals with
a 'this'. But per accidens sight sees universal colour, because this individual colour
which it sees is colour; and this individual a which the grammarian investigates
is an a. (1087al0-21)
The contrast between potentiality and actuality in this passage is
admittedly problematic. Walter Leszl supposes that it is the contrast
80
To be sure, Heinaman can say that the relevant universals cannot themselves be objects of actual2 knowledge since there are no other propositions knowledge of which could constitute potential knowledge of
them.39 Nevertheless this seems, if anything, to establish their priority,
not their lack of it.
35 Leszl (1972), 312.
37 Heinaman (1981b), 70.
39 Heinaman (1981b), 70, 67.
81
82
universals do not exist apart from individuals, to hold as Leszl does that
the content of knowledge is the universal is to preserve what Cherniss
has called "the discrepancy between the real and the intelligible,"43 and,
as I said in Chapter I, it is this very discrepancy that Aristotle worries
about in the aporia in Metaphysics III. His solution, I have claimed, is
that Socrates' form, a specimen of the kind human being, is, by virtue
of being numerically the same as Socrates, a particular and yet knowable
just because it is indistinguishable from other specimens of the same
kind. But if knowledge is of the universal only in the sense that it is
indeterminable which specimen of a given kind it is that one knows,
then Metaphysics XIII 10 and the passages in Posterior Analytics and De
V
The Separation of Substance
For Fine that Aristotle should claim separation for his substances
while criticizing Plato for separating the Forms in no way presents a
problem. Platonic Forms, as Fine understands Plato, are universals, and,
according to her, are thought by Aristotle to be so.1 Aristotle's objection
to Plato, as she sees it, is that universals should not be said to have a
capacity for independent existence. Substance, on the other hand, which
is understood by Fine to be for Aristotle the composite, can be said to
have just this characteristic, existing independently of other substances
as well as of accidents.2
1 Fine was discussed in Chapter I. For Forms as universal, see (1984), esp. 45.
2 Fine (1984) argues that independent existence can apply to the composite in the sense
that composites are separate from their attributes. By this she means that although
83
84
But Charlotte Witt also takes substance to be the composite and holds
that for Aristotle substances are separate while nonsubstances are not;
she understands the separation of substances to mean that nonsubstances and only they exist in subjects, thereby making nonsubstances
ontologically dependent on substances.3 If Witt also believes that Platonic Forms are universals (a question on which she offers no opinion),
then one can assume that she would find no conflict between Aristotle's
criticism of Plato for separating the Forms and his views about the
separation of substance. That is to say, if Forms are such as to be
predicable, then Aristotle is justified in saying that they too must exist
in subjects, while substances do not.
Fine and Witt therefore offer views on which there is no difficulty
about the fact that Aristotle both criticizes Plato for separating the
Forms and, in the same sense, claims separation for his substances.
However, on my interpretation, as has no doubt become apparent,
Aristotle's claim to separation in the case of substances is problematic.
For what I have said is that Aristotelian substances, being forms in the
sense of specimens of natural kinds, are in fact rather like Platonic
patterns except as regards separation. I will need to propose, therefore,
a way of understanding the separation of substances which is different
from the separation had by Platonic Forms. To do so will be the task
of this chapter.
SEPARATION OF SUBSTANCE
85
that VII 1 makes use of the categories and in the Categories an appeal
to independent existence is found: "Thus all the other things are either
said of the primary substances as subjects or in them as subjects. So if
the primary substances did not exist it would be impossible for any of
the other things to exist" (2b4-6).
Thus VII 1 alludes to some doctrines from the Categories. On the
other hand, there are problems with importing views about separation
from the Categories into the Metaphysics. In the first place, while the
view expressed in the Categories clearly states the priority of substance,
it does not state that substance could exist without the other categories,
only that they cannot exist without it, and if the former claim is intended,
it is hard to see how it can be sustained. In the second place, whatever
the view in the Categories and whatever its justification, it cannot be
assumed that Metaphysics VII 1 is simply a restatement of it; in the
Categories, after all, views about substances were uncomplicated by matter. For both these reasons, therefore, it seems that we cannot rely on
the Categories to explicate VII 1. It is likewise with Physics i85a31-32,
where Aristotle says, "For none of the others can exist independently
except substance; for everything is predicated of substance as subject";
since in Metaphysics VII 3 matter, form, and the composite are all said
to be a sort of substratum (1029a2-3), appeal to that notion does not
clarify the meaning of 'separation'.
It seems then that there is no recourse but to determine from the
Metaphysics itself what the separation of substances can be,5 yet in the
Metaphysics exegesis of the sense in which substances must be separate
is complicated by Aristotle's own terminology. Although Aristotle sometimes distinguishes separation in space from separation in definition,6
and the latter from simple or unqualified separation (VIII 1 1042a2831), at other times he just says 'separation' (1028a33-34). In VII 3
matter is rejected as a candidate for substance on the ground that it is
neither tode ti nor choriston (1029a27-28). Nevertheless here too what
that which is separate7 is to be separate from and in what way it is to
be so are not clarified.
5 Metaphysics XIII 10 1086b 14-20 might seem promising. Aristotle says: "Let us now
mention a point which presents a certain difficulty both to those who believe in the
Ideas and to those who do not. . . . If we do not suppose substances to be separate,
and in the way in which particular things are said to be separate, we shall destroy that
sort of substance which we wish to maintain; but if we conceive substances to be
separable, how are we to conceive their elements and their principles?" But unfortunately, how particular things are to be separate is, as we have seen, itself disputed.
6 De Anima 4i3bl4-15, 429al 1-12, 432*20; Metaphysics ioi6b2-3, 1048bl5, 1052bl7;
On Generation and Corruption 32ob24, 320b 12-14; Nicomachean Ethics no2a28-31; Physics i93b4-5.
7 Some interpreters prefer to translate choriston as 'separable' rather than separate. I do
not find it necessary or useful to make a distinction here, but some of the varieties of
separation that I consider would be described by others as varieties of separability.
86
SEPARATION OF SUBSTANCE
87
saying of this chair what it is, one need not say of the coat lying on the
chair what it is; having a coat thrown on it is not part of what a chair
in itself is. But for both Plato and Aristotle, the more interesting issue
is, as it were, the "vertical" separation of what is ontologically and
epistemologically fundamental from what is in flux, and it is on this
relation that I will focus in explicating and defending my interpretation.
