¹ The metaphor of the poet's ‘voice’, here perhaps suggesting both the idea of oral tradition and that of personal authority, appears at Plato, Prt. 347e3 (a passage I discuss later).

² Aristotle's own quotations from poetry are analytically indexed in Moraitou, D., Die Äusserungen des Aristoteles über Dichter und Dichtung ausserhalb der Poetik (Stuttgart, 1994), 130–42Google Scholar; but Moraitou's discussion of the material (120–4) is superficial.

³ Murray, P., Plato on Poetry (Cambridge, 1996), 230–1Google Scholar, unnecessarily suspects that the quarrel was largely of Plato's own making; so too, for more complex reasons, does Nightingale, A. W., Genres in Dialogue: Plato and the Construct of Philosophy (Cambridge, 1995), 19CrossRefGoogle Scholar, n. 15 and 60–7: see my remarks in Ancient Philosophy 17 (1997), 455–6.

⁴ Brandwood, L., A Word Index to Plato (Leeds, 1976)Google Scholar, supplies a very useful ‘Index of Quotations’ (991 1003) which lists 397 Platonic and pseudo-Platonic passages (section A), and 275 passages of Greek literature (section B), the great majority from poetry; all references to ‘Brandwood’ in the following notes are to this index. However, Brandwood gives no criteria for inclusion of a passage in section A, and his list embraces a variety of partial or allusive references to poetry, as well as overt quotation and verbally close paraphrase, but not (n.b.) citations that lack specific verbal elements from the original (e.g. Rep. 1.331d-e, Crat. 392b-d, Symp. 190b7-cl). Moreover, Brandwood omits some passages which contain certain or possible verbal allusions to poetry: see e.g. Prt. 310dl (óὒπνος ảνῆĸεν, a discreet allusion to e.g. II. 2.71), Rep. 5.457a6–7, 458d5, 8.546el, and Phaedo 95b7–8 (where Homeric allusion is signalled, but the phrase is not attested).

⁵ I will not, however, be concerned with questions of the textual accuracy of Platonic quotations (nor does my use of the term ‘text’ carry any implications about the state or availability of written copies). The fullest handling of this subject is Labarbe, J., L'Homère de Platon (Liège, 1949)Google Scholar, who argues for the influence of an ongoing performance tradition; cf. Nagy, G., Poetry as Performance (Cambridge, 1996), 142–6Google Scholar (in the context of a highly controversial thesis). The older treatment by Howes, G. E., ‘Homeric quotations in Plato and Aristotle’, HSCP 6 (1895), 153–237Google Scholar, may be right to minimize the idea of error due to faulty memory, but his arguments are not always sufficiently rigorous. Tarrant, D., ‘Plato's use of quotations and other illustrative material’, CQ 45 (1951), 59–67CrossRefGoogle Scholar, at 62, thinks some inaccuracy in Platonic quotation is due to imitation of conversational informality. Much Platonic ‘rewriting’ is diagnosed in the three-part investigation by Lohse, G., ‘Untersuchungen über Homerzitate bei Platon’, Helikon 4 (1964), 3–28Google Scholar; 5 (1965), 248–95; 7 (1967), 223–31.

⁶ My case does not depend on a comprehensive hypothesis about the extent of interest in poetry within Greek societies: it is sufficient that Plato both depicts, and writes for, people who shared this interest. However, the evidence of both texts from and the cultural institutions of democratic Athens establishes that the ability to quote and/or recognize substantial amounts of poetry was widely spread among male citizens: for oratory, see Perlman, S., ‘Quotations from poetry in Attic orators of the fourth century B.C.’, AJP 85 (1964), 155–72Google Scholar; Ober, J., Mass and Elite in Democratic Athens (Princeton, 1989), 177ff.Google Scholar; Ford, A., ‘Reading Homer from the rostrum: poems and laws in Aeschines’ Against Timarchus’, in Goldhill, S. and Osborne, R. (edd.), Performance Culture and Athenian Democracy (Cambridge, 1999), 231–56.Google Scholar Comedy calls for separate treatment in its own right: its evidence is complicated by numerous distortions; thus e.g. Strepsiades’ expectations at Clouds 1354–65 may be filtered through an element of symposiac parody.

