¹ cf., for example, Kenney, E. J., Lucretius – De rerum natura Book 3 (1971),Google Scholar at 30: ‘Of all the books of the D.R.N., Book III appears to be the most highly finished, neatly constructed, and the best able to stand on its own.’

² cf. Kenney, op. cit. (n. 1), at 222 on 11. 978–1023.

³ On the complexity of the notion of fear of death see Striker, G., ‘Commentary on Mitsis’, Proceedings of the Boston Area Colloquium for Ancient Philosophy 4 (1988), 323–8CrossRefGoogle Scholar.

⁴ cf. their editions and commentaries on this section (1896–8 and 1947 respectively).

⁵ Rand, E. K., ‘La composition rhétorique du troisième livre de Lucréce’, RPh 60 (1934), 243–66;Google ScholarRambaux, C., ‘La logique de l'argumentation dans le De Rerum Natura’, REL 58 (1980), 201–19Google Scholar.

⁶ I discuss this passage at length in my ‘The speech of Nature in Lucretius' DRN 3931–971’, CQ 52 (2002), 291–304;Google Scholar on the soul-vessel metaphor in this speech and its use to convey the idea of Epicurean ‘katastematic pleasure’ see ibid, and below.

⁷ See Heumann, H. G. and Seckel, E. (eds), Handlexikon zu den Quellen des römischen Rechts (9th edn, 1906)Google Scholar, s.v. lis no. 1, give as its meaning ‘Rechtsstreit, Prozess in Zivilsachen, bürgerliche Rechtsstreitigkeit’; but cf. Mommsen, Th., Römisches Strafrecht (1899), 392 n.4.Google Scholar

⁸ cf. Kaser, M. and Hackl, K., Das römische Zivilprozessrecht (1996), at 60Google Scholar: ‘Als Parteien stehen sich der Kläger (actor, petitor, is qui agit) und der Beklagte (reus, is cum quo agitur, is a quo agitur) regelmässig mit kontradiktorischen Behauptungen gegenüber. Bei der streiteinsetzenden legis actio in personam behauptet der Kläger ein Recht zum Zugriff auf die Person des Beklagten aus dessen Haftung; der Beklagte bestreitet dieses Recht. Auch bei der actio in rem schliessen sich die Behauptungen der Parteien uber das Recht, den Streitgegenstand für sich haben zu dürfen, gegenseitig aus. Da bei all diesen Klagen der Kläger etwas vom Beklagten oder eine Partei etwas von der anderen begehrt, kann der Sachentscheid nur dahin lauten, ob dieses Begehren berechtigt oder unberechtigt ist. Imersten Fall wird der Kläger – bei der actio in rem einem Prätendenten – ein Weg freigegeben, der zu seiner Befriedigung führt, im zweiten wird ihm der Weg versagt.’ It is compatible with this trial-like situation that Nature's words are characterized as invective (1. 932 increpet, 1. 963 incilet); a glance at Cicero's Pro Caecina may convey an impression of the tone of proceedings in Roman private trials of the period.

⁹ See the commentary on this section in Th. Kurth, Senecas Trostschrift an Polybius (1994), 124–30, and Schiesaro, A., ‘Lucrezio, Cicerone, l'oratoria’, MD 19 (1987), 29–61,Google Scholar at 60–1.

¹⁰ On the very Roman colouring of the descriptions of punishment see Jocelyn, H. D., ‘Lucretius, his copyists and the horrors of the Underworld’, AClass 29 (1986), 43–56Google Scholar.

¹¹ On this kind of narrative device, its use in the tradition of diatribe, and its rôle in the wider scheme of psychagogics see B. Wehner, Die Funktion der Dialogstruktur in Epiktets Diatriben (2000), at 79–105.

¹² cf. Conte, G. B., ‘Il trionfo della morte in Lucrezio’, SIFC 37 (1965), 114–32;Google Scholar C. Segal, Lucretius on Death and Anxiety (1990), ch. 8, at 180. The final section of Book 3 is full of arguments which were originally devised for consolatory contexts but then transformed by Lucretius so as to apply to other concerns, e.g. fear of one's own death. A convenient list of consolatory topics is in Scourfield, J. H. D., Consoling Heliodorus: a Commentary on Jerome, Lette 60 (1993), 253Google Scholar.

¹³ I suggested this in passing in my earlier article (op. cit. (n. 6)), at 300 n. 29.

