1 This may reflect Plato's view that the first principles of his philosophy are incapable of being put down in a treatise (VIIth Letter, 341b-e); at any rate, it is a feature of his writing that makes the extrapolation from it of his metaphysics so controversial and difficult.

2 In part, Plato's rehabilitation of the phenomenal world is the result of a change in attitude, rather than doctrine. His description of the cosmos as the best of the things that have come to be (29a); as complete, comprehensive, unified, and free from illness and age (30c-33a); as a living being endowed with soul and reason (30b-c); as everlasting in its existence (38b-c) - all these are indications of the higher value Plato places on the phenomenal world in the Timaeus than in earlier works. Also indicative of this change is the high praise he accords to vision (47a-b) and the repeated use he makes of mathematics, which for him is a paradigm case of rational knowledge, in explaining the cosmos (cf. esp. 31b-32c, 35b-36d, and 53c-57d). To what extent this shift in attitude is the result of the introduction of a divine creator of the cosmos I do not know; but certainly Plato's earlier contempt for the phenomenal world would be difficult to reconcile with his expressed view in the Timaeus that the cosmos is the product of an intelligent and benevolent craftsman-deity.

3 Philebus 15b-c likewise suggests that the nature of participation remains unresolved in the later dialogues.

4 The relevance of the discussion of the ‘elements’ to the explication of the nature of Necessity is not immediately obvious. Necessity is the force in nature that opposes or restricts the plans of the Demiurge. Plato conceives this role as analogous to that of a somewhat refractory material on which an ordinary craftsman works. If the four ‘elements’ were in Plato's view the ultimate material with which the Demiurge works, they would be suited to play this role; their unsuitability in his eyes occasions the resulting search for more elementary components of the cosmos, in the course of which the Receptacle is introduced.

5 As Lee, E.N. notes in ‘On Plato's Timaeus, 49D4-E7M,’ American Journal of Philology, 88 (1967) 14Google Scholar, the translation proposed by H.F. Chemiss, ‘at different times in different places,’ in ‘A Much Misread Passage of the Timaeus (Timaeus 49C7-50B5),’ American Journal of Philology, 75 (1954) 114, is unwarranted. This passage is difficult, and translations proposed for it have varied widely. I have generally opted for the traditional reading (given by Comford and, with variations, by others) to the radically different interpretation proposed by Cherniss and followed, with reservations, by Lee. Both readings have their awkward moments, but neither seems to be impossible Greek; therefore, my objections to the Cherniss reading are primarily philosophical and not philological. It is well known that the Cherniss reading commits Plato to a four-fold division of entities (Forms, Receptacle, immanent character, and instance of immanent character), wheras Plato insists in the passage on a three-fold classification (51e-52b; cf. K.W. Mills, ‘Some aspects of Plato's Theory of Forms: Timaeus 49c ff.,’ Phronesis, 13 [1968)153-54, 170). Moreover, Cherniss explicitly claims it as a consequence of his interpretation that ‘Phenomena cannot be distinctively denominated, because no part of the phenomenal flux is distinguishable from any other’ (128). According to this view, the immanent characters can be named, but not their instances. Against this view it should be pointed out that Plato does not make anything of the distinction between immanent character and instance (cf. p. 130, below); that he shows in Tim. 51b that he has no difficulty naming instances of immanent characters; and that such a view would make the wisdom of the philosopherking of little value in the cave (cf. p. 124-5, above) and the distinction between right and wrong opinion at least problematic, at any rate as regards particulars.

6 ‘hekastote’ (49d5-6) parallels ‘aei’ 49d4, 7) and contrasts with ‘mēde … pote' (49d7). Cherniss (115) finds this redundant, but we must remember that Plato delights in pleonasm and parallel construction.

7 As Alex Mourelatos has pointed out to me, ‘to toiouton’ has a demonstrative as well as a relative use, and it is the former that is found here. There is no corresponding demonstrative use of ‘such’ in English; hence, the translation is somewhat awkward.

