² Numerous conjectures are reported in the Prinz-Wecklein edition (Bd. I 6 ed. N. Wecklein, 1898). More recent editions are: A.C.Pearson (Cambridge, 1903), N. Wecklein (Leipzig/ Berlin, 1907), G. Murray (OCT vol. iii, 1909, 1913), G. Italie (Groningen, 1949), A. Y. Campbell (Liverpool, 1950), H. Grégoire (Paris, 1950), K. Alt (Teubner edn, 1964), A. M. Dale (Oxford, 1967), R. Kannicht (Heidelberg, 1969). Other studies referred to by author's name only are: Wilamowitz-Moellendorff, U. von, Griechische Verskunst (Berlin, 1921), pp. 561–6Google Scholar; Zuntz, G., An Inquiry into the Transmission of the Plays of Euripides (Cambridge, 1965), pp. 211–48Google Scholar; Lloyd-Jones, H., review of Zuntz in CR 16 (1966), 158Google Scholar; Schmiel, R., ‘The Recognition Duo in Euripides Helen’, Hermes 100 (1972), 274–94Google Scholar; Young, D. C. C., ‘The Text of the Recognition Duet in Euripides' Helena’, GRBS 15 (1974), 39– 56Google Scholar; Diggle, J., ‘On the Helen of Euripides’ in Dionysiaca: Nine Studies…presented to Sir Denys Page… (Cambridge, 1978), 159–77.Google ScholarBiehl, W.explores the Duo's metrical patterns in Helikon 20– 21 (1980 1 [1983]), 257–92Google Scholar, but his metric is as unconvincing as his textual decisions.

³ For the points of contact, see Matthiessen, K., Elektra, Taurische Iphigenie und Helena (Göttingen, 1964), pp. 134–8CrossRefGoogle Scholar. For the late dating of Sophocles' Electra (after E. El. and not long before S. Phil.), cf. my commentary on Orestes (hereafter referred to as ‘ comm. Or. p.…’ or ‘comm. on Or …’), Introd. p. Ivi n. 91.

⁴ Cf. also Alc. 244–72 (Alcestis, Admetus), Andr. 825–65 (Hermione, Nurse), Tro. 235–92 (Hecuba, Talthybius), H.F. 1178–1201 (Amphitryon, Theseus), Ph. 103–92 (Antigone, Paedagogus). The list can be greatly extended with exx. where the interlocutor is the chorus-leader, the longest being the Phrygian's aria in Or. (1369–1502).

⁵ Two singing actors: Hec. 177–215 (Polyxena, Hecuba), El. 1177–1232 (a trio with chorus, on the model of A. Ch. 315–465). The exx. are too few for the formulation of rules; but there are a number of analogies if we include duets for actor and chorus (e.g. Or. 1246–86. which has speech and song for both parties).

⁶ Diggle…τηλύγετον «cúμενον» | χθον⋯c ⋯πο «πρ⋯» πατρίδοαc 'Aργάθεν, ὦ ϕίλο (2δ). But I include 830 within the obeli, suspecting that the terminal ϕίλοc, anticlimactic after ὦ ϕίλτατ…ϕίλτατοc, may conceal ϕΑΟΣ cf. Ion 1439, S. El. 1224, etc. (see below). τηλύγετον (‘latest-born’)πατρίδοc ϕάοc would be a plausible phrase; but it is hard to know what other words to add or subtract.

⁷ Against Diggle, I accent cέ here (‘et ego te’), as also in Hel. 630 and 658; and I put the comma before, not after, τήν θανοûcαντήν θανοûcαν.

⁸ Bauer; so Diggle. Mastronarde, D. J. (Contact and Discontinuity [Berkeley, 1978], p. 56Google Scholar) prefers the transfer of 832 only (Lohmann, Maas); but all his parallels have syntax split between two singers. Monodists never (so far as I am aware) leave their syntax to be completed by a speaking interlocutor, though their sentences may be variously broken into or supplemented. He seems to have overlooked Hel. 659 in objecting to ‘the anomaly of a lyric line in the male role in the duet’ (also a dochmiac dimeter, and similarly in an embrace).

⁹ Coordination with a relative pronoun is conjectural in 834, but rightly regarded as probable by Diggle, who proposes ȏν ἔτι (after Bergk τόν ἔτι).

¹⁰ The pattern is spoilt, pace Kannicht and Lloyd-Jones, by Kretschmar's transfer of 656 to Men. (in conjunction with Lachmann's wrong assignation of 654–5 to Helen). L's assignations are correct here, as argued further below.