¹⁴ A precedent in the epic tradition for such a ‘virtual katabasis’ (and perhaps one source of inspiration for Lucretius) is to be found in the Argonautica of Apollonius of Rhodes; in the Argonautica, the heroes do not undertake a katabasis, but rather their whole voyage to the Black Sea is likened to one through a wide range of devices: see R. L. Hunter, The Argonautica of Apollonius: Literary Studies (1993), 182–9; Kyriakou, P., ‘Katabasis and the underworld in the Argonautica of Apollonius of Rhodes’, Philologus 139 (1995), 256–64;CrossRefGoogle Scholar D. Nelis, Vergil's Aeneid and the Argonautica of Apollonius Rhodius (2001), 228–55. One crucial difference in De rerum natura is of course that here it is the reader (or so I argue) who undertakes the katabasis.

¹⁵ Hesiod was an important model for the philosophical didactic tradition from Parmenides onwards, for obvious reasons: he was the most prominent representative of non-philosophical attempts to explain the world in verse. Cf. Burkert, W., ‘Das Proömium des Parmenides und die Katabasis des Pythagoras’, Phronesis 14 (1969), 1–30,Google Scholar at 16–17; on the wider aspects of pre-Socratic philosophers positioning themselves within the didactic tradition see S. Broadie, ‘Rational theology’, in A. A. Long (ed.), The Cambridge Companion to Early Greek Philosophy (1999), 204–24.

¹⁶ Apart from the references already given, any poet who has read Aristophanes' Frogs may be tempted to place some shadowy creatures at the entrance to the underworld. That allegorical representations of Curse, Envy and Riot (and similar ‘creatures’) featured in pictorial representations of the underworld in the Hellenistic era emerges from [Dem.] 25.53; cf. also the description of Polygnotus' painting of the underworld in Paus. 10.25–31.12, especially the characters at the entrance to Hades (10.29.1), and Robert, C., Das Hadesbild des Polygnot, 16Google Scholar. Hallesches Winckelmannsprogramm (1892). Wöhrle, G., Hypnosder Allbezwinger (1995), at 34Google Scholar, points out that in statues Hypnos is sometimes represented as holding a jar filled with water from the underworld stream Lethe, and that the bat's wings he is sometimes said to have suggest an underworld association as well (e.g. Lucian, ver. hist. 2.32–3; cf. Od. 24.6–9 where the souls of the dead suitors are likened to bats; surely the etymology of νυκερίς plays a role here); on representations of Hypnos as a demon of death see ibid., 32. A very comprehensive survey of ancient underworld descriptions is Dieterich, A., Nekyia – Beiträge zur Erklärung der neuentdeckten Petrusapokalypse (1893)Google Scholar; see esp. 46–62 for a survey of minor characters located there.

¹⁷ See, for example, Clay, D., Lucretius and Epicurus (1983), at 87–95 and 109–10Google Scholar. Cf. Hardie, P., Vergil's Aeneid — Cosmos and Imperium (1986), at 196Google Scholar: ‘In Lucretius one may also distinguish internal and external aspects of inversion; the external inversion by which Epicurus brings man from oppression to supremacy is matched by the continuous upsetting of received ideas and hierarchies in the poem, a didactic device intended to lead the reader gradually from illusion to clear vision. Irony and allegory are the verbal equivalents of the narrative device of inversion.’

¹⁸ Burnyeat, M., ‘First words: a valedictory lecture’, PCPS 43 (1997), 1–20,Google Scholar at 5–8 shows that in his Republic Plato uses the katabasis motif to highlight a crucial aspect of his enterprise of outlining the ideal city, and that a reference to the Odyssey, which effectively makes the katabasis ‘Odyssean’, plays acrucial role in this. I believe this further strengthens the case for the possibility of an ‘Odyssean reading’ of the katabasis in De rerum natura 3, given the broader context of Lucretian interaction with Platonic dialogues (see below).

