Phaedo

Socrates is not maintaining an absolute veto upon suicide. On the contrary, with the words 'until God sends some necessity, such as the one now before us' (c7—8), he implies that his own death will be self- inflicted. In his case, at least, self-destruction would be not merely permissible, but a religious duty. That such acts, when required by the state, were viewed by Plato as exceptions to the general rule is clear from the provisions for treatment of suicides in the Laws (873c—d). Indeed, a still broader range of exceptions is there envis­aged, since a suicide is subject to punitive burial only if 'no state has required it of him, no stress of cruel and inevitable calamity driven him to the act, and he has been involved in no desperate and intol­erable disgrace'. This is a far cry from the 'condemnation of suicide in every circumstance and form' for which Geddes (201) admires the present passage, but which is not to be found in it. Socrates is not denouncing suicide at large; he is trying to explain why the philosopher's desire for death would not justify him in procuring it for himself. See previous note.

'Kill' at 62c2 and 62c7 could possibly be interpreted, with Loriaux (69), to mean 'try to kill', on the ground that punishment proper could be inflicted only upon something still living. However, the words 'if you had any punishment at hand' (c3—4) are perhaps meant to concede that for a successful suicide no punishment would be practicable. Moreover, posthumous 'punishments', such as dis­honouring the corpse, were sometimes imposed. No distinctive penalties in the afterlife are specified in the closing myth, although culpable acts of suicide would no doubt have been included among 'wrongful acts of killing' (113e, cf. Republic 615c), and punished accordingly.

On the whole subject see J. M. Rist, Stoic Philosophy (Cambridge, 1969), Ch.13, and, for a broader study, A. Alvarez, The Savage God (London, 1971).

2.2 Socrates' Defence (63e8—69e5)

This section contains a passionate apologia for the philosophic life. It resolves the contradiction with which Socrates had been faced. The philosopher's whole life is a preparation for death. He should therefore welcome death when it comes.

64c2—9. The translation 'death is something' (c2) preserves the Greek idiom. Less literally: 'there is such a thing as death'. Socrates means, non-technically, that 'death occurs' or 'there is death'. A 'Form' of Death will be needed later (105d9) but Socrates can hardly be referring to it here.

In defining death at 64c4—5 as 'the separation of the soul from the body', he seems to be treating it as the 'event' in which their separation is effected—cf.67d4, d9—10, and Gorgias 524b. At 64c5— 8, however, it is treated no longer as an 'event', but as the 'state' of 'being dead', the separated condition of soul and body. Cf.66e6— 67a2. See also note 4 and on 71d5-e3. It is to that 'state' that Socrates seems to refer when he asks at 64c8: 'Death can't be anything else but that, can it?'

Several difficulties arise here. (1) In what sense 'is there' such a thing as death? People die. But do they exist when dead? Or are 'the dead' simply those who no longer exist at all? Socrates avoids these questions by assuming that a living being is a body conjoined with a soul, and by defining death as the separation of one from the other. For a person to be dead is for his soul to be separated from his body and vice versa. This leaves it unclear what is the proper subject of the predicate 'dead'. Is it only the man who is to be called 'dead' when soul and body are parted, or may 'dead' be predicated of soul and body separately? See on 105el0-107al (p.221). Socrates will generally avoid speaking of 'dead souls'—although the soul is twice said to 'die' (77d4, 84b2)-perhaps because this would produce a conflict with the conclusion of the dialogue, that the soul is 'death­less': 'it won't admit death, nor will it be dead' (106b3-4). He will speak, rather, of 'the souls of the dead' (72a7, 72d9). But the question now arises whether 'the souls of the dead' exist.

(2) By defining death as he does, Socrates seems to prejudge this question in favour of the soul's survival. Hackforth (44, n.l) tries to defend him against this objection: 'all that Socrates here wants is an admission that we can properly think and speak of soul "apart" from body; whether soul continues to exist when thus apart is the question at issue.' However, 'being dead' is taken to include 'the soul's being apart, alone by itself, separated from the body'. From this definition, conjoined with the admission that there is such a thing as death (c2), it follows that the soul does exist apart from the body. If it did not, there would be no such thing as death, in the sense given to the word at 64c5—8. It therefore seems hard to acquit Socrates of pre­judging the issue at this point. See also on 70c4—8 (p.106), 71al2—b5.

A definition of death in terms of the relation between soul and body provides, it need hardly be said, no useful criterion for determining when a person is dead. But even if this medical question is judged irrelevant, the definition gives a questionable account of the concept. For it provides that the body has 'come to be apart, separated from the soul, alone by itself. But the body's post mortem existence is not, in fact, a necessary condition for death, whether death be viewed as an event or as a state. The bodies of those long dead do not exist. And after certain kinds of death the body may not remain 'alone by itself at all. Nor, in saying that a person is dead, need any reference be made to the existence or state of his body.

The definition of death poses some difficulty for the inter­pretation of 'deathless', as that term will later be applied to the soul. If 'death' means 'the soul's being separated from the body', how can the soul be 'deathless' in that sense? So far from denying that it is separated from the body, Socrates affirms precisely this. For the word 'deathless' or 'immortal' see on 72e7—73a3,95b5—e6,105dl3— e9. As applied to the soul it must mean the capacity to persist through separation from the body. The soul survives the transition from attachment to the body to a state of detachment from it. It survives, we might say, the 'event' of death into the 'state' of death. In this sense it could be held 'deathless' in a manner consistent with the present definition. It is, however, a further question whether the term 'death' is itself used consistently with this definition throughout the dialogue. See on 71d5-e3, 91d2-9, 105el0-107al (p.221), and cf.88al—b8,95b8—el, 106b2-4,106e5-7.

Behind the above difficulties lies a more fundamental one. Is the view that living beings are a composite of body and soul tenable? The assumption that an independent entity, 'the soul', animates the body is implicit throughout the discussion, and is made explicit at 105c9—d5 (cf.79b 1—2). Yet for many readers the question of the soul's immortality is pre-empted by that of its existence. Socrates' conception of it will later be challenged by Simmias, with the counter-hypothesis that the soul is an 'attunement' of bodily elements (85e—86d). By refuting that alternative (91e—95a), Socrates clears the way for his own view. More positive support for the assumption can be found in other sections of the dialogue. See on 76cll—13 and 105el0-107al (p.220). A philosophical reading of the dialogue should examine the basis for it. See next note.

64e4—65a3. Socrates now pursues the implications of the soul- body dichotomy for the philosopher's life: his attention will be turned away from the body as far as possible, and 'towards the soul'. The traditional 'soul' has been used throughout for psyche. But the dialogue contains no explicit account or consistent usage of this term, and it therefore remains unclear exactly what it is whose immortality Socrates is seeking to prove. No short account can do justice to this huge subject. All that can be attempted is an outline of some aspects of it that are especially relevant to the Phaedo.

The soul is that element in us whose good condition con­stitutes our true well-being. Its 'care' is therefore paramount, and overrides all other human concerns. This teaching, familiar from other Platonic dialogues, is elaborated in the Phaedo in two ways. First, the body is seen as a constant hindrance to the care of the soul, its demands incessantly conflicting with the soul's interests (66b—d). Hence the repeated stress upon 'purifying' the soul (67a—c, 69c), 'releasing' it from the body as far as possible (67c—d), and 'admon­ishing' or 'punishing' the body (94b—d). Hence, also, the stern disparagement of bodily pleasures and material possessions (64d— 65a). Nowhere else in Plato is asceticism so uncompromisingly extolled. Secondly, the nurture of the soul is viewed as of vital importance, not only for this life but for the life to come. This gospel, proclaimed with unequalled eloquence, imbues the dialogue with its distinctive ethos of religious fervour and earnest moral concern.See especially 63b5-9, 69d2-e2, 80d5-84b8,90d9-91al, 107cl—5, 108a6—c5, 114dl-115a3.

The soul is 'the true self, 'the real person*. Hence in caring for his soul, the philosopher is caring for 'himself. In this vein Socrates will urge his friends to take care of'themselves' (115b5—cl). After his death they should not speak of burying 'him', but rather 'his body' (115c4—116al). 'Socrates' is to be distinguished from his body, and is thus implicitly identified with his soul. The use of 'our souls' for 'us' will be critical for the argument at certain points. See on 70c4—8 (p. 105). An argument for identifying the person with his soul is presented in the (possibly spurious) Alcibiades I (129a- 130e).

The soul is often equivalent to 'intellect' or 'reason', a 'thinking faculty' or 'cognitive principle' by which the quest for wisdom is pursued. In this quest, the soul is referred to as 'attaining the truth' (65b9), as 'reasoning' (65c2-5), 'getting knowledge' (76c6), and 'possessing wisdom' (76c 12-13). It seeks 'vision' or 'grasp' of its objects, analogous to the seeing or touching of material things. It is an organ of intellectual sight or touch, or a 'subject' by which the truth is apprehended. At 67c3 'soul' is replaced by 'intellect', as if that term were a natural variant for it. Cf.65e7, 66a2, 79a3, and see on 65c2-4.

The soul also functions as 'the rational self in opposition to emotions and bodily desires. This opposition is not between 'reason' and some other 'part' of the soul, but simply between soul and body. No distinction is drawn in the Phaedo between 'reason', 'spirit', and 'appetite', which Plato treats elsewhere as separate 'parts' of the soul— see Republic 435a—441c andPhaedrus 246a—b, 253c—e. Pheno­mena viewed in those dialogues as 'mental conflicts' are described in the Phaedo as struggles between soul and body: the soul opposes 'the bodily feelings' (94b7—cl, c9—e6), which include not only hunger and thirst (94b8—10), but also passions and fears (94d5). In opposing them, the soul thus performs the role played in the Republic (439c2—d8, 441b2—c2) by the 'reasoning element' only.

In several passages the soul is treated as the subject not only of rational thought but of a wide range of conscious states. Thus Socrates can speak of 'the pleasures that come by way of the body' (65a7), and of hearing or sight, pleasure or pain, as 'bothering' the soul (65c5—9). Later (79c2—8), the soul is said to be 'confused and dizzy' when it studies things through the senses, to be 'intensely pleased or pained' (83c5—6), and to 'share opinions and pleasures with the body' (83d7). Pleasure and pain are viewed not merely as changes in the body, but as psychic states (cf. Philebus 33d2-el, 35c3—d7, 55b3). In such places the soul is depicted as an inner recipient or 'owner' of sense impressions and feelings—or as what will, in a later era, be called 'the mind'. Cf. Theaetetus 184b—186a.

The soul is often conceived as a 'life-principle', or 'animating agent', bringing life to the body which it occupies. This idea, although largely absent from the English 'soul', is close to the root sense of psyche, which is cognate with a verb meaning 'to breathe'. It under­lies the definition of death already given at 64c, and will appear in both the first and the final arguments for immortality (72all—d5, 105c9—d5). This view of the soul is not restricted to human beings but applies to all living things. Cf.70d7—9. Note also that since the universe as a whole is regarded by Plato as a living thing (cf. Timaeus 30d, 92c), soul as life-principle may be conceived as animating not only individual organisms, but also the universe itself. For the etymology of psyche see Cratylus 399d-400b.

One feature of 'soul' is easily obscured in translation. The word may be used not only of individuals, but also in a generic sense to connote a kind of 'stuff, just as 'body' may be used to mean not only individual bodies but also 'matter'. See on 70c8—d6 and 80c2— d7. It is not always clear whether 'soul-stuff or 'the individual soul' is meant. The distinction is critical for immortality. For the idea that 'soul is immortal', merely in the sense that there exists a permanent quantum of 'soul-stuff, would no more imply the immortality of individual souls than the notion of a permanent quantum of matter implies the immortality of individual bodies. Clearly, the latter notion is compatible with the view that no individual body is immortal, and that there may be continual variation in the total number of bodies existing at any given time. Might not the same hold true of souls? At Republic 611a it is argued that the total number of souls must be 'always the same'. This would imply, assuming that no soul can be shared by several organisms, a fixed limit to the number of things that could be simultaneously living.

The foregoing notions of soul are immensely varied. They range, on a descending scale, from the intellectual and spiritual functions of a small class of human beings to that which is shared by living things in general. This variation poses several problems. First, it may be asked whether arguments appealing to such widely divergent con­cepts of soul could establish the immortality of the same thing. The notions of soul as 'intellect' and soul as 'life-principle' will, between them, largely monopolize the coming arguments. But these 'souls' can hardly be identified. Soul as intellect is said to be 'similar to what is unvarying' in virtue of its kinship with the change­less objects which it knows (80b2—3); whereas soul as life-principle is an agent of change in living bodies, and will, through its association with the body, be subject to change itself (e.g. 79c2—8, 81bl-c6, 83d4-e3).

Secondly, it remains uncertain which psychical characteristics are supposed in the Phaedo to survive death, and whether any others could, consistently with its arguments, be supposed to do so. Much of the argument suggests survival of the intellect or rational self only; and it may be asked whether this would be sufficient for, or even compatible with, the faith in 'personal' survival which Socrates affirms. If, for example, such features as memory or emotion are required for personal immortality, but are, at least implicitly, excluded from survival by the philosophical arguments, then personal survival not only goes unproven in the Phaedo but is actually ruled out.

Finally, the wide range of notions of 'soul' gives rise to a series of images of the soul-body relationship that can hardly be reconciled with one another. We may ask, for example, how the soul can at once 'bring life' to the body (105 c—d), 'rule and be master' of the body (80a, 94b—d), and yet be a 'prisoner' within the body, co­operating in its own captivity (82e—83a). The dialogue contains no single, logically coherent 'doctrine' that might answer such questions. As E. R. Dodds has said (G.I. 179), 'the Classical Age inherited a whole series of inconsistent pictures of the "soul" or "self'.' Several of these are amalgamated, no doubt consciously, in Plato's thought. But the more clearly they are distinguished, the more usefully the philosophical arguments can be explored.

See, further, E. R. Dodds, op.cit., Ch.7, and T. M. Robinson, P.P., Ch.2.

6Sbl—7. Plato's language for sense experience is often hard to interpret. The words translated 'sight' and 'hearing' at 65b2 may mean either, as here, the senses of sight and hearing, or, as at 74dl and 98d7, visual and auditory sensations. Both words recur at 65c6, where they seem best taken in the latter sense. For the meaning of 'sight' at 65e8, where the text is uncertain, see note 9.

The generic word for 'perception' is similarly ambiguous between 'sense' and 'sensation'. At 65b5 the phrase translated 'the bodily senses' means, literally, 'the senses around the body', suggesting that they are thought of as having bodily location. 'Sense' has been used in the translation also at 75bll, 79a2, 79c4, 79c5, 83a5, 96b5-6, and 99e4. At 65dl 1 'sense-perception' has been used, since the phrase 'those that come by way of the body' suggests that sensations are meant. This translation has been used also at 73c7 and 76a2. The meaning is uncertain at 66al, 75a7, 75all, 75bl, 75b6, 75e3, and 76d9. See also on 74c7-d3 and note 6.

The senses are continually disparaged—cf.65c5—10, 65e6—66a6, 79c2—8, 83a4—7, 99el-4. Yet just how they are 'neither accurate nor clear' (b5) is never properly explained. Can Socrates be thinking of misjudgements of size due to distance, or of refraction or other sources of visual error? His talk of the soul being 'taken in by the body' (65bll, cf.83a4—5) might suggest this. Cf.Protagoras 356c5— 357a2, Republic 602c—e. Such 'deceit', however, occurs against a background of perceptual judgements that are generally correct. Moreover, the senses themselves play an indispensable part in measurement, and thus in the correction of error. Mistaken judge­ments of size or shape would therefore fail to illustrate the theme that 'we neither hear nor see anything accurately' (b3—4), unless, indeed, the poets are to be credited with some version of the modern 'argument from illusion', to show that we never perceive physical objects as they really are.

Socrates' quarrel with the senses appears more radical. It is not merely that they misrepresent the physical world, but that they never present anything else. They hamper the soul's access to the real objects of its understanding, the 'Forms', that will be introduced at 65d4—5 below. They give no indication that there are any such objects, and strongly suggest that there are not. Clear philosophic understanding can therefore be achieved only when normal sensory awareness is suspended.

At 65b3, if any specific poets are meant, the reference may be to Epicharmus' line 'intelligence sees; intelligence hears; the others are deaf and blind' (DK 23 B 12).

65c2—4. 'The things that are'-, here, as often, Plato uses the present participle of the verb 'to be' with the definite article, to denote the object of the soul's understanding. The singular phrase has generally been translated 'that which is', and the plural 'the things that are'.

The Greek verb 'to be' has both (1) an 'incomplete' use, in which it requires a complement, expressed or understood, and (2) a 'complete' or 'absolute' use, in which it stands by itself. In use (1) the verb may express identity of subject with complement, or it may predicate the complement of the subject. In some contexts it is not certain which of these is meant. Nor is it clear whether any such distinction was recognized by Plato when this dialogue was written. Not until the much later Sophist is any systematic clarifi­cation of 'being' attempted. The distinction between the 'is' of identity and that of predication sometimes affects the interpretation of a phrase (see on 75c7—d6, p.131), or the assessment of an argument (see on 93d 1—5, p. 162).

In use (2), which survives only vestigially in English, the verb commonly means either 'to be true' or 'to exist'. It has often been rendered in one or other of these ways. In the present passage, however, it is not clear whether 'the things that are' are thought of as 'truths' or 'existents'. They are said to become clear to the soul in its reasoning. This might suggest that, as Burnet holds (note on 65c2—4), 'truths' are meant. Yet Socrates will shortly go on to speak of 'the hunt for each of the things that are' (66a3) in connection with 'Forms', which are introduced in terms suggestive of existents (see next note). It should be borne in mind that Plato constantly treats thought processes as analogous to perceptual ones. Sight and touch are his standard models for intellectual discovery and understanding. Such understanding, even when expressed as vision or grasp of 'objects', may well be thought of as including the apprehension of truths. If so, a sharp dichotomy between 'truths' and 'existents' need not be drawn at this point. Cf. C. H. Kahn (V.B.A.G. 457): 'If we recognise some interaction between the old use of the participle to refer to facts or events and the new use to designate whatever things there are in the world, this will help us to understand the persistent Greek refusal to make any sharp distinction between states of affairs or facts with a propositional structure, on the one hand, and individual objects or entities on the other. For the Greeks, both types count as "beings".'

Sometimes 'the things that are' clearly refers to Forms (78d4, 83b2), sometimes clearly not (79a6, 97d7, 99d5). More often the reference is indefinite: cf.65c9, 66a8, 66c2, 74dl0, 82e3-4, 90c4, 90d6, 99e6, 100a2, 101e3, 103e9.

65d4—e5. The so-called 'Theory of Forms' is introduced at this point. As this theory is of central important in Plato's thought, and is more prominent in this work than in any dialogue usually regarded as of earlier date, a general note on the Forms may be useful here. The following account will be based mainly upon the Phaedo, and will be limited to (1) the terminology of the Theory, (2) its philosophical basis, and (3) its relationship with the proofs of immortality.

(1) Two of Plato's commonest semi-technical terms for Forms are eidos (102bl, 103e3, 104c7, 106d6) and idea (104b9, 104d2, 104d6, 104d9, 104el, 105dl3). Both derive from a common root which appears in the Greek verb 'to see', and they are thus connected with the visual appearance of a thing. Ordinarily, they connote 'shape' or 'figure', and also 'sort' or 'species'. These associations are preserved in the conventional translation 'form', which has been adopted for both words in their Form-referring use. A third word, morphe, also normally meaning 'shape', has been translated 'charac­ter' (103e5, 104dl0). As used for Forms, the three words seem inter­changeable. See note 72.

The English 'idea' is an exact transliteration of one of the above words. But its associations are misleading. The Forms are in no sense 'psychological' or 'subjective' entities. They are not thoughts or concepts, existing only 'in the mind' of a thinker. Nor are they mind-dependent objects. They may be thought of by the mind, 'viewed with the soul' (66el—2, and see previous note), but they do not depend upon being thought of for their existence. Just as physical objects exist independently of eye and vision, so Forms exist independently of mind and thought.

Individual Forms are typically designated by a neuter singular adjective, occasionally standing alone, but more often with the definite article or the pronoun 'itself, or with both. Such expressions have been translated as literally as possible. Thus, phrases such as 'the equal', 'the beautiful itself, have generally been retained. For the present passage, however, see note 5, and on 74a9—bl, lOObl— c8. Abstract nouns are frequently used for Forms, and have been translated by the corresponding English abstracts, e.g., 'health', 'largeness', 'strength' (65dl2—13). For the special problem posed by

Forms for numbers, see on 101b9—c9.

