1. See Fantham (1982a), 116ff., for a discussion of the manuscript tradition.

2. Troas is unlikely in this context to mean ‘the Trojan Woman’. Who in this play would it refer to? There is no single unifying character.

3. My argument might seem to suggest that Troas is the correct title, as being more accurate. I think Troas is more accurate but not necessarily therefore more authentic. The two titles overlap. Troades equals Troas with added implications: both imply place, but Troades is more suggestive about occasion, the likely protagonists and dramatic model. Troades is a misleading title, but that hardly makes it an impossible one.

4. See on the matter of Greek models Tarrant (1978), esp. 217, Fantham, (1982a), 71–75, Calder (1970).

5. Note that Fantham (1982a), 263, says the second chorus of Seneca’s Troades is not to be understood as sung by Trojan women!

6. The Euripidean parodos and the third ode in Seneca’s Troades speculate on likely destinations for the captive women.

7. I speak about Seneca’s Troades as if in performance for the following reasons: a) it is almost impossible to discuss drama otherwise; b) there is nothing in this particular play which in performance would be either impossible or incomprehensible; c) I do not accept the hypothesis that Senecan tragedy was not written for performance of some sort. The discussion of this issue has been long and fruitless. The points made by Walker (1969) in her excellent review of Zwierlein (1966) have never to my knowledge been countered. The possibility of ‘private performance’ as raised by Calder (1975) cannot be lightly dismissed.

8. On the supernatural in Senecan tragedy and the Troades, see Braginton (1933), Childress (1981), and Jackson Knight (1932).

9. See the discussion of the play’s structure in Lawall (1982), 250.

10. Tarrant (1978), 228, goes too far in seeing an absence of organic coherence as a result of abrupt transitions within the action. The disorientation produced in the spectator or reader is only temporary and serves dramatic ends of surprise and uncertainty.

11. Though this produces unity of a different kind, it is a genuine unity. Tarrant argues (1972), 196: ‘To demonstrate unity of structure, it is not enough to posit one or even several isolated points of contact between a scene and a recurrent theme; many scenes in Seneca bear on the central issues of their play without ceasing to be self-contained units, elaborated primarily for their own sake.’ In answer, it can be maintained that there is nothing in the self-sufficiency of the unit incompatible with its subordination to a greater design. Secondly, the relationship between scene and theme is never, at least in this play, a matter of ‘one or even several isolated points of contact’, but one of complete interdependence. For many of these recurrent themes and motifs see Steidle (1972) and Schetter (1972).

12. See Aristotle Poetics ch. VIII.

13. It has almost become a commonplace of criticism that Seneca pays too little attention to overall dramatic structure because his interest is mainly in the scene or speech at hand. See Mendell (1941), 16; Costa (1974), 101, 103.

14. ‘The agon of Ulysses and Andromache fighting to discover and conceal the child’s hiding place would itself be a whole episode of Greek tragedy, and is treated by Seneca on that scale’, Fantham (1982a), 271. I outline what I take to be the structure of the third act later in this paper.

15. That a major break occurs here has been remarked upon before. For instance see Fantham (1982a), 240. Tarrant (1978), 228, uses this break as an example of scenes ‘juxtaposed without concealing material’.

16. Hence the futility and unfairness of comparisons like that of Mendell (1941) between the Oedipus plays of Sophocles and Seneca.

17. Longinus ch. 19. On the theory of rhetorical figures as being founded upon particular states of psychology and emotion see Vickers (1970), 93ff.

18. A handy guide to the characteristics of the pointed style and Seneca’s use of it is given in the Introduction of Summers (1962): this discussion is marred, however, by the author’s view of the style as aiming mainly at wit and a show of ingenuity. In the words of Herington (1966), 437: ‘Every age has its own forms and conventions, which cannot be judged as good or bad in the abstract, but only one way in which the individual writer employs them.’ See also Canter (1970).

19. See, for instance, Watling (1966), 25: ‘If we look among the idiosyncrasies of Seneca’s tragic style for “faults”, we can find plenty: excess of rhetoric, irrelevance, iteration, banality, bathos … Such lapses are the by-product of the labour of striving to extract the utmost effect from the spoken word.’

