Some Problems about Being and Predication in Plato's

Sophist 242-249
William B. Bondeson

Journal of the History of Philosophy, Volume 14, Number 1, January

1976, pp. 1-10 (Article)
Published by Johns Hopkins University Press
DOI: 10.1353/hph.2008.0639

For additional information about this article

http://muse.jhu.edu/journals/hph/summary/v014/14.1bondeson.html

Access provided by Universidad Complutense de Madrid (2 Sep 2015 13:22 GMT)

Some Problems about Being and

Predication in Plato's
Sophist 242-249
WILLIAM BONDESON

ONE OF THE CENTRAL TASKSwhich Plato sets for himself in the Sophist is to say what
being ( ~ 5v) is. In doing this he makes a variety of philosophical moves. The first is to
show that non-being in a very restricted sense of the term (T6 g~qSag~g ~v) is an impossible and self-contradictory concept. 1 This occupies the first part (237A ft.) of the central
section of the Sophist. After discussing some puzzles concerning deceptive appearances
(240 B) and falsehoods (240 D), Plato turns to a discussion of being at 242B. In this
section of the dialogue Plato claims to show that the attempts of previous philosophers to
define being have failed and he makes his own first attempt in the dialogue to define being
(cf. 242C and 247E). 2 In this paper I am concerned only with this section of the Sophist
(242-249), and I want to show first that Plato's notion of being here is ambiguous, the
term x6 5v shifting between "being" and "what has being," between the form and those
things which participate in it. Second, I want to show that the definitions of being at 248C
and 249D are not only compatible with one another but also that, when properly understood, they make sense of Plato's use of motion and rest in the Sophist. And finally, I
want to show that Plato is caught in the snares of self-predication when he talks about
being and other Forms of the same ontological level. This is due to the way in which he
formulates the difference between statements of identity and predication in the argument
against Parmenides in this section of the Sophist.
Plato begins his discussion of being in the Sophist in a quasi-historical way by dealing
with his predecessors' attempts to characterize it. The statement at 242C4-6 serves as an
outline of the arguments which follow: Parmenides and the other philosophers have attempted to determine the number (~ttoa) and character (~o{a) of the things which are
called T& 5wa. Those who have tried to determine the "number" of T& 5wct are the
monists and pluralists in 2 4 3 A - 2 4 5 E ; those who have tried to determine their "character" are the materialists and "idealists" in 2 4 6 A - 2 4 9 D . Plato appears to be arguing that
1 Cf. my "Non-being and the One: Some Connections between Plato's Sophist and Parmenides," forthcoming in Apeiron. My view is somewhat different from that of G. E. L. Owen's
"Plato on Not-Being" in Plato, A Collection of Critical Essays, ed. G. Vlastos (Garden City, New
York: Doubleday and Co., 1971), vol. I.
2 Cf. Owen, ibid. p. 229, n. 14. Owen presents a convincing case that Plato is giving a definition (as opposed to a mark or sign) of being. However, Owen also seems to take the view, for
example against Moravcsik in Being and Meaning in the Sophist (Acta Philosophica Fennica,
XIV [1962]), that little of philosophical significance happens in 242-249. I hope to show in this
paper that this is not the case.

[1]

