¹ Nietzsche's recognition in Birth of Tragedy (ch.14) is grudging – contrast Bakhtin, M., ‘Epic and the novel’, in The Dialogic Imagination (Austin 1981)Google Scholar, and Rohde, E., Der griechische Roman und seine Vorläufer (Leipzig 1876)Google Scholar. These conflicting perspectives all derive from F. Schlegel, Kritische Ausgabe II, Lyceum, 26: ‘Novels are the Socratic dialogues of our day.’ The influence dialogue exerted on the formation of fiction in sixteenth-century Italy, examined by Snyder, J., Writing the Scene of Speaking (Stanford 1989)Google Scholar, is analogous.

² Existing treatments including Stewart, J., The Myths of Plato (London 1905)Google Scholar; Gaiser, K., Platone come scrittore filosofico (Naples 1984)Google Scholar; Gill, C., ‘Plato on falsehood – not fiction’, in Gill, C. and Wiseman, T.P. (eds.), Lies and Fiction in the Ancient World (Exeter 1993) 38–87CrossRefGoogle Scholar; Nightingale, A. Wilson, Genres in Dialogue (Cambridge 1995)CrossRefGoogle Scholar, and now Morgan, K., Myth and Philosophy from the Presocratics to Plato (Cambridge 2000)CrossRefGoogle Scholar, show how variously the philosophical use of fiction in Plato can be interpreted.

³ Cicero, De Finibus 2 is a parallel critique of the Epicurean view of the ethical issues raised by the story of Gyges, but does not engage with the story itself. Vander Waerdt, P.A., ‘The justice of the Epicurean Wise Man’, CQ 37 (1987) 402–22CrossRefGoogle Scholar, argues that Epicurus has been misrepresented by such criticism and offers a reconstruction of his position.

⁴ R. Waterfield's translation of the Republic (Oxford 1993) of δικαιοσύνη as ‘morality’ might better signal the debate's relevance to modern readers, conveying something broader than the English ‘justice’: see e.g. Arist. Nic. Eth. 5.1129a-34; Vlastos, G., ‘The theory of social justice in the Republic’, in North, H. (ed.), Interpretations of Plato (Leiden 1977) 1–40Google Scholar, discusses the problems. But ‘morality’ is not really a category pertinent to Plato and certainly not an ἀρετή (as δικαιοσύνη is) and ‘justice’ no less conveys the community of the Platonic debate with the concerns of Panaetius and Cicero set out in §111 below.

⁵ The opening chapters of Herodotus – which may have been in oral circulation – have long been deemed to be parodied in Aristophanes’ Acharnians (425 BC): see (e.g.) How, W.W. and Wells, J., Commentary on Herodotus 1 (Oxford 1912) 448Google Scholar. Asheri, D., Erodoto: Le storie 1 (Fond. Lorenzo Valla 1988) xvGoogle Scholar, dates Herodotus’ death at 430 BC or after. Pichler, R., Die Gygesgeschichte in der griechischen Literatur und ihre neuzeitliche Rezeption (diss. Munich 1986)Google Scholar treats circulation of the Gyges story in antiquity. Composition of the Rep. is generally set in the 370s BC; the dramatic dates in Plato can rarely be fixed with certainty.

⁶ See Adam, J., The Republic of Plato 1 (Cambridge 1902) 126–7Google Scholar; Asheri (n.5) 269 notes: ‘il tentativo a unire e integrare le versioni è vano’; Dyck, A., Commentary on Cicero De Officiis (Ann Arbor 1996) 539–40Google Scholar; Erbse, H., ‘Die Funktion der Novellen im Werke Herodots’, in Kurz, G., Müller, D. and Nicolai, W. (eds.), Gnomosyne: Festschrift für Walter Marg (Munich 1981) 251–69Google Scholar; Meyer, E., Geschichte der Altertums 3 (Stuttgart 1937) 133 n.1Google Scholar; ‘Gyges und sein Ring’, in K. Reinhardt, Vermächtnis der Antike (1966) for an interpretation of the stories in Herodotus and Plato; Stein, H., Herodotus 1 (Berlin 1893)Google Scholar.

