17

The Watchman's bed is not ‘supervised’ (13 ἐπισκοπουμένην) by dreams; instead, fear ‘stands in attendance’ (14 παραστατεῖ). The images are medical. He is ill; dream-filled sleep would be a good doctor, but the bad doctor fear is already on the job, preventing him from sleeping well.Footnote ¹

The thought and medical imagery are resumed after an anacoluthon:

Apparently, straight after complaining that fear prevents sleep, the Watchman describes how he sings to stave off sleep; whereas dreams were to be a welcome doctor, now sleep needs a remedy.

There is independent evidence that ὕπνου is intrusive.Footnote ²POxy. 18 2178 (+2nd c.) preserves one letter-top and two accents for line 17 – an acute followed shortly by a grave. The set of Aeschylean papyri from Oxyrhynchus gives a moderate number of diacritics, though very few grave accents. Usually, it seems, accents give guidance where pronunciation could be difficult.Footnote ³ The most plausible reconstruction is that ἀντίμὸλπον was written to stop readers from seeing ἀντὶ μολπ- then having to backtrack. If so, the spacing in the ancient book implies that the line began τόδ' ἀ]ṿ[τ]ί̣[μ]ὸ̣[λπον.Footnote ⁴

West comments ‘vocem fort. om.’, but the papyrus may have continued ◡◡◡ ἐντέμνων ἄκος. For example, τόδ' ἀντίμολπον <αὐτὸς> ἐντέμνων ἄκος would produce the following logic: ‘Sleep normally cures cares, but I cannot sleep; I therefore try to cure them myself.’ The lack of a genitive defining ἄκος may have caused the corruption. As Σ_vetAg. 16–17, wrongly inspired by the proverb ‘to sing on the watch’, glossed ἀντίμολπον as τὸ ἐναντιούμενον πρὸς τὸν ὕπνον, so an early reader may have written an explanatory ὕπνου in the left-hand margin.Footnote ⁵ This would have subsequently intruded, with αὐτός ousted to preserve metre.

This point affects, though not decisively, the interpretation of ἀντίμολπος. This is normally derived from ἀντιμέλπω, to mean ‘involving μολπή-in-opposition’ rather than ‘substituting for μολπή’. A possible objection is that the Watchman's singing is a long way from the usual sense of μολπή – a public ritual song with dancing.Footnote ⁶ However, if his singing is meant to comfort, and to ward off fear rather than fatigue, its overlap with the traditional set of functions of μολπή is greater.

Finally the Watchman's singing, if ὕπνου is replaced, is designed to cheer but cedes quickly to pessimism. This accords with a pattern pervading the whole tragedy, including the Watchman's later comments at 32–5.

69–71

Fraenkel's commentary is very useful, but unconvincing in its final suggestions. The manuscripts offer:

ἄπυρος ‘uncooked’ is in the context of sacrifices a technical term referring to unburnt offerings of fruit, cakes, etc.Footnote ⁷ With this sense, the ἄπυρα ἱερά cannot attach to ὀργάς.Footnote ⁸ The most plausible place to emend is δακρύων, which looks like an intrusive gloss on ὑποκλαίων. In that case, our emendation need not stay palaeographically close to δακρύων. Simply excising οὔτε δακρύων (so e.g. Page) has two problems: firstly, it presupposes the incorporation of a more complicated gloss; secondly, taking ἀπύρων ἱερῶν as partitive genitive after –λείβων would limit the ἱερά to libations, which the parallels do not warrant.

ὑποκλαίων, whether glossed by δακρύων or not, is itself likely to be an early error. Casaubon's ὑποκαίων produces a far better disjunction with the libations and ἄπυρα ἱερά, where a reference to animal sacrifice is much needed.

If we have straightforward references to burnt sacrifice and to libation, I think we should favour a similarly direct emendation at οὔτε δακρύων. This tells against Fraenkel's suggestion οὔτ' ἐπαοιδαῖς. The suggestion οὐδ' ἀκροθίνοις (H. Funke, ‘Aischylos Agamemnon 70’, MH 33 [1976], 122–4) is better semantically, though overly led by the ductus litterarum of δακρ−. I suggest οὔτε or οὐδὲ τραπέζαις. Unburnt offerings were frequently placed on τράπεζαι, and the word would interact with the τράπεζα of Menelaus' hospitality (401–2, 701–2), abuse of which has necessitated such attempts at atonement.Footnote ⁹ For οὔτε … οὔτε correlating a participle and something other than a participle, cf. 496, Septem 664–6, Eumenides 411–12.

