¹ For the view that Plato distinguishes different senses of einai in the Sophist, seeAckrill, J. H., ‘Plato and the Copula: Sophist 251–9’, JHS 77 (1957), 1–7CrossRefGoogle Scholar; Frede, M., Prädication und Existenzaussage, Hypomnemata Heft 18 (Göttingen, 1967)Google Scholar; Owen, G. E. L., ‘Plato on Notbeing’, in Plato, i, ed. Vlastos, G. (London, 1972), 223–67Google Scholar. For the view that Plato confused different senses of einai in earlier works seeOwen, , ‘Notes on Ryle's Plato’, in Ryle, ed. Wood, O. P. and Pitcher, G. (London, 1971), pp. 341–72Google Scholar; Taylor, C. C. W., Plato's Protagoras (Oxford, 1976)Google Scholar; Gallop, D., Plato's Phaedo (Oxford, 1975)Google Scholar. References to Owen are to ‘Plato on Notbeing’, unless otherwise specified.

² See Owen, pp. 256–8; Frede, pp. 12–28.

³ We normally translate to mega as ‘what is big’. I consequently translate to on as ‘what is being’ and to me on as ‘what is notbeing’, to preserve the parallel in the Greek.

⁴ As Moravcsik, J. M. E. notes, ‘Being and meaning in the Sophist’, Ada Philosophica Fennica 14 (1962), 23–78 (p. 60)Google Scholar, this new task is not strictly necessary. If everything that is being is also notbeing (and vice versa), then a separate argument that statement, etc. commune with notbeing is otiose.

⁵ Two examples of this may be found in Wiggins', D. ‘Sentence meaning, negation and Plato's problem of notbeing’ in Plato, ed. Vlastos, G., 1. 268–303Google Scholar, andKeyt's, D. ‘Plato on falsity: Sophist 263’ in Exegesis and Argument, Studies in Greek Philosophy presented to G. Vlastos, ed. by Lee, E. M. et al. , Phronesis Supp. 1 (Assen, 1973), 285–305Google Scholar. Both these commentators think that 263 directly answers the problems in 237–41. Moravcsik, by contrast, sees well enough that 237–1 is answered at 257 (see p. 25). But in common with all commentators before Frede and Owen, he radically misinterprets the dialogue because he assumes without argument that einai means ‘exist’. This causes him problems when he considers the claim at 258e that Parmenides has now been refuted (see p. 66).

⁶ See Owen, pp. 258–9.

⁷ See below, pp. 120–2 for a discussion of this view, which has been adopted byMalcolm, J. in his ‘Plato's analysis of Ὄν and Μ ὂν in the Sophist’, Phronesis 12 (1967), 130–46CrossRefGoogle Scholar, among others.

⁸ Owen (pp. 241–4) takes the view that the problem Plato sets himself is transformed in the course of the puzzles. Owen's view is that Plato initially equates ‘what is notbeing’ with ‘nothing’, but then goes on to set puzzles about ‘nothing’ that are not, and are not confused with, puzzles about what does not exist. What is wrong with this view is that Plato focuses on the expression ‘what is notbeing’ and not on the expression ‘nothing’. The equation of ‘what is notbeing’ with ‘nothing’ is indeed in the first puzzle – but only because of the assumption that ‘what is notbeing’ is opposed to ‘what is being’. When the assumption is rejected, the identification of ‘what is notbeing’ with ‘nothing’ will no longer seem plausible. (Owen does seem to express this view of the first puzzle too. See pp. 247–8 of his article.)

⁹ cf. also Owen, p. 249: ‘in denouncing the sophist’, ‘we seemed to assume’ ‘the incompatibility j, of what is with what is not’, and p. 260: ‘the man who speaks falsely does after all and without paradox ascribe being to what is not or notbeing to what is’.

¹⁰ For different views about late learners, see Owen, p. 251 n. 48; Frede, pp. 61–7; Moravcsik, pp. 56–9.

¹¹ Frede also deals with it on p. 71; but his treatment of it there adds nothing to what he says earlier.

¹² Indeed, Frede himself notes on p. 83 that Plato tends to analyse negative propositions as of the form ‘x is (not-y)’. See alsoMcDowell, J., ‘Falsehood and notbeing in Plato's Sophist’, in Language and Logos, Studies in Ancient Greek Philosophy presented to G. E. L. Owen, ed. Schofield, M. and Nussbaum, M. C. (Cambridge, 1982), pp. 115–34 (pp. 116–21)CrossRefGoogle Scholar. McDowell, however, first recognizing the position of the negation sign in 255–7, concludes that when in place the ‘not’ will ultimately, for Plato, make no difference (p. 121). Plato, who remarks in 257bc that the negation sign operates on the words that follow it, is at least not explicitly drawing McDowell's conclusion.