⁷ Smith, K.F., ‘The Tale of Gyges and the King of Lydia’, AJP 23 (1902) 261-82, 361–87, at 263Google Scholar; cf. Pedley, J.G., Literary Sources on Sardis (Harvard 1972) 16–18CrossRefGoogle Scholar.

⁸ The sixth book of Nicolaus Damascenus' Universal History, which reputedly draws from the Lydian logographer Xanthus, is preserved in Porphyrogennetus’ abstract – cf. Krumbacher, K., Geschichte der byzantinischen Literatur (Munich 1897) 252Google Scholar, and Hanfmann, G., ‘Lydiaka’, in HSCP 63 (1958) 65–88Google Scholar. Ptolemaeus Chennus' Καινὴ Ἱστορία is abstracted by Photius. Philostratus' Apollonius 3.2 refers to the dragon stone (found inside a dragon's head) which is ‘invincible even against the ring they say Gyges possessed’.

⁹ Smith (n.7) 383-5 provides the full version epitomized here. Again in Damascenus' account (which very roughly resembles Herodotus), Gyges kills the king Sadyattes because he has fallen in love with his queen.

¹⁰ Smith (n. 7) 268 n.2. Cf. the passages assembled in Slings, S., ‘Critical notes on Plato's Politeia, II’, Mnemosyne 42.3-4 (1989) 382Google Scholar.

¹¹ As well as Asheri, Stein and Meyer (n.6), see Gould, J., Herodotus (New York 1989) 19–41Google Scholar.

¹² Dyck (n.6) 539-40 notes: ‘It has long been recognized that Plato's folktale version… is original and Herodotus' version a secondary rationalization…’

¹³ The significance of the Lydia connection in relation to the Myth of Er is suggested below (§II); ‘Gyges’ is discussed here.

¹⁴ Proclus refers to the ‘narrative about Gyges' ancestor’ (τῶι καιὰ τὸν Γύγου πρόγονον διηγήματι) – the text is in Kroll, W. (ed.), Procli Diadochi in Platonis Rem Publicam Commentarii 1 (Leipzig 1901) 111Google Scholar.

¹⁵ Perhaps the best justification comes from Wiegand, , ‘Aehrenlese der Kritik und Erklärung der sieben ersten Bücher des platonischen Staats’, Zeitschrift für die Altertumswissenschaft 107 (1834) 863Google Scholar, who suggests (on the basis of Cicero, De officiis 3.38 as well as Rep. 612b) that Γύγου in 359d is a gloss: the reference there would then be to Gyges as an (unnamed) ancestor of the Lydian [sc. Croesus]. Cf. Smith (n.7) in the reconstruction summarized above. Others prefer the recentiores: Γύγηι τῶι for τῶι Γύγου. Although Γύγηι τῶι Λυδῶι, offered as a ‘trial balloon’ by Slings (n.10) 381-3, follows a salutary discussion, the Proclan reading may still be compatible with what Socrates says. See discussion of 612b in §11 below.

¹⁶ Frutiger, P., Les mythes de Platon (Paris 1930) 235Google Scholar. Note how even here Frutiger has lapsed into regarding Gyges and not Gyges' ancestor as Glaucon's protagonist.

¹⁷ Frutiger (n.16) 235 n.2. An explanation for this remark in 612b will be offered below (§11).

¹⁸ The note on Alexander Aetolus fr. 9, 5-6: καὶ Μούσας ἐδάην Ἑλληνίδας, αἵ με τυράννων / θῆκαν Δασκύλεω κρείσσονα καὶ Γύγεω (= Plut. De exilio 599e) in J. Powell, Collectanea Alexandrina (1925) 127, confirms that this later testimony has no bearing on the case made here.

¹⁹ Erbse (n.6) treats the role of the Gyges episode in the grander scheme of Herodotus’ project; Schubert, P., ‘L'anneau de Gygès: réponse de Platon à Hérodote’, Antiquité Classique 66 (1997) 255–60CrossRefGoogle Scholar, uses his perception of the parallels to examine the implications of H.'s Lydian history for the formation of Plato's political theory in the Republic as a whole. See also Fowler, D.P., ‘On the shoulders of giants: intertextuality and classical studies’, MD 39 (1997) 13–34Google Scholar (= Fowler, D.P., Roman Constructions (Oxford 2000) 115–37)Google Scholar.

