1. Verse-dialogue: Thy. is far more flexible than Ag. in the handling of speaker-changes and in the manipulation of the trimeter line — see Fitch (1981), 289ff. Lyric Metre: Ag. with its polymetric odes, based on the principle of deriuatio (as in Oed.), is more complex and ambitious metrically than Thy, Prologue Integration: Thy. prologue is more integrated thematically and imagistically with its play than Ag. prologue — see Calder (1976a), 29f.; Shelton (1977), 33ff. Chorus-Act Interrelation: Ag.’s choruses are less dramatically/thematically functional than Thy.’s — Ag.’s final chorus is only of minor relevance. Choral Persona: no choral persona is developed in Ag. until the second ode — the first ode is of the ‘authorial voice’ type, see Tarrant (1976), 180ff.; in Thy. choral persona is established immediately. Dramatic Language and Imagery: more consistently structured and developed in Thy. Messenger’s Speech: the lengthy, uninterrupted, leisurely messenger’s speech of Eurybates in Ag. Act 3 contrasts sharply with the interrupted, episodic messenger’s speech in Thy. Act 4, the parts of which are designed so as to form their own mini-drama. Dramatic Structure: Ag. has a diptych structure, is episodic, discontinuous; Thy. has a more continuous dramatic movement — with a crescendo effect from the end of the third chorus through the Messenger’s speech and the Star chorus into the climax of the final act. On dramaturgical variety in Senecan tragedy as a whole see Boyle (198-), Section I, n. 16.

2. Thyestes is virtually a two-actor play (only in Act 3 are there three speaking roles) and has a single chorus of Argive nobiles (119). Agamemnon has two choruses — Argive girls (who presumably also speak the opening ‘authorial voice’ chorus) and Trojan women; its final scene involves four speaking parts.

3. For an excellent study of Seneca’s indebtedness to postclassical dramatic conventions see Tarrant (1978), 217ff.

4. See Boyle (198-), Section III.

5. Shelton’s interpretation (1977), 36f., of the ghost of Thyestes as ‘a character who represents the psychological force which motivates the plot action’ seems to simplify the dramatisation (note the self-loathing, 4, 23ff.; the preference for the grim pools of Hell over the upper world, 12ff.) and to focus on only a part of the ghost’s function. Tarrant (1976), 158, takes the ghost as primarily ‘a vehicle for the rhetorical expression of character’.

6. The text of Agamemnon is from Tarrant’s edition (1976). All translations in this essay are my own.

7. Despite Tarrant (1976), ad loc, alterno does not mean ‘retributive’, although it may be used of retributive situations. It means ‘alternating’, ‘answering’, ‘reciprocal’, etc. — see OLD.

8. Contrast Henry and Walker (1963), 5: ‘No attempt is made … to relate the action of the play to the working out of any demonic design by Thyestes or the di inferi, or indeed to regard the horrors of the present as influenced by the horrors of the past’ (my emphasis).

9. On the development of the association of sea-storm and fortune in Agamemnon see Pratt (1963), 225f.

10. The only tolerable explanation of 239–309: so Croisille (1964), 466–69, and Calder (1976), 31f. The need for Clytemnestra’s testing of Aegisthus has been dramatically prepared for (48–52, 226–33). Contrast Tarrant (1976), pp. 217 and 230, who regards the sudden reversals of 239ff. and 307f. as indicative of Seneca’s absence of interest ‘in the realistic depiction of change or development’. Tarrant’s own thesis (ad 302ff.), that Aegisthus is manipulating Clytemnestra, seems to give insufficient consideration to the following: i. The prior presentation of Aegisthus as fearful and weak (48–52, 226–33) and its confirmation both in the dialogue with Clytemnestra itself (his strident rhetoric seems to betoken fear rather than ‘superiority’) and in Clytemnestra’s account of Aegisthus’ role (the semiuir) in Agamemnon’s murder (890f., 904ff.); ii. The prior presentation of the extent and force of Clytemnestra’s loathing of Agamemnon and its grounds (esp. 162ff.); iii. The prior focus on Clytemnestra’s decision to resort to feminei doli (116); iv. The suddenness and the commanding tone of the resolution to action with which Clytemnestra, not Aegisthus, closes the scene. A final rhetorical question: what member of what audience could possibly take seriously Clytemnestra’s opening remark to Aegisthus, amor iugalis uincit (‘marital love vanquishes me’, 239), after her preceding outburst to the Nurse at 162ff., itself a development of Clytemnestra’s memory of another ‘marriage’, Iphigenia’s ‘marriage’ with death (iugales filiae memini faces, 158)? Agamemnon has shown her what amor iugalis and fides mean.

