Herrero ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 2017 Miguel Herrero de Jáuregui Line Transposition in Euripides Alcestis 1094–1096 Miguel Herrero de Jáuregui DITORS AND COMMENTATORS have long been distressed by two lines in the stichomythic dialogue between Admetus and Heracles in the final scene of the Alcestis (1094–1095). Wilamowitz suggested deleting them and most editors put them between square brackets.1 The text offered by the manuscripts is as follows (textual variants in line 1094 will be discussed below): {Ηρ.} γυνή σε παύσει καὶ νέοι γάµοι πόθου. {Αδ.} σίγησον· οἷον εἶπας. οὐκ ἂν ᾠόµην. {Ηρ.} τί δ’; οὐ γαµεῖς γὰρ ἀλλὰ χηρεύσῃ λέχος; {Αδ.} οὐκ ἔστιν ἥτις τῷδε συγκλιθήσεται. 1090 {Ηρ.} µῶν τὴν θανοῦσαν ὠφελεῖν τι προσδοκᾷς; {Αδ.} κείνην ὅπουπερ ἔστι τιµᾶσθαι χρεών. {Ηρ.} αἰνῶ µὲν αἰνῶ· µωρίαν δ’ ὀφλισκάνεις. [{Αδ.} ὡς µήποτ’ ἄνδρα τόνδε νυµφίον καλῶν. {Ηρ.} ἐπῄνεσ’ ἀλόχῳ πιστὸς οὕνεκ’ εἶ φίλος.] 1095 {Αδ.} θάνοιµ’ ἐκείνην καίπερ οὐκ οὖσαν προδούς. {Ηρ.} δέχου νυν εἴσω τήνδε γενναίων δόµων. {Αδ.} µή, πρός σε τοῦ σπείραντος ἄντοµαι Διός. {Ηρ.} καὶ µὴν ἁµαρτήσῃ γε µὴ δράσας τάδε. 1 G. Murray, Euripidis Fabulae I (Oxford 1902), reports Wilamowitz’s suggestion (per litteras) in his apparatus, although in his own translation of the play he actually deleted 1093–1094: cf. L. P. E. Parker, Euripides. Alcestis (Oxford 2007) 269. A. M. Dale, Euripides. Alcestis (Oxford 1954), J. Diggle, Euripidis Fabulae I (Oxford 1984), D. Kovacs, Euripides (Cambridge [Mass.] 1994), and Parker put 1094–1095 between square brackets. Only L. Méridier, Euripide I (Paris 1926), and G. A. Seeck, Euripides Alkestis (Berlin 2008), keep the manuscript text unaltered. For other proposed emendations of these lines see nn.6 and 8 below. E 590 LINE TRANSPOSITION IN ALCESTIS 1094–1096 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 H.: A woman and a new union will put an end to your longing. A.: Hush! What a shocking thing you have said! I should never have thought it of you. H.: What? Will you never marry but keep a widower’s bed? A.: No woman shall ever lie beside me. 1090 H.: Do you suppose you are doing your dead wife any good that way? A.: Wherever she is, she must be held in honor. H.: I commend you, truly. But you deserve the name of fool. [A.: You will never call this man a bridegroom. H.: I commend you for being faithful to your wife.] 1095 A.: May I die if ever I betray her, even though she is gone! H.: Take this woman, then, into your generous house. A.: I beg you by Zeus who begot you, do not ask this! H.: And yet you will be making a mistake if you do not. (transl. Kovacs) The main problem with line 1094 is that the only way to understand the ὡς sentence is as depending on an implicit ἴσθι (a usual construction, suggested by a manuscript variant). However, on such an interpretation the ὡς clause should be constructed not with a participle but with a finite verb.2 Furthermore, apart from grammatical faultiness, it is also stylistically awkward, since Admetus’ assertion refers directly to Heracles’ αἰνῶ µὲν αἰνῶ in 1093, obviating the second part of the line (µωρίαν δ’ ὀφλισκάνεις). Besides, both 1094 and 1095 have been often criticized as adding nothing to the progression of the dialogue.3 Only a few scholars focusing on the theme of philia in the play have defended 1095, arguing that far from 2 Parker, Alcestis 269: e.g., in the same play, Alc. 800–802; also Soph. Aj. 39, Phil. 117. There are instances with ἴσθι (or similar verbs) + participle negated with µή (e.g. Eur. Heraclid. 