1 In this paper I use self-predication in the conventional sense to refer to the proposition that a Form has the property that it is. According to this usage not just any sentence of the form The F itself is F is a self-predication.

2 It is clear that (S2) does not contradict (S1) without some additional assumptions, which can be captured by emending motion and rest to what is in motion and at rest. The significance of this will become apparent in III, below. For now, it is enough to note the ES's indifference to the distinction between nominative and predicative forms.

3 (P1) evidently does not follow from what precedes without erasing the distinction (as in n.2 above) between nominative and predicative forms of Motion and Rest. See III, below.

4 I take the solution of this problem to be as central to Plato's objectives in the Sophist as the analysis of false statements, which has enjoyed greater attention in the scholarly literature.

5 Even though he intends the question at hand to be generalized, the ES appears to be discounting the question of naming as it applies to sensory particulars. For of course the question of how a particular can have more than one name had already been answered by the early theory of Forms.

6 The arguments at 243d8-244a2 and 244b6-d13, for example, are governed by the model that follows.

7 Conforming to Plato's usage, I treat the following propositions as equivalent: (a) F and G are one (244a 1-2), (b) F is the same as G (245b8), (c) F is (not other than) G (254e3, 255a4-5). All three are equivalent to (d) F and G have the same meaning (255b8-c2).

8 J.L. Ackrill, Plato and the Copula: Sophist 251-259, Journal of Hellenic Studies 77 (1957) 31-55, has shown that Plato's usage in the Sophist distinguishes between symmetrical and non-symmetrical relations between Forms. Participation is a nonsymmetrical relation such that if F participates in G, it does not follow that G participates in F. Blend and combine capture the more generic concept of being connected, which itself is symmetrical but includes both symmetrical and nonsymmetrical relations among Forms.

9 True as far as it goes, but not the whole story. The self-identity of a Form could be threatened by koinnia, so the ES will later invoke the vowel Forms, Being, Sameness, and Difference to secure the natures of Forms that blend among themselves. See IV, below.

10 This is not the notion of natural names in the Cratylus. To say that a Form has a name in virtue of its nature is not to say that there is something about the name, e.g. etymology, that makes it appropriate. The Form's nature determines what its name is, but only on the assumption that there is already an established practice governing use of the name. Given such a practice, the nature of Motion, what Motion is, makes Motion the name₁ of the Form.

11 For the textual evidence that general terms are proper names of Forms see T.W. Bestor, Plato's Semantics and Plato's Parmenides, Phronesis 35 (1980) 43-5 and n.5, 73.

12 The relevant phrase is kata tn hautou physin (250c6). It is clear that kata has explanatory force in this passage from the similar lines at 255e4-6, where kata is replaced by dia with the accusative: for each one is different from the others, not in virtue of its own nature but in virtue of partaking of the form of the Different.'

13 This is not to say that the Sophist does not contain the conceptual apparatus for disambiguating esti and for clarifying the sentence form The F itself is F. The work of Ackrill, Plato and the Copula: Sophist 251-259, M. Frede, Prdikation und Existenzaussage (Gottingen: Hypomnemata Heft 181967), J. Malcolm, Plato's Analysis of to on and to me on in the Sophist, Phronesis 12 (1967) 130-46, and G.E.L. Owen, Plato on Not being in G. Vlastos, ed., Plato I (Garden City, NY: Anchor 1971), 223-67, has shown that it does, indeed that it provides more than one avenue to accomplish it. The point is rather that the text of the Sophist makes it doubtful whether Plato himself thought that The F itself is F stood in need of clarification.

14 Termed the Non Identity Assumption by Gregory Vlastos in his influential paper, The Third Man Argument in the Parmenides, Philosophical Review 63 (1954), 319-49. Benson Mates, Identity and Predication in Plato, Phronesis 24 (1979) 225, writes that allowing the large to be large by virtue of itself, would seem to be [Plato's) only real alternative in combatting the Third Man Argument. I concur in that judgment, but without specifically endorsing Mates schema for interpreting the Platonic is.'