11. See Lohikoski (1966), 63ff.; Seidensticker (1969), 119ff.; Anliker (1972), 451ff.; Lefèvre (1972a), 470ff.; Calder (1976), 31ff.

12. For the strongest case (too strong) in support of Agamemnon’s guilt see Lefèvre (1973), 64–91. Agamemnon’s guilt has, however, to be understood within the context of history’s determinism.

13. Cf. Seidensticker (1969), 128 n. 157, who sees in Eurybates’ narrative an account of the destruction of the Greek fleet ‘als erster Vergeltungsschlag für die Zerstörung Troias und als Präludium für die Ermordung Agamemnons’.

14. Anliker’s thesis (1972), 454, that the Strophius scene is ‘grausamer Hohn’ makes little sense and ignores the obvious truth-value of the scene’s symbolism. Orestes will in fact triumph.

15. Contrast Henry and Walker (1963), 9: ‘It is asserted as it is in the closing speeches of the Agamemnon that destructive and insane forces will prevail over any principle of justice and order in things.’

16. The relationship between human responsibility and history’s determinism was a notorious difficulty in Stoic philosophy. For an intelligent discussion see Long (1971), 173–99. Long notes (p. 187) that while character to the Stoics was a proximate cause of action and good and bad men were ‘men who deliberately in consequence of their characters choose actions meriting praise or blame’, an ‘individual’s character’ was itself ‘determined since it follows from his particular nature and upbringing’, ‘the fully developed disposition’ being ‘a necessary consequence of heredity … and environment’. See also Sandbach (1975), 1O1ff.: ‘the determining causes’ of action (assent, impulse, character, the condition of the psyche) ‘are themselves determined’; ‘Fate … determines what he [a man] is’ (103).

17. As in the prose-works where nature, god, fate command and man perforce obeys: ducunt uolentem fata, nolentem trahunt (Ep. 107.11). For the interchangeability of the names, fatum, fortuna, deus, natura, Iuppiter, for the controlling force of the universe see De Ben. 4.7–8, Ep. 107.7ff. The argument of Lefevre (1972a) that the first half of Agamemnon is ‘eine Leidenschaftstragödie’ (p. 474), in which Agamemnon is responsible for his own fate and all external influences on Agamemnon’s actions have been eliminated (see esp. pp. 461, 464), and the second part ‘eine Fatumstragödie’ focussing on Cassandra and built on new premisses about man’s responsibility for his own fate, seems extraordinarily wrong-headed. It ignores the dramatic and thematic function of the prologue, misinterprets the first choral ode and Clytemnestra’s speech of 162ff., and misses the significance of Agamemnon’s Schicksal as product not simply of the sacrifice of Iphigenia and of Agamemnon’s behaviour at Troy, but of the unimpedable rotation of fortuna and Thyestes’ monstrous feast. It is worth noting that contrary to Lefèvre (pp. 460ff.) the three lines on which his case essentially rests, 87–89, do not mean that human beings are responsible for their own downfall, as the conclusion of the appropriate section of the first chorus (100f.) makes clear.

18. Tarrant (1976), 231, who denies Agamemnon ‘organic structure’ (p. 4), seems to ignore the obvious, ironic connections between this victory ode and the events of the play. See, however, Seidensticker (1969), 131 n. 163, who correctly observes the ‘tragischen Ironie’ of this ode, even if some of the particular ironies which he notes are less than self-evident.

19. Calder’s thesis (1976), 33f., that Ajax is ‘the antithesis of Cassandra’ and intended exemplum of uitae dints amor (590) is difficult to accept. The treatment of Ajax is heroic, its emphasis upon Ajax’ ability to endure (inuictus, ‘unvanquished’, 532; nil ille motus, ardua ut cautes, ‘utterly unmoved, like a towering crag’, 539 — cf. Aeneid 7.586 of Latinus); and the connection between his behaviour and, on the one hand, the Trojan chorus’ praise of contemptor leuium deorum and, on the other, the presentation of Cassandra seems marked. Nor seems there anything in the text to support Calder’s claim that all Ajax wins is ‘the briefest extension of an intensely painful existence’ (my italics). Lefèvre’s interpretation (1973), 82 — ‘Er ist somit der Prototyp des in voller Verantwortlichkeit handelnden hybrishaften Menschen’ — misses the tragic splendour (caeco mari/ conlucet Aiax, ‘in the dark sea/ Ajax shines’, 542f.) of Ajax’ triumphal contempt.

