1. Green, J. R., Theatre in Ancient Greek Society (London and New York, 1994)Google Scholar and Wiles, David, Tragedy in Athens: Performance and Theatrical Meaning (Cambridge, 1997)CrossRefGoogle Scholar are perhaps the first major studies to do otherwise.

2. Goldhill, Simon, ‘Modern Critical Approaches to Greek Tragedy’ in Easterling, P. E. (ed.), The Cambridge Companion to Greek Tragedy (Cambridge, 1997), 324–47, at 339CrossRefGoogle Scholar: ‘The convention of masking has been much discussed with regard to what it might imply or deny about characterisation.’ Even more remains unexamined, however.

3. Green, , op. cit., 77–8Google Scholar: ‘We have no written discussion of masks surviving from the fifth century and so attitudes to masks in the early period of theatre can only be guessed at.’

4. Sommerstein, Alan H., Aeschylean Tragedy, (Bari, 1996), 41Google Scholar.

5. There is no distinction between masks worn by speaking actors and those that will not speak. All characters are open-mouthed, which is why it is a genuine surprise when Cassandra and Pylades speak in the Oresteia. This idea is opposed by Walton, J. Michael: ‘such non-speaking actors were to all intents and purposes invisible, since they did not wear masks’ (Greek Theatre Practice [London, revised edition 1991], 146Google Scholar; italics in original). In Hecuba, the elderly serving woman returns as an unspeaking character, as does Tecmessa in Ajax.

6. Harrison, Tony, ‘Facing Up to the Muses’, Proceedings of the Classical Association 85 (1988), 7—29, at 17—22Google Scholar; Rehm, Rush, Greek Tragic Theatre (London and New York, 1992), 41CrossRefGoogle Scholar.

7. Wiles, , The Masks of Menander: Signs and Meaning in Greek and Roman Performance (Cambridge, 1991), 142Google Scholar. Calame, Claude, ‘Vision, Blindness, and Mask: the Radicalization of Emotions in Sophocles' Oedipus Rex’ in Silk, M. S. (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford, 1996), 17—37, at 26—31Google Scholar, finds considerable significance in the holes in the mask and the senses the openings represent. Buxton, Richard, ‘What Can You Rely on in Oedipus Rex? Response to Calame’ in Silk, , 38–48, at 38–9Google Scholar, prudently urges caution with this.

8. Cf. Webster, T. B. L., Monuments Illustrating New Comedy, third edition rev. and enlarged by Green, J. R. and Seeberg, A., BICS Supplement 50 (London, 1995), vol. 1Google Scholar, 3 for references, to which could be added Plato Comicus fr. 151 PCG (142 K) and Suda, s.v. Thespis ( Walton, , op. cit., 33)Google Scholar. I disagree with Webster on the fragility of masks: cf. Green, , ‘Dedications of Masks’, Révue Archéologique 1982, 237–48Google Scholar and Halliwell, S., ‘The Function and Aesthetics of the Greek Tragic Mask’, Drama 2 (= Intertextualität in der griechisch-römischen Komödie, ed. Slater, Niall W. and Zimmermann, Bernhard, Stuttgart, 1993), 195—211, at 202 and n. 22Google Scholar. It is possible but unlikely that cork, leather, and wood were occasionally used. It is very convenient to experiment with stuccoed linen today, because of its cheap availability from medical suppliers (where its intended purpose is making casts).

9. Simon, Erika, The Ancient Theatre (London and New York, 1982), plates 4.1, 6.1, 9Google Scholar.

10. Jones, John, On Aristotle and Greek Tragedy (London, 1962), 59Google Scholar.

11. This point is well made by Johnson, Martha, ‘Reflections of Inner Life: Masks and Masked Acting in Ancient Greek Tragedy and Japanese Noh Drama’, Modern Drama 35 (1992), 20–34Google Scholar; it is a point missed by SirPickard-Cambridge, Arthur, The Dramatic Festivals of Athens, second edition revised by Gould, John and Lewis, D. M. (Oxford, 1968), 149, 171—4 (an important discussion nevertheless), and 190–5Google Scholar. ‘It's only when a mask is being worn by a skilled performer that the expression changes’ ( Johnstone, Keith, Impro: Improvisation and the Theatre [London, 1981], 185)Google Scholar; ‘The first second of wearing is easy; it is as if the mask switches itself on. The difficulty lies in letting it sustain its expression. In the early stages of learning the gaze will soon be lost, the mask betrayed as inanimate by the wearer's desire to make it express something which comes from his or her own experience rather than the characteristics of the mask’ ( Rudlin, John, Commedia dell'Arte: an Actor's Handbook [London and New York, 1994], 40)CrossRefGoogle Scholar; ‘The art of the full Mask lies in moving the mask in such a way that the attention is never distracted away from the face, by the body’ ( Johnstone, , op. cit., 196)Google Scholar. This emphasis on the presence of an audience is necessary: ‘The Masks live in the eye of the beholder, not of the actor’ ( Rudlin, , op. cit., 42)Google Scholar; ‘a theatre audience revises and reconstrues a mask's physiognomy, when the circumstances, attitudes, and emotions of the character change’ ( Rehm, , op. cit., 41)Google Scholar.

