Published online by Cambridge University Press: 08 November 2013
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.
I thank Chris Pelling, Richard Rutherford and CQ's reader for helpful and encouraging advice, but above all David Raeburn: all these ideas stem from my conversations with him; some have now been sketched, with less detail, in our The Agamemnon of Aeschylus: A Commentary for Students (Oxford, 2011). The following works are referred to by author's name: J.D. Denniston and D.L. Page, Aeschylus: Agamemnon (Oxford, 1957), E. Fraenkel, Aeschylus: Agamemnon (Oxford, 1950), D.L. Page, Aeschyli tragoediae (Oxford, 1972), M.L. West, Aeschyli Tragoediae (Stuttgart, 19982). I have not incorporated literature published since the article's submission in July 2011. I use West's sigla.
1 In 15, τὸ μὴ + infinitive expresses ‘fear preventing sleep’ (cf. Kühner–Gerth II 217) not ‘fear of sleeping’ (Bollack, J., L'Agamemnon d'Eschyle: Le Texte et ses interprétations, vol. 1 [Lille, 1981], 20–1).Google Scholar The fear is left ominously vague.
2 This evidence also suggests a different course from Housman, who tentatively suggested πόνου for ὕπνου: ‘The Agamemnon of Aeschylus’, reprinted in Diggle, J. and Goodyear, F.R.D. (edd.), The Classical Papers of A. E. Housman: Volume I 1882–97 (Cambridge, 1972), 55–90Google Scholar, at 55 [orig. JPh 16 (1888): 244–90].
3 For this set by scribe A3, see Johnson, W.A., Bookrolls and Scribes in Oxyrhynchus (Toronto, 2004), 18–19Google Scholar, 61, and PSI 15 1472. For accents defining word division see e.g. Turner, E. and Parsons, P.J., Greek Manuscripts of the Ancient World (London, 1987 2), 11Google Scholar.
4 So too West in his apparatus, correcting Lobel's editio princeps.
5 For the proverb see also Σ Ar. Av. 721a–b: watchmen sing either to ‘palliate their wakefulness’ (παραμυθεῖσθαι … τὴν ἀγρυπνίαν, apt for Ag.'s watchman) or to ‘fend off sleep’ (ἀποπερισπῶντες τὸν ὕπνον: only as a variant in MS E).
6 Wilson, P. and Taplin, O., ‘The “aetiology” of tragedy in the Oresteia’, PCPhS 39 (1993), 169–80Google Scholar, at 171.
7 See Pind. Ol. 7.48 + Σ 71b, 86a and Philostr. Imag. 2.27.3 (Rhodians to Athena); Eur. fr. 912 θυσίαν ἄπυρον πάγκαρπείας to chthonic Zeus; Inscr.Cos ED 241.24 (4th c.) ἄπυρα of oil and wine for Athena, apparently placed in cups rather than poured out; Plut. De gen. 578b ἀπύρους ἱερουργίας performed secretively at the tomb of Dirce; Diog. Laert. 8.13 Delian altar ἄνευ πυρός, a few lines after a reference to ἄπυρα food.
8 ἀπύρων ἱερῶν ὀργάς is difficult to construe even if ἄπυρος could mean ‘failing to burn’. ‘Anger at non-burning sacrifices’ will not work, since the gods' anger causes a burnt sacrifice to fail, rather than being caused by it. ‘Anger as expressed in non-burning sacrifices’ strains the genitive. ‘Anger of non-burning sacrifices’ requires a rare personification.
9 For τράπεζαι see Gill, D., Greek Cult Tables (London, 1991).Google Scholar
10 See Jameson, M.H., Jordan, D.R. and Kotansky, R.D., A Lex Sacra from Selinous (Durham, NC, 1993), 30–1Google Scholar, 70–3. They also (p. 71) support ὑπολείβων in Ag. 69.
11 The other manuscripts (< τ) have mistakenly swapped εὖ and γάρ.
12 See West, M.L., Studies in Aeschylus (Stuttgart, 1990), 178–81CrossRefGoogle Scholar: his text – I think convincingly – produces the latter interpretation.
13 For γ' ἄρα see Lowe, J.C.B., ‘Γ' ἄρα, γ' ἆρα and τἄρα’, Glotta 51 (1973), 34–64.Google Scholar He does not mention our passage, but (p. 60) tentatively supports γ' ἄρα in those mentioned from Euripides and Aristophanes. CQ's reader helpfully pointed out that J.F. Martin suggested εὖ γ᾽ ἄρ᾽ εἴη in his Observationes criticae in Aeschyli Oresteiam (Posen, 1837)Google Scholar, 4, though he did not explain the difficulties of γάρ.