20. He seems to have read the same handbook as Atreus(Thy. 247f.)and Lycus(H.F. 511f.).

21. There is no need to punctuate with a comma after furiosa, as Tarrant (1976) and others. Even so the epithet furiosa seems more plausibly construed as Clytemnestra’s description of Cassandra’s present behaviour rather than some general reference to Cassandra’s experience of prophetic furor (as Tarrant).

22. As Calder (1976), 36, astutely observes: ‘Furor picks up furiosa and means “what you call madness, the desire for death, will in time come upon you both”. In the Senecan manner the metus of the regicides will teach them in time the meaning of mors libera.’ Cf. Anliker (1972), 452f. Of course Cassandra’s death-wish is not the ‘unreflecting tendency towards death’ (ad moriendum inconsulta animi inclinatio) criticised in Ep. 24.25. What the latter passage suggests, however, which both Petronius and Tacitus confirm, is that the death-wish, though a debating topic in the schools (Sen. Elder Contr. 9.6.2), was also index of an age. See Boyle (198-), Section V.

23. Not stabbed as Anliker (1972), 453 n. 7: ‘Es hindert übrigens nichts … dass Klytäm-nestra mit diesen Worten furiosa morere Kassandra ersticht.’ Cassandra is in the process of moving from the stage (trahite, ‘drag her off’, 1003; nihil moramur, rapite, ‘I do not delay, take me’, 1010), as Calder (1976), 35, observes.

24. It is certainly not the nadir of Senecan pessimism that Lefèvre (1972a), 475, argues for: ‘Dergestalt ist die Deutung des Agamemnon-Schicksals eine der pessimistischsten Aussagen, die Seneca in seinen Tragödien gemacht hat … des überaus negativen Weltbildes des “Agamemnon”.’

25. The text of Thyestes is that of Giardina (1966).

26. This is not to say that all Senecan drama focusses on this, as Poe (1969), 360, seems to suggest.

27. Steidle’s criticisms (1972a), 497–99, of the inconsistencies in Tantalus’ behaviour do not bear inspection. The subjunctives of 19f. do not imply that Tantalus wants his descendants to perform their abdominable deeds. He knows what the future will be and sees his responsibility for it. He nevertheless attempts to resist that future, as 67ff. attest.

28. The Tantalus figure is of course a negation of the De Ira image of man as by nature inclined towards kindness, harmony, mutual love and succour (see, e.g., De Ira 1.5.2–3). Certainly there are those kinds of impulses in Tantalus (90–95), but they prove useless against the constraints of fames infixa. Against those who would see the prologue’s ghost as an emblem in the play of ‘unnatural man’ along the lines of De Ira, the dominant association of Tantalus’ and Atreus’ furor with natural human appetites, fixed and, in Atreus’ case, inherited, seems telling, as does Tantalus’ status as proto-man (hence his association in certain contexts with Prometheus: e.g. Horace Odes 2.13.37f., 2.18.35ff.) and both protagonists’ instantiation of the Tantalus paradigm. Aetema fames and aeterna sitis (149f.) seem presented in the play less as human aberration than as defining human attributes. See also note 30.

29. The view, for example, of Detienne (1972), 71–113. Cf. also Segal (1977), 104: ‘Sacrifice creates a series of mediations between god and beast and god and man and thereby asserts an orderly distinction of these planes of existence in the biological and alimentary codes.’

30. There is another matter. Poe (1969), 364f.: ‘The very length of the description of Atreus’ furor shows that Seneca sees in it not just a matter of individual wickedness to be condemned objectively or modified upon, but an instinct which he to some degree shares and which he expects his audience to share.’ There are of course those who see no important dramatic function in the ‘piling up of horror’: e.g., Costa (1974), 110.

31. The beast’s triumph would not surprise Freud. See especially Civilization and its Discontents (Strachey ed. 1963), 48f.

32. Henry V, Act 4, Sc. i.

33. The Satellite’s remark at 334f. means precisely the opposite of what it says. As Calder (1976/77), 8f., notes, a slight pause before the delivery of the final word, fides, would make it clear that it is the other word, timor, which is meant.

