¹ An early version of this paper was read at a research seminar in Bristol, and I am grateful for comments and suggestions received on that occasion. The version presented here is substantially that offered to the Society at its meeting in February, 1986, but has benefited from discussions begun or continued at that meeting and since: especially with Myles Burnyeat, Geoffrey Lloyd, Mary Margaret Mackenzie (judiciously performing two separate roles, one as editor of the Proceedings, the other as protagonist in an important part of the argument), and Neil O'Sullivan.

2. See Allen, Michael J. B., Marsilio Ficino and the Phaedran charioteer (1981)Google Scholar. This is the view implicitly adopted e.g. by Nussbaum, Martha C. in her paper ‘Poetry, goodness, and understanding in Plato's Phaedrus’ (in Plato on beauty, wisdom, and the arts, ed. Moravcsik, J. and Temko, P. (1982)Google Scholar; reprinted in revised form as chapter 7 of her new book The fragility of goodness (1986)). Elegant though the paper is, its conclusions are undermined by the absence of any account of the latter part of the dialogue, which contains Plato's own reflections on the performances of Lysias and Socrates.

3. I.e. at 265c-d (collection and division). The lesson is in fact drawn from both speeches, which are for the moment treated as one: collection and division are introduced as the means by which ‘the λόγος’ was able to ‘pass over’ from castigating love, as Socrates did in his first speech, to praising it, as he did in his second (265c 5-6).

4. The claim is that the second speech ‘discovered’ divine love by dividing divinely-inspired madness into four parts (266a 6-7, 265b 2-5). What it actually did (or purported to do: see pp. 110 and 117-19 below) was to bring in the other three ‘parts’ of madness as analogues of the fourth, in order to show in a preliminary way the possible benefits of a madness which comes from the gods. For the point behind the redescription in 265-6, see p. 113.

5. Especially if, as is usually supposed, the Phaedrus is written against the background of the Symposium. Moore, John D.'s arguments for placing Phaedrus before Symposium (‘The relation between Plato's Symposium and Phaedrus’, in Patterns in Plato's thought, ed. Moravcsik, J. M. E. (1973), 52–71CrossRefGoogle Scholar) are not persuasive. But the eroticism of the Platonic Socrates – the Lysis notwithstanding – is well established even apart from the Symposium (Protagoras 309a-c, Meno 76c). For some brief remarks on the general question of the dating of the Phaedrus, see pp. 120-1 below.

6. Cf. Mackenzie (article cited in n. 9 below) 66.

7. W. H. Thompson, The Phaedrus of Plato, with English notes and dissertations.

8. Cf. n. 2. Most favoured are Socrates' second speech (as in Nussbaum's case), which seems to offer rich and easy pickings over a wide range of topics; and the last few pages, in the context of the debate on the ‘unwritten doctrines’. This is a bad way to treat any work, but – as I shall suggest – more than usually unfortunate in the case of the Phaedrus.

9. Among the members of this category are Helmbold, W. C. and Holther, W. B. (‘The unity of the Phaedrus’, University of California publications in classical philology 14.9 (1952) 387–417)Google Scholar; Mackenzie, Mary Margaret (‘Paradox in Plato's Phaedrus, PCPS n.s. 28 (1982) 64–76)Google Scholar; and Mackenzie, Mary Margaret, Plato's Folly, Cambridge Ph.D. dissertation (unpublished), 1983)Google Scholar. I think it fair to say that at the various levels of explanation at which these scholars work many aspects of the structure of the dialogue remain puzzling. (On Dr Mackenzie's paper – to which she has now published a sequel, in ‘Putting the Cratylus in its place’, CQ n.s. 36 (1986) 124–50CrossRefGoogle Scholar – see pp. 114-15 below). The two commentaries which have appeared since Thompson's, those of Hackforth (1952) and de Vries (1969), are strangely reticent about larger structural issues; more helpful is Robin's introduction to the Budédition (1933). The main groundwork for the analysis offered in the present paper is provided by my own commentary on the Phaedrus, to be published during 1986.

10. 303e 10-304a 2; 304c 10-d 2.

11. Understood as ‘telling stories about Justice and the other matters upon which you [i.e. Socrates, habitually] discourse’: Thompson xvi, xxii (see below).

12. Thompson xvii.

13. xvii – xviii.

14. xviii n. 1.

15. 277c 2-3.

16. xviii n. 1, with 163. Not all of what I attribute to Thompson here is explicitly said by him; but it is, I think, a legitimate interpretation or extension of what he says.

17. xxi.

18. 276e 1-5.

19. xxii.

20. These are, of course, the things which Socrates habitually talks about. My point is merely that Phaedrus does not here clearly refer to them under that description.

21. Cf. n. 4 above.

22. I shall myself suggest a different explanation of this aspect of what Phaedrus says (p. 112). The point of comparison between written works (in general) and μῦθοι is just that –as Socrates has argued – writers ought to regard themselves, like story-tellers, as composing ‘for the sake of amusement’. But there will then be important consequences for the way in which their products are to be read: see p. 120.