12. This is despite what happens to Benin's kings ( Jones, , op. cit., 45)Google Scholar who are not relevant: not all masking traditions serve similar societal functions. Even in Athens, there are differences between acting and, say, the archon basileus ‘becoming’ Dionysos at the Anthesteria. Examples exist of many masks being given to one character, and of many masks for one actor, each of which demonstrates the impossibility of Jones's claims.

13. This is attested only for Hellenistic acting: Pickard-Cambridge, , op. cit., 195–6Google Scholar; Simon, , op. cit., 10, plates 2–5Google Scholar; Rehm, , op. cit., 154 n. 22Google Scholar.

14. Easterling, , ‘Form and Performance’ in Easterling, , op. cit., 151–77, at 153Google Scholar. Cf. Wiles, , op. cit., 133Google Scholar; Pickard-Cambridge, , op. cit., 138—9Google Scholar. Modern audiences have no difficulties accepting rolesharing, even when the actors remain unmasked ( Halliwell, , op. cit., 199–200)Google Scholar.

15. Cf. Pavlovskis, Zoja, ‘The Voice of the Actor in Greek Tragedy’, CW 71 (1977), 123–33Google Scholar. Here notions of ‘illusion’ and ‘disguise’ are often unhelpfully confused.

16. Understanding Comics: the Invisible Art (Northampton, Mass., 1993), 36Google Scholar, with a full discussion of iconic faces at 28–37 (and more generally 24–59). McCloud is discussing cartoons, but makes explicit links with masks at 34. In the citation, I have substituted the word ‘mask’ for McCloud's ‘cartoon’, but the principle remains the same.

17. Rehm, , op. cit., 41Google Scholar. Cf. Halliwell, , op. cit., 203Google Scholar and Taplin, Oliver, ‘Comedy and the Tragic’ in Silk, , op. cit., 188–202, at 189Google Scholar: ‘The tragic mask is, in fact, rather blank and expressionless, somewhat solemn perhaps, waiting to take its “expression” from the events of the play.’

18. Wiles (n. 1), 169. The offstage decision that motivates Neoptolemus' entry at Philoctetes 1222 is the rule-proving exception in this case. That this is a late play (produced in 409) shows the beginning of a development towards a more psychological understanding of character, perhaps triggered by the rise of more literary genres designed for a reading public.

19. True and False: Heresy and Common Sense for the Actor (New York, 1997), 9Google Scholar: ‘The actor does not need to “become” the character. The phrase, in fact, has no meaning. There is no character. There are only lines upon a page …’

20. Halliwell, , op. cit., 199, 201–3Google Scholar.

21. Cf. Rehm, , op. cit., 40Google Scholar. In contrast, Green (n. 1), writes ‘I would suspect, however, that the underlying motive for the dedication of the masks in the earliest days of the theatre was a more serious one, to leave behind with the god in his sanctuary the “otherness” created in his honour, and not to take it out into normal society’ (79). This may be, but it is not essential to masked acting, particularly with institutionalized role-sharing.

22. Halliwell, , op. cit., 197—9Google Scholar also rightly downplays many false assumptions about masks and the god Dionysos.

23. Taplin, Oliver, The Stagecraft of Aeschylus: the Dramatic Use of Exits and Entrances in Greek Tragedy (Oxford, 1977), 33—4Google Scholar blurs the difference between what is actually on stage (which because it is in a theatre will never be fully like life), and how the audience perceives what is on stage (which can be illusionistic – ‘life-like’ – or not). The anecdote is found in the Life of Aeschylus [ed. Page, ], p. 2Google Scholar, lines 10–13, with a translation in Csapo, Eric and Slater, William J., The Context of Ancient Drama (Ann Arbor, 1995), 260CrossRefGoogle Scholar.