34. He is not the ‘stoischer sophos’ of Gigon (1938), 176ff., nor the noble Stoic of Herington (1966), 458ff., nor the ‘wandering ascetic sage’ of Poe (1969), 360. His Stoic attitudes are product of his circumstances, not index of his nature: the past which he acknowledges (513ff., 1103), the present which he displays (404ff.), the future which he reveals (520ff., 780ff., 909ff., 948, lllOf.) testify to that. Even Seidensticker (1969), 104ff., who rightly attacks the ‘stoischer sophos’ interpretation and observes how Thyestes’ opening words in Act 3 seem to confirm the prologue’s presentation of him and Atreus’ own evaluation (289), maintains that Thyestes’ subsequent words in this scene show Atreus to have been wrong (‘dass Atreus sich geirrt hat’, p. 105) and reveal the opening words themselves as but a momentary lapse into previous ways (i.e. not evidence of a deep-seated impulse). Seidensticker’s thesis seems contradicted both by the details of Thyestes’ behaviour in this act and the latter’s relationship to the prologue. Knoche’s analysis (1972), 483–86, is much better. Steidle’s criticisms of Knoche (1972a), 494 n. 14, ignore the text they cite, as does his attribution to Thyestes of ‘eine verzweifelte Resignation’ (p. 493). Poschl (1977), 230, similarly misdescribes.

35. So at the end of the play Thyestes does not deny his earlier scelus (1103).

36. The attempt of Paratore (1957), 56, in what is nevertheless a most stimulating essay on Senecan tragedy, to read Thyestes’ behaviour as prompted by his fatherly concern (‘dalla brama disperata … di stornare dal loro capo la vendetta [quale che sia] che sente incombere’) seems to make little sense and ignores Thyestes’ opening soliloquy and the prologue; as does Steidle’s explanation (1972a), 495, that Thyestes is overcome by Atreus’ words of greeting (‘die für echt genommenen Begrüssungsworte des Atreus’).

37. Significantly perhaps Thyestes’ behaviour dramatises the futility of Seneca’s own advice at Ep. 8.3: ad omne fortuitum bonum suspiciosi pauidique subsistite; et fera et piscis spe aliqua oblectante decipitur. munera ista fortunae putatis? insidiae sunt. (‘Halt before every good that chance brings with suspicion and fear; wild beasts and fish are tricked by tempting hopes. Think these things fortune’s gifts? They are traps.’) Thyestes knows all this (see esp. 472: errat hic aliquis dolus, ‘some trick strays hereabouts’) and more. It affects his behaviour not a jot. He becomes the beast in the trap (plagis … fera, 491). If it is true, as Marti (1945), 240, claims, that ‘his suffering in exile has taught him much’, it is also true that such teaching has no significant effect on Thyestes’ actions and behaviour. See Knoche’s sensitive analysis (1972), esp. 484f.

38. See Poe (1969), 369.

39. Contrast Knoche (1972), 488: ‘Aber im letzten Vers, den der vernichtete Thyestes spricht, kündet sich die Wahrheit wieder an.’ This is simply false, and negates Knoche’s thesis concerning the ‘morality’ of the play. It is pointless to argue, as Knoche does, that for Atreus to have renounced ‘die menschliche Natur’ is the worst punishment of all, for two simple reasons: i. The play subverts traditional notions of what constitutes ‘die menschliche Natur’ and presents Atreus as exemplum of something innate in man (fames infixa); ii. It is clear that Atreus is satisfied (1096ff.) and does not — and will not — suffer.

40. On the play as ‘contest’ see Braden (1970), 30ff.

41. Knoche’s Stoicising thesis (1972), 478ff., that Atreus inhabits a world of delusion (‘Welt des Wahnes’) misreads. In the play Atreus is presented as in important respects not deluded. Though his fears concerning his brother in Act 2 are dramatised as manic — Atreus is in a state of mental transport — his evaluation of Thyestes’ regni furor (302; cf. 288ff.) and accusations against him are confirmed by Thyestes’ words and behaviour in Act 3 (and Act 5), his vengeance brings him satisfaction and triumph (1096ff.), and, unlike Thyestes, he predicts the truth (1112). Even if the Stoic standpoint dictates that he live in a world of total delusion, the play asserts the contrary; just as it asserts that Atreus, not the second chorus’ Stoic paradigm, is the true king.

42. Not so Knoche (1972), 489, for whom surprisingly the Fury of the prologue is ‘das Recht der Weltordnung’.

43. Note the ironic verbal counterpoint here with the prologue: arceat … scelerum recall Tantalus’ futile arcebo scelus (95); altemae … ne … uices recall the Fury’s prophetic command, alterna uice … (25). The counterpoint points up the naivety of the choric prayer.