20. Summers (1962) xxxvi n. 1, observes: ‘Sometimes suspicio is very like what we mean by “dramatic irony”.’

21. Compare the discussion of General Irony of Muecke (1970), 68: ‘The universe appears to consist of two systems which simply do not gear together. The one functions, and can only function, in terms of meanings, values, rational choices, and purposes; the other seems not to be comprehensible in these terms. And yet, though the two systems are incompatible, they are also interinvolved; the alien system extends its dominance into the very centre of the “human” system and the “human” system feels obliged to find meanings, values and purposes in the non-human, in short to reduce the duality to a unity.’ In the Trojan Women, however, the ‘alien system’ seems not purposeless to the reader or spectator but wilful.

22. Adopting the A reading.

23. See Owen (1970), 367: ‘It is not in the nature of either the funeral occasion or its ritual to set a point of departure, but rather to signal an end, a full stop. Psychologically and dramatically, its vision is retrospective.’

24. Using Lattimore’s translation (1951). For Seneca’s familiarity with the Homeric epics, see Fantham (1982a), 20.

25. Fantham (1982a), 237 (note on 193f.) and Seidensticker (1969), 170, are among those who have pointed to allusions in Act II to Book 1 of the Iliad. However the connection seems to me to have been insufficiently emphasized; it is not a mere matter of allusions. Iliad 1 provides the rationale for Act II of Seneca’s play.

26. Allusions to Iphigenia in Act II have often been pointed out. See e.g. Schetter (1972), 233–41; Seidensticker (1969), 166f.

27. Fantham (1982a), 260: ‘The huge cost of Calchas’ utterance (ingenti mercede) recalls again the sacrifice of Iphigenia (cf. quo solent pretio 360).’

28. Lawall (1982), 252: ‘Pyrrhus’ murder of Polyxena at the end of the play repeats in an even more cruel and senseless act his murder of Priam described at the beginning.’

29. See Cardinal (1981), 176, whose terminology I adopt here.

30. For the dependence of paradox on irresolution, see Colie (1966), 508–20.

31. Colie (1966), 518: ‘Paradox is not evolutionary, nor yet logically sequential … it has no end: or, to switch meanings a little, its means are always ends.’

32. It is wrongly assumed by Bishop (1972), 335, that Andromache forces the information out of Helen.

33. For a different approach to Helen’s characterization and role and a discussion of the treatment given her in pre-Senecan literature, see Tsirpanlis (1970).

34. See Fantham (1982a), 337.

35. Dryden describes it as comparable to Shakespeare at his best. The comments of Costa (1974), 103, are unworthy of him: ‘There is a cold intellectuality in Andromache’s soliloquy which reflects the debating schools rather than the anguished decision she has to make, whether to give up her son to death or to allow her husband’s tomb to be desecrated (Tro. 642ff.).’

36. The importance of death in Act III is clear: the whole point of the act is the attempt to save Astyanax from execution. Fear and hope (see choral ode line 399), particularly the former, dominate the characters in Act III: fear 423, 425, 426, 431, 437, 477, 496, 505, 513, 515, 530, 548, 551, 576, 586, 588, 592f., 609f., 612, 618, 626, 631, 632, 642, 662, 742, 767, 790; hope and loss of hope 425, 490, 741, cf. Andromache’s speeches 461–76 and 768ff.

37. The play as a whole can be seen as a similar recessed panel structure: witnessing and mourning the physical end of Troy (Act I); Polyxena (Act II); Astyanax (Act III); Polyxena (Act IV); witnessing and mourning the human end of Troy (Act V).

38. Seneca’s prose writings support the retention of life when one is needed by others, e.g. Ep. 104.3f.

39. To complain about Seneca’s dramaturgy because Andromache in fact has no real choice is misguided. See Owen (1970a), 119; Fantham (1982a), 302 (note on 642–62). Note the following: a) Andromache can hardly be expected to give up on the spot; she calls Ulysses’ bluff and stalls for time, as she sorts out her thoughts, b) Andromache herself recognizes that she has no real choice (686f.), but comes to this understanding gradually. Seneca uses the futility of the choice as a means of resolution of the conflict in the woman’s mind. For a critic to take the resolution of a dramatic conflict as evidence to prove retrospectively the conflict was unnecessary in the first place, is idle, c) Thematic gains are made. Seneca illustrates the crippling effects of fear; cements the identification of father and child (utrimque est Hector, 659); plays upon moral terminology — fides (665f.), nefas (668), furor (670, 679); and exploits irony and paradox, d) It is hardly ‘psychologically absurd’ that a character should fret over a course of action even though the character has, in the end, no real choice. People do it all the time.