That Plato's Forms have "independent being" in relation to phenomena is clear enough. To say what The Shuttle Itself is one need not
- and indeed cannot - refer to its phenomenal images. The problem
with Plato's Theory, I take Aristotle to be arguing, is that Plato erroneously supposes that that which has independent being in relation to
phenomena must also be numerically distinct from them. This point is
made most clearly, I believe, in Metaphysics VII 6. Beginning with the
Platonic Forms, Aristotle proceeds in VII 6 to undercut the case for
the Forms by arguing that each self-subsistent thing and its essence are
the same. The argument is as follows. If the essence of good is to be
different from the Idea of Good, there will be substances besides those
postulated - that is to say, the essence of the Idea of the Good will be
a substance numerically distinct from the Idea of the Good and prior
to it (1031a31-b3).
Using the example of the essence of horse, Aristotle also expands
the argument in VII 6 in such a way as to create a regress. If the essence
of the essence of horse is different from the essence of horse (and if
the essence of horse is different from horse, on what grounds would
the essence of the former not be different from it?), the same result namely, the existence of substances prior to presumed substances - will
again obtain (1031b28-30). But regress, Aristotle tells us, is not the
only difficulty. For if the Idea of the Good, for example, is numerically
distinct from its essence, it will be unknowable (1031b4). That is to say,
whatever we know, we will not know the Form (the presumed knowability
of which was one of the chief motives for its postulation), but something
else, its essence.
In the case of the Forms, since there is nothing that is a property
of the Form and not a property of its essence or conversely, to deny
the numerical distinctness of a Form and its essence is to assert their
identity. Presumably to do so is the course Plato would want to take.
But, Aristotle claims, his argument need not apply to Forms alone. That
is to say, whether there are Platonic Forms or not, all primary and selfsubsistent things are, Aristotle claims, the same as their essences
(1032a4-6), a position that, if defensible, would seem to eliminate the
need for Forms. Nevertheless it is clear that if Aristotle intends, as I
think he does, to imply that Socrates and to be Socrates are the same
(1032a8), they cannot be identical; unlike a Platonic Form and its es-
88
SEPARATION OF SUBSTANCE
89
said to lack is complicated by disagreement among interpreters of Aristotle on how the notion of proximate matter is to be understood. One
might suppose that the bronze of the statue - I do not mean the bronze
from which a statue can be made, that is, the preexisting matter, but
the bronze from which an existing statue has been made - is an example
of matter and that as such it has various familiar properties, including
size, weight, hardness, color, location, and, of course, shape. Nevertheless this view may in fact not be Aristotle's. For many recent interpreters, including Balme, Dancy, and Michael Frede, believe that matter
as such has no properties. As Dancy explains it:
In Z 3, we could turn directly to the bronze, which is something (the elements
mixed in the ratio X:Y) in its own right, but thinking of it that way is not thinking
of it as the material cause of both the statue and the lump into which the statue
is beaten. Thinking of it in this latter way is not thinking of it in its own right
at all, but as what used to be pale and charioteer-shaped and is no longer either.
When we consider the transmutation of fire into air, the case is clearest. For
here, there is no answer to the question "what is it that used to be hot and dry
and is now hot and wet?" at all: there can be no question of turning directly
to the stuff and discussing its essential structure. But this is relatively unimportant. It was at the top of Aristotle's hierarchy that the distinctness of form
was clearest; still, it was distinct from the matter all the way down. So also here,
at the bottom of the hierarchy, the inseparability of the matter is clearest; yet
it is inseparable all the way up.8
If one wonders why Aristotle might have wanted to say that it is not
the bronze as such which is matter but rather the bronze understood
in relation to the statue, the motivation would seem to come from cases
where it is not so easy to identify the matter of the composite with the
preexisting material. Ackrill explains the problem that arises if one tries
to apply a distinction readily made in the case of many artefacts to living
things:
The timber, hinges, and screws can still be seen when the cupboard is built,
but the eggs and sugar are lost in the cake. If, as a result of cooking, a and b
combine to form a homogeneous stuff c, a and b are no longer there to be
picked out. We can refer to the a and b we started with, and perhaps we can
recover the a and b again by some process. But a and b are present now, if at
all, only potentially. Actual bricks constitute an actual wall, though those very
same bricks might not have done so. But here is quite a different story: potential
a and b are 'in' actual c, though they might have been actual a and b. Chemical
change, in short, which yields a new sort of stuff, cannot easily accommodate
an account tailor-made for other operations. . . . This is the difficulty for Aristotle with the basic living materials such as flesh and bone. They are produced,
as he explains in detail in the biological works, by processes like cooking. . . .9
8 Dancy (1978), 408.
90
SEPARATION OF SUBSTANCE
91
a view fits well with the account of separation I want to give. For, as
we have seen, in VII 3 Aristotle rejects matter as a candidate for substance on the grounds that it is neither tode ti nor chbriston. One might
suppose that by 'not separate' he means that it is not numerically distinct
from the composite, and On Generation and Corruption I 5 ("It is therefore better to suppose that in all instances of coming-to-be the matter
is inseparable, being numerically identical and one, though not one in
definition" (320bl2-14; cf. 320b22-25)) seems to suggest as much. But
Dancy says of Aristotle: "when he talks about separability he need not
be, and generally is not, thinking about questions of individuation, divided reference, or countability, about whether or not stuff can occur
in the absence of 'things'."13 Rather, according to Dancy, what is inseparable about matter is that, as matter, there is nothing that it is "in
its own right."14 As the potential for change in something, matter does
not "stand on its own";15 in other words, there is nothing that it is. But
if not having an essence is the ground for the inseparability of matter
as matter, then the sort of separation that matter is said to lack, the
sort of separation that is had by substance, would seem again to be the
ontological correlate of separation in definition.
Ill
In the first section of this chapter I argued that the separation of substances, which I have taken to be specimens of natural kinds, could be
understood as independence in being, a sort of separation not had by
attributes, matter, and (except in relation to accidents) the composite.