¹⁹ For full details cf. Gale, M., Myth and Poetry in Lucretius (1994), 119–25Google Scholar, who at p. 124 writes: ‘[the reader] … even makes a katabasis in 3.978–1023’, which is the only (partial) anticipation of the katabasis idea I could find. I sample some of the material Gale has collected. Passages connected with Epicurus ∼ Odysseus: (i) metaphor of the journey, imposed on the image of the ‘flight of the mind’: 1.62–79; (ii) 1.66 Epicurus introduced as Graius homo cf. ἀνὴρ πολύτροπος Od. 1.1–2, with names delayed until 3.1042 and 1.21, respectively; (iii) 6.8 ‘[Epicurus] cuius … iam ad caelum gloria fertur’, cf. Od. 8.74, οἴμης τῆς τότ' ἄρα κλέος οὐρανὸν εὐρὺν ἵκανε, and 9.20. Passages connected with reader ∼ Odysseus: (i) he is on a voyage towards the sapientum templa serena (2.8), guided by the poet and divine Epicurus. Cf. Buffière, F., Les Mythes d' Homère et la pensée grecque (1956) 365–91Google Scholar, for allegorical readings of the Odyssey as a spiritual voyage, from the late Hellenistic era onwards; (ii) Epicurus brings the reader rescue from the storm (5.8–12); cf. Ino in Od. 5.333ff.; (iii) various characters try to hinder the reader, e.g. a vates in 1.102 or quidam in 1.371; (iv) the reader is subject to many errores, cf. 1.332, 1.393, 2.82, 2.132, 3.105, 4.824 and passim; for Odysseus' wanderings as errores cf. Cic., Off. 1.113, Manil. 2.4 (e.g.). On a polemical association of Odysseus and Epicurus, made by opponents of the Epicureans, see Kaiser, E., ‘Odyssee Szenen als Topoi’, MH 21 (1964), 109–36 and 197–224,Google Scholar at 220–3; Gordon, P., ‘Phaeacian Dido: lost pleasures of an Epicurean intertext’, CA 17 (1998), 188–211Google Scholar.

²⁰ De rerum natura is pitched as a novice's introduction to Epicureanism. See Kleve, K., ‘What kind of work did Lucretius write?’, SO 54 (1979), 81–5;CrossRefGoogle Scholar G. B. Conte, Generi e lettori (1991), ch. 1.; Clay, op. cit. (n. 17), 212ff.; Mitsis, P., ‘Committing philosophy on the reader: didactic coercion and reader autonomy in de Rerum Natura’, in Mega Nepios. Il destinatario nell' epos didascalico, MD 31 (1993), 111–28Google Scholar.

²¹ Parmenides: cf. Burkert, op. cit. (n. 15), at 28: ‘Um nun aber nochmals auf Parmenides zurückzukommen: die Fahrt — auf dem Weg zur Sonne? —ins Jenseits zur geheimnisvollen Göttin, die Verkündigung der Göttin über die Wahrheit von Sein und Nichtsein, dies hat sein Vorbild nicht nur bei Hesiod, Epimenides, Sibylle, sondern gerade auch in der Katabasis des Demeter-Hierophanten Pythagoras, des Verkünders der Seelenwanderungslehre.’ Empedocles: frg. 110.2DK, which may lie behind De rerum natura 1.1114–17 (to be quoted below); and see P. Kingsley, Ancient Philosophy, Mystery, and Magic (1995), 230–2.

²² cf. Ep. ad Hdt. 36.83; SV 52; Metrod. frg. 37 (Clem. Alex., Strom. 5.138) and 38 (Plut., Adv. Col. 1117a); a full survey of the vocabulary of initiation in general is to be found in C. Riedweg, Mysterienterminologie bei Platon, Philon und Klemens von Alexandrien (1987).

²³ See Riedweg, op. cit. (n. 22), 17–21 and in particular Soph. 230b4 and Phd. 69b8; Schefer, C., ‘Platons Lysis als Mysterieneinweihung’, MH 58 (2001), 157–68Google Scholar.

²⁴ On the depiction on vases see pls 70–3 and the commentary in Bianchi, U., The Greek Mysteries, Iconography of Greek Religions xvii.3 (1976)Google Scholar.

²⁵ Graf, F., Eleusis und die orphische Dichtung Athens in vorhellenistischer Zeit (1974), 138 and 140Google Scholar.

²⁶ The evidence is collected in Riedweg, op. cit. (n. 22), 122–3 with nn. 32–3; see also idem, ‘Die Mysterien von Eleusis in rhetorisch geprägten Texten des 2./3. Jahrhunderts nach Christus’, ICS 13 (1988), 127–33.

²⁷ R. Merkelbach, Roman und Mysterium in der Antike (1962), esp. 37–53, with the review by Turcan, R., ‘Le roman initiatique: à propos d'un livre recent’, RHR 163 (1963), 149–99;CrossRefGoogle Scholar see also R. Beck, ‘Mystery religions, aretalogy and the ancient novel’, in G. Schmeling (ed.), The Novel in the Ancient World (1996), 131–50.

