¹ References to Frogs are to the Budé edition of Coulon, V. (Paris, 1923–1930)Google Scholar. The following will be cited by author's name: Burkert, W., Homo Necans (RGVV 32, Berlin and New York, 1972) = (Engl.ed., Berkeley and Los Angeles, 1983)Google Scholar; Graf, F., Eleusis und die orphische Dichtung Athens in vorhellenistischer Zeit (RGVV 33, Berlin and New York, 1974)Google Scholar; Nilsson, M. P., Geschichte der griechischen Religion³ 1 (Munich, 1967).Google Scholar

This paper originally formed part of a contribution to a colloquium on Greek drama held at the University of Western Ontario, 24 March 1990. For encouragement and advice I am grateful to Professors D. E. Gerber, R. D. Griffith, B. C. MacLachlan, E. Robbins, and, especially, Sir Hugh Lloyd-Jones. In addition, I am indebted to Professor A. Henrichs, who not only commented critically on a draft of this paper but also generously sent me a copy of the paper discussed below, n. 31.

² It is unclear whether we are to imagine Empousa as real or simply a figment of Xanthias' imagination devised to terrify Dionysus; Xanthias' fear seems real enough at 296. The question, however, has no direct bearing on the present discussion.

³ Dover, K. J., Greek Homosexuality (Cambridge, MA, 1978), p. 139Google Scholar, notes that ‘sexual opportunism and uninhibited arousal are characteristic of the comic stage’.

⁴ The paradosis in fact transmits both Hegelochus' γαλ and Euripides' γαλ'; the former is more humorous and is almost certainly correct in the present passage. For some speculation on the nature of Hegelochus' mistake, see Daitz, S. G., CQ 33 (1983), 294–5CrossRefGoogle Scholar. In addition to the Frogs passage, this slip provided comic material for Sannyrion (fr. 8 PCG) and Strattis (frs 1 and 63 PCG).

⁵ A full collection of material has been assembled by Waser, O., RE 5 (1905), 2540–3Google Scholar, S.V. ‘Empusa’; see also Herter, H., Rhein. Jahrb. f. Volkskunde 1 (1950), 112–43Google Scholar = Kl. Schr. pp. 43–75, a detailed discussion of evil spirits and related beings in early Greek popular belief (119f. = 50f., on Empousa).

⁶ Borthwick, E. K., CQ 18 (1968), 200–6CrossRefGoogle Scholar (quotation from 206).

⁷ The earliest explicit reference is Pl. Phdr. 250b–c, εἶδ τε κα τελοτο τ⋯ τελετ⋯ ἣθμις λγει μακαριωττη … κα εὐδαμoα ϕσματα μυoμεo τε κα πoπτεoτες αὐγῇκαθαρᾳ …. Plut. fr. 178 Sandbach speaks of σεμτητας κουσμτων ἱερ⋯ κα φασμτω γω; Aristid. Or. 22.3 (Keil) speaks simply of the worshippers τoîς ρρτoις φσμασι, and Procl. Resp. 2.185.4 Kroll of φσματα … γαλης μεστ. We also learn from Origen, Cels. 4.10 of φσματα κα δεματα in Bacchic mysteries. Cf. Burkert, p. 317 n. 64 = p. 288 n. 64. Most suggestive for our discussion is the Plutarch fragment (see below).

⁸ Graf, pp. 40ff., who also provides a thorough survey of earlier views on the question.

⁹ Cf. the well-known story of Aeschylus' prosecution for revealing the Mysteries on stage (the sources are collected by Radt as T 93a–d, 94). Lefkowitz, M. R., The Lives of the Greek Poets (Baltimore, 1981), p. 68Google Scholar, is rightly sceptical of Heracl. Pont. fr. 170 Wehrli = T 93b Radt, but that the more reliable Aristotle preserves the story (Eth. Nic. 1111a8 = T 93a) suggests that it has some historical basis. To the bibliography assembled by Radt, add Ostwald, M., From Popular Sovereignty to the Sovereignty of Law (Berkeley and Los Angeles, 1986), pp. 528–30.Google Scholar

¹⁰ For Eleusinian eschatology, see Graf, pp. 79ff. The most likely vehicle for the propagation of such views was epic poetry (see below).