983; Soph. Ant. 1063, Phil. 253, 415, OC 1154–1155), but in all these cases the imperative verb of knowledge is ex- plicit, never omitted. Some early editors attempted to emend the negative participial clause to introduce the more usual οὐ (cf. n.6 below), but ὡς followed by participle may also be constructed with negative µή (cf. n.10). 3 Cf. Dale, Alcestis 127: “to say nothing of the grammatical tour de force of passing over µωρίαν δ’ ὀφλισκάνεις, this is all rather feeble and pointless.” Parker, Alcestis 169: “these two lines are a very flat repetition of what both have already said.” MIGUEL HERRERO DE JÁUREGUI 591 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 being redundant, it emphasizes Admetus’ loyalty and love (not so evidently synonymous as moderns often think).4 However, suppressing these two lines undoubtedly makes 1096 a neat and meaningful answer to 1093.5 As an alternative to deletion, suggestions have been made to emend the participle in 1094 to καλεῖς, or to accept the sec- ondary variant καλεῖν, which surely comes from the syntactic difficulty of the sentence opening with ὡς, but does not solve it. Following a manuscript in which ἴσθι is written above ὡς as an explanation, some conjectures have inserted it into the sen- tence. Unsurprisingly, none of these suggestions have been suc- cessful among most editors.6 A third possibility is altering the order of the lines, supposing an instance of line transposition, “an extremely frequent error 4 S. Schein, “Φιλία in Euripides’ Alcestis,” Mètis 3 (1988) 201 n.69; G. R. Stanton, “Φιλία and ξενία in Euripides’ Alkestis,” Hermes 118 (1990) 44 n.8. 5 Dale, Alcestis 127, “1096 is more effective as a rejoinder to 1093 than to 1095”; Parker, Alcestis 269, “while 1096 does not follow well on 1095, µω- ρίαν δ’ ὀφλισκάνεις in 1093 invites Admetus to restate his position.” 6 H. W. Hayley, The Alcestis of Euripides (Boston 1898) 59, selects some conjectures to 1094: “ἴσθ᾽ οὔποτ᾽ — καλῶν Wakefield, ὡς οὔποτ᾽ — καλεῖς Herwerden, οὐ µήποτ᾽ — καλεῖς Kviçala, ὡς µήποτ᾽ ἴσθι τόνδε νυµφίον καλῶν Weidner (rejecting ἄνδρα). Kirchhoff conjectures that two verses have fallen out before this line.” There are other possible solutions: my col- league Felipe Hernández Muñoz calls my attention to the fact that ὡς µήποτε is an uncommon construction particularly used by Euripides (six times but none in the other tragedians), while ὡς µήποτε + participle seems to be a syntactical hapax in classical times, and to the parallel (not only syntactical, but also situational and even phonetically similar) of Phaidra’s utterance in Eur. Hipp. 420 ὡς µήποτ’ ἄνδρα τὸν ἐµὸν αἰσχύνασ’ ἁλῶ (“so that I will not be caught shaming my husband”): he therefore conjectures the subjunctive καλῶ instead of the participle καλῶν, which would clear away the grammatical difficulty with an easy correction of a common aural error. However, in my opinion the verb καλῶ (or ἁλῶ) in the first person would fit strangely with Admetus referring to himself as ἄνδρα τόνδε as object of his own action (compare the same expression at Alc. 1084). 592 LINE TRANSPOSITION IN ALCESTIS 1094–1096 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 in poetry.”7 A. Garzya recovered a forgotten proposal made in 1868 by K. Schenkl in a footnote in a review:8 “die Verse 1093–5 also angeordnet werden müssen: 1095, 1094, 1093, indem nur so ein richtiger Zusammenhang und eine entspre- chende Construction hergestellt wird.” This means transposing Heracles’ two utterances (I keep original line numbers for the sake of clarity). However, even if the ὡς sentence of 1094 be- comes grammatically acceptable (see next paragraph), saying ἐπῄνεσ’ before αἰνῶ µὲν αἰνῶ seems counter to the logic of the speaker, and the progression in Heracles’ thought is absurdly zigzagging: after his surprised question in 1091 he would be praising Admetus’s faithfulness (1095) and then he would recall that nevertheless he will be considered mad (1093). Schenkl was on the right track, but it is Admetus’ lines, not Heracles’, which must be transposed. Transposing lines 1094 and 1096 solves all linguistic and literary issues. This is the re- sulting dialogue: {Αδ.