⁷ The citation of poetry would be, of course, precisely one way of sharing it and giving others an opportunity to acquire or renew familiarity with it: in Plato, note esp. Prt. 339b4, where Protagoras, having started to quote Simonides’ song, asks Socrates if he knows it (Socrates replies that he knows it well) and whether he needs to quote it in full; cf. ps.-Pl. Theages 125dl 1—12. On the traditional availability of poetry by performance and memorization, see Herington, J., Poetry into Drama (Berkeley, 1985), chs. 1–2.Google Scholar Poetry is one, though not the only, thing that Isocrates (if he is the author of Ad Demonicum) has in mind in recommending the storing up and maintenance of a mental stock of useful logoi: Isoc. 1.18–19 (cf. 51).

⁸ Some variations in the frequency of quotation are noted by Tarrant (n. 5), 59 (though some of her figures should be adjusted in the light of Brandwood section A). I note that all thirty-four quotations from Laws in Brandwood section A belong to the Athenian (a subtle cultural colouring?); yet we are left to infer that the Spartan and Cretan can at least appreciate his use of poetry (a point specifically made at 1.629b).

⁹ Quotation is dispensed with at e.g. Crat. 392b6–8, Laws 3.682al-2 (second reference); elsewhere, we are often given no more than a token indication: e.g. Laws 1.629a-b ff.

¹⁰ For what it is worth, Xenophon too depicts a Socrates extensively au courant with poetry; note Mem. 1.2.56–9, where Xenophon defends Socrates against the charge of having quoted the poets to support heterodox views.

¹¹ Tarrant (n. 5), 59, distinguishes between integral and embellishing quotations; cf. Vicaire, P., Platon critique litteraire (Paris, 1960), 81–2.Google Scholar But see n. 17 below.

¹² Brandwood A78 (with B131 for three other allusions to the same line, but not to the verb) refers only to II. 6.211, but the verb of course occurs elsewhere too.

¹³ One example of this is Homer's οὐ γáρ ảπò δρνóς … οὐδ’ ảπò πέτρησ, ‘neither from oak nor rock’(Od. 19.163), quoted at both Apol. 34d4–5 and (loosely, without any metrical shape and without the negatives) at Rep. 8.544d7–8: for this and cognate expressions, see West, M. L., Hesiod, Theogony (Oxford, 1966), 167–9Google Scholar, on line 35. Plato himself marks proverbial status, in another case, at Laws 7.818bl–3, for a phrase which is cited, never in the same wording, twice more in Laws (5.741a4–5, 7.818el–2). Tarrant (n. 5), 62–4, lists poetic ‘tags’ which are akin to proverbs.

¹⁴ For Socrates’ relationship to Euthyphro, see the humorous hints at Crat. 396d, 399a, 409d, 428c.

¹⁵ Cf. the adverb χαριέντως, in connection with poetic quotation, at Prt. 344b2, Rep. 1.331a3; but in neither of these cases is χóρις divorced from the meaning of the text.

¹⁶ The significance of this point needs ideally to be set within the larger framework of poetic influences on Platonic writing—influences which, at an extreme, help to generate the production of full-blown muthos of Plato's own.

¹⁷ I do not mean to claim that there is a sharp dividing-line between quotations which do and those which do not address the significance of the source text. At Phlb. 47e8–9, for example, it would be artificial to decide whether Socrates quotes II. 18.108–9 (on anger ‘sweeter than dripping honey’) because it contains a memorable simile or because it lends weight to the idea of mixed psychological pleasures and pains. Different again is Rep. 5.457b2–3, which adapts a phrase of Pindar's (fr. 209 Snell-Maehler) both to play piquantly with a poetic image and in order to reply to the criticism of philosophers in the original context: see my commentary, Plato Republic 5 (Warminster, 1993), ad loc.

