1. This paper developed from a lecture given at the JACT Greek Summer School at Bryanston in 1992 to introduce the production of the play directed by David Langslow. The production was a memorable and exciting one, and my thanks go to the director and cast for convincing their audience that Euripides always knows best. In a later incarnation, it was given to the Classical Association AGM held at St. Andrews in 1995, to the Classics Research Seminar at Manchester University, and to the Oxford University Classical Society. I am most grateful to all the participants in these sessions for their patience and their insightful comments, especially to Malcolm Heath and David Bain.

2. This is indeed probably the only possible way to construe this remark, given its position in the hypothesis: on the regular shape of these see Page, D. L. (ed.), Euripides: Medea (Oxford, 1938), liii–lvGoogle Scholar. Brown, A. L., ‘The Dramatic Synopses Attributed to Aristophanes of Byzantium’, CQ 37 (1987), 427–31CrossRefGoogle Scholar, who doubts the authenticity of many of these hypotheses, describes the end of the hypothesis of And. as containing material ‘of “Aristophanic” type’, but does not commit himself further.

3. Verrall, A. W., Essays on Four Plays of Euripides (Cambridge, 1905), ‘A Greek Borgia’, 1–42Google Scholar. The quotation is from p. 3.

4. Erbse, H., ‘Euripides’ Andromache', Hermes 94 (1966), 276–97Google Scholar, repr. in Schwinge, E.-R. (ed.), Euripides: Wege der Forschung (Darmstadt, 1968), 275–304Google Scholar.

5. Norwood, G., Essays on Euripidean Drama (Berkeley, 1954), 46Google Scholar; Garzya, A., Euripide. Andromaca (Naples, 1953), introduction pp. 23–4Google Scholar.

6. Kovacs, D., The Andromache of Euripides: An Interpretation (Chico, 1980), 75–7Google Scholar,83. A variation on this approach is put forward by Aldrich, K. M., The Andromache of Euripides, University of Nebraska Studies 25 (1961), 10–22Google Scholar, esp. 13, and 59, who argues that unity is achieved through the interrelation of contrasting pairs of characters.

7. Boulter, P. N., ‘Sophia and Sophrosyne in Euripides’ Andromache', Phoenix 20 (1966), 51–8Google Scholar. Aldrich (n. 6 above) sees Troy as a unifying theme (77–8); Kitto, H. D. F., Greek Tragedy (London, 1954), 240–7Google Scholar, saw hatred of Sparta as the point, and unifying theme, of the play.

8. Storey, I. C., ‘Domestic Disharmony in Euripides’ Andromache’, G&R 36 (1989) 16–27Google Scholar, repr. I. McAuslan, and Walcot, P. (eds.), Greek Tragedy (Oxford, 1993), 180–92Google Scholar. Aldrich (n. 6 above), 67–9 also stresses the importance of the marriage of Neoptolemus and Hermione in motivating the action.

9. Hartung, I. A., Euripides restitutus II (Hamburg, 1844)Google Scholar, Friedländer, P., ‘Die griechische Tragödie und das Tragische,’ DieAntike 2 (1926), 101Google Scholar. See also Vellacott, P. (tr.), Euripides: Orestes and Other Plays (London, 1972), 26–43Google Scholar, esp. 30.

10. Pohlenz, M., ‘Der Ablauf der Handlung in der Andromache des Euripides’, Anzeiger der österreichischen Akademie der Wissenschaften phil.-hist. 84 (1947), 115Google Scholar; Friedrich, W. H., Euripides und Diphilos, Zetemata 5 (Munich, 1953), 47Google Scholar.

11. Others have been willing to ascribe some limited importance, and even unifying force, to the character of Neoptolemus, but always with careful qualification. Kamerbeek, J. C., ‘L’Andromaque d'Euripide', Mnemosyne 11 (1943), 47–67Google Scholar, esp. 67, seems to play with the idea only to reject it: ‘For the figure of Neoptolemus, even though he remains in the wings, contributes in a peculiar manner (singulièrement) to the unity of the tragedy (I am far from wishing to maintain, as Friedländer does, that Neoptolemus is fundamentally the character who unites the different motifs of the piece)…’ See also Lee, K. H., ‘Euripides’ Andromache: Observations on Form and Meaning’, Antichthon 9 (1975), 6–14CrossRefGoogle Scholar, esp. 6: ‘Less tangible, but no less real, is the function in this connection (sc. links between the actions at both the thematic and the formal level) of the person of Neoptolemus. Though he takes no active part in the play, he is mentioned repeatedly, and can be seen as one of the causes from which each action springs… his is, of course, not to say that he is the hero of the play…, since he is presented as a figure whose actions are of importance only insofar as they impinge on the lives of others.’ Cf. p. 14. See also Lloyd, M. (ed.), Euripides: Andromache (Warminster, 1995), 4Google Scholar: 'Neoptolemus is central to the myth on which Andromache is based, and the play itself focusses on his return, but it would be an exaggeration to describe him as its “hero”.' I am not looking for a hero, a very problematic concept (see n. 16 below), but for a unifying central figure.