If this account of separation is right, the dispute between Plato and
Aristotle is not over whether what is ontologically fundamental is, in
some sense, separate, but over the implications of requiring separation
in the sense of independent being of that which is ontologically fundamental. In the case of the Forms, Aristotle tells us that Socrates sought
definitions but that it was Plato who was responsible for separation. Of
course, from the Socratic dialogues we learn that Socrates regularly
rejected definition by example, and in the Euthyphro he is even made
to talk of what he is seeking not just as what is common to all cases but
as that by which pious things are pious and that to which we can look as
a standard (6d-e). Nevertheless, even if we assume that Socrates made
claims such as these, it seems not implausible to believe that he left it
unclear what he meant by them;16 that it was Plato, not the historical
13 Dancy (1978), 401.
14 Dancy (1978), 407.
15 Dancy (1978), 400.
16 Thus I disagree with Allen (1970), 133-36, 147, 149. That Socrates did not separate
the forms might, he says, mean that Socrates did not distinguish forms from their
sensible instances or that, although he distinguished them, he did not regard them
as individuals or even that though he took them to be individuals he did not regard
them as independent of and prior to their instances. What Allen argues is that forms
were for Socrates separate in all these ways and that Socrates differed from Plato only
in that separation did not for Socrates make sensibles deficient.
92
SEPARATION OF SUBSTANCE
93
to this: that Aristotle believes that the separation of what is real and
knowable from flux can be achieved without its numerical distinctness
from what is in flux.
IV
My interpretation of separation in Aristotle differs from Fine's with
regard to the sort of separation that Aristotle attributes to Plato and
also the sort that is had by substances. My view also stands in sharp
contrast to the interpretation of the separation of Aristotelian substances offered by Donald Morrison. Morrison argues that in Aristotle
separation should be understood in terms of being outside the "ontological boundaries" of something else - or, equivalently, according
to Morrison, as numerical distinctness from another thing.17 But if a
thing is separate from that which is outside its ontological boundaries,
it is separate from that in relation to which it does not have ontological
influence. If this is so, then, as Morrison says, what substances are
separate from is other substances; they are not, that is to say, separate
from the accidents that inhere in them.18
On Morrison's view, therefore, separation as it applies within Aristotle's ontology is what I have called a horizontal relation; it is the
separation, as it were, of cats and dogs, and indeed of one dog from
another. Moreover, because a substance is naturally prior to its accidents
but is not separate from them while it is on the other hand separate
from every other substance but not prior or posterior to them, the
connection for Morrison between separation and natural priority is
more accurately a connection between natural priority and nonseparation. On my view, by contrast, since substances are said to be separate
in the sense of having independence in being from matter and accidents
and the composite, it is easy to see why, for Aristotle as for Plato, being
separate and being prior should be linked. Indeed, as I said at the close
of Section III, my interpretation makes Aristotle's view satisfyingly parallel to Plato's. For Aristotle recognizes that substances do need to be
separate from sensible flux, and he disagrees with Plato not about the
need for separation but about its nature.
Thus my account, unlike Morrison's, preserves the link between separation and priority as well as making evident the connection between
the Platonism Aristotle deplores and the Platonic requirements he
nevertheless endorses. There is also another difference. For it follows
from his account of the nature of separation that, as Morrison says,
17 Morrison (1985), 128, 138-39.
94
SEPARATION OF SUBSTANCE
95
I have argued that, when Aristotle criticizes Plato for separating the
Forms, he is arguing that Plato makes them numerically distinct from
22 Loux (1991), 263.
24 Loux (1991), 262, n. 28.
96
SEPARATION OF SUBSTANCE
97
98
SEPARATION OF SUBSTANCE
99
would not be substances is clear. Indeed that Aristotle's view is far from
Suppe's is very evident if a claim of Owen's is correct. For what Owen
says is that in Aristotle it is a man or a plant that is a prime example
of substance, where a man and a plant "carry no namely-rider such as
Viz. Socrates or Callias or . . ,',"27 and this proposal is, I think, much
the same as the position which I have defended; Socrates-qua-human
being and Callias-qua-human being, being indistinguishable, might very
well be described as x-qua-human being - entities that are, that is to
say, without a "namely-rider."
To be sure, Owen's rendering appears no less paradoxical than mine;
that 'a man' means 'some man' and that 'some man', if taken to refer
at all, must refer to Callias or to Socrates or to whomever seems as
certain as that 'Socrates-qua-human being' refers to Socrates. And, of
course, in a way it does. Nevertheless Aristotle, unlike Suppe, has reason
to resist the impulse toward ontological reduction. What I intend
'Socrates-qua-human being' to exhibit is not that specimens of kinds
are psychological entities, in other words, ways of thinking about things
and thus idealizations or abstractions from them, but rather that specimens of kinds are not Platonic entities, existing in addition to sensible
objects.
Of course it is one thing to say that in Aristotle's ontology it is entities
such as 'a human being' which are substances, that these are numerically
the same as but not identical with sensible objects, and indeed that they
are ontologically more fundamental than the latter; it is another thing
to show that such a view is defensible or even intelligible. How can
there be numerical sameness without identity? Then too, even if one
grants that there are natural kinds, how can Aristotle take specimens
of these kinds to have ontological priority over the sensible objects with
which they are numerically the same? In the next chapter I will begin
by presenting a contemporary example of numerical sameness without
identity which is also suggestive for addressing the second of these
questions.
27 Owen (1978), 20. Owen's suggestion is endorsed by Lennox (1985), 83, in a paper
written from the perspective of the biological works.
VI
Substance and Teleology
In this book I have argued that the assumption that Aristotle distinguishes numerical sameness from identity provides a wide-ranging explanation of referential opacity in his works and makes possible an
interpretation of substance that sees Aristotle's theory as a response to
what he takes to be the flaws in Platonism. I have not attempted to
defend distinguishing between numerical sameness and identity on philosophical grounds or even to consider the philosophical implications
of such a view; as I said in Chapter II, the logic of a metaphysics that
confounds counting has to be, to say the least, problematic. It may be,
of course, that Aristotle adopted a position that cannot be made coherent or attractive, although such a conclusion would be disappointing.
Although I will not in this final chapter try to offer a philosophical
analysis or defense of the distinction, I will nevertheless describe an
interesting occurrence of it in the recent philosophical literature. But
the primary goal of this chapter is to argue that substances, understood
as specimens of natural kinds, can defensibly be said to be ontologically
prior to the sensible objects with which they are numerically the same,
and for that argument too the example now to be offered will prove
useful.