²⁰ ὀφθαλμοὶ γἁρ τῶν ὤτων ἀκριβέστεροι μάρτυρες, fr B101a (Diels-Kranz); cf. n.23 below.

²¹ For discussion of propriety in Hdt. 1.8.3, see Raubitschek, A.E., ‘Die schamlose Ehefrau’, Rh.Mus. 100 (1957) 139–40Google Scholar = Raubitschek, , School of Hellas (Oxford 1991) 330–1Google Scholar.

²² Herodotus and Plato use different words for seeing; moreover Plato's ἀφανής and φανερός do not occur in Herodotus. Shell, M., The Economy of Literature (Baltimore 1978) 30–6Google Scholar, discusses this, pointing out an association in Greek thought between invisibility and tyranny. To supplement Shell's cross-cultural comparisons: the New York Times for 20 April 1999 reported that the renovated Reichstag was ‘topped with a glass dome to symbolize the political transparency on which Germany has based its post-war revival’.

²³ Prodicus' myth of Heracles at the crossroads is related in Xenophon, Memorabilia 2.1.20-4 (fr. 84B2 Diels-Kranz). See Untersteiner, M., The Sophists (New York 1953) 216–21Google Scholar. The road as life figures in Rep. 328e2 and 364d: see Adam (n.6) ad loc. for comparison with Hesiod, Works and Days.

²⁴ In addition to Smith (n.7), see Bernhardy, G., Suidas' Lexicon (Halle 1853)Google Scholar s.v. Γύγου δακτύλιοςFauth, W., ‘Zum Motivsbestand der platonischen Gygeslegende’, Rh.Mus. 113 (1970) 1–42, fails to prove it was not Plato's invention.Google Scholar

²⁵ The etymologies in Plato's Cratylus (e.g. 437b) often depend on what Ahl, F., Metaformations (Cornell 1985) 54Google Scholar, calls ‘anagrammatic rearrangements’. Ahl discusses the wordplays in Cratylus 439c at 286; Cratylus 405d-e at 129 n.; Plato, Minos 315c at 73-4, and Rep. 620a at 190-1. Sedley, D., ‘The etymologies in Plato's Cratylus’, JHS 118 (1998) 140–54CrossRefGoogle Scholar, provides an important defence of Plato's etymologies. Many involve words with fewer sonic or literal elements in common than there are between Δασκύλος and δακτύλιος.

²⁶ Derrida, J., ‘La pharmacie de Platon’, in Dissémination (Paris 1972) 69–197Google Scholar. The significance of the juxtaposition of Διός and δῖον in Phaedrus 252e is discussed in Nussbaum, M., The Fragility of Goodness (Cambridge 1986) 228–9Google Scholar.

²⁷ The earnest response of Plato's ancient readers to his etymologies would support this. The art of etymology, Sedley (n.25) points out at 143, was ‘an exercise, not in linguistic science, but in the recovery of ancient thought’.

²⁸ For instance, the possibility that Critias the elder referred to in Timaeus 20e might be Plato's great-grandfather has been recognized by K. Morgan, ‘Designer history: Plato's Atlantis story and fourth-century ideology’, JHS (1998) 118, at 101-2 n.3, and Osborne, C., ‘Creative discourse in the Timaeus’, in Gill, C. and McCabe, M.M. (eds.), Form and Argument in Late Plato (Oxford 1996) 179–211, at 182 n.8Google Scholar.

²⁹ Phaedrus' accusation that Socrates makes up his stories from Egypt and elsewhere (ῥαιδίως σὺ Αἰγυπτίους καὶ ὁποδαποὺς ἂν ἐθέληις λόγους ποιεῖς) in Phaedr. 275b3 is pertinent: Socrates playfully (?) replies that the provenance of a story (τίς ὁ λέγων καὶ ποδαπός) is not as important as its truthfulness. On Egyptian ‘authority’, see Gill, C., ‘Plato's Atlantis story and the birth of fiction’, Ph&Lit. 3 (1979) 75Google Scholar, and again Morgan (n.28) 104, 110. This exchange in the Phaedrus confirms my view of the hopelessness of trying to answer an interesting question put to me by Greg Woolf: does Glaucon's tale give any indication of Plato's conception of Herodotus’ truthfulness?