²⁸ Fauth, W., ‘Divus Epicurus: zur Problemgeschichte philosophischer Religiosität bei Lukrez’, ANRW 1.4 (1973), 205–25,Google Scholar esp. 220–4; Gale, op. cit. (n. 19), 193–6; D. Fowler, ‘The didactic plot’, in M. Depew and D. Obbink (eds), Matrices of Genre: Authors, Canons, and Society (2000), 205–19 and 299–302, at 212–17. On the subject of mystery language in Lucretius in general see E. Norden, Agnostos Theos: Untersuchungen zur Formengeschichte religiöser Rede (1913), 100–1.

²⁹ cf. Fowler, op. cit. (n. 28), at 301 n. 36, and Hippolytus, Ref. omn. haer. 5.8.40.

³⁰ Note that this shift of rôles is analogous to the one observed in connection with the Odyssean theme above: Epicurus may be likened to Odysseus, and the reader to Odysseus.

³¹ Graf, op. cit. (n. 25), 137 ends his discussion of the passage in this way: ‘Damit stellt sich uns der Mysterienvergleich aus “De anima” dar als eine Schilderung des Aufstiegs der Seele, welche lediglich in dem Sinne auf das Mysterienerlebnis Bezug nimmt, daß der Stimmungsverlauf während der Initiation mit demjenigen während des Todes zusammengestellt wird, ohne daβ die einzelnen Riten, welche einen solchen Erlebnisverlauf bewirken, konkretisiert würden.’ Even if the reference was to mysteries strictly speaking, we could not tell if it was to specific rituals like, e.g., those of Eleusis.

³² Relevant evidence is gathered and analysed in M. P. Nilsson, The Dionysiac Mysteries of the Hellenistic and Roman Age (1957), 116–32, esp. 122ff.; Graf, F., ‘Katabasis’, in Der Neue Pauly 6 (1999), 327–30,Google Scholar at 328. For the classical period see Seaford, R., ‘Dionysiac drama and the Dionysiac mysteries’, CQ 31 (1981), 252–75,CrossRefGoogle Scholar esp. 261–2 and earlier literature ibid.; on the probable rôle of an Orphic ἱερὸς λόγος περὶ τῶν ἔν Ἄιδου or περὶ τῆς Ἄιδου καταβάσεως see C. Riedweg, ‘Initiation – Tod – Unterwelt. Beobachtungen zur Kommunikationssituation und narrativen Technik der orphisch-bakchischen Goldplättchen’, in F. Graf (ed.), Ansichten griechischer Rituale (1998), 359–98, at 378–9, with R. Parker, ‘Early Orphism’, in A. Powell (ed.), The Greek World (1995), 483–510, at 484–7, and W. Burkert, Ancient Mystery Cults (1987), at 70, on the use of books in such initiations. A testimony from the imperial era is Lucian, Catapl. 22, where a man newly arrived in the underworld remarks that his experience closely resembles initiation at Eleusis (he is suddenly faced with Tisiphone); on the passage see Brown, C. G., ‘Empousa, Dionysus and the Mysteries: Aristophanes, Frogs 285ff.’, CQ 41 (1991), 41–50,CrossRefGoogle Scholar at 46. On the katabasis in the oracular cult of Trophonius and its relation to mysteries see P. Bonnechere, ‘Trophonius of Lebadea: mystery aspects of an oracular cult in Boeotia’, in M. B. Cosmopoulos (ed.), Greek Mysteries the Archaeology and Ritual of Ancient Greek Secret Cults (2003), 169–92, at 174.

³³ Georgics: Scazzoso, P., ‘Riflessi misterici nelle Georgiche di Virgilio’, Paideia 11 (1956), 5–28;Google Scholar L. Morgan, Patterns of Redemption in Virgil's Georgics (1999), 184–97; Hardie, A., ‘The Georgics, the Mysteries and the Muses at Rome’, PCPS 48 (2002), 175–208Google Scholar. Aeneid: Luck, G., ‘Virgil and the mystery religions’, AjfPh 94 (1973), 147–66Google Scholar. Note that it is possible to read the whole of De rerum natura 3 as mapped onto an initiation ritual. I observed above that, while at the beginning of the book it is Lucretius who assumes the rôle of the initiand, with Epicurus as mystagogue, in 3.930ff. one would have to assume a shift of röles: there Lucretius becomes the mystagogue, and the reader the initiand. Accordingly, the journey from darkness to light, perhaps the most significant single theme of initiation imagery, is performed twice, once by Lucretius in the prooemium, and once by the reader in the course of the katabasis which finally releases him to the upper world. Entirely in keeping with this is the fact that the first half of the book includes extensive instruction on Epicurean psychology; for there is good evidence that in various mystery cults the initiand had to undergo preparatory learning (παράδοσις); on this see Burkert, op. cit. (n. 32), at 69 with n. 14; Riedweg, op. cit. (n. 22), 6–19. So there would be another way to read Lucretius' invitation to apply the knowledge acquired in the first part of the book to one's fear of death (3.830 igitur).