} κείνην ὅπουπερ ἔστι τιµᾶσθαι χρεών. {Ηρ.} αἰνῶ µὲν αἰνῶ· µωρίαν δ’ ὀφλισκάνεις. {Αδ.} θάνοιµ’ ἐκείνην καίπερ οὐκ οὖσαν προδούς. 1096 {Ηρ.} ἐπῄνεσ’ ἀλόχῳ πιστὸς οὕνεκ’ εἶ φίλος. 1095 {Αδ.} ὡς µήποτ’ ἄνδρα τόνδε νυµφίον καλῶν. 1094 {Ηρ.} δέχου νυν εἴσω τήνδε γενναίων δόµων. A.: Wherever she is, she must be held in honor. H.: I commend you, truly. But you deserve the name of fool. A.: May I die if ever I betray her, even though she is gone! 1096 H.: I commend you for being faithful and loving to your wife. 1095 A.: Because you will never call this man a bridegroom! 1094 H.: Take this woman, then, into your generous house. (my translation, adapting Kovacs’) 7 L. D. Reynolds and N. G. Wilson, Scribes and Scholars (Oxford 1968) 159. Transposition, however, must be justified, see n.20 below. 8 K. Schenkl, “Literarische Anzeigen: Euripidis fab. et. trag. v. Kirchhoff, Pflugk u. Klotz,” ZÖstG 19 (1868) 348 n.8; A. Garzya, Euripides. Alcestis (Leipzig 1980), follows Schenkl’s proposal, but emends 1094 with the sec- ondary variant καλεῖν. MIGUEL HERRERO DE JÁUREGUI 593 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 Thus ὡς … καλῶν is much better understood. There is no need to postulate an implicit verb ἴσθι (or any other implicit imperative like ἔπαινε), which would present the afore- mentioned grammatical difficulty of having a participle instead of a finite verb. Admetus simply continues the sentence of Heracles in the previous line.9 He does so with a participle that conveys the cause (from Admetus’ subjective viewpoint, under- lined by ὡς) of Heracles’ epainein.10 The negation with µή instead of the more usual οὐ in a (future) participial clause starting with ὡς has some celebrated parallels in tragedy;11 and the µήποτ’ of 1094 can be easily justified with any of the much- debated explanations that these syntactical constructions have received (µή has been said to underline the subjectivity of the character, to introduce a generalising statement, and to be 9 E.g. Eur. Ion 312–313 quoted below. L. Schuren, Shared Storytelling in Euripidean Stichomythia (Leiden 2015) 39, analyses some cases of “completion of one interlocutor’s utterance by the other” (e.g. Eur. Or. 775–776, {Ὀρ.} εἰ λέγοιµ’ ἀστοῖσιν ἐλθών … {Πυ.} ὡς ἔδρασας ἔνδικα, although in this case ὡς has a finite verb). 10 The causal sense of µή + participle seems derived from the more common conditional sense of this sort of construction (with or without the reinforcing ὡς): cf. H. S. Smyth, Greek Grammar (Cambridge [Mass.] 1956) 618, 621, and the bibliography quoted in nn.12–13 below. From “in so far as you are not calling this man a groom,” the slight step to “because you are not calling” is easy. 11 The closest is Soph. Phil. 935, ἀλλ’ ὡς µεθήσων µήποθ’, ὧδ’ ὁρᾷ πάλιν (“he looks away as one determined never to give it up”): cf. J. M. Stahl, Kritisch-historische Syntax des Griechischen Verbums der klassischen Zeit (Heidelberg 1907) 777 (“vereinzelt wegen der voluntativen Bedeutung des Futurums”); also Ion 312–313, {Κρ.} ἡµεῖς σ’ ἄρ’ αὖθις, ὦ ξέν’, ἀντοικτίροµεν. {Ἴων} ὡς µὴ εἰδόθ’ ἥτις µ’ ἔτεκεν ἐξ ὅτου τ’ ἔφυν (CR: “so in my turn, stranger, I pity you.” ION: “Because I do not know who gave me birth and from whom I descend”); Eur. Heraclid. 