23. 257 c.

24. For a more extended justification of this interpretation of 277c 2-3, see my forthcoming commentary ad loc.

25. ἀναγαῖον, e 5.

26. Szlezàk, Thomas Alexander, Platon und die Schriftlichkeit der Philosophie: Untersuchungen zu den frühen und mittleren Dialogen (1985)CrossRefGoogle Scholar; especially Anhang I.

27. ‘Paradox in Plato's Phaedrus’ (n. 9 above) 65.

28. Real intellectual progress can take place only in the context of the interchange between the knowledgeable dialectician and his pupil, who must possess a ‘fitting’ soul (276e 6) – i.e., presumably, one that is ‘philosophical by nature’ (252e 3); and Socrates suggests that there will only be a small minority of such souls (see especially 249e – 250b). It seems to follow that most of us will in fact be strictly unteachable.

29. If it were not, the whole long discussion of rhetoric would be without point (which is, incidentally, how Mackenzie's reading seems to leave it). The proposed new science of speaking will be a science precisely of the production of conviction (270b).

30. That is, not only will he be unable to teach (which is in any case beyond the capacity of his medium), but he will not even be able to persuade anyone. Persuasion, unlike teaching, is a one-sided process, requiring input only from the speaker or writer: see especially 261e - 262c.

31. I claim, in fact, that Plato consistently treats (public) speaking and writing, in their ideal form, as essentially the same activity – not only sharing the same function, and governed by the same rules, but subject to the same limitations (see pp. 111-12 above). This claim is integral both to Thompson's interpretation and to mine.

32. 275c 6-7 (paralleling the description of the proper attitude of the author at 277d: p. 112 above).

33. That is, it will still need to be questioned, and will be incapable of answering by itself. So for example when he lays out the rules for the composition of a λόγος, (264b), or for skilful speaking/writing as a whole, these rules may be good ones, but the apprentice will still need to ask how they are to apply in the particular case (268a – 272b).

34. Plato warns us that these items – which constitute the special subject-matter both of rhetoric (259e – 260a) and of dialectic (276c) – are inherently slippery (263a - c); statements about them will then perhaps be particularly insecure (cf. 277d 6 - e 2). We are reminded of what the Politicus says about the shortcomings of law as a basis for government (Politicus 294a - 299e).

35. Mackenzie 69.

36. Mackenzie 72.

37. 275 c.

38. That is, to all λόγοι (spoken or written: see n. 31 abve) other than those of the dialectician.

39. Quoted (in Black's translation) by Muecke, D. C., Irony (1970) 18Google Scholar.

40. de Vries, G. J., A commentary on the Phaedrus of Plato (1969) 190 (on 258 e 4-5)Google Scholar.

41. 242a - b, 257a; 262d.

42. Cf. 228b.

43. 265b.

44. 253a.

45. 231d, 241a; 244a-b, 245a.

46. 250a.

47. 251d-e, 250a-b.

48. 256a.

49. 249c-d.

50. 249e 1-2.

51. Cf. especially Ion, and Meno 99c-d.

52. 242b-d.

53. 237a-b.

54. 238d, 241e.

55. 244b 1-3.

56. 244a-d.

57. Cf. Dodds, E. R., The Greeks and the irrational (1951) 70Google Scholar.

58. See especially 261a, one of many echoes of the argument of the Gorgias.

59. It will follow that I reject Nussbaum, Martha's claim (The fragility of goodness, 201Google Scholar) that ‘the style of Socratic philosophizing now [i.e. in the Phaedrus] fuses argument with poetry’. If Plato rehabilitates poetry, it is as part of a new programme for rhetoric, not for philosophy (dialectic).

60. 246d 6 - e 2.

61. That is, if Burnet and others are right in taking Ψυχ ή in d 8 as a gloss (as they surely are).

62. My intention is simply to stress the way in which Plato seems consistently to exploit this aspect of ‘story-telling’ (while also combining it with deeper levels of meaning, which no ordinary ‘story’ would contain). It is as if he were simultaneously challenging us to believe and to disbelieve: see p. 120 below.

63. 247a.

64. References in the commentaries of Hackforth and de Vries, ad loc.

65. Further examples (on which see my forthcoming commentary): 251c 1-5; 252b 1 - c 2; 252c 4-7; 255c 1-2 (with 251c 6-7); 256e 1-2.

66. There is of course no implication that it – and books – are not worth reading at all; just that we should not expect more from them than they are capable of giving.

67. 276e 5-6.

68. For the most recent discussion of this topic, with full references, see Nussbaum, , The fragility of goodness 470–71Google Scholar (n. 5 to ch. 7).

69. See my commentary ad loc.

70. Cornford, F., Plato's cosmology (1937) 112 n. 1Google Scholar.