¹³ We do not yet have a satisfactory account of how the paraphrases offered in 256a 12–b4 resolve the apparent contradiction involved in 256b 10–11. Owen, who rejects the view that different meanings and/or uses of ‘is’ are involved here, has made two different (and incompatible) tentative suggestions as to what is going on here. In note 63 to his ‘Plato on Notbeing’ (p. 258) he suggests first of all that (1) and (2) are in no way paraphrases intended to clarify ‘Motion is the same and not the same’. This approach to the difficulty must be rejected: there can be little doubt that the hopotan of a 12 and the hotan of b 1 pick up the hotan of a 11, and that (1) and (2) each expand on one limb of the clause ‘whenever we call it the same and not the same’ (a 11–12). We do, then, have paraphrases on our hands.

Owen's second suggestion is more helpful. The suggestion is basically that we do have paraphrases here, but paraphrases of ‘the same’ rather than paraphrases of ‘is’. Owen simply suggests that in ‘Motion is identical’, ‘identical’ signifies ‘partaking in the Form Identical’, but that in ‘Motion is not identical’, ‘identical’ signifies ‘the Form Identical’. And of course we would have to say much the same about the other cases – that Being, Rest and Non-identity here sometimes signify the Forms Being, Rest and Non-identity, but sometimes signify ‘partaking in the Forms Being, Rest and Non-identity’.

But there remains a problem with this way of seeing 255e–257a: it does not explain why the Eleatic Visitor feels able to conclude at 256e4–5 that there is infinite notbeing with reference to each of the Kinds. None the less, I think we are forced to accept this reading of the passage faute de mieux.

Lewis, F. A. discusses three other possible readings of the passage in his ‘Did Plato discover the estin of identity?’, California Studies in Classical Antiquity 8 (1975), 113–43 (pp. 130–6)CrossRefGoogle Scholar. One is that Plato is here exploiting the relational nature of the term ‘other’ – but he thinks that ‘external considerations go against… such a reconstruction’ (p. 132). Another would turn on Plato's use of the definite article. Given the complications that arise on this reading, where tauton and thāteron are concerned, it seems very implausible. The third suggestion is that Plato has in mind here a distinction between two different senses of ‘not’. One problem with this suggestion is that only one of each pair of prima facie contradictory propositions need be paraphrased (see Lewis, p. 134); but Plato does seem interested in paraphrasing both propositions in each pair.

¹⁴ See Cornford, F. M., Plato's Theory of Knowledge (London, 1935), e.g. p. 208Google Scholar, where Cornford recognizes two senses of notbeing. Malcolm makes some necessary modifications to Cornford's position (p. 135 n. 13). He also explicitly equates ‘what is in no way being’ with the opposite of being (p. 135), unlike Owen.

¹⁵ For Owen's view, see further pp. 234–5 and p. 259. Frede also uses the two expressions (opposite of being, and what is in no way being) interchangeably (see pp. 74–7). He understands ‘what is in no way being’ in the same way as do Malcolm and Owen. Thus on p. 75, he suggests that in a proposition of the form ‘x is mēdamōs on’, x must allow itself to be characterized in no way, if the proposition is to be meaningful.

¹⁶ See Owen, p. 235 n. 24; Keyt, p. 302; Wiggins, p. 269 n. 1 and p. 284.

¹⁷ Owen locates this assumption only in the last two puzzles. The word enantion does figure most prominently there. But we also find it at 238e8 in the third puzzle. And if two enantia cannot be compresent, then the assumption is present throughout the puzzles. It simply becomes clearer as they progress.

¹⁸ Barnes, J., The Presocratic Philosophers, 1Google Scholar. Thales to Zeno (London, 1979), p. 170Google Scholar.

¹⁹ See Barnes, pp. 169–70.

²⁰ The question how ‘being’ relates to ‘is’ for Plato is discussed by Owen, pp. 232–3; Malcolm, p. 143. Malcolm draws a comparison with Aristotle. Aristotle explicitly remarks in a number of places that e.g. badizei may be expanded at will into esti badizon (e.g. Met. 1017a 19–30). Plato seems to expand esti to esti on in just the same way.

²¹ By Keyt, for example.

²² See Lewis, F. A., ‘Plato on “Not”’, California Studies in Classical Antiquity 9 (1916), 89–115 (pp. 97–9)Google Scholar, for a full discussion of kai pōs here.

²³ cf. Owen, , ‘Proof’, p. 306 n. 1Google Scholar. Lewis, , ‘Plato on “Not”’, also takes this view (pp. 96–7)Google Scholar.

²⁴ This line is taken by Bluck, R. S., in his ‘False statement in the Sophist’, JHS 11 (1957) 181–6 (p. 185 n. 16)Google Scholar.

²⁵ Wiggins, p. 295 n. 15a: ‘the whole project is surely to contrive the explanation of negation with the apparatus provided by the great Kinds’.

²⁶ See Wiggins, p. 301; Keyt, p. 297.

²⁷ It has now been vigorously defended by McDowell, however – though along somewhat different lines from those suggested here.

²⁸ For the translation of hōs here as ‘that’, see Frede, p. 52; Keyt, pp. 287–293.

²⁹ This point was also noted by Bluck, p. 183.

³⁰ I would like to thank Lesley Brown, Geoffrey Lloyd and Malcolm Schofield for helpful comments on earlier drafts of this paper.