44. Thyestes’ second prayer is not, however, for a ‘cataclysm’ (so Poe [1969], 371 and 375).

45. As the fulfilment of Medea’s prayers and the relation of Jason’s suffering to ‘cosmic crime’, its position in a complex, quasi-moral scheme of cause and effect, both cosmic and human, attest. It is not accidental that the first and last word of Medea are the same.

46. On Seneca’s transformation of ‘celestial commonplace’ into ‘dramatic symbol’ in Thyestes see Owen’s intelligent remarks (1968), 296–300.

47. An interpretation of course articulated in the Stoic doctrine of sympatheia tōn hōlbn.

48. The reaction of nature is dramatically underscored by Seneca’s presentation of the sun’s disappearance through four separate experiences of it, that of the Messenger, the Chorus, Atreus and Thyestes — see Owen (1968), 297. In each case the reaction is all. Nothing follows.

49. See also Poe (1969), 364.

50. With Thy. 313f. cf. especially the emphasis on the imperatives of Aegisthus’ birth at Ag. 48 and 233. Cf. also Phaedra’s enigmatic quod uiuo (‘That I live’) reply to Theseus’ question at Phd. 879f., on which see Boyle (198-), Section V.

51. On this see Poe (1969), 366f. Relevant here is the Thracian nefas motif. As Tantalus’ crime is regenerated with increased horror, so Thracian nefas is re-enacted with greater number (maiore numero, 55f.; cf. 272ff.).

52. Of the endings of extant Senecan plays that of Medea is closest, but still not close. Its theatrical closure and assertion of at least ambivalent ‘cosmic’ or ‘moral’ order contrast with Thyestes’ end. Signally Medea triumphs at the cost of her humanity; Atreus triumphs because of his.

53. Cf. Fantham’s attempt (1982) to establish a pattern of cross-references from Agamemnon to Troades, which, if Troades is the earlier play, would give added significance to particular features of Agamemnon. Fantham’s more recent opinion (1982a), 14, that Agamemnon may in fact be earlier than Troades, points up an aspect of the difficulty of the task.

54. There are those who have denied this. Fantham (1982), 121: ‘There is little coincidence of phrasing between the two plays [Agamemnon and Thyestes].’

55. Herzog (1928) also attempts to provide a chronology for Seneca’s plays based upon topical allusions, which has however attracted little credence.

56. Calder (1976), 28–30.

57. Leo (1878), 133. Tarrant (1976), 7, also regards the early dating of Agamemnon as ‘inherently plausible’.

58. Shelton (1977), 39: ‘I think that the opening scene of Agamemnon represents an early stage of Seneca’s work as a dramatist. We can see in this play the first attempts to control techniques which he refined and developed, and therefore employed more successfully, in Thyestes.’

59. See note 1 above.

60. Some examples: the important feast/banquet image of the prologue (11, 21, 48), which reappears in Cassandra’s description of Agamemnon’s death (875ff.), is not developed (as in Thyestes) in the intervening scenes; the storm image, important in the first half of the tragedy (63ff., 90ff., 109, 138ff., 465ff., 593ff.), is not even employed in the prologue and disappears after the Trojan chorus, failing to reappear even in the account of Agamemnon’s death (contrast Pratt’s uncritical evaluation of the storm image [1963], 225f.); the sacrifice image, which is developed in the main body of the play (162ff., 219, 364ff., 448, 584f., 644f., 775ff., 806ff., 898ff.), is hardly used in the prologue at all — daturus coniugi iugulum suae (‘to yield his neck to his wife’, 43) is the only possible occurrence, and even that is not clearly an occurrence: see Tarrant (1976), ad loc.

61. Most notorious is the minor thematic relevance of the Hercules chorus (808–66) at the climax of the play — despite the obvious verbal act-chorus ‘joins’ or Stichworte (Argolica Iuno, 806, Argos … nouercae, 808f.; totidem diebus/Troia quot annis, 865f., … par annis decern, 867). See Tarrant (1976), ad loc. On the other hand, the imagistic, thematic and dramatic functionality of all Thyestes’ choruses seems evident.

62. See Shelton (1977), 33–43.

63. Fitch (1981). On Fitch's case see Fantham (1982a), 14.

64. Attempts, however, to identify Seneca with Thyestes or with the Satellite (or with both — so Pöschl [1977], 233) seem misguided. Thyestes is no crude biographical allegory. The issues it dramatises, the levels of significance at which it operates, signify the play’s symbolic, not allegorical nature.