1 Phil. Rev., 1956, pp.324-341.

2 In their paper “On Rational Philosophy of Language: The Programme in Plato’s Cratylus Reconsidered”, Mind, 1967, pp.1-20. They discuss this argument on pp.5-6.

3 In his “Making Sense of the Cratylus”, Phronesis, 1970. pp.5-25. His remarks on this argument are on pp. 14-15.

4 Some passages in earlier literature which may contain attempts to deal with a theory of this sort, are discussed inf. on pp. 14a-14b.

5 Cf Smyth, H. W. Greek Grammar (Cambridge: Harvard University Press, 1959), §1607Google Scholar.

6 Cf. LSJ. s.v. IX.1 = expression, utterance, speech; by contrast to IX.3.b = sentence, complete statement.

There is no precise English equivalent for as Plato is using the term throughout this passage. The closest sense is probably that expressed by the French ‘le langage’, as that term is used in recent linguistic studies. Cf. Ullmann, S. Semantics: An Introduction to the Science of Meaning (Oxford: Black-well, 1964). pp.19sqqGoogle Scholar. ‘Le langage’, Ullmann notes (p. 19, n.1), “embraces the faculty of language in all its various forms and manifestations”. This notion includes la langue—the vehicle of communication or “code”—, and la parole—the encoding of a particular message on a given occasion.

In this part of the Cratylus Plato is trying to shift attention away from la parole, from the user and the circumstances surrounding the occasion of an utterance (i.e. away from the pragmatics of language), but he does not have the technical notion of language as a code. On this see further the discussion of immediately following the presentation of the second stage of the argument.

7 , like English ‘say’, has inter alia the two senses: (i) to say or speak, i.e. to utter an articulate sound, and (ii) to mean or intend, i.e. to have a definite signification. Cf. LSI. s.v. III, esp. 1 and 9. The sense mean is probably predominant in this occurrence of , but the ambiguous ‘says’ is used in this translation as it has a better sound in English at this point.

8 As reported in Sextus Empiricus, Adv. math. VIII, 83-87.

9 Sic. Fowler in the Loeb edition; and similarly Méridier in the Budé, Buccellato in his edition, and Ast in his Latin translation, to give but a few examples. Otto Apelt, however, retains the literal sense of the original: “Also ist es möglich dutch die Rede das Seiende zu bestimmen und auch wieder nicht?” [Emphatic spacing added.]

10 Most translators, however, render this passage incorrectly: even Apelt, who alone of those referred to in the preceding note, translated the similar occurrence in 38SB accurately, gives the last two lines here as: „Oder heisst lügen nicht so viel als das Nichtseiende sagen? [Emphatic spacing added.) MéVidier translates the present passage correctly:

but perhaps only because of the French idiom; for in his note ad the second line he remarks: (et non, où) s’explique comme se rapportant non à l’ensemble de la phrase, mais (cf. plus loin ).” But surely he cannot be referring to in the fourth line! yet there are no other occurrences of with any negative in the rest of the dialogue!

11 In the Sophistes, both these cases generate problems: the negation of ’δv’seems to imply the existence of not-being, although the expression cannot be applied to any existing thing (237A-D); and the very attempt to formulate it is taken to result, not in speaking but saying nothing , but in not speaking at all was apparently not understood on the analogy of .

12 Cf. the similar ellipsis in 387C of the Cratylus, where the adversative-negative expression is used to distinguish the conditions for unsuccessful speaking, from those for successful. The latter are given in the complex statement and the former by the negation of this whole set of conditions in the expression quoted. Yet no one would expand the statement of them as

13 Professor A. Mourelatos of the University of Texas, who looked at an earlier draft of this paper, pointed out that cannot automatically be taken as adverbial, and called attention to the alternative interpretation according to which the is taken as “veridical” in C H. Kahn’s sense. (This sense of is discussed inf., in note 20.)

On p.253 of his article “The Creek Verb ‘to be’ and the Concept of Being” (Foundations of Language, 1966), Kahn renders a summary statement of the Creek concept of truth, conflated from this passage in the Cratylus, and from Sophistes 263B, and Aristotle Met. 1011 b” as:

But his n.5 ad loc. seems to reverse this position, and thereby support the interpretation given in this paper. In that note he remarks that “The translation given in (his) text reflects the natural syntax of , e.g., in Protagoras or Aristotle. Plato, however, often seems to play on the alternative construction (taking as adverb rather than as conjunction) and thus to take the formula as meaning: “to speak of the things which are just as they are.. .”.”[Italics added.)