HISTORY OF PHILOSOPHY

defining or characterizing x& ?Jvx&is the same as stating what entites are the proper subjects of discourse. In another way such a characterization is an attempt to determine what
things can be said to be. This amounts to a definition of being itself. It seems to be at least
part of the purpose of the following arguments to separate out not only the various senses
of "is" (in the limited way in which Plato does this), but also to shed some light on the
distinction between being (x6 ~v) and things which have being (z& ?Jr-ca).
That this latter distinction is not yet made is shown from the double way in which the
basic question in this section is raised; i.e., to ask what the number and character of
r& ~wa are is not distinguished from asking what the term xb ?Jr signifies. For the present,
a conceptual terminology will serve to state the problem: Plato views his predecessors'
work as attempts to define being in terms of a particular concept or set of concepts.
Thus the pluralists (243D6-224B4), those who want to define being in terms of a pair
of concepts, are represented as maintaining that "everything" or "the universe" (T&
;xrvrct) is hot or cold. What the "is" means is the difficulty. It could mean, as a physical
theory, that everything "is made up of" these two elements (which might be more accurate historically), but Plato interprets these pluralists as maintaining the identity of hot
and cold with being. Or in another way, they have attempted to limit the meaning of "to
be" to "to be hot" and "to be cold. ''8 If these do not have the same meaning, then, Plato
argues, hot, cold and being are distinct and everything wiU be more than the original pair.
The argument is brief and not a good one. Whatever distinction there is to be made depends on giving an account of being and this is just what is sought. It should be noticed
here that the distinction rests upon common sense, a feature of this and succeeding arguments which will be discussed later in this paper. But the argument as it stands is scarcely
effective against disjunctive definitions. If being is identified with hot and cold, Plato
argues, then they are one and no longer two. Being must be some third thing. As Moravcsik points out, this would not convince his opponents. 4
Plato does not spend much effort on the pluralists, probably because his own account
of being is pluralistic, although in a wider sense which will be discussed shortly. He takes
much greater pains with the monists and, in particular, with Parmenides in 244B6245E2. This section can be divided into two arguments; the first (244B9-D13) is directed against a literal interpretation of Parmenides' monism, the second (244D11245E2) is directed against Parmenides' account of his unitary entity as identical with
being. The second argument is concerned also with distinguishing being, wholeness and
unity as concepts and, ultimately, as distinct Forms.
The first argument can be easily paraphrased. Parmenides posits a unique entity but in
so doing he admits two names or descriptions for that entity. It can be called "being"
and "one." But to say this entails that there are two terms and a single entity described
by them, in effect three things instead of Parmenides' original one. 5 This bears a resemblance to the first hypothesis of the Parmenides, i.e., if there is only a single entity, and
that entity is a pure unity, then that entity cannot have a name (cf. Parm. 142A3-6 and
8 Cf. A. E. Taylor, trans, and Intro., Plato: the Sophist and Statesman, ed. R. Klibansky and
E. Anseombe (London: Thomas Nelson and Sons, 1961), p. 138, n.
4 Being and Meaning, p. 29.
5 Cf. Taylor's paraphrase, Sophist and Statesman, p. 139, n.; Moravcsik's comments, Being
and Meaning, pp. 30-31; and J. H. Ackrill, "~TMIIAOKH EIA~N," reprinted in the Vlastos
anthology (supra.), p. 204.

BEING

AND PREDICATION

1 6 4 A 7 - B 2 ) ; there can be no discourse a b o u t such an entity. ~ T h e argument in the

Sophist also rests on a c o m m o n sense distinction between t e r m and thing, one which recurs throughout the Sophist. ~
T h e second argument is m o r e complex; it begins with the disjunction: P a r m e n i d e s '
unitary being either is or is not a whole ( 2 4 4 D 1 4 - 1 5 ) . 8 B u t the Eleatic Stranger argues
f r o m the hypothesis that being is a whole, leaving aside whatever connections P a r m e n i d e s
might have maintained between being and his unique entity.
1. If being is a whole, then it must b e a whole which has parts. ( 2 4 4 D 1 4 244E7)
This follows from the description of being in P a r m e n i d e s ' q u o t e d statements. But " b e ing" is ambiguous. It might m e a n either what has being or being itself. Thus statement 1
m a y m e a n either that everything which has being is a whole and has parts or that being
itself is a whole and has parts. Although P l a t o does not yet d r a w such a distinction, this
does not hinder his distinguishing being, unity, and wholeness because, as will be p o i n t e d
out later, it is ultimately based on the distinction between w h a t " t o b e , " " t o b e one," a n d
" t o be a whole" mean. Parmenides, in his confused way, seems to have thought that
these three (no matter in what w a y the distinction is stated) were identical and this i m plied for him, because he did not distinguish between identity and predication, that his
unique entity (whether we call it "being" or "the one") was the only genuine entity. Plato,
- - 6 0 w e n points out that the first hypothesis of the Parmenides can also be construed as Plato's
own criticism of his earlier theory of Forms: "So the deductions of the second part [of the
Parmenides] begin with a negative movement because the general effect of that movement is to
show the bankruptcy of one way of dealing with unity which had been characteristic of the theory
of Forms brought up by Socrates in the first part (129A-E; cf, Rep. 525E). Just as a dyer's sample
of vermilion might be a piece of cloth having that and no other color, so it had been thought
that in a higher world unity could be represented by a Form so paradigmatically unitary as to
have no sort of plurality in it at all. The notion was helped by identifying the Form with the
number 1 (Rep. 525E). How can I be another number of anything? But then how can it even be
defined by any conjunction of properties? The question belongs to the lumber-room of philosophy
partly because the movement IA [Owen's number for the first hypothesis of the Parmenides] was,
inter alia, the necessary clearing operation" ("Notes on Ryle's Plato," in Ryle: A Collection of
Critical Essays, ed. O. P. Wood and G. Pitcher [Gerden City, New York: Doubleday and Co.,
1970] p. 344).
r Moravcsik is correct in maintaining against Cornford that "in attacking Parmenides Plato
(has not) presupposed his own theory of Forms. This would have been rather odd. What Plato
does assume is that there is knowledge and that we apply terms to entities. The activities of the
average intelligent person presuppose a world which consists of a plurality of intelligible entities.
In taking this as his starting point, and in some of the subsequent arguments, Plato bases his case
on common sense. Throughout the dialogue he fights claims that go against common-sense. Such
claims are: that there is no plurality, that there are no meaningful statements, and that there is no
meaningful falsehood. It is characteristic of Plato that he thought it necessary to construct elaborate metaphysical defenses of common-sense truths" (Being and Meaning, pp. 30-31). But it does
not follow from this that Forms are not very close to the surface. For Plato to distinguish between terms or concepts is also to distinguish between Forms. Moravcsik's remarks and mine are
to be applied to the second argument against Parmenides also.
8 This statement as it stands is ambiguous. It might mean either that Parmenides' unitary entity is or is not to be identified with wholeness or that it is or is not a whole in the predicative
sense. The Stranger takes it predicatively and argues from this point. This is one instance where
Parmenides can be refuted because he has not distinguished between the "is" of identity and the
"is" of predication.