12. The obsolete heroism of Neoptolemus and the contrast he presents with the Spartans is also stressed by Kovacs (n. 6 above), p. 74, and Aldrich (n. 6 above), 74–5, who thinks he is kept off-stage precisely because of his status as a Homeric hero of the old school. See also below.

13. Kovacs (n. 6 above), p. 50, comments: ‘The value of this double plot i s … that it allows the poet to deal more exhaustively with a single set of themes.’ Rather, it allows him to knit several very different themes closely together.

14. Heath, M., Unity in Greek Poetics (Oxford, 1989), 8–9Google Scholar.

15. Aristotle, Poetics 1451b33ff., tr. Halliwell, S., The Poetics of Aristotle (London, 1987), 41–2Google Scholar and see 105–12 for his commentary on the passage.

16. See Flashar, H., ‘Die Poetik des Aristoteles und die griechische Tragödie’, Poetica 16.1 (1984), 1–23Google Scholar for discussion of the problematic aspects of relating Aristotle to fifth-century tragedy; Lee (n. 11 above), 8 does not feel the need to look for strict Aristotelian unity in the play.

17. Indeed, precisely what Aristotle meant by these terms, particularly ‘necessary’, is very unclear: see Halliwell, S., Aristotle's Poetics (London, 1986), 98–106Google Scholar.

18. Cousineau, Thomas, Waiting for Godot: Form in Movement (Boston, 1990), 53–5Google Scholar.

19. Marcel Frère: ‘Who is Godot?’ Roger Blin: ‘That's the secret of the play.’ (Marcel Frère, Combat, 7 January 1953). Alan Schneider: ‘Who or what does Godot mean?’ Beckett: ‘If I knew, I would have said so in the play.’ (Schneider, Alan, ‘Working with Beckett’, in Samuel Beckett: The Critical Heritage, ed. Graver, Lawrence and Federman, Raymond (London, 1979), 173–88Google Scholar.

20. Jong, I. J. F. de, ‘Three Off-Stage Characters in Euripides’, Mnemosyne 43 (1990), 1–21CrossRefGoogle Scholar.

21. Christie, Agatha, Five Little Pigs (1942)Google Scholar, Pt. 3, ‘Reconstruction’.

22. Aristotle famously complained about this: Poetics 1461bl9–21, if it indeed refers to Aegeus in Medea and not to some problem connected with the lost Aegeus.

23. Others have also found Orestes' account of himself convincing dramatically: see e.g. Grube, G. M. A., The Drama of Euripides (London, 1941), 209–10Google Scholar, and Stevens, P. T. (ed.), Euripides: Andromache (Oxford, 1971), p. 7Google Scholar. But Kovacs (n. 6 above), 52–4 raises a number of queries, most of which seem rather of the number-of-Lady-Macbeth's-children type, though his argument that the postponement of Neoptolemus' death until after the recovery Hermione by Orestes helps to knit the play together more closely is plausible. Lloyd (n. 11 above) on 879–1008 sees Orestes' reticence as ‘hardly explicable in terms of his own motives’, but intelligible because it ‘allows Hermione to complete the exposition of her despair’.

24. So also e.g. Aldrich (n. 6 above), 61, and Lloyd (n. 11 above) on 1008.

25. Heath, M., The Poetics of Greek Tragedy (London, 1987), 92–5Google Scholar; de Jong (n. 20 above), 14. Lee (a 11 above), 4–16, has a similar understanding of the play to Heath's and de Jong's but he does not make use of the same terminology.

26. Heath defines ‘focus’ thus (n. 25 above), 91: ‘The audience is required to concentrate its attention on key figures in the action, and to do so sympathetically, that is, in such a way that they are involved with and respond to the fortunes and feelings of those characters. I shall use the term “focus” henceforth in this technical sense, of any character who is serving as a centre of sympathetic attention.’ But this definition begs a number of questions: see below.

27. Lloyd (n. 11 above), 3, sees the focus of the final part of the play as resting with Neoptolemus.

28. Heath (n. 25 above), 207–8.

29. These are both Heathian terms: see Heath (n. 25 above), 93, 97.

30. Although Heath's definition of focus is clearly quite different from the use of terms like ‘focalization/focalized/focalizer’ in narratological criticism (the fundamental text is Genette, G., ‘Discours du Récit’, in Figures III (Paris, 1972)Google Scholar, tr. Jane E. Lewin (New York, 1980), followed by Nouveau discours du recit (Paris, 1983)Google Scholar, tr. Lewin (Ithaca, 1988), our understanding of focus in drama might in fact be improved if we were able to develop a similarly nuanced set of terms which might be applied to a dramatic text as a whole. Jong, I. J. F. de, Narrative in Drama; the Art of the Euripidean Messenger Speech (Leiden, 1991), uses these concepts to great effect in her analysis of the narratives of the messenger speeches, but does not bring them to bear on the play as a whole: the messengerGoogle Scholar speech is seen as something qualitatively different from the rest of the play. In one sense of course, it is qualitatively different (she rightly [p. 117] quotes Aristotle, Poetics 1449b 24–7, who specifically contrasts the dramatic mimesis with narrative [απαγγελíα]). But narratological techniques have successfully been applied to film (see Chatman, S., Story and Discourse: Narrative Structure in Fiction and Film (Ithaca, 1978))Google Scholar, and I see no intrinsic reason why it should be impossible to describe how one character is used to throw light on the actions and emotions of another using an analogous set of terms.

