¹ In speaking of Socrates' dialectical method here I intend two limited historical claims: (i) that Socrates had an identifiable method, the method we see exhibited in Plato's early dialogues; (ii) that this method is recognizably practised in the Theaetetus, albeit on a larger scale and with a content that reflects Plato's own later concerns, including a concern with methodology, not the views of the historical Socrates, (i) is by now relatively uncontroversial; the case for (ii) will be made in the following pages to the extent that I succeed in offering a perspective in which the Theaetetus and the early dialogues can usefully be compared. The aim of such comparison will not be to add directly to our knowledge of the historical Socrates, but to further the understanding of Plato's conception of the Socratic method and of problems of philosophical methodology generally.

² One might think it enough to know that at least some of a certain range of examples were genuine, without knowing, in advance of settling on a definition, which they were. This is a complication which both Socrates and Geach ignore, but since the knowledge in question, if it really is knowledge, would need to rest on some general principle about the status of examples vis-à-vis definition, it will not in the end affect matters if we ignore it too.

³ Geach, P. T., ‘Plato's Euthyphro: An Analysis and Commentary’, Monist 50, No. 3 (10 1966)CrossRefGoogle Scholar, cited from his Logic Matters (Oxford: Blackwell, 1972), 33–34Google Scholar. Others have attributed (A) to Socrates without pursuing its consequence (B), e.g. SirRoss, David, Plato's Theory of Ideas (Oxford: Clarendon Press, 1951), 16Google Scholar; Robinson, Richard, Plato's Earlier Dialectic (Oxford: Clarendon Press, 1953), 50–51Google Scholar. On the other hand, Anderson, Albert, ‘Socratic Reasoning in the Euthyphro’, Review of Metaphysics 22, No. 3 (03 1969), 462–465Google Scholar, and Santas, Gerasimos, ‘The Socratic Fallacy’, Journal of the History of Philosophy 10, No. 2 (04 1972)CrossRefGoogle Scholar, impressed with the fact that examples are after all used in Socratic inquiries, deny the attribution of both (A) and (B); cf. also Nehamas, Alexander, ‘Confusing Universals and Particulars in Plato's Early Dialogues’, Review of Metaphysics 29, No. 2 (12 1975), 287–306Google Scholar. We shall see that the mere use of examples is not to the point where (B) is concerned, although it is relevant to Geach's conditions for sensible discussion.

⁴ Wittgenstein, Ludwig, The Blue and Brown Books (Oxford: Blackwell, 1958), 20Google Scholar: ‘When Socrates asks the question “what is knowledge?” he does not even regard it as a preliminary answer to enumerate cases of knowledge’.

⁵ The claim is explicit in the parallel passage of the earlier Philosophical Grammar, trans. Kenny, Anthony (Oxford: Blackwell, 1974)Google Scholar, §76: ‘… our answer consists in giving such an enumeration and a few analogies’.

⁶ Thus Nahknikian, George, ‘Elenctic Definitions’, in Vlastos, Gregory (ed.), The Philosophy of Socrates (New York: Doubleday, 1971), 144–148Google Scholar, goes seriously wrong, from the point of view of historical accuracy, when he speaks of pre-analytically accepted facts or data to which a Socratic definition is expected to conform.

⁷ This is relevant to Geach's further objection (op. cit., 34–35) that ‘the definition “Knowledge is sense-perception” could have been dismissed at once by looking to Theaetetus' examples of knowledge’.

⁸ Santas, , op. cit., 129–133.Google Scholar

⁹ Santas, , op. cit., 134–139Google Scholar, arguing against the attribution of (A) to Socrates, insists that the latter group of passages do not actually say that only a definition will settle the questions about examples. True enough. But they strongly suggest it, giving no hint of any other way of coming to know the answers; and Meno 71ab (which Santas does not mention) is explicit that a definition of virtue is necessary for knowing whether virtue is acquired by teaching. Moreover, Santas does not take account of the evidence of Theaet. 147ab, to be discussed shortly, and its predecessor Meno 79bc.

¹⁰ Thus Soph. 243b and, probably, Theaet. 184a. As these two passages illustrate, the Greek verb sunienai is used not only for plain linguistic competence (Charm. 160a, Prot. 325c, e, Laws 791e) but also for deeper kinds of understanding which presuppose the linguistic one (Ion 530c, Prot. 339a and, in connection with definition, Phdr. 249b).

¹¹ The interpretation may be compared with that of McDowell, John, Plato: Theaetetus (Oxford: Clarendon Press, 1973), 114Google Scholar. He does not distinguish levels of understanding, but takes it that Socrates demands articulate knowledge, formulated in a definition, as a condition for any understanding of an expression. McDowell then rightly objects that the condition is not plausible, even less so when extended to expressions like ‘carpentry’ which do not themselves contain the word ‘knowledge’. On my interpretation, the condition for (philosophical) understanding is reasonable, the crucial extension is not, but we shall see how Plato could mistakenly think it was.