HISTORY OF PHILOSOPHY

on the other hand, is attempting to sort out the meanings of these terms, or, in the material mode, to sort out the relevant properties. Whether he can legitimately do this without distinguishing between being, unity, and wholeness and the entities which are one,
whole, etc. will be discussed later.
2. Something may have the attribute (nd0og) of unity in so far as it is one
whole and sum of its parts. (245A1-3)
3. But this cannot be identical with unity because "to be one" does not mean
"to be many" (or "to be made up of many parts"). (245A8-B2)
These two steps serve to distinguish unity from plurality. This should not be taken as implying that unity (as might be inferred from 245A5-6) is one "simply" or "truly." Such
an entity seems to be ruled out in the first hypothesis of the Parmenides. The more general point is that everything is both many and one but not, of course, in the same sense.
The next steps distinguish being and wholeness in the form of a disjunction.
4a. Either being has the attribute (7r60o;) of unity and is one and a whole
or
4b. being is not a whole (and therefore not one whole). (245B4-5)
5. If being has the ~xt0o; of unity, it is not identical with unity and there is
more than one entity. (245B7-9)
Statement 5 gives an instance of a rule which appears to be basic to many of Plato's arguments both here and elsewhere, i.e., if x has the :tt0o; of y, then x is not identical with y.
Although some commentators have maintained that Plato did not distinguish between
identity and predication, it is not difficult to see that the distinction is being used in statements 2 and 3 and in the consequent of 5. The rule thus allows for a distinction between
being and unity and this amounts to a refutation of Parmenides' original thesis that there
is only a single entity. Hence to this point there is a distinction, in the first place, between
unity and wholeness (as is implied by statements 2 and 3), and second, there is a distinction between being and unity. What remains is a distinction between being and wholeness
and the confirmation of the original premise that being is a whole but not identical with
it.
6. If being is not a whole by having the :t~0o~ of wholeness and there is such
a thing as wholeness, then being will turn out to be incomplete (i.e., it will
not be a whole). (245C1-6)
7. If being and wholeness each have their own nature and are not identical,
then there are more things than one. (245C8-9)
The point of statement 6 is a repetition of 4b, but this is clearly what Plato wants to
deny. It appears that 245C1 l - D 1 0 is designed to establish this denial on the grounds
that any entity is a whole in some sense or else quantitative predicates could not be applied to it. This not only denies statement 6, but it also affirms the original hypothesis of
statement 1. Being, then, is a whole and has the ~t&0og of wholeness. And, by the rule
stated above that if x has the ~td0og of y, then x and y are not identical, then if being has
the 7t~t0og of wholeness they are non-identical and each has its own "nature" (statement
7). From this follows the refutation of Parmenides' unitary entity. 9
The above argument contains many difficulties which must be explored. But before
doing this, I shall discuss the arguments against the materialists and idealists in 2 4 6 A 249D since here Plato lays down a definition of being. The ease against the materialists
is not a sound one. They admit that souls are ~;v~t (246E6) and they also admit that
9 This analysis owes much to Moravesik (Being and Meaning, pp. 3 liT.) but differs from his
at:count in several points of detail. The fundamental difference is in the statement of the "rule."