¹¹ See Süss, W., Ethos: Studien zur älterer griechischen Rhetorik (Leipzig, 1910), p. 248Google Scholar (cf. also pp. 249f., on the rhetors' use of sexual innuendo), Jacoby on FGrHist 338 F 2, and Wankel on Dem. 18.130.

¹² Alciphron (3.26.3) alludes to Dem. 18.130 in a description of a lustful, old serving-maid. At Philostr. V.A. 4.25 Empousai are described as man-eating creatures, addicted to sex (έρσι…άφροδισίων), who use sexuality to ensnare their victims. It is noteworthy in this regard that one of the manifestations of Empousa in Frogs is a γυν | ὠραιοττη τις (290f.); cf. also the libidinous old woman at Eccl. 1056f.

¹³ The principal ancient support for this view is Str. 10.3.18 (quoted below, n. 16). The fundamental modern discussion is Lobeck, C. A., Aglaophamus (Königsberg, 1829), i.646–59Google Scholar (who detects ‘Orphic’ elements as well); cf. Rohde, E., Psyche^9/10 (Tübingen, 1925), ii.110Google Scholar with n. 1. More recent scholars have identified the rites as those of Sabazius, but without ‘Orphic’ elements: e.g. Jacoby, on FGrHist 338 F 2Google Scholar; Nilsson, p. 836; Burkert, p. 317 n. 64 = 288 n. 64, and Ancient Mystery Cults (Cambridge, MA, 1987), p. 141 n. 34Google Scholar (quoted below, n. 14); Graf, p. 105 with n. 53; Wankel on Dem. 18.259; Koster, S., Die Invektive in der griechischen und römischen Literatur (Meisenheim am Glan, 1980), p. 89Google Scholar; Parker, R., Miasma (Oxford, 1983), pp. 302fGoogle Scholar.; West, M. L., The Orphic Poems (Oxford, 1983), pp. 26fGoogle Scholar. Dieterich, A., Nekyia² (Leipzig, 1913), pp. 81fGoogle Scholar. and RhM 48 (1893), 279–80Google Scholar = Kl. Schr. pp. 121–2, saw in Demosthenes' description on ‘Orphic’ rite of expiation (influenced by Asian cult), a view which won a certain amount of acceptance (bibliography in Graf, p. 105 n. 53).

¹⁴ ‘This is the most detailed description of telete in the classical period’, Burkert, , Anc. Myst. Cults (previous note), p. 141 n. 34.Google Scholar

¹⁵ Harding, P., Phoenix 41 (1987), 25–39, pp. 30fCrossRefGoogle Scholar. on the present passage. Cf. also Nisbet's remarks on the similar habits of Roman orators in his edition of Cic. in Pis., p. 193.

¹⁶ This is patently true of Strabo, whose identification of the cult is clearly an inference based on the ritual cries (10.3.18): τν δὼ Φρυγων [SC. ίερν μμνηται] Δημοσθνης, διαβλλων τν Aἰσχνου μητρα κα αὐτν, ὡς τελοσῃ τῇ μητρ συνντα κα συνθιασεοντα κα πιφθεγγμενον εὐοῖ σαβοῖ, πολλκις κα ὓης ἄττης, ἄττης ὓης· τατα γρ στι Σαβζια κα Μητρῷα. Cf. also Amphitheos, , FGrHist 431 F 1bGoogle Scholar, where the cult cry is emphasized.

¹⁷ The earliest passage in which the identification is made explicit seems to be Amphitheos, , FGrHist 431 F 1Google Scholar, but his date is uncertain: see Jacoby, , FGrHist 3b (Komm.) i.258.Google Scholar

¹⁸ See Wankel on Dem. 18.259 (τᾐ μητρ τελοσῃ). Cf. the Ὡραι of Aristophanes, probably written some time after 420, in which the ‘new’; gods are tried, condemned and deported (frs 577–89 PCG): cf. Cic. Leg. 2.37, ‘nouos uero deos et in his colendis nocturnas peruigilationes sic Aristophanes, facetissimus poeta ueteris comoediae, uexat, ut apud eum Sabazius [mentioned in fr. 578] et quidam alii dei peregrini iudicati eciuitate eiciantur’. It is noteworthy that Sabazius is singled out alone among the dei peregrini. Sabazius is also referred to at Vesp. 9, Au. 873, Lys. 388; his worship may have been mentioned in the Βπται of Eupolis (cf. fr. 94 PCG), but it is uncertain how it was treated. There seems to have been greater conservatism regarding θεο ξενικο in the fourth century than in the fifth: see Nilsson, pp. 837f.