31. Lysias, , On the Olive Stump, 41Google Scholar.

32. I am arguing here for the principle that the audience can be interested in more than one person at once, not for the oikos as the unifying theme in the Andromache (for which see the views of Kovacs and Aldrich in n. 6 above): of course the oikos is important in the play (as in numerous Greek tragedies), but I think it is insufficient to explain it.

33. See also Aldrich (n. 6 above), 69: ‘One senses throughout the play Euripides’ determination to keep the image of Neoptolemus in the minds of the audience. How he might have accomplished this more effectively without actually bringing the young warrior on stage is difficult to suggest… Cf. Lloyd (n. 11 above), 3–4.

34. See a 12 above. Garner, R., From Homer to Tragedy: the Art of Allusion in Greek Poetry (London, 1990), 132–4Google Scholar, sees his death as alluding to the death of Hector. This is not impossible, since the Trojan context is important as a contrast between Neoptolemus (and Peleus) and the Spartans, not in order to emphasize any differences between Greeks and Trojans.

35. Eur. HF 514 (Megara sees him approaching in the distance), 524 (he arrives).

36. See Lloyd (n. 11 above), 3–4 on the play as a nostas-drama. There is the more room for doubt in the audience's mind as to the precise turn events will take because it seems clear that Euripides has innovated in combining the stories of the concubinage of Andromache with the recovery of Hermione by Orestes and the death of Neoptolemus: see Friedrich (n. 10 above), 47–9, Stevens (n. 23 above), 1–6.

37. Menelaus' character has found no defenders that I know of: see also Conacher, D. J., Euripidean Drama (Toronto, 1967), 178–9Google Scholar; Kovacs (n. 6 above), 61–3.

38. HF 626–36: there is even a similar general statement (631–6) of the importance of children: compare 418–20.

39. See also Kovacs (n. 6 above), 74–5.

40. Eur. And. 738–46: παρώ δε πρòs πα παρόѵτας έμϕαѵώς γαμβροѵς διδάεομαι λόγοѵς. κἂν μὲν κολάϚηι τήνδε καί τὸ λοιπόν ἡι σώϕρων καθ ἡμάς, αὼϕρον ὰντιλἡΨεται θνμούμενος δἑ τεύξεται θνμονμένων. τοὺς σοὺς δὲ μύθονς ῥαιδίως έΥὼ ϕέρω σκιἀ Υὰρ ὰντίστοιχος ὡς ϕωνὴν εχεις, ὰδύνατος ούδὲν άλλο πλἠν λέΥειν μόνον. (‘When both I and my son-in-law are present, I will instruct him openly, and he me. And if he punishes this woman and in future is reasonable to me, he will find me reasonable in turn, but if he is angry, he will find me angry too. I take your words lightly; for like an opposing shadow you have a voice, but are incapable of doing anything other than talk.’ The heavy use of polyptoton stresses the reciprocity of Menelaus' relations with Neoptolemus and, while making Menelaus sound splendidly pompous, it also succeeds in making Neoptolemus seem very real.

41. See also Kovacs (n. 6 above), 72.

42. See also Kovacs (n. 6 above), 74–5.

43. See above, nn. 12 and 37, and Borthwick, E. K., ‘Trojan Leap and Pyrrhic Dance in Euripides’ Andromache’, JHS 87 (1967), 18–23Google Scholar.

44. See Verrall (n. 3 above), 42, Vellacott (n. 9 above), 43, and a 1 above.

45. And not just in tragedy: Odysseus in the Odyssey is an obvious absent husband whose family have problems without him. Christopher Pelling has suggested tq me that Odysseus in the Telemachy is an excellent example of another absent character on whom much attention is focussed: see Klingner, F., ‘Über die vier ersten Biicher der Odyssee’, Akad. d. Wissensch. z. Leipzig, Phil.-hisl. Kl. 96 (1944), 1–55Google Scholar. In Andromache, the family is destroyed both by war and by the disastrous Spartan alliance: but the two causes cannot be fully separated out because Helen, the cause of the war, is also Spartan, and Hermione is her daughter. This may be why the abuse of that Spartan character is not modified or qualified in the way that Euripides usually redefines initial stereotypes. Peleus' outburst against Spartan women, and Andromache's abuse of Spartans, cannot be used to date the play, but they are nothing if not convincing in the context of the damage the Spartan dynasty has done to the house of Peleus. On Sparta in the play see also e.g. Aldrich (n. 6 above), 58–9; Conacher (n. 37 above), 171–2; Kovacs (n. 6 above), 63–4.

46. Salacrou, Armand, headline of review of En Attendant Godot, Arts #400, 01 27, 1953Google Scholar.