οὔθ' ὑπολείβων is generally altered to οὔτ' ἐπιλείβων (Schütz). However, ὑπολείβω is now known from the contemporary sacred law of Selinous, A.9–14:

These libations are poured underground through a hole or pipe in the roof of a structure which represented the tomb of ancestor spirits.Footnote ¹⁰ Something similar is possible in Agamemnon, and introduces a reference to ‘chthonic’ deities which broadens the range of sacrificial activities envisaged. This characterizes the sacrificer as more desperate, and ties in with Agamemnon's concentration on the Erinyes as Zeus' angry and implacable agents.

217

According to MV, Agamemnon ends his deliberation speech εὖ γὰρ εἴη.Footnote ¹¹ The main uses of γάρ would require that the content and/or form of the preceding words can be explained by the wish ‘May things be well’. However, the preceding words, which say either that the army are right to demand Iphigenia's sacrifice or that Right forbids it, seem to fail this requirement.Footnote ¹² One possibility might be that Agamemnon (as re-enacted by a chorus member) is supposed to nod assent to Calchas' μῆχαρ βριθύτερον πρόμοισιν (199–200) immediately before the word εὖ. But the insistent rhythm – whether a run of choriambs or ionics – suggests otherwise, and it is rash to specify how the Chorus presented Agamemnon's speech.

At Euripides, Cyclops 261 L's γάρ has been emended by Kirchhoff (followed recently by Diggle, Seaford and Kovacs) to γ' ἄρ', emphasizing the preceding word in a wish: κακῶς γ' ἄρ' ἐξόλοι'. This is also the least intrusive emendation of R's τ' ἄρ' at Aristophanes, Thesmophoriazusae 887, adopted in Wilson's recent OCT: κακῶς γ' ἄρ' ἐξόλοιο. This particle combination is preferable in Agamemnon's wish too: it does not explain his previous words, but emphatically tries to quash his despair.Footnote ¹³

783: STAGING

Cassandra enters, sitting (cf. 1054 θρόνον) in a wagon (1039 ἀπήνη, 1054 ἁμαξ–); she is in prophetic garb, and carries a staff (1264–5). It is becoming a commonplace that she entered in the same vehicle as Agamemnon, that this reflected the contemporary Attic practice of a groom driving his bride home beside him in a cart, and that this formed a significant part of what motivates Clytemnestra to kill Cassandra.Footnote ¹⁴ However, at least one ancient staging tradition, known to the author of the hypothesis, had Cassandra enter in a second wagon with Agamemnon's spoils. I hope to show that the ‘same wagon’ view deserves a good deal of caution, though certainty about how things were done in 458 b.c. is of course impossible.

Cassandra is actually a captive concubine, and nobody – including Clytemnestra – speaks of her as a rival bride. Aeschylus had plenty of opportunities to introduce verbal imagery or reference to Cassandra's ‘marriage’ to Agamemnon, but did not.Footnote ¹⁵ The point that a strong visual allusion to marriage would have lacked coherence with Aeschylus' verbal concerns may be aggravated by a further speculation. Agamemnon's entry is brilliantly prepared by the end of the second stasimon, particularly his cue πᾶν δ' ἐπὶ τέρμα νωμᾶι [sc. Δίκη]. The focus on the vain superficiality of wealth in the Chorus's preceding lines (773–81) packs more punch if Agamemnon entered with some spoils.Footnote ¹⁶ This might well call for a second wagon, in which Cassandra, as a form of booty, would ‘belong’. The odd bridal imagery if she instead entered in Agamemnon's wagon would be all the more egregious.

Secondly, the focus of attention up to 950 should be almost entirely on Agamemnon and Clytemnestra, and on the battle of wits to which Aeschylus devoted such a subtle passage of stichomythia – the only one between two actors in the play. Cassandra's august presence cannot be totally occluded, but I find it detrimentally distracting if she is centre stage beside Agamemnon.

Thirdly, Agamemnon's wagon needs to be close to the centre of the orchêstra, so that he can step directly on to the path of crimson garments which leads to the central door, as Clytemnestra bids at 905–6, and so that the path is long enough for his exit to last from 957 to 972. Unless the wagon is subsequently removed by attendants, it will restrict the Chorus during the third stasimon. But, if Cassandra is still in it, it cannot be fully removed: it would, contrary to any narrative logic, have to be moved to the edge of the orchêstra for the stasimon, and then off during the fourth episode. The ‘same wagon’ view thus entails that the Chorus might dance in a restricted orchêstra, and/or that verisimilitude was not observed in removing the wagon. Neither implication is absurd, but it is worth noting explicitly that something as unfathomable to us as an ancient choreographer's vision might well have needed Cassandra to be in a second wagon, not central, though in a position which can become focal when Clytemnestra re-emerges.