¹⁸ For the noun ῥῆμα see esp. Prt. 342e-3d, and e.g. Rep. 1.336al; in this usage the term implies an especially striking or pregnant expression (cf. e.g. Pindar, Pyth. 4.278, Aristoph. Frogs 97). ῥήματα in this sense are akin to, and part of the same larger cultural phenomenon as, the forms of verbal wisdom considered by Russo, J., ‘Prose genres for the performance of traditional wisdom in ancient Greece: proverb, maxim, apophthegm’, in Edmunds, L. and Wallace, R. W. (edd.), Poet, Public and Performance in Ancient Greece (Baltimore, 1997), 49–64.Google Scholar Note also passages such as Rep. 5.466c2–3, Laws 3.690e2–3, where Hesiod's wisdom is cited in a saying which was almost certainly an independent proverb: see Works & Days 40, with West's note ad loc. Closely related to the idea of a ῥῆμα is that of a γνώμη), a concise and memorably turned ‘judgement’ or statement of ‘opinion’: see Prt. 340b7 for a poetic instance of the concept; cf. Isoc. 2.44 (in connection with the idea of poets as ‘guides for life’, ibid., 43), Aeschin. 3.135. All such sayings presumably played a part in the process of excerpting and anthologizing poetry which Plato himself attests at Laws 7.811a: on the history of gnomologies and related ancient collections, see Chadwick, H., ‘Florilegium’, in Klauser, T. (ed.), Reallexikon für Antike und Christentum (Stuttgart, 1969), 7.1131–60.Google Scholar

¹⁹ The idea of poets as ‘witnesses’ occurs in Plato at e.g. Symp. 196e3–4, Laws 1.630a3, 3.680d2–4; cf. e.g. Arist. Met. 995a7–8 (cited above in my text), Rhet. 1375b28–9, with n. 46 below. On its more general occurrence, see Dalfen, J., Polis und Poesis (Munich, 1974), 41ff.Google Scholar

²⁰ πεóθεσθαι: see Rep. 1.331d5, 2.365e6, 3.408c2, 5.468d7, e8, 469a3, 7.536dl; cf. the negative ảπιστεîν at Rep. 5.331e5–6.

²¹ Cf. Laches 201b, where the line is again cited approvingly by Socrates. Likewise at e.g. Euthyphro 12a-c, where Socrates quotes and disagrees with some anonymous lines of epic (Cypria F24 Davies) which contain the statement that ‘where there is fear there is also inhibition (αảδώς)’: here, not only is the dramatic status of the words ignored, but Socrates illegitimately takes the remark to have universal applicability to all cases of fear and αἰδώς, when the context (slender though it is) suggests that what is meant is limited to fear of persons. Similarly contestable extrapolation can occur even when dealing with a first-person voice which has some claim to be equated with the poet: see e.g. Socrates’ disagreement at Rep. 7.536dl-2 with Solon (fr. 18W) over the view that a person can learn many things in old age, which apparently depends on taking as a generalization what in its context is a personal claim (‘I am always learning many things in my old age’); the same line is used positively by Laches at Laches 189a4–6; cf. ps.-Plato, Amatores 133c6.

²² There are seven passages of poetic citation in Callias’ speech: see Brandwood A83–9 for references. On Callias's use of Euripides’ Antiope, in particular, see Rutherford, R. B., The Art of Plato (London, 1995), 166–8Google Scholar; Nightingale (n. 3), 73–87.

²³ For references, see Brandwood A254–64.

²⁴ My formulation of this double model of meaning is deliberately provisional and designed to promote further elucidation of the principles active in Platonic citations from poetry. Such a model has, of course, been a focus of contention in much modern theory of (literary) interpretation, from the historically intentionalist hermeneutics of E. D. Hirsch, with its distinction between ‘meaning’ and ‘significance’ (e.g. The Aims of Interpretation [Chicago, 1976], 1–13), to the (qualified) anti-intentionalism of Jacques Derrida's deconstruction, with its insistence on the indeterminacy of ‘context’ and the necessary ‘iterability’ of language (see esp., in this connection, ‘Signature Evénement Contexte’, in Marges de la philosophie [Paris, 1972], 367–93).

²⁵ On this factor see my edition, Plato Republic 10 (Warminster, 1988), on 605dl, 606b2; for its place in Plato's larger views of mimesis, see my ‘Plato and the psychology of drama', in Zimmermann, B. (ed.), Antike Dramentheorien and ihre Rezeption (Drama 1: Stuttgart, 1992), 55–73CrossRefGoogle Scholar, at 56–64 (revised version forthcoming in my book, The Aesthetics of Mimesis [Princeton, 2000]).

²⁶ On the importance of Achilles to Plato, see Tarrant (n. 5), 60.

²⁷ This link is never overtly made in the treatment of inspiration in Ion, even though Ion himself draws attention (abortively) to the poet's dramatizing imagination at 540b3–5.

