On this model, and the alternative Greek model, in which reflection is a prerequisite for virtue, see Gill 1996b, 4.6, 5.7; also, on the characteristic structures of thinking in Greek ethical philosophy, Annas 1993, esp. chs. 1 and 22. Here, as elsewhere, I use 'her/him' and 'he/she' indifferently as indefinite
personal pronouns, even when summarizing (though not translating) ancient authors who use only male forms for this purpose. 2 For the reading of Aristotle's EN assumed in these comments, see Gill 1996b, 5.6, and, on parallels between the Aristotelian pattern and that of Plato's Republic and of Stoicism, see 4.2 and 365-9. 3 See esp. Williams 1985, chs. 4-5, also 1981, ch. 1; MacIntyre 1981, chs. 4-6.
4 For this view, applied esp. to the Greek conceptions of 'nature,' as an ethical ideal, see Gill 1996b, 6.4-5. 5 See e.g., EN iv 5, esp. 1125b31-1126a1; on the Hellenistic debate, see e.g., Annas 1993, 59-66; Nussbaum 1994, esp. chs. 1-3. 6 Much of Nussbaum 1986 and 1994 is, in effect, centered on this issue, typically formulated in terms of the relationship between Aristotelian and Platonic/Stoic ethical approaches; for a summary of her thinking on this, see Gill 1995c.
Thus, 'before reason comes' means, I think, 'before complete reason [post-reflective as well as pre-reflective] comes,' not 'before any rational functions develop'; see further 199, esp. n.15, 208-9, 212 below. Translations are mine unless otherwise attributed. 8 For this reading of the two-stage educational program, see Gill 1996b, ch. 4, esp. 4.2, 4.4-5. 9 Rep. 430c3-5. The alternative interpretation is that what is described here is courage 'within the polis,' whereas what is to follow is courage 'within
the psyche.' But the alternative interpretation fails to account for the claim that 'we shall examine it in a finer [K6XXLOVI form,' c4-5, since the analysis of psychic courage in 442cl-3 is closely parallel to that of 429-30, and is not conspicuously 'finer.' The courage of 442c might be described as 'finer' because it depends on (one's own) reason rather than socially-based beliefs. But the only educational basis for 'reason' discussed prior to 441-444 is social habituation (cf. Gill 1985, 13-15, Gill 1996b, 268); thus, if 442 implies a 'finer' (more reflectively-based) conception of courage, its superiority is not explained before 486a-b. The characterization of C1'wCPPOC1'vvr¡ in 430e-432a can also be understood as a pre-reflective (and less complete) version of the virtue, i.e., the restraint of one part by another, rather than the unforced redirection of desires ascribed to the philosophical nature (cf. 431 a-b with 485d-e). t� The relationship between the 'internalization' of ethical norms in the psyche and the 'extemalization' of them in social structure is well brought out by Lear 1992. 11 see esp. 500c-d; also the description of the virtue not based on reflection as 'demotic' (d8). The philosopher's knowledge is not at this point defined as `post-dialectical'; this explanation of what full philosophical knowledge means (and how it fits into the educational program) is explained later in Rep.; see 202 below.
12 Williams's view is indicated in 1981, esp. 16-18. See also 1985, chs. 6- 8, esp. 129-30, 140-8; Gill 1996b, 285-7. t3 See e.g., 412d-e, 413c-414b, 429e-430b; also 503a, 537e-540d, esp. 539d. 14 For two versions of this reading see Irwin 1977, esp. 202-3: '[the first phase of education] only trains someone to take pleasure, suffer pain, feel shame, anger and pride in the right objects before he acquires reason .... He still regards virtue as a source of pleasure or honour, and that is why he chooses it... he does not choose it for its own sake'; and 1995, 235: 'it matters to [the person so trained] that he is honoured for his willingness to do the brave action without further incentive,' though this still falls short of choosing the fine action 'for its own sake,' i.e., without reference to honor (233-5).
