¹ A History of Greek Philosophy, iv (1975), p. 236Google Scholar. Guthrie hesitates about the relative date of Meno and Gorgias, and discusses Meno first; but he recognizes that most scholars place it after the Gorgias.

² Irwin, T., Plato's Moral Theory: the Early and Middle Dialogues (Oxford, 1977), pp. 2, 102, 131Google Scholar. For his chronology see pp. 291–3, n. 33. The area of agreement diminishes, however, if we look beyond recent British and American scholarship. Consider, for example, the five lists cited by Ross, in Plato's Theory of Ideas (p. 2)Google Scholar: almost the only relevant point on which they all agree is in placing the Gorgias after the Protagoras. Even on this point the ‘consensus’ was challenged by Taylor, A. E. (Plato: the Man and his Work, pp. 20, 235)Google Scholar, Grube, (Plato's Thought, p. xii)Google Scholar, and by Ernst Kapp, who (writing after 1942, published only in 1968) regarded Apology, Crito, Gorgias as the three earliest dialogues and the only ones to be ‘attributed with practical certainty to the period between 399 and 389’ (‘The Theory of Ideas in Plato's Earlier Dialogues’, in Ausgewählte Schriften, p. 80). In his recent Studies in the Styles of Plato (Acta Philosophica Fennica xx, 1967), H. Thesleff partially follows Böhme, R. (Von Sokrates zu Ideenlehre, 1959)Google Scholar in placing the Gorgias earlier than the Protagoras (p. 21 with n. 1).

More recently still, J. Kube, on the basis of a careful study of Plato's theory of technē, has proposed an order which largely anticipates the chronology suggested here: his arrangement of the first five dialogues is exactly the same as mine (Ap., Crito, Ion, Hi. Mi, Gorgias), and the four dialogues of definition also follow in my proposed order (La., Ch., Lysis, Euthyphro). Only in regard to the three major ‘pre-middle’ dialogues (Protagoras, Meno, Euthydemus) do we diverge, since Kube inserts these among the four dialogues of definition. See TEXNH und APETH (Berlin, 1969), pp. 122 ff., with Table of Contents, p. x. Kube does not claim chronological accuracy for his arrangement, but he does suggest (p. 121) that it should correspond roughly to the chronological sequence.

³ Lutoslawski thought otherwise, but he was mistaken, as Ritter pointed out. See Ritter, C., Platon i, pp. 246, 261Google Scholar. For Ritter's unsuccessful later attempt to subdivide the group on stylometric grounds see Dodds, E. R., Plato's Gorgias, p. 19, nn. 1 and 2Google Scholar.

I am indebted to a recent letter from Dr Leonard Brandwood, confirming ‘that stylometry has so far been unsuccessful in indicating any chronological order within this [early] group beyond the probability that certain works, e.g. Phaedo, Symposium, Cratylus, are at or near the end of it’. See his Ph.D. thesis, ‘The Dating of Plato's Works by the Stylistic Method – a Historical and Critical Survey’, University of London (1958)Google Scholar.

⁴ I have no doubt of the authenticity of the Seventh Letter (which may or may not entail that of the Eighth, but surely none of the rest: see Thesleff (op. cit. in n. 2 above) p. 16, n. 1). If, on the other hand, one shares the scepticism of Edelstein and others, we are left with no information on Plato's early career and hence no basis for dating the Gorgias beyond the thematic and emotional links to the Menexenus, whose date is secure. These links point, however, to roughly the same date as the evidence from the Seventh Letter, i.e. the early 380s.

⁵ For the connection between the Seventh Epistle and the Gorgias, see Kapp, ‘Theory of Ideas’, pp. 98 ff., as well as Dodds, , Gorgias, pp. 25–31Google Scholar and Guthrie, , History iv, p. 284Google Scholar.