I
In an extraordinarily interesting book in the philosophy of art, The
Transfiguration of the Commonplace, Arthur Danto puzzles about the re-
lation between works of art and physical objects, putting the problem
into its sharpest focus by considering works of art that are perceptually
indistinguishable from ordinary things. To take Danto's example, Duchamp's Fountain is perceptually indistinguishable from a urinal. Following Danto, let us call the sensible object, in this case a given urinal,
100
101
What "detaches objects from the real world" and makes them part of
the art world, according to Danto, is that art works belong to a world
of interpreted things or representations, and the properties of a work
of art are just those that contribute to its interpretation. 3 For example,
whether gleams, a property of the porcelain, is a property of Fountain
would depend upon its relevance to Fountain's representational content;
even whether Fountain has a color is, according to Danto, to be decided
on the same grounds.4
In the present context it is not necessary to consider what properties
are lost and gained in the "transfiguration" of mere things into art or
even whether Danto's theory of art is correct. What is interesting to
me is that Danto distinguishes the properties of Fountain from the properties of its material counterpart. Fountain is witty, daring, and irreverent, but the urinal is not; the urinal is white, but Fountain may or
may not be so. It must be emphasized that such differences will arise
regardless of whether the material counterparts are everyday objects
such as a urinal or (another of his examples) a snow shovel5 - that is,
objects that have a use outside art - or simply painted canvases and
chunks of marble. Examples of the former sort contribute a certain
vividness to Danto's presentation of his theory, but that is all.
If a work of art and its material counterpart differ in properties, the
relation between them - a relation that Danto dubs the 'is' of artistic
identification6 - is not identity, given that Leibniz's Law is taken to
define that relation. Whether one describes it as a nonstandard identity
relation of some sort or whether one says that there are relations of
numerical sameness other than identity is perhaps not a matter of extreme importance. However, in specifying the extent to which for Danto
a work of art and its material counterpart differ, it is important to notice
that he has taken a stance that is more radical than the nonstandard
1 Danto (1981), 104. See also 99.
2 Danto (1981), 93-94.
3 Danto (1981), 135; see also 125. Danto argues that art works are distinguished from
other representations (for example, maps) by "[their use of] the means of representation
in a way that is not exhaustively specified when one has exhaustively specified what is
being represented" but which, in addition, "expresses something about [their] content"
(148).
4 Danto (1986), 38.
5 Danto (1986), 26.
6 Danto (1981), 126.
102
identity advocated by Allan Gibbard, for example, who uses the same
sort of case to argue that contingent identity can obtain between the
referents of proper names. What Gibbard maintains is that Goliath (a
statue) and Lumpl (a piece of clay), which come into existence simultaneously (by sticking two halves of the statue together) and cease to
exist simultaneously (by being smashed), are identical. Nevertheless,
Gibbard says, the identity is contingent. This is so because, although
the objects in question in fact begin and come to an end simultaneously,
had, for example, the wet clay once stuck together been squeezed into
a ball, the times at which the statue and the piece of clay ceased to exist
would have been different. That is to say, because Goliath could have
been destroyed while Lumpl survived, Goliath and Lumpl differ in
modal properties.7
Certainly Danto and Gibbard do not disagree merely about how the
relationship between Goliath and Lumpl, in the circumstances Gibbard
has described, is to be labeled. Rather what is evident is that Danto
would take issue with Gibbard's description of the case. Although Gibbard has said that Goliath and Lumpl have all the same actual properties,
for Danto, as we have seen, only Goliath would have representational
properties, and, furthermore, Goliath would have only representational
properties. On the other hand, in describing the relation between the
work of art and its material counterpart, Danto is no Platonist.8 Thus
since he does not take the work of art and its material counterpart to
be two distinct things and yet he denies their identity with each other,
Danto has offered a contemporary example of numerical sameness without identity.
The comparison between Danto and Aristotle will bear pressing. But
first it is useful to notice that, despite the term 'material counterpart',
if matter in Aristotle is understood as "in its own right" lacking properties, Danto's material counterpart is not, in Aristotle's sense, matter;
the material counterpart, so Danto tells us, has properties that are not
properties of the form and are unrelated to its potentiality for form.
On the other hand, there seems to be no reason why the material
counterpart should not be said to be a sensible object. In the case of
the urinal and the snow shovel, that it is so is obvious. To be sure, the
material counterpart of a painting is not assignable to any ordinarily
recognized kind since were it said to be a painting, it would have the
representational properties Danto denies to it. But even if the fact that
painting is a kind dependent on certain practices that endow its members
with (indeed on Danto's view with only) a very unusual sort of properties,
namely, representational ones, means that what the painted canvas must
be said to be is just a painted canvas and even if, as a result, the material
7 Gibbard (1975), 190-91.
103
104
105
erties of a work of art, one could say that many different works of art
are artistically identified with the same material counterpart, as many
as there are interpretations. Nevertheless, even though the 'is' of artistic
identification does not mandate any distinction concerning priority and
does not make the selection, as we have seen, it does provide some
grounds for doing so. In particular, because the properties of the work
of art and the properties of its material counterpart are related so as
to make plausible the claim that certain properties of the one are explanatory of and indeed determinative of certain properties of the other
in light of the artist's intention, where art is concerned it is reasonable
to say that specimens of the kinds whose defining properties are of the
first sort are ontologically prior.
In this section the relation between a work of art and its material
counterpart has been used to explore the relation of being numerically
the same as but not identical with, a relation that I believe Aristotle
takes to hold between specimens of kinds and sensible objects, and to
consider the possibility that, given such a relation, one might be justified
in holding that ontological priority lies with specimens of certain kinds.
What I have said about the maker's intention determining the properties
of a work of art could be said of any more ordinary artefact, of course;
the reason for using art as my example has been only that Danto in his
discussion of that subject actually proposes a relation of numerical sameness without identity.