15 See Irwin 1977, 232-46; 1995, 245-8, 288-97. Thus, for this reading too, 'when reason comes' (Rep. 402a) also signifies 'when reflective reason comes' (see n. 7 above); but it does so on the assumption that, until the capacity for reflective reason is acquired, 'reason' (as a significant locus of moral motivation) is completely undeveloped. 16 See e.g., Rep. 522a, taken as a comment on 413c-414b, 429e-430b; also (apparently stressing stability of character, rather than correct beliefs, as a precondition for dialectic), 539d, taken in the context of 537e-540a. 17 See Gill 1996b, ch. 4, esp. 4.3-5, also 326-7, 338-40, on the combination of Kantian and modem (esp. post-Kantian) thinking informing Irwin's approach. See esp. 268-71 for a contrasting characterization of pre-reflective virtue. 18 See EN 1095b3-8, 1098a33-b4, EE 1216b35-1217al0; also Burnyeat 1980, 71-3, Sherman 1989, 196-7. On the relevance of Aristotle's distinction to Rep., see Gill 1996b, 272-5.
19 On the belief-knowledge distinction, see Rep. 476d-480a; on the relevant distinctions in the divided line and cave images, see 509d-517e, esp. 5 lid, 517b, also 519c-d, 520c. For interpretations of this distinction which avoid, as I would wish to, a radical 'two-worlds' view, see Annas 1981, ch. 8, esp. 209- 15 ; Fine 1978, 1990. 20 am grateful to Victor Caston for suggesting to me this way of explicating the distinction. t See 197 above. Another form of explanation was suggested to me in discussion at Brown University during the BACAP visit. Music and gymnastics implant (in a direct, unmediated way) a structure in the psyche whose intelligible basis (which can be analyzed in terms of mathematical or harmonic proportion/ratio) is understood through the combination of mathematical and dialectical education at the second stage. However, I focus here on a type of explanation which centers on the conversion of belief into knowledge. See further nn. 25, 56 below. See esp. 534b9-c3: the philosopher's È7TLCTT�P.'T/ is expressed in his ability 'to define the Form of the Good in argument, distinguishing it from everything else, and, as though in battle, surviving every critical examination [�\€�)(oy], determined to analyze [ÈÀiYX£LV] by reference to essence [or 'being,' o�o-t'al, not opinion [866a].'
On psychic harmonization, see 197 above and 212-214 below. For artistic media in the ideal state as expressing sets of virtues, see 395c- d, esp. c4-5, 399a-c, esp. 399cl-3, 402c. The linkage in language between 402c ('copies' of the virtues) and 500d, 501b, invites us to see the former passage as describing a pre-reflective version of the order (K60-iAog, 500c) embodied in the latter. The virtues as a set provide the political structure of the ideal state (427d-434d). On poetry as conveying a pre-reflective world-view, see 208 below. 26 See n. 23 above, taken with 508e-509d, 517b-c, 532a-c. See 500c2-5, c9-d2, d4-8, 501b9-c2, cited in 197 above. Such knowledge is not explicitly connected with dialectic in 500-1; but, when the role of dialectic in the educational program is explained in 531d-534e, its contribution to providing this structured understanding is made clear (see refs. in n. 26 above). On pre-reflective rationality, see 208, 212 below.
Thus, Dustin takes Rep. 518b-d, esp. b, as implying a rejection of the two-stage model (more precisely, of the idea that the first stage could make a substantive contribution), in favor of a model in which there is a single, progressive movement towards the Forms. As noted in Section I, text to nn. 5-6, my primary concern here is not so much the two-stage model of ethical education, which I see as common to Plato's Republic and Stoicism, but rather the problem of how to locate the development of passionless virtue within this model. 3� See Plato, Symp. 21 Oa-21 I e7, also (though less clearly marked as a single process), Phaedrus 250a-256e. 32 See e.g., Rep. 486a5-6, 490a8-b7, noted by Dustin; also SOSdI l-e2. 33 Dustin cites the myth of the cave as an instance of continuous movement towards the Forms; but, as indicated in 200 above, this movement is also correlated with movement between cognitive states; and those states are themselves linked with different types of discourse, within the myth (S15b4-5,
dl-5) as well as subsequently (532a-d, 534b-d); see also, on the ethically shaping power of public discourse (here, to bad effect), 492b-d. See nn. 26-7 above. 35 On this issue, see e.g., White 1988, 252-3; Sayre 1988, 95-7; and, on the other side, Gadamer 1988, esp. 264-5; also, discussing Plato, Epistles \\\ , 340- 5, Gill 1992, 158-60. On the late Platonic dialogues, see Gill 1996a, discussing the issues raised in Gill and McCabe 1996. On the relationship between types of dialogue/dialectic and cognitive or ethical states in Laws, see Bobonich 1996. On the parallel between Plato's Rep. and Aristotle, see 200 above. See 197-198, 200-202above. Of the passages Dustin sees as rejecting or transcending the two-stage model, 518b-c, esp. b7-c2, does not seem to be directed at the first stage in Rep.; but this stage does seem to be referred to in 518d9-e2, as it is, explicitly, in 522a, where it is presented as preliminary to philosophical (mathematical and dialectical) education. For this point of interest, see Section I.