³⁴ cf. Clinton, K., ‘The Eleusinian mysteries: Roman initiates and benefactors, second century BC to AD 267’, ANRW 11.18.2 (1989), 1499–1539Google Scholar.

³⁵ cf. also Cic., Leg. 2.36: ‘Nam mihi cum multa eximia divinaque vide<a>ntur Athenae tuae peperisse atque in vitam hominum attulisse, turn nihil melius illis mysteriis, quibus ex agresti immanique vita exculti ad humanitatem et mitigati sumus, initiaque, ut appellantur, ita re vera principia vitae cognovimus, neque solum cum laetitia vivendi rationem accepimus, sed etiam cum spe meliore moriendi’, ‘For among the many excellent and divine institutions which your Athens has brought forth and contributed to human life, none, in my opinion, is better than those mysteries. For by their means we have been brought out of our barbarous and savage mode of life and educated and refined to a state of civilization; and as the rites are called ‘initiations’, so in very truth we have learned from the beginnings of life, and have gained the power not only to live happily, but also to die with a better hope' (translation Keyes). And earlier Plato, Rep. 364b–365a.

³⁶ Privileged in more than one sense: one would expect a katabasis close to the middle of an epic.

³⁷ There are many other such correspondences, e.g. in the progression and development of themes from Books 2 to 3 and 5 to 6 respectively; see also Mewaldt, J., ‘Eine Dublette im Buch IV des Lucrez’, Hermes 43 (1908), 186–95,Google Scholar who argued that Lucretius originally intended to place Books 3 and 4 in reversed order, a view which still finds some supporters.

³⁸ On the interpretation of De rerum natura 6 fin. as a test whether the conversion to Epicureanism was successful see Clay, op. cit. (n. 17), 257–66; Gale, op. cit. (n. 19), ch. 6. Alii alia.

³⁹ cf. two of the three other occurrences of Tartar-^* in De rerum natura (3.42 wher e Tartara is tellingly qualified by leti; 5.1126 where the contrast with summum indicates that the underworld in general is at issue; at 3.1012 Lucretius refers to the creature Tartarus) and, e.g., Verg., Aen. 11.396–8 (Turnus speaking): ‘haud ita me experti Bitias et Pandarus ingens ׀ et quos mille die victor sub Tartar a misi, ׀ inclusus muri s hostilique aggere saeptus.’

⁴⁰ See the material collected by Scherling, O., ‘Tartaros’, RE II.8 (1932), 2440–5Google Scholar and the references in Bremmer, J. N., The Rise and Fall of the Afterlife (2002), at 136 n. 39Google Scholar. Kenney, op. cit. (n. 1), notes on ‘barathrum nee Tartara … atra’: ‘… “the black pit of Tartarus”, a hendiadys’. It seems barathrum here serves to disambiguate Tartara. D. N. Sedley, Lucretius and the Transformation of Greek Wisdom (1998), at 60–1, has an interesting digression on this passage. He starts off by observing that in 3.955, where Nature is haranguing an old man who is unwilling to die, baratre seems problematic, given its meaning (Ammonius, De diff. adf. voc. B3.29 βάραθρος μὲν γὰρ ὁ βαράθρου ἄξιος ἄνθρωπος); as he says, ‘hardly a productive way of conveying Lucretius’ principal message, that there is no pit of hell to fear’. So he suggests as a conjecture the objective genitive barathri, i.e. ‘aufer abhinc lacrimas, barathri et compesce querellas’ (‘Away with your weeping, and curb your complaining about the pit of hell’). Sedley continues: ‘The proposal has one immediate advantage. It supplies a piece of information which is otherwise left unstated, that the old man — whose words were not actually quoted — has been complaining partly about the prospect of hell. And without an indication to that effect, one might be left wondering why Lucretius, at 966–7, should offer his rationalistic denial of hell as directly confirming Nature's rebuke.’ Baratri in 3.955 seems an intriguing suggestion, but, as I have argued, we do not need it to make sense of what Lucretius says in 3.966–7. For an alternative suggestion concerning 3.955 see Smith, M. F., ‘Lucretius 3.955’, Prometheus 26 (2000), 35–40Google Scholar.