²⁸ see Lucas, D. W., Aristotle, Poetics (Oxford, 1968)Google Scholar, on 61a31–61b9: Aristotle's examples are not of straightforward self-contradiction, but his language seems clearly enough to cover such (alleged) cases too. In his Homeric Problems, a much larger work which probably lies behind the compressed material of Poetics 25, Aristotle seems to have discussed numerous alleged contradictions: see esp. frs. 146, 149, 153 Rose (= O. Gigon, Aristotelis Opera III [Berlin, 1987], frs. 370, 373, 377). For a further approach to the issue of poetic contradictions, see the reference to Antisthenes and Zeno the Stoic at Dio Chrys. 53.5.

²⁹ See Rep. 3.390d, 4.441b-c: on the Homeric passage see my ‘Traditional Greek conceptions of character’, in Pelling, C. (ed.), Characterization and Individuality in Greek Literature (Oxford, 1990), 32–59Google Scholar, at 38–42.

³⁰ Other occurrences of this idea are Prt. 341e8, 347a4 (see my text below), Ion 530b10, c4, Lysis 205b2, Rep. 1.332cl; cf. the definition of διóνοια as inner dialogue, to be expressed outwardly in logos, at Soph. 263e. This point warrants a qualification on the claim of Asmis, E., ‘The poetic theory of the Stoic “Aristo”’, Apeiron 23 (1990), 147–201CrossRefGoogle Scholar, at 148, 153, 156, that the idea of poetic διóνοια was a conceptual innovation of the Hellenistic period. Basic pre-Hellenistic occurrences of the διáνοια/λέξις distinction occur at e.g. Lys. 10.7, Isoc. 9.11.

³¹ A later but interestingly explicit formulation of such an interpretative principle is offered by ps.-Long. Subl. 4.7, criticizing Herodotus for revealing mikropsuchia in the sentiments he gives to the Persians at 5.18.

³² But see the last part of my text for a possible defence of one of the citations found in this section of the work.

³³ ‘Making’ (ποιειῖν) is of course the standard term for creating/composing poetry, but it serves to convey implicit responsibility in such passages as Rep. 3.388a6, b3–4, 8, 390a8. For poets ‘saying’ or ‘telling’ (λέγειν), see e.g. 2.377d4–6, e7, 379d2, 380al; and cf. Ion 531a5–6 (Homer and Hesiod sometimes ‘say the same things’). A passage which reveals that such ‘saying’ can be a matter of conveying a (supposedly) implicit moral/message is Rep. 2.378b2–5, where the verb λέλειν shades from ‘telling', i.e. reciting, a poetic story into ‘asserting’ and endorsing its point. The technique of paraphrase later known as metathesis (cf. Greenbert, N., TAPA 89 [1958], 262–70)Google Scholar was designed in part to identify the logos within a text: note Plato's allusion to such techniques at Rep. 10.601b2 (with my note ad loc).

³⁴ See ‘Plato and the psychology of drama’ (n. 25 above), 57–8, and ‘The Republic's two critiques of poetry’, in Höffe, O. (ed.), Platon: Politeia (Berlin, 1997), 313–32Google Scholar, at 317–22. Another recent discussion of these issues, which emphasizes Plato's lack of a determinate concept of fiction, is C. Gill, ‘Plato on falsehood—not fiction’, in Gill, C. and Wiseman, P. (edd.), Lies and Fiction in the Ancient World (Exeter, 1993), 38–87.CrossRefGoogle Scholar

³⁵ See Rep. 2.378d6 (with Xen. Symp. 3.6) for the idea of poetic ὑπóινοια or ‘(allegorically) concealed meaning’: this is not a sufficient defence against the dangers of poetry to the young, but Plato does not rule it out as a more general hermeneutic strategy. Cf. the references to poetic ‘double meaning’ (αἰνíττεσθαι) at Rep. 1.332b9 (ironic), Tht. 194c8.

³⁶ Phaedo 61e2 seems to use μυθολογεῖν, to ‘tell stories’ or ‘mythologize’, in anticipation of the work's dialectical discussion of the soul's destiny after death, not just its concluding ‘myth’. In any case, the dialogue as a whole patently combines things which can be called logos or muthos. Poetry, by the same token, is said to comprise logoi at Grg. 502c.