BEING AND PREDICATION

whatever can be said to be "present to" or "absent from" something are 5v-re. If the materialists use the criteria of visibility and tangibility to determine their set of entities
(60~bv x&~ &nzdv, 247B3), they will be forced either to say in what sense the moral
qualities can be fitted to these criteria or to give more adequate criteria for something to
be material. As it stands their criteria are too crude, but this is all the argument directed
against them can hope to show. The intention of Plato's argument is to show how the
materialists have tried to limit the class of xh 5vTct. But the argument could succeed only
if it could be shown that at least some of the things which the materialists have called
corporeal (o(b~t~t~ov,247B8) are really incorporeal (~odJrt(~ov, 247D1). They may have
a difficulty but they are not refuted.
There is an interesting historical aspect to this argument. First, there is no hesitation
here to call material things ~wct. They are given what appears to be the same status as
that given to the Forms in the earlier dialogues; the "being-becoming" distinction was coordinate with (roughly, since I do not have the space for a detailed treatment of the earlier
dialogues) a distinction between those entities which "are eternally" and those which were
in continual flux. The former were the only obiects of knowledge, the Forms, and the
latter the objects of some lower form of cognition, e.g., opinion. The former had "true
being" and because the latter were continually changing they could not be said to "be" in
any sense but only to "become." This may not be an entirely fair representation of what
occurs in the earlier dialogues but if any sort of view like this were still held by Plato, this
point in the Sophist would seem a natural place to introduce it. But it is an obvious consequence of the argument against the materialists that sensibles and/or material objects
are ~vm. This is supported by the way the "friends of the Forms" are dealt with in the
following section.
However, out of this argument comes an "account" of being which is relevant to many
of the arguments in the dialogue. Being or x& 5wa are characterized by their "power"
(~a3va~ttg, 247E3-4). But it should be noticed that this is qualified by what immediately
precedes, i.e., xh ~vza are characterized by their power of "acting" and "being acted
upon" (:toter and :t(x0~v, 247E1). This recalls the analysis of sensation in the Theaetetus (156Aft.) where the eye and the object seen are both active and passive with regard
to one another. However, a wider sense of "action" and "passion" seems to be intended
here. It has already been pointed out above that having the ~t~0og of x means "having
the predicate x." In Platonic language this seems to mean that x "affects" or "acts upon"
that which has the :r~0og of x and that what has this property is "acted upon" or "is
passive." Thus if rh ~v~ct are characterized by their capacity to act and be acted upon,
this seems to mean that any subject of discourse, in so far as it is spoken about and various properties are predicated of it, falls in the class of xh 5vza. But it also follows from
this that if x has the property of y, then, by this account, both x and y can be genuine
subjects of discourse, a~ In what sense this is a definition of x& 8 w a will be discussed
shortly.
The above account of -c&~vra has been played down in significance by some commentators n but I shall show that it is consistent with what has already been maintained in the
arguments against the pluralists and monists and that it harmonizes with the account of
lo Cf. Moravcsik, Being and Meaning, p. 27.
al E.g., F. M. Cornford, Plato's Theory of Knowledge (New York: Liberal Arts Press, 1957),
pp. 238-239, and A. E. Taylor, Plato: The Man and His Work ('New York: Meridian Books,
1957), pp. 384-385. Owen, however, takes this to be a definition (ef. my n.2 above).