¹² The expansion of ‘a knows x’ into ‘a knows what x is’ is standard and can be observed at Meno 79c8–9, Theaet. 147b2–5.

¹³ Consider from this point of view Soph. 231b–236c, the survey which leads to the idea that the key notion needed in defining the sophist is the notion of an image—one of many passages relevant to the issue, which it would take us too far afield to explore here. McDowell, , op. cit., 115Google Scholar refers to Phil. 12c–18d as allowing a connection, in some cases at least, between an account of what X is and an account of the instances or kinds of X.

¹⁴ Even at Laches 190C, where Socrates claims that one who knows what something is should be able to say what it is, he at once makes clear that he envisages a considerable inquiry to lie ahead (190cd); it is Laches who thinks he can say straight off what courage is (190e4). Cf. also Hipp. ma. 295a.Google Scholar

¹⁵ Santas, , op. cit., 134, note 6Google Scholar, says he does not dispute that Socrates holds this version of (A); yet it is sufficient for (B)'s rejection of examples, which Santas does dispute.

¹⁶ Cf. also his dilemma at 154cd and the way it is resolved.

¹⁷ Illustrations from the useful discussions of, respectively, Robinson, Richard, Definition (Oxford: Clarendon Press, 1950), 113–114Google Scholar, and Lakatos, I., ‘Proofs and Refutations’, British Journal for the Philosophy of Science 14, No. 53 (05 1963)–14, No. 56 (02 1964).Google Scholar

¹⁸ Cf. Gasking, Douglas, ‘Clusters’, Australasian Journal of Philosophy 38, No. 1 (05 1960), 30–31CrossRefGoogle Scholar, from whose discussion the illustration is borrowed.

¹⁹ Santas, , op. cit., 138Google Scholar, notices the difference between morals and mathematics, and he too uses it to impugn the generality of Geach's (A)—but at the same time he virtually concedes that (A) does hold for moral concepts.

²⁰ As is emphasized at Euthyph. 7b–8e (a passage Geach goes on to take issue with (op. cit., 35–7)) and Phdr. 263ac, which adds love as another disputed item; cf. also Gorg. 451 ff.Google Scholar

²¹ Moore, G. E., Some Main Problems of Philosophy (London: George Allen & Unwin; New York: The Macmillan Co., 1953), 119–126.Google Scholar

²² The first paper is to be found in Moore, G. E., Philosophical Studies (London: Routledge & Kegan Paul, 1922)Google Scholar, the argument under discussion being on p. 228, the second in his Philosophical Papers (London: George Allen & Unwin; New York: The Macmillan Co., 1959).Google Scholar

²³ Schilpp, Paul Arthur (ed.), The Philosophy of G. E. Moore (New York: Tudor Publishing Co., 1942), 668–669; cf. also 673–674.Google Scholar

²⁴ Some Main Problems of Philosophy, op. cit., 124–125.Google Scholar

²⁵ Philosophical Papers, op. cit., 53.Google Scholar

²⁶ Cf. ‘Four Forms of Scepticism’ in Philosophical Papers, op. cit., 225Google Scholar, ‘Certainty’, ibid., 243, and an entry on ‘Immediate Knowledge’ in Lewy, Casimir (ed.), The Commonplace Book of G. E. Moore 1919–1953 (London: George Allen & Unwin; New York: The Macmillan Co., 1962), 173–178Google Scholar. On the basis of Moore's indications in the Preface to Philosophical Papers, the two papers are to be assigned to the period 1940–44, although they were not published until later; the third item is from a notebook of circa 1941–42.

²⁷ ‘Certainty’, Philosophical Papers, op. cit., 244–251Google Scholar; compare and contrast ‘Proof of an External World’, ibid., 149.

²⁸ Cf. the editor's note, Philosophical Papers, op. cit., 251.Google Scholar

²⁹ Wittgenstein, Ludwig, On Certainty (Oxford: Blackwell, 1969), §125.Google Scholar

³⁰ On Certainty, op. cit., §10.

³¹ This paper was written during my tenure of a Radcliffe Fellowship. I should like to acknowledge the generosity of the Radcliffe Trustees and to thank University College, London for allowing me to take up the Fellowship. I have greatly benefited from discussion of earlier drafts at meetings in London, Oxford and Birmingham and from suggestions and criticisms by James Dybikowski, Terence Irwin, Malcolm Schofield, Richard Sorabji, Gregory Vlastos, Richard Wollheim and Michael Woods. Above all I would like to thank Roderick Firth, who first drew my attention to the problem I have been concerned with and set me thinking on the lines followed in the paper.