Individual Forms are referred to in these Notes as, e.g., 'the Form Equal', 'the Form Beautiful', 'the Form of Three', or as 'Equality', 'Beauty', etc. The variable F is sometimes used to stand for the adjectival component—'the Form F\ 'the F', 'the F itself, 'F-ness'. Note that phrases of the form 'the F' are liable to be ambiguous. They need not refer to a Form, but may simply mean 'that which is F', i.e. a thing characterized as F. See on 70e4—71all (p.108), 103a4—c9. It is sometimes uncertain whether a Form is being referred to or not. Typographical distinctions for Forms of any kind have therefore usually been avoided in the translation.

Forms are designated by several other expressions. 'That which is' and 'the things that are' are sometimes, although not always, so used—see previous note. Plato also uses ousia, an abstract noun ordinarily meaning 'property' or 'estate'. This has been translated 'Being', to distinguish it from the present participle 'being', on which it is formed—see note 7. Sometimes a Form F is referred to as 'the F in nature'. See on 103a4—c9. In several places an exceptionally puzzling expression for Forms is used: literally 'what F is', or perhaps 'what is F', or 'the F which is'—with or without the emphatic 'itself. Sometimes F is omitted, leaving only 'what it is'. The phrase is still in need of elucidation. Cf.65dl3-el, 74b2,74d6, 75bl—2, 75b5—6, 7Sd2, 78d3-5, 92d9. See on 75c7-d6 (p.130), and notes 7,21,24-6, 28, 31,50.

(2) Platonic Forms are conceived as timeless and non-spatial objects, immutable entities set over against the changing world of sensible things, a realm of 'things that are' transcending the 'things that come to be'. They are the realities of which sensible objects are mere appearances or 'phenomena'. Each Form is unique, a single 'one' set over its 'many' sensible instances. Each is a perfect original, of which sensible things are imperfect copies. Forms are invisible to the eye, but 'intelligible'—accessible to 'intelligence', the mental organ that 'sees' them. In terms of a contrast that Plato draws in the Republic between 'knowledge' and 'opinion', they are the only objects of true 'knowledge'. See on 84a2—b8.

Plato writes of Forms, such as Equality, Largeness and Health, as if they existed 'alone by themselves', in splendid isolation from the familiar world of equal sticks, large men and healthy children. It is this 'separation' of Forms from sensible objects that distinguishes the fully-fledged version of the Theory found in the Phaedo. Nowhere is the contrast between Forms and sensible things drawn more sharply than here. The soul-body dualism already postulated (64e— 65a) is one aspect of the same dichotomy. In effect, Plato adapts metaphysical dualism to his present purpose by equating the domain of Forms with the 'next world' of religious faith. Thus the Forms not only afford the ground for believing in the immortality of the soul; they are also its destination.

They are sometimes spoken of in terms suggestive of direct acquaintance or even mystical experience. See, e.g., 79dl—7,84a8— b4. Since it is also suggested that such direct acquaintance is unattain­able during this life (66e2—67b2), it might be supposed that their existence has to be accepted purely on faith. The Theory is, indeed, everywhere assumed rather than proved. Nevertheless, it is not adopted arbitrarily or merely to buttress the case for immortality. Rather, it offers a solution to a number of philosophical difficulties, which will now be considered.

The Theory has its origin in Socratic inquiries into value concepts such as 'just', 'beautiful', 'holy', and 'good'. Plato depicts these in several earlier dialogues, whose prime question is 'What is FT, where 'F' stands for a value term in common use that is felt to require clarification. No such inquiries occur in the Phaedo, but they are familiar to the speakers, and the Theory is explicitly connected with them. Cf.75dl—3 and 78dl—2, where the 'questions and answers' mentioned evidently concern the essential nature of each of the concepts referred to, the attempt to specify what each of them is. This enterprise is envisaged in the Republic as a prospective science called 'dialectic', which would aim, in ethical and meta­physical matters, for the objectivity and rigour achieved in mathe­matics. It would also aim to secure the foundations of the mathe­matical sciences themselves.

The programme requires Forms to be posited as objects of scientific knowledge. They provide the sciences with a subject-matter about which their statements are made, and of which these state­ments hold true. 'The true' is not clearly distinguishable in Greek from 'the real', nor 'that which is the case' from 'that which exists'. See previous note and on 66b 1— c5. The requirement that what is known must 'be true' thus gives rise to the view that the objects of knowledge must be 'realities' or 'existents'. And, since the truths attained in mathematics and dialectic are timeless or eternal, the 'realities' are credited with this feature as well. The Forms thus serve as timeless realities discovered in mathematics and dialectic. Their existence is a necessary condition for there being scientific knowledge at all.

For certain mathematical concepts, such as squareness, cir­cularity, and equality, the Forms are treated as ideal exemplars of properties that are never perfectly embodied in physical objects. No piece of wood is perfectly square, no stone is perfectly circular, no two sticks are exactly equal. The perfect Square or Circle, or

Absolute Equality, may thus be thought of as a standard, to which objects in the sensible world approximate but of which they fall short. The Form serves as a paradigm for such properties. Indeed 'paradigm' derives from a word Plato sometimes uses for Forms in this role. It connotes an individual used as a pattern, sample, or model for a given property, to determine whether or how far other individuals possess it. Just as the length or weight of a thing is determined by appealing, ultimately, to the standard metre or pound, so—according to the Theory—there is an implicit appeal to the Form F when an individual is judged to be F.

This idea is fairly naturally extended to concepts of value. No man or society is perfectly just or good. Yet to say that individuals fall short of perfect justice or goodness is to suggest that these ideals exist independently of any particular man or society, and may be more or less closely approached by them. Moreover, since the ascription of these properties is often uncertain or controversial, the Form is supposed, once again, to provide a paradigm or measure against which individuals can be assessed. In the Republic Plato portrays the perfectly just man and society with this express purpose.

There is, however, a crucial difference between Platonic para­digms and the standards used in weighing or measuring physical objects. The standard metre is itself a sensible item, whereas a Form is not. It is an important feature of the Theory that no sensible item can provide an ultimate standard against which others may be measured. Sensible Fs are inherently defective, and must therefore fail to be 'really' F. Only a non-sensible F can be properly so called. The Form cannot be identified with any of its sensible instances. The various grounds for this claim will be discussed below. See on 74b7— c6.

(c) The Forms posited in the Phaedo are mainly those for the quantitative and value concepts studied in mathematics and dialectic. But the Theory has a further dimension: it offers a general account of predication and naming. Forms are the designata of common adjectives and nouns predicated of individual subjects. Thus at 102b2 (cf.78e2, 102cl0, 103b7) it will be said that things called 'large', 'small' and 'beautiful' are 'named after' the relevant Forms. The prime bearer of the name ^kF' is the Form F. It is the one thing common to each of the many Fs, the single feature F-ness shared by its many instances, the meaning of their common name. In this capacity, it functions as what later came to be called a 'universal'.

In the Republic (596a) Forms are said to be posited for each plurality of things 'to which we apply the same name'. This formula is so broad that it has sometimes been doubted whether Plato would seriously have posited Forms for every item it covers. Forms for artificial objects, for example, seldom appear. Such Forms as Shuttle (Cratylus 389b), Bed, and Table {Republic 596b) are introduced only for special purposes, and may have played no part in the Theory in its original form. Yet, clearly, a far wider range of Forms is demanded by a theory of 'universals' than by the concerns of mathematics and dialectic. This implication is faced in the Parmenides (130c—d), where misgivings are expressed over extending the Theory to such classes as Hair, Mud, and Dirt.

It should be noted that the 'universal' role of Forms is distinct from their scientific and paradigmatic roles. Immense difficulty is incurred if these roles are conflated by treating all Forms alike as postulates of a single, comprehensive 'theory'. Some of the resulting problems will be considered further below. See on 74d4—8 and 102a 10—d4.

(3) The Theory of Forms and the belief in immortality were called by F. M. Cornford the 'twin pillars' of Platonic philosophy. But the metaphor misrepresents the logical relationship between them. For the doctrine of immortality is logically dependent upon the Theory. All of the major arguments except the first are based upon it. This dependence is made explicit at 76e—77a, 92d—e and 100b. The Theory is clearly fundamental.

It is, however, nowhere defended, but is simply accepted without argument by all parties (65d6, 74b 1, 78d8-9, 92d6-e2, 100cl-2, 102al0—bl). Its status is well described in Simmias' words at 85c8— dl: it is 'the best and least refutable of human doctrines'. These words, though they do not, as uttered by Simmias, refer to the Theory, are so closely echoed by Socrates later that they may be taken to express his own position. Cf.99c6—100a8 and see on 84d4— 85dl0. The Theory is 'the theory he judges to be strongest', adopted provisionally and for the sake of argument. It offers the most plausible solution he can find to the problems outlined above.

Socrates will urge that the validation of a first principle should not be mixed up with the testing of its consequences (lOlel—3). Similarly, criticism of arguments based upon the Theory should be distinguished from criticism of the Theory itself. The arguments for immortality must remain inconclusive without a defence of the Theory, as Socrates recognizes at 107b4—6. But the Theory is, in this dialogue, subordinated to the discussion of immortality and cannot be further investigated here.

A lucid exposition of the Theory of Forms is given in G. Ryle's article 'Plato', in the Encyclopaedia of Philosophy, ed. P. Edwards (New York, 1967) vi. 320-4.

66b 1— c5. For the text and translation at 66b4 see note 10. The words translated 'that, we say, is truth' (66b7) mean, more literally, 'we say the true to be that', the referent of 'that' being 'what we desire'. Cf.67bl. 'The true' combines a neuter singular adjective with the definite article, not to designate an individual Form, but to characterize the Form world in general. Many other adjectives are so used. Cf., e.g., 79d2, 80bl-2, 81a5, 83e2, 84a8. 'The A' in such passages does not mean 'the Form A' but 'that (domain) which is A', where 'A' is an attribute of Forms as such—'pure', 'divine', 'immortal' and the like.

The adjective translated 'true' can also mean 'real', as well as 'genuine' or 'truthful'. It has an attributive use, in which 'true T means 'genuine X or 'X properly so called'—e.g. 'true goodness' (69b3). Plato commonly uses 'the truly X" in this sense—e.g. 'genuine philosophers' (64b9), Note also that 'true' or 'real' admits of degrees—cf., e.g., 65el—2, 83c7. See note 8, and on 83b2—3.

'True' has generally been used, and 'truth' for the abstract noun. But 'real' has seemed necessary at 81b4,83b3,83c7, and 83d6. Like the verb 'to be' (see on 65c2—4), 'the true' embraces that which exists as well as that which is the case. Since the Platonic Form world is 'the real', and knowledge of it affords 'the truth', the two senses of the word naturally converge in references to it. 'Truth' in this sense is perhaps meant at 65b9, 66a6, 66d7, 67b 1, and 99e6. Sometimes the meaning seems more general, as at 91cl—2, 102b9, and 115al. At 65b2 'truth' approaches 'veracity', and at 89d6 'true' approaches 'truthful'. See also on 89dl-4 and 90b4-91c5.

The saying referred to at 66c4—5 is unknown. For the verb there translated 'think' cf.96b4. It is cognate with one of the nouns for "wisdom' discussed at 69a6—c3, and is thus linked with apprehension of the Forms.

68a3—b7. The analogy between wisdom and human loved ones demands that the literal meaning of 'philosopher' be brought out at 68b2—3. He is a 'lover of wisdom', contrasted with other kinds of 'lover' at 68cl-2 (cf.82bl0-c8). Philosophy is thought of as the quest for wisdom, not the attainment of it. See 66e2—67a 1, and note 12. Cf. also Symposium 204a—c.

At 68a4—5 there may be a reference to Jocasta or Evadne, whose suicides occur in Euripides' Supplices (985ff.) and Phoenissae (1455ff.). Evadne's was committed for the sake of reunion with her husband.

68b8-69a5. This passage develops a contrast between the gen­uine goodness of the philosopher and the spurious goodness of other men. Non-philosophers withstand evils only for fear of worse alter­natives, or forgo some pleasures only to secure others. Their so-called 'bravery' and 'temperance' are thus no more than an intelligent hedonism.

The hedonistic position is elaborated more fully, perhaps ironically, in the Protagoras (351b—360e), where a calculus of pleasures and pains is expounded in terms similar to those used here. Its role in the Protagoras, and the extent of Plato's commitment to it in that dia­logue, which is probably earlier than the Phaedo, are much disputed. However, the view proclaimed in it by Socrates is essentially similar to the one he is made to disparage here. In other dialogues the thesis that 'good' is equivalent to, or definable in terms of, 'pleasure' is generally rejected. See Gorgias 495e—500a, Republic 505c—d, Philebus 20c-21d, 53c-55c.

Socrates deals here with 'bravery' and 'temperance' only. No parallel argument is given for 'justice', despite its inclusion at 69b 1— c3 and 82all—b3. These qualities, together with wisdom, make up the quartet of 'cardinal virtues', as they later came to be called. They are analysed more fully in the Republic (427e—444e), where the central concept is justice. The Phaedo, concerned as it is with the distinctive nature of philosophers, assigns a special role to wisdom. Bravery and temperance are explored in the Laches and Charmides respectively.

At 68c6 the translation 'people of the disposition we have des­cribed' slightly expands the text, so as to make clear that it refers to philosophers, and not to the non-philosophers who have just been mentioned (cl—3). 'Bravery' could not be meant to apply to the latter, for it must be parallel to 'temperance', which is ascribed at 68cl0 to philosophers only. 'What is named "bravery"' (c5) is best taken, with Burnet, to be paralleled at 68c8 by 'what most people name "temperance"'. In respect of both qualities, a distinction is being drawn between 'so-called' and genuine goodness, which Socrates will proceed to sharpen at 69a6—c3. Cf.82bl—2. Here, as often, 'what they call JP suggests that in ordinary usage the term is misapplied. See on 107cl—d2.

Note that Socrates does not, as might be expected, simply ascribe 'so-called' goodness to the ordinary man, while reserving true good­ness for philosophers. What he suggests is that 'bravery' and 'temper­ance', even as commonly understood, apply especially to philo­sophers, or to them alone. He is questioning the non-philosophers' title not only to 'true bravery' and 'true temperance', but also to 'bravery' and 'temperance' so-called. A weaker claim is made for 'bravery' than for 'temperance': the former is attributed to philo­sophers 'especially' (c6), the latter to them 'only' (clO). In arguing that philosophers deserve the title of 'so-called' bravery and temper­ance, Socrates does not, of course, mean to deny 'true' bravery and temperance to them. See 69b2, cl—2, and Hackforth, 57.

The arguments about (1) bravery and (2) temperance are rather different. In both cases, however, it is claimed that the qualities are 'strange' (d2—3) or 'illogical' (dl2—13, cf.e3): non-philosophers are 'brave through cowardice', 'temperate through intemperance'. They are thus F by reason of G-ness, where F and G are opposites. This is paradoxical, and also—it may be noticed—conflicts with the account given later (100c—e) of the 'reasons' by which a subject can be characterized by one member of a pair of opposites.

(1) For 'bravery' the argument runs as follows (68d5—el): (A) non-philosophers count death a great evil (d5—6); hence, (B) 'brave' ones abide death through being afraid of greater evils (d8—9); hence (C) they are 'brave' through fearing and fear (dl 1—12); yet (D) it is illogical that anyone should be brave through fear and cowardice (dl2—13).

The argument hinges partly upon an assumed equivalence between 'brave' and 'fearless', 'cowardly' and 'fearful'. Philosophers are 'brave' in this popular sense, since, unlike other men, they do not regard death as an evil, and are therefore 'fearless' of it. The noun translated 'being afraid' in (B) differs from that used in (C) and (D) for 'fear'. The latter is cognate with the verb for 'fearing' in (C) and also with the noun translated 'cowardice' in (D). This facilitates the dubious slide in (D) from 'brave through fear' to the full-blown paradox 'brave through cowardice'.

The move from (A) to (B) assumes (i) that men faced with a choice between evils will take the one they regard as lesser; and (ii) that what they regard as evil they also fear. For these assumptions see Protagoras 358d—e. It is worth noting that they could be used to represent any agent who chooses what he considers a lesser evil as doing so from 'fear' of a greater one, and therefore as acting 'from cowardice'. Even philosophers may be motivated by a desire to avoid consequences which they regard as evil, and which they there­fore 'fear' to incur—see, e.g., 83b5-c3. If so, they too might be represented, along the lines of the present argument, as acting 'from cowardice'.

Conversely, it could be shown that non-philosophers may perform cowardly actiorts'frombravery': those who shun death do so through being 'fearless' of the alternatives. They are 'fearful through fearless­ness', hence 'cowardly through bravery'. In this way it might be argued that the 'cowardice' of non-philosophers is no less paradoxical than their 'bravery'. Yet Socrates does not speak of 'so-called cowardice'. Nor is it anywhere suggested that the terms for bad qualities, like those for good ones, might be systematically misapplied.

(2) The word rendered 'temperate' is notoriously difficult to translate. It can mean either (a) 'sensible' or (b) 'self-controlled' in the sphere of bodily pleasures. Its opposite in the latter sense, here translated 'intemperate', means literally 'unchastened', 'Temperance' in sense (b) signifies moderation rather than abstinence, control rather than elimination of the desires. It is only the 'true temperance' of the philosopher that entails a purifying or total purging of the emotions (69c 1).

The argument at 68e2—69a5 runs thus: (A) 'well-ordered' non- philosophers, being afraid of being deprived of certain pleasures, and desiring them (e5—6), (B) abstain from some pleasures, being over­come by others (e6—7); yet (C) 'to be ruled by pleasures' they call 'intemperance' (e7—69al); therefore (D) they overcome some pleasures, being overcome by others (69a2); hence, (E) they achieve temperance, after a fashion, as a result of 'intemperance' (a3—4).

The participles in (A), (B), and (D), 'being afraid' (e5), 'desiring' (e6), 'being overcome' (e7, a2) have all been translated with causal force. This is essential to support the claim that temperance is achieved 'because of intemperance. The agent abstains from certain pleasures only 'because' he desires others, e.g. those of health or saving money.

'Well-ordered' non-philosophers are so called in (A) as a variant for 'temperate'. They are characterized in (B) as 'abstaining from pleasures', and thus in (D) as 'overcoming pleasures'. In (C) 'intemperance' is equated with 'being overcome by pleasures (e7—al). Thus, a man who restrains himself for reasons of health or economy can be represented as temperate 'because of intemperance', since he is 'overcome' by the pleasures of health or saving money (cf.82c5— 8, 83cl—2). Yet the meaning of 'intemperance' has here been artificially stretched, as Socrates recognizes with the phrase 'a kind of intemperance' (e3), and by adding the qualification 'after a fashion' at 69a3. By a parallel extension of 'temperance', it might be argued that a drunkard is 'intemperate because of temperance'. For he is overcome by the pleasures of the bottle, through having overcome those of a healthy liver and bank-balance. This is parallel to the difficulty raised about bravery in (1) above.

There is a parallel problem here also in keeping Socrates' argument from impinging upon the philosopher. He too, it might be objected, is gratifying one kind of desire only through yielding to another, his passionate desire for truth and wisdom (66e, 68a—b). He too, therefore, overcomes some pleasures through being overcome by others. At 114d8—115al Socrates will speak of one who rejects bodily pleasures and devotes himself to those of learning, adorning his soul with temperance, justice, and bravery. It might be asked what distinguishes the virtues of such a man from those of an enlightened hedonist? This raises the question of the role, of wisdom in relation to the other virtues, and of the ethical position of the Phaedo more generally. See next note.

69a6—c3. This eloquent but difficult sentence forms the climax of Socrates' defence of the philosophic life. For the text and grammar see note 13. The key concepts are those of wisdom, goodness, and pleasure. The relations between these call for some comment.

Two words are usually translated 'wisdom': sophia and phronesis. The former ordinarily means 'skill', 'knowledge', 'learning', or (in a pejorative sense) 'cleverness', 'subtlety' (96a7, 101e5). The latter normally means 'practical judgement', 'prudence', or 'good sense'. In Plato's usage, however, the words are not sharply differentiated. Here, and generally in this dialogue, phronesis is a solemn term for the condition of the soul for which the philosopher yearns (66e3, 68a2, 68a7, 68b4), attainable only in communion with the Forms (79dl—7). It has been translated 'wisdom' throughout. See note 14.