Cornford remarks on the ambiguity of the expression, but observes that “the difference is grammatical rather than substantial” (PTK., p.310). To put it another way, the original Creek is identical in form, whichever function it may have, and our contemporary distinction between the conjunction ‘that’ and the adverb ‘as’, cannot be urged with equivalent force in Classical Greek. But in any case, evidence that the expressions function adverbially in the present ococurrences, is provided by parallel contexts: in the Euthydemus, where Plato first raises the issue (as discussed at the end of Part I of this paper), the expressions occupying the corresponding position in the statement suggesting how false speaking should be described, are , and can only be adverbial: also, later in the Cratylus where Plato examines the relationship between speaking (and naming), and reality, by mean* of the craft analogy (3B7A sqq.), he uses adverbial expressions in similar positions in his statements assessing the various ways that speaking and naming can be carried out (cf. e.g. .

14 Something of this sense is reflected in the derivation of , from a privative and the stem of escaping notice; according to which derivation, truth is literally not-escaping notice, not-concealed, un-hidden, etc. Cf. LSJ s.v. .

The etymological approach to interpreting this notion has lost favour since Heidigger pushed it to a non-Greek extreme, claining that the “original” meaning of the term was the “unhiddenness (Unverb-orgenheit) of Being”, which he construed as a purely ontologica! characteristic, opposed to the epistemologica! one of “correctness of apprehension”. Cf. his Sein und Zeit, pp.219 sqq., and Platons Lehre von der Wahrheit, pp.46 sqq.

But there is no justification for taking the etymology in an exclusively ontological sense: as P. Friedländer points out in his discussion of this issue (Ch. XI of his Plato: An Introduction), the single case, surviving from the early Creek period, in which is understood as (viz. Hesiod, Theogony 233), has to do with the [epistemologica!) characteristic of not-forgetting or not-neglecting. He notes, moreover, that the essential contrary to truth was, for the Creeks, always, lying, deception, conscious distortion, etc.—in short, everything that disturbs, distorts, slants, or conceals the true and real.

The extent to which the etymology does, in fact, provide a useful insight into the conception of truth in the Ancient Greek world, will become clear from the outline of the Ancient Greek world-view, of which it was an integral part, as given in the text of this paper.

15 This Weltbild can be reconstructed from descriptions and usages in Homer, Hesiod, the Lyric poets and early, dramatists. It is based initially on some suggestions and ideas from Professor L. E. Woodbury of the Department of Classics, University College, Toronto. On another occasion I hope to present it in greater detail; it is not possible here to do more than sketch the relevant aspects of a much-oversimplified picture.

16 Thus Solon, in the story told by Herodotus in Bk. I, 30-33, was unable to say that Croesus was the happiest of mortals, without first discovering whether, in addition to his present fortunate circumstances, he would end his life well too .

17 This is illustrated by the story of Cassandra, whose gift of prophecy was vitiated by the fact that no one ever believed her. The story required that kind of twist in order to be consistent with what the Greeks felt could be brought within human power.

18 Cf. II.27-28

19 Cf., for example, the sentiment in the first part of frgt. B234 of Democritus, where he observes that “men ask in prayers for health from the gods and do not know that they have in themselves the power for this .

The trend toward human independence in matters of truth may have begun with Xenophanes, if, in his frgt. 818, there is a genuine opposition between revealing , which the gods are responsible for, and discovering , which mortals may undertake. But these may be just two aspects of the traditional disclosure which is wholly within the purview of the gods alone: for in B34 Xenophanes seems to be suggesting that even if a mortal were. Quite by accident . to utter the most complete [truth] he would not even realize it himself

20 As Professor Mourelatos pointed out, the conception of truth as involving a correspondence between knowledge and reality, or language and reality, is present earlier than Plato’s treatment of it in the Cratylus, especially by implication, in, for example, Herodotus. F. M. Cornford, in his Plato’s Theory of Knowledge, describes the view as current and popular (p.310); cf. also J. V. Luce, “The Date of the Cratylus”, Am. J. Philol., 1964, p.151.