24. The importance of this date in the development of Athenian acting is rightly noted by Kaimio, Maarit, ‘The Protagonist in Greek Tragedy’, Arctos 27 (1993), 18–33Google Scholar.

25. I am glossing over several issues of dating and authenticity that cannot be addressed here (such as with Prometheus Bound and Women of Trachis). Suffice it to say that I believe the claim to be true as it stands.

26. Halliwell, , op. cit., 209Google Scholar(and cf. 205) also argues for a minimalistic mask, ‘one functioning component of an actor's appearance’. In Webster, contrast, ‘The Poet and the Mask’, Classical Drama and its Influences (New York, 1965), 5–13Google Scholar suggests that in the fifth century individual masks were made for individual characters, and this great diversity only became conventionalized by Hellenistic times ( Hunter, R. L., The New Comedy of Greece and Rome [Cambridge, 1985], 11)CrossRefGoogle Scholar. There is a sense in which this remains possible: that masks were made individually but conformed to one of the six types I am about to describe; that individual masks were created according to minimalistic designs. This is not what Webster means, however.

27. MacDowell, Douglas M., Aristophanes and Athens: an Introduction to the Plays (Oxford, 1994), 258Google Scholar.

28. Sommerstein, , op. cit., 47Google Scholar; cf. Stone, Laura M., Costume in Aristophanic Poetry (Salem, 1984), 22–7Google Scholar. Sommerstein continues, suggesting how this could be ‘darkened’ in Aeschylus' Suppliant Women, while still maintaining a light/dark distinction: ‘the swarthy complexions of the Danaids, at once African and unfeminine – and surprisingly appropriate, it transpires, to these maidens who first reject marriage and then, when compelled to submit, murder their husbands – and the still darker features (cf. Supp. 719, 745, 888) of the Egyptian pursuers.’ In practice, ‘dark’ skin could reasonably be a reddish colour, much like the colour of red-figured pottery. Similarly, there is no need to assume any especial paleness in the masks of Sophocles' Tyro (fr. 648) or Euripides' Ino (Wasps 1413–14).

29. Arnott, Peter D., The Ancient Greek and Roman Theatre (New York, 1971), 45Google Scholar.

30. The Greek equivalents given here are to show that the six basic mask types for which I am arguing do map naturally onto the Greek language. In no way am I trying to be prescriptive of the use of the words themselves in the dramatic texts.

31. ‘The Mask Types’ ( Webster, , Monuments Illustrating Old and Middle Comedy, third edition rev. and enlarged by Green, BICS Supplement 39 [London, 1978], 13–26)Google Scholar does not qualify the difference between (comic) young women and middle-aged women (masks S, SS and T, TT) but the difference illustrated is only the presence of some facial lines and the absence of a part in the hair of the middle-aged masks. In tragedy the difference need not be more than this.

32. Halliwell, , op. cit., 200–3Google Scholar. I often remind students that this is a world before corrective lenses, and for many viewing at a distance would be difficult: this is an undervalued reason for the importance of the chorus in fifth-century drama.

33. Cf. Wiles (n. 7), 179. By Hellenistic times there was an increased desire to represent physiognomic factors. While the concept in Greek thought that character is denoted in appearance goes back to Homer (e.g., Thersites in Iliad 2), I can see no indications of its influence on fifthcentury masking. For Hellenistic interest in physiognomies and masks, cf. Wiles, ibid., Theophrastus, Characters, and ps.-Aristotle, Physiognomônika. The minimalist view of masks offered here is, in fact, latent in Pollux (a scholar from the second century A.D., whose list is available in translation at Csapo, and Slater, , op. cit, 398—402)Google Scholar, in Pickard-Cambridge, , op. cit., 193—5Google Scholar, and in Webster (n. 31), 13—26: masks in the more extensive lists are grouped according to categories of age and sex. The reason for the further subdivisions of mask types in all these lists is that it is now the eye of a taxonomist, rather than that of a theatre-goer, that is looking at the masks.

34. Vase-painting suggests ‘tragic masks were used by the principals in the satyr play, which helps to reinforce the belief that the satyr play was primarily self-parody’ ( Walton, , op. cit., 158)Google Scholar, but artistic conventions are not always prescriptive. I believe that there were significant differences between comic and tragic masks, but precisely what these were remains uncertain because each vase-painter had his own iconographic idiom. Taplin's reflections on the Choregoi vase (n. 17, 189–90) might not be quite so straightforward as a result.