³⁷ For what poets ‘say’ (λ⋯γειν), see n. 33. It is interesting that Isoc. 11.38, 40 refers to poetic mythology as λóγους…εἱπεῖν in a context where he voices moralistic criticism of the kind developed by Socrates in Rep. 2–3.

³⁸ For the motif of poetry speaking about ‘the greatest things', see Prt. 347a2, Rep. 10.599c7.

³⁹ A recent discussion, citing earlier scholarship, is G. W Most, ‘Simonides’ ode to Scopas in contexts’, in I. J. F. De, Jong and Sullivan, J. P. (edd.), Modern Critical Theory and Classical Literature (Leiden, 1994), 127–52Google Scholar: Most contends that literary interpretation calls for both internal and external contextualization.

⁴⁰ On the immediate point of self-contradiction, see n. 28 above.

⁴¹ Socrates’ exegesis of ‘punctuation’ at 343d6 ff., 345e ff., 346e, corresponds to the method of solving problems listed at Arist. Poet. 25.1461a23–5; the verb διαλαμβáνειν (346e2–3), for ‘dividing’ up word-groupings in or between clauses, corresponds to διαστíζειν at Arist. Rhet. 3.5.6,1407b13. Socrates’ interpretation of a particle at 344a4 ff. is another kind of argument from lexis (Poet. 1461a9 ff.): for particles as part of lexis see Rhet. 3.5, 1407a19 ff.

⁴² Compare Laws 1.629b-e, where the Athenian imagines his companions questioning Tyrtaeus about his views. This hypothetically contravenes the idea that poets cannot be questioned about their meaning (cf. nn. 47–8).

⁴³ The ‘as if’ of b6 refers not to the sense of λέγε, which we have seen applied to poetry elsewhere (n. 33 above), but to the hypothetical notion of Simonides making his utterance in person to Pittacus.

⁴⁴ Note that at Prt. 344b4 Simonides’ response to Pittacus is itself described as elenchos: this fits in with the characterization of the poem as a rebuttal of Pittacus's supposed wisdom; but it also helps in due course to heighten the force of Socrates’ remark at 347e3–7 (see my text).

⁴⁵ One should note here that the verb νοεῖν can be used of the meaning of words without necessarily implying conscious intention on the part of a speaker: see e.g. Crat. 397e2–3, with the sophistry of Euthd. 287c-e.

⁴⁶ The verb ༐πóγεσθαι at 347e4 means to ‘introduce’ as a witness: cf. Lysis 215c7, Hp. Mj. 289M, Rep. 2.364c5, Arist. Met. 995a8, with n. 19 above. The judicial metaphors of Prt. 347e probably presuppose the older Athenian practice, which prevailed until c. 380 (when it was replaced by written depositions), of requiring witnesses to testify orally in court and to be open to cross-examination.

⁴⁷ See also Hp. Min. 365c-d, with the comments of M. Whitlock Blundell, in Klagge, J. C. and Smith, N. D. (edd.), Methods of Interpreting Plato and his Dialogues, OSAP Suppl. Vol. (Oxford, 1992), 151Google Scholar, 167–8. Cf. n. 42 above.

⁴⁸ Contemporary poets are, of course, the exception, as Apol. 22a-c famously shows: Socrates asked the poets ‘what they meant’ or ‘what they were saying’ (λέγειν) in their works. Their inability to withstand this elenchos is there used to ground the inference that poets create by inspiration, not by knowledge or ‘wisdom’(sophia).

⁴⁹ Brandwood omits the passage from his index (see n. 4 above). Cf. Page PMG 642 (Simon, fr. 137)

⁵⁰ θεῖοσ, ‘godlike', is elsewhere used of poets/poetry at e.g. Ion 530b10, Meno 81b2, 99c, Phaedo 95a2, Laws 1.629b9, 3.682a3: in some of these contexts inspiration is implied, but this does not seem to be invariably the case.

⁵¹ Cf. the verb διανοεῖσθαι at Rep. 1.332cl, with n. 30 above.