¹⁹ Harding (above, n. 15) 30. I would not, however, characterize the innuendo as ‘light’; it seems to me to be very heavy indeed. Cf. also Süss (above, n. 11), p. 248 (‘…das sie eine Veranstalterin von Winkelmysterien von verdächtigem Charakter gewesen sei’).

The majority of commentators, it seems, have been beguiled by the wealth of detail in Demosthenes' description into believing that an actual cult is being documented. It was, of course, his intention to persuade his audience, but this does not mean that his account is true. The description seems to be composed of features of conventional Dionysiac cult laced with a few details hinting at more obscure practices. The various details are all fully discussed by Wankel in his notes to the passage (although it should be noted that he believes the cult to be that of Sabazius). One practical result of a more agnostic view of the Demosthenes passage is that individual details should now be treated with some suspicion. It is possible that Demosthenes, drawing on his knowledge of cult practices, found every detail in some cult (not, of course, necessarily the same cult); but it is equally possible that some are merely plausible fictions. A problem in this regard is that many of the late lexicographical sources brought to bear on the interpretation of the passage are ultimately derived directly from it or from later discussion of it (e.g. Strabo, quoted above, n. 16), that is to say that these late sources have little value for establishing the cultic significance of any particular in the description. For example, the cry φυγον κακν, εὗρον, ἄμεινον is elsewhere said to be a marriage formula and is not attested in mystery cult (Graf, p. 106 n. 54, where he also notes that marriage is a kind of initiation, but this point does not prove that the formula was ever used in mystery cult). The combination ὑς ἄττης ἄττης ὑς seems suspect as well, ὑς may derive from an epithet of Dionysus (Cleidemus, FGrHist 323 F 27 with Jacoby ad loc.; cf. Pherec. FGrHist 3 F 90, where Ὕη is a name for Semele), or it may be a name of a foreign god (Ar. fr. 908 PCG); as Wankel notes, the connection with Sabazius in late sources seems to have been prompted by the Demosthenes passage. Ἄττης, on the other hand, is apparently an eastern figure (Neanthes, FGrHist 84 F 37)Google Scholar. As Wankel points out, what little evidence there is for a link between Ἄττης and Dionysus-Sabazius is late. Accordingly, it is possible that the combination in our text rests on no religious authority. Demosthenes himself may have fashioned a plausible-sounding cult jingle from two unrelated elements (on the form of the jingle, cf. Fraenkel, E., JRS 51 [1961], 48 = Kl. Beitr. ii.119)Google Scholar. In short, I would argue that, despite its richness of detail, Demosthenes' description of the religious activity of Aeschines' mother should not be accepted uncritically as evidence for cult.

²⁰ For the importance of darkness in the Greater Mysteries, see Burkert, p. 304 = p. 276.

²¹ In view of the centrality of the tradition of Heracles' katabasis, the oath here may well be pointed.

²² For the general view of the work, see Helm, R., RE 13 (1927), 1732fGoogle Scholar. Lucian quotes or alludes to Frogs at Philopatr. 25, Cont. 24, Cat. 14, Fug. 28. Other references to Aristophanes are collected by Householder, F. W. Jr., Literary Quotations and Allusions in Lucian (New York, 1941), pp. 4–5Google Scholar (cf. also pp. 7–10, for references to uncertain comic fragments). I am grateful to Professor C. P. Jones for advice on Lucian.

²³ ‘Much the commonest use of γον is to introduce a statement which is, pro tanto, evidence for a preceding statement’, Denniston, J. D., The Greek Particles² (Oxford, 1954), p. 451Google Scholar (with copious examples).