For that matter, a correspondence theory seems to be built into the very conception of being which Kahn, in the article referred to in n.13 sup., identifies as underlying the most fundamental value of (and of the Indo-European *es). This “veridical” usage, as he designates it on p.252, is not simply to be or to exist, but to be so (p.250); and is essentially ambiguous, meaning to be true with respect to statements made in words, and to be the case with respect to facts or situations in the world (p.252). Kahn further explains that the ambiguity is likely based on the one-to-one correspondence between what is the case and the truth of the statement that it is the case (p.252).

21 Most translators, in fact, render correctly: e.g. Fowler by ‘speech’, Mendier by ‘le discours’, Apelt by ‘die Rede’. It is commentators, such as those mentioned at the beginning of this paper, who create the erros in their analyses of the argument. This is evidence that the errors are the result of interpretation, i.e. of bringing presuppositions to the reading of Plato, rather than looking at what he actually said.

22 Most commentatos who construe the argument as having a plural subject, ignore this fact. But D. D. Heath, who takes the argument as being about propositions, even inserts the term into his discussion, as though it occurred in the original text. This is surprising in that his article (“On Plato’s Cratylus”, J. Philol., 1888, pp.192-218) is one of the more insightful studies of the Cratylus that has appeared.

23 This is what is usually investigated in contemporary philosophical studies of language. N. L. Wilson, for example, in The Concept of Language (Toronto: University of Toronto Press, 1959) argues that ‘language’ is an “individuative” rather than a “bulk” term (cf. ‘water’), and language a thing or entity of a certain kind, not a kind of stuff (pp.3-4). His book is an investigation of the structure and functioning of language; Plato, by contrast, in the Cratylus, is investigating the functioning of discourse or speech, and is concerned with “stuff”, not structure. (But cf. n.25 inf.)

24 cannot here mean whole in the sense of structured eniity or organic whole, for then the rest of step five would not make sense. Meridier, perhaps aware of this, renders, ‘dans l’ensemble’; Apelt, ‘als Ganzes’.

Plato does not seem to be perfectly clear about the different ways in which bits of stuff, and parts of structured entities, are totalled into an “all”, and a whole, respectively. Even in the Theaetetus (203E-205A) he identifies the notion of all the bits of something, and that of the totality —204E), and, though suggesting that the whole differs in being a unitary conception distinct from the sum of the parts —204A), he argues that since both totality and whole are that from which nothing is lacking, they are identical (205A)! This is, of course, not to deny that he may have been aware of the distinction without working it. out explicitly. It is presupposed, for example, in the doctrine of the or virtues in the Republic, and figures in the difference between moderation and justice: the former is the appropriate or peculiar virtue of each citizen and each division of the soul individually, and permits each to keep to its own function; the latter is the appropriate virtue of the lot collectively—the state and the soul—and consists in facilitating circumstances in which each individual or division can perform its own function.

25 These ranges must differ in degree only, not in kind, if the characteristic of truth or falsity is to be equally applicable to both. In the Sophistes 261D-262E, Plato argues that is composed of combinations of two kinds of words: and and the smallest such combination is a , the shortest form of discourse —262C). But at this point in the Cratylus he seems to treat both sentences and words as the same in being Dits of discourse, and differing only with respect to the size of the bits. He does not appear to attach any special significance to the fact that the bigger bits are composed of a minimum of two different types of smaller bits, nor even to distinguish such types, though later in the dialogue, at 431B, he does say that in that passage may be sentences.

26 From what argument is given, however, it is clear that names should state individual things as they are if true, and not if false. Since is that which and an can bespoken in both kinds of it would seem that

(a) an in would and

(b) one in would also .

In fact, later in the dislogue, a correct is discovered to be of the sort which reveals the nature or the being of things, and in 430O Plato equates the correct assignment of names to things, with their true assignment, and incorrect with false.

27 G. J. de Vries ad loc., in his “Notes on some Passages of the Cratylus” (Mnemosyne, 1955, pp.290-297), notes that here the partitive genitive seems to stress the as being a part of the But he, too, describes the argument as involving fallacious reasoning, because Socrates seems to “be attributing logical correctness [sic.] not to judgement, but to

28 Cf. 400D of the Cratylus, where Socrates notes that mortals know nothing of the gods, nor of the names the gods use, but that it is clear that they use the true ones .