⁴¹ Strictly speaking, Hades is not envisaged in Od. 11 as being under the surface of the earth; but there are passages in the Iliad suggesting that the alternative view was also familiar (3.276, 19.258).

⁴² Indeed, the notion that the two Homeric epics reveal privileged knowledge of the kind made accessible through initiation rites (or the different notion that Odysseus' voyages can be read as a kind of initiation) can itself be found in antiquity; see Buffière, op. cit. (n. 19), 36–9, 49, 413–18.

⁴³ In general, Platonic dialogues can often be identified as targets of Epicurus and Epicureans, notably the Timaeus; cosmology and physics are of course areas where Epicureans and Platonists disagree sharply. Cf. Sedley, op. cit. (n. 40), 75–81; idem, ‘Epicurus and his professional rivals’, in J. Bollack and A. Laks (eds), Etudes sur l'épicurisme antique (1976), 119–59; and earlier Solmsen, F., ‘Epicurus and cosmological heresies’, AJPh 72 (1951), 1–23Google Scholar and ‘Epicurus on the growth and decline of the cosmos’, AJPh 74 (1953), 34–51;Google ScholarLacy, P. De, ‘Lucretius and Plato’, in Carratelli, G. P. (ed.), ΣΥΖΗΤΗΣΙΣ– Studi sull' Epicureismo Greco e Romano offerti a Marcello Gigante, vol. i (1983), 291–307Google Scholar.

⁴⁴ Indeed Neo-platonic commentators were to call Platonic myths like this one Nekyia later on, if they were concerned with the soul; see H. Tarrant, R. Jackson and K. Lycos (eds), Olympiodorus: Commentary on Plato's Gorgias (1998), at 294–5; Vorwerk, M., ‘Mythos und Kosmos: zur Topographie des Jenseits im Er-Mythos des Platonischen Staates (614b2–616b1)’, Philologus 146 (2002), 46–64,CrossRefGoogle Scholar at 46 n. 2.

⁴⁵ Some of the relevant passages are to be quoted below (notably 49267–493d3), on which see E. R. Dodds, Plato – Gorgias (1959) and Graf, op. cit. (n. 25), at 108, 120, 140.

⁴⁶ Apart from the commentators see W. Görier, ‘Storing up past pleasures’, in K. A. Algra, M. H. Koenen and P. H. Schrijvers (eds), Lucretius and his Intellectual Background (1997), 193–207; De Lacy, op. cit. (n. 43).

⁴⁷ See Dodds, op. cit. (n. 45), 1–5 and 58–66; Tarrant, H., Plato's First Interpreters (2000)Google Scholar, ch. 9: ‘From False Art to True. A Neoplatonist History of the Interpretation of Gorgias’.

⁴⁸ cf. Görier, op. cit. (n. 46), at 196.

⁴⁹ On the Epicurean theory of pleasure and the distinction between katastematic pleasure and pleasures in motion cf. Long, A. A. and Sedley, D. N. (eds), The Hellenistic Philosophers (repr. 1997), sec. 21Google Scholar; Algra, K. et al. (eds), The Cambridge History of Hellenistic Philosophy (1999), 648–66CrossRefGoogle Scholar; Striker, G., ‘Epicurean hedonism’, in idem, Essays on Hellenistic Epistemology and Ethics (1996), 196–208CrossRefGoogle Scholar.

⁵⁰ Dodds, op. cit. (n. 45), 297: ‘Socrates does not claim to know, and we cannot know, the identity of the κοψὸς ἀνήρ. He is not a philosopher but a teller of myths (μυθολογῶν, as); this rules out Olympiodorus' suggestion of Empedocles.' As the Strasbourg papyrus shows, in particular by placing Empedocles frg. 139DK in the prooemium of the Physika, demonology had a firm place in Empedocles' physical theory (see Primavesi, O., ‘Editing Empedocles’, in Burkert, W. et al. (eds), Fragmentsammlungen philosophischer Texte der Antike (1998), 62–88, esp. 80–6)Google Scholar; so Dodds seems to rule him out on what appear now to be insufficient grounds. While we should perhaps be hesitant to see a reference to Empedocles here (cf. Graf, op. cit. (n. 25), 108), it seems quite probable that Lucretius would have seen it. For in Empedocles there are of course ideas which fit the general context of our passage rather well; frg. 115DK, for instance, where Empedocles talks about his fate as a fallen demon who has been forced to return to earth, may be considered a predecessor to the ‘hell on earth’ motif we find in Gorgias and De return natura. For the Empedoclean theory of transmigration of the souls as a target of Epicurean criticism see Diog. Oen. frg. 42.ii.7–14 Ferguson Smith.