¹¹ Diggle rightly defends the assignation of 692–3 to Helen (see below); but his statement ‘Menelaus does not sing lyrics in this duet’ needed more exact formulation (either ‘in 660–97’ or ‘enoplian verses’).

¹² Zuntz refers no less irrelevantly to Men's ‘deuteragonist’ role. Many tragic persons sing in one scene only, including ‘second actor’ male persons (e.g. Theseus in Hipp.), and the second actor does all the solo singing in Or. (Electra, the Phrygian). Conversely, Ion does not sing in the Ion duo, though he is a singing character elsewhere.

¹³ Leg. ὦ μέλεοc άμέρα? ίώ for exclamatory ώ is a frequent error (comm. Or. pp. 140, 244, 362); and (cretic-paeonic) is likelier than either (cf. T. C. W. Stinton, BICS 22 (1975), 88ff).

¹⁴ ὢc «c'» Brodaeus (Porson's ὡc είc έμάc «c'» is no better). Canter's ἥ c', preferred by Murray and Dale, may seem more natural. But ὡc is likely to be right. Kannicht cites Ba. 130; but for closer causal-exclamatory parallels cf. Or. 90 ὡ μέλεοc ή τεκοûcάθ', ώc άπώλετο and ibid. 130 θεοί cε μιιcήcειαν, ὢc μ' άπώλεcαc ‘…for having (thus) destroyed me’ (with comm.).

¹⁵ A conventional feature, cf. I.T. 842, S. El. 1281. Young neglected the parallels (‘she plays to the gallery of the chorus…’).

¹⁶ So Kannicht (also Young), with appeals to Dale's, Lyric Metres (2nd edn, pp. 115f.Google Scholar) and Conomis, N. C., ‘The Dochmiacs of Greek Drama’, Hermes 92 (1964), 23– 50, at 28–30Google Scholar. Dale herself had no doubt here (‘the syllables must all be short’). The time is ripe for a reassessment of the ‘dochmius kaibelianus’ in tragedy; cf. n. 19 below.

¹⁷ Cf. comm. on Or. 1502 ἒλαβε τ⋯ν Έλέναγάμον(and ibid. 1565; Collard on Su. 536); for the combination with ἂcμενοc cf. Or. 776 (with comm).

¹⁸ That hypothesis is consistent with L's erroneous verse-division here after χέρα ϕίλιον seen. 56 below.

¹⁹ See Barrell (who favours τά κρύπτ' έκ). For other suspect instances(Hec. 1083–1, H.F. 1070, Or. 330/345, 1247/1267), see comm. Or. pp. 138, 288 and CQ 38 (1988), 96.

²⁰ Cf.Tarrant, D., ‘Greek Metaphors of Light’, CQ 10 (1960), 181–7CrossRefGoogle Scholar.

²¹ P. Oxy. 2336, ed. Roberts, C. H. in The Oxyrhynchus Papyri 22 (1954), 107Google Scholar; see Zuntz, pp. 217ff. and p1. xvi.

²² Schmiel in general follows Zuntz, but goes further: his Menelaus is denied any ‘terms of endearment’ (276), and at the end of the Duo ‘there has been no reconciliation’ (280); cf. n. 65 below.

²³ As argued in CQ 38 (1988), 314Google Scholar, Med. 377 should be followed, after a pause, by a direct question (πότερον ύϕάψω…;).

²⁴ Cf. also the standard rhetorical hesitation ‘How shall I begin?’ (as old as Homer; Denniston on El. 907–8 cites Od. 9.14).

²⁵ It may be that corruption of K to c is commoner than the reverse; but that is a slight consideration. Lloyd-Jones censures Zuntz's ‘bias in favour of the papyrus’; but bias in favour of L would be more reprehensible.

²⁶ The negative view of ‘the famous husband’, as developed in Andr. 456ff., 590ff., 629–31, surely has an ancient heritage (II. 17.588 μαλθακòc αἰχμητήIlias Parva fr. 17, Ibycus 296 Page), alongside more heroic views of Menelaus' martial prowess; cf. comm. on Or. 682–716, 742, 754.

²⁷ For criticisms of Zuntz's suggestions, see Dale, Lloyd-Jones and Kannicht.

²⁸ Cf. also Ph. 21 ήδον⋯ι δού(codd.; 'νδούMarkland), PI. Phaedr. 250e ήδον⋯ι παραδού

²⁹ Cf. comm. Or. p. 105.