³⁰ Such manipulation is a recurrent feature of our own category of fiction. Examples abound of ‘original’ works conceived after models: John Fuller, Flying to Nowhere epitomizes The Name of the Rose; the title of Ulysses signals a relation to the Odyssey; Kate Chopin, The Awakening, Fernando Alas, La Regenta and Henry James, Portrait of a Lady emulate Madame Bovary.

³¹ Compare the observations on Plato's story in Tatum, J., Xenophon's Imperial Fiction: On the Education of Cyrus (Princeton 1989) 266Google Scholar.

³² Cf. Dover, K. (ed.), Plato: Symposium (Cambridge 1980)Google Scholar 80 on this practice. Brisson, L., Plato the Mythmaker (Chicago 1998) 149–51Google Scholar, usefully lists occurrences of the declarative μυθολογέω. Osborne (n.28) 183 notes the number of such declarative verbs in the Timaeus which suggest verbal (oral) narration.

³³ See Laird, A., Powers of Expression, Expressions of Power (Oxford 1999) 76–8Google Scholar, on this feature in Platonic dialogue.

³⁴ Diog. Laert. 1.12 recounts Pythagoras was the first to use the term philosophia (love of wisdom) because no mortal can actually be wise. This suggests religious and ethical overtones to the word. Cicero's account of the same conversation with Leon (Tusc. Disp. 5.3.8) reports that Pythagoras also likened those engaging with philosophy to the audience of a spectacle (cf. the points to come and nn.37-9 below): ‘so there were a few rare people who counting all else as nothing studiously scanned the nature of things' (rerum naturam studiose intuerentur). Pythagoras' notion has been contrasted with Croesus’ neutral use of the word in Hdt. 1.30, which seems to be about acquiring knowledge. Croesus juxtaposes φιλοσοφέων with θεωρίη in that very passage.

³⁵ Socrates' view of the philosopher in Phaedo 62c-69e can be read as a development of the usage in Hdt. 1.30.

³⁶ Rutherford, R., The Art of Plato (London 1995)Google Scholar, and Kahn, C., Plato and the Socratic Dialogue (Cambridge 1998)Google Scholar, demonstrate the importance of literary artistry for appreciation of Plato's philosophy; the essays in Griswold, C. (ed.), Platonic Writings, Platonic Readings (London 1988)Google Scholar, point out the drawbacks of divorcing literary and philosophical readings of Plato. Nightingale (n.2) goes further in showing Plato's role in transforming philosophy into a genre.

³⁷ Consider Annas, J., An Introduction to Plato's Republic (Oxford 1982) 1Google Scholar: ‘[Plato] is too good a philosopher not to raise difficult and important philosophical issues…’!

³⁸ ‘Thought experiment’ is the translation of Rep. by R. Waterfield (n.4). Rescher, N., ‘Thought experimentation in Pre-Socratic Philosophy’, in Horowitz, T. and Massey, G. (eds.), Thought Experiments in Science and Philosophy (Maryland 1991) 31Google Scholar, defines the term as ‘an attempt to draw instruction from a process of hypothetical reasoning that proceeds by eliciting the consequences of an hypothesis, which for aught that one actually knows to the contrary, may well be false. It consists in reasoning from a supposition that is not accepted as true – perhaps is even known to be false – but is assumed provisionally in the the interests of making a point or resolving a conclusion.’ There are further accounts in Horowitz and Massey; see also Nails, D., Agora, Academy and the Conduct of Philosophy (Philosophical Studies Ser. 63) (Dordrecht 1995)Google Scholar on the Derveni papyrus.

³⁹ There are many discussions: see e.g. Halliwell, S., Aristotle's Poetics (London 1986) 9, 56–9.Google ScholarFord, A., Early Greek Terms for Poetry: Aoide, Epos, Poiesis (diss. Yale 1981)Google Scholar, shows usage of ποιητής for poet (as distinct from ἀοιδός) converged with the gradual conception of poetic representation as a τέχνη or professional art.

⁴⁰ Contrast e.g. Rep. 402d4 and 480a where the verb has a more innocent sense of ‘gaze’ with the curious wordplay of Rep 511c8: ἀναγκάζονται… θεᾶσθαι οί θεώμενοι, replayed in 611c (for contemplation of the soul).