HISTORY

OF PHILOSOPHY

x& ~v-ca in the arguments against the "friends of the F o r m s " which follow. These philosophers (whether they be Plato's actual opponents, his students, or his earlier self will
not be discussed) want to maintain a rigid distinction between the realms of o6o(,~ and
y~wlOtg, the latter apprehended by perception (a~o0rlotg) and the former apprehended by
reason (~oyto~t6g, 2 4 8 A 7 - 1 3 ) . T h e y allow for some sort of interaction (
2 4 8 A 1 0 ) between each of these realms and the knower or perceiver, but they will not
admit the "action-passion" description for the realm of o~o(a ( 2 4 8 C 6 - 8 ) . T h e y do admit
that knowing is a kind of acting and therefore that what is known is "acted u p o n " or
"affected." F r o m this it is argued (in a highly informal way) that since they admit that
there is such a thing as knowledge they must also admit the existence of soul, life, and
motion. More generally, they must admit that motion is necessary in order for there to be
knowledge (249B5-6). But if there were motion in an unqualified sense, then there
would be no knowledge, since knowledge (or mind, voa)g) also requires that what is
known be, in some sense, at rest. T h e argument ends with the account of being as "whatever is moved and u n m o v e d " (~oa a
215
249D3).
It is difficult to trace the thread of the argument here and to determine how much
weight Plato considered it to have, but it appears to be connected with that account of
being given at the end of the arguments against the materialists. TM If we say that when y
is predicated of x, x is "affected b y " y, it can, in a metaphorical sense at least, be maintained that y "acts u p o n " x. In another sense, if y is predicated of x, then x must be, in
some way, locateable or, in Plato's language, "at rest." T h e "action-passion" terminology
can be linked in this way with the "motion-rest" terminology. Some commentators TM have
read this section as maintaining that what is known, e.g., Forms, are "affected" in so far
as dated temporal propositions are true of them. However, both of these sections seem
to be maintaining the m o r e general view that 5 w a must be at "rest" in order to have
properties predicated of them and that, in so far as properties are predicated of them,
they are "in motion." There is no need to restrict this analysis to temporal propositions.
Needless to say, " m o t i o n " and "rest" are used in a wide sense. But there is a further
point which can at least be suggested, if not fully defended here. I have attempted elsewhere to show that Plato, in the later dialogues and particularly in the Theaetetus, main12 This passage has been subject to a wide variety of interpretations. For a survey of some
recent work el. Harold Cherniss, Plato (1950-1957), Lustrum, 4, 185-188. The main problem is
the interpretation of xb nctwe~.(7~; 5v; it recalls Republic 477A where Plato argues that
xb ntt~xe~.tS~ 5v is nctvze~.tS; V~,o~orbv.But does this passage necessarily have to hark back to the
ontology of the Republic? Such a question is too elaborate for discussion here except for some
tentative comments. Clearly one of the main conclusions to be drawn from the arguments against
the monists, pluralists, materialists, and idealists is that their ontologies have been too restricted.
And, in particular, the idealists have not recognized that certain active or living entities should
be included in any adequate list of what is. Thus the passage could be construed as stating that
life, soul, mind, etc. "really are" in the sense that they have being, are 5~tt and are subjects of
discourse just like any Form or material object. And it is an interpretation of 248E along these
lines which I should prefer; I take this also to be the implication of what appears to be an earlier
but related passage 247A8-9. Such an interpretation has to face the fact that in the lines immediately following the 248E passage, what is ~c~vxe~ ~ is said to "have" mind, life, etc. (of.
249A4, 7, and possibly 9 if the Schleiermacher emendation is accepted). For my interpretation
to be successful, the "have" cannot be taken literally. This is helped a bit by the following lines
where the conclusion is drawn that motion and what is moved should be considered d~g ~vxa
(249B1-2), i.e., as among those things which are or have being.
la Cf. Moravcsik, Being and Meaning, p. 40 and W. G. Runciman, Plato's Later Epistemology
(Cambridge, 1962), pp. 79ff.