693, ὡς µὴ µενοῦντα τἄλλα σοι λέγειν πάρα (“talk on, if you like, I will not stay to hear the rest”). For this last case see the comments of J. Wilkins, Euripides. Heraclidae (Oxford 1993) 139; and J. Diggle, “Further Notes on the Heraclidae of Euripides,” PCPS 28 (1982) 60– 61 (repr. Euripidea [Oxford 1994] 225; his conjecture µενοῦντι would not affect the grammatical parallel). 594 LINE TRANSPOSITION IN ALCESTIS 1094–1096 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 influenced by an implicit imperative).12 Most clearly, the speakers’ intentional subjectivity emphasized by µή is patent in Admetus, who projects his will into Heracles’ future (negative) action. Besides, this sentence can also be taken as yet another instance of the tendency of the Greek language to expand negations with µή in substitution for those with οὐ.13 Besides the grammatical advantage, this minimal alteration of the text gives these two lines a more dense literary meaning that assures their authenticity: θάνοιµ’ answers directly to µωρίαν δ’ ὀφλισκάνεις, which was part of the solution both in Wilamowitz’s deletion and Schenkl’s transposition. After this climactic demonstration of Admetus’ self-assertion that leaves no doubt about his inner change after Alcestis’ death (now he is really able to sacrifice his life), the stichomythia runs smoothly, with each line echoing the previous one and paving the way of the following ones. Heracles starts by being persuasive: “I praise you because you are faithful and loving to your wife.” With this sentence he reassures Admetus that he does not con- sider him mad, but a good husband, thus implying that his upcoming petition that Admetus host the woman is in fact (as 12 The participles negated with µή instead of οὐ have been much dis- cussed, often singling out use with ὡς as a particular construction: Kühner- Gerth, Ausführliche Grammatik II 200–203; A. C. Moorhouse, “On Negativing Greek Participles, Where the Leading Verbs are of a Type to Require µή,” CQ 42 (1948) 35–40 (38 n.1 for constructions with ὡς); A. F. Braunlich, “Euripides Medea 239 and 815: µή with the Causal Participle,” AJP 77 (1956) 415–418; W. K. Pritchett, “Μή with the Participle,” AJP 79 (1958) 392–404; P. W. Wallace, “Μή with the Participle in Longus and Achilles Tatius,” AJP 89 (1968) 321–333; A. C. Moorhouse, The Syntax of Sophocles (Leiden 1982) 330–333. Moorhouse holds generic characterization to be the main explanation, while Pritchett (396, the only one to mention Alcestis 1094) thinks it is due to an implicit imperative. 13 Pritchett, AJP 79 (1958) 392–404; Wallace, AJP AJP 89 (1968) 321– 333; Moorhouse, Syntax of Sophocles 331, whose explanation of these cases is “this would be an early part of the general extension of the use of µή with participles which is seen in later Greek. It is notable that generic (charac- terising) sense appears in a number of the Classical examples, and this may be held responsible for the growth of the causal use.” MIGUEL HERRERO DE JÁUREGUI 595 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 the audience knows) respectful of his promise.14 Admetus is encouraged by this praise to recall the literal terms of his promise in 330–331 (κοὔτις ἀντὶ σοῦ ποτε / τόνδ᾽ ἄνδρα νύµφη Θεσσαλὶς προσφθέγξεται).15 But for the audience now the setting is ready for the apparently shocking order that pre- cipitates the happy ending of the play: “now receive this one inside your noble house” (1097). As a result of the transposition, there is a remarkable