41 Although this passage begins with a reference back to the prescriptions for poetic education in the ideal state (603d4-5, referring to 387d-e, esp. e3-7), the reactions described (here as elsewhere in the Book x discussion of poetry) are conventional, real-life ones, not correlated with either stage of the ideal educational scheme. 42 'Do you think that he will fight and resist the pain more when he is observed by his equals, or when he is alone and by himself? ... When he is on his own, I think he'll dare to say many things, which he would be ashamed of if anyone heard him, and do many things, which he would not want anyone to see him doing ....' (604a1-8). a3 See 211 below.
On ideas of this type in the early, 'Socratic,' dialogues, see e.g., Irwin 1977, 100-1, 1995, 73-5; Vlastos 1991, ch. 8, esp. 224-31. 45 Contrast Book x, 603e-604a, in which (in spite of the reference back to 387e in 603e4-5), the strongest reaction ascribed to the good person (after the rejection as 'impossible' of the thought that 'he will not grieve at all') is that 'he will be moderate fJ.¡£TpLåCT£I] as regards his grief,' subsequently explained as that he will 'fight and resist the pain' (603e7-8, 604a2); see further 211 1 below. The 'Socratic' assumption that emotions depend directly on beliefs (knowledge or its absence) is most evident in the (apparent) denial of akrasia in Protagoras 352b-357e. Nussbaum 1994, 92-3, also notes the Socratic character of the ideas referred to in Rep. 387d-e, and contrasts the implied view of the ethical status of emotions with that of Aristotle (on this contrast see n. 6 above).
Thus, such beliefs are, presumably, among those referred to in Book IV, as those (true or right, op9rj) beliefs 'about what and what sort of things are to be feared [8ema]' whose retention constitutes 'citizen's courage' (429c7-8, 430b2-c5), though their content is not specified there. It is more puzzling that the Book x passage contains the idea, 'that nothing in human affairs is worth great concern' (604bl2-cl); cf., in Book vi, that 'human life is not something important' (486a9-10). I see no obvious explanation for this, apart from the suggestion that (if the reactions in Book x are consistently located at the conventional level) this idea is considered in some way to have become part of conventional thinking about death. The idea that life (under certain unacceptable conditions) can cease to be worth living is presented as conventionally available in Rep. 406d-407a, but without the further thought that 'human life is not something important.' 8 On post-retlecdve understanding, see further 209-10 below. 49 See further 212-13 below. I do not pursue here the question whether the beliefs (and reasoning) that inform emotions and desires in this psychological model are functions of the two lower parts or of the reasoning part acting in collaboration with the two other parts. 50 For this suggestion, see n. 21 above. 51 See 208 below.
52 See 40pd11-el (for the general point, see 400c-402c), taken with 395c-d, 396c-d, 399b-c). Beliefs such as those specified in 387d-e are explicitly linked with the promotion of virtues such as courage (386a-b). 53 The qualification refers to the idea, developed in 213 below, that post- reflective virtue brings with it a deeper level of psychic harmony. See 399a8-b3 and 387d6. The idea that the first phase is presented as a 'rehearsal' in beliefs and feelings is often noted (see Rep. 395c-d, 396c-d, 399b-c, and e.g., Janaway (1995), ch. 4); what needs to be added is that this stage is a rehearsal in passionless virtue. 5 See 401e4, 402a1-2, and 387d11-el, taken with 197, 206 above.