To be sure, there are limits to the analogy - even if a work of art
or a specimen of some artefact-kind can be prior to the sensible object
that is numerically the same as it, works of art and specimens of artefactkinds are hardly ontologically prior simpliciter. Nevertheless, the implications are clear. If Aristotle could claim that some properties, more
precisely, the essential properties, of sensible objects are determined
by the properties of specimens of certain kinds, he could reasonably
maintain that specimens of those kinds are ontologically prior to the
sensible objects with which they are numerically the same and also to
specimens of other kinds, in other words, to accidental unities. In particular, if such an argument were successful, Aristotle might be able to
conclude, for example, that specimens of the kind human being are and
that specimens of the kind builder are not ontologically prior to individuals such as Callias and Socrates.11 But, of course, on Aristotle's
account, nature is rather like a work of art - or at least like an artefact
11 Though his account of substance is different from mine, Irwin (1988) too makes a
case for the ontological priority of some subjects over others. He says: "If we see the
connexion between potentialities and persisting subjects of change, we can also see
why some subjects (e.g. men) are more genuine subjects than others (e.g. musical
men). A persistent subject needs a persistent potentiality explaining the changes that
happen to it. If we find that rather little is explained by reference to permanent
potentialities of musical men as such, and more is explained by reference to permanent
106
ditions of his existence, or, if we cannot quite say this then the next thing to
it, namely, that it is either quite impossible for a man to exist without them,
or, at any rate, that it is good that they should be there. And this follows: because
man is such and such the process of his development is necessarily such as it is; and
therefore this part is formed first, that next; and after a like fashion should we
explain the generation of all other works of nature, (emphasis added)
As this passage and many others make clear, the presence of teleological claims in Aristotle's writings is undeniable. However, the analysis
potentialities of men as such, we have some reason for thinking that men are more
genuine subjects than musical men are. Aristotle claims in Metaphysics vii 4-6 that
some, but not all, apparent subjects are thises and primary subjects, identical to their
107
108
human activities and artefacts. But the view that teleological claims are
metaphorical is often also the starting point for unpacking and justifying
their explanatory worth in terms of features other than the presence
of a mind. Working from the assumption that for Aristotle intention
or mind is to be excluded from the account, in the remainder of this
section I will consider one such externalist account, 13 that of Larry
Wright, in order to see how it might be adapted to Aristotle's needs.
Wright, who takes himself to be giving the content of what he says
is a dead metaphor, analyzes goal statements as follows:
S does B for the sake of G iff
1) B tends to bring about G
2) B occurs because (i.e., is brought about by the fact that) it tends to
bring about G.14
Thus, for example, spiders spin webs for the sake of catching flies if
and only if spinning webs tends to bring about the catching of flies and
spinning webs occurs because it tends to bring about the catching of
flies. Function statements are analyzed in similar fashion:
The function of X is Z iff:
1) Z is a consequence (result) of X's being there
2) X is there because it does (results in) Z.15
But Aristotle, as we have seen, also thinks of organic development teleologically; still for such cases it seems that Wright's schema could be
extended in the following way:
E exists for the sake of O iff:
1) E tends to develop into O
2) E exists because it tends to develop into O.
Wright's analysis, according to which "X is there because it results in
Z," sounds not unlike Aristotle's claim that the fittest mode of explanation is to say that such and such parts are necessary for an organism's
existence. Additionally, Andrew Woodfield argues persuasively that,
since externalists such as Wright in their talk of goal-directed systems
identify only some outcomes as G or Z (for example, those that contribute
to the survival of the organism), the outcomes in question have been
assumed to be good. 16 In fact Wright does talk of "resultant advantage," 17 and, as we will see, that Aristotle would hold that goodness
cannot be a coincidental feature of the outcome is beyond doubt. Thus,
for an analysis at least of Aristotle, something about goodness needs to
be added to the schemata, along the following lines:
13 The distinction between internalist and externalist analyses comes from Woodfield
(1976), 104-6.
14 Wright (1976), 39.
15 Wright (1976), 81.
16 See Woodfield (1976), 100-11, 113-23, 130-40, 203-6.
17 See Wright (1976), 85.
109
As for the case of organic development, since the very notion of development seems to assume the worth, at least from some point of view,
of the resulting organism, although one could add as a third claim "E
develops into O because its development into O is good," one could
argue also that it is not necessary.
It is one thing, of course, to produce an analysis of how an author
uses teleological claims, another to agree that he is entitled to use them
as he has. In Wright's analysis, this question turns on whether 'because'
is justified. Woodfield attributes our willingness to say that X occurs
because it results in Z to our imagining practical reasoning to have occurred.19 Wright himself, on the other hand, justifies its use on the
ground that explanation is contextual.20 But neither of these justifications is available to Aristotle, who argues, after all, that teleological
explanation is for nature the fittest mode of explanation. Nor is the
currently popular evolutionary justification of teleology available to
him,21 given his belief in the permanence and unchangingness of species
(Generation of Animals II 1 731b24-732al).22 To be sure, it could be
that Aristotle's attempt to make nature purposive in the absence of
purpose in fact fails. However, in the following sections I will argue
that Aristotle is aware of the need to justify teleological explanation in
nature and addresses himself to the question in a way that, given certain
assumptions about goodness, is successful. In Section IV I will consider
how benefit and agency are connected in the craft analogy.
18 Cooper (1982), 197, says: "[Aristotle] understands by a goal (hou heneka) whether
natural or not, something good (from some point of view) that something else causes
or makes possible, where this other thing exists or happens (at least in part) because
of that good."
19 Woodfield (1976), 135.
20 See Wright (1976), 61.
21 See Millikan (1984), 17-49. The same is true also of Bigelow and Pargetter (1987)
insofar as their propensity theory (". . . what confers the status of a function is not
the sheer fact of survival-due-to-a-character, but rather, survival due to the propensities the character bestows upon the creature" (192)) is superimposed on the assumption of biological change across generations.
22 See Cooper (1982), esp. 202-5.
110
IV
Although Aristotle knows that teleological explanations in nature at least
appear to imply agency, he seems to reject the implication. In Physics
II he says:
Now action is for the sake of an end; therefore the nature of things also is so.
Thus if a house, e.g., had been a thing made by nature, it would have been
made in the same way as it is now by art; and if things made by nature were
made not only by nature but also by art, they would come to be in the same
way as by nature. The one, then, is for the sake of the other; and generally art
in some cases completes what nature cannot bring to a finish, and in others
imitates nature. If, therefore, artificial products are for the sake of an end, so
clearly also are natural products. The relation of the later to the earlier items
is the same in both. This is most obvious in the animals other than man: they
make things neither by art nor after inquiry or deliberation. (199al 1-21)
Again, he declares:
It is absurd to suppose that purpose is not present because we do not observe
the agent deliberating. Art does not deliberate. If the ship-building art were in
the wood, it would produce the same results by nature. If, therefore, purpose
is present in art, it is present also in nature. The best illustration is a doctor
doctoring himself: nature is like that. (199b26-32)
Since crafts are practiced by agents with conscious desires and plans
and since this is precisely the sticking point in the application of teleology to nature, the comparison of nature to craft is usually dismissed