35. Jones, , op. cit., 45Google Scholar; this is echoed by Taplin (n. 23), 35.

36. Wiles (n. 7), 131.

37. It is worth at least mentioning the effect this has on anonymous characters. When we read a play today, we see lines attributed to ‘Messenger’ or ‘Shepherd’ (or even ‘Chorus’) and we treat them differently than we do those of named characters. With only six basic mask types, audience expectations for every character that comes on stage is at the same high level. Perhaps part of the surprise with Euripides' Telephus in 438 can be explained because costume and body language were not consonant with ‘Telephus’, and there was nothing to individuate the character in the mask. Students and actors react differently to characters if they have to cross out the word ‘Messenger’ in their translations, and call him (or her?) by a real name. Taplin (n. 23), 82 n. 1 makes a similar point about the costumes Messengers wear.

38. One could add ‘script’ as a fourth element. Johnstone, , op. cit., 181–2Google Scholar emphasizes the nonrepeatability of performance when mask and text are combined, which provides unexpected corroboration for the single performance of Athenian plays in competition.

39. Caedel, E. B., ‘The Division of Parts among the Actors in Sophocles' Oedipus Coloneus’, CQ 35 (1941), 139–47Google Scholar.

40. Dearden, C. W., The Stage of Aristophanes (London, 1976), 126Google Scholar is needlessly concerned with the possibility of two characters wearing the ‘same’ mask type.

41. Taplin, , Greek Tragedy in Action (Berkeley, 1978), 89CrossRefGoogle Scholar; Walton, , op. cit., 158Google Scholar. Halliwell, , op. cit., 206 n. 36Google Scholar doubts blind masks were used. Other blind masks are suggested for Sophocles' Thamyras and at least one of his plays called Phineus (for Phineus and his blinded children). Sophocles acted the lead in Thamyras (he played the lyre), so this is an early play, before his retirement from performance ( Life of Sophocles 4; translation available at Csapo, and Slater, , op. cit., 225)Google Scholar; as such, it likely predates the institution of the actor's contest c.449. In this early blinding, it remains just possible (in a way that it would not be later, I would contend) that a vertically bisected half-blind, half-sighted mask was used: ‘The story that [Thamyras] has one blue and one black eye was explained by Lessing as deriving from the mask worn by the actor who played the part, which had one blue eye, which he presented to the audience before the blinding, and one black eye, which he presented after it’ ( Lloyd-Jones, Hugh, Sophocles: Fragments [Cambridge, Mass., 1996], 102–3)Google Scholar; this would be very restrictive for movement, with the audience on three sides of the performance area: nevertheless, from time to time comedy seems to have achieved a similar effect, so it is not impossible.

42. Cf. Calame, , op. cit., 26–31Google Scholar and Buxton, , op. cit., 38–9Google Scholar.

43. Richard Seaford describes the same effect being achieved differently in a modern production: ‘In the Oxford production of 1976 … the mask was simply turned upside down, so that Pol[yphemus'] gory mouth became his wounded eye’ ( Euripides: Cyclops [Oxford, 1984], 220)Google Scholar.

44. Cf. Marshall, C. W., ‘Idol Speculation: the Protean Stage of Euripides' Helen’, Text & Presentation 16 (1995), 74–9, at 76–7Google Scholar. It seems likely that the chorus of Libation Bearers also had gashes on their cheeks ( Sommerstein, , op. cit., 47Google Scholar; Pickard-Cambridge, , op. cit., 192Google Scholar suggests this for the chorus in Aeschylus' Suppliant Women too). In this same category might also be placed the bruised face of Sophocles' Tyro ( Pickard-Cambridge, , op. cit., 195)Google Scholar. There is no need to assume any special reddening to explain Silenus in Euripides' Cyclops 228.

45. The only instance where I can see value in having a mask depicting an age category other than the basic three (not counting children, which I am not considering here) is with Jocasta, in Oedipus Tyrannus. Oedipus himself is a mature man (he has been king for about ten years and has children), but his wife is also his mother (‘This woman [gunel] is his wife and mother … of his children’ says the chorus at 1. 928, with ironic caesura). Certainly the old woman mask would be ridiculous in this context (old women are always barren in tragedy, e.g., Creusa in Ion, or Hecuba), but a mask that was clearly older would be desirable: might Jocasta be going grey? It is possible that even this is too specific, and that the gunê mask would be worn by any character aged, say, 30–50.