8 Cherniss (117) thinks the subject of ‘pheugei’ is ‘touto. ‘Lee thinks the subject is 'simply one of those individual things which we so often point at and talk about’ (6). I do not wish to limit the subject to individuals; otherwise, I concur with Lee. There is no grammatical difficulty in taking ‘hosa’ (49d7) as the antecedent of ‘pheugei,’ since neuter plurals regularly take singular verbs.

9 Many commentators have noted the difficulty of giving a sense to ‘kai tēn tōide' (49e3), and some have simply not translated it. I take it that Plato's point is that inflections of ‘tode’ and not Just the nominative singular imply stability; thus he includes a dative singular as an example. The sense of the phrase is of secondary importance.

10 This is the reading of Cornford, F.M., Plato's Cosmology (London: Routledge & Kegan Paul 1937; reprint ed., New York: Liberal Arts Press 1957), 179Google Scholar. Taylor, A. E., in A Commentary on Plato's Timaeus (Oxford: Clarendon Press 1928), 318,Google Scholar notes a parallel between this passage and 50a4, which supports this reading.

11 Cherniss (120-3) places heavy emphasis on the meaning of ‘houtōi’ and chides translators who omit it. I have rendered it ‘thus,’ and have altered the word order so that it refers to a phrase that follows it, as ‘thus’ ordinarily does in English, rather than a phrase which precedes it, as ‘houtōi’ does in Greek. I take it that the adverb modifies ‘kalein’ and refers to the phrase, ‘to de toiouton … homoion’ (49e5).

12 With Gulley, Norman, in ‘The Interpretation of Plato, Timaeus 49 D-E,' American Journal of Philology, 81 (1960) 54Google Scholar, I take it that ‘to toiouton’ is predicated of ‘fire,’ and not the reverse.

13 I borrow this term from Gregory Vlastos, ‘Reasons and Causes in the Phaedo,' in Vlastos, , ed., Plato I: Metaphysics and Epistemology (Garden City, NY: Doubleday 1971), 140Google Scholar. I do not think, as Vlastos does, that the distinction between the Form and immanent character is to be found in the Phaedo, but it seems clearly to be present in the Timaeus. It may not seem clear how the instances of the immanent character are related to the character itself (I would suggest that it is as a part to a whole, or as an individual member of a class to the totality of members) or why Plato feels he needs both the character and its instances. The instance of an immanent character seems to be the result of the interaction of the character with a particular region or portion of the Receptacle; Plato apparently thinks that this interaction particularizes not Just the complex entity, area of space + character, but the portion of the character itself. In any case, the distinction between immanent character and instance (of which, as I note in the body of the paper, Plato makes little use) seems to have a parallel in the distinction between individual and universal accidents, which the Scholastic tradition traced back (whether rightly or wrongly is a matter of dispute) to Aristotle's Categories.

14 Cherniss, 128-30

15 Cornford, 181

16 Cornford discusses the phrase, ‘eph’ hōi,’ in a note (370-1); but he fails to consider this sense of ‘epi'with the dative, which is well attested: cf. Liddell, Scott, and Jones, , A Greek-English Lexicon (Oxford: Clarendon Press 1940)Google Scholar. s.v. ‘epi.' B. 1. g.

17 Aristotle objects (Metaph. A. 9, 99lbl-2) that the Forms could not possibly be the essences of things, since they were supposed to exist in separation from things. This is one of the many points on which the ontologies of Plato and Aristotle are in partial agreement (both agree that Form is essence) and partial disagreement.

18 It is simply not clear whether we are to regard the phenomenal object as merely the image of the Form as reflected in the Receptacle, or as the image + the portion of the Receptacle in which it occurs. When Plato talks of Becoming in and of itself he tends to speak of the phenomenal images as objects, but when he speaks of the Receptacle he tends to include it as a component of some sort in the makeup of phenomenal reality (cf. 51b, quoted above, p. 133). Although it is difficult to determine whether the various portions of the Receptacle are to be included as parts of individual phenomenal objects, it is hard to doubt that the Receptacle as a whole is taken to be part of the phenomenal cosmos.