BEING AND PREDICATION

tains that knowledge and belief are propositional in character24 If m y view is correct,
then knowing x consists in knowing that x has a certain characteristic, that it has the
~d0og of y. Hence, in so far as something is known, it is known as having a certain
characteristic and in this sense is "affected." It would also have to be, in some sense,
identifiable or "at rest." We could then distinguish two senses of "motion" (and, correspondingly, of "rest"). First, in so far as dated propositions are asserted about x (that it is
y at t~, that it is z at t 2, etc.), x is "in motion." Second, in so far as something is known
via a proposition or series of propositions, it is known qua being "in motion" or "affected." These two senses are not contradictory but I believe that m y sense of "motion"
and "rest" is the more fundamental one and serves to bring the two accounts or definitions of z& ~iv~ct closer together. This extended sense of motion and rest helps to explain
why they are included among the "greatest kinds" later in the Sophist.
I have maintained above that the purpose of the section of the Sophist under discussion here (241C--249E) is to examine the doctrines of various philosophers about the
" n u m b e r " and "nature" of Th ~;vrct (242C). Plato's predecessors have tried to limit
r& 5vzct to certain numbers or kinds of things. Plato, on the other hand, seems to be maintaining that being cannot be so limited. Rather, it appears to be universally applicable or
topic-neutral because he directs his arguments against various attempts to limit it. I n an
earlier passage in the Sophist Plato argues from the fact that something does not have
oSo~a or is not an ~v to the conclusion that it cannot be a subject of discourse (cf. 2 3 7 A 239B). This is good evidence that he considers T& ~vra to cover the entire range possible subjects of discourse. 15 But there is also evidence that this is not the only characteristic which subjects of discourse must have. It seems possible to construe the final
hypothesis of the Parmenides as deriving disastrous conclusions from the hypothesis that
there is no such thing as unity (cf. Parm. 165E2ff.) and to infer f r o m parts of the second
hypothesis that being and unity cover the same range of entities (cf. Parm. 144Blff.).
This is supported by the sections of the Sophist discussed above. Thus every subject of
discourse is, in some sense, one and has being. Plato also seems to be maintaining that
14 Cf. my "Perception, True Belief and Knowledge in Plato's Theaetetus," Phronesis, XIV, 2
(1969), 111-122; "The 'Dream' of Socrates and the Conclusion of the Theaetetus," Apeiron, HI,
2, 1-13; and "Plato's Sophist; Falsehoods and Images," Apeiron, VI, 2, 1-6.
25 Cf. Moravcsik: "The outcome of Plato's criticisms [i.e., of the monists, pluralists, materialists, and "idealists"] is that one cannot withhold Existence from any class of entities. In showing
that Existence is necessarily all-inclusive Plato leads us to see that Existence is not an attribute.
It does not separate a certain class of entities from all else; for there is nothing to separate existents from. From Plato's arguments we can also see that Existence is topic-neutral. Its ascription
to an entity does not tell us what characteristics that entity has. Thus Existence cannot function as a sort- or value-concept. This leads to the important metaphysical conclusion that
the evaluation of types of entities and the determination of their ontological priority cannot be
done by assigning Existence partially to some entities while withholding it from others" (Being
and Meaning, p. 28). Moravcsik's statements seem to me to be correct with three exceptions.
First x6 6v, throughout the Sophist, should be translated as "being" rather than as "existence."
Cf. J. Malcolm, "Plato's Analysis of xb 6v and "~b Ix~l6v in the Sophist," Phronesis, XII, 2 (1967),
130-146; M. Frede, Priidikation und Existenzaussage, Hypomnemata, 18; and the articles by
Owen cited in notes 1 and 6 above for strong arguments against an existential sense of e[va~ in
the Sophist. All of these authors agree on the topic- neutrality or, as I should prefer to put it, the
universal applicability of being. Second, ascribing being to something does tell us some of the
other characteristics which it has; i.e., it must have all those characteristics which are universally
applicable characteristics. Third, as I shall point out shortly, there is a sense in which being is not
an attribute or characteristic but there is also another sense in which it is.