⁵¹ cf. Kenney, op. cit. (n. 1), at 227–8.: ‘… the Danaids must be taken as a type of behaviour rather than of suffering … It is suggested by Heinze that this is Lucretius' reason for not naming them, but it is a curious fact that they are not named by Plato in the Gorgias (4933–d), where the allegory of the leaky jar first occurs, though they are clearly alluded to.’ For further suggestions as to which features of the Danaids' description in Gorgias and Lucretius connect the two, against other instances of this image, see Keuls, E., The Water Carriers in Hades: A Study of Catharsis Through Toil in Classical Antiquity (1974), at 106Google Scholar.

⁵² In Cicero's De orat. 1.47 (written in 55 B.C.; dramatic date 91 B.C.) one of the protagonists (L. Licinius Crassus) claims to have studied it under the supervision of the Academic philosopher Charmadas (admittedly with particular attention to the critique of rhetoric); in the following fifty paragraphs the discussion is largely about topics raised in Gorgias, and it is clear that familiarity with the dialogue on the part of the reader is assumed. How widely read the Gorgias became in the imperial era can be gleaned from the index of testimonia in Dodds, op. cit. (n. 45), 397–8.

⁵³ I relegate to a footnote a selective survey of correspondences between De rerum natura 3 fin. and the final myth. Most conspicuous is of course the general structure of both texts, i.e. the sequence ‘judgement of the dead’ followed by a survey of the ‘mythical sinners’; it is remarkable that neither Homer nor any other epic prior to Lucretius nor the end myths in Plato's Phaedrus and Republic share this feature (Phaed. has a reference to the judgement in 107d–e, but no detailed treatment). Further, both Plato and Lucretius underline the truthfulness of their account, Plato by having Socrates stress that he believes the myth to be true and not a μῦθος, as Callicles would term it (52331–3), Lucretius by commenting as epic narrator on the harangue of Nature (3.951); in a situation where Lucretius may be seen as trying to assert Epicurean authority over a powerful Platonic tradition, this seems significant. By the same token, both texts emphasize the eternal validity of their accounts; in Plato, it is said that a judgement of the dead καὶ ἀεὶ καὶ νῦν ἔτι ἔστιν ἐν τοῖς θεοῖς (52336) in Lucretius the same is said about the exchange process of matter which, on Lucretius' view, as I have argued, could be the only subject of contention in any final judgement (3.967–71). In the Gorgias Socrates goes on to emphasize the Homericness of his underworld description, a somewhat obfuscating move, given that Plato diverges considerably from Homer and follows him only on minor details (cf. Dodds, op. cit. (n. 45), at 373); for a reader of Lucretius, however, this may facilitate the integration of the model Gorgias into the reading of De rerum natura 3 fin. as an Odyssean katabasis. And it is interesting that at Gorg. 524b2–4 Socrates says: ‘Death, it seems to me, is in fact nothing other than the separation of two things, the soul and the body, from each other’ — a view Lucretius has argued for in Book 3 as a whole and which he is driving home in the speech of Nature, but which in Gorgias obviously implies the continued existence of the soul; so this could be seen as a clever example of recontextualization (cf. n. 17 above on Lucretian inversion techniques).

⁵⁴ The quotations are from p. vi of Most's preface in Most, G. W. (ed.), Commentaries — Kommentare (1999)Google Scholar, but similar analytical categories are employed by I. Sluiter in her article in the same volume: ‘Commentaries and the didactic tradition’, 173–205. See also Sluiter, I., ‘The dialectics of genre: some aspects of secondary literature and genre in antiquity’, in Obbink, and Depew, , op. cit. (n. 28), 183–203. And further J. Assmann and B. Gladigow (eds), Text und Kommentar (1995)Google Scholar; Gibson, R. K. and Kraus, C. S. (eds), The Classical Commentary: Histories, Practices, Theory (2002).CrossRefGoogle Scholar

⁵⁵ cf. Most, op. cit. ( n. 54), xi, on ‘empowerment’ as the essential function of commentary.