sym- metry of line 1097 with the preceding one (1094): νυν echoes µήποτ’, opposing Admetus’ absoluteness with Heracles’ matter- of-factness; τήνδε echoes τόνδε, with the ironical juxtaposition of both spouses;16 and the wedding theme suggested by νυµφίον is echoed by δέχου εἴσω δόµων, “receive into your house,” which is precisely what a groom does with the bride.17 In this way, Euripides starts building the “renewed wedding” of Admetus with his wife, to which he will allude repeatedly in the 14 Seeck, Alkestis 199, rightly defends 1095 as more than a doublet: “Herakles will daraufhin darstellen, dass er nicht die Treu als solche (scil. Torheit) tadle.” Moreover, the distinction between philos and pistos full makes sense in the context of the play, cf. n.4 above. 15 For Dale, Alcestis 127, these lines would be the cause of the interpola- tion of 1094. Yet its full meaning is that Admetus recalls his earlier promise, precisely adopting a more internal tone (from “calling” to “being called”), to underline his own sacrifice. The play is full of internal references that echo in these last dialogues those in the first part of the play: cf. M. Herrero de Jáuregui, “The Meanings of σώιζειν in Alcestis’ Final Scene,” Trends in Clas- sics 8 (2016) 205–225. 16 Schuren, Shared Storytelling 182–183, shows some instances in which ὅδε is anaphoric, referring to what has been said before, instead of pointing to a present person: in this case τήνδε unites ironically both kinds of deixis: it is in both senses Admetus’ ἄλοχος. 17 Schuren, Shared Storytelling 63–66, analyses the ambivalence of γυνή as “woman” and “wife” precisely with the example of the last scene of Alcestis. Although her conclusion that the woman is not the real Alcestis is uncon- vincing, she shows how the dialogue plays with deictic and non-deictic meanings. 596 LINE TRANSPOSITION IN ALCESTIS 1094–1096 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 following lines.18 Tragic irony is patent: line 1097 is taken by Admetus as a blunt rejection of his previous noble position, but Heracles and the audience know that this order is precisely a consequence of the faithfulness to his matrimonial promise that Heracles and the audience have just witnessed. The transition in Heracles’ utterances is entirely logical for the audience (though not for Admetus), so no lacuna needs to be posited before 1097.19 It remains to be explained how this transposition could have occurred.20 An accident in transmission may have a variety of causes, and there are many examples in classical literature of omission and transposition of lines for which there is no apparent palaeographical cause. This is even more common in dramatic texts in which actual performances of the piece may have caused alterations in the text. So there is no imperative need to find a specific cause that justifies the transposition. However, a possible reason might be at least suspected in the fact that the affected lines have a similar ending, which is the most typical cause for a copyist omitting some of them, which 18 On the nuptial imagery in the last scene see M. R. Halleran, “Alkestis Redux,” HSCP 86 (1982) 51–53, and “Text and Ceremony at the Close of Euripides’ Alcestis,” Eranos 86 (1988) 123–129; Parker, Alcestis 275–276; H. P. Foley, “Anodos Drama: Euripides’ Alcestis and Helen,” in R. Hexter and D. Selden (eds.), Innovations of Antiquity (New York/London 1992) 133–160; R. Buxton, Myths and Tragedies in their Ancient Greek Contexts (Oxford 2013) 201– 218; Herrero de Jáuregui, Trends in Classics 8 (2016) 205–209, 218–219. 