³⁰ Zuntz proposed: ΕΛ ώ π όι, ὧ Φιλτάτα πρόόοψιΜΕ. «τ ⋯ υόν, γύναι, πρ όθυμον» ούκ έμέμΦθην but neither justified the choriambic verse nor accounted for the lacuna. Dale humorously commented: ‘no unmanly tenderness there!’

³¹ A cretic, of course, not a dactyl: cf. Ion 1470 ⋯ τ έκνον, τίί ήιΦ;… There is nothing wrong with metrical pause at such a comma (a fortiori, colon); cf. Stinton, ‘Pause and Period…’, CQ 27 (1977), 27–66, at 27ff.. 36.

³² For this confusion, cf. Or. 1039 (with comm.). S.0.T. 1061; for the reverse corruption, cf. 652(?) below. Andr. 427, El. 870.

³³ Note also that λέκτρα ἒχειν + gen. feminae elsewhere simply means ‘to have as wife’ (Ph. 14, etc.), even as λέκτρα λαβείν(οϒ γαμείν)+gen. feminae simply means ‘to marry’ (‘obtain the hand-in-marriage of’).

³⁴ It is vain to appeal to Med. 140 τόν μ⋯ν γ⋯ρ ἔχει λέκτρα τυράννων and ibid. 594 γ⋯μαί με λέκτρα βαιλέων ἃ νȗν έχω (βαιλεων II, Elmsley; -έωcodd.); cf. Sen. Med. 56 regum thalamos. There too the gen is the usual gen. conjugis (n. 33 above). A princess is ‘royal persons’ according to a standard Greek generalizing idiom, used when a single person is regarded as the representative of a type (Barrett on Hipp. 49).

³⁵ That is the most that Zuntz (followed by Young) establishes in his defence of the text. For the interpretation of Agathyllos ap. Dion. Hal. Antiq. Rom. 1.49.2, see also Lloyd-Jones and Kannicht. One wonders whether Agathyllos was influenced by corrupt texts of Euripides: the false reading βαιλέωMed. 594 (n. 34 above) could well be of sufficient antiquity.

³⁶ No known ex. in Euripides, though cf. A. Ag. 223/233. As L. P. E. Parker observes in CQ 26 (1976), 21, ‘most of the examples of mid-verse bacchiac are in Aeschylus’.

³⁷ On ‘catalexis’, ‘pause’ and ‘period-end’, see the cited articles by Stinton (n. 31), esp. 39f., and Parker (last n.). ‘Elision at period-end’ is a contradiction in terms.

³⁸ I am indebted to Dr Diggle for the following list of catalectic iambic trimeters in Euripides and Sophocles: Ale. 223/234, 272, Held. 892/901, Andr. 1032/1043, Hec. 634, 642, 656, Ion 1459, 1463, 1464, 1492, 1493, Tro. 1088/1106, 1290, 1292/1299, 1303/1318, 1316/1332, El. 1206/1215, Hel. 632, 633, 636, Ph. 1713/1738, Phaethon 86/94; S. Tra. 954/963, Ant. 592/603, O.T. 192/205, 202/215, 865/875, 889/903, 891/905, El. 163/184, 1276, 1277, O.C. 541/548, 1672/1699. E. Tro. 1296 has no caesura, but Diggle rightly obelizes the whole of 1295–7 (also 1289).

³⁹ Campbell's ἔ χω τά τ⋯Δι⋯ά λέκτρα γήμαwas somewhat contorted; but for the characteristically Euripidean idiom λ έκτρα (λέχο) тІνόγαμείν, cf. Med. 594 (n. 34), I.T. 538, Or. 20 1, etc.

⁴⁰ Bacchei tend to come in multiples of two. For runs of four, cf. comm. on Or. 1294–5.

⁴¹ Only Young in recent decades has attempted to defend L's attributions here (and numerous other transmitted anomalies in the Duo, see below). I make no apology for passing over in silence most of his arguments and metrical interpretations (cf. Diggle, p. 175 n. 14).

⁴² Zuntz seems not to have considered the possibility that II had something other than av at the beginning of the sentence. His very different inference, namely that the clauses αν…ωλβιαν and ωλβιαν εμε εε ετ ματαν must have been split in II (or II's archetype) between different singers, produced an intolerable arrangement as to style (see Dale, pp. 172–3), further refuted by the consideration that ωλβιαν εμε ε тε ματαν would naturally have been written as a separate verse (ΙΚ). following the hypothetical speaker-change.