as unhelpful. But recently Sarah Broadie has argued that the analogy
could be made interesting if it were agreed that craft is not animated
by desire. Just as a tree that desired to be a tree would be more than
a tree, so, the analogy goes, a builder qua builder already pursues the
builder's ends.23 On this view of it, the purpose of the analogy is not
to attribute desire to nature but rather, having denied its presence in
craft, to show that nature like craft exhibits regularity and also that it
makes correct moves - that craft qua craft does not err.24
Thus Broadie defends the craft analogy from the usual charge against
it. She also contends, however, that even so, the comparison fails. The
reason why it fails, she argues, is that Aristotle in his use of craft to
illuminate nature has so distorted the notion of craft that it is almost
unintelligible; whereas the real craftsman aims at a good that answers
some human need or interest recognized as a result of practical wisdom
or reflection and refines his product accordingly,25 to serve the purpose
of the analogy, craft knowledge must be thought of as developed with
a view to its purpose only and not as a capacity that could be put to
23 Broadie (1987), 43-44.
25 Broadie (1987), 48.
111
112
Xing is a good. Indeed when Broadie criticizes the craft analogy on the
ground that without a context of human needs and interests there is
no background against which to judge the value of the craft activity,
perhaps her point can be put by saying that, insofar as the craft analogy
is intended to be its support, Aristotle's appeal to final causes in nature
fails; without a context, nature too becomes no more like good doctoring than like bad doctoring and no more like doctoring of any sort
than like thievery.
But even if the craft analogy is not very helpful for the justification
of his use of teleological explanation in nature, the ways in which the
notion of craft must be stretched in order for there to be an analogy
are themselves illuminative, and it is not self-evident that Aristotle's
view of nature is false. The health of an oak tree is, after all, something
that is objectively determinable, and insofar as health (over a certain
span of years) is the norm and disease often overcome, surely nature,
in this case the nature of oaks, can be said to be more like good doctoring
than like inept or corrupt doctoring. Of course, it might be objected
that, since healthy ivy flourishes at the expense of trees, in the case of
ivy nature is rather like thievery, but even here it is not clear that
Aristotle would have no recourse. For thieves are persons who work
against the flourishing of their fellow human beings and indeed, on an
Aristotelian view, against the flourishing of their own true selves, while
ivy is just doing what ivy flourishes by doing; though thieves have an
unactualized good, ivy is not separated from the good for ivy. To put
the point another way, only if thieves were, like human beings, dogs,
and ivy, a natural kind would the objection be well taken.
I did not promise a defense of Aristotle's view that there are natural
kinds, and beyond what was suggested in the second section of this
chapter I could not deliver on such a promise; neither do I have any
novel criteria for judging of any given kind whether it is natural. But
that Aristotle thinks there are natural kinds, that he gives some thought
to which they might be (even if he does not give us entirely successful
criteria), and that we in reading him can see, at least for many cases,
what kinds he thought natural is true nonetheless. Metaphysics VII 4 is
a case in point. There Aristotle, needing to say what sorts of things have
an essence, distinguishes compounds such as white man from things
that are something in their own right (kath' hauto) and tries to explain
what is characteristic of the former. He says:
One kind of predicate is not said of a thing in its own right because the term
that is being defined is added to something else, e.g. if in denning the essence
of white one were to state the formula of white man; another because something
else is added to it, e.g. if 'cloak' meant white man, and one were to define cloak
as white; white man is white indeed, but its essence is not to be white. (1029b311030a2)
113
114
In Physics II, as we have seen, Aristotle says that the accidental includes
the conjunction of a substance and an accidental property (the musical
man) or of a pair of properties (musicality and housebuilding), as well
as the causal connection between an agent described by means of some
conjoined property and an outcome (the fluteplayer builds a house).
As he says of this last case, 'Tor just as a thing is something either in
virtue of itself or accidentally, so may it be a cause. For instance, the
housebuilding faculty is in virtue of itself a cause of a house, whereas
the pale or the musical [human being] is an accidental cause" (196b2427).
But chance too, Aristotle thinks, is part of the accidental (198a5-7)
since it involves both causation and conjunction. More specifically, Aristotle classifies chance as that part of the accidental where there is an
outcome that is unintended but beneficial. As he says, "Hence it is clear
that events which belong to the general class of things that may come
to pass for the sake of something, when they come to pass not for the
sake of what actually results, and have an external cause, may be described by the phrase 'from chance"' (197bl8-20). Unlike the narrower
category of luck (he tuche), for something to occur by chance (to automaton) it is not necessary that it be done by or happen to a being capable
of choice; if a man goes to the marketplace and happens to meet his
debtor, that is luck (196b33-197a3), but if a tripod falls and lands
upright (197bl6-18) or a stone falls and strikes someone (197b30-32),
it is chance.30
With this analysis of chance, Aristotle is ready to argue against those
who claim that the structure of plants and animals is due to the random
29 Kahn (1985), 197.
30 Though like Ackrill (1981), 36-41, I prefer 'luck' and 'chance', it should be noted
that the Greek terms are variously translated. Barnes (1984), uses 'chance' where there
is a being capable of choice and 'spontaneity' for the broader category where there
need not be. Charlton (1970) contrasts 'luck' with 'the automatic'. I find 'luck' very
natural for the case involving an agent; however, I am less happy with either 'the
automatic' or 'the spontaneous' for the broader category because neither captures
the central point: that something has happened which because it has a good result
has the appearance of having been done for the sake of that result.
115
116
117
teleology in Aristotle this fact is important. For it has long been noticed
that Aristotle frequently offers two explanations, one involving the characteristics of matter, the other those of form. For example, he says that
the deer sheds its horns both for relief and because of their weight
{Parts of Animals III 2 663bl2-14), that human heads are hairy both
for protecting the brain from excessive heating and chilling and because
of the fluidity of the brain and the sutures in the skull, which are such
that heat and fluid produce the outgrowth of hair (II 14 658b2-10),
and that eyelashes exist for the protection of the eyes and because they
are located at the end of small blood vessels where moisture comes off
(II 15 658bl4-26). Likewise he holds that the ability of serpents to
turn their heads backward enables them to guard against attacks from
the rear and is a necessary consequence of their cartilaginous and flexible vertebrae (IV 11 691b31-692a5) and that the webbed feet of waterbirds are useful in enabling them to swim and are a necessary consequence of their residual earthy substance (IV 12 694a23-bl2).
