^{page 29 note 1} See Moreschini, C. (ed.), Platonis Parmenides Phaedrus (Rome, 1966).Google Scholar

^{page 30 note 1} Taylor, A. E., The Parmenides of Plato (Oxford, 1934), 73.Google Scholar

^{page 30 note 2} Cornford, F. M., Plato and Parmenides (London, 1939), 137, n. 1.Google Scholar

^{page 30 note 3} Cornford took his construction of theGreek to have the same effect as Taylor's: he understood his ‘is asserted to belong to’ as equivalent to ‘is predicatively attributed to’ (loc. cit.).

^{page 31 note 1} For what it is worth, I note that AyeTat is often used in the second part of the Parmenides without any ‘about …’-clause following to mean just ‘is said’. An example near by is 142 a 5; a particularly pleasing instance occurs at 160 e 2–7 (where Parmenides seems prepared to use as a synonym: 161 a 3). The sentence in Plato of which I am most strongly reminded by this clause containing is at Crat. 385 C 10: .

^{page 31 note 2} Cf. Anscombe, G. E. M., ‘The New Theory of Forms’, The Monist 1 (1966), 409, whose analysis of this argument is keener than that suggested by Cornford's translation or by Taylor's.Google Scholar

Another possible rendering which has been suggested to me involves taking the genitives not as possessive but as absolute: ‘If “is” is said inasmuch as the one is and “one” inasmuch as what is is one’. This suggests substantially the same interpretation of the argument as does the version I adopt. But I hesitate over taking gen. abs. And I am inclined to read the collocations of words (cf. 1 4.4 e 5) in the same way throughout our passage.

^{page 32 note 1} There is general agreement that Parmenides is made to say first , then TOO OVTOC gvOs, only to give emphasis to a different word in the former expression from that which one would naturally express in using that expression. Each expression, that is to say, is taken to have the same reference. But I differ from Taylor and Cornford in supposing that it is which Plato wants to take the stress in the second expression, in the first.

^{page 32 note 2} Together with the possessive adjective.

^{page 33 note 1} Owen, G. E. L., ‘Plato on Not-Being’, in Plato: A Collection of Critical Essays, i. ed. Vlastos, G. (New York, 1970), 265.Google Scholar

^{page 33 note 2} Taylor took this clause as a genitive absolute (albeit lacking a verb): ‘whereas the subject of the postulate, the existent one, is self-identical’; but Cornford pointed out that the genitive must be possessive, as in the first clause of the protasis (op. cit. 137, n. I).

^{page 34 note 1} The Parmenides of Plato (Glasgow, 1894) 125.Google Scholar

^{page 34 note 2} Inasmuch as ‘the one that is’ is a definite description, and so bears logical and grammatical resemblance to a substantive, ‘the whole one that is’ seems to be a possible, if unusual, expression, in English and in the Greek original; it can be treated as parallel to such phrases as (Prot. 329 e 1–2), ‘the whole face’ or ‘the face as a whole’.

^{page 35 note 1} This is what Taylor and Cornford, who accept Burnet's text, try to maintain he is saying. The former writes: ‘Ergo, any existing one is a whole, and also has a part’; the latter: ‘Therefore, any “One that is” is a whole and also has parts’. But this strains the Greek. And in any case, Parmenides makes no appeal to the idea that ‘the one that is’ has that sort of generality in the regress which immediately follows, although that argument entails the correctness of the idea.

^{page 35 note 2} Especially if the erring scribe, after writing , expected to find the verb indicative, balancing. (as, e.g., at 137 d I–2).

^{page 36 note 1}

^{page 37 note 1} Here and above I use the symbols for MSS. adopted by Moreschini. On Venet. 185, see his preface, especially pp. 9–II.

^{page 37 note 2} This interpretation of the identity of the sub-part A (as really just the original A) receives support from the explanation of why (2), (3), and (4) are supposed to follow later in the argument: ‘for one always possesses being and being one’ i.e., any A always has with it a B, any B an A.

^{page 38 note 1} Plato and Parmenides, 539.

^{page 38 note 2} See ‘Notes on Ryle's Plato’, in Ryle: A Collection of Critical Essays, ed. O. P. Wood and G. Pitcher (New York, 5970), 352–3 369–70.

^{page 38 note 3} I must here add some words of qualification. In using Owen's paper on the Parnzenides I have asked questions of his interpretation which he does not set out to answer there; and consequently I have had to try to construct answers out of things that he does say. I may very well have got hold of the wrong end of the stick, and if so, I gladly apologize. My only defence is that his interpretation of the regress seemed to present a challenge to my own which had to be met. And without pressing that interpretation quite hard the issues between us would have remained unclear. My reasons for not examining Cornford's account of the argument are two: first, it is not very clear; second, I am inclined to think that in his commentary he meant to adopt the sort of interpretation given by Owen, while his translation implies the sort of view I have been urging (except that his rendering of 142 e 1–2 opts unambiguously for an ‘unrestricted’ interpretation: ‘unity can never be lacking to the part “being”, nor being to the part “unity”’)-so I am not sure that further discussion of his views would add much.

^{page 39 note 1} Op. cit. 369–70.

^{page 39 note 2} Op. Cit. 353.

^{page 39 note 3} But I am uncertain on this point. All that now follows should be regarded as directed at what someone tempted by Owen's account of the regress might say, rather than at Owen's own position.

^{page 42 note 1} At 144 c 4 I follow Moreschini in accepting Hermann's conjecture .

^{page 43 note 1} Parmenides' favourite device for indicating that he is drawing on a previous argument is to use expressions like e.g. 149 c 5 (cf. 146 d 1–5; 147 a 3-b 3)159 c 5 (cf. 157 c 1–4); e.g. 148 a 8 (cf. 146 d 5–147 b 6).

^{page 43 note 2} Op. cit. 352–3.

^{page 43 note 3} Op. cit. 369–70.