The word translated 'goodness' at 69a7, 69b3 and 69b7 (arete) can connote non-human as well as human goodness. Hence the inadequacy of the traditional translation 'virtue', which is mainly confined to good qualities in human beings, and a limited range even of these. 'Virtue' has a convenient plural, however, and is sometimes useful as a generic term. It stands for the 'whole' of which bravery, temperance, justice, and wisdom are elsewhere treated as 'parts'. Each of them is a virtue, and also part of virtue as a whole. See Protagoras 329c—330a and Meno 73e—74b. For the virtues generally see F. E. Sparshott, Monist 1970,40-65.

What is the relation between wisdom and virtue? Are they identi­cal, or is the former a means to the latter? If they are identical, is the whole being equated with one of its parts? Or is one of the parts supposed to be a necessary and sufficient condition for each of the others, and thus for virtue as a whole? If wisdom is a means to virtue, regarded as something distinct from it, why is it represented (a6—10) as the sole object of value? And how can it be 'sold' (b2) without being given up? These questions turn upon the interpretation of the 'money' metaphors at 69a6—b6 and of 'purification' at 69cl—3. See Bluck, 154-6, Hackforth, 191-3, J. V. Luce, C.Q. 1944,60-4, and P. W. Gooch, J.H.P. 1974,153-9.

In treating wisdom as the only 'right coin' (a9—10), Plato assigns to it an intrinsic worth that he denies to pleasure. The latter is supplanted by wisdom as the sole standard of value. In this context 'pleasure' is implicitly restricted to that of the body (cf.83b—d). The distinctions Plato will draw in other dialogues between different kinds of pleasure (see on 60b 1— c7) enable this stark and misleading opposition between wisdom and pleasure to be moderated. At Republic 505b—d rival accounts of 'the good' in terms of wisdom or pleasure alone are both rejected. Cf. Philebus 20e—21e. The oppo­sition between them is evidently artificial, since there can be pleasure in the pursuit of wisdom itself— cf.59a3.

On a sufficiently broad interpretation of 'pleasure', it might be argued that the philosopher pursues it as much as anyone else. He too seeks to maximize his own pleasure, now and hereafter. His practice of virtue in this life, it might also be said, is motivated ultimately by self-interest in the next. See, e.g., 107cl—d5, 114c6— 8. But the spirit of the present passage, and of the dialogue as a whole, is against this interpretation. The philosopher is not merely one whose prudential judgements take account of the afterlife. He seeks wisdom and virtue for their own sakes, and not merely as a means to eternal reward. They are, in some sense, their own reward. See on 81d6-82d8 and 107cl-d2.

3. THE CASE FOR IMMORTALITY (69e6—107bl0)

Cebes' objection that the soul may not survive death leads to the series of arguments that form the core of the dialogue. Each of these will be labelled for convenient reference, and they will be separately analysed in the following notes. But it is important to observe the connections between them. Plato does not offer a set of discrete, self-contained proofs of immortality, but a developing sequence of arguments, objections, and counter-arguments. As the dialogue unfolds, the earlier arguments are criticized, refined, or superseded, until Socrates' belief in immortality is finally vindicated. It is in this process that the intellectual power of the work is largely to be found.

The closely integrated arguments of the Phaedo contrast sharply with the solitary, and quite different, proofs of immortality in the Republic (608c-61 la) and Phaedrus (245c-246a).

3.1 The Cyclical Argument (69e6— 72el)

Socrates begins by appealing to the principle that opposites come from opposites, and positing cyclical processes between every pair of opposites,

69e6—70c3. The popular idea that the soul may be dispersed at death rests on a misunderstanding of its nature. Later Socrates will return to this idea (77d5—e2), and will argue that the soul is not the kind of thing that can be 'blown away by winds' (84b4-7). In the Cyclical Argument he is concerned, rather, with the possibility of its separate existence. The other half of what Cebes requires, 'that it possesses some power and wisdom' (70b3—4), will be supplied by the Recollection Argument (72e—78b). For the relation between these arguments see on 77a6—d5. For the translation 'wisdom' see note 14.

Socrates' denial that he is 'talking idly' (70c 1—2) may be an allusion to Aristophanes' caricature of him in the Clouds. For the gibe cf. Republic 489a, Gorgias 485 d—e. As if in answer to charges of'irrelevance', the close connection between the present inquiry and Socrates' own situation is stressed again and again (76bl0-12, 78al—2,80d7—8, 84c6-85b9,89b, 91a-c, 98c-99a).

The Cyclical Argument has defects that have often enough been pointed out. It is better construed as an opening dialectical move than as an argument to which Plato was seriously committed. But it deserves more credit for ingenuity and subtlety than it is usually given. It serves to introduce some of the concepts that will be of central interest, and gives fertile ground for philosophical argument. See, e.g., J. Wolfe, Dialogue 1966,237-8, J. Wolfe and P. W. Gooch, I.A.C.P. ii. 239-44, 251-4, and C. J. F. Williams,Philosophy 1969, 217-30.

70c4—8. To appraise the argument it is essential to keep in mind the different functions of the verb translated 'come to be' or 'be born'. This verb (gignesthai), which is quite unconnected with Greek words for either 'come' or 'be', may be used both with a complement, meaning (1) 'become', and absolutely, meaning either (2) 'come into being' or (3) 'be born'. In uses (1) and (2) it has generally been rendered 'come to be'. The distinction between these is not, indeed, clear-cut, since a subject's coming to possess a property may also be thought of as that property's coming to exist in the subject. Thus 'x comes to be F may also be expressed as *F comes to be (in x)'. The grammar of sentences containing 'come to be' is not always clear—cf.97a—b, 101c, and see notes 58 and 65. In the present argument, when opposites are said to 'come to be' from each other (70e5—6, 71a9—10), it is meant that things come to be characterized by one member of a pair of opposites from having previously been characterized by the other. See 103b 1—c2. Note also that the Greek verb is liable to be misleading in translation, (i) It may be thought that the things in question are 'generated' from distinct sources, as eggs 'come' from hens, or offspring from their parents. In this sense it would not, of course, be true that the living 'come to be from the dead'. Nor would the claim that 'opposites come to be from oppo- sites' have any plausibility as a general principle at all. But this is not Plato's meaning, (ii) The word 'come' may suggest a local point of departure 'from' which the things in question 'arrive'. The argu­ment can, indeed, be represented as exploiting this idea, Hades being thought of as the 'place' where souls begin their journey. But the verb under discussion carries no connotations of the kind suggested by the English 'come'.

The distinction between uses (2) and (3) is also critical for the argument. As applied to the soul, the verb is not generally used as in (2), since the soul is not, on Socrates' view of it, subject to 'coming into being'. Rather, the verb is used to mean that the soul 'is born'. In this use it is frequently linked with talk of the soul's 'entering' the body (cf., e.g., 77dl—2, 95c5—d2). It means, in fact, 'become incarnate'. Thus, in the next few lines (70c8—d2) Socrates will argue that our souls 'could hardly be born again, if they didn't exist'. The translation 'be born' is clearly demanded here. For it could not be argued that 'our souls could hardly come into being again, if they did not exist'.

Note, however, that 'being born' permits inference to an earlier, discarnate existence, only if it is the soul, as distinct from the living thing, that is said to 'be born'. For it is plausible to hold that, for a living thing, its birth, or (where this is distinct) its conception, is its coming into being. An inference from birth or conception to an earlier existence would then be as unwarranted as an inference from its coming into being to an earlier existence. It is only if 'be born' is predicated of the soul, and taken to mean 'be born into a body' or 'become incarnate' that the required inference to an earlier existence can be drawn.

The actual subject of 'be born' varies. Often Socrates will speak simply of 'our' being born (e.g. 76a9, 76e4, 76e6, 77al, 77al0-bl, 77c2). But at critical points (70c8—dl, 72a6—8) he will speak, rather, of 'our souls' being born. At 73a 1—2 the verb is used explicitly to mean 'become incarnate'. Later (83dl0—el) the soul will be said to 'fall back into another body, and grow in it as if sown there'. It is 'bound' or 'imprisoned' in the body (82e2-83al, 83d 1-5, 84a5, 92al, cf.67dl—10). These metaphors sustain a distinction, vital for the argument, between the soul and the living thing that it animates. Yet this use of 'the soul' as subject of 'be born' is logically suspect. For it insinuates a view of 'birth' in which the soul's discarnate existence is already covertly assumed. And since that is precisely what the argument purports to prove, the very concept of incarnation can be seen to beg the essential question.

This central objection to the Cyclical Argument may be restated. Life and existence, it may reasonably be held, both begin for a living thing at birth or conception. Yet the argument treats the predicate 'alive' as- if it stood for an attribute capable of being acquired by an antecedently existing subject, and 'birth' as if it were something undergone by such a subject, rather than the coming into being of something that did not previously exist. A wedge is forced between 'being born' and 'coming into being' by predicating the former of a supposedly independent subject—'the soul'. Yet whether there is any such subject is just what has to be shown.

Much the same applies to the predicate 'dead'. 'That which is dead', says Socrates, 'comes to be from that which is living' (71dl0— 11, 72a5—6). This is, in a way, undeniable. Yet it fails to prove the posthumous existence of that which is dead. For if 'death' consists precisely in a living thing's ceasing to exist, then when someone passes from being alive to being dead, he will not, in the latter state, enjoy discarnate existence, but will have ceased to exist altogether. 'Socrates is dead' does not, on this view, ascribe a property to a persisting subject, but says merely that someone, who once lived and existed, no longer does so. Here again, a wedge might be driven between 'being dead' and 'ceasing to exist' by treating Socrates' soul as a separate subject, distinct from Socrates himself, and alternating between incarnate and discarnate states. But this would be, once again, to assume what has to be proved. See also on 64c2—9.

70c8—d6. Reference is made at 70dl to 'our souls', i.e. the souls of individual human beings. It is not merely 'soul' in a generic sense whose survival is at issue. Nor, again, is it merely a 'universal' or 'cosmic' soul, into which individuals might somehow be 'absorbed' at death. Throughout the dialogue the speakers are concerned with the fate of their own souls (e.g. 63cl-5,69d7-e2, 88b6-8,95d4- el, 115d2—4), and conclusions are continually drawn in terms of those (e.g. 71e2, 76cll, 107al, 114d8). See also on 64e4-65a3 (p.89-90).

Hackforth points out (59, n.2) that if the first part of the 'ancient doctrine'—'that they do exist in that world entering it from this one'—were included as a premiss of the argument, Socrates would be taking for granted the existence of the soul in the other world after death, which is exactly what he has to prove. He thinks, however, that the argument actually rests only upon the second part of the 'ancient doctrine'—'that they re-enter this world and are born again from the dead'. It is true that after 'if this is so' (c8), only the second part of the doctrine is taken up again in the words 'if living people are born again from those who have died' (c8—9). Perhaps, however, the ancient doctrine should be read not as stating the premisses of an argument, but rather as formulating in toto what still has to be proved, that the soul passes alternately from incarnate to discamate existence.

It is important to note that the proof of our souls' existence in Hades rests upon their being 'born again from the dead¹. 'The dead' are those who 'have died', and who have therefore already lived. Every incarnation is thus a reincarnation. The shift of tense from 'the dead' (70c8, d4) to 'those who have died' (70c5, c9) suggests this—see Loriaux, 122—6. The soul could not be 'born again', it is assumed, unless it persisted in Hades between incarnations. Cf.77d3— 4. Here, once again (see previous note), it is apparent that 'be born' must be taken as predicated of the soul, and not of the living things. For a living thing can be born, in the literal sense, only once.

70d7— e4. The argument is here extended to plants and animals as well as human beings. The scope of'all things subject to coming-to- be' (d9) is not very clear. It could be interpreted broadly to include all things subject to any kind of change. That it is not limited to the genesis of living things becomes apparent when Socrates starts talking (70el—6) about the coming-to-be of opposites. For some of the properties mentioned are not peculiar to organisms or even to physical objects.

The noun here translated 'coming-to-be' (genesis) corresponds to the verb discussed above (see on 70c4—8), and shares its ambiguity. 'Coming-to-be' is a key concept, and will be more fully explored later—see on 95e7—96a5 (p. 170). In the present argument the noun will be used several times from 71al3 onwards for a process of change between each of a pair of opposite states. It has there been translated 'process of coming-to-be' or simply 'process'.

The concept of an 'opposite', though critical for the argument both here and later (102b—106e), is never defined, but is illustrated with the pairs 'beautiful' and 'ugly', 'just' and 'unjust'. These pairs are contrary rather than contradictory opposites. They may not both be truthfully asserted of a given subject at the same time, but they may both be truthfully denied. With such pairs inferences cannot be made from the denial of one member to the affirmation of the other. Cf. Symposium 201e-202b, Sophist 257b. Plato notices at 90a that such properties may lie upon a range. The denial of a predicate at one end of the range does not license the affirmation of a predicate at the other.

How does this affect the argument? 'Living' and 'dead' are not extremes, lying at either end of a range. Nor can things be 'very living' or 'very dead'. None the less, 'living' and 'dead' are contrary opposites. They may both be truthfully denied of a subject. For their logical asymmetry, compare 'married' and 'divorced'. To argue that what is living must earlier have been dead, and therefore still earlier living, is like arguing that someone who is, at a given time, married, must at some earlier time have been divorced, and therefore still earlier married. Yet just as someone not yet married cannot be divorced, so something that is not yet living cannot be dead. See also next note.

70e4—71all. All the adjectives used here are in comparative form. Hackforth (64) thinks that such pairs as larger and smaller 'weaken the argument, since they are not genuine opposites'. Since opposites are never defined, it is not clear what would constitute 'genuine' ones. But the use of comparatives seems actually to strengthen the argument at this point. Socrates could hardly infer from a thing's coming to be weak that it must previously have been strong, for it might have been neither. But he can plausibly argue that if a thing comes to be weaker (than it was before), it must previously have been stronger (than it is now). This inference is valid, provided that both comparatives are filled out in the appropriate way. An inference from, e.g., 'Simmias comes to be weaker than Socrates' to 'Simmias was stronger than he now is' would be invalid. Note also that where the comparative after 'comes to be' is filled out by 'than it was before', the difference between 'x comes to be F-er' and 'x is F-er' becomes unimportant. For the latter implies that x has come to be F-er, and therefore that it was G-er (than it is now). It is, perhaps, assumed that a thing's coming to be large is equivalent to its coming to be either larger than something else or larger than itself at some previous time. For the question whether the distinction between simple and comparative forms is recognized in the Phaedo, see on 100e5-101b8, 102al0-d4.

The next stage of the argument will enable Socrates to dispense with comparatives altogether. But no argument relying entirely upon them could, in fact, support the claim that 'all things come to be in this way, opposite things from opposites' (71a9—10). For some opposites, such as 'odd' and 'even', have no comparative form. And no argument based solely upon comparatives could cater for 'living' and 'dead', since these do not admit of degree.

'That which is weaker' and 'that which is faster' (71a3—4) have been used here for the definite article with a neuter singular adjective in comparative form. Greek uses 'the F' to denote either the property F-ness, or a thing characterized by that property, or the class of things so characterized. The second use is relevant here. A particular thing that comes to be weaker (or faster) must previously have been stronger (or slower). This point is implicit in the use of 'things' (alO), and will be made explicit later (103a-b). See on 103a4-c9.

Socrates' 'law of opposites' may look like an inductive general­ization based on the instances he gives. But it should be recognized that it owes any plausibility it has to the concepts used in its formulation. Bluck (18—19, 57) regards the Cyclical Argument as relying on 'mechanistic' principles, and suggests that Plato offers it only for the benefit of those who find mechanistic arguments impressive. But the present law is not genuinely 'mechanistic', nor is it derived from 'the study of physical phenomena'. It takes no empirical study to discover that what comes to be larger must have been smaller, or that what comes to be worse must have been better. Nor do these truths apply only to physical phenomena.

71al2—b5. A new phase of the argument begins here. Between each of a pair of opposites there is a process of coming-to-be. Every such process requires the postulation of a converse process. For the noun here translated 'process of coming-to-be' see on 70d7—e4 and 95e7—96a5 (p. 170).

This passage seems to introduce a further argument to show that the living are born from the dead. After concluding at 71e2 that our souls exist in Hades, Socrates will insist that there must be a process of coming to life again, and will continue (72a4—6): 'In that way too, then, we're agreed that living people are born from the dead.' This suggests that the appeal to reciprocal processes is meant as a second, independent argument for the conclusion.

In fact, however, the claim that there must be processes in both directions is vital for any effective use of the principle that opposites come from opposites. For the principle that whatever comes to be F must previously have been G requires that there actually be such a process as 'coming to be F\ If, for a given opposite F, there is no such process as 'coming to be F', the argument will break down. If, for example, there is no such process as 'coming to be unripe', it cannot be inferred from a thing's coming to be unripe that it was previously ripe. For this reason the appeal to reciprocal processes is not really a separate argument, but is essential for the working of the Cyclical Argument as a whole. It is not, indeed, clear whether Socrates recognizes it as such. But if he does, this might explain why it is introduced here, and interwoven with the reasoning that ends at 71e2.

Is there, in fact, such a process as 'coming to be alive'? In one sense there clearly is. Things come to be alive when they come into being at birth or conception. But from a thing's coming to be alive in this sense, the proper inference is not that it was previously dead, but that it did not exist previously at all. The sense of 'come to be alive' required for the argument is not that in which a living thing comes into being, but that in which a soul 'becomes incarnate' in a living body. Yet it cannot do this unless it already exists before birth or conception. And whether it does so or not is just what is at issue. See on 70c4-8 (p. 105).

71b6—d4. The examples of complementary processes have a Heracleitean ring about them. Cf., e.g., DK 22 B 88, 126. That 'there are' such processes as cooling and heating is true enough. It does not follow, of course, that they continually recur in each individual thing subject to them. What has cooled down may never warm up again.

The parallel between dying and going to sleep owes its appeal, in part, to a superficial resemblance between sleep and death. Sleep, however, is a temporary state, whereas the permanence of death could not be denied without begging the question. There is, more­over, a basic difficulty in the analogy between waking up and coming to life again. Someone who wakes up exists before waking, whereas someone who comes to life, in the sense of being born or conceived, has not previously existed. A thing cannot be said to 'come to life again' in the sense required by the argument, unless the persistence of an independent subject, 'the soul', is already assumed. Yet this is just what has to be proved. See on 70c4—8 and previous note.

Note also that 'coming to life again' implies a previous life, whereas 'waking up' does not imply a previous waking state. The argument exploits a purely formal resemblance between the two Greek verbs. The prefix common to both of them has a quite different function in either case, shown in the words italicized. See T. M. Robinson, Dialogue 1969,124-5.

What is meant by 'even if in some cases we don't use the names' (71b7—8)? Is it, as most editors assume, that not all processes have names, as do those just mentioned? Or is it that not everything that is in fact a 'process of coming-to-be' would normally be so called? On the first interpretation, the point will be that the available terminology for processes is defective; on the second, that the ordinary use of 'coming-to-be' is unduly narrow. For the relation between names and realities see on 107cl—d2.

71d5—e3. The noun 'death' is ambiguous in Greek as in English. It may refer either to (1) the process of dying, or (2) the event terminating that process, or (3) the state succeeding that event. The verb translated 'be dead' at 71c5 and 71d6 is ambiguous between (2) and (3)—see 64a6 and note 4. Here, however, (3) is clearly the sense required, both for noun and verb. For the opposite of 'death' in senses (1) and (2) would not be 'life' but 'birth' or 'conception'.

'That which is living' and 'that which is dead' have been used at 71dlO—13 for the definite article and neuter singular participle. Note that when Socrates shifts to plural participles at 71dl4—15, he uses both masculine and neuter forms, and thus distinguishes between 'living people' and 'living things'. The masculine form is used by itself at 70c9, 70d3-4, 72al, 72a4-5 and 72d9, and the neuter by itself at 71dl0—13, 72c7-d2 and 77c9. The argument was extended to non-human beings at 70d7—9. It is nowhere suggested, however, that human and non-human beings are animated only by the souls of their own respective kinds. On the contrary, see 81e—82b.

71e4—72al0. The word here translated 'obvious' (e5) may mean either 'observable' or 'certain'. Thus the meaning may be either that people can be observed to die, or that it is certain that they die. The process of 'coming to life again' is, as Socrates seems to recognize, neither observable nor certain. It is not, of course, made any more believable by the fact that dying is both observable and certain. There is, moreover, a peculiar difficulty in understanding what 'coming to life again' means. Socrates is thinking, evidently, not of 'resurrection' or 'rising from the dead', in which bodily identity is maintained, but rather of 'reincarnation', in which the soul is con­ceived as entering a fresh body. Yet the whole idea of the soul's 'entering' a body is perplexing. For it seems to require that the body be thought of as already existing, prior to incarnation, just as a gaol exists prior to a man's imprisonment in it. Yet there is, of course, no separate body existing before conception in the way that a corpse exists after death. The genesis of a living thing cannot, therefore, consist in the animation of an already existing, but previously inanimate, body. Nor is it easy to suppose that the soul, if it is a 'life-principle', enters the body at any point after conception has occurred. For it is natural to associate the start of life with concep­tion itself.