⁶ For example, the first chapter of Isocrates' Helen mentions with contempt a doctrine of the unity of the virtues in knowledge, which it is natural to read as a reference to Plato's Protagoras (and perhaps to the Meno as well, though this is much less certain). Unfortunately the date of the Helen is unknown; recent scholars place it anywhere from 390 to 380. Since no one doubts that the Protagoras was written before 380 (and probably the Meno as well), this bit of external evidence is of no real use to us. It has been argued that the apparent reference to Isocrates at Euthydemus 304d, 305b ff. must be later than the Panegyricus of 380 bc, which would be more surprising (but not impossible, in my view): see Matthieu, G. in Mélanges Glotz ii, pp. 558–60Google Scholar. But there is too much guesswork involved for this kind of (possible) cross-reference to count as historical evidence.

⁷ The mention of Ismenias of Thebes at Meno 90 a could reflect any date after 395; but Ismenias' career came to a dramatic end in 382, and it may be that event which recalled his name to Plato's attention (cf. Guthrie iv, p. 236, n. 4). This is only guesswork, of course; but the later date fits well with Dover's carefully argued case for dating the Symposium after 385 and before 378 (Phronesis 10 (1965), 1–20).

⁸ No reason, that is, beyond the mere fact that Plato puts it in his mouth. As a principle of historical interpretation, this leads straight to the Taylor-Burnet view of Socrates as author of the theory of Forms (not to mention the fact that a similar principle would saddle Socrates with all the foolish things that Xenophon makes him say). Of course there is a special attraction to the principle in this case: why should Plato wish to mislead us concerning Socrates' motives for such a momentous decision? But we must distinguish between (a) Socrates' moral stance, and (b) the theories and arguments by which it is defended. I suggest that Plato had every reason to represent the former as faithfully as he could, but that with regard to the latter he felt free (or even obliged) to provide the strongest arguments available. Hence in the Phaedo Socrates' attitude in the face of death will be grounded in the theory of Forms. In the Crito and the Gorgios we have the same tendency at work; only the theory is less elaborate.

⁹ See Flashar, H., Der Dialog Ion as Zeugnis Platonischer Philosophie (Berlin, 1958), pp. 96–100Google Scholar. Note that the ‘dramatic date’ of the dialogue would be 394–393 bc (or soon thereafter), to judge by contemporary allusions, e.g. to Athenian rule over Ephesus. (So Flashar, and similarly Méridier in the Budé Plato, vol. v. 1, p. 24.) In fact such allusions can only count as evidence for the date of composition: since Socrates was not available in 394–393, the dialogue has no dramatic date! And of course the same is true for the Menexenus, where Socrates alludes to the King's Peace of 386. The Gorgias is also notoriously indeterminate with regard to a dramatic date. (See Dodds, , Gorgias, p. 17Google Scholar.) Plato is not yet consciously recreating the quasi-historical background characteristic of the Protagoras and the Symposium, where it makes sense to speak of anachronisms. I suggest that the careful invocation of a particular time and place in the Laches, Charmides, Lysis, and Euthyphro should be recognized as an artistic innovation, like the used the reported form. The earlier use of a specific setting in the Crito is quite different, since that is not a free invention on Plato's part, given the topic of the dialogue.

¹⁰ My doubts on the Hippias Major are essentially the same as those expressed by Thesleff, p. 13, n. 4. For the early date of Ion and Hippias Minor, see Thesleff's view, p. 19, and Méridier (Budé, Plato v. 1, pp. 27 f.), who cites Wilamowitz and others.

¹¹ See Dodds, , Gorgias, p. 24Google Scholar; Guthrie iv, p. 213.

¹² For dating Phaedo, Republic, etc. in the 370s, cf. Dodds, , Gorgias, p. 25Google Scholar; Kapp, ‘Theory of Ideas’, pp. 90 f.