19 Cornford, 181

20 When Aristotle gives his own resolution of the Third Man Argument (De Sophisticis Elenchis, 178b36-39) he relies on a distinction only verbally distinct from Plato's between ‘this’ (tode ti) and ‘such’ (toionde ti), a distinction of fundamental importance for Aristotle's own ontology and one which is elaborated elsewhere (e.g. at Metaph. Z. 8, esp. 1033b19-1034a8). It is essential to recognize, however, that the similarities in language and thought between Aristotle's explicit solution to the Third Man and the solution I find implicit in Plato are accompanied by equally important differences. Plato uses the this/such distinction to divide all phenomena from their spatial underpinning, whereas Aristotle uses it to distinguish concrete individuals, which he regards as substances, from their attributes. Given Plato's views about the instability of all phenomena, there could be for him no phenomenal substances; thus, he is forced to regard the individual substances of Aristotle's ontology as 'suches’ rather than ‘thisses,’ and to treat them indifferently from their attributes. In addition, there is no place in Aristotle's ontology for the separate Forms that make up the third category in the ontology of the Receptacle passage. Although Aristotle does take seriously in Metaphysics Z the possibility that essence or Form might be primary substance, it is the enmattered form of the concrete individual that he considers, and not the abstact Platonic Form, which he treats as a universal.

21 Cf. Gregory Vlastos, The Third Man Argument in the Parmenides,’ in Allen, R.E., ed., Studies in Plato's Metaphysics (London: Routledge & Kegan Paul 1965), 236-7,Google Scholar 242-3.

22 It is the view of G.E.L. Owen, whose argument I shall discuss in the next section, that only the interpretation of the Forms as paradigms is refuted by the Parmenides, and not the entire Theory of Forms (cf. The Place of the Timaeus in Plato's Dialogues; in Allen, ed., 321-2, n. 3). If the second version of the argument is a sound objection to paradeigmatism, however, it would seem that the first version must be a sound objection to the theory as a whole. This fact makes Owen's view in all likelihood untenable.

23 Cf. e.g. Colin Strang, ‘Plato and the Third Man; in Vlastos, ed., 184-200.

24 Vlastos gave the self-predicational interpretation of ‘F-ness is F’ in ‘The Third Man Argument.’ Alternative treatments of the statement-schema have been given by, e.g., R.E. Allen, ‘Participation and Predication in Plato's Middle Dialogues,’ in Allen, ed., 43-7; H.F. Cherniss, The Relation of the Timaeus to Plato's Later Dialogues; in Allen, ed., 369-74; Vlastos himself, The Unity of the Virtues in the Protagoras,’ in Vlastos, , Platonic Studies (Princeton: Princeton University Press 1973), 259-64Google Scholar; and Nehamas, Alexander, ‘SelfPredication and Plato's Theory of Forms,’ American Philosophical Quarterly, 16 (1979) 93–103Google Scholar.

25 For a list of such statements, cf. Anders Wedberg, ‘The Theory of Ideas,’ in Vlastos, ed., 41, n. 18.

26 The Greek term ‘paradeigma’ may be translated ‘exemplar’ or ‘pattern’ depending on context. It is the former translation (suggested perhaps by the English expression, ‘paradigm case’) that leads to the assumption of self-predication and the resulting absurdities that, e.g., the Form of Large is a large thing, and the Form of Living Being a living being (cf. Vlastos, ‘The Unity of the Virtues,' 261-2). Yet there is no passage in Plato which concerns the Forms in which the troublesome translation is required; furthermore, in light of the generality Plato attributes to his Forms on many occasions, it seems preferable to treat his 'paradigm’ Forms as patterns rather than as exemplars.