72all—d5. The logical role of this passage is not very clear. It is sometimes treated as merely ancillary to the main argument. Note, however, that it is introduced to justify 'admissions' made earlier (72all—12). These have rested partly upon the postulate that there is such a thing as coming to life again (see on 71al2—b5). Socrates has so far gained assent to this postulate by simply rejecting the possibility of nature's being 'lame' as an obviously unpalatable alternative (71e9). As if the point were not yet firmly established, he has said (71el4—72a2) that if there is such a thing as coming to life again, it will consist in a process of coming-to-be from dead to living people. The present passage may be meant, therefore, to secure the postulate: a process of coming to life again must be assumed, to prevent everything from ending up dead. Taken in this way, the passage supplies a fundamental principle for the whole argument.

But several uncertainties remain.

At 72b5 the translation 'they would cease from coming to be' differs from Hackforth's 'the coming into being of things would be at an end', and from Bluck's 'everything ... would cease from being born'. The supposed consequences of a linear universe may be not merely that things would cease from 'being born' or 'coming into being', but that they would cease from 'coming to be anything', i.e. from acquiring any new character at all. For they would remain permanently frozen in one of a pair of opposite states.

The idea that in a linear universe all things would end up having 'the same form' (72b4) is illustrated with the examples of all things being asleep, and Anaxagoras' 'all things together' (72b7—c5). It is not clear, however, whether these suppositions are meant merely to lead up to the crucial case of life and death (72c5—d3), or whether they are thought of as absurd in themselves. Is it being suggested that a perpetually dormant universe is inconceivable? If it is, Socrates does not say why he finds it so.

The argument at 72dl—3 is too cryptic to be reconstructed with confidence. The translation 'if the living things came to be from the other things' follows Burnet's text. But what are 'the other things'? They cannot be 'the dead' (Bluck: 'the other world'). For the hypothesis that living things come from them seems to be counterfactual, whereas Socrates holds that living things actually do come from the dead. Burnet takes 'the other things' to mean 'things other than the dead', and Hackforth (62, n.5) emends the text to obtain this sense. Alternatively, the hypothesis may be that living things come from 'things other than the living', i.e. from non-living sources. In that case, if living things were to die, without subsequently being reborn (cf.72c7—8), then when the 'other', non-living sources were exhausted, all possible sources would be used up. Everything would then be 'spent in being dead' (72d3). Note, however, that if 'the other things' are conceived as 'sources', from which new living things are generated, there has been a shift from the sense in which opposite things were originally said to 'come to be' from each other. See on 70c4-8 (p. 104).

It is apparently assumed without argument that the possible sources of life are finite and constant. Support for this assumption is sometimes found in Republic 61 la, where it is argued that the num­ber of souls must be 'always the same'. There, however, constancy in the. number of souls does not support, but rests upon, acceptance of their immortality. For it is adduced as a rider to the conclusion that the soul is immortal. The prool of this rider assumes, without further explanation, that everything's 'ending up immortal' is inconceivable.

(5) The present argument makes the complementary assump­tion that everything could not conceivably end up dead. This also is left unexplained; and it is natural for a modern reader to ask why the implication that all life would eventually be exhausted should discredit a view that gives rise to it. For it seems to us conceivable, and even likely, that this will happen. It has to be recognized that Plato's view of the physical world as a whole is organic. The universe is itself a living thing. The present assumption may be connected with the principle that 'nothing can come from nothing or disappear into nothing'. See on 105el0-107al (p.220).

72d6—e2. The three main propositions which the Cyclical Argu­ment purports to have established are resumed here in order of logical priority. The translation follows Burnet and most editors in omitting the final words (72el—2), 'and it is better for the good souls, and worse for the bad ones', as an edifying 'enrichment' of the text. In substance they recall 63c6—7, but they are logically and grammatically out of place here.

3.2 The Recollection Argument (72e3— 78b3)

It is now argued that what we call 'learning' consists in 'recollection', the recovery of knowledge possessed in a former existence. The soul must, therefore, have existed before birth.

72e3—7. The doctrine of Recollection (anamnesis) is one of the best-known of all Platonic themes. It is prominent in the Meno (80d-86c), and is briefly mentioned in the Phaedrus (249e-250c). It is introduced here as a familiar tenet (e4—5), and its significance goes well beyond the present context. Both here and in the Meno the immortality of the soul is inferred from it, but its philosophical interest is largely independent of that conclusion.

Although the doctrine states broadly that so-called 'learning' is recollection (73b5, 75e5, 76a6—7), it does not cover everything to which the term 'learning' would ordinarily be applied. It does not, for example, include the learning of factual information, learning by rote, or the learning of skills. It should be borne in mind that the word translated 'learn' can also mean 'understand'. For it is with coming to understand certain concepts, in particular those that give rise, to the Theory of Forms, that the doctrine of Recollection is concerned.

'Recollection' has been used throughout the translation, and in these notes, for the noun anamnesis, and 'be reminded' has been used for the cognate Greek verb. Note, however, that 'recollection', as ordinarily used, is too broad a term for the context and is used here only for want of a suitable English noun cognate with 'remind'. The key element in the process that Socrates describes at 73c—74a is that on perceiving one thing, a person should think of another. But 'recollection' need not be occasioned by any such experience. 'Be reminded' is closer to the Greek verb both in grammar and sense. In Greek as in English, one is 'reminded of Y by JC, where Y and X are the thing remembered and the reminding item.

The translation and the notes on this section of the dialogue owe numerous points to the valuable article by J. L. Ackrill in Exegesis and Argument, 177—195.

72e7—73a3. The Recollection doctrine is here taken to imply that the soul existed before being born into human form. The reasoning has some force. If 'we', who are now reminded of certain things, are to be identified with 'we' who formerly learned them, and if, as the coming argument will try to show, 'we' learned them before entering human form, then 'we' who learned them cannot be identified with the whole human being, the present composite of body and soul, since before our entry into human form our present bodies did not exist. Hence prenatal learning requires a pre-existing subject—'the soul'. See also on 76cl 1—13.

The Recollection Argument partly resembles the Cyclical Argu­ment. Both turn upon the idea that if, at any given time, we are re-Xed ('re-born', Ve-minded'), we must already have existed before that time. For if we had not thus existed, we could never have been Xed, and therefore could not possibly be re-Xed. But the Recollection Argument escapes an objection raised earlier against the previous argument. 'Being reminded' is a familiar predicate, intelligible with­out special assumptions. But 'being reborn' is not intelligible unless a previously existing soul is presupposed (see on 70c8—d6). The Recollection Argument cannot be convicted in this way of assuming what has to be proved.

At 73a2 the word translated 'immortal' occurs for the first time in the dialogue. For the difficulty of interpreting it consistently with the earlier definition of 'death', see on 64c2—9 (p.87). It may surprise a modern reader that the soul's prenatal existence should be taken as evidence for its immortality. 'How', it has been asked, 'can pre-existence be evidence for post-existence?' (E. J. Furlong, Hermathena 1940, 65). But 'immortality' must cover more than 'post-existence'. Cebes will say later (77cl—2) that only half what is needed has been shown. The Recollection Argument proves only the prenatal 'half of the total immortality thesis. Cf. Meno 81b3—6, 81c5—7, where 'immortality' is understood not merely as 'post-existence', but as persistence through a series of incarnations. See also on 95b5-e6, 105dl3-e9.

73a4—blO. For the translation at 73a7-b2 see note 19. These lines, although not necessarily an allusion to the Meno, clearly refer to the method there followed. Socrates questions a slave boy, previously unversed in geometry, and elicits from him the right solution to a geometrical problem. The present argument for Recollection is offered as an alternative to a proof of that sort (b3—4), and differs from the 'reminding' of the slave boy in several respects, (i) It makes no reference to any mathematical problem, but mentions only judgements about the deficiency of sensible things, (ii) It is concerned with the understanding of concepts, rather than with the proof of propositions, (iii) It does not introduce the Recollection doctrine in the context of a 'what is FV inquiry, or to overcome an apparent obstacle to such an inquiry, (iv) In the Meno no stress is placed upon the use of the senses, whereas Recollection will here be said to be occasioned by their use (74b4—5, 75al 1—b2, e3—4). (v) No mention is made in the Meno of Forms, whereas in the Phaedo they are of central importance.

Nevertheless, both passages are concerned with the mathematical and value notions that give rise to the Theory of Forms (see on 65d4—e5, p.95). Both may be viewed as concerned with 'a priori knowledge'. See R. E. Allen, R.M. 1959-60, 165-74, N. Gulley, C.Q. 1954, 194—213, G. Viastos, Dialogue 1965,143-67.

'Knowledge' and 'correct account' (a9—10) are interrelated, the former involving the ability to provide the latter—see on 76b4—clO. At 73b 1 the word translated 'diagrams' could mean 'proofs'. Cf. Cratylus 436d2, and see Gulley, op.cit. 197. 'Diagrams' would fit the context well enough, however, and the words 'or anything else of that sort' can then be easily explained as referring to solid models (cf. Hackforth, 67, n.l). The words 'that this is so' (b2) are consistent with either the 'diagram' or the 'proof interpretation. They may refer to the fact that people can answer well-framed questions correctly, or to the inference that 'knowledge and a correct account' are present within them. Either of these things could be made clear by introducing them to proofs or to diagrams.

73cl—3. Socrates here begins to formulate conditions for 'being reminded'. Are these conditions intended as necessary or sufficient, or as both? This point is critical for the structure of the Recollection Argument. Necessary conditions would enable inferences to be drawn from the Recollection doctrine. Sufficient conditions would enable the doctrine itself to be inferred. Unfortunately, the conditions seem to vary in this respect. The one given here, that what one is reminded of one.must previously have known, is represented, like the one given at 74a5—8, as necessary. But the requirements that on perceiving one thing one think of another (73c4—dl), and that one should have forgotten the latter (73el-4), seem meant as sufficient ('whenever' 73c4, 73d5, 73e2). The application of these conditions to the case in which we think of the Form Equal on perceiving its sensible instances is therefore problematic.

J. Gosling (Phronesis 1965, 155) has suggested that all of the conditions are meant as both necessary and sufficient. But there is nothing in the text to suggest this, nor could it reasonably be main­tained. Thinking of one thing upon perceiving another is obviously not a necessary condition for being reminded of the former. At 73el it is said to constitute 'a kind of Recollection, i.e. only a single form of it. And Simmias' joking use of the word at 73b7 shows that it is possible to be reminded of things in a very different way, namely by verbal prompting from another person. Cf. also 60c9, 105a5.

73c4—dl 1. Two further conditions for 'being reminded' are now given: (1) on seeing, hearing, or getting some other sense-perception of a thing, one 'not only recognizes that thing, but also thinks of something else' (c6—8); and (2) what is thought of must be the object 'not of the same knowledge but of another' (c8).

(1) The meaning of the first condition depends on the sense of the verb translated 'recognize' at 73c7 and 73d7. This is an ordinary word for 'know', but is distinct from the verb that will mainly be used for 'knowing' the Forms. It can, like the English 'know', mean 'recognize'; and it is plausible to hold that if one is to be reminded of Y on perceiving X, one must, indeed, not merely perceive X, but must also notice features of X by reference to which the association of Y with it in thought can be explained. Cf. J. L. Ackrill, E.A. 182-3.

At 73d5-8 lovers are said to 'recognize' their favourite's lyre, and get in their minds the form of its owner. Does 'recognizing' here consist merely in recognizing the instrument as a lyrel Or do they recognize it as its owner's lyre? If the latter, it may be asked if their thinking of the boy is a separate mental act, distinct from recognizing the lyre. Are they not already thinking of the boy, in recognizing the lyre as his? A similar question arises in the crucial example of Simmias and his picture (73e9—10). In recognizing the picture as one of him, it seems that we should eo ipso necessarily be thinking of Simmias himself. If so, the question would arise whether the thing thpught of and the thing perceived are, in this case, objects of 'another knowledge', i.e. whether the second condition here laid down for 'being reminded' has been met. This bears on the later suggestion that sensible equals 'remind' us of the Form Equal (74c7—d3). For it might be objected that in recognizing sensible equals as such, we are eo ipso thinking of Equality. The Form and its instances would not then be objects of 'another knowledge', and the conditions for our being reminded of the former by the latter would not be met.

(2) Hackforth (67, n.4) rightly rejects Burnet's view that the point of the second condition is merely to exclude cases where one member of a pair of opposites, such as odd and even, or darkness and light, reminds us of the other. Although the idea that opposites are objects of the same knowledge will appear later (97d5), it has no relevance here. But Hackforth's own explanation is also dubious. He suggests that Socrates means to exclude cases in which perceived features of a man might remind us of things we know about him but do not perceive. For the latter features would form part of 'our total knowledge of the man', and would therefore be objects of 'the same knowledge'. But, in that case, how should 'our total knowledge of the man' be defined? Why should it not include his lyre or his cloak, so that these too, contrary to the suggestion of 73d3—4, would be objects of 'the same knowledge'? Moreover, why should the cases Hackforth refers to be excluded as irrelevant for Socrates' purpose? The point of the exclusion remains obscure.

At 73d3—4 it is agreed that 'knowledge of a man is other than that of a lyre'. This might be taken to mean merely that the concept of a man is different from that of a lyre. But such an interpretation would not suit all the examples of 'being reminded' that follow. One may be reminded of Cebes on seeing Simmias (d9). Yet these two, since both are men, would presumably instantiate the same concept, and thus, if concepts were meant, would be objects of'the same knowledge'. If the knowledge of Cebes is different from that of Simmias, what makes it so? Not the mere fact that they are numer­ically distinct, for it would then be otiose to add the condition at all. It would be met automatically in any case where the thing thought of was numerically distinct from the thing perceived.

Ackrill has suggested (op.cit. 184—5) that knowledge of Y is 'other' than knowledge of X if the thought of the former is not already given in the recognition of the latter. Thus, one may see and recognize Simmias, or a picture of Simmias (73e6—7), without eo ipso thinking of Cebes. If this is the meaning, then, as Ackrill says, the condition will be infringed by the case of Simmias and his picture, since one cannot recognize a picture of Simmias without eo ipso thinking of Simmias himself. It is unclear whether the condition, as thus interpreted, would be satisfied in the case of the Form Equality and its sensible instances. See (1) above. In any case, no use appears to be made of it in what follows. Socrates will make no effort to show that the Form Equality is, in fact, the object of a different knowledge from its sensible instances. He will show only that it is non-identical with them (74b7—c6).

73el—74a4. The further examples of Recollection given here are so arranged as to lead up to the case of Simmias and his picture (a9—10), which best illustrates the relation between Forms and sensible particulars. But it is hard to see why so many examples are given, and in particular why 'being reminded by dissimilars' is illustrated at such length, since cases of this sort will play no part in the coming argument about Forms and particulars. Why does Socrates insist that there is Recollection from dissimilar things (74a2—3, cf.74cll—d2, 76a3-4), as well as from similar? Hackforth (68, n.l) suggests that he does so because 'a particular is obviously "like" a Form, and yet may be said to be unlike it because they belong to different orders of existence'. But the case of Simmias and his picture would seem well suited, and sufficient, to make a point of this sort, whereas the case of a lyre and its owner seems quite unsuited to it. For those objects, although unlike each other, both belong to the same order of existence.

As J. L. Ackrill has pointed out (E.A. 188-9), recollection from similars is, in one important respect, not parallel with recollection from dissimilars. One may think of Simmias on seeing his picture because it is like him. But one does not think of a boy on seeing his lyre because it is unlike him. Similarity may be part of the associative mechanism, whereas dissimilarity is not. See on 74c7—d3 and 75d7—76b3.

74a5—8. It is here agreed that whenever one is reminded of a thing by something like it, one must consider whether the reminding object 'is lacking at all, in its similarity, in relation to what one is reminded of. This assertion could hardly be defended as a necessary condition for all cases of being reminded from similars. Only in the special cases in which the reminding item is not merely like, but also a likeness of, what one is reminded of, does it have any plausibility.

But the claim is dubious even in these cases. J. Gosling (Phronesis 1965, 151—61) has argued that it would be implausible to hold as a necessary, or even as a common, condition of 'being reminded' of a person by his picture, that one must assess it as a likeness of him. He therefore suggests that a picture is judged as 'lacking' simply in virtue of its being an image, and thus, precisely because it is an image, failing to possess some features of its original (cf. Cratylus 432b—c). Thus, one who is reminded of Simmias by his picture must judge that it 'falls short', e.g. in virtue of its being two-dimensional.

The point is crucial for the interpretation of the later assertion (74d4—8) that sensible equals appear to us to 'fall short' of the Form. On Gosling's view they do so, not because they are not as equal as the Form, or because they fail to be exactly equal, but because they are mere sensible instances of Equality, and must, as such, differ from the Form itself.

This interpretation is attractive. It frees Socrates from making the unsupported claim that sensible equals, unlike the Form, are never exactly equal. However, (i) is it, in fact, necessary that, in being reminded of someone by his picture, one should consider whether it falls short of him, and always hold that it does so? Must one, if one is to be reminded of somebody by a pencil sketch of him, attend to the fact that it is black-and-white and only a few inches long? Are not such things simply taken for granted and disregarded, rather than consciously noticed? (ii) At 74e2—5 it is asserted that whenever one judges that this X falls short of that Y, one must have known Y previously. But this could hardly be held necessary forjudging that a picture 'falls short' of its original in Gosling's sense. One may readily judge that a picture falls short of its original in this way, whether one has previously known the original or not. Only if the picture has to be compared with its original, as Socrates says we 'compare' sensible particulars with Forms (76e2), is prior knowledge of the original required.

74a9—bl. The Form Equal is introduced here, and will be used as a specimen Form in the coming argument. For the Theory of Forms generally see on 65d4—e5.

The words translated 'we say, don't we, that there is something equal' (74a9—10) might be rendered 'we say, don't we, that equal is something', in conformity with the Greek idiom 'X is something' for 'there is such a thing as X* (see on 64c2—9). However, the contin­uation 'I don't mean a log to a log, or a stone to a stone' follows more naturally if 'equal' is taken predicatively. The translation follows the explanation of these lines given by G. E. L. Owen, A.D. 114-5. See also note 5, and on 100bl-c8.

74b2—3. More literally, 'we know it, what it is'. The pronoun 'it' refers to the Form Equal, and functions both as direct object of 'know' and as subject of the 'what' clause. Cf., e.gRepublic 354c, and see note 21. Verbs of knowing, in Greek as in English, can govern a direct object or an indirect question. Note that 'he knows the man who is on duty' could mean either 'he is acquainted with the man who is on duty' or 'he knows which man is on duty'. These are logically independent statements (neither entails the other), but their meanings converge in the common Greek construc­tion in which a noun is taken both as direct object of 'know' and as subject of the interrogative clause. This construction helps to assimilate 'knowing what F is' to 'knowing the Form F\ The assimilation is natural, given Plato's use of 'what F is' as a standard designation for the Form F (see on 75c7—d6), and given also his tendency to treat the quest for definitions as a search for Forms (see on 65c2—4).

The assertion that 'we know what the Form Equal is', assuming its equivalence to 'we know the Form Equal', appears to be contradicted at 76b4—c3. It is there denied that all men know the Forms, on the ground that they cannot give an account of them, and that the ability to do so is necessary for knowledge (see on 76b4— clO). This contradiction could be avoided in one of two ways, (i) The sense of 'know' might be less strict here than at 76b—c. Here it might be meant merely that we have the concept of equality, 'know what it is' for normal human purposes; whereas the 'knowledge' referred to at 76b—c would require the ability to give a philosophical analysis of it. This, however, would involve (a) an unacknowledged shift in the meaning of 'know', and (b) a shift in the reference of 'we' from 74b 1, where it means 'we philosophers', to 74b2, where it would mean 'people in general'. Alternatively, (ii) 'we' might be taken to mean here not 'people in general', but 'we philosophers' (cf,74a9—bl, 75d2, 76d8, and 78dl). This would limit the scope of the Recollection Argument to the souls of philosophers. But no such limitation is indicated in the Recollection Argument itself. Indeed, 76c4 says that even those who cannot 'give an account' of the Forms are reminded of them. This suggests that Recollection is not a philo­sopher's privilege, but, as in the Meno, is possible for human beings generally. See also next note.