²⁴ See Nilsson, p. 725, and Herter (above, n. 5 ) 133f. = 65f. Empousa's association with Hecate is central to Borthwick's discussion of the superstitious background to the passage (see above, n. 6). For Hecate, see the standard discussion by Kraus, T., Hekate (Heidelberg, 1960)Google Scholar; cf. also Nilsson, pp. 722–5; Wilamowitz, , Der Glaube der Hellenen² (Basel, 1956), i.169f.Google Scholar; Johnston, S. I.. Hekate Soteira (Amer. Class. Stud. 21, Atlanta, 1990).Google Scholar

²⁵ The use of κ το in line 440 is aetiological: see N. J. Richardson ad loc, and Clay, J. S., The Politics of Olympus (Princeton, 1989), p. 257.Google Scholar

²⁶ See Richardson, on H.Dem. 24–6Google Scholar; cf. also ARV ² 1012.1 with the remarks of Bérard, C. in A City of Images (Princeton, 1989), p. 115Google Scholar (fig. 159). Clinton, K., OAth 16 (1986), 45Google Scholar, has argued that Hecate did not play a rôle in Eleusinian cult, noting that ‘nowhere over a span of ca. 1,000 years does the name Hekate appear at Eleusis’. This position is in keeping with Clinton's attempt to minimize the connection between the Hymn and the Mysteries. Although his paper is in many ways a healthy corrective to the work of those scholars who treat the Hymn as a kind of blue-print for the Mysteries, his position on the rôle of Hecate seems unlikely. In the first place, Hecate seems to have been closely associated with Artemis (perhaps attested as early as Hes. fr. 23a.26 M–W), who had a place in Eleusinian cult (cf. Richardson, , AND Farnell, L. R., The Cults of the Greek States [Oxford, 1896–1909], ii.560f., 569)Google Scholar. It is also possible that her name and function in the Mysteries were considered part of the ἄρρητον, and this would account for the relative silence of our sources. It is, however, striking to note Melanthius, FGrHist 326 F 2, who mentions in his work On the Eleusinian Mysteries that the τρλη and the μαινς were sacrificed to Hecate (cf. Antiph. fr. 68 Kock, and Apollod. FGrHist 244 F 109 for these offerings). Clinton (45 n. 17a) objects that Melanthius does not say explicitly that this sacrifice was made in the Mysteries, but it seems likely, especially in view of the evidence of the Hymn, that Hecate's worship would not be discussed in Melanthius' treatise if it were not relevant to the Mysteries. Clinton's judgement has recently been endorsed by Clay (previous note), p. 218 with n. 62, who prefers to play down the relevance of external considerations to the understanding of the Hymn, holding in this case that ‘The text itself sufficiently elucidates Hecate's role as a critical intermediary in the gradual process of revelation’ (at p. 257 she is more sympathetic to an aetiological reference to line 440). Yet this position does not explain why it is Hecate and not some other intermediary figure.

²⁷ She is called κουροτρϕος at Orph. Hy. 1.8 (Quandt). See Friis Johansen and Whittle on Aesch. Suppl. 676, where her function as κουροτρος is discussed. Cf. H.Dem. 24, ταλ φρονουσα with Allen–Sikes–Halliday as well as Richardson. It may be relevant to note that West prints Ἔκτα (Badham) at Aesch. Ag. 140, a passage much concerned with the nurture of young.

²⁸ See Kannicht on Eur. Hel. 569–70, and Johnston (above, n. 24), pp. 34–5, 135–6.

²⁹ Cf. also ΣAp. Rh. 3.861 (p. 242 Wendel), …φσματα πιπμπειν τ καλομενα ‘Εκαταῖα (Brimo and Empousa are mentioned); Bekker, , Anec. Graeca i.249.27f.Google Scholar, Ἔμπουσα φσμα στ τν ὑπ τς Ἑκτης πεμπομνων.

³⁰ It is suggestive in this regard to note the frequent association of Hecate with dogs (Johnston, op. cit. [n. 24], pp. 134ff.); in Frogs it is only after Xanthias says that the apparition has changed into a κων that Dionysus identifies it as Empousa (293).