⁴³ For this double ace. construction with a neuter pronoun, cf. S. Aj. 552 καίτοι ε καί νûνν τοûτό γε ζηλοûν ἒχω, Ar. vesp. 588 τουτί γάρ τοί ε μόνον (Reiske, for εμνόν) τούτων ὦν εἴρηκαμακαρίζω.

⁴⁴ Cf. S. ΕΙ. 164 ΗΑ. ὂν γ' έγώ… (Hermann, for ὅν ἕγωγ').

⁴⁵ Dale posed the question ‘What kind of line is (ωλβιαν ωλβιαν εμε ε τε ματαν) meant for?’ The best that can be made of it is δία with a horrid split resolution and overlap at the join. Split resolution of the third longum of a δ occurs only before another δ and is virtually limited to the pattern described on p. 59 below, apart from Tro. 253 παρθένον άι γέραόχρυοκόμα…

⁴⁶ Zuntz proposed έμέ «τε» έ τε μάταν. Kannicht's suggestion έμέ «δε» έ τε μάταν. begins a new sentence, but μάταν(if authentic) must go with ὢλβιαν. At one time I considered έμέ έ τ' «οÙ» μάταν.

⁴⁷ For the false division now after a Ik, cf. 670f. (and n. 56 below).

⁴⁸ Hdt. 1.31 ⋯μακάριζον τ⋯ν τήν ṕώμην cf. Tro. 229 εὔανδρον όλβίζων γ⋯ν.

⁴⁹ Analysed in comm. Or. p. 321 as T² ()+ba

⁵⁰ Menelaus reacts similarly with interrogative iteration to a surprising remark at 675 (ME 'ΗΡα;…); cf. Diggle, Studies on the Text of Euripides (1981), pp. 50f. (but the emotion is often surprise, rather than indignation or incredulity). Ba. 1177 (ΑΤ. ΚιθαιΡών … ΧΟ. [τι] Κιθαιρών;)is an instance of bacchiac iteration. Note that [προαλλα]ν (bis) will not do, since τ⋯δε then has no referent.

⁵¹ The space before N in II's 641a is a little longer than the space before N in 641 (nearly the width of the letter N); but the same words written twice are not always exactly the same length. Or did II perhaps have [τι το προθε]ν?

⁵² The lineation of 641–2a is rational if Aristophanes read τύχαν, with a plausible ‘dochmiac compound’ (comm. Or. p. 106) framed between 2ba verses. Zuntz overlooked that, if Aristophanes had read υμϕοράν he could scarcely have failed to divide after έλαύνει for the run of bacchei.

⁵³ τ⋯ κακ⋯ν άγαθόν is a juxtaposition of opposites superficially similar to τό καλόν ού καλόν at Or. 819; note that in both passages it is incorrect to supply a copula. But Or. 819ff. is otherwise very different (see comm.).

⁵⁴ I agree with Kannicht in preferring δέ. against Lloyd-Jones (cf. n. 25 above). The position is similar to that in 642 (γ' II, δ' L; see above).

⁵⁵ A surprising oversight, since it was Zuntz who taught us to take account of such colometric indications. It is credible that the Editor scanned 661a and 662a as ia -x - (another iambicizing misinterpretation), following an indeterminate ‘extra-metric’ exclamation (cf. 166 7 γδπένθειν:ἕἔ: πτεροϕόροι).

⁵⁶ L here attests a doubly divergent lineation: πο-έμ-ἓχ-ἔχ⋯ν ἔμ-ἔμ-έκ Τρ-πολ-μολ-. The verse-end after Τροίαis likely to be ancient, though probably not Aristophanic. The other division at ἓχομενὅνis not an isolated scribal aberration, pace Zuntz (p. 229), but of a piece with L's divisions in 628–9 (ϕίλιον/έν μακράι), 634–5 (ἔβαλον\ήδονάν), 694–5(κακόποτμον\άραίαν), 696–7(ἔλιπον\ ού λιπού'). Such consistently wrong treatment of dochmiacs, typically creating a false ia–tr dimeter or glyconic, must go back to erroneous colometry in antiquity. II may have shared some of these lineation errors (cf. 638f., ?661f, 670f.), but certainly not all (cf. 634–5). Some misinterpretation and corruption of dochmiacs probably goes back to Aristophanes (and earlier still); but we must also recognize that the Editor's division of cola was by no means uniformly transmitted in the sub-Aristophanic tradition. Zuntz's reference to the division of cola attested in L as ‘due to the Alexandrian editor’ (p. 212) needed some qualification.