¹³ The first to be taken in was apparently Aristotle, whose account of Socrates' position on definition, epagōgē, akrasia, and the like seems largely based upon a reading of the Platonic dialogues as historical documents. In this sense, the conception of Socratic dialogues which I am attacking can be traced directly back to Aristotle (as Pierre Aubenque has reminded me). In effect, Aristotle gives Socrates credit for everything in the dialogues before the doctrine of Forms and the theory of Recollection. When he arrived in Athens 33 years after Socrates' death, the memory even of those who had known the man personally would inevitably be coloured by the brilliant literary portrayal that had been presented in the meantime. Of course, there would still be informants enough to remind Aristotle that the theory of Forms was an innovation, and he could see for himself that it was not there in the more typically ‘Socratic’ dialogues.

Aristotle had no real taste or talent for the history of philosophy, and I see no reason to take his account of Socrates and Plato at face value. (What sensitive reader of the Cratylus can believe that the namesake of that dialogue was a teacher from whom Plato believed he had learned something of great importance?) Aristotle systematically neglects the crucial role played by Parmenidean ontology in the formation of Plato's theory, and emphasizes instead a debt to the Pythagoreans which must have been fashionable to proclaim at the time (see Jaeger, W., Aristotle, tr. Robinson, R., p. 97)Google Scholar, but whose historical reality is extremely doubtful. There is generally a kernel of historical truth in Aristotle's statements about his predecessors, but in order to extract that kernel we must first be able to interpret their doctrine on independent grounds. He is certainly not a model to follow on how to read a Platonic dialogue!

For the documentation and discussion see Deman, Th., Le Témoignage d'Aristote sur Socrate (Paris, 1942)Google Scholar, and Ross, W. D., Aristotle's Metaphysics i, xxxiv ffGoogle Scholar.

¹⁴ I do not deny a recognizably Socratic normative ethics (above all in the Apology and Crito), with some more general claims (no one does wrong willingly, the unity of virtue in knowledge). But how far and in what directions Socrates himself developed these claims theoretically is anyone's guess. What I deny is that Plato's theoretical developments (in Gorgias, Charmides, Protagoras, etc.) should be regarded primarily as attempts to reproduce Socrates' own thought.

¹⁵ ‘Plato on the Unity of the Virtues’, in Facets of Plato's Philosophy, ed. Werkmeister, W. H. (Phronesis suppl. vol. ii, 1976)Google Scholar.

¹⁶ Of course the Gorgias and the Republic have one major theme in common (the theory of justice and the defence of morality), but the latter is in no sense a sequel to the former: Plato begins over again from scratch, with the attack by Thrasymachus (reformulated by Glaucon and Adeimantus) as a replacement for Polus and Callicles.

I cannot deal here with all the arguments brought forward to show that the Gorgias must be later than the Protagoras (e.g. Dodds, , Gorgias, p. 21 fGoogle Scholar. with reference there to Rudberg, 's article in Symbolae Osloensis 30 (1953)Google Scholar). To answer only two points: (1) the arguments which claim that the depiction of Socrates in Protagoras is more lifelike (‘with no wart or wrinkle smoothed out of the portrait’, in Vlastos' memorable phrase) and more aporetic, whereas in the Gorgias he is more idealized and more dogmatic, all rely for their chronological inference on the assumption that Plato began by an historically faithful portrayal, which gradually changed into something else. But this is precisely the historicist assumption I wish to challenge. Furthermore, even the pursuit of historical verisimilitude seems to be a secondary development, part of the literary project that begins after the Gorgias. (See n. 9, above, on the notion of a dramatic date.) (2) It is true that the contrast between pistis and mathēsis (or epistēmē) at Gorgias 454c-455a partially anticipates the distinction between true opinion (doxa) and knowledge which is first drawn at Meno 97b ff., but here again there is no thematic or conceptual continuity. On the one hand, pistis in the Gorgias belongs to the language of rhetoric: it is the subjective state of ‘persuasion’, whatever the audience is convinced of, with no intrinsic connection with the truth; since the audience is gullible, pistis is more likely than not to be false. The term plays no part in the positive account of knowledge in the Gorgias, which is built on the contrast between techē and empeiria (463b ff., 465a, 501 a). The standard Platonic distinction, on the other hand, opposes knowledge to true opinion, where the latter characterizes not the ignorant mob but the statesman (Meno 98c ff.) and even the philosopher (Symp. 202a with 203d – 204b), or is an essential point of contrast for the cognition of Forms (Rep. 477b ff., Timaeus 51 d–e). Nothing connects the discussion of pistis in the Gorgias with the theory of doxa in the Meno, in the way that the latter is directly continued by the treatment of doxa in the Symposium and later works.