27 Plato uses the language of resemblance most frequently not to describe what we should call paradigm cases of resemblance, such as the relation between identical twins or two sets of equal objects, but to describe the relation that holds between a man and his portrait or between the Form of Equality and a set of equal objects. Rather than assuming that resemblance essentially involves the sharing of a property, as Parmenides insists in the second version of the Third Man Argument, Plato insists that when two objects resemble each other in the manner that images resemble originals, the images lack the properties of their originals (Cratylus 432d). Plato knows that the portrait of a man does not resemble its original by being a man; he also knows that the Form of Equality does not resemble its participants by being a pair of equal objects. For a good example of Plato's use of the language of resemblance in a context where the sharing of a property is ruled out, cf. Tim. 37c-38c, where the cosmos is said to resemble its archetype in that the archetype is atemporal and motionless, whereas the cosmos exists forever in time and is in constant motion.

28 The fact that the Forms are not in space means that they cannot be the causes of things by virtue of being in them. This point is raised by Aristotle as an objection to the Theory of Forms (Metaph. A. 9, 991blff; cf. 992a24ff.); yet imma nent causation seems to be a feature of the Theory of Forms in the early Euthyphro (5d1-5), and it is at least arguable that Plato depicts the Forms themselves and not Just their corresponding immanent characters as ‘in’ things in Phaedo 100cff., where he outlines what might be called the ‘official’ position of the middle dialogues on the causal role of the Forms (for contrasting views on the ontology of the Phaedo passage, cf. Vlastos, ‘Reasons and Causes,' 139-43, and Gallop, David, Plato Phaedo (Oxford: Clarendon Press 1975), 195-6Google Scholar; as noted above (n. 13), I side with Gallop on this point). Thus, the docrine of the Timaeus on the causal role of the Forms appears to be different from that of the Phaedo. This change may be the result of Plato's reflection on the arguments against the Forms in the Parmenides; whether or not he accepted the arguments against immanence given at Parm. 130e-131e or the Two-Worlds Argument’ of Parm. 133a-134e at face value, he seems to have realized that he could not insist both on the transcendence of the Forms and on their immanence. His solution to this problem, at least as far as the Timaeus is concerned, is to adhere to the view that the Forms are transcendent, and to relegate their causal role to that of paradeigmata, patterns. The causal function performed by the Forms in the Phaedo is taken over by the immanent characters that correspond to them and, more importantly, by the Demiurge. It is surely significant that, whereas the investigation of causation in the Phaedo led Socrates to posit the Forms, the search for a cause of Becoming in the Timaeus (and, incidentally, in the Philebus, 26e-30e) leads Timaeus to introduce the Demiurge (28c-29a). Thus, though the Forms continue to play a crucial role in the cosmological scheme of the Timaeus, it is the role of patterns or models from which the creator shapes the cosmos, and not that of immanent causes.

29 Vlastos, The Unity of the Virtues,’ 259. Presumably, these self-predicational statements are legitimized by the denial of the other crucial assumption of the Third Man: the non-identity assumption.

30 This is not to say that the Sophist is later than the Timaeus, though I suspect that it is. The theory of the Sophist does not appear in the Timaeus (except for a brief allusion at 35a) because the cosmological project of the Timaeus essentially involves relations between Forms and phenomena, as the Sophist does not, but does not involve relations of predication among Forms, as the project of the Sophist does.

31 Guthrie, W.K.C., A History of Greek Philosophy, v. 5 (Cambridge: Cambridge University Press 1978), 43-4Google Scholar has argued that what is predicated of the Form is not the phenomenal property which its participants derive from the Form, but a corresponding intelligible property. I am not sure what Guthrie means by this, but perhaps his point is the same as mine. In either case, it should be noted, the regress of the Third Man does not arise.

32 ‘The Place of the Timaeus,’ 318-22

33 The results of earlier stylometric studies are summarized in Ross, W.D., Plato's Theory of Ideas (Oxford: Clarendon Press 1951). 2–10Google Scholar; for a more recent study with the same results, cf. Brandwood, L., A Word Index to Plato (Leeds: W.S. Maney & Son 1976)Google Scholar, xvi-xviii.