56 On 'falsehood' in the first stage of education, see e.g., 377e, 382a-e, esp. 382a7-8, bl-9, 391d-e; also Gill 1993, 42-7. The idea that justice leads to happiness appears in negative form (as what conventional poetry fails to convey) in 392a-b. For this way of reading Plato's Rep., see e.g., Nussbaum 1986, ch. 5, esp. 138, 154-8, 160-3; also Williams 1985, 110. See 412c-e, 413c; these prerequisites are referred to in 503a (and possibly 484b9-cl); see also n. 13 above.
For post-dialectical knowledge of Forms as yielding a systematic and inclusive perspective on the nature of reality, see 202 above. The idea that an understanding of the permanent, unchanging character of the Forms brings an altered perspective on human life (stressed in e.g., Symp. 211b, d) is highlighted in 486a5 ('always'), a8-9 ('survey all time'). On 500-1, see 197 above; and, on the linkage with post-dialectical knowledge, 202 above. 61 A further issue which arises from this point (one not taken up here) is whether the Republic's conception of ethical development entails that one comes to value wisdom more than its practical realization, an issue highlighted in the account of the philosopher-rulers' attitude towards re-entering the cave (519-21); for one approach to this crux, see Gill 1996b, 4.6-7. For Stoic thinking which is comparable to the interpretation of 500-1 offered here, see 223 below.
See e.g., Annas 1981, 130-1, 338-40; Price 1995, 68-9. For this view, see Gill 1996b, 4.2, esp. 245-6, 259-60; key texts for this view are Rep. 442clO-dl, 554cl2-d2, d9-e4. As suggested in Gill 1996b, 252, 'reason,' in Plato, and in Greek thought generally, serves to denote (1) a function, (2) a norm, and (3) a mode of desire; and the idea of 'reason's rule' in Rep. iv, viii-ix, is, essentially, the dominance of reason as a norm. However, these senses of 'reason' are often compounded, thus contributing to the (immense) significance of this notion in Plato's Rep.; see further Cooper 1984, esp. 6-8; Kahn 1987, esp. 84-91. See 604b6-7, d5-10; also 603a, 606a; also 604al0 (X6yog and v6�LoT), taken with 205 above). See Rep. 604d8-10 (also 606a3-7); on the tripartite model, see 436a- 441b, also 441d-442d, and 212-13 below. It corresponds to the language of 'control' in 442a4-b3, rather than that of 'agreement' in 442dl; on the two kinds of language, see Gill (1996b), 4.2.
67 See 603e-604b, esp. 'will moderate' (e8); 'fight and resist' (604a2); 'resist' (604al0); 604d; 606a-b, esp. 'the part which is forcibly restrained' (606a3); 'relaxes its guard' (606a8); 'restrain' (606b8); 606c-d, esp. 'restrained' (c6). For the alternative model ('agreement'), see n. 63 above. See 205 above. For (what is presented as) a conventional model of ethical psychology, namely a complex of struggling parts (knowledge, pleasure, fear, etc.), see Prot. 352b-c, 353c). For further suggestions about the considerations shaping the psychological model of Rep. 10, see Ferrari 1989, 132-41; Lear 1992, 208-15. See e.g., Rep. 439d5-8, 440cl-d3, 441e4-6, taken with 580d-e; for the model assumed, see Gill 1996b, 251-2. On the question of the scope of the cognitive functions allocated to the two lower parts, see n. 72 below. 'o Cf. 441e8-442b3 and 442d1 with 429b-430c; see further Gill 1985, 8-12, 1996b, 268-70. For the idea that 'death is not terrible' as a feature of pre- as well as post-reflective courage, see 206-7 above. The idea that the appetitive
part is 'silenced' (an idea borrowed from McDowell 1980, 370) is not explicit in these passages, but is compatible with the language of internal dialogue used to characterize intrapsychic relationships in the Republic; on this language, see e.g., Moline 1978, esp. 9-10, 13-14; Annas 1981, 129-34; Gill 1996b, 252-9. 71 See Rep. 440a-e, on anger; also Annas 1981, 339. In other words, the responses of 'spirit,' like Stoic passions, involve the belief 'that it is right to react, with (e.g.,) anger or grief; seeRep. 440cl-d3, and, on this feature of Stoic passions, Long and Sedley (1987) (=LS) 65 B, Cicero, Tusculans. 3.76 (on grief). 72 For these two ideas, see Rep. 554c-e, esp. d2, 586d-e; and 485d-e, esp. dlO-12, respectively. Several recent accounts have suggested that the first idea is more intelligible if we assume that the appetitive part, like spirit, is capable of means-end reasoning, as well as of responding to beliefs, and that it is 'persuaded' (in a quasi-personal way) that its desire for pleasure will be best served by responding to reason's direction. See e.g., Kahn 1987, 84-6; Irwin 1995, 219-22; Price 1995, 60-4.