HISTORY OF PHILOSOPHY
every subject of discourse in some sense is a whole and has parts or aspects; the remaining
sections of the Sophist add sameness and otherness to this list and various sections of the
Parmenides suggest that similar-dissimilar be added as well. Thus all these are terms
which seem to indicate characteristics which any subject of discourse must have and
thus can be called universally applicable.
But, there are serious philosophical problems here. I have already pointed out that
there is no distinction made between the terms Tb 5v and T& ~v~a and further that
Tb 5v is itself ambiguous. It might stand for either what has being or being itself. Problems
arise in two ways. First, how is being like or unlike other characteristics? Second, is being, when talked about as a Form or characteristic, self-predicable and, if so, are there
any difficulties involved?
With respect to the first question, Plato seems to be arguing that every subject of discourse has being. Some commentators take this universal application of being to mean
that being is topic-neutral and therefore they ascribe to Plato the doctrine that being is
not a predicate (e.g., Moravcsik, el. n. 15). But this involves the view that if some characteristic is universally applicable, then it cannot be a predicate or attribute, since to be
a predicate or attribute is to be a characteristic which can be opposed to other characteristics which fall within a certain range of application. For example, having a shape may
be considered a predicate because it is distinguishable from other characteristics and
further because not having a shape can be applied as well. Being is not a predicate in this
sense for Plato because its opposite, ~b ~ t ~ a ~ t ~ ~v, is rejected as self-contradictory and
because being is universally applicable.
Even though being is not a predicate in this sense, it does appear to be distinguishable
from the other universally applicable characteristics. The argument which serves to distinguish between being, wholeness and unity as well as the elaborate arguments distinguishing the "greatest kinds" (254-259) would seem to indicate this. We can ask whether
Plato would distinguish between having the :t~0o~ of being and having, say, the ~t0o~ of
redness. By the rule which I have discussed above, if x has the ~t~t0o; of being and also
has the :t~t0o; of redness, it is not identical with either of these characteristics. But would
Plato maintain that these are characteristics in the same sense? The fact that one is universally applicable and the other is not would seem to call for a negative answer. Thus
he would argue that if x has the :t~0og of being, this does not entail that x has the ~r~0o;
of redness but having the n~t0og of redness does entail that x has the ndt0o~ of being. In
the language of the section on -cb ~t~l~a~t~~ ~Jv,redness is an 5v, a member of the class -c&
~vxa. Tb Ix~l~a~t~g8v can be translated as "that which has no characteristics at all" on
the grounds that this logical fiction is "removed from x& 8v~t" (of. 237D). Redness, like
any other characteristic, is a member of this class. But we have to ask whether being is
itself a member of this class, or in the terminology of the Greek, is xb 5v or o ~ [ a a member of the class of x& ~wa? Or, in another way, is being itself an entity?
The answer to this question is complex. Let me reiterate that being is a universally applicable Form or characteristic. Every subject of discourse has it. This must be coupled
with the following observations. Plato alternates between the material and formal modes
of speech. He can talk about being in terms of ordinary subject-predicate statements (as
in his discussion of the "greatest kinds") as though it were no different either from other
characteristics or from more ordinary objects. But he earl also use the term "being" to
stand for the class of subjects of discourse as well as for the characteristics which subjects
of discourse have. He does so in 237A-239D. Further, his arguments in many places are
also designed to discriminate the meanings of terms. Thus a distinction between, say,

BEING AND PREDICATION

being and unity is a distinction between two Forms or characteristics while at the same
time it is a distinction between the meanings of "being" and "unity." My question can
now be answered on the basis of the above considerations. Although it has a unique
ontological role in Plato's scheme, being is a member of the class of x& ~vzct. It is a Form
or characteristic which can be a subject of discourse like any other subject of discourse.
In both the Sophist and the Parmenides, Plato takes ~ e a t pains to distinguish being
as a subject of discourse from the other universally applicable characteristics and arguments have already been polnted out which run: being has the :t~t0o; of x, but is not
identical with x. In order to argue in this fashion, Plato has to hold that being is distinguishable, if not from an intelligible opposite, at least from certain other terms which
appear to be of universal application as well. This leads to a comparison of the way in
which "being" operates with the operation of such pairs of terms as "part-whole," "onemany," and "same-other." Plato seems to be arguing, in the Parmenides about the second pair, and in the Sophist about the first and third pairs, that both members are universally applicable but cannot be applied to an entity in the same sense at the same time;
e.g., "x is the s a m e . . . " and "x is other t h a n . . . " cannot have the same completion. But
every subject of discourse, since these terms have universal application, is in some sense
one, many, whole, part, etc. However, the implication-relations between these terms are
different from those between lower-level terms. Plato argues that "x is a whole" (any
other universally applicable characteristic can be substituted for "whole") entails and is
entailed by "x is y" where y is any other characteristic of this same level. In doing this,
Plato is well on the way to a doctrine of formal concepts and to the realization that these
concepts are incomplete. And in this sense, Plato realizes that the universally applicable
characteristics are quite different from ordinary characteristics.
But the fact that he persists in talking about these universally applicable characteristics
in the same way as he talks about other characteristics causes him to have virtually insurmountable problems which center around the notion of self-predication. If unity and
sameness, say, are universally applicable, in Plato's technical language we should be able
to say that unity has the ~t~t0o~of sameness and also that sameness has the ~xgt0o~of unity.
From these follow two statements of non-identity distinguishing sameness and unity. But,
i~ in the paradigm ca.se of a statement, x is y, it is possible to apply all the universally applicable characteristics to both of the variables, then we should be able to maintain that
sameness is the same (as itself) and that unity is one.
The issue of self-predication has been shown, both by Plato and his commentators, to
have severe difficulties. To say that triangularity is triangular implies that there is a threesided figure in some region of space of which every ordinary triangle is an instance; this
leads to the question whether triangularity and an ordinary triangle are "triangular" in
the same sense. If they are triangular in the same sense, it is a short step to an infinite
regress. I shall not deal with self-predication as applied to lower-level Forms because
there are sufficient difficulties when these questions are raised with regard to those Forms
discussed in the Sophist. One way out of the regress difficulties is to distinguish between
what is identical with x and what is predicatively x. Forms are defined as "what is identical with x." On this account, ordinary triangles are triangular but the Form triangularity
is identical with triangularity and is not predicatively triangular. This may or may not be
a way out of the regress difficulties for lower-level Forms but it clearly leads to difficulties for those Forms which are universally applicable. 16 Going back to unity and same16 CL Harold Cherniss' discussion of various analyses of the Third Man Argument in "The