⁵⁷ None of Zuntz's parallels (p. 229 n. §) contains such a sequential trio. More pertinent is Or. 149 κύταγε κύταγε, πρόιθ' άτρέμαάτρέμαϊθι but the χ⋯μα Εύριπίδειον obviously makes a difference there, also the aabccb pattern.

⁵⁸ As to ordinary adjectives, cf. comm. Or. p. 252. Doubled έμόoccurs at Hec. 710 ⋯μό⋯μόξένοΘρήικιο…, H.F. 1190 έμ⋯έμόöδε γόνοό πολύπονο…and EI.1149 ἔπεεν έμόέμόύρχέτμin each case preceding the noun and as the only anadiplosis; so also, perhaps, Ph. 153 öέπ' έμάν «έμάν» πόλιν ἔβα πέρων(Diggle).

⁵⁹ Cf. Conomis, art. cit. 45, and L. P. E. Parker, ‘Split Resolution…’, CQ 18 (1968), 241–69, at 267 8. Ph. 1295 (with άχήω Elmsley) and I.A. 1285 (with ἔβαχε«ν») may be further exx.

⁶⁰ Poetic personification of τύχη need not, of course, imply deification; see, for example. Barrett on Hipp. 818–21. But editors of Euripides could afford to give Τύχη a capital letter more often. See in general G. Busch, Untersuchungen Zum Wesen der Τύχη in den Tragödien de.s Euripides (Diss. Heidelberg. 1937).

⁶¹ Dale observed that 656 ‘follows badly on the previous line’. She failed to make it clear that this is true only when the speaker-assignations are altered.

⁶² δοκοȗαν may include the imperfect sense ‘(previously) believed (by me)’; but IT. 831 shows that Menelaus might equally have said μολούαν ώδοξάζεται.

⁶³ For πρόθεών followed by a question and introducing a new λόγοcf. Or. 92 and 579.

⁶⁴ Cf. Andr. 857, 862, Ion 1480(?), 1494, Hyps. fr. 64.94; H.F. 1190 (…üδρα) and Ion 1487 (…κύκλωι) have anceps penult.; for the related , cf. El. 586, 588, 590. It is a matter of opinion whether these verses are best regarded as compounded of an+~-*- or as prolongations of (both may be correct: cf. the ‘sub-dochmiac’ 2ia verse comm. Or. p. 112; ‘prolongation’, ibid. p. 288). But we can certainly reject the frequent analysis an + irregular δ (), since ‘only “iambic-types” of metres are combined internally with dochmiacs’ (Conomis, art. cit. 48).

⁶⁵ It has evidently misled Schmiel, who supports his argument for a cold, ‘unreconciled’ Menelaus (n. 22 above) by the very fact that he ‘interrogates’ his wife. Menelaus has warmly embraced and been embraced by Helen for thirty-five lines before seeking to satisfy his curiosity: it is then Helen who makes the running and has the last word.

⁶⁶ Period-end at οỉονοỉον is out of the question, and οỉον έοίομυι cannot therefore be a dochmius, whose anceps first syllable (pace Dale) can only follow a longum or biceps. The telesillean occurs exceptionally in the rising part of enoplian dicola (Hipp. 1269, S. O.T. 1096/1108), but not in the close.

⁶⁷ For these clausulae, variously following ρ χ , Τχ or Αχ see comm. Or. pp. xxi, 113, 288.

⁶⁸ Classified by Mastronarde (Contact and Discontinuity, pp.56ff.)as ‘suspended syntax with intervention encouraging completion (lyric and imabic)’.

⁶⁹ As to the latter, it is the person, not the emotion, that πέτεται in the passages cited by Kannicht as parallels.

⁷⁰ ‘Either’ may also be deferred: cf. Med. 846ff. πώούν ίερ⋯ν ποταμών\ἥ πόλιἥ ϕίλων\πόμπιμοε χώρα…; (where Elmsley compared Ar. Av. 420).

⁷¹ On supposed dochmiacs with two shorts for initial anceps, see Barrett, Hippolvtos, p. 434, and Diggle, III. Cl. Stud. 2 (1977), 123, and Studies 54.

⁷² Both LSJ and Allen-Italie cite these passages under έπελαύνω; and the silence of Dale and Kannicht seems to imply acquiescence. Dr Diggle drew my attention to the truth.