¹⁷ So Allen, R. E., Plato's ‘Euthyphro’ and the Earlier Theory of Forms (1970)Google Scholar, and ‘Plato's Earlier Theory of Forms’, in Vlastos, G., ed., The Philosophy of Socrates (1971)Google Scholar.

¹⁸ I think myself that Plato will have had the metaphysics of the Symposium–Phaedo in mind when he wrote the Euthyphro. There is nothing in the text to show that it must be so; but the Parmenidean vein is opened by the characterization of the ‘forms’ in terms of self-identity, self-similarity, and contrariety (compare Euthyphro 5d 1–3 with Parmenides frr. 8, 29 and 55–9). In the case of Lysis and Euthydemus, there are more unmistakable signs of things to come, as we shall see. And in the Meno it seems to me certain that Plato has in mind eternal and intelligible Forms as objects previously cognized by the disembodied soul and recollected here, since (a) the immortal soul requires an eternal object, and (b) unless prenatal cognition was different in kind from ordinary learning it could not help to solve the paradox of inquiry. But Plato does not tell us this in the Meno; he is still not ready to reveal his mysteries, though he has begun the initiation by invoking the authority of learned priests and priestesses (Meno 81 a 10). And the proleptie reading of the Meno is confirmed by Plato's explicit backward reference on the topic of recollection at Phuedo 73a–b.

¹⁹ Or perhaps one should say: at roughly the same time as the Euthyphro. Both dialogues presuppose the Laches, and both are presupposed by the Meno. But the priority of Euthyphro to Protagoras is suggested by two considerations (other than the sheer scale of the latter): (i) Euthyphro, like Laches and Charmides, prepares for the thesis of unity without mentioning this thesis (cf. 12d, the designation of piety as a ‘part’ of justice): and (ii) the semi-technical use of οὐσ⋯α is motivated by the ⋯τι πoτ' ἔστιν question at Euthyphro 11 A 7 but simply taken for granted at Protagoras 349b4.

²⁰ At Gorgias 509e1. 510a4. dynamis is used as an equivalent for technē, as at Hi. Mi. 375d–376a. At Gorgias 472b6 and 486c 1 ούσíα means ‘property’, in the former instance with a play (pace Dodds) on the sense of ‘truth’ or ‘reality’, but no reference to ‘essence’ or ‘entity’ as at Prot. 349 b 4. Charmides 168d2.

One might find the more technical usage of δύναµις prefigured at Gorgias 447 c 2 (τ⋯ς ή δύναµις της τéχνης τού àνδρός, in connection with the τ⋯ ⋯στιν question); but the same phrase is used by Gorgias in a colloquial sense (‘the power of rhetoric’) at 455d7 (cf. 456a5. a8, c6, etc.). Of course the Gorgias does draw a distinction between ποία τις and τίς (at 448E6; cf. 462c 10–d1) which provides the basis for the πάθoς–ούσία opposition in the Euthyphro. But the terminological precision of the latter reflects a theoretical interest which is lacking in the Gorgias.

²¹ The absence of theoretical concern with definition in the Gorgias may also be contrasted with the Charmides. Although the latter does not focus our attention on formal conditions for correct definition, it clearly describes sōphrosynē as a psychic dynamis at 160d6 ff. (without using the term) and discusses the conditions of competence that should permit the interlocutor to discover a definition (158e7–159a 10). elaborating the point made at Laches 190c4–6. Both dynamis and ousia are used in their semi-technical sense at Charmides 168b3–5, d1–2, e5; cf. 169a3.