15 Bestor, Plato's Semantics, 57-8, and n.12, 73-4, plausibly construes the One Over Many Principle as inclusive of non-identity, but he goes on to claim that the former cannot apply to Forms because it is a deliberately restricted principle. I find the latter claim unconvincing. Bestor argues that to say that the Form F could be a participant in some further Form F is really to say that the Form F could be a thing whose proper name is F and at the same time take that name after a named thing, and that, he thinks, is absurd. But Bestor's argument merely shows that the One Over Many principle is incompatible with the claim that general terms are logically proper names of Forms. It does not show that Plato initially put forth a restricted version of the One Over Many principle. The Third Man Argument derives its force precisely from the fact that Plato had previously never had to take seriously the question in virtue of what the F itself is F. So another Form F in virtue of which the F itself and the other F-things are F (Parm. 132a6-8) was not yet explicitly ruled out. There seems little doubt that Plato always believed that general terms name Forms, even name them naturally (cf. nn.10, 30), but the implications of natural naming were not clearly articulated at the outset.

16 The sense of is at 255a5 is given at 254e2-3.

17 255a11-b1 makes participation in its enantion the reason why Rest changes to the enantion of its nature. But the point cannot be generalized. Sameness and Difference, which participate in each other, do not thereby change into the enantia of their natures. Or are Sameness and Difference not really enantia? The latter is not plausible. At 259d2-5 Plato treats them on a par with the Large and the Small (called enantia at Phaedo 102d-103c) and the Like and the Unlike, and he ridicules those who enjoy producing such enantia in arguments. Parmenides 159e-160a calls Likeness and Unlikeness enantia, and the argument is extended to include Sameness and Difference. It is preferable to suppose that some enantia do blend without changing to something enantia to their natures. That seems to be the point of the entire passage at 259b8-d7. I flag the distinction by translating opposite or incompatible as the context requires.

18 The occurrence of gar at 255a11 shows that (v) is a reason for (vi), hence a premiss on which the conclusion is based.

19 Note the frequent occurrence of the explanatory dia with the accusative and poiein in the passage articulating the relationships between the most important Forms at 255e-256e, 259a.

20 The p in the protasis of the conditional cannot justify construing it as categorical, thereby allowing that Motion somehow participates in Rest. The conditional is counterfactual. G. Vlastos, An Ambiguity in the Sophist, reprinted in Platonic Studies (Princeton, NJ: Princeton University Press 1973) 293-4 is surely right that we are not being told that there is a sense in which Motion partakes of Rest, but that if there were, then Motion is resting would not be absurd.'

21 It is hardly an argument. The principle begs the question against the late learners, as Ackrill, Plato and the Copula, and others have noted. But the ES's strategy is to draw a distinction, which justly undermines an assumption about naming embedded in the late learners position.

22 It is sometimes maintained that Plato thought the sentence absurd because it ascribes the property of being at rest to Motion, but that would be no more absurd than asserting that Motion has the property of being in motion. Since Motion itself is a property, it makes no sense to say that Motion has the property of being at rest or in motion. On the other hand, if Plato did not know that Motion is a property, he might just have thought of the Form as an entity to which the property of being in motion could be ascribed. But if this implies that he thought Motion had the property of being in motion, it would undercut his motivation for positing the Form of Motion in the first place. If he had fully understood what it is to ascribe a property to something, he would have had no reason to adduce the Form to explain how it is that some individual thing can be in motion. Could not Plato then have thought that Motion is at rest was absurd because it does in fact ascribe a property to Motion, even though he did not actually think of it in that way? It is not altogether clear what this would mean other than: (1) Plato might have thought of Motion as an entity at rest without thinking of its immobility as a property of it. This is surely possible, but there is nothing absurd about it, since being at rest is one of the definitive features of Forms; (2) Plato might have thought of Rest in Motion is at rest as analogous to fast, slow, and circular, which are properties of Motion even though, strictly speaking, he did not yet think of them in this way. If so, the sentence would surely be thought to be absurd, because being at rest (for Motion qua Motion) is not analogous to being fast, slow, or circular. However, if he had been thinking along these lines, he would also have thought of Motion is in motion as absurd, since being in motion (for Motion qua Motion) is no more analogous to being fast, slow, or circular than being at rest. Instead, Plato appears to have thought that Motion is (in) motion was obviously true.