These examples and others are too numerous to be ignored. Nevertheless from such passages alone it is unclear exactly what Aristotle
takes the relation between the two explanations to be. For example, if
the fluidity of the brain is a sufficient explanation for eyelashes, could
it not be objected that even if eyelashes did not serve the animal's good,
they would exist anyway? But if it were true that eyelashes would exist
regardless of their usefulness, then, even given that eyelashes are in fact
useful, the appeal to benefit would seem to have no force. To this
objection - namely, that where there is a sufficient material condition,
the value of the result is causally irrelevant - there are several possible
lines of reply. That Aristotle intends teleological explanation to be, like
chance, compatible with mechanical necessity and yet, unlike chance,
more fundamental than explanation in terms of matter will be argued
in the next section. There I will argue also that it is final causality that
grounds the ontological priority of specimens of natural kinds.
VI
In response to the worry that given a necessitating material cause the
teleological explanation is not really explanatory, Martha Nussbaum at
one time proposed that double explanations are pragmatic. Against his
mechanist predecessors she imagined that Aristotle would have argued
along the following lines: "If you were a shepherd in charge of the
flocks, which account [Democritus's or Homer's] would give you more
information that was relevant to your plans and precautions? From eight
lines of Homer I learn more that is general and valuable about the
118
But Aristotle does not believe merely that final and material causes
are compatible (assuming that he does believe them to be so); he believes
teleological explanations to be more fundamental than material explanations. That is to say, beyond holding that the features explained by
a physical account and the features explained by a teleological account
are distinct, Aristotle claims further that material necessity is hypothetical; it is because the organism is what it is that it must be made of matter
having certain characteristics. Of Aristotle's view Charles says:
33 Nussbaum (1978), 71.
34 Balme (1939), 137-38. See also Cooper (1982), 211.
35 Charles (1988), 38. Charles falls roughly into the functionalist camp (even if physical
states are sufficient for the outcome the behavior could in principle have been differently realized). Concerns of the sort Ackrill (1973) presents need not undermine
this view. See Chapter V, n. 10, of the present book.
37 Charles (1988), 40.
36 Charles (1988), 38-39.
119
It is in virtue of possessing these goals that man has the nature he has; and it
is because he has this nature that he has the potentiality to <f> and must <p in C.
For if he failed to do so, he would fail to be the creature he is. So far from
teleological causation resting on efficient causation, it is rather the reverse. The
presence of goals makes the organism what it is, and its being that organism
explains why it must <>
/ in C. If this is correct, teleological causation is not
explained in terms of efficient causal necessitation. The one is irreducible to
the other. Teleological goals are taken as primitives. These latter concerns show
why Aristotle preferred the downwards-perspective. The nature of the kind is basic,
and this is fixed by primitive teleological factors.38
120
not refer to either this or that member of the kind, but to any member
of the kind, taken as representative of the others. The end - the work
of art - does not include that which differentiates one print from another.
In short my proposal is that, just as in the case of art there are
preferred explanations of the features of the work of art which explain
some but only some of the features of its material counterpart, so for
Aristotle an adequate explanation of the goodness of certain activities,
structures, and patterns of development is an explanation only of what
is characteristic of specimens of natural kinds, and not of the more
finely differentiated activities and structures of individual organisms.
But just because this is so, teleological explanation can reasonably be
said properly to apply to specimens of natural kinds rather than to the
individual organisms with which they are numerically the same. To be
sure, living a lion's life is done by some particular lion in some particular
way; this lion is at this moment stalking this zebra. Nevertheless such
detail is not part of the specification of the end, and just because it is
not, the final cause remains at a level of generality commensurate with
the generality of specimens of a kind, thereby in the case of specimens
of natural kinds, justifying their ontological priority.
VII
In these chapters I have defended a version of the view that substance
is form. On my view of Aristotle, what there most fundamentally is in
the universe is form; ontological priority belongs to specimens of natural
121
Bibliography
The Greek texts that I have used for Plato and Aristotle are the Oxford
Classical Texts.
Ackrill, J. L. 1963. Aristotle's Categories and De Interpretatione. Trans, with notes.
Oxford: Clarendon Press.
1973. "Aristotle's Definitions of Psuche." Proceedings of the Aristotelian Society
73: 119-33.
1981. Aristotle the Philosopher. Oxford: Oxford University Press.
Albritton, Rogers. 1957. "Forms of Particular Substances in Aristotle's Metaphysics." Journal of Philosophy 54: 699-708.
Press.
1961. "Aristotle." In Anscombe and P. T. Geach, Three Philosophers.
Balme, David M. 1939. "Greek Science and Mechanism: 1. Aristotle on Nature
and Chance." Classical Quarterly 33: 129-38.
1984. "The Snub." Ancient Philosophy 4: 1-8.
Barnes, Jonathan. 1975. Aristotle's Posterior Analytics. Trans, with notes. Oxford:
Clarendon Press,
ed. 1984. The Complete Works of Aristotle: The Revised Oxford Translation. 2
124
BIBLIOGRAPHY
Charlton, W. 1970. Aristotle's Physics Books I and II. Trans, with introduction
and notes. Oxford: Clarendon Press.
Cherniss, Harold. 1944. Aristotle's Criticism of Plato and the Academy. Vol. I.
versity Press.
Dooley, W. E., trans. 1989. Alexander of Aphrodisias, On Aristotle's Metaphysics
1. Ithaca, N.Y.: Cornell University Press.
Driscoll, John. 1981. "EIAH in Aristotle's Earlier and Later Theories of Substance." In Studies in Aristotle. Ed. Dominic O'Meara. Washington, D.C.:
Catholic University of America Press.
BIBLIOGRAPHY
125
126
BIBLIOGRAPHY
Clarendon Press.
1986. "Aristotelian Actions." Phronesis 31: 68-89.
1988. Aristotle's First Principles. Oxford: Clarendon Press.
Kahn, Charles. 1981. "The Role of nous in the Cognition of First Principles in
Posterior Analytics II 19." In Aristotle on Science: The 'Posterior Analytics'. Ed.
Press.
Loux, Michael. 1979. "Form, Species and Predication in Metaphysics Z, H, and
0." Mind 88: 1-23.
1991. Primary 'Ousia': An Essay on Aristotle's Metaphysics Z and H. Ithaca, N.Y.:
Cornell University Press.