74b4—6. The claim that we thought of the Form Equal, or got the knowledge of it, from sensible things recurs several times (74c7— 10, 75a5—7, 75al l-b3, and cf.75e3-5), and is evidently vital to the argument. Yet its meaning depends upon that of the question at 74b4 'where did we get the knowledge of it?', and this is far from obvious. Is Socrates asking (i) from what source we (people in general) acquired the concept equality, i.e. how we learnt, for ordinary purposes, to recognize things as equal, and to describe them as such? Or is he asking (ii) from what sources we (philosophers) became acquainted with the Form Equal, i.e. how we came to acknowledge its existence and to discover its nature? Is he concerned with ordinary concept formation, or with philosophical clarification? Uncertainty on this point pervades the whole argument.

There is difficulty on either view, (i) The assertion that we acquired the concept of equality 'from' sensible equals, or 'from' sensing them, although perhaps congenial to common sense, would be a jejune answer to a complex empirical question. Clearly, much more would need to be said about how we learnt to compare and measure sensible things in order to judge them equal. The role of counting would have to be considered. And it would need to be shown that we could have acquired the concept in no other way (75a6) than through sensing.

Alternatively, (ii) if Socrates is talking about the clarification of concepts, his insistence that we could acquire knowledge of the Form Equal only from sensing particular equals must seem surprising, in view of his continual disparagement of the senses elsewhere (65a9— 66al0, 79a, 83a—b). They are denounced as nothing but a hindrance in the quest for Forms (see on 65b 1—7). Socrates will later tell how he gave up using them in his own inquiries (99d4—e6). How then could we, in using the senses (75e3), regain the very knowledge to which they have been said to deny us access?

74b7—c6. This passage is of the utmost importance for under­standing the Theory of Forms, since it argues that the Form Equal, and by implication many other Forms (cf,75cl0—d5), must be non- identical with their sensible instances. It should therefore throw light upon the motives from which the Forms are 'separated' from sensible things in the fully-fledged version of the Theory adopted in this dialogue (see on 65d4—e5, p.94). Unfortunately, the argument itself is much vexed, and has become a notorious philosophical crux. See especially J. L. Ackrill, P.R. 1958, 106-8, and K. W. Mills, Phronesis 1957, 128-47,1958,40-58.

The non-identity of the Form Equal with its sensible instances is inferred from the fact that the latter do, but the former does not, possess a certain property. Thus, the argument is of the following form: sensible equals have the property P; but the Form Equal does not have that property; therefore sensible equals and the Form Equal are not the same thing. However, (1) what exactly is PI And (2) why does Socrates use the expression 'the equals themselves' (plural) at 74c 1, unparalleled elsewhere in the dialogue, instead of 'the equal itself, his usual expression for the Form, which reappears in the conclusion at 74c4—5?

(1) The property ascribed to equal stones and logs at 74b8—9 has been interpreted in four different ways. Socrates may be suggesting (a) that equal stones and logs, while remaining the same, seem equal to one observer but not to another; or (b) that they seem equal to one thing but not to another; or (c) that they seem equal in one respect (e.g. length) but not in another (e.g. weight); or (d) that sometimes, while remaining the same, they seem equal at one time, but not at another, (a), (b), and (c) are all based upon Burnet's text, (d) is based upon a variant reading with good MS. authority (see note 22).

Each interpretation has its supporters, and none can be con­clusively disproved. It is, however, doubtful whether the Greek will bear sense (c), and it will not be further considered here, (d) is defended by Verdenius, who glosses the lines thus: 'under certain conditions, e.g. when you are walking around two equal stones, they in turn seem equal and unequal, whereas under other conditions, e.g. when you are sitting in front of them, they do not show such a variation in their mutual relation.' Verdenius argues that the temporal reference at 74c 1 ('did the equals themselves ever seem ...') requires a temporal contrast at 74b8—9. But just such a contrast is provided by the word 'sometimes', so the variant reading is not needed to obtain it. Nor does the fact that Socrates refers at 74cl to only a single observer ('to you') tell in favour of the variant reading, as Verdenius supposes. For if Burnet's text is not interpreted in sense (a), as referring to different observers, the fact that only a single observer is mentioned at 74c 1 would not tell against it at all. (d) is therefore a possible interpretation, but by no means necessary.

The present translation, 'sometimes seem equal to one, but not to another', is deliberately ambiguous between (a) and (b). With (a) the point will be that two different observers could disagree as to whether logs or stones were equal, whereas no such disagreement would be possible with respect to the Form Equal. With (b) the point will be that logs and stones can have contrasted predicates in different relations, whereas the Form Equal cannot.

Both interpretations are grammatically acceptable, but neither is without logical difficulty. For (a) see Mills, op.cit. 129—31. For(b) see N. R. Murphy, I.P.R. Ill, n.l, answered by Hackforth, C.R. 1952, 159. It is, indeed, a common enough Platonic reproach to sensible Fs, that they are only relatively F, and may, when viewed in a different relation, appear not to be F, or to be the opposite of it, G. Cf., e.g., Republic 479-80, 523c-524c. However, (i) since any given sensible equal is always unequal to something, it is somewhat odd to say that this seems to be the case only 'sometimes', (ii) It is very hard to suppose that the Form Equal is itself free from this defect, if this means attributing equality to it (see on 74d4—8, p.128).

And (iii) the names of the Forms mentioned at 75c 11—dl, to which the argument in general is supposed to apply, could not, on inter­pretation (b), be directly substituted for 'equal' at 74b8 without any further changes of wording, whereas on interpretation (a) they easily could. None of these objections to (b) is decisive, however, and the argument, if taken in this sense, can readily be adapted to the other relevant Forms—see G. E. L. Owen, S.P.M. 306. Either (a) or (b) therefore remains a defensible reading of Burnet's text.

Note that all the defects of sensible equals that feature in the above interpretations are implicitly ascribed to instances of the Form Beauty at Symposium 21 lal— 5. The doubt as to which particular defect is here in question should not be allowed to obscure the basic case for 'separating' Forms from sensibles to which they all contri­bute. The essential point, on any interpretation, is that no sensible F can be the true nominatum of *F\ can be 'that which F is', because all sensible Fs admit also of being called ^lG\ Yet 'F' and '(?' are names for two different and opposite things, so anything that admits of both names cannot be 'what F is'. The Form F must therefore be something 'other than', distinct from, its sensible instances. See Mills, op.cit. 145-7, and R. E. Allen, R.M. 1961, 328-9.

(2) Plato's use of the unusual plural, 'the equals themselves', has been variously interpreted. Burnet, Hackforth (69, n.2), Bluck (67, n.3), and W. D. Ross (Plato's Theory of Ideas, 23-5, 60) take it to refer to non-sensible mathematical objects, such as the angles at the base of an isosceles triangle. But equals of that sort would be irrele­vant to the conclusion that sensible equals differ from the Form Equal, which is what the argument purports to prove. Moreover, as Ackrill has pointed out (op.cit. 108), the premiss would on this interpretation be false, since one might, if unfamiliar with the relevant Euclidean theorem, suppose that the angles at the base of an isosceles triangle were not equal, or not suppose that they were.

R. S. Bluck (Phronesis 1959, 5—11) proposes to read back into the present passage the distinction drawn later (102d6—103b5) between the Forms Large or Small and the Large or Small 'in us'. The Forms 'in us', which he calls 'Form-copies', may also, he thinks, be exemplified by 'the equals themselves' of 74c 1. But it seems unlikely that a distinction not drawn till much later, having no obvious relevance to the present context, should suddenly be intro­duced here without explanation. See J. M. Rist, Phronesis 1964, 28-9.

If the premiss at 74c 1—3 is to be both relevant to the argument and true, 'the equals themselves' is best understood as an alternative designation for the Form Equal, used, like 'equality' later in the same line, as a variant for 'the equal itself. Cf. Parmenides 128e— 129b, and see Rist, op.cit. 29-30. The plural designation is some­times attributed to the fact that 'equal' implies at least two terms. It has been further suggested that this Form was conceived by Plato as a set of two (or more) Equals, a bipartite (or multipartite) entity, that could appropriately serve as a standard to which sensible equals approximate. See P. T. Geach, P.R. 1956, 76, Mills, op.cit. 49-50, and G. Vlastos, S.P.M. 287-91. But this explanation of the plural is neither free from difficulty nor necessary. 'Equal' has just been used (74b8) in the plural of logs and stones. As Owen has suggested (A.D. 114—15), the word is quite naturally picked up in its plural form when Socrates switches attention to the Form. In any case, whatever the explanation of the plural may be, there need be no shift of subject in the two halves of Socrates' question at 74cl—2. When he adds the second half, 'or (did) equality (ever seem to you to be) inequality?', he is asking a question about the same entity as in the first half, namely the Form of Equality.

It may still be asked, however, how the predicate terms 'unequal' and 'inequality' in the two halves of the question are to be under­stood. Do they both alike stand for the Form of Inequality? If so, the second half of the question will simply be a doublet of the first, the 'or' between them having the force of 'i.e.' or 'that is to say'. In both halves, Socrates will be scouting the possibility of identifying Equality with Inequality. But this view appears to render the whole question inappropriate to the argument. For the predicate which will then be denied of the Form Equality in the second premiss will differ from that which was affirmed of the logs and stones in the first. They, it was said, can seem to be 'unequal', whereas the Form of Equality, it is now said, cannot seem to be 'inequality'. Yet the predicate in question needs to be the same in both premisses if the argument is to go through.

It might therefore be supposed that the first half of the question is scouting the possibility of the Form of Equality's being unequal, whereas only the second is denying that it could be identified with Inequality. This would make the first half of the question relevant to the argument. But the second half would still need to be explained. For the alleged impossibility of identifying Equality with Inequality would then be left without a clear role in the reasoning. ,

Robin translates the lines as follows: 'Mais quoi? L',Egal en soi s'est-il en quelque cas montre a toiinegal, e'est-a-dire l'Egalite, une inegalite?'. Such a translation would enable 'inequality' to be under­stood as meaning 'art inequality', i.e. an instance of inequality. This would bring the second half of the question into line with the first. In effect, Socrates would be asking, once again, whether the Form of Equality has ever seemed unequal. But 'inequality' is not easily taken in this way. For given that 'equality' has just been used as the name of a Form, 'inequality' is naturally read likewise as standing for a Form, rather than for an instance of one.

What is needed for the argument is a premiss to the effect that the Form of Equality can never (seem to) be unequal. The required premiss is of the form 'the Form F can never (seem to) be G\ Now, as noted earlier (see on 64d4—e5, p.93). Forms are designated with apparent indifference by adjectives and abstract nouns alike. We may therefore suppose that assertions of the form 'the Form F can never (seem to) be G' were not felt to be as distinct from those of the form 'the Form F can never (seem to) be G-ness' as they appear to us to be. Failure to distinguish between (A) 'Equality can never (seem to) be unequal' and (B) 'Equality can never (seem to) be Inequality' would be understandable if the Form Inequality was taken to be the true bearer of the names 'unequal' and 'inequality' alike. It would be all the more intelligible if the predicative function of the verb 'to be' was not yet clearly distinguished from its role as an identity sign (see on 65c2—4). For it might then be thought that (A) and (B) were simply equivalent ways of denying that equality could ever (seem to) be unequal. On this view the second half of the question at 74cl— 2 would appear as a perfectly natural variant for the first. For an explanation along these lines see Mills, op.cit. 48—9.

Note, however, that Inequality, if it is to be construed as a Form, cannot be understood in a manner altogether parallel to Equality, as the latter Form figures in the ensuing argument. For sensible un- equals could hardly be held to 'fall short' of Inequality in the way that sensible equals may be held to fall short of Equality. On a common interpretation of 74d4—e8, it will be suggested that equal logs and stones exemplify Equality imperfectly. But unequal logs and stones would seem to exemplify /^equality as well as could be desired. If so, it would be strange to postulate a paradigm of Perfect Inequality. This illustrates a general difficulty about the scope of the Theory of Forms, which arises if their roles as 'universals' and 'paradigms' are confused. For certain pairs of opposites, notably for value concepts such as 'good' and 'bad', 'just' and 'unjust', it is only one member of the pair that is naturally thought of as an ideal, of which particular things fall short. It may therefore be doubted whether 'paradigmatic' Forms are admissible for certain concepts whose title to 'universal' Forms is as good as any. See on 65d4-e5 (p.97).

74c7-d3. For the interpretation of 74c7-10 see on 74b4-6. Here, as above, Socrates speaks simply of our getting knowledge of the Form from particular sensible equals. He draws no distinction between acts of sensing and objects sensed. But he later uses a variety of expressions (75a7, 75al 1, 75bl, 75b6, 75e3, 76d9) which suggest that it is, strictly, our sensing of things, rather than the things themselves, that occasions our thought of the Form. It is, however, unclear in several of these passages whether the translation 'senses' or 'sense-perceptions' more nearly expresses the meaning. See on 65b 1-7.

At 74cll—13 it is noted that the Form may be either similar or dissimilar to its sensible instances, but that this 'makes no difference': provided that from seeing X we think of Y, recollection has occurred, whether X and Y are similar or dissimilar. The point of insisting on this here is not clear, but there seems no reason to follow Archer-Hind in excising the lines. It has not yet become apparent that being reminded of a Form by its sensible instances is a case of being reminded by similars. And the claim that one may be reminded either by similars or dissimilars will be reiterated at 76a3—4.

Bluck translates 74cll: 'Presumably either because it is like them, or else because it is unlike?' But it is incorrect to take 'being' in this line as a causal participle. It would be inappropriate to suggest that we might be reminded by sensible equals of the Form because it is unlike them (see on 73el—74a4). Bluck glosses the line: 'because they resemble it, or because they are associated with it only in thought' (67, n.4). But 'being associated with it only in thought' is not equivalent to 'being unlike it', nor does it correctly express the relation between unlike objects, one of which reminds us of the other. The boy at 73d5—8 is not associated with his lyre or cloak 'only in thought', but because he has been observed to use them. Any causal explanation of someone's being reminded of him on seeing them would have to mention this fact.

74cl3—d3, like 73c4-dl, represents 'thinking of Y from seeing X as a sufficient condition for 'being reminded of F. But one might, surely, think of Y from seeing X, without being reminded of the former, e.g. when one imagined Y or made it up. The condition mentioned is not, in fact, sufficient for 'being reminded', unless it be stipulated that Y is something previously known. But, with respect to the Form Equal, this is precisely what has to be proved. See J. L. Ackrill, E.A. 186-7.

74d4-8. Socrates now makes the point, to be repeated several times, that we think of sensible equals as 'falling short' of the Form Equal. The Form is here referred to as 'what it is itself. For this phrase see on 75c7-d6 (p.131) and note 24. The phrase translated 'the instances in the logs' contains no word as specific as 'instances'.

More literally: 'the things in the logs'.

Why, and in what sense, are we said to think that sensible equals 'fall short' of the Form, or that they 'strive', unsuccessfully, to be like it (75a2—3, bl-2, b7-8)? The usual interpretation of these remarks is that the Form is never perfectly realized in its sensible instances. Cf. Burnet's notes on 65d4, 74a9, 74e9, 75cll, and see on 65d4—e5 (p.96). The Form F serves as a paradigm, a perfect exemplar or standard F, to which sensible Fs can only approximate, and against which they may be judged: no sensible equal is ever exactly equal; only the Form Equal is so. On this view, 'F' not only names the Form F, but can be predicated of it. Indeed, strictly, it is only of the Form F that it is predicated correctly.

The Theory of Forms, as thus interpreted, gives rise to a number of acute difficulties, (i) It is hard to imagine how the Form Equal could function in the way proposed, or to believe that in practice it ever really does so. To suppose that we might determine whether one log is equal to another by comparing either or both of them with a non-sensible Form, to see if they share its character, seems a travesty of what we actually do. To find out if they are equal, or how nearly equal they are, we measure them not against Equality but against each other, or against some other sensible object, such as a ruler.

The judgement that two logs are not exactly equal is no doubt of a kind that we do sometimes make. We know that our senses or instruments are inaccurate, and we may therefore suppose that closer measurement would always reveal inequalities that had previously escaped us. But in judging that two logs are not exactly equal, we do not take ourselves to be comparing them with a non- sensible Form, and finding that they lack a property which it alone possesses. Indeed, we may make such judgements without supposing that any such entity exists.

It is debatable whether, in order to be able to make such judgements as 'these logs are not exactly equal', we must previously have been acquainted with something which is exactly equal. More generally, it is not clear that judgements to the effect that 'X is not exactly F' require prior acquaintance with an x that is exactly F. Must we have encountered perfection in order to recognize imperfection?

As noted at 65d4—e5, the Form F functions, in some contexts at least, as a 'universal': it is the character F-ness common to all particular F things. But if the Form is construed as a universal, it seems, for many values of F, nonsensical to say that it, and it alone, is perfectly F. Equality cannot itself be equal; no more can smallness be small, or largeness large. Forms, as universals, cannot, in general, have the characteirs that they are. The paradoxes incurred by attributing the character F to the Form F were recognized by Plato and explored in the .later Parmenides (132a-b, 132c—133b). They can easily be generated if the paradigmatic and universal roles of the Forms are confused. Whether Plato himself confused them in the Phaedo and other dialogues of his 'middle period' has been a major crux of recent scholarship. See S.P.M., Chs. 4, 12—14.

(v) If 'equal' is predicated of the Form Equal, even in its purely paradigmatic role, it may be asked whether it is equal to anything, and if so, to what. It could hardly be equal to everything. If it is equal to some things but not to others, it will suffer from the very defect from which Forms are supposedly immune, and which, on one plausible interpretation of 74b8—9 above, was said to distinguish sensible equals from it. If it is equal only to itself, it will fail to be a genuine paradigm for items that are said to 'strive to be like it'. Should we conclude, then, that the Form Equal is not equal to anything, either to itself or to anything else, but is just equal simpliciterl On this view, it would fail, once again, to be a genuine paradigm for the things supposed to approximate to it. And to treat 'equal' as a non-relational attribute seems a patent miscon­struction, deserving to be characterized, as a case of 'Greek mis­treatment of "relative" terms in the attempt to assimilate them to simple adjectives' (G. E. L. Owen, S.P.M. 310).

Does the present passage, in fact, improperly predicate 'equal' of the Form? Undeniably, it implies that the word 'equal' is grammati­cally predicated of it. For it must be taken to mean that sensible equals seem to us to fall short of being equal in the same way as the Form Equal is equal. The italicized words, though not in Burnet's text at 74d6, clearly have to be understood. See note 24 and cf.l00c4-5, 102e5. But in what way 'is' the Form Equal equal? Is it possible to interpret 'the Form Equal is equal' without raising the difficulties mentioned above? A possible solution is to under­stand it not as attributing equality to the Form, but simply as an identity statement: the Form Equal is (identical with) Equal. Sensible equals are therefore not equal in the way that it is. For they are not identical with the Form, but only (to use the termin­ology introduced later) 'participate' in it. They 'fall short' of it, not in failing to be exactly equal, a claim for which the present passage has provided no argument whatever, but rather in that they are non- identical with it, as argued at 74b7—c6. They 'strive to be like it', but they fall short. For they suffer from the defects that character­ize sensible particulars as such.

On this view, the judgement that sensible equals 'fall short' of the Form Equal will not be that of a plain man confronted with logs that he regards as not quite equal. It will be the judgement of a philo­sopher, who has recognized that the Form is distinct from its sensible instances, and who, on sensing the latter, reflects upon how different they are from the former. This interpretation avoids some difficulties in the traditional view. See H. F. Cherniss, S.P.M. 368-74, R. E. Allen, P.R. 1960, 147-64. It is not clear, however, that it will fit all contexts. In particular, it runs into difficulty in the later discussion of Largeness and Smallness. See on 102al0— d4 (p. 194).

74d9—75c6. It is now argued that on perceiving sensible parti­culars, we 'refer' them to the relevant Form (75b7, cf.76d9-e2), and judge that they 'fall short' of it (74d9-e2, 75al-3, 75al 1 -b3). Such judgements are held to require prior knowledge of the Form (74e2— 75a4), and consequently (75a5—b9) knowledge of it before we began to use the senses, which was at birth (75b 10— 11). For the expression 'what equal is' see notes 25, 26, and on 75c7—d6 (p.131).