40. Seneca does not miss the opportunity to play upon this paradox in the dialogue. See 489f.; 509–12.

41. Note the repetition of misereri (‘have pity’, 694; 703; 762; 792).

42. The importance of death in the Troades has justifiably been heavily stressed recently: see Fantham (1982a), Introduction and 78–92, and Lawall (1982), 244: ‘The theme of death unifies Seneca’s Troades’. But ‘survival’ also needs emphasis, because it is not so much ‘death’ which unifies the play but a series of parallels and contrasts, including the contrast between those who are going to die and those who are going to live.

43. Owen (1970), 367.

44. The enemy are still gathering up spoils (18f.); the fires have not yet gone out (16f.).

45. These first half dozen lines also look ahead to Agamemnon’s remarks on power and fortune at 258–75 and may also be meant to allude obliquely to the disastrous return voyage of the Greek forces.

46. This passage has other implications as well. The Trojan war ends as it began in a vision of fire. Hecuba’s recollection of her prophetic dream introduces the supernatural into the play and establishes her credentials as a predictor of the future, so that when she imprecates disaster on the returning Greek fleet (1005–08), we know it will happen.

47. This is especially clear between 83–98 and 99–116. See Owen (1970), 366.

48. See Bishop (1972), 330: ‘For modesty presupposes a stabilized society in which modesty has value.’

49. I see no need to assume that this ode is to be understood as ‘outside the dramatic action’ and ‘not spoken by the Trojan women’, as advocated by Fantham (1982a), 85, 263; see also Schetter (1972), 271. The preference for death over life and the treatment of death as paradoxical are consistent elements between the first and second odes.

50. Especially Lucretius DRN. 3.830ff.

51. Ibid, lines 436; 455f.; 583.

52. See Fantham (1982a), 262–64; Lawall (1982), 247; Owen (1970a), 125.

53. This seems to be the strategy adopted by Fantham (1982a), 263.

54. Lawall (1982) has recently illustrated the presence of both views of death in the play. As he rightly claims the contradictory views are also to be found in Seneca’s prose works, e.g. Ad Marc. 19.4f. and 25. 1f. The presence of these inconsistencies in the prose works suggests to me that trying to resolve them is not a major concern of Seneca in the Troades.

55. Fantham (1982a), 356–58, seems to see the ode as predominantly positive in its theme of ‘the pleasure in sorrow shared’ (356). Lawall (1982), 249, describes the theme as ‘misery loves company’, and finds in the ode first ‘a device for distancing and removing attention from the horror of Troy’s final hours’, and then in the last section a sense of ‘almost pleasurable nostalgia’.

56. The relation of the fourth to the third ode is well-established and is clearly set out by Fantham (1982a), 324. The correspondence between these odes in length and metre does not, of course, rule out correspondences between the fourth ode and other parts of the play. The second and fourth odes share the Lucretian echoes and a more universal subject than is found in either the first or third. The images of smoke and clouds being dispersed in the air are repeated in the fourth ode (1053f.) from lines 392–95 (another recollection of Lucretius).

57. Compare Seneca’s dulce … dulce … lenius (‘Sweet … sweet … more gently’, 1009–11) with Lucretius’ suave … suave … nil dulcius est … quam (‘Pleasant … pleasant … nothing is sweeter … than’, DRN. 2.1–7).

58. These passages are compared by Fantham (1982), 126. Both describe the fading from sight of Ida (Ag. 457; Tro. 1049). Both refer to the smoke of Ilium seen in the distance (Ag. 459; Tro. 1053f.). Both have a suggestion of clouds in the distance also (Ag. 462f.; Tro. 1054). The hypothesis of Fitch (1981) on the relative dating of the plays supports the priority of the Agamemnon.

59. I retain the reading of E for line 1042. Questum is not inappropriate as is argued by Fantham (1982a), 362; the ‘complaint’ or ‘grievance’ refers back to the state of mind described in the preceding parts of the ode.

60. It is passivity which differentiates the ritual and lamentation of the last act from the ritual and lamentation at the close of Act I. However there is meant, I think, to be seen here a formal reiteration of the motif from the start of the play.

61. What is stressed by Seneca is not the difference in these scenes between Trojans and Greeks, but their involvement in the same emotions and responses despite their differences. Hence I cannot agree with La wall (1982), 251: ‘Further complexity is introduced by the careful distinction between the vulgar gawking and petty emotionalism of the Greek spectators and the piteous terror that grips the Trojans who watch the final Act in the death of their city.’

62. I should like to thank the editor of this volume and Kate Kornicki for their most helpful criticisms of an earlier draft of this essay.