Finally, the third episode of Agamemnon is recast in Euripides, Electra 988–1146, which appears to involve the entrance of two vehicles – a splendid chariot carrying Clytemnestra which is consistently called an ὄχος (966, 999, 1135–6), and a more humble wagon for her slaves (998: Clytemnestra bids ἔκβητ' ἀπήνης as she had bid ἔκβαιν' ἀπήνης τῆσδε at Ag. 906, 1039). There are numerous correspondences, to which the rest of this paragraph is devoted.Footnote ¹⁷ Euripides' chorus welcomes Clytemnestra with extreme, ϕθόνος-inspiring praise (994, containing σεβίζω), not as Aeschylus' chorus had welcomed Agamemnon (Ag. 785–7 πῶς σε σεβίξω | μήθ' ὑπεράρας μήθ' ὑποκάμψας καιρὸν χάριτος;) but as Clytemnestra herself had (895–905). Clytemnestra's maids are Τρωιάδος χθονός | ἐξαίρετ' (El. 1001–2), as Cassandra was ἐξαίρετον (Ag. 954). The question of how Clytemnestra will leave her vehicle is raised, as it was for Agamemnon, and Electra urges a solution that is similarly charged with symbolism – that she herself will help her mother down as a sign of Electra's ‘exile’ into a slave-like state (El. 1004).Footnote ¹⁸ The slaves are, like her, ‘left orphaned of their father’ (1010), to which Clytemnestra retorts ‘Well, such (τοιαῦτα) were the plans your father laid against those of his ϕίλοι who were least suitable’ (1011–12). Her intended meaning is that Agamemnon deserved the assassination to which Electra had alluded, but there is dramatic irony: since Electra is currently plotting against Clytemnestra, the audience may hear τοιαῦτα … βουλεύματα as a reference to Electra's plot, which is in fact a close analogue of Clytemnestra's own. Euripides asserts his independence from the Aeschylean model in delaying Clytemnestra's entrance with a characteristic agôn, where bold mythological innovations are put to the service of argumentation.Footnote ¹⁹ But Clytemnestra's exit does return to the model, with Electra's ambiguous talk of a sacrifice to be performed (1125–42; cf. Ag. 1056–9), reference to fine clothes besmutted (1140, also recalling the δύσκαπνα δώματα where Justice is to be found at Ag. 774), and gloating when Clytemnestra finally disappears into the house where she will die (1143–6). The following dochmiacs (beginning ἀμοιβαὶ κακῶν) explicitly draw attention to the reworking of Clytemnestra's assassination of Agamemnon. This allusive engagement is strengthened further if Aeschylus had used two vehicles.

984–7

MVG are lacking. For τ one can reconstruct:

My aim is not to solve this crux, but to suggest two specific possibilities for 984, starting from a widely held assumption about the more intractable 985.

Fraenkel argued convincingly that χρόνος ἐπεί alone does not mean ‘It is a long time since …’. I do not think it beyond Aeschylus to say that ‘Time has passed its prime since’ a certain thing happened, meaning that the event is many years in the past and hinting that its effects on the present are enfeebled.Footnote ²⁰ If so, the structure χρόνος παρήβησεν ἐπεί … also suits the flow: the Chorus should be able to drop their anxieties about the Greek expedition now that it is over, yet they find themselves unable to.

That does not prove that χρόνος παρήβησεν and ἐπεί are right, but this is the working assumption I shall make for the paragraphs which follow. We now need a verb for the temporal clause, and the obvious location is ψαμμίας ἀκάτα, which neither makes sense nor responds metrically with 998 (εὔχομαι δ' ἐξ).Footnote ²¹ Page plausibly explained the corruption via incorporation of an interlinear variant.Footnote ²² I prefer to think that these syllables started with a form of ψάμμος rather than ψαμμίον, or ψάμμι(ν)ος whose existence is shadowy. The main problem with this approach is that all proposals remain difficult palaeographically, but let us proceed with due hesitation to focus on the other words.