19 Parker, Alcestis 270 (supposing 1097 comes after 1096) says that it “is a glaring non sequitur” and proposes exempli gratia at least two lost lines of this sort: “Heracles: ‘do you believe I am your friend and wish you well?’ Ad- metus: ‘yes, I certainly do.’ Heracles: ‘then receive this lady’.” This misses all the ironic effect, but in any case with the transposition proposed here, line 1097 cannot be said to be a non sequitur. 20 Cf. M. D. Reeve, “Interpolation in Greek Tragedy, III,” GRBS 14 (1973) 149 n.9: “transposition of a self-contained and dispensable passage is never preferable to deletion unless some palaeographical cause can be found for the error.” In this case both conditions (palaeographical con- sistency and indispensability) are sufficiently met to defend transposition against deletion. MIGUEL HERRERO DE JÁUREGUI 597 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 would entail a later correction, by the same or another copyist, in which the lines are reinserted in the wrong order. Con- secutive lines 1093, 1096, and 1095 end with a sigma, though admittedly a single letter is not a very compelling argument to suspect confusion. However, the possibility of a slip of the eye in this passage is strengthened by the similarity of the strings of last letters in the crucial lines (in uncial form, taking into ac- count the respective similarity of angular and round letter shapes): 1092 -ΕΩΝ, 1093 -ΝΕΙϹ, 1096 -ΔΟΥϹ, 1095 -ΛΟϹ, 1094 -ΛΩΝ, 1097 -ΜΩΝ. A possible saut du meme au meme be- tween lines 1092 and 1097 would explain the omission of lines 1093–1097. And another likely possibility is the confusion of 1093 -NEIϹ (with round epsilon) and 1095 -ΛΟϹ, which would explain the omission of 1096 and 1095.21 A later correction, by the same or another copyist, adding the omitted two lines in the margin or below, would have caused their insertion after 1094. Whether or not caused by these palaeographical reasons, once the transposition took place, the new position of line 1094 would have led to the aforementioned variant καλεῖν and the gloss ἴσθι because of the awkwardness of the resulting syntax. Besides the linguistic difficulty, the transposed text presented a literary problem, since Heracles’ transition to 1097 was not ex- plained: that is why editors such as Kirchhoff or Parker posited a lacuna. The two transposed lines were indispensable, and their absence from their original place caused awkwardness. 21 I am grateful for this idea to my colleague Álvaro Cancela. He also suggests a possible confusion of 1096 -ΔΟΥϹ and 1095 -ΛΟϹ that could cause the omission of one line. These proposals resist well West’s judgment: “the palaeographical argument is looked up to as an ideal by many whose understanding of palaeography is minimal and who think that in order to make a conjecture palaeographically plausible it is only necessary to print it and the transmitted reading in capitals”: M. L West, Textual Criticism and Editorial Technique (Stuttgart 1973) 59 n.11. Besides, Cancela underlines that many omissions of lines in poetic texts have no apparent cause (e.g. Soph. El. 1244 is omitted in T). 598 LINE TRANSPOSITION IN ALCESTIS 1094–1096 ————— Greek, Roman, and Byzantine Studies 57 (2017) 589–598 Line transposition is a common phenomenon in poetic (and particularly dramatic) texts that, furthermore, can find here palaeographical justification. But above all, it makes this pas- sage grammatically sound and literarily consistent. We can be confident, therefore, that the original order of the lines was 1093/1096/1095/1094/1097. May, 2017 Universidad Complutense de Madrid miguelherrero@filol.ucm.es