23 Once it has been determined that Forms can blend, whether or not those which can actually do combine can be established from the truth or falsity of the resulting statement, as in Man understands, to be contrasted with, e.g., Man is winged, Man flies.'

24 Plato does not know the Aristotelian notion of paronymy in the Sophist. Contra R.E. Allen, Plato's Euthyphro and the Earlier Theory of Forms (London: Routledge 1970) 126, and J. Malcolm, Semantics and Self-Predication in Plato, Phronesis 26 (1981) 287, Plato's use of eponymy is not equivalent to paronymy. Aristotle draws a distinction between grammatically different forms of terms distinguished by a difference of ending, one of which is derivative from the other (Categories 1a12). Paronymous is a term he applies to things whose names are thus derivative. For Aristotle the grammatical distinction is reflected semantically in different functions of predicates that are in and said of a subject. The importance of Aristotle's contribution is to locate the different ways in which something can be named in the different functions of the predicate. Neither the Phaedo doctrine that particulars are named after Forms, nor the Sophist's two ways of naming Forms, takes that important step. Aristotle's discovery emerges naturally as a result of Plato's protracted considerations of how something is named, and it rests finally in the distinction between predicative and nominative functions of the predicate.

25 At Theaetetus 182a3-b8 Plato makes what looks like the distinction we are looking for in the course of expounding a secret doctrine attributed to Protagoras: The agent comes to be qualified in a certain way (poion ti) not a quality (poiots) not hotness or whiteness but hot or white and similarly with all the rest'; cf. 156e2-8. But there is no indication in the text that this important distinction is yet associated with any semantic content. It is introduced in a metaphysical context to explain the subtleties of a theory of perception held by certain more refined advocates of the flux theory. In Plato the distinction between being a quality and having a quality, corresponding to the original distinction between Forms and their participants, just adds one more crucial piece to a puzzle about meaning and predication that was not to be completed and precisely articulated until Aristotle.

26 This interpretation is at variance with the standard view summed up by Owen, Plato on Not being (261), as follows: It is commonly agreed that (B) [my (Pl)] is proved by an illicit move from X is not identical with either Y or Z to X is not characterized by either Y or Z; and that subsequently the ES blocks this move by distinguishing identity-statements from predications . I am contending that the disambiguation of esti was not a concern of Plato in the dialogue, and that it is not necessary in order to resolve the paradoxes of Being. This is not to say, of course, nor does it follow, that Plato confused attribution and identity.

27 In the Phaedo it would be more accurate to say that if a Form has two names by nature, one will entail the other but they will not necessarily be synonymous. At Phaedo 104a the number three is said to have the name odd in addition to its own name by nature. The Sophist makes no explicit mention of how Forms get the names of necessary attributes. The kind of case in which subject and predicate terms apparently name different things, yet predicate is implied by subject, could be problematic for the Sophist's model of naming. Also in question would be the status of things designated by terms that define a form. How are they related to each other and to the form defined? Is the definition a unity'? Answers to all these questions would doubtless require some modification of the simple model of naming adumbrated in the Sophist. It must be remembered, however, that the enterprise of the Sophist is very different from that of Phaedo, which was not inquiring into the matter of how Forms can be called by more than one name. In the Phaedo relationships among Forms were merely assumed as a premiss of the immortality argument. The Sophist very deliberately calls into question that assumption in order to examine its consequences. We should therefore not expect the Phaedo's use of by nature necessarily to coincide with language in the more comprehensive and rigorous examination of naming found in the Sophist.