ὑπ' Ἴλιον suggests not only motion towards Troy, but approaching it from below and so nearby; I.C. Cunningham usefully compared the Iliadic formula ὑπ' Ἴλιον ἦλθεν⁺.Footnote ²³ The ἐπεί-clause would therefore have described the Greeks' arrival at Troy, and in particular – it seems – the effect on sand of something involving sterns or stern cables. Ships were rowed up stern-first on to the beach, the anchors thrown out, then other stern cables thrown out and tied up. If we retain πρυμνήσια ‘stern cables’, the smallest viable change of ξυνεμβόλοις is ξυνεκβολαῖς (‘simultaneous overboard-throws’) or perhaps ξὺν ἐκβολαῖς. Alternatively, if Aeschylus could use πρυμνήσιος to qualify something other than a rope, then πρυμνησίαις ξυνεμβολαῖς could refer to the sterns simultaneously ramming into the shore.Footnote ²⁴ ξυνεμβολαῖς would easily be corrupted to ξὺν ἐμβόλοις (perhaps still in capitals, before word division), given that ἔμβολον is a nautical term. With either ramming or throwing out ropes, the sand might be (for example) compacted or, as in Wilamowitz's conjecture ψάμμος ἄμπτα, splashed upwards.Footnote ²⁵

πρυμνησίαις ξυνεμβολαῖς prompts a further speculation. The word ναυβάτας may mean ‘(currently) travelling on a ship’ or ‘which travelled on a ship’. The εὖτε-clause may correspondingly describe the navy's arrival in the Troad as like the land army's subsequent assault on the city walls. With ἐμβολή, the ships ram the shore as a land army tries ramming the citadel's gates. Possibly, if Wilamowitz's ψάμμος ἄμπτα is right, the curious focus on sand now finds an explanation: it contributes to the conflation, by echoing the dust cloud thrown up by an approaching land army.

1001–4

First compare τ with, in the right-hand column, the scansion of the responding verses (1018–21), which appear to be sound:

As observed in West (n. 12), 207–9, τᾶς πολλᾶς is probably a gloss on ἀκορέστου which intruded and disrupted the whole passage; γάρ (1003) is frequently added by scribes at an asyndeton. West would replace ὑγιείας too, but it is conceivable that the diphthong scanned short (cf. 147 Πᾰιᾶνα).

At its most general, the underlying thought of this difficult strophic pair appears to be that both health and wealth are dangerous in excess. In 1003–4 health and sickness are next-door neighbours; sickness pushes against their dividing wall, as health will if it gets too big for itself. There is therefore a permanent danger of disease encroaching. Commentators have compared Hippocrates, Aphorisms 1.3 ἐν τοῖσι γυμναστικοῖσι αἱ ἐπ' ἀκρὸν εὐεξίαι σϕαλεραί, ἢν ἐν τῶι ἐσχάτωι ἔωσιν. A.Y. Campbell (The Agamemnon of Aeschylus [Liverpool, 1936]) suggested introducing σϕαλερόν into Aeschylus, alongside characteristically radical changes (τὸ θαλίας ἀκορέτου σϕαλερὸν τέρμα). A less drastic measure might be:

τέρμα is both the ‘limit’ of health-as-quality and the ‘boundary’ of health-as-householder, so its adjective should mean both something like ‘dangerous’ and something like ‘fragile’. σϕαλερόν satisfies this criterion, and as a metaphor interacts with ἐρείδει, which carries the implication of a sturdy footing.

1063: STAGING

Why is Cassandra's manner ‘like that of a newly captured wild animal’? Several recent scholars have inferred that her body is quivering (~ the animal's spasms of fear), and so that she is entering a prophetic seizure.Footnote ²⁷

I am uncertain that a Greek audience would have understood the manner of a newly captured wild animal to be spasmodic. Arrian's Cynegeticus, for example, repeatedly refers to hunted animals realizing that they cannot escape and giving up: a hare cowers (16.5); a deer gawps (23.4); a wild ass is netted and led home (24.3). More importantly, there are overwhelming reasons for Cassandra to remain motionless. This staging clarifies the crucial dramatic point that Cassandra (alone) is impervious to Clytemnestra's persuasion.Footnote ²⁸ Secondly, immediately after Clytemnestra's request that Cassandra use gesture (1061), the latter's supposed prophetic twitches – which have not occasioned comment before and so are presumably just starting – would hardly generate the response ‘She seems to need an interpreter’. Thirdly, if νεαιρέτου is right, then Clytemnestra picks up on it two lines later in a correction of the Coryphaeus' words: Cassandra is not like a θὴρ νεαίρετος, but like a horse which has been taken from a πόλις νεαίρετος but is not itself submissive. Clytemnestra's picture of Cassandra as a wild and proud horse is sharper repartee if the Coryphaeus' image, supported by Cassandra's actions, is one of a still and submissive animal (cf. Arrian) rather than one of a twitching and fearful one.