10

HISTORY OF PHILOSOPHY

hess, their non-identity depends on the rule I have pointed out above and each has the
~ t 0 o g of the other. But if, in x is y, both x and y can be subjects of discourse and the
universally applicable characteristics apply to every subject of discourse, then both x and
y can be said to have (among other characteristics) the ~t~t0og of unity and sameness. The
result of this is that unity, say, is both identical with unity and also has the ~ f 0 o g of
unity. But by the rule, if unity has the ~t~t0og of unity, it cannot be identical with unity.
Thus, this rule, which was used to discriminate the universally applicable characteristics
from one another, must fail in the attempt in spite of the fact that it does distinguish
identity and predication and m a y have some value when applied to more ordinary cases.
Plato, then, is forced to accept self-predication at least for this special set of Forms.
And, if he admits this once again he must also admit that being is itself a member of the
class of "c& ~vra. If "being" were only a shorthand way of talking about the class of subjects of discourse, some way out of his dilemma could be found. But as long as he attempts to distinguish being from the other characteristics of this same level by the distinction between identity and predication, difficulties are bound to result. And, of course,
these same difficulties arise for all the other characteristics which operate in this same
way.
Has Plato, then, defined being? If a genus-species definition is required, clearly he has
not and cannot define being because of its universal application as shown by the universally applicable descriptions given of it. Plato probably realized that a definition in
this sense was out of the question. But he clearly wanted at least to distinguish it f r o m
the other characteristics. However, such a distinction is not possible by means of the
argument he has advanced. Perhaps he would fall back on some sort of common-sense
distinction between "to be," "to be one," "to be the same," etc. but this is not elaborated.

University of Missouri, Columbia

Relation of the Timeus to Plato's Later Dialogues," American Journal of Philology, 78, 225-266:
"It can be shown that Plato was well aware of the difference between such an assertion of identity
and an attribution and in this awareness consciously denied what Vlastos [cf. G. Vlastos, "The
Third Man Argument in the Parmenides," Philosophical Review, vol. 63, 1954, pp. 319-349] believes he unconsciously assumed without understanding its implications. In the second part of the
Parmenides, for example [Cherniss cites Parm. 158A], he uses the following argument: If there
are parts, since 'each' signifies 'one', each part must participate in Unity (xb ~v); but its participation in Unity implies that it is other than one, for otherwise it would not participate but would
itself be one, whereas only Unity itself (~t~Tb ~b ~v) can be one. So each part, like the whole of
which it is a part, 'is one' only by participating in Unity. Here Plato clearly distinguishes two
meanings of 'is x', namely (1) 'has the character x' and (2) 'is identical with x'; assumes that whatever 'is x' in one sense is not x in the other; and states that a6~b xb and only ct~zb xb x 'is
x' in the second sense. As applied to the statements quoted by Vlastos and to the doctrine of ideas
generally, this is to say: 'the idea of x is x' means 'the idea of x and x are identical and therefore
the idea of x does not "have the character x ..... (pp. 258-259). Both the Cherniss article and the
Vlastos article have been reprinted in the anthology, Studies in Plato's Metaphysics, ed. R. E.
Allen (London: Routledge and Kegan Paul, 1965). In "Self-Predication in Plato's Later Period,"
Philosophical Review, LXXVIII (1969), 74-78 (reprinted in his Platonic Studies, Princeton: 1973)
Vlastos replies to Cherniss, agreeing with J. Moravcsik ("The 'Third Man' Argument and Plato's
Theory of Forms," Phronesis, VIII, [1963], 50-62), that Plato is committed to self-predication for
those Forms which are universally applicable. As I read the rule in Sophist 245B" if x has the
n~t0og of y, then x and y are non-identical, there are restrictions or special conditions laid upon
it and thus it is equivalent to what Vlastos, in the famous 1954 article cited above in this note,
called the full strength non-identity assumption. Cf. Frede, Pri~dikation, p. 72.