This may be regarded as a stronger version of the 'silencing' of fear described in 212 above. Another possible line of thought is that the feelings of pleasure and pain become attached to the objectives of (reason-ruled) spirit and reason to the point where they correspond to the standards recognized by these other parts. At the pre-reflective level, 401e-402a, esp. 'enjoy fine things' (ra . .. rcaJ�a ... Xaipwv, 401e4) might be taken to point in this direction; see also the idea of 'desiring' (sexual pleasure) in a way that is 'self-controlled and musical' (403a7-8). 74 The most obviously relevant Stoic ideas are: (1) that the full recognition that virtue is the only real good brings complete peace of mind or absence of passion (daa9e�a), and (2) that emotions or passions (7r&017) derive from taking ethically 'indifferent things' (such as health, continued life, or wealth) to be goods. See 217 below.
75 See e.g., LS 61 L, taken with 65 J esp. (1); also D.L. vn.89, Cic. Tusc. III.2. 76 This view underlies one of Chrysippus' explanations for wrongdoing: 'the influence of those with whom one associates' (D.L. 7.89). 77 On the (complex) question of the history of Stoic thinking about the ethical status of actual (as distinct from ideal) societies, see e.g., Schofield 1991, Vander Waerdt 1991, 1994. See LS 59, esp. D(3-4), E(2), F, Q; also Kidd 1971b, 1978. See further on this passage, and on the Stoic theory of ethical development (01KEtwo-Lg), Gill 1990, 144-8; Engberg-Pedersen 1990, 80-97; Annas 1993, 166-70, 263-4.
See also D.L. vn.88, cited in 223 below. 81 See Rep. 485d-486b, 500-1, and 197, 209-10, 213 above. See LS 26; and see further (with differing views about the extent to which the idea of nature figures as a significant one within Stoic ethical philosophy, and, if so, in what sense, White 1985; Striker 1991; Annas 1993, ch. 5; Cooper 1995; Annas 1995. See e.g., Epictetus Discourses not.2.6-17. For the idea that a proper understanding of 'the norms of nature' depends on developing an appropriate type of ethical character, either pre-reflectively or through reflection, see Gill 1990, 1996b, 6.5 (on Aristotle and Stoicism), and 1995b, ch. v (on Greek thought in general).
This claim must be qualified by acknowledging that the most common exemplars for wisdom (or 'sagehood'), Heracles and Diogenes the Cynic have, clearly, reached this state by a less reflective route, as is noted by Kerferd 1978, 127. This kind of linkage between psychological parts and phases of education is suggested by Rep. 522a, taken with 518c-519c, esp. 518e9-519al. See LS 65, esp. B-D, F, L; see further Inwood 1985, ch. 5, esp. 173-5; Frede 1986, esp. 99; Engberg-Pedersen 1990, ch. 5; Nussbaum 1994, ch. 10; Price 1995, ch. 4; Brennan forthcoming.
See LS 65 I, K, M-R; also Galen De Placitis Hippocratis et Platonis v.5.32-5, De Lacy (1978-84), vol. 1, 324-5; see further Kidd 1971a; Cooper (forthcoming); Gill (forthcoming a), sect. III. See refs. in nn 70, 72 above. See Section III, esp. 206-8, 212 above. On Epictetus, see Gill 1995a, Introd., xviii-xxiii, with refs. to previous work.