Is it legitimate to infer, as Socrates does at 75cl—6, that we possessed knowledge of the Forms before birth? At 75b4—9 he argues that we must have gained knowledge of the Form before we began to use our senses. Yet it has so far been shown only that we gained knowledge of the Form before the first occasion on which we referred sensible equals to it (74e9—75a4). Might this occasion not have been some time after we began to use the senses? Cf. J. L. Ackrill (P.R. 1958, 108): 'One could admit that we saw and heard from birth, and that referring what one sees and hears to standards implies prior knowledge of the standards, and one could still deny that we had prenatal knowledge of standards. For we may have done a good deal of infantile seeing and hearing before we began to refer what we saw and heard to any standards (in fact we certainly did).'

Socrates would no doubt ask when and how prior knowledge of the standards was acquired. If referring sensible things to standards requires 'direct acquaintance' with those standards, then it would follow, given his assumptions, that the referring of sensible things to standards implies prenatal acquaintance with the latter. For we can have had no direct acquaintance with them since birth, given (a) that there is no direct acquaintance with them through sense experience (65d9-e5, cf.79al-5), and (b) that it is only through sense experience that we have thought or could think of them at all (75a5—7). But this line of thought is not made explicit. The argument is therefore defective as it stands.

The inference at 75cl—6, that we must have gained knowledge of the Equal before we were born, would suffice to prove the soul's prenatal existence directly. Yet it takes a further page of argument, proving the Recollection doctrine, before prenatal existence is inferred at 76cll—13. What is the role of this further argument? If, as seems to be the case, the Recollection doctrine is premissed upon the conclusion drawn in the present lines, and prenatal existence is then derived from it, the reasoning from 75c7 to 76cl3 will be circular. See on 76cl4—d6.

75c7-d6. Hackforth (71, n. 1) rightly notes that the clause 'if, having got it before birth, we were born in possession of it' (c7—8), like that at 75d7 below, expresses a hypothesis that Socrates does not accept: it will shortly be argued that we do not possess know­ledge of the Forms at birth (cf.76d2-3). But although the protasis of the present sentence will prove false, the apodosis is clearly held to be true. It extends the conclusion that we knew the Form Equal before birth to the whole range of Forms posited in dialectical question and answer. The correctness of this extension cannot be meant to depend upon the truth of the 'if' clause.

Note that the claim that a Form's instances 'fall short' of it, as often interpreted in connection with Equality, would be untenable with respect to some of the Forms mentioned here. The interpre­tation considered at 74d4—8 was that sensible equals are judged only approximately, never exactly, equal. But it could make no sense to suggest that on seeing two logs, one larger than the other, we judge the former to be only approximately, never exactly, larger than the other. It may be answered that there is no question of postulating a Form of Larger, as distinct from that of Large. The language is, as Hackforth says (71, n.2), loose, and the Forms of Large and Small are meant. But even so, do we judge sensible objects to be only approximately large in the way that such objects may be judged only approximately equal? In what sense could anything be, or fail to be, 'exactly large'? The interpretation clearly breaks down in this case. Yet a satisfactory interpretation of'falling short' of the Form should preferably fit as many of the Forms mentioned here as possible.

The phrase at 75d2 translated 'what it is' is a standard Platonic expression for the Forms. It occurs, with variations, at 65dl3—el, 74b2, 74d6, 75b 1-2, 78d4-5, 92d9, and often in other dialogues. The phrase is here recognized as semi-technical—'we set this seal' on the Forms. Often the name of the relevant Form is included, but sometimes, as here and at 92d9, no name is specified. The phrase then consists simply of the neuter singular relative pronoun followed by 'is'. The word 'itself', though sometimes added, seems not (despite Burnet's note on 75d2) to have been an integral part of the phrase. For the departure from Burnet's text at 75d2, see note 28. Cf. also notes 25 and 31.

Where the name of a Form 'F' is included, the phrase may be taken in three different ways:-

that F which is;

that thing which F is;

that thing which is F.

In (a) 'is' will be the 'complete' or 'absolute' use of the verb 'to be'. In (b) and (c) it will be 'incomplete' (see on 65c2-4). In (b) it must function as an identity sign. In (c) it could be taken either (i) as an identity sign, or (ii) as attributing to the relevant Form the character F. If the 'is' is an identity sign, the choice between (b) and (c) (i) will be unimportant. On interpretation (c) (ii), however, the Form will be 'self-predicated' in the manner that gives rise to the difficulties discussed above (see on 74d4—8).

Interpretation (a) can be ruled out in some passages on grammat­ical grounds. See notes 7 and 25. Other passages are less clear, however, and it cannot, perhaps, be assumed that occurrences must be construed uniformly, even within a single dialogue, or with respect to a single Form. K. W. Mills (Phronesis 1957, 146) takes the phrase in sense (b), as correlative to the question 'What is FY, where this is understood to mean 'What is the thing of which "F" is the name?' Similarly, G. F. Else, H.S.C.P. 1936, 43-4. This fits the reference in the present passage, and at 78dl— 5, to dialectical questions and answers, which may indicate the provenance of the phrase. It could well have arisen from asking and answering the question 'What is FY—the Form being designated, in answer to that question, as 'the thing that F is'. Mills's view of the phrase has been adopted throughout the present translation. Occurrences of the phrase without 'F' have been rendered 'what it is'. Alternative renderings for these occurrences, corresponding to (a) and (c) above, would be 'that... which is' and 'that which is...' respectively.

See, further, H. Cherniss, S.P.M. 372, A. R. Lacey, C.Q. 1959, 51, G. Vlastos, R.M. 1972,452—8, and R. Loriaux, E.F.P., reviewed by K. W. Mills, Gnomon 1957, 325-9.

75d7—76b3. Simmias is now presented with two hypothetical statements, the first (ifP, Q) at 75d7—el, and the second (ifR, S) at 75e2—8. He is then offered a choice (76a4—7) between their resp­ective consequents, Q and S. Q will be rejected, for reasons to be given in 76b4—c3. S will then be inferred at 76c4—5. The argument down to 76c5 is therefore of the following form:

IfP, Q (d7—el).

If R, S (e2—8).

So (3) Either Q or S (a4-7).

But (4) Not-Q (cl—3).

So (5) S (c4—5).

In this argument, step (3) follows if, but only if, it is assumed that:

(2*) Either P or R. In terms of the actual argument, this is the contention that at birth, either (P) we did not forget our prenatally gained knowledge of the Forms, or (R) we lost it. In view of the equivalence between 'losing it' and 'forgetting' (dlO-11), P = Not-R, so that (2*) is an appli­cation of the Law of Excluded Middle. Its presupposition, that we did gain knowledge of the Forms before birth, is built into (1) and (2) by the wording at 75d7 and 75e2. Both antecedents, P and R, begin with the words 'having got them'—i.e. having acquired the knowledge prenatally. If we gained knowledge before birth, it is supposed that we must either have retained it at birth or lost it. But Simmias' suggestion at 76cl4-15 will cast doubt upon whether we did get it before birth. His point does not seem to be satisfactorily answered. See below, and on 76c 14—d6.

At 75d7 'on each occasion' must refer to each occasion on which the soul becomes incarnate. At 75e3 the meaning of the phrase translated 'using the senses about the things in question' is far from clear. See on 65b 1—7 and 74c7—d3.

At 76al-4 Socrates recalls points agreed earlier (73c6—dl, 74cl3—d2): one may, on perceiving X, think of Y, which one had forgotten, whetherX and Y are similar or dissimilar. Note, especially, the condition that one should have 'forgotten' Y, repeated from 73el—3. The claim that we 'forgot' the Forms at birth is an import­ant element in the argument summarized in the first paragraph of this note. However, since prenatal knowledge of the Forms is pre­supposed in this claim, such knowledge cannot be inferred from it without begging the question.

Hackforth translates at 76a3—4: 'something with which the first object was connected, whether by resemblance or contrast'. The suggestion that dissimilarity is itself the source of the association is no more implied in the Greek here than at 74cl 1. See on 74c7—d3. The nature of the association is not clearly specified in the verb translated 'was related', but the idea is, perhaps, that dissimilar items have frequently been found together.

76b4-cl0. The ability to 'give an account (logos)' is said at 76b5—7 to be a necessary condition for knowledge. 'Correct account' was also linked with knowledge at 73a9—10.

What is it to 'give an account'? The phrase can mean 'give a defin­ition', and this would be a natural way of taking it here, where the things about which an 'account' has to be given are Forms. On this interpretation, no one has knowledge of the Form F unless he can give a definition of F, say what F is. See also 78dl—5, and cf. Republic 534b. However, 'give an account' can also mean 'give proof' (cf.95d7), and it may be the having of grounds for a pro­position, or the ability to prove it, that Socrates has in mind as requisite for knowledge. Reasoning of this sort is said to distinguish knowledge from correct belief at Meno 98a. On this interpretation, 'giving an account' about the Forms will mean defending proposi­tions about them by rational argument. 'Giving an account' in this sense, and giving one in the definitional sense, are, of course, related, the former being required for the latter.

Inability to give an account is agreed at 76b8—c3 to preclude most men from knowing the Forms. For the conflict with 74b2 see on 74b2—3. Hackforth (76) would reconcile the present lines with 74b by saying that it is moral Forms, as distinct from mathematical ones, that Socrates here has in mind. But there is no hint in the text of any such restriction. On the contrary, 'moral' and 'mathematical' Forms are expressly said to be on a par (75cl0—d2).

76cll-13. The inference is drawn here that our souls existed before entering human form 'apart from bodies'. The most natural interpretation of this last phrase is 'apart from bodies altogether'—i.e. in a discarnate state, cf.ll4c3. It might be objected that on this interpretation the inference does not, strictly, follow. For our present incarnation might have ensued immediately after a previous death, without any discarnate intermission. The most that could then be claimed would be that our souls existed apart from our present bodies, incarnate in some previous one. This seems a less likely inter­pretation of 'apart from bodies'. But even if it were correct, the required point could still be made. 'We' are not to be identified with our present bodies, since we already existed before they did. Nor can 'we' be identified with our previous bodies, since we still exist when they no longer do. Hence, 'we' who once knew the Forms, and are now reminded of them, must be identified with a persisting non-bodily subject. See on 72e7—73a3.

There can be no doubt, however, that the argument does envisage us as having known the Forms in a discarnate state. As W. D. Ross pointed out (P. T.I. 25), the Recollection doctrine would not explain how we can now come to know Forms on the suggestion of sensible things, merely by referring back to a previous knowing of them in the same way. For this previous knowledge would require explanation no less than does our knowing them on the suggestion of sensible things now. The Recollection doctrine therefore points to a direct and immediate knowledge of the Forms. But since we have no such knowledge of them while the soul is incarnate (66d7-67b2), the Recollection doctrine, if it is to explain what it is introduced to explain, requires that the soul must at some time have existed in a discarnate state.

76cl4—d6, Simmias tries to resist the conclusion that our souls existed before birth, by suggesting that we might acquire the relevant bits of knowledge at the time of birth. He is silenced by being asked at what other time we could lose them.

But he gives in too easily. For why should it be inferred from our lacking knowledge when we are born, that we must 'lose' it at that time? An analogy may be useful here. We should not think of a child born blind as losing his sight at birth, since he never had it. No one would argue from his lack of sight at birth that he must have lost it at that time, and therefore possessed it prenatally. Someone born without sight has never possessed it, and therefore could not be said to have 'lost' it at all. Nor, if he should later acquire sight, could he be said to 'regain' what he had never had.

If, then, we are born lacking knowledge, the inference that we 'lose' it at birth, and subsequently 'regain' it, is justified only if it is already assumed that we once possessed it. Why should this be assumed here? Only, it would seem, because, the idea that we acquired it before birth is contained in the disjunctive hypothetical at 75d7—11 and 75e2—7. Both disjuncts assume prenatally acquired knowledge. See on 75 d7—76b3. But this assumption is precisely what Simmias is here questioning.

It is true that Socrates has already, at 75c4—5, obtained the admission that we gained knowledge of the Form Equal before birth. But once this is granted, he already has all he needs to prove pre- existence, and the ensuing argument for the Recollection doctrine will be otiose. Alternatively, if the Recollection doctrine is meant to be a ground for holding the thesis of prenatal knowledge, the reasoning seems circular. For, as argued above, the thesis of prenatal knowledge is used at 75d7-76c5 in establishing the doctrine. Simmias is here challenging the thesis. But he is rebutted only by appealing to the very assumption his objection calls in question, that we did in fact gain knowledge before birth. It should therefore be asked whether prenatal knowledge and existence are logical conse­quences of the Recollection doctrine, or whether it is a consequence of them. It cannot function both as premiss and conclusion of the same argument.

76d7—77a5. Socrates refers to the Forms as 'what was formerly ours' (el—2), meaning that they were objects of our knowledge

before birth. Cf.92d8-9 and note 50.

At 76e2—7 the existence of the Forms and the pre-existence of the soul are said to be equally necessary. Socrates then adds: 'and if the former don't exist, then neither did the latter'. Simmias' reply shows that he too thinks of the Forms' and the soul's existence as logically interdependent. But the nature of the relation is not entirely clear. The passage might be taken to authorize a bi­conditional: 'If the Forms exist, our souls existed before birth; and if they don't exist, our souls didn't exist before birth.' But Socrates can hardly mean that the pre-existence of the soul and the existence of the Forms are materially equivalent, so that either can, without further premisses, be inferred from the other. He means, more likely, that the existence of Forms necessitates the pre-existence of the soul in virtue of the Recollection Argument; and that their existence is also necessary for the success of that argument. 'If they don't exist, this argument will have gone for nothing' (76e4—5). The words 'if the former don't exist, then neither did the latter' are perhaps only a somewhat inaccurate restatement of the latter point. The existence of the Forms is, indeed, essential to the Recollection Argument, and is the ruling hypothesis of the dialogue. Cf.92d, and see on 65d4-e5 (p.97).

At 77a3—5 Simmias says that the Forms are 'in the fullest possible way'. Degrees of 'being' may sound paradoxical to a modern reader; but the use of 'are' is 'complete' here, and seems undeniably existential, in view of the connection with the pre-existence of the soul. See on 65c2-4 and 83b2-3.

77a6—d5. Simmias and Cebes object (bl—c5) that only half what is needed has so far been shown—the soul's pre-existence. Socrates replies (c6—d5) that its post-existence has been proved as well, if the Recollection and Cyclical Arguments are 'combined'. But how exactly is this to be done?

Archer-Hind (45, n.3) takes the arguments as two halves of a single proof, one showing the pre-existence and the other the post- existence of the soul. But what Socrates here claims to have been proved by conjoining the two arguments is simply 'post-existence'. For 'that's been proved' (c6) must refer back to the claim that our soul 'will exist after we've died' (c3—4). It is not, therefore, the whole of Socrates' thesis for which the two arguments are to be combined, but only the post-mortem half of it. But if so, how? As E. J. Furlong has argued (Hermathena 1940, 62—72), the Cyclical Argument, whatever its defects, seems equally relevant to pre-existence and to post-existence. It would 'prove' either without combination with the Recollection Argument. On the other hand,

the Recollection Argument has no bearing upon post-existence at all.

It is usually supposed that whereas the Cyclical Argument is meant to prove the bare existence of the discarnate soul, the Recollection Argument meets Cebes' further demand to be shown that it possesses 'some power and wisdom' (70b3—4). This latter inference has indeed been drawn, with respect to pre-existence, at 76cl2—13. But its bearing upon the present problem is doubtful. It is not mentioned here, nor would it be to the point. The objection Socrates faces is not that our souls have not been proved to exist after death possessed of power and wisdom, but simply that they have not yet been proved to exist after death at all. He would merely evade this objection by replying that they possessed power and wisdom before birth.

At 77c9—d5 Socrates explains the combination of the arguments as follows: the soul 'must also exist after it has died, given that it has to be born again' (d3—4). This is inferred from a conjunction of (1) 'the soul does have previous existence' (c9—dl), and (2) 'when it enters upon living and being born, it must come from no other source than death and from being dead' (dl—3). (1) and (2) appear to invoke the Recollection and Cyclical Arguments respectively. Yet, as noted above, (1) is derivable from the Cyclical Argument alone. So if Socrates thinks that the Recollection Argument is needed at this point, he is in error.

Hackforth (80, cf. C.R. 1925, 12) interprets Socrates as wanting to show that 'the soul as conceived in the Recollection Argument, the soul which apprehends the Forms, exists after death as well as before birth'. Hackforth claims that the principle 'all that is living comes from that which is dead' (c9), derived from the Cyclical Argument, justifies us in regarding the conclusion of the Recollection Argument—'the soul does have previous existence' (c9—dl)—as involving the corollary that the soul must also exist after it has died. For, he suggests, 'the time before birth is the time after death'.

If this is the reasoning, it supports the required conclusion by a mere sophism. For to say that 'the time before birth is the time after death' would be true only in the sense that the time before the birth of one living thing is the time after the death of an earlier one. The argument would show that a soul animating any given living thing must have survived the death of some previous one. It would not show that it will survive the death of the living thing in which it is now incarnate. So the argument gives the speakers no assurance that their own souls will survive death. Essentially the same objection will be raised by Cebes later (88a—b, cf.95c4—el).

Note that at 77d4 Socrates speaks of the soul as 'dying'. This, although formally inconsistent with its 'immortality', must mean merely that it becomes separated from the body. See on 64c2—9 (p.86) and 84a2-b8.

3.3 The Affinity Argument ( 78b4-84b8)

Socrates next argues that the soul must be immortal in virtue of its affinity to the unchanging and eternal Forms. He then speculates upon the after-life of purified and unpurified souls, and renews his advocacy of the philosophic life.

78b4—10. 'What kind of thing is not liable to it?' (78b7): the translation follows Burnet in supplying 'not', which is omitted in the MSS. at this point. The passage sets out concisely the programme that will be followed in the coming argument. Socrates asks (1) which class of thing is liable to destruction and which is not, and (2) to which of these classes soul belongs. (1) will be considered at 78cl— 79all, and (2) at 79bl—80a9. The conclusions will be drawn at 80al0—81a3.

78cl—9. Socrates here links (1) incompositeness with indestruc­tibility (cl—4), and (2) constancy with incompositeness (c6-8).

The supposition that incomposite things are indestructible has a strong intuitive appeal. For the destruction of a material thing seems to require the separation of its component parts. How an incomposite material thing could be destroyed it is not easy to conceive. So it seems that if the soul were material and incomposite, it could not be dispersed 'like smoke'. But according to the present argument, the soul is essentially immaterial. And it is doubtful whether the principle that incompositeness entails indestructibility applies to immaterial things. Simmias will later (85e—86a) give an example of something immaterial—the attunement of a lyre—that is clearly destructible, but (although composite) is not destroyed through the separation of its own components. Perhaps, then, immaterial things, even if incomposite, might nevertheless be des­troyed. In that case, the soul's incompositeness would not show its indestructibility.

At 78c3—4 it is suggested that 'if there be anything incomposite, it alone is liable, if anything is, to escape this (sc. destruction)'. The first 'if clause need not be taken to imply doubt as to the existence of something incomposite. Nor need the second 'if clause be taken to cast doubt upon there being anything indestructible. Cf.l06d2—4.

There seems no difference in meaning between the phrases translated 'constant' and 'unvarying', or their respective opposites. At 78e5 Cebes uses one of these phrases to answer a question framed in terms of the other. Both mean 'being in the same state', 'admitting of no kind of alteration' (cf.78d6—7). For the kinds of alteration envisaged, see on 78dlO- e5.

At 78c6—7 it is not clear which of the phrases 'the things that are constant and unvarying' and 'the incomposite' is to be taken as subject and which as complement. Is it being said that unvarying things are likely to be incomposite, or that incomposite things are likely to be unvarying? Or are the classes of unvarying and incom­posite things thought of as coextensive? On the translation adopted, Socrates is saying that unvarying things are likely to be incomposite. Logically, this makes better sense, since he is arguing that the Forms, being unvarying, are likely to be incomposite, and he therefore needs the premiss that unvarying things are likely to be incomposite rather than the converse. Hackforth's translation makes him say that incom­posite things are likely to be unvarying. But this, since its converse would not hold, unless the classes were assumed to be coextensive, would not further the argument.