Mabbott, J. D. 1926. "Aristotle and the XQPISMOS of Plato." Classical Quar-
BIBLIOGRAPHY
127
128
BIBLIOGRAPHY
Ross, W. D. 1949. Aristotle's Prior and Posterior Analytics. A revised text with
introduction and commentary. Oxford: Clarendon Press.
1951. Plato's Theory of Ideas. Oxford: Clarendon Press.
Sellars, Wilfrid. 1957. "Substance and Form in Aristotle." Journal of Philosophy
54: 688-99.
1967. "Aristotle's Metaphysics: An Interpretation." In Philosophical Perspectives. Springfield, 111.: C. C. Thomas.
Sharpies, R. W. 1985. "Species, Form, and Inheritance: Aristotle and After."
In Aristotle on Nature and Living Things. Ed. Allan Gotthelf. Pittsburgh:
Mathesis Publications.
Smith, J. A. 1921. "TOAE TI in Aristotle." Classical Review 35: 19.
Sorabji, Richard. 1980. Necessity, Cause, and Blame: Perspectives on Aristotle's
Theory. Ithaca, N.Y.: Cornell University Press.
1992. "Intentionality and Physiological Processes: Aristotle's Theory of Sense
Perception." In Essays on Aristotle's 'De Anima.' Ed. Martha Nussbaum
and Amelie Rorty. Oxford: Clarendon Press.
Strang, Colin. 1963. "Plato and the Third Man." Proceedings of the Aristotelian
Society, Suppl. Vol. 37: 147-64.
Suppe, Frederick, ed. 1974. The Structure of Scientific Theories. Edited with introduction. Urbana, 111.: University of Illinois Press.
Vlastos, Gregory. 1954. "The Third Man in the Parmenides. "Philosophical Review
63: 319-49.
1965. "Anamnesis in the Meno." Dialogue 4: 143-67.
1991. Socrates: Ironist and Moral Philosopher. Ithaca, N.Y.: Cornell University
Press.
White, Nicholas. 1971. "Aristotle on Sameness and Oneness." The Philosophical
Review 80: 177-97.
1976. Plato on Knowledge and Reality. Indianapolis: Hackett.
Wiggins, David. 1967. Identity and Spatio-Temporal Continuity. Oxford: Basil
Blackwell.
Wilkes, Kathleen. 1992. "Psuche versus the Mind." In Essays on Aristotle's 'De
Anima.' Ed. Martha Nussbaum and Amelie Rorty. Oxford: Clarendon Press.
Witt, Charlotte. 1989. Substance and Essence in Aristotle: An Interpretation of 'Metaphysics' VII-IX. Ithaca, N.Y.: Cornell University Press.
Wollheim, Richard. 1968. Art and Its Objects: An Introduction to Aesthetics. New
York: Harper & Row.
Wolterstorff, Nicholas. 1970. On Universals: An Essay in Ontology. Chicago: University of Chicago Press.
Woodfield, Andrew. 1976. Teleology. Cambridge: Cambridge University Press.
Woods, M. J. 1967. "Problems in Metaphysics Z, Chapter 13." In Aristotle: A
Collection of Critical Essays. Ed. Julius Moravcsik. Garden City, N.Y.: Doubleday.
1975. "Substance and Essence in Aristotle." Proceedings of the Aristotelian
Society 75: 167-80.
1982. Aristotle's Eudemian Ethics Books I, II, and VIII. Trans, with commentary. Oxford: Clarendon Press.
Wright, Larry. 1976. Teleological Explanations. Berkeley: University of California
Press.
Index
flux
knowledge, and, 19, 78
related to separated Forms, 8-11, 92
substances, and, 96
form
and essence, 456
indistinguishability within kinds, 55, 61
predication (Lewis, Loux), 53-5
senses of, 45-6, 55
specimens of natural kinds, 47, 51
Form of the Good (Plato)
teleology of Aristotle's counterpart to,
116
Forms (Plato)
Aristotle's criticism of separation of,
5-20, 92
capacity for independent existence,
7-14
importance for Aristotle of their
unchanging nature, 3, 121
obstacles to knowledge, 16-19, 63-8,
87
as particulars and universals (Aristotle's
interpretation), 14-19
regress arguments against, 15-16, 38,
87
Third Man Argument, 15-16
as types, 61, 77
universal or particular, 57-61
Frede, Michael, 89
Frege, Gottlob, 24-5, 35
Furth, Montgomery, 44
Geach, P. T., 59
geometrical entities, 29
Gibbard, Allan, 102
129
130
INDEX
INDEX
Sorabji, Richard, 26
specimens of kinds
accidental unities as, 30-2, 34
definition, 48-9
knowability, 2, 50, 55-6, 61, 63
knowledge of, 77-82
sensible objects, and, 30-1, 37, 88
specimens of natural kinds, and, 30,
104-5
specimens of natural kinds
as epistemologically fundamental, 50
as ontologically fundamental, 3, 31,
97_9, 103-6, 119-21
substances as, 40, 50, 55-6, 61-3, 86,
104, 120-1
See also specimens of kinds
Strang Colin, 58
substances
alleged inconsistency of view, 45, 51-5
candidates for, 44-6, 83-6
in the Categories, 40-3, 56
criteria for, 44-5, 83-5, 91
as specimens of natural kinds, 40, 50,
55-6, 61-3, 86, 97-9, 120-1
See also specimens of natural kinds
substitutivity
accidental sameness, and, 25-8, 31-5,
37
essential sameness, and, 36-9
failures of, 22-8, 34-7
as principle governing identity, 22
Suppe, Frederick, 98-9
131
teleological explanation
art, and, 105
chance, and, 114-17
craft analogy, 110-12
downward perspective, 119-20
externalist accounts, 108-9
the good, and, 111, 113-16
internalist accounts, 106-7
material causes, and, 117-19
natural kinds, and, 112-13
Third Man Argument
as in Alexander of Aphrodisias, 15-16
nature of Forms, and, 57-60
separation, and, 56-7
types
Forms as, 60-1
substances as, 61-2, 74-88
universals
knowability of, 14, 19, 51, 61
knowledge of, 74-82
separated Forms as, 7, 9-10, 14-16,
57-60, 83-4
unknowability
of accidental characteristics, 14, 16-18,
46-7
recollection, and, 18-19, 63-8
Vlastos, Gregory, 13
White, Nicholas, 24, 65
Witt, Charlotte, 84
Woodfield, Andrew, 108-9
Woods, M. J., 57
Wright, Larry, 108-9