⁴¹ See e.g. Ar. Frogs 2 and Clouds 518, and Hdt. 8.116. Work is forthcoming from Andrea Wilson Nightingale on Platonic theoria, and from Ian Rutherford on theoria in Greek culture and religion.

⁴² Mitchell, W.J.T., Iconology: Image, Text, Ideology (Chicago 1985) 5-6, 158Google Scholar.

⁴³ Ad Herennium 3.16-24; Cic. De oratore 2.86.3516; Quint. Inst. Or. 11.2.17-22. Cf. Blum, H., Die Antike Mnemotechnik (Hildesheim and New York 1969)Google Scholar; the articles entitled ‘Memory and the study of classical antiquity’ by Small, J.P. and Tatum, J. in Helios 22.2 (1995) 149–77Google Scholar, with bibliography at 174-7; Small, J.P., Wax Tablets of the Mind: Cognitive Studies of Memory and Literacy in Classical Antiquity (London 1997)CrossRefGoogle Scholar; on the Cave, see Gaiser, K., Il paragone della caverna: variazioni da Platone a oggi (Naples 1985).Google Scholar

⁴⁴ See Zeitlin, F., “The poetics of Eros: nature, art and imitation in Longus' Daphnis and Chloe’, in Halperin, D. et al. , (eds.), Before Sexuality (Princeton 1990) 417–64Google Scholar. On meta-ekphrasis in Horace, Ars poetica 14-19, see my discussion in Eisner, J. (ed.), Art and Text in Roman Culture (Cambridge 1996) 91–4Google Scholar. Embedded openings in Greek fiction recall the beginnings of Platonic dialogue (cf. n.33 above and Perry, B.E., The Ancient Romance (Berkeley and Los Angeles 1967) 325)Google Scholar.

⁴⁵ Annas (n.37) 69.

⁴⁶ E.g. Adam (n.6) 126-7; Frutiger (n.16) 235. This passage along with Cic. De officiis 35 was the basis for Wiegand's conjecture (n.15 above).

⁴⁷ See Leaf, W. (ed.), The Iliad 1 (Amsterdam 1971) 251Google Scholar: ‘The name Ἀΐδησ here evidently preserves something of its original sense, the Invisible (Ἀƒίδησ).’ Contrast Socrates in Cratylus 404b: ‘And the name “Hades”, Hermogenes, is not at all derived from the invisible (ἀιδοῦς), but far more likely from knowing (εἰδέναι) all fine things.’

⁴⁸ The importance of characterization is noted in two recent studies: Kahn (n.36) and Beversluis, J., Cross Examining Socrates: A Defence of the Interlocutors of Plato Early Dialogues (Cambridge 2000)Google Scholar.

⁴⁹ For the idiom, Shorey, P. (ed.), The Republic 2 (Cambridge, MA 1935) 485 nGoogle Scholar. compares εἰ μὴ ἀδικῶ earlier at 608d, and on 608d ad loc., compares Rep. 430e, Charmides 156a and Menexenus 236b. In Rep. 430e the wordplay expressly features in a discussion of justice (δικαιοσύνη).

⁵⁰ The scholia to Lucian, Ver. Hist. Praef. 1.3-4 interpret the passage as a retort to Plato's use of myth in Rep. 10: 614a. Cf. Macrob. In Som. Scip. 1.1.8-2.5 for Colotes' attack on Plato: a philosopho fabulam non oportuisse confingi.

⁵¹ The Greek title περὶ τοῦ καθήκοντος (‘On appropriate action’) corresponds to De officiis. Dyck (n.6) 17-28 gives a valuable assessment of the relation between Panaetius' lost treatise and Cicero's work.

⁵² For a standard accounts of focalization and its relation to ‘point to view’, see Genette, G., Narrative Discourse (Oxford 1980) 161–211Google Scholar. Genette made the important distinction for narrative theory between ‘who speaks’ and ‘who sees’. Hornblower, S., Greek Historiography (Oxford 1994) 131–6Google Scholar, and Gribble, D., ‘Narrator inventions in Thucydides’, JHS 118 (1998) 4167CrossRefGoogle Scholar, consider focalization in Thucydides.