Socrates claims only that unvarying things are 'likely' to be the incomposite. It therefore could not be certain that members of any sub-class of unvarying things, such as Forms or souls, are incomposite. Nor is the basis for the alleged 'likelihood' explained. Perhaps change is thought of as depending upon rearrangement of parts within a complex whole. The characterization of the Forms as incomposite is, indeed, often taken to imply that they are without parts, and therefore incapable of such change. 'Incomposite' may, however, mean merely 'not having been put together', and might thus be ascribed to something that has possessed parts from all eternity. See K. W. Mills, Phronesis 1958, 45—7, answered by R. S. Bluck, Phronesis 1959, 5, and J. M. Rist, Phronesis 1964, 32-3. See also on 80al0-cl.

The question whether Forms can have parts, and if so in what sense, raises the question whether the soul, in virtue of its kinship with the incomposite Forms, is supposed to be without parts, and if so, whether this account of it can be reconciled with the teaching of the Republic and the Phaedrus. See R. W. Hall, Phronesis 1963, 63-82.

For many different ways of understanding 'composite' see Wittgenstein, Philosophical Investigations, 1,47.

78dl-5. The grammar and sense of the words translated 'the Being itself, whose being we give an account of' (dl—2) are uncertain. 'The Being itself' clearly refers to the domain of Forms (see note 7). But 'giving an account' could mean either 'giving a definition' or 'giving a proof; and the use of 'being' may be either 'incomplete' or 'complete' (see on 65c2—4). 'Giving an account' of a Form's 'being' may therefore mean either 'defining its essential nature' or 'proving that it exists'. Loriaux (152—3, 164—5) defends the existential interpretation. See also E.F.P. 26—34. Similarly Hackforth. For the view adopted here see Bluck and Burnet. The non-existential reading has been preferred, in view of the reference to 'asking and answering questions'—cf.75d2—3. It seems far more natural to associate this with the Socratic quest for definitions than with proofs of the Forms' existence. To 'give an account of the being of F' is to answer the question 'what is F?'. Questioning and answering of this sort are familiar in Plato's writings, and typical of dialectical inquiry (cf., e.gRepublic 538d6—e3), whereas question­ing and answering to prove that the Forms exist can hardly be said to occur in the dialogues at all.

The phrase 'what... is' has been translated accordingly at 78d4 and 78d5. See note 31.

78dl0-e5. For the text at 78dl0-el see note 33. At 78e2 Socrates speaks of 'all things that bear the same name as those objects'. 'Those objects' are, of course, the Forms, here treated as the prime bearers of their names. Particulars are 'named after' them. Cf.l02b2, 102cl0, 103b7, and see on 65d4-e5 (p.96), 102al0-d4.

In contrasting sensible particulars with Forms, two kinds of variation are mentioned (e3): particulars may vary in relation either 'to themselves' or 'to one another'. Hackforth (82, n.l) rightly rejects Burnet's view that the latter phrase refers to things presenting different appearances to different people. But there seems no need to suppose, with Hackforth, that the words 'or to one another' may be added because equality requires two terms. The point is simply that particulars may become more or less F at t₂ than they were at t_{ (they vary 'in relation to themselves'), or they may become more or less F than some other particular thing (they vary 'in relation to one another').

The stress here is upon the mutability of sensible particulars, their liability to be characterized by different attributes at different times, rather than (as at 102b—d) their being characterized by opposite attributes at the same time in relation to different things. They are also, of course, mutable in that they are ephemeral: particular Fs come into being and pass away, whereas the Form F is conceived as eternal, 'that which is always existent' (79d2).

79al— bl7. 'Two kinds of beings' (a6): more literally, 'two kinds of the things that are'. See on 65c2—4 and note 34. For the inaccess­ibility of the Forms to the senses (al—4) cf.65d9—e4.

At 79a9—10 the 'invisible' is said to be always constant, and the 'seen' never constant. This is best taken as referring to the Forms and the sensible world as such, rather than as asserting, quite generally, that whatever is unseen is constant, and whatever is seen is inconstant. In effect, it recapitulates the premisses of 78dl—8 and 78dl0—e5, enabling the Form world to be designated in what follows either as 'the invisible' (79b 16) or 'the unvarying' (79e4).

The assumption that we are 'part body and part soul' (bl—2), upon which the argument of the whole dialogue rests, is simply posited and accepted without demur. For the concept of 'soul' see on 64e4—65a3. The inference that it is 'more similar' to the invisible than body (bl6—17) rests simply on the claim that it is invisible (al4).The 'seen' is defined (b9—10) with reference to human nature, perhaps to forestall the objection that the soul is not invisible to the gods.

'Similarity' is not defined. But if 'being more similar' means 'having more features in common', the fact that the soul shares with the Forms a given feature that the body lacks would not show that it is 'more similar' to them than is the body. Even if this were shown, it would not follow that the soul has all features in common with the Forms that the body lacks. Taken as an analogy, the argument is weak. It would be stronger if the feature common to the soul and the Forms were to entail the other relevant features, i.e. if invisibility were to entail invariability, invariability incomposite­ness, and incompositeness indestructibility. A chain connecting these concepts can, indeed, be derived from 78cl—9 and 79a9—10, but its links are weak. See on 78cl-9.

79c2-e7. A tacit assumption underlying the argument here is the Empedoclean doctrine that 'like knows like' (see DK 31 A 86, B 109). The soul's knowledge of the unvarying Forms shows its likeness to them; its confusion when it is in contact with varying sensibles shows that it is unlike them. Hence soul is 'more similar to what is unvarying than to what is not' (e3—5). Note that this conclusion is of a different form from the earlier 'soul is more similar than body to the invisible' (79b 16). That conclusion was of the form 'A is more similar than B to C\ This one is of the form 'A is more similar to C than to D\

Socrates claims only that the soul is 'more similar to the unvary­ing', not that it is unvarying. He could hardly say this after just saying that in using the senses it is 'dizzy, as if drunk' (c7—8). This recalls his earlier criticism of the senses (65a—b). But he does not explain their defects any more clearly here than before—see on 65b 1-7. Why should use of the senses make the soul 'dizzy'? In dizziness objects are experienced as changing when they are really at rest. Yet if the sensible world is, in fact, changing, the senses would not be deceiving us by representing it as doing so. Cf. Cratylus 411 b—c.

There is, perhaps, a further reason for Socrates' reluctance to say that the soul is unvarying. That the soul is, in fact, subject to change seems an inescapable consequence of (1) its liability to incarnation, and of (2) its role as a 'life-principle'. For (1) apparently requires that such properties as 'being in Socrates' body' should be truly predicable of the soul at one time but not at another. And (2) if, as a life-bearing agent, the soul is itself characterized by the property it imparts, and if life entails change, then soul must be subject to change. For the resulting tension with the present view of it as akin to the changeless Forms, see T. M. Robinson, P.P. 30.

For 'wisdom' (79d6—7) see on 69a6—c3.

79e8—80a9. The soul's kinship with the divine is derived from its being 'naturally adapted' (a4—5) to rule the body, i.e. to discipline it and resist its desires (cf.94b—e). This may seem odd, in face of Socrates' continual stress upon the soul's bondage to the body (e.g. 66b-d, 83c—e). If 'rulership' indicates divinity, and 'being ruled' mortality, why should the soul's servitude to the body not suggest that it is mortal and the body divine? No doubt the con­ception of 'nature', here as elsewhere in Plato, is normative. What 'nature ordains'(80a 1) is what ought to happen, not what usually does. The soul's 'natural' fitness to rule the body does not mean that it al­ways does so, just as in the Republic (430e—43 la) the 'natural' super­iority of reason does not mean that it is actually in control. Even so, the notion of the soul's 'rulership' sits awkwardly with the theme of its imprisonment within the body. See on 64e4—65a3 (p.90).

'Rule' is more naturally attributed to gods than to Forms. But the Form world is virtually identified with the gods—see on 81a4—11. For human service to the gods, cf.62b-63a and see on 58a6-c5.

80al0—cl. The contrasts that have been drawn between Forms and the sensible world are summarized at 80b 1—5. Forms are called 'intelligible' (bl) versus 'non-intelligible' (b4). The word translated 'non-intelligible' can also mean 'unintelligent', and there may be a hint of this sense here: body is 'stupid'. For 'body' without the article see next note.

At 80b2 the Forms are said to be 'uniform', by contrast with the material realm, which is called 'multiform' (b4)—cf.78d5, 83e2. Hackforth (81, n.2) understands 'uniformity' to mean 'the denial of internal difference or distinction .of unlike parts'. If the Forms' incompositeness is taken to imply that they have no parts (see on 78c 1—9), then no distinction of parts of any kind, whether like or unlike, could be allowed them. However, the notion of parts may be irrelevant here. For 'uniform' may be explained as meaning 'of just one character'. See K. W. Mills, Phronesis 1958, 45—6 and W. F. Hicken, S.P.M. 191.

At 80b3—5 soul and body are said to be 'most similar' (super­lative) to their respective domains, and not merely, as at 79al6 and 79e4, 'more similar'. 'Most similar' could be intensive—'very similar', or could mean 'more similar than anything else'. These interpretat­ions of the superlative are logically distinct. Neither entails the other, and they call for different kinds of support. On neither interpretation would it follow that the soul shares all the features of the Forms, or that it shares any particular feature with them, or that it shares every feature that the body lacks. These inferences would be stronger on the intensive interpretation. But the claim that the soul is 'very similar' to the Forms is hardly warranted by the fore­going argument, which has shown merely that it has a few properties in common with them.

At 80b9—10 Socrates concludes that soul must be completely indissoluble 'or something close to it'. It is not clear whether this last phrase is meant to qualify the soul's indissolubility, or to signal reservations about the argument. Does Socrates mean that the soul may be nearly (but not quite) indissoluble, or that the argument nearly (but not quite) proves the soul indissoluble? Or are these two things confused? No suggestion that the soul might be 'nearly indissoluble' appears elsewhere, nor would it be consistent with the final conclusion that soul is 'immortal and imperishable' (106e9- 107al). On the other hand, 'nearly indissoluble' seems the more natural way of reading the present text.

80c2—d7. The meaning and text at 80c2—9 have been much disputed—see note 35. But the main argument is clear: even the body lasts for a considerable time after death, so the soul, being of superior nature to the body, may be expected to last longer still. The best comment on this argument is supplied by Simmias' and Cebes' parodies of it at 86a6—b5 and 87b6—c5.

It is somewhat odd that 'the body' should here be said to last for 'a fairly long time' (c5—6) after death, when only a few lines above (80b8) 'body' was said to be 'quickly dissolved' (cf.87e5). 'Body', which occurs at 80b8 without the article, as often (79b 16, 80a 1, 80b5, 81e2, 91d3, 91d7, 105c9), might be used generically ('matter'), or might perhaps mean 'a body' (i.e. 'a material thing'). However, 'the body' normally means 'the human body' (cf.79c2—6), and is explicitly equated with the corpse at 80c2—4. If its meaning varied in the course of the argument, there would be serious equivocation. For what is true of 'body' in the generic sense need not be true of individual 'bodies'. The same goes for 'soul'—see on 64e4—65a3 (p.89-90).

At 80d5—7 (cf.81a4, 81cl 1) there is a play on the words 'Hades' (haides) and 'invisible' (aides). The Form world is 'Hades in the true sense', i.e. 'unseen' but an object of thought. For Hades as the place of departed souls, cf.58e5,68bl, 70c4, 71e2,83d9, 107al. The place is named after the god whose realm it is. This suits the identification of Forms and gods as the soul's destination—see on 81a4—11, The derivation of 'Hades' from 'unseen' is rejected at Cratylus 404b, but may be sound—see L.S.J. s.v. q8rj<; and Bluck, 197.

80d8-81a3. 'Blown apart' (80dl0) is a reference to the material­ist view of the soul under attack. Cf.70a2-6, 77d7-e2, 84b4-8. The cumbersome sentence beginning 'Far from it' (80el) has been broken up in the translation, which assumes a comma after 'shunned it' (e4). The sentence breaks at 80e5, and the sense is not completed until Socrates' next speech, the 'suppose' clause beginning at 80e2 being picked up by 'if it is in that state' at 81a4.

81a4—11. 'The divine and immortal and wise': Hackforth (88, n.4) says that these epithets belong properly to the gods or God (cf.80d7) rather than to the place of the departed soul. But in this section of the dialogue, God (or gods) and the Forms are spoken of interchangeably as the soul's destination. They are, in effect, identi­fied, and divine attributes are applied to both alike. See 79d2, 83e2, 84a8, and on 80c2-d7.

81c4—d5. Unpurified souls are here portrayed not as immaterial substances, but as phantoms or insubstantial wraiths. They are described in terms that could not literally apply to the soul in its essential, incorporeal nature. How could an incorporeal thing be 'interspersed with a corporeal element' (c4—5), be 'weighed down' (clO), or 'fall back into another body, and grow in it' (83dl0—el)? Such language, taken literally, describes interaction between one material substance and another. Even the purified soul of the philo­sopher is said to 'gather itself together' (83a7-8, cf.67c8, 80e5), which suggests, as Hackforth says (52, n.3), the spatial diffusion of a vital fluid throughout the body. The fact is that, despite the tenor of his argument for the soul's nature, Socrates does not consistently speak of it as immaterial. Note also that at 83d7 he speaks of the soul as 'sharing opinions and pleasures with the body', implying, strictly, that the body as well as the soul can have pleasures and opinions. He may be speaking 'loosely' here, as Hackforth says (93, n.l), but his language at least is not that of a Cartesian dualist, for whom soul and body could have no attributes in common.

81d6—82d8. The notion of reincarnation in animals is Pythagor­ean (cf. DK 21 B 7), and is developed here with savage irony. Cf. Phaedrus 248e-249b (ed. Hackforth, 88-91), Republic 619e-620c, and Timaeus 41e-42d, 91d-92c. See also 113a5 below.

How can the soul be essentially rational in virtue of its kinship with the Forms, and yet be reborn in animal bodies? This is a further point of conflict (see previous note) between the view of the soul's nature espoused by the Affinity Argument and what is actually said about it. E. R. Dodds remarks: 'The notion of reincarnation in animals was in fact transferred from the occult self of Pythagoreanism to the rational psyche which it did not fit' (G.I. 229, n.43, cf.215).

At 82a7—8 the meaning may be 'And it is clear that the way every other class also would go depends upon the nature of their previous practices' (Bluck). With this translation, what is clear is the general principle that souls will be reborn into appropriate forms of life, rather than the form of life that each soul will actually enter.

At 82c 1—8 the philosopher is said to practise restraint not for ordinary prudential reasons (cf.83b8—c2), but to become purified, and thus escape from reincarnation. This should not be taken to mean, or to imply, that virtue should be practised only for the sake of reward in the after-life. See on 69a6-c3 and 107cl-d2.

83b2-3. 'And not to regard as real what it observes by other means and what varies in various things': what the soul observes 'by other means' is, of course, the sensible world, observed by means 'other' than the soul itself, i.e. the senses (cf.83a3—5). The soul's mediated observation of sensible things is contrasted with its unmediated observation of the Forms. Note the stress (83b 1—2) upon the soul's being 'alone by itself', which is correlative to the state of the Forms (cf.66e—67a).

The sensible world is characterized, literally, as 'what is other in other things'. Bluck (82, n.l) argues for the meaning 'appearing variously in various other things', on the ground that the emphasis is not so much upon the invariability of the soul's object as upon its purity, corresponding to that of the soul itself, when it is 'alone by itself'. However, the variability of sensible things is itself one among several features that render them 'impure', and the phrase 'other in other things' is too vague to be tied closely to any single feature.

The word translated 'real' can also mean 'true' (see on 66b 1— c5), but 'real' is clearly required here, as at 81b4, where the unpurified soul was said to regard nothing but sensible things as 'real'. Cf.83c7 below, where pleasure and pain are said to dispose the soul to regard their objects as 'most clear and most real' (see also 83d6). The superlatives there may be merely intensive ('transparently clear and utterly real', Hackforth). But Plato elsewhere speaks of Forms as 'more real' than sensible things, notably in the Republic 514a— 516a, 585b—e.

What sort of 'reality' is at issue when sensible things are said not to be real, or not to be the only, or the most, real things there are? G. Vlastos (N.E.P.A. 1-19,A.P.A. 1965-6, 5-19) has argued that in questioning the reality of sensible things, Plato does not mean to deny their 'existence'. Nor, in denying that they are 'most real', does he ascribe to them a lesser degree of 'existence' than to Forms. Rather, the Form F is 'more real' than sensible Fs in that 'it yields a better disclosure of what an F is' (N.E.P.A. 5). Would such an interpretation fit the present passage? Philosophy tells the soul not to regard what it observes through the senses as 'real'. Could this mean 'not to regard, e.g., sensible equals as really equal'? Or does it mean 'not to regard sensible things as real beings'? The latter surely suits the context better. The error to which sense-bound souls were said to be prone at 81b4 is not that of taking objects of sensual enjoyment to be the only real Fs, but that of taking them to be the only 'real beings'. In the wider context, the point required is that sensible bodies are not the only real beings, and the survival of the soul therefore need not depend upon the continued existence of the body. It would be irrelevant to this point to insist that sensible things are unreliable exemplars of attributes such as equality or beauty. The essential point would be to maintain that they are not the only 'realities' or 'things that are\ It is the Forms that 'are' in the fullest possible way. Cf.77a2—5, where their 'being' seems undeniably existential, and see on 76d7—77a5.

If this is correct, the point of denying that sensible things are (fully) real, or of asserting that Forms are more real than sensible things, is that the latter are more properly said to be 'real' or to 'be'. The so-called 'realities' and 'beings' of the sensible world are, strictly, miscalled. High standards for the proper use of 'being' are thus imposed, to which Plato does not always adhere himself. See, e.g., 79a6 and note 34.

83b5—e3. Bodily pleasure and pain are reproached here for com­pelling belief in the reality of their objects. There is no suggestion that they are themselves unreal or not fully real. But see on 60b 1— c7. For the alleged 'greatest and most extreme of all evils' (83c2—3) see on 89dl-4.

84a2—b8. The phrase 'not the object of opinion' applied to the Form world at 84a8 glances at the important Platonic contrast between 'opinion' and 'knowledge'. These are differentiated in the Republic (476d—480) according to their 'objects', Forms being the only objects of true knowledge, and sensible things being the objects of opinion. The distinction is not drawn in these terms in the Phaedo. But its essence is everywhere present in the contrast between philosophers and other men—e.g. 82al 1—b3.

At 84b2 the soul is said to 'die'. Strictly, this is inconsistent with the thesis that it is 'deathless'. The meaning, however, is clearly that it ends its :association with the body. Cf.77d4 and see on 64c2—9 (p.86).

Socrates ends (b5—8) by scouting all fears that the soul could be 'blown away by winds'. This reminds us that the Affinity Argument has been concerned to correct the popular misconception of the soul's nature. See on 80d8—81a3.

3.4 Simmias' and Cebes' Objections (84cl—88b8)

Counter-arguments are now advanced by Simmias and Cebes. Simmias likens the soul to the attunement of an instrument, and Cebes compares it with the weaver of a series of cloaks. These images suggest that it perishes at death.

84d4—85dl0. Socrates' comparison of himself with his fellow servants, the swans, illustrates the theme of mortal obedience to divine rule (80a). See on 58a6—c5. For 'eleven Athenian gentlemen' (85b9) see on 59e6-60al.

Simmias' remarks about philosophical inquiry (85cl—dlO) are notable on several counts. In recognizing that certainty is either difficult or impossible to attain in this life (85c3—4, cf.l07a8—b3), he echoes what was said at 66e—67a and 68a—b. The reference to a 'divine doctrine' (85d3-4) may be to Orphic teaching, but Simmias seems to entertain little hope of a revelation. He conveys a clear sense of the difference between philosophic inquiry and religious dogma. His characterization of fallible human reason as a 'raft', and his 'voyage' metaphor, are expressive of the open-minded tone of the discussion. Indeed, Simmias is committed to the very method that Socrates will later describe as his own (99c-100a). The parallel may escape notice in translation, because uniform rendering of logos, a key word in both passages, is impossible. Among its commoner meanings are'discussion', 'account', 'definition', 'reason', 'argument', and 'statement', none of which will suit the present contrast between the human and the divine (85c9, d3). 'Doctrine' has there­fore been "used here. Later, where Socrates speaks of his own re­course to logoi (99e5, lOOal), it has been translated 'theories'. But the close similarity between the passages should be observed. Cf. especially, 85c7-d2 with 99c8-dl. See also on 90b4-91c5,92c 11- e3, and 100a3-9 (p. 181).