46. Walton, , op. cit., 189Google Scholar.

47. Mills, S. P., ‘The Death of Ajax’, CJ 76 (1980), 129–35Google Scholar.

48. Peter Burian, ‘Myth into Muthos: the Shaping of Tragic Plot’ in Easterling, , op. cit., 178–208, at 198 n. 34Google Scholar. Cf. Taplin (n. 41), 98–100; Rehm, , op. cit., 40Google Scholar, Wiles (n. 1), 174.

49. In Euripides' Andromeda (and Sophocles' earlier version?; cf. Green [n. 1], 20–2) Medusa's severed head would have appeared as a mask. In Sophocles' Phineus, a mask may have been used to represent the decapitated head of Phineus' second wife (fr. 707a; Lloyd-Jones, , op. cit., 334–5)Google Scholar. In Aeschylus' Isthmiastae, a satyr play, satyr-actors, wearing satyr masks, appear on stage holding satyr masks; this must have been a striking effect (cf. Green [n. 1], 45 and 79, Taplin [n. 23], 420–2). Similarly, I am doubtful of the specific description of Taplin's, ‘Old Man Playing Antigone’ (Comic Angels and Other Approaches to Greek Drama through Vase-Painting [Oxford, 1993], 83–8): it is certain that the old man himself is already wearing a (comic) gerôn mask. For whatever reason, the man is wearing a dress and (though it is hard to see) carrying a female (korë?) mask. Might this rather be the sort of scene one would expect in a play set ‘behind’ a dramatic festival, such as Aristophanes' Proagon?Google Scholar

50. A modern production of Orestes (Euripides, Orestes, dir. Alexander Gelman, trans. John Peck and Frank Nisetich, prod, and dramaturg James T. Svendsen: the Twenty-Six Classical Greek Theatre Festival of the University of Utah, September 1996) accomplished the same effect by establishing an initial tableau with Orestes' body covered by a white sheet, drawing on modern iconic associations of murder scenes (familiar to the audience from television). When Orestes did eventually stir, the modern audience felt the same surprise that the ancient one would have felt.

51. There is ‘possibly a somewhat feminine appearance for the unmanly Aegisthus, who is more than once called a “woman” by enemies’ ( Sommerstein, , op. cit., 47)Google Scholar. For Aristophanes, MacDowell, , op. cit., 258 and n. 19Google Scholar lists Agathon in Thesmophorizusae, the students in Clouds, Chaerephon in Wasps 1412–14 (and therefore also in the first production of Clouds, if he appeared as a character?). Contrast the woman in Lysistrata who tried to tan herself to appear more male (MacDowell, ibid., 304). Similarly, the Relative in Thesmophoriazusae once shaved is still darkfaced, as he tries to pass as a woman (MacDowell, ibid., 309); this is clearly another instance of a special mask, with a detachable beard (detachable? not necessarily: with a single performance it need only be a beard that can be easily cut during the show; note the presence of stubble in a nearcontemporary illustration of the play – cf. Taplin [n. 49], 36–41 and plate 11.4, Eric Csapo, ‘A Note on the Würzburg Bell-Crater H5697 (“Telephus Travestitus”)’,Phoenix 40 [1986], 379–92)CrossRefGoogle Scholar.

52. In this context, it might be possible to make sense of the claim that in Phaethon Phaethon's corpse appeared smoking and charred (fr. 78.1 [= line 214]; unfortunately the fragmentary and corrupt nature of the text makes it unclear). While a baroque literal realization of this is not beyond Euripides, I suspect that the claim could have originated in an unusual death mask. Rather than representing the corpse with a white mask, because the youth was consumed by fire it seems plausible that he might be given a specially-created black mask. The body does not, at any rate, need to be kept off stage (cf.Collard, C., ‘Phaethon’ in Collard, , Cropp, M. J., Lee, K. H. (edd.), Euripides: Selected Fragmentary Plays [Warminister, 1995], 202—3)Google Scholar, and this would be sufficient to account for the meaning of the line in production.

53. Taplin (n. 41), 84–5, 127. There is no indication that the material, form, purpose or construction of choral masks differed in any way from that of actor masks.