⁵³ Abel, K., ‘100 Jahre hekaton-Forschung’, Würzburger Jahrbücher N.F. 13 (1987) 111Google Scholar, concurs that Cicero drew directly from Plato for his account of the story rather than from any intermediate source. Cicero's ut ferunt fabulae as a translation of μυθολογοῦσι confirms this.

⁵⁴ Dyck (n.6) 541 has a less favourable view of Cicero's manipulation of Plato's version here: ‘Cicero inserts this piece of information only at the point where it becomes vital to the story to explain Gyges' connection with the court, whereas both Herodotus and Plato had more artfully explained Gyges' position upon first mention of his name…’

⁵⁵ Numerous studies of fictional narrative such as Cohn, D., Transparent Minds (Princeton 1978)Google Scholar, and Genette (n.52) provide circumstantial evidence to bear this out; accounts like Riffaterre, M., Fictional Truth (Baltimore 1990)Google Scholar, and Cohn, D., The Distinction of Fiction (Baltimore 1999)Google Scholar, are more specifically concerned with identifying actual formal features of fiction.

⁵⁶ Ancient narrative practice bears this out even if ancient literary criticism is reticent about it: epic and Greek prose romance employ soliloquy, focalization and indirect discourse to present characters’ thoughts in contrast to historiography and ‘factual’ reportage. Cf. Hågg, T., Narrative Techniques in the Early Greek Romances (Stockholm 1971), and Laird (n.33) 102–10Google Scholar.

⁵⁷ This criterion for distinguishing fictional from factual narrative is meant to complement the more traditional kinds set out in Gribble (n.52) 49-50, based on responses and expectations of readers and hearers.

⁵⁸ Compare Cic. De Amicitia 1.3: Quasi enim induxi loquenter ne inquam et inquit saepius interponerentur, atque ut tanquam a praesentibus coram haberi sermo uideretur (‘I have as it were brought the speakers on stage in person, avoiding a frequent insertion of “I said” and “he said”, so that the conversation might seem to be held by people as if they were actually present.’). Powell, J.G.F. (ed.), Cicero: On Friendship and the Dream of Scipio (Warminster 1990) 78Google Scholar, notes that induxi (‘I brought [the characters] on [stage]’) is a theatrical word.

⁵⁹ Shell (n.22) 14 n., who at 11-88 offers a Marxist allegorization of the Gyges story, lists writers (including Rousseau) who treat the myth: Hans Sachs, Montaigne, La Fontaine, Friedrich Hebbel, Quevedo y Villegas, Théophile Gautier, Addison, Beaumont and Fletcher, Hugo von Hoffmanstal and André Gide. More recently, Mario Vargas-Llosa and Frederic Raphael have provided novelistic retellings with elements from Herodotus as well as Plato.

⁶⁰ Perry (n.44) has been influential in eschewing the philosophical dimensions of Greek romances; their interpretation as religious texts by (e.g.) Kerényi, K., Die griechische–orientalische Romanliteratur (Tübingen 1927)Google Scholar, and Merkelbach, R., Roman und Mysterium (Munich and Berlin 1962)Google Scholar, has enjoyed far more currency.

⁶¹ For a translated selection of fragments of Aristotle's dialogues, see Barnes, J. (ed.), Complete Works of Aristotle 2 (Princeton 1984) 2389–426Google Scholar. An Aristotelian ‘myth’ in which Silenus converses with Midas – fr 44 in Rose, V. (ed.), Aristotelis Fragmenta (Leipzig 1886)Google Scholar – is better termed ‘dialogical fiction’. Ar. Poetics 9.1451b5 (διὸ καὶ φιλοσοφώτερον καὶ σπουδαιότερον ποίησις ἱστορίας ἐστίν) is a classic articulation of the philosophical value of fiction.

⁶² Examples range from Boethius to Thomas More and Nietzsche. T. Nagel, What is it Like to be a Bat? will appear fictional to the ‘man on the Clapham omnibus’, who risks being fictional himself. As well as thought experiments (n.37 above), theories of ‘possible worlds’ involve fabrication; whilst Vaihinger, H., The Philosophy of As-If (London 1924)Google Scholar, suggests that traditional discourses of philosophy are themselves ‘fictions’. Richard Rorty's neo-pragmatism can be seen as a more recent version of this position.