14 So e.g. Taplin, O., The Stagecraft of Aeschylus (Oxford, 1977), 304–6Google Scholar; Rehm, R., Greek Tragic Theatre (London, 1992), 84CrossRefGoogle Scholar; Mitchell-Boyask, R., ‘The marriage of Cassandra and the Oresteia: text, image, performance’, TAPhA 136 (2006), 269–97.Google Scholar
15 There is a stark contrast with (e.g.) the treatment of Andromache and Hermione in Eur. Andr.
16 The spoils were also brought up by the Herald at 578–9.
17 I have not been able to find a detailed list elsewhere.
18 Whether Electra gets her way, and when Clytemnestra actually descends, are left unclear by the text.
19 e.g. Clytemnestra claims to have killed Agamemnon only after Cassandra was introduced into the house; Electra reveals that Clytemnestra had been flirtatious even before Iphigenia's sacrifice; Clytemnestra expresses remorse.
20 For time ageing in Greek, see e.g. Judet de la Combe, P., L'Agamemnon d'Eschyle: Le Texte et ses interprétations, vol. 3 (Lille, 1982), 224–5.Google Scholar J.F. Gannon does not think the parallels are precise enough to credit χρόνος παρήβησεν ἐπεί … here: ‘Aeschylus, Agamemnon 984–6, 998’, CQ 40 (1990), 46–53CrossRefGoogle Scholar, at 48–9.
21 In 998, the fact that the subject ‘these fears’ has to be understood from the previous sentence has bothered some scholars; e.g. Denniston–Page and Gannon (n. 20), 50 prefer εὔχομαι <τά > δ' ἐξ …, but this cavalierly introduces an asyndeton.
22 See Denniston–Page, with the suggestion ΨΑΜΜΙΑCAICΑΤΑ, with ΑΙC > ΑΚ. Page observed that the hypothesis of interlinear –ΑΙC would suggest that a feminine dative plural noun preceded.
23 ‘Aeschylus, Agamemnon 984–6’, JHS 86 (1966), 166–7.CrossRefGoogle Scholar Gannon (n. 20), 47 too quickly assumes that ὑπό + accusative can simply mean ‘to’.
24 J. Lavery suggests that πρυμνήσια here and in Eur. IT 1395 are simply ‘sterns’, but the latter passage is itself badly corrupt: ‘Agamemnon 984f.: an “inhibition”?’, BICS 42 (1997–8), 57–61.Google Scholar
25 Lavery (n. 24) suggests ψάμμος ἐσκάϕη, and follows Denniston–Page in emending 998 (cf. n. 21).
26 The first three words are unclear. T begins μάλα γέ (with γάρ above) τοι δή and F μάλα γάρ τοι.
27 e.g. Ewans, M., Aeschylus: The Oresteia (London, 1995)Google Scholar, Collard, C., Aeschylus: Oresteia (Oxford, 2002)Google Scholar, Sommerstein, A.H., Aeschylus II: The Oresteia (Cambridge, MA, 2008)Google Scholar, ad loc.
28 cf. Fraenkel p. 485: ‘As long as Cassandra remains motionless, she is unassailable’.
29 See in particular 1087–92, where Cassandra translates the Chorus' description of the palace into a different register of Greek, and 1254 where she claims to know Greek ‘too well’ for the Chorus. Montiglio, S., Silence in the Land of Logos (Princeton, 2000)Google Scholar, 214 makes a similar point: ‘Clytemnestra wishes to believe that Cassandra is not grasping her words, but it is she who is not grasping Cassandra's silence.’
30 e.g. Solon fr. 34.2–3 West κἀδόκεον ἕκαστος αὐτῶν ὄλβον εὑρήσειν πολύν, | καί με κωτίλλοντα λείως τραχὺν ἐκϕανεῖν νόον.
31 Aeschylus uses αἰνῶ + dat. + inf. at Cho. 715. ὅρκους θεμένη, now with commas on either side, may imply a prior bargain. Alternatively, for the coincident aorist participle see Schwyzer–Debrunner 301, Barrett, W.S., Euripides. Hippolytos (Oxford, 1964)Google Scholar, 214. Neither cites instances with a present indicative; Goodwin, W.W., Syntax of the Moods and Tenses of the Greek Verb (London, 1878 6)Google Scholar, 52 cites Thuc. 8.50.2 πέμπει [historic present] … ἐπιστείλας; Ag. 97–8 has coincident λέξασ' with present imperative αἴνει.