28 Noted by G. Vlastos, The Unity of the Virtues, reprinted in Platonic Studies, 238. Vlastos cites Phaedo 103e2-5 to illustrate the second use of onoma in Plato. Examples in the Sophist can be found at 251a5, 251b3, 257c. Onomata at 26ld2, 4 includes both kinds of names; but epnymia occurs at 257d9, and rhmata at 257b7, instead of onoma.

29 Thus Vlastos writes: To use a rule of language without the least awareness of the rule is an all too common phenomenon of linguistic behavior (ibid., 240-1). But Vlastos (239, n.49) nevertheless believes that Plato implies the distinction between predicative and nominative functions of onomata in the Sophist.

30 Plato does assign a descriptive property to terms in the discussion of natural names in the Cratylus, where the etymology of the name reveals (dloz) its proper nominatum. But the natural names of the Cratylus include both proper names and common names or qualifying predicates and thus bear no relation to the distinction between reference and description or to the two kinds of naming in the Sophist.

31 Contra Vlastos, The Unity of the Virtues': To take note of the distinction between referring and predicative expressions he resorts to two different words, onoma, rhma, instead of saying that the same word, onoma, can be used in each of these distinct ways (239).

32 A. Nehamas, Participation and Predication in Plato's Later Thought, Review of Metaphysics 36 (1982) 343-74

33 He has already demonstrated indirectly at 255a-b that the Different cannot be identical with Motion or Rest. What has not been shown (but is implicit in the same argument) is its difference from Sameness. 255e3-6 contrasts the Different with the other four Forms, thus affirming its distinctness from them, but it is not an argument that establishes its uniqueness.

34 Vlastos, Self-Predication in Plato's Later Dialogues, reprinted in Platonic Studies (340, n.13), has pointed out the strong adversative force of alia. He takes autn to exclude the Different and refrains from endorsing the paradox of selfparticipation, even though he allows that the Different might be different in virtue of its own nature and by participating in itself. Vlastos rightly concludes, however, that if Plato had intended that meaning, he would surely have expressed the outcome in some clear and unambiguous way. Nehamas, Participation and Predication (352ff.), on the other hand, boldly endorses self-participation. Taking autn to be inclusive of the Different, he claims that the Different is (1) what it is to be different in virtue of its own nature and (2) different from other things by participating in itself. But Nehamas does not explain what it would mean for a Form to participate in itself. To support his interpretation he merely calls attention to 256a7-8 where it is said that Motion is the same because everything participates in the Same. Now it is true that participation (metechein) is at least semi-technical in Plato's metaphysical vocabulary, so he would have been free to alter its meaning as he wished. But without independent evidence that he revised the meaning of participation in the Sophist so as to make it reflexive, the latter passage is no more convincing than 255d-e as evidence for self-participation. In the absence of some reason to construe 256a7-8 otherwise, the quantifier will automatically range over everything except the Same. (Similarly, unless one explicitly redefines the relation taller than, one will be correctly understood to exclude oneself from the scope of everyone in saying, I am taller than everyone in the room.) As additional evidence for self-participation Nehamas cites Parmenides 162a6-bl, and he offers a reinterpretation of the conflicting passage at 158a3-6. But the second part of the Parmenides is a risky place to look for positive doctrine.

35 Robert Renehan has remarked that use of the feminine gender (autn) to refer to to thateron at e3 (introduced by the periphrasis tn thaterou physin at d9) highlights the fifth Form, setting it apart from the others and foreshadowing the ou alla at e4-5, where the same distinction in gender is maintained between hen hekaston (e4) and ts ideas ts thaterou (e5-6).

36 Consider: Jones is fifth among the important novelists we are picking out and he respects them all. The group of novelists now includes Jones, but we do not understand the sentence to mean that Jones respects himself along with the rest of the group. Instead we take them to refer only to the other four writers.

37 This distinguishes my view from any interpretation which claims that Plato, either intentionally or in effect, marked out a distinction between predication and identity statements in the Sophist.