One might object that Clytemnestra's raving-horse comparison itself entails that Cassandra is raving. However, Clytemnestra's words are interesting and effective if they are a disparaging, visibly unfair stereotype of the wild madness of eastern seers, who listen not to authority but to voices or ideas in their heads (1064 κακῶν κλύει ϕρενῶν) and who foam at the mouth (cf. 1067 ἐξαϕρίζεσθαι). The words then signal the frustration which Clytemnestra states in 1068 just before storming out, and on which the Coryphaeus comments in 1069 (ἐγὼ … οὐ θυμώσομαι). Clytemnestra has lost control, not only over other people (she cannot ‘break in’ this ‘horse’) but also over her emotions. Her blatant ‘misreading’ of Cassandra as haughty and less-than-rational would demonstrate the same qualities in herself, and thus enhance the irony which is clear in the passage generally – namely that both Clytemnestra and the Coryphaeus, in assuming that Cassandra fails to comprehend and needs an interpreter, themselves fail to comprehend that it is her command of Greek which will need clarification and interpretation.Footnote ²⁹

In spite of this argument, one should not exclude that νεαιρέτου is after all wrong. Its recurrence in the same sedes two lines later, and the obscurity of what ‘like a newly-captured animal’ might mean, do at least raise suspicion. Doubtless there are many possible replacements, but ἀϕράσμονος would suit the context of incomprehension.

1525

Clytemnestra says that Agamemnon acted against ἐμὸν ἐκ τοῦδ' ἔρνος ἀερθέν. Editors feel forced to ascribe a very unusual sense to ἀερθέν. Fraenkel cites Nicander fr. 108 G–S as the one parallel he could find for the meaning ‘conceive’; Denniston and Page cite it for the meaning ‘rear’. (Either translation fits that passage: Helen conceived or reared Paris' child.) Herodas 9.13 ]κεῦσί σ' ἤειρα may or may not have meant ‘I reared you for your parents’: we do not have enough context to know.

However, the normal sense of αἴρω is apt: ‘my scion from him, raised aloft’. The fact that Iphigenia was lifted over the altar was singled out for special mention in 234 (she was manhandled ἀερδήν: echoed here); cf. Euripides, Electra 1022 ὑπερτείνασ', Iphigenia among the Taurians 27 μεταρσία ληϕθεῖσ', Lucretius 1.95 sublata uirum manibus. Artists also depicted Polyxena and Iphigenia being manhandled into the air: cf. LIMC Iphigenia 2 = A.N.R.W. Prag, The Oresteia: Iconographic and Narrative Tradition (Warminster, 1985), pl. 38 (Attic, c. 640), Polyxena 26 = Prag pl. 41b (Timiades Painter, c. 560), Polyxena 37 = Prag pl. 41a (Etruscan, c. 460), and many later representations from Italy.

1568–73

As the text stands, στέργειν must be prolative after ἐθέλω, and the tone of acquiescence implies that δύστλητα means ‘hard to bear’.

There are two problems. The first is the abrupt change of construction at ἰόντα, though there are occasional parallels.Footnote ³⁰ The second is how Clytemnestra (and observe that ἐγώ is emphatic) can present the circumstances as something which she finds hard to bear but to which she will acquiesce as a graceful concession. Even in the exodos, where her foreboding has sunk in further, her comment that the business involved κακά (1654–5) falls well short of remorse. In any case, Clytemnestra's acquiescence is no genuine bargaining chip against the daimôn, but her tone cannot be wholly cynical given that her preceding words acknowledged that the Chorus ‘hit on an oracle with truth’ in warning her about retribution. Her shift in attitude towards the conciliatory, crime-weary tone of the exodos has its origins here, and our knowledge that she has begun to fear for her future is a crucial foil for the portrayal of Aegisthus' short-sightedness.

If the daimôn were the subject of στέργειν and δύστλητα meant ‘dreadful to dare’ (cf. 408 ἄτλητα τλᾶσα), the sense would be corrected. His side of the bargain will be to leave a dreadful crime unpunished, and Clytemnestra will in return offer a large proportion of her possessions (as implied in 1575). This syntax can be achieved without simply assuming a harsh anacoluthon, by replacing ἐθέλω with a verb which can take a dative + infinitive as well as an accusative + infinitive; this also mitigates the change of construction at ἰόντα.

I suggest αἰνῶ ‘recommend’.Footnote ³¹ Aeschylus may have recalled his phrasing at Persae 642–3 δαίμονα … ἰόντ' αἰνέσατ' ἐκ δόμων, even if the sense of αἰνῶ there is ‘assent to’. ἐθέλω, on my reconstruction, would be an imprecise but not stupid gloss on αἰνῶ, which intruded into τ. This does not seem far-fetched given that Σ_B (perhaps Thomas Magister) gave the gloss θελήσατε in the passage of Persae just mentioned.