⁷³ Campbell gilded the lily (more suo) by writing ε τ⋯νδ' ἔθηκε ϕάρμακον for οι ϒώνδ' ἔθηχ' Ηρα κακών without thinking it necessary to offer a parallel for the sense ‘scapegoat’ (see LSJ). Perhaps that is why Diggle's survey did not include the plausible conjecture ἔκκριτον.

⁷⁴ Young is too conservative, as elsewhere: ‘That she [Hera] might deprive of Kypris [sexual pleasure]…Paris, to whom she [Kypris as goddess of sex] had assigned me…'. No one could be expected to understand that.

⁷⁵ Metre guarantees ὦ τλήμον at Hec. 775 and ὦ τλάμων ύμεναίων at Hipp. 554 (both obviously pitying). In exx. of exclam. nom. (cf. comm. on Or. 90, 160, 1527, 1537–8, and Stevens on Andr. 71) it is normally easy to supply the appropriate name or pronoun in the nominative case.

⁷⁶ In all senses τλήμων and τáλαc are characteristically applied in tragedy to huma n beings, whose ‘audacious’ conduct ma y be due to divine affliction (cf. τλήμων Ορ⋯cτηc); cf. also ταλαίπωροc, μ⋯λεοc, δ⋯cτηνοc. Dale needed better support than Alc. 1, where Apollo uses the verb ἔτλην for his ungodlike ‘submission’ to servile status. Dr Diggle draws my attention to Ion 905, where (in his text) Creusa reproaches Apollo as τλᾱμον; that is doubtless right (after the precedent of Med. 990 cύ δ', ὧ τáλαν, ὧ κακáνυμϕε), but it is likely that the use of τλȃμον addressed to Apollo was as bold a novelty as Orestes' rhetorical hyperbole ⋯κείνον ⋯γείcθ ⋯νócʟον καί κτεʟ'νετε at Or. 595.

⁷⁷ So Diggle, , citing Parker, CQ 16 (1966), 12Google Scholar .

⁷⁸ Cf. Breitenbach, W., Unlersuchungen zur Sprache der euripideischen Lyrik (Stuttgart, 1934), pp. 221–5Google Scholar . Paregmenon of ‘pitying’ words is especially frequent.

⁷⁹ The iambelegus (cf. Or. 1264/1284) makes its appearance more often than not in Euripides’ ‘enoplian dochmiac’ scenes; and for the verse , see comm. Or. p. 303.

⁸⁰ For the pejorative-negative use of ⋯- and ⋯πο- words, see in general Fehling, D., Hermes 96 (1968), 142ff., and commGoogle Scholar. on Or. 162–5.

⁸¹ Cf. H.F. 883–5, 1082–6, 1205–13. Tro. 886–92, IT. 895–9, Ion 1494–6. Or. 181–6/202–7, 1256–65/1276–85, 1363–5/1546–8. There is no excuse for attempts to make dochmiacs out of 692–3 an unexceptionable dicolon as transmitted, like Med. 992–3/999–1000.

⁸² Euripides exploited the imprecision of δʟώλετο and δɩολ⋯cανταc rather differently at Or. 1512 and 1566; see comm.. also CQ 1988. 96 on H.F. 1021–4.

⁸³ L has … ⋯π⋯ κακóπυτμον⋯ραίαν ἒβαλε θεóc…; see n. 56 above.

⁸⁴ Kannicht rightly rejects the vulgate ὂτε (Dobree). ‘When I left home…’ feebly ends the Duo with a less relevant temporal point, and (more seriously) loses the echo of the traditional accusation in μ⋯λαθρα… ἒλιπον.

⁸⁵ Pearson was wrong in principle (‘“I left and did not leave‘…”, i.e. I seemed to do so’), though followed by Dale (‘when I left-without-leaving…’) and cited approvingly by Kannicht. 611 does not mean ‘thinking that Paris seems-to-have Helen’. In most of these positive-negative quibbles (there are several species) there is an implicit λóγωɩ/ἒργωɩ antithesis, e.g. (λóγωɩ μ⋯ν μήτηρ (ἒργωι δ⋯) ⋯μήτωρ. In 138 τεθνȃcι κοὐ τεθνȃcι (obviously not ‘they seem to be dead’) there are explicitly ‘two λóγοι’.

⁸⁶ Cf. H.F. 1072–8, where ‘the double-edged use of (οὐ) ϕεύγω complicates the thought, but the point is clear enough’ (CQ 1988, 97).