²² Compare Friedländer, , Plato, tr. Meyerhoff, , ii, pp. 102–4Google Scholar.

²³ So Burnet. Better read (with Gallop) το⋯το τò ⋯ ⋯στι.

²⁴ See parallels in Dodds' note on Gorgias 449c2. For confirmation of Socratic practice, cf. Xenophon, , Mem. 4. 6. 13–15Google Scholar.

²⁵ Cratylus 390b5–10, echoing Euthydemus 289b10–c4; note the mention of Euthydemus at Crut. 386d.

²⁶ So Robinson, R. in the first edition of Plato's Early Dialectic (p. 121)Google Scholar, corrected in the second edition, pp. 117 f.

²⁷ Robinson, (Plato's Early Dialectic, 2nd ed. pp. 137 f.)Google Scholar denies any connection between ἱκανόν τí in the Phaedo and τò àνυπόθετον in Republic 6, but he reads ἱκανόν τɩ as if it were equivalent to ἱκανήν тινα (sc. ὑπόθεσɩν), which it is not.

²⁸ See, e.g. Taylor, A. E., Socrates, p. 144 n.Google Scholar, Plato: the Man and his Work, pp. 258–61; Grube, G. M. A., Plato's Thought, p. 61Google Scholar; and the detailed exegesis by Sullivan, J. P. in Phronesis 6 (1961), 10 ffGoogle Scholar. For recent interpretations which ascribe the hedonism to Socrates (or to Plato) see Irwin, Plato's Moral Theory,ch. iv, and Taylor, C. C. W., Plato's Protagoras (1976), pp. 208–10Google Scholar (with bibliography). Other views in Gulley, N., The Philosophy of Socrates (1968), pp. 111–18Google Scholar; Vlastos, G., ‘Socrates on Acrasia’, Phoenix 23 (1969), 75 fGoogle Scholar.

²⁹ It is true that the many are first described as unready to accept the hedonist view (351 c3), just as Protagoras himself is unwilling to do so (351 c7 ff.). Discussion and clarification is required to get him to assent to it first in the name of the many (354c–e), and then in agreement with the others (358 b 2–3). So the procedure is not exactly that of the Meno (as Geoffrey Lloyd reminds me), where the argument from the hypothesis is given first (87 b–c), and the hypothesis itself defended afterwards (87 c11 ff.). (Note that the Meno ends by regarding its own hypothesis as at least doubtful, since virtue may rest on true belief as well as on knowledge: the hypothesis remains a problematic assumption.) But the function of hypothesis is essentially the same in Meno and in our argument from Protagoras: a useful preliminary (προὔργου, Meno 87a 2) permitting one to solve a problem that cannot be attacked directly.

³⁰ Thus I very much doubt the historicity of the Symposium as an account of Agathon's party, but I do not doubt that Socrates sometimes stood still in the street or in the snow at Potidea, and that Alcibiades was passionately attached to him in the way his speech indicates. The historicity of the meeting with Protagoras in the house of Callias (before Plato's birth…) is at least doubtful. The encounter with Parmenides and Zeno is almost certainly pure fiction.

For those who feel that Plato's artistic simulacra must have some prima facie claim to historical accuracy, 1 refer to Gill, Christopher, ‘The Death of Socrates’, Classical Quarterly n.s. 23 (1973), 25–8Google Scholar.

³¹ This is a revised text of the paper read to the Cambridge Philological Society on 1 November 1979. It has since been presented to several different audiences, including the Société Française de Philosophie in Paris, February 1980. I am indebted to a number of friends, colleagues and auditors for helpful criticism, and in particular to Julia Annas, Pierre Aubenque, Jonathan Barnes, and Anthony Long.