³¹ Cf. Hesychius S.V. Ἔμπουσα, Ἀριστοφνης δ τν Ἐκτην ἔφη Ἔμπουσαν. A. Henrichs has challenged the usual view of Ar. fr. 515 in part 3 of ‘Namenlosigkeit und Euphemismus: zur Ambivalenz der chthonischen Mächte im attischen Drama’, forthcoming in the proceedings of a colloquium held in honour of S. Radt. Henrichs argues that Hecate is a goddess of public and private cult, but that Empousa is a creation of the popular imagination without any rôle in cult. On his view, speaker B in terror undercuts the first speaker's solemn invocation of Hecate through the comic substitution of Empousa for Hecate, which makes his fear concrete on the level of popular superstition. This view seems to me to be problematic in a number of respects. That Empousa has no cultic identity is difficult to maintain in the face of Idomeneus F 2 (discussed above), a passage which Henrichs does not discuss. On any view, Idomeneus seems to suggest a link between Empousa and mystic initiation (of course, if the usual understanding of this passage were correct, Ἔμπουσα would stand as a cult-title), and this is not surprising. Empousa's association (though not identification) with Hecate is well attested in other sources (see above, n. 29), and it is therefore plausible to argue that Empousa had a place in certain cults as an extension of the figure of Hecate. Hecate is closely connected with what we generally describe as popular superstition (in which Empousa is regularly defined), and it is highly unlikely that these associations are irrelevant to Hecate's cultic identity. In this light I would argue that Burkert, W., Le sacrifice dans l'antiquité (Entretiens sur l'Antiquité classique 27, Vandoeuvres–Geneva, 1980), p. 118Google Scholar, is correct in saying with reference to Ar. fr. 515, ‘Sie [i.e. Hekate] kann selbst als ‘Ungeheuer' erscheinen’. Speaker A refers to Hecate in terrifying terms; speaker B uses the name Empousa, and thus specifies the relevant aspect of Hecate. Henrichs sees a discontinuity between the two speakers, arguing that they are differentiated in ‘Stilhöhe’. But there is nothing in the text of the final line to suggest that there is any stylistic contrast with the paratragic solemnity of speaker A (the question τ καλεῖς is not alien to tragic usage: e.g. Soph. Ph. 737, OC 1459; Eur. El. 1123, IT 780, Ph. 617, 849).

³² According to Paus. 1.28.6, Aeschylus was the first to present the Erinyes with snakes in their hair. It became common in later sources. Cf. Garvie on Aesch. Cho. 1049–50. For the Erinyes in general, see Wüst, E., RE Suppl. 8 (1956), 82–106Google Scholar, and Lloyd-Jones, H., REG 102 (1989), 1–9CrossRefGoogle Scholar = Craik, E. M. (ed.), ‘Owls to Athens’: Essays on Classical Subjects Presented to Sir Kenneth Dover (Oxford, 1990), pp. 203–11Google Scholar. On the iconography of the Erinyes, see Sarian, H. and Delev, P., LIMC 3.1 (1986), 840fGoogle Scholar.

³³ See Nilsson, p. 656, and Richardson on H. Dem. 47. Note especially the prominence given to torches in the parodos of Frogs (340ff., 351ff., 448).

³⁴ Norden on Verg. Aen. 6.309–12 (and passim); Lloyd-Jones, H., Maia 19 (1967), 206–29Google Scholar. Further discussion in Boardman, J., JHS 95 (1975), 6–10CrossRefGoogle Scholar; Clark, R. J., Phoenix 24 (1970), 244–55CrossRefGoogle Scholar, and Catabasis: Vergil and the Wisdom Tradition (Amsterdam, 1979), pp. 79ff.Google Scholar; Graf, pp. 142–50, 182–5; Robertson, N., Hermes 108 (1980), 274–300Google Scholar (who argues that the katabasis formed part of the Hesiodic Aegimius [frs 294–301 M–W], which he would attribute to Cecrops of Miletus). This poem seems to lie behind the summary of Heracles' journey to the Underworld preserved in [Apollod.] Bibl. 2.5.12, and to have influenced Bacch. 5, Pind. fr. dub. 346 Maehler, Ar. Ra., and Verg. Aen. 6. Scepticism has been expressed by Thaniel, G., Phoenix 25 (1971), 237–45CrossRefGoogle Scholar (a response to Clark's paper in the same journal), and Austin, on Verg. Aen. 6.309ff.; but neither of these scholars actually offers any arguments against the hypothesis.