⁵² There is an interesting sense in which Plato's construal of the relationship between philosophy and poetic ‘testimony’ may reflect the relative importance of litigants and witnesses in the Athenian judicial system: for the place of witnesses in that system, see esp. S. C. Todd, ‘The purpose of evidence in Athenian courts’, in Cartledge, P.et al. (edd.), Nomos: Essays in Athenian Law, Politics and Society (Cambridge, 1990), 19–39Google Scholar, at 23–31.

⁵³ For more detail on Rep. 468c-9a, see the notes in my edition (n. 17 above).

⁵⁴ See n. 50 above.

⁵⁵ I leave aside the further, and far from rhetorical, consideration that it might be thought psychologically implausible to accuse a supreme writer of fictional dialogues of an under-developed sense of narrative and dramatic form.

⁵⁶ It is highly likely that this inclination had already established strongly ‘anti-contextualist’ habits in general Greek attitudes to poetry, especially Homer, before Plato's time: Ford, A., ‘The inland ship: problems in the performance and reception of Homeric epic’, in Bakker, E. and Kahane, A. (edd.), Written Voices, Spoken Signs (Cambridge, MA, 1997), 83–109Google Scholar, at 91–101, assembles evidence for the archaic and early classical practice (sometimes on the part of poets themselves!) of treating the Homeric poems as sources of self-standing utterances rather than as larger artistic wholes.

⁵⁷ But Aristotle's relationship to Plato is not straightforward, since some of this chapter's considerations are anticipated in Plato's own writings; see nn. 28,41 above.

⁵⁸ See my ‘The importance of Plato and Aristotle for aesthetics’, Proceedings of the Boston Area Colloquium in Ancient Philosophy v [1989] (Lanham, MD, 1991), 321–8, at 339–7.

⁵⁹ Even so, it is interesting that a modern critic working with strongly contextualist methods can still arrive at a concept of authorial responsibility which has much in common with the one I have attributed to Plato: see the (moral) notion of the ‘implicit author’ used in Booth's, WayneThe Company We Keep: An Ethics of Fiction (Berkeley, 1988)Google Scholar, whose motivation I would describe, however oxymoronically, as a kind of ‘liberal Platonism', but which can also be said to balance considerations of muthos and logos much more finely than Plato seems interested in doing.

⁶⁰ The quotation is from Aesch. fr. 154a.15–16 Radt (fr. 273 Mette); the most recent text is that of Diggle, J. (ed.), Tragicorum Graecorum Fragmenta Selecta (Oxford, 1998), 20–1.Google Scholar Plato is convicted of unfairness by Fraenkel, E., PBA 28 (1942), 239Google Scholar, and the point is repeated by Lloyd-Jones, H., The Justice of Zeus (Berkeley, 1983), 207Google Scholar, n. 70 (where the reference to Fraenkel is doubly inaccurate).

⁶¹ It is a revealing coincidence that at Men. Aspis 412–13 Daos quotes exactly the same words from the Niobe: that Daos is here striking a ludicrously overwrought ‘tragic’ attitude does not detract from the reflection of a cultural mentality which can find imposing thoughts encapsulated (as gnomai, cf. γνωμολοεῖς, Aspis 414, with n. 18 above) in a poetic text.

⁶² Of later ancient readers (who obviously cannot be documented in detail here), Plutarch would make the most interesting case-study within the Platonic tradition itself (cf. Kidd, I. G., ‘Plutarch and his Stoic contradictions', in Burkert, W.et al. [edd.], Fragmentsammlungen philosophischer Texte der Antike [Göttingen, 1998], 288–302Google Scholar, for Plutarch's treatment of philosophical quotations): although he follows Platonic patterns in many respects, even taking over poetic quotations sometimes from the text of Plato, his De aud. poet, makes a number of modifications to the idea that a character's words can be equated with the poet's own views, esp. in the comments on characterization at 18b-f For an especially interesting Stoic case in which we can trace the importance of quotation within a moral view of literature, see Rutherford, R. B., The Meditations of Marcus Aurelius: A Study (Oxford, 1989), 26–33.Google Scholar The fullest and most wide-ranging study of an individual ancient reader is Stock, B., Augustine the Reader (Cambridge, MA, 1998)Google Scholar; and for the evolution of readings of a single author, see Lamberton, R. and Keaney, J. J. (edd.), Homer's Ancient Readers (Princeton, 1992). But there is much more work to be done on what methods and patterns of quotation can reveal about habits of reading.Google Scholar