⁵⁶ The work against Gorgias was by Metrodorus; Zeno (Pap. Here. 1005 col. 2) knew reasons for suspecting the attribution of its second book to Metrodorus. See K. Kleve, ‘Scurra Atticus — the Epicurean view of Socrates’, in ΣΥΖΗΤΗΣΙΣ, op. cit. (n. 43), 227–53, at 229 (mistakenly assigning the Against Gorgias to Zeno himself); D. Obbink, Philodemus on Piety (1996), 379–89 on col. 25.701–8.

⁵⁷ We have knowledge of books by Epicurus against rhetoric (and other arts; cf. D. L. Blank, Sextus Empiricus: Against the Grammarians (1998), at xxx–xxxi), but given the Gorgias' overt hostility towards rhetoric in its conventional form, I find it difficult to believe that it is the topic of rhetoric which made Epicureans turn to the Gorgias.

⁵⁸ For reconstructions of the Epicurean view of virtue and justice in particular see Waerdt, P. A. Vander, ‘The justice of the Epicurean wise man’, CQ 37 (1987), 402–22;CrossRefGoogle Scholar A. Alberti, ‘The Epicurean theory of law and justice’, in A. Laks and M. Schofield (eds), Justice and Generosity – Studies in Hellenistic Social and Political Philosophy (1995), 161–90.

⁵⁹ Another text which suggests itself for comparison is the Derveni Papyrus, on which see A. Laks and G. W. Most, Studies on the Derveni Papyrus (1997). In this text the speaker, who is obviously concerned with physical theory, provides an allegorical explanation of an Orphic theogony, which includes a section on punishments in the underworld; see Janko, R., ‘The physicist as hierophant: Aristophanes, Socrates and the authorship of the Derveni Papyrus’, ZPE 118 (1997), 61–94Google Scholar. The author attempts an allegorical explanation of this theogony and likens himself to an hierophant. I have no intuition whether the Derveni text or something like it was available in Rome in the first century B.C. (on the question of its earlier distribution see Obbink, D., ‘A quotation of the Derveni Papyrus in Philodemus' On Piety’, CE 24 (1994), 111–35)Google Scholar, but surely the similarities are remarkable. On the actual format of the commentary section see Lamedica, A., ‘Il Papiro di Derveni come commentario. Problemi formali’, in El-Mosalamy, A. H. S. (ed.), Proceedings of the XlXth International Congress of Papyrology, Cairo 5–9 Sept. 1989 (1992), vol. i, 325–33Google Scholar.

⁶⁰ cf. the Roman epigram in IG XIV 1746; Cic., Tusc. 1.10–14; Sen., Ep. 24.18; Juv. 2.149–52; Plin., N.H. 2.158.

⁶¹ The revival of Pythagoreanism in first-century B.C. Italy plays a rôle here too; Neo-Pythagoreans often claimed that Plato (and Aristotle) had stolen their doctrines from Pythagoras, and the subject matter of the Gorgias will have secured it additional attention in that connection. See e.g. C. H. Kahn, Pythagoras and the Pythagoreans: a Brief History (2001), ch. 7: ‘The Pythagorean Tradition in Rome’.

⁶² Lucretius' reading of Gorgias is tendentious in the sense that, by highlighting the connection between the pleasure passage and the end myth, he neatly pushes into the background a reading of Gorgias which would take into account that the discussion of pleasure is in Plato intertwined with a discussion of the good. Clearly, that is an aspect of the Platonic argument which an Epicurean must want to avoid.

⁶³ cf. e.g. Marincic, M., ‘Der “orphische” Bologna-Papyrus (Pap. Bon. 4), die Unterweltsbeschreibung im Culex und die lukrezische Allegorie des Hades’, ZPE 122 (1998), 55–9Google Scholar.

⁶⁴ An earlier version of this material was read to the Oxford Philological Society on 22 November 2002; the discussion afterwards helped me to make improvements, and I am grateful to Lesley Brown, Carlotta Dionisotti, Philip Hardie, Leofranc Holford-Strevens, Richard Sorabji, and Martin West for their questions and suggestions. For comments on earlier drafts I am indebted to Stephen Harrison, Nicholas Richardson, Michael Winterbottom, and Jim Adams, as well as the Editorial Committee of JRS; for suggestions on particular points I am grateful to Stephen Heyworth, Oliver Taplin, Matthew Leigh, Peter Parsons, Christina Kraus, and Michael Frede.