³⁵ Lloyd-Jones (previous note) 219, who also discusses other details that may reflect the influence of the Heracles epic. It might be objected that it is unlikely that Aristophanes could have been influenced by both Eleusinian cult and the lost Heracles poem. But I see nothing implausible in positing complementary influences: Aristophanes was in no way constrained to follow one ‘source’.

³⁶ Cf. Pind. Ol. 13.63, Pyth. 10.47; Aesch. Cho. 1049–50; [Aesch.] PV 799; [Apollod.] 2.4.2; Ov. Met. 4.771. In general, see Ziegler, K., RE 7.1 (1912), 1630–55Google Scholar. For possible links with Hecate and the Erinyes, see Fontenrose, J., Python (Berkeley and Los Angeles, 1959), pp. 287fGoogle Scholar.

³⁷ The poem seems to have contained an αἴτιον for the preliminary initiation in the Lesser Mysteries: see Lloyd-Jones (above, n. 34) 211f. with n. 8, and Boardman (above, n. 34) especially 9f.; cf. Burkert, pp. 292ff. = pp. 265ff.

³⁸ There is the possibility that the experience was aural. At some point, it is possible that the μσται were blindfolded or at least veiled (Burkert, pp. 303f. = p. 275). It may be that at this time they were told of an approaching phantom. This view can be reconciled easily with the passage in Frogs, where Dionysus is told of Empousa, but does not himself see her. Nonetheless, I regard this possibility as less likely, since the other passages discussed point to something seen.

³⁹ Cf. H.Dem. 480–2 (with Richardson ad loc), Pind. fr. 137a Maehler, Soph. fr. 837 Radt, and terms like ποπτεα.

⁴⁰ The language here makes it very probable that the reference is to the Eleusinian Mysteries. Against the scepticism of Mylonas, G. E., Eleusis and the Eleusinian Mysteries (Princeton, 1961), pp. 264ffGoogle Scholar., see Graf, pp. 132ff.

⁴¹ With the image of the λειμνες, cf. Frogs 326 and 374, where meadows are the goal of the mystic procession; such places of beatitude may have been the end of the way travelled by the initiate in the Underworld: cf. [Diph.] fr. 136.6 PCG; Posidip. fr. 705.22 SH (with Lloyd-Jones, H., JHS 83 [1963], 93f.)CrossRefGoogle Scholar. It is, of course, possible that the image in the parodos of Frogs is due simply to a general association of λετμνες with Demeter (for which see Richardson's edition of H.Dem., p. 143), but this seems unlikely. The meadow has a firm place in the traditional picture of the Underworld, which was developed in the eschatologies of various mystery cults: cf. Od. 11.539, 573, 24.13; Pind. fr. 129.3 Maehler; PI. Resp. 614c2, 616b2, Gorg. 523e4, 524a2; Verg. Aen. 6.637ff. (with the commentaries of Norden and Austin); Gold Leaf A 4 Zuntz (= Orph. fr. 32f. 6 Kern); Orph. frs 222.3 and 293 Kern; perhaps Emped. Kath. fr. 5 Zuntz = 31 B 121 DK. In general, see Norden's edition of Aen. 6, pp. 25f.; Dodds' edition of Pl. Gorg. p. 375; Dieterich, Nekyia (above, n. 13), pp. 19ff.; Zuntz, G., Persephone (Oxford, 1971), pp. 253f.Google Scholar

⁴² I propose to discuss elsewhere the relevance of the parodos to our understanding of cultic αἰσχρολογα.

⁴³ Cf. especially 448ff., a passage which has strong affinities with Pind. fr. 129 Maehler: see Graf, pp. 82ff., and Lloyd-Jones, H. in Pindare (Entretiens sur l'Antiquité classique 31, Vandoeuvres–Geneva, 1985), pp. 245–83Google Scholar, on the nature and context of Pindar's eschatological poetry.

⁴⁴ See Burkert's compelling reconstruction of the initiation in the τελεστριον (pp. 303ff. = 274ff.)