38 This is the issue raised at Parmenides 129e over the prospect of someone showing that the separately existing Forms might be combined and separated among themselves. What happens to the identity of the subject when the F itself is both F and not F, or F and also G?

39 To that end, all that is required is a distinction between affirmative predication and negative identity, which effectively neutralizes the contradictions that otherwise would result when a Form is said to be F and not F. For example, Motion is the same and not the same (256a-b), different and not different (256c), being and not-being (256d).

40 Cf. R. Heinaman, Self-Predication in the Sophist, Phronesis 26 (1981) 55-66; G. Vlastos, An Ambiguity in the Sophist 270-317.

41 Heinaman, ibid., 55

42 It is important to distinguish the analysis of a statement from what follows if the statement is true. Thus (a) Motion is the same as itself is paraphrased as (b) Motion participates in Sameness with respect to itself (256b). If true, (a) would presumably say (263b) things that are that they are about Motion. But it would be a mistake to infer from the discussion of true and false statements that some such statement as Sameness is with respect to Motion is therefore the correct analysis of (a).

43 The substance of this criticism is made by G. Vlastos, On a Proposed Redefinition of Self-Predication in Plato, Phronesis 26 (1981) 76-9. Vlastos is responding specifically to A. Nehamas, Self-Predication and Plato's Theory of Forms, American Philosophical Quarterly 16 (1979) 93-103.

44 Cf. the similar distinction at Euthyphro 10e9-llb1 between statements that give the ousia of a thing and those that give a pathos, depending on whether the things designated by subject and predicate expressions are the same (tauton) or altogether different from each other (pantapasin heter onte allln). In that context it is clear that Socrates considers the first sort of statement to be the more important of the two.

45 Using the criterion of what can and cannot be said to determine possible combinations among Forms, one must conclude that Motion does not participate in Rest, because (1) Motion is (at) rest is absurd. But if (2) Every Form is (at) rest is true by Plato's definition of Form derived from epistemological considerations, it follows that (1) is true. Employing different criteria to deny (1) and affirm (2) no doubt obscured this puzzle for Plato, but it need not cause a problem for him. The two criteria do not actually conflict. A Form's nature determines its capacities for participation, but being a form as such is not part of any Form's nature. Its nature distinguishes it from other Forms not from particulars. So, on those grounds, (2) would not imply that Motion participates in Rest.

46 For example, it would not be necessary for every sentence of the form The F itself is G to imply participation between Forms. This might be so if not all predicate terms denoted Forms or if sentences assigning features to Forms as a category of being did not import participation (cf. Aristotle, Met. 990b27-991a1). If the latter, such a sentence might imply no relation at all between nominata of subject and predicate terms or a relation other than a semantic one. In any event, the question would justly arise as to how such a sentence should be construed, if not as affirming a participation relation between two Forms.

47 What is called for is a distinction between different levels of discourse. G.E.L. Owen, Dialectic and Eristic in the Treatment of the Forms in G.E.L. Owen, ed., Aristotle on Dialectic (Oxford: Oxford University Press 1968) 103-25, pointed to the distinction between different types of predicate drawn by Aristotle in the Topics (cf. 137b3-13, 148a14-22) and argued that in some contexts (Met. 990b27-34, Topics 137b3-13) Aristotle concedes recognition of the distinction to the Platonists. I am less confident of the attribution than Owen was, but nevertheless believe that Plato himself could have drawn the distinction, that nothing in the Sophist rules it out, and that his position requires it. Vlastos, The Two-Level Paradoxes in Aristotle, reprinted in Platonic Studies 331, claims that the distinction is ruled out for Plato by his doctrine of the absoluteness of his Idea, but that appears to overlook the Sophist's solution to the problem of Not-Being, which allows that a Form can be both F and not-F without compromising its privileged status.

48 I owe thanks Richard Bosley for discussing portions of this paper with me, to the Editors of this journal for critical comments, and to Julia Annas for criticism of an earlier draft.