91 Epict. Diss. m.2.1-2 (also 3-5); see also L4.11, n.17.14-18, hi.12.13-15, Handbook 52; trans. Hard in Gill (1995a). More precisely, the protreptic is to more technical philosophy, but if and only if this is conducted with a view to, and in the light of, developing ethical understanding; see e.g., Diss. 1.4, esp. 6-12, m.2. For other ancient categories of practical ethics, see e.g., Stobaeus Eclogue. n.39-45 (Philo and Eudorus), Seneca, Ep. �xxxtx.l4. See e.g., Diss. 1.1.7, IL1.4; see further Inwood 1985, 84-6; Long 1991, 111-20. A related theme is that we all have natural 'preconceptions' (7TpOÀ�"I£L<;), which enable us to carry out this process in line with the recognition that virtue is the only good (e.g., 1.22, n.17).
94 For the Socratic ideal of 'the examined life' (Plato, Apology 28e-30b), see e.g., Diss. 111.1.19-23. 95 Cf. Hi.2 with (e.g.,) L2, 11.10, 111.3.5-9. See e.g., 1.4, esp. 9-11, n.17.14-18, tn.12.15, Handbook 52. I am grateful to Tad Brennan and Richard Sorabji for clarifying this point. 97 See e.g., 1.4, esp. 3-4, 18-19; 1.6.37-43; see 216 above.
See m.3.5-6: ... the good is thus preferred above every form of relationship [0IK£LÓT11<;]. My father is nothing to me, only the good.—Are you so hard-hearted? Such is my nature, and such is the coin which god has given me. If, therefore, the good is different from the noble and just, off go father and brother and country and everything of that kind' (trans. Hard). That the same point applies to one's death comes out in a closely related passage in Diss. ttt.24.88: 'Henceforth, when you take delight in anything, bring to mind the contrary impression. What harm is there while you are kissing your child to say softly, "Tomorrow you will die"; and likewise to your friend, "tomorrow either you or I will go away, and we will see each other no more'" (trans. Hard).
100 Anaxagoras' much-reported dictum, on his son's death, that 'I knew I had begotten a mortal,' was typically interpreted by Stoics in line with this view; see Gill 1994, 4613-14. On the sage's understanding, see 223 below. 101 But the distance from conventional ethics should not be exaggerated: the idea that 'we are all mortal,' underlying the advice, is the most common theme in the Greco-Roman 'consolation.' 102 See further on Diss. 1.2 as exemplary of (orthodox) Stoic practical ethics, Gill 1988, 187-91. 103 The relationship of such language to Stoic debate about responsibility and determinism, and the fact that Epictetus presupposes (but does not seek to develop) Chrysippus' position, is brought out better by Bobzien forthcoming than by Dobbin 1991, 119-21.
104 For 'freedom,' in this sense, see Diss. iv.1; for the idea that desire (5peetv) should be directed only at the good (not preferable indifferents), and that, if you cannot do this, you should 'delay' desire, see 1.4.1, Hand6ook 2; see further Inwood 1985, 115-26. 105 Cf. t.17.27-8 with 1.1.27, n.6.9-10, m.1.99-102. See further, on 1.17.15- 29, Hahm 1992, 40-3; on the sage's understanding of nature, White 1990, 49- 55.
106 On the issues raised for Stoicism by the idea of divine providence, see LS 54. 107 See Seneca, EpisC xcv.57, cited in 216 above (esp. 'tranquillity'); D.L. vn.88, cited above (esp. 'smooth flow of life'), taken with refs in n. 105 above and Epictetus Diss. 1.4.3. 108 This linkage is implied in the connection between 1.2.12-21, summarized above (also e.g., t.1.17-32) and 1.17.27-8). 109 See 216 above. It is striking, however, that Epictetus' most frequent paradigms for sagehood are Diogenes the Cynic and Socrates, neither of whom reached wisdom by this route (see also n. 84 above). 110 See 1.2.19-21, taken with 221-2 above. For related suggestions about 'vernacular apatheia' and links between Stoic and conventional Roman thinking on reason and passions, see Gill (forthcoming b), sect. iv.
111 See Section in, esp. 209-10 above. 112 See 206-7above. a3 On ethics (and politics), see e.g., Vander Waerdt 1994; and, on psychology, Gill (forthcoming a), sect. IV. 114 The paper prepared for the BACAP lecture at Brown University has been revised in the light of Christopher Dustin's commentary (Section II), written comments from the BACAP Proceedings referee and Victor Caston, and other valuable responses at Brown, Princeton, and the University of Pennsylvania, for all of which I am most grateful.