54. Dover, K. J., ‘Portrait-Masks in Aristophanes’ in Boerma, R. E. H. Westerdorp (ed.), ΚΩΜΩΙΔΟΤΡΑΓΗΜΑΤΑ: Studia Aristophanea uiri Aristophanei W. J. W. Koster in honorem (Amsterdam, 1967), 16–28Google Scholar.

55. Dover, ibid., 26.

56. This explanation helps clarify the story in Aelian (Varia Historia ii. 13) that when foreign audience members at the original production of Clouds asked who this Socrates was, the real Socrates (aged 45 in 423) stood silently in answer. The story makes sense whether there is a portrait mask or not (so Dover, , Aristophanes: Clouds [Oxford, 1968], xxxiiiCrossRefGoogle Scholar; he acknowledges disagreement about what exactly the point is at xxxiii n. 1), but the opportunity to avoid confusion in the audience is maximized by using an established theatrical mask (the silenus from satyr-drama) in a novel context (for a character in Old Comedy). The meaning of the mask (‘Socrates looks like a silenus’) is foregrounded or ‘marked’ by its context in a way that would not occur either with a tradition of portrait masks or with a standard mature male mask. As for other Old Comic ‘portraits’, again there is no need to go beyond conventional signs established for use in masks.

57. Cf. fr. 60; Storey, Ian C., ‘The War Between the Poets’, delivered at CAMWS (Omaha, Neb., 04 21, 1995)Google Scholar.

58. Though I do not agree with every aspect of the reconstruction, the point is made very clear by the computer model available at http://didaskalia.berkeley.edu/stagecraft/mask_mm/rotmaskl.html [URL is correct as of 10.1.99].

59. Dodds, E. R., Euripides: Bacchae (Oxford, 1960), 113–14Google Scholar.

60. Pollux does list Actaeon among his ‘special’ masks: if it is based in fact (and not an ‘educated’ guess), it could derive from a Hellenistic tragedy. Similar doubts could be raised about a mask representing ‘Euhippe daughter of Cheiron changing into a horse in Euripides’ ( Csapo, and Slater, , op. cit., 400)Google Scholar. Precise as this entry appears, there is no known play in which these characters occurred.

61. Cf. Csapo, , ‘Deep Ambivalance: Notes on a Greek Cockfight’, Phoenix 47 (1993), 1–28, 115–24CrossRefGoogle Scholar, esp. 1–8, 120–4, with the extensive bibliography on p. 1.

62. Dearden, , op. cit, 125Google Scholar.

63. This is suggested by Heracles, 883–41. For what it is worth, Lyssa does appear on Pollux's list of special masks. Based on Plutus, 422–4, it is possible Poverty was also given an Erinys mask (cf. Halliwell, , op. cit., 204–5)Google Scholar.

64. Dodds, , op. cit., 131Google Scholar. The idea has been accepted most recently by Gredley, Bernard, ‘Comedy and Tragedy – Inevitable Distinction: Response to Taplin’ in Silk, , op. cit., 203–16, at 203 and 214 n. 2Google Scholar.

65. An alternative is that the character has two masks, one for his incarnation as the ‘human’ Stranger, and a second for the divine Dionysos, but this seems not to be preferable. However, without a better understanding of how Odysseus was staged in Euripides' Philoctetes and Hera in Aeschylus' Semele, we cannot be certain (cf. Taplin [n. 23], 428 n. 1).

66. I presume that both characters (like most of the heroes at Troy after the ten-year expedition) are middle-aged men. Even here, this is clearly part of a deliberate effect, for there are three other such characters in the play: Menelaus and Agamemnon (both played by the same actor) and Odysseus.

67. Cf. Walton, , op. cit., 70Google Scholar for a preliminary discussion.

68. Vernant, J.-P., ‘The God of Tragic Fiction’ in Vernant, and Vidal-Naquet, P., Myth and Tragedy in Ancient Greece (New York, 1988), 181–8Google Scholar; Halliwell, , op. cit., 197Google Scholar.

69. Cf. Friedrich, Rainer, ‘Everything to Do with Dionysos? Ritualism, the Dionysiac, and the Tragic’ in Silk, , op. cit., 257–83, at 268–9Google Scholar.

70. Versions of this paper have been delivered to the Department of Classics, Linguistics, and Modern Languages at Concordia University in Montreal, March 1998, and to the Classical Association of Canada, in Ottawa, May 1998. I would like to offer thanks to the audiences of these occasions, to Jim Svendsen and Andrew Sherwood, and to Ian Storey from whose conversations I always leave the richer.