85e3—86a3. Harmonia has been translated 'attunement' rather than 'harmony', except in the joking allusion to Harmonia, wife of Cadmus, who personifies Simmias' argument at 95a4. In a musical context 'attunement' is less misleading than 'harmony', whose modern musical sense the word never bears. 'Attunement' is also conveniently related to 'tune', which is needed for the verb at 93a—94a. The noun is formed on this verb, whose primary meaning is 'fit together' or 'join'. It has a basic sense of 'fitting together', 'adjustment', or 'arrangement', in which Simmias applies it to 'all the products of craftsmen' (86c6—7). It can thus mean the 'adjustment' of the parts of an instrument, and specifically the tuning of its strings. Hence it comes to mean a musical scale or mode, or, more broadly, music. See L.S.J. s.v. apfiovia. For various Platonic uses, cf. Symposium 187a—188a, Republic 398e—400a, 430e3—4, 431e8, 443d5-e2, 531a-b, The interpretation to be placed updn it in connection with Simmias' theory of the soul is far from clear. See on 86b5—d4.

86a3—b5. Simmias here parodies the argument of 80c2—dlO. The properties ascribed earlier to the soul—'unseen', 'incorporeal', 'divine'—belong also to an attunement. But just as an attunement does not outlast lyre and strings, so the soul's possession of these properties need not entail that it outlasts the body. This is a power­ful objection, and, as Bluck notes (22, 86), the force of the attune­ment analogy as a criticism of the Affinity Argument is never denied. However, the parallel between attunement and soul, as the latter was represented in the Affinity Argument, is not exact: (i) The case for the soul's indestructibility rested upon its supposed incompositeness. But an 'attunement', however interpreted (see next note), is evidently not incomposite. Socrates later calls it 'a composite thing' (92a8, cf.93al). (ii) The soul was said to resemble the divine by virtue of its ruling the body (80a). But, as will be argued later (94b4—e7), an attunement is incapable of ruling its constituent elements. So 'divine' must apply differently to soul and attunement (cf.94e5—6). (iii) An attunement, although 'unseen' (85e5), is-on at least one interpretation—capable of being heard, whereas the soul is not accessible to the senses at all.

86b5—d4. The attunement analogy is now developed further.

Simmias says that something of this sort is 'what we actually take the soul to be' (b6—7): the soul is a blending or attunement of the bodily elements. Here he goes beyond criticism of the Affinity Argument. The attunement theory of the soul is not merely con­sistent with its destruction at death, but actually entails it. Two problems arise here: (1) Who are 'we' who are said to hold the theory? (2) What exactly does it maintain?

Does Simmias mean, as might seem natural for a disciple of Philolaus (see on 61d3—e9), 'we Pythagoreans'? It is true that the theory has some appeal for the Pythagorean Echecrates (88d3—6). But it is flatly inconsistent with the Pythagorean orthodoxy that seems implicit in Philolaus' teaching on suicide, if this is correctly represented by 61e-62b (cf. DK 44 A 23, B 14). Simmias must be associating himself either, as Burnet suggests (note on 86b6), with a heterodox Pythagorean group, or with 'people in general'. The latter seems preferable for the reasons given by Hackforth (102—3). Cf.92dl—2 and see note 49. The attunement theory is acutely criticized, from very different perspectives, by Aristotle (De Anima 407b27—408a30) and Lucretius (De Rerum Natura iii. 94-135), but neither refers to the Phaedo or identifies the theory with any parti­cular school. Aristotle simply calls it 'persuasive to many' (407b27— 8), and ascribes a wholly different theory of the soul to the Pythag­oreans (404a 16-20). For the view that the attunement theory is of Pythagorean origin, see F. M. Cornford, C.Q. 1922, 145-50. For a full discussion of the relation between the present passage and later versions of the attunement theory, see H. B. Gottschalk, Phronesis 1971, 179-98.

Is Simmias identifying the soul with (a) a ratio according to which bodily elements are combined; or (b) the condition of the body when they are thus combined; or (c) some complex product, analogous to the 'music' of a lyre, yielded by the bodily parts?

A. E. Taylor (P.M. W. 194) takes the theory in sense (c): '"mind" is the tune given out by the body, the music made by the body.' He compares it with the epiphenomenalism of T. H. Huxley. Harmonia could, perhaps, bear this meaning (see on 85e3—86a3), and occasional phrases give the interpretation some support: 'very lovely and divine' (85e5—86al) seems well suited to music; at 86c6-7 Simmias refers to attunements 'in musical notes', which might be regarded as elements in a musical scale or melody (cf. 'notes' at 92cl); and at 94c5 an attunement is assumed to depend not only upon the tension of the strings, but upon their being struck (cf.93a9).

Nevertheless, this interpretation, which is unmentioned by Aristotle, is probably mistaken: (i) The soul is expressly said (86b9, d2—3) to be a 'blending and attunement' of the bodily elements, the hot and cold, wet and dry, which are more closely analogous to the physical components of a lyre than to musical notes, (ii) Simmias refers to attunements 'in all the products of craftsmen' (86c7). These presumably include such things as painting, weaving, building, and furniture (cf. Republic 400dll—401a8), which exhibit 'struc­ture', but produce nothing analogous to the music of an instrument, (iii) 'Attunement' will be contrasted several times with 'non-attune- ment' (93c5, 93e4, 93e8, 94a3), which means 'being out of tune' rather than 'lack of music'. Above all, (iv) Simmias is comparing the soul with something that is destroyed as soon as the tension of the strings is altered: 'when our body is unduly relaxed or tautened by illnesses and other troubles, then the soul must perish at once' (86c3—5). Clearly, it is the tuned state of the strings that 'perishes' in this way rather than the music. For music could hardly be said to 'perish', even though it can no longer be produced, as a result of relaxing or tightening the strings. It is true that the tuned state, once lost, may be restored by restringing, and retuning the instru­ment, whereas the soul cannot be restored to a dead body. But equally, a musical scale or melody can be replayed many times on the same instrument. The analogy breaks down at this point on any interpretation.

There remain the alternatives (a) and (b), that are distinguished and separately refuted by Aristotle. The distinction between these is rather a fine one, and may not even have been recognized by Plato himself. But if a choice has to be made, (b) is perhaps preferable. The notion of the soul as a mathematical ratio is rather abstruse, and would be less likely to have appealed to 'most people' (92d2). It is true that if 'attunement' means the tuned state of an instrument, it will be hard to understand the suggestion at 93al 1—12 that an attunement itself could be tuned. But this is a difficulty on any view.

However interpreted, the attunement theory is a notable ancestor of the many accounts of the soul, from that of Aristotle onwards, which have denied its independent existence either as causally im­possible or as logically absurd. It makes the essential point, central to all such accounts, that the soul is not an entity separate or separable from the body. The theory thus challenges a basic assump­tion of Socrates' arguments, that we are 'part body and part soul' (79b 1—2). In confronting him with it, Plato recognizes it as a plausible and popular rival to his own view (cf.92d).

86e6—87a7. Cebes here accepts the conclusion of the Recollect­ion Argument that our souls existed before birth (al—4). His own account of the soul will be consistent with that view, whereas that of

Simmias is not (cf.91e—92a). Unlike Simmias, therefore, he can preface his objection by harking back to the state of the argument at 77b5—c5.

At 87a5—7 he dissociates himself from Simmias' objection 'that soul isn't stronger and longer-lived than body'. Here he speaks as if Socrates had asserted and Simmias had denied that soul is 'stronger' and 'longer-lived' than body. Neither of them had used those words. But Socrates' argument at 80c—d was based on the supposition that soul is more durable than body, and Simmias had countered by comparing the soul with an attunement. Cebes here rejects the implications of that comparison. He goes on (87a7—b2, cl—5) to parody the argument of 80c—d in his own way.

87b4—8. It is essential to Cebes' analogy, as Hackforth rightly argues against Burnet, that the weaver should both weave and wear his own cloaks (b7—8). This introduces a new and striking picture of the relation between soul and body: the latter is causally dependent upon, 'made by', the former. In this respect Cebes' theory is anti­thetical to Simmias', exactly reversing the order of causal dependency between body and soul. If the soul, in some sense, 'makes' the body, it must exist before it. Makers exist before their products. But they need not survive them. Indeed, many kinds of product, such as statues and buildings, last longer than men. Cloaks are not among these, however, and the special implications of weaving are developed further below. See on 87d3—e5.

Since cloaks are relatively short-lived, it is of some importance that the weaver should die in old age (b5). An old weaver would quite likely die before his last cloak wore out, whereas a younger one might be expected to outlive the cloaks he was currently making. Accordingly, the soul would have less chance of surviving its last 'cloak', if the weaver were old. His age would not matter, of course, if it were stipulated that he goes on weaving and wearing cloaks till he dies. For he would then have to predecease his last cloak in any event.

87cl—3. 'Which class of thing is longer-lived, a man, or a cloak in constant use and wear': the terms of Cebes' parody suggest a distinction between classes and their members. From the premiss that Xs as a class are more durable than Ys as a class, it cannot be inferred that this X will last longer than this Y. Even if souls as a class are stronger than bodies as a class, the fact that the body is preserved for some time after death does not show that the soul survives also. Men as a class last longer than cloaks in constant use. But some cloaks last longer than some men, and even the longest-lived of men will perish before a cloak finished on his dying day. It should be noted that men and cloaks alike have a finite life-span, and that a member of the longer-lived class may thus predecease a member of the shorter-lived one. This is why Socrates must show not merely that soul is 'stronger' or 'longer-lived' than body, but that it is 'completely immortal and imperishable' (88b5—6).

87d3—e5. Cebes now deploys his analogy in a more damaging way. If it is correct, the soul's immortality is not only not proven, but is actually disproved. The soul does not survive the last of the 'many bodies' that it weaves and wears out (d7—8). These 'bodies' are, it should be observed, the series of tissues that are successively wasted and rebuilt in the course of a single life-time. At 87d9 and 87e5 'the body' reverts to its usual meaning. For Cebes, it is, of course, only a series of 'bodies' in the special sense of 87d8. On his theory, our ordinary concept of 'the body' as a single, persistent entity is, strictly, a fiction. Cf. Symposium 207d—e. The body is in constant flux, and must be continually 'rewoven' by the soul(87el, cf.91d7). Death occurs when, and because, the soul itself perishes (e3—5). The soul is conceived here as a source of vital energy, an agent of continuous organic renewal, finally exhausted at death. Thus, it is later suggested that death might be 'just this, the perishing of soul'. See on 91d2—9.

Note that Cebes does not, as yet, raise the possibility that the soul might occupy a succession of 'bodies' in the ordinary sense, i.e. become incarnate in one living body after another (95d5). The words 'especially in a life of many years ... while the man is still alive' (87d8—9) show that his 'weaving' image cannot be taken in that way. The idea of a series of incarnations is not broached until 88a (see next note).

Cebes' statement of his theory shouid be compared with Socrates' restatements of it at 9Id and 95b—d. Significantly, it is never refuted in the sequel. Unlike the attunement theory, it is not recanted or denounced as an impostor (92dl—4).'Socrates will, indeed, argue that the soul, unlike the weaver, is imperishable. But otherwise he never disputes the theory as a model for understanding the relation between body and soul. Perhaps, therefore, it expresses his own view of the soul's animating function (cf.l05c9—d5), though it is incompatible with the notion of soul as a 'prisoner' in the body.

88al—b8. For the text and grammar of this passage see note 43. It might be granted, Cebes suggests (a4—7), that some souls could have undergone many previous incarnations. But unless the soul can be proved 'completely immortal and imperishable', no one can feel assured that his own soul is not incarnate for the last time, and will not perish at his death (b3—8).

To admit that reincarnation is possible would, of course, preclude inferring from the mere fact of death that the soul had perished, as was suggested above (87e4—5). But even so the soul, conceived as a source of vital energy, might eventually wear out. Successive incarnations might have weakened it (a8—9), or incarnation might be the start of a 'terminal illness' for it (cf.95dl—4). Socrates will try to meet this objection by arguing that soul is not the kind of thing that can 'wear out' at all. The effect of Cebes' remarks here is to show that previous incarnations, even if admitted, would give no 'inductive' grounds for a belief in survival. For we can never tell how many previous incarnations a soul has undergone, or what its present condition is.

'No one can know this death or detachment from the body which brings perishing to the soul—since none of us can possibly perceive it' (alO—b3). Hackforth (100, n.2) translates and interprets these words as if they supported the viewpoint of someone who grants, improbably, that the soul 'does not suffer in its many births', i.e. one who holds that the soul could survive many deaths and rebirths without ill effect. But they are much more likely to support the sceptical position for which Cebes himself is arguing, the view that, for all we can tell, our soul's present incarnation may be its last. It is not, indeed, very clear exactly what it is that we cannot 'perceive', since the object of that verb is not expressed. But perhaps Cebes means that, since we cannot perceive the soul, we cannot tell how badly 'worn' it is, and therefore cannot gauge its chances of survival. Hence, no one can know whether his own death, or—more generally—whether any particular death, will bring perishing to the soul.

At 88b 1 the words translated 'detachment from the body' are taken by most translators to mean 'dissolution of the body'. The present version takes 'from the body' as genitive of separation, and makes the phrase explanatory of 'this death' which precedes it. It will then be parallel with 'its present disjunction from the body' at 88b7—8. It shows that 'death' is here used in the sense established at 64c2—9, and suggests that the soul's repeated 'deaths', like its 'births' (a6—9), may be traumata, one of which proves fatal to it. The usual interpretation would imply that the soul might be des­troyed by 'dissolution of the body', i.e. decomposition. But this would be inapposite. For decomposition does not ensue immediately at death (80c4—5), and there is no reason, on Cebes' theory, to suppose that it affects the soul at all.

The words 'completely immortal and imperishable' (b5—6) are significant. Socrates' answer to Cebes will consist in trying to show that soul has just these properties (cf.95cl, 106e9—107a 1). Hackforth (100,n.3,cf.l22, n.l, 164, n.l) thinks that 'immortal' and 'imperish­able' are 'as yet' used synonymously. But Socrates will first argue that soul is 'immortal' (102b—105e), and then that it is 'imperish­able' (106a—e). D. O'Brien (C.Q. 1968, 97) rightly argues that the distinction is anticipated here. See on 91d2—9, 95b5—e6, 105el0— 107al (p.217).

3.5 Reply to Simmias (88c 1—95a3)

Socrates prefaces his reply with a warning against 'misology', the hatred of arguments. He then counters the objection of Simmias by refuting the attunement theory.

88c 1—89c 10. Echecrates' dismayed reactions to Simmias' and Cebes' objections introduce the parallel between arguments and people that occupies this section. Note that the argument for immortality is itself personified at 89b9—cl. Socrates is determined that it shall not 'die', that it shall be 'revived'. The fate of the argu­ment thus parallels his. He is the human counterpart of his own thesis.

At 88d8 Echecrates wants to be convinced that when a man dies, his soul does not 'die with him'. Comparison of this passage with 88a4—6 reveals a shift in the usage of 'die'. In the earlier passage Cebes grants that when we have died, nothing prevents our souls from being born and 'dying over and over again'. 'Dying over and over again' must mean 'being repeatedly separated from the body'. This is consistent with the definition of death at 64c2—9, and with the soul's being said to 'die' at 77d3-4 and 84b2. But in the present passage, the compound verb translated 'die with him' cannot mean merely 'be separated from the body'. It must mean something like 'cease to exist (along with the man himself)'.

At 89c5—6 Phaedo says: 'even Heracles is said to have been no match for two'. Heracles, while fighting the hydra, was attacked by a large crab, and called his nephew Iolaus to his aid (cf. Euthydemus 297c). The 'two' opponents meant here must be (1) Simmias and (2) Cebes, and not, as K. Dorter suggests (.Dialogue 1970, 570), (1) their joint case against immortality, and (2) the consequent threat of misology. Since Socrates does not mention misology till 89dl, Phaedo could not be alluding to it at 89c5. 'While there's still light' (89c7—8), however, glances forward to Socrates' execution, which was due at sunset (cf.61e4,116el—2).

89dl—4. Socrates here says that one could suffer no greater evil than 'misology', whereas at 83c2—9 the greatest of all evils was said to be taking the sensible world to be fully real. These alleged evils, although surely not the same, may, in Plato's view, be related. One who has lost all faith in rational argument will not recognize Forms as the true 'realities'. He will assume that the sensible world alone is real, and will thus be 'deprived both of the truth and of knowledge of things that are' (90d6-7).

'Misologist' and 'misology' have been used to parallel 'misanthro­pist' and 'misanthropy'. The Greek words are rare, but cf. Laches 188c6, Republic 411d7. 'Arguments' suits the emphasis in this context upon rational discussion, but is not altogether satisfactory (see next note). At 90b6—8 and 90c9, arguments are said to be 'true' and 'false'. These words do not go naturally with 'arguments', but have been kept to preserve the required parallel between argu­ments and people. 'Skill in arguments' (90b7) perhaps includes not only 'knowing how to assess arguments' (Hackforth), but the capacity to handle arguments in live discussion. This is not only parallel with 'skill in human relations' (89e5—90a2) but actually requires it. Socrates is shown as a master at handling people and arguments alike (88e4-89a7, 95a8-b4). See also on 84d4-85dl0.

90b4—91c5. The misologist's distrust of arguments is extended at 90c3 to the 'things' that they concern. Cf.88c6—7. Finding argu­ments neither true nor secure, the misologist supposes that 'things' lack these properties as well. Thus, since arguments seem to him 'now true and now false' (d2—3), there can be for him no fixed truths: 'the things that are' share the arguments' instability, like things fluctuating in the rapid currents of the Euripus (c5—6).

The assumption that 'arguments' and 'things' have properties in common underlies Socrates' own resort to 'arguments' (logoi) at 99e—100a, where it has been translated 'theories'. In studying in them 'the truth of the things that are' (99e5-6, cf.90d6-7), he supposes that they reflect that truth. Indeed, the Theory of Forms, in one aspect, is precisely the assumption that truth is discoverable through philosophical arguments. The Forms are the 'truth' that 'true arguments' express. Here the inadequacy of the translations 'true' and 'argument' becomes apparent. 'Truth' not only belongs to 'arguments', but also characterizes or designates 'the things that are'. It is a property not only of arguments, but of what they express. And 'arguments' are not merely pieces of reasoning, to be assessed for their internal logic, but are characterizations of a 'reality'external to themselves. These nuances defy translation, and are a constant source of difficulty. For 'truth' and 'reality' see on 66b 1— c5 and 83b2—3. For 'arguments' see on 84d4-85dl0, 100a3-9 (p. 178),

and previous note.

At 90c 1 the translation 'antinomies' preserves the sense of 'arguments for and against' a given proposition. For the prevalence of such argumentation, see G. Ryle, Plato's Progress, Ch.4, and Hackforth, 108—11. Good specimens are to be found in the treatise Dissoi Logoi, 'Arguments Both Ways', trans. R. K, Sprague, Mind 1968, 155—67. The style of argument is parodied by Plato in the Euthydemus. 'Contradiction-mongers' has been used to translate the same notion at 101e2, where it is applied to those who trade in arguments of this kind, caring nothing for truth but only for victory. Plato denounces them elsewhere for inducing intellectual or moral scepticism (cf., e.g., Republic 538d—539e). The motive to which they appeal, the desire to be proved right and to prove others wrong, is endemic in philosophy. Socrates may even be semi-serious in confessing to it himself (91a—b). But his primary concern to convince himself (91a7—bl) is genuine. It is a mark of the true philosopher, and shows continually in what he says (96b 1, 96e7, 97b4,100d9).

91d2—9. In restating Cebes' position, Socrates asks 'whether death might not be just this, the perishing of soul' (d6—7). D. O'Brien (C.Q. 1968, 98-100) interprets this as a new definition of 'death': instead of meaning, as at 64c2—9, 'separation of soul from body', it now means 'perishing of soul'. Consequently, soul must now be proved not only 'immortal' but 'imperishable'. For it might survive separation from the body once or many times, as Cebes had allowed (88a4—7), and yet be unable to survive all such separations. To be shown capable of this, it must be proved immune from 'death' in the new sense, i.e. imperishable.

O'Brien (op.cit. 98, n.2) links the present passage with 88al0— b2 and 95d2. But neither of those passages points to a redefinition of 'death', and both use the term in its familiar sense. The former (see on 88al—b8) expressly recalls the sense established at 64c. The latter suggests that incarnation might be the start of the soul's perishing, and that it might finally perish 'in what is called death' (95d4), i.e. in 'death' as that term is commonly used. Could Socrates here be operating with the new sense of 'death' that O'Brien suggests? If so, he would be speculating, that 'death', in this new sense, might begin with the soul's entry into the body, and be completed at the time of 'death' in the familiar sense. This would be a confusing way of redefining the term.