^{page 51 note 2} Page, p. xviii.

^{page 52 note 1} When elsewhere in Euripides stands a long way from what it refers to, the structure of the whole passage makes its reference clear (e.g. Ion 384, 645). At Helen 454 πονvov naturally lacks a reference, and at Medea 1117 was doubtless accompanied by a gesture.

^{page 52 note 2} Weil on 1058: Ces most sont en contradiction flagranu avec les vers suivants … ’; Bethe, 9.

^{page 53 note 1} Cf. Matthiae: ‘ut quod sentio dicam, de eo, quod homini fixum et est, usurpari non memini, quum dicatur.quae divinitus vel fati necessiisic destinata sunt.’ Wilamowitz (p. 494) is with the usual sense, but fate has place in Medea's deliberations.

^{page 53 note 2} ‘Anyway the thing is done—i.e. the murder of the princess with its inevitable conseqoteces’ (Verrall). It should be added that must refer sideways to the situation the speech rather than forwards to 6 from which it is cut off by καì δή.

On lldt. 9. III., 5 adduced by Lenting, see Stein. The passages collected by Matthiae are all apodotic kind illustrated by Kühner-corth § 384. 5, who say ‘die Bezichung auf Zukunft geht entweder aus der Kon des Satzes oder aus dem ganzen der Rede hervor’. Verrall before stopped short of the truth when he this interpretation of breed’.

^{page 53 note 3} Sec the Concordance of Allen and Italic (California and Cambridge, 1954).Google Scholar Wilamo (p, 494) took the verb as passive, a use in Attic.

^{page 53 note 1} Editors who remore 1062–3 and still want to refer to the murder of the children arc surely asking too much of the audience. If they dislike the repetition of they would do better to delete 1064. Cf. Kvíčala, 269: ‘Retten lässt sich nur auf eine Weise, nämlich wenn man diesen Vers nach 1066 setzen würde. Dann würde sich auf das der Glauke von Medeia bereitete Schicksal beziehen und zu wäre als Subject aus 1066 νύμøŋ zu ergänzen. Doch sobald man an dicse Möglichkeit denkt, dann stellt sich auch eine andere Möglichkeit heraus, nämlich dass V. 1064 von einem Interpolator an den Rand geschrieben wurde mit der Bestimmung nach 1066 eingeschoben zu werden, und dass er sodann in den Abschriften hinter 1063 gerieth.’

Valckenaer (on Phoen. 1286, 1282 in modern texts) removed 1240–1 (= 1062–3) but gave no reason beyond the repetition. He should be followed, because the lines merely repeat the content of the previous four. Cf. Leo: ‘Nur sollte man 1062.63 mit Valckenaer Porson Elmsley an ihrer Stelle lassen und die den Ton leidenschaftlicher Kürze in drängendster Gefahr unterbreehenden Verse 1240.41 streichen’.

^{page 54 note 1} If he does not go, there, is always a clear reason: sometimes, for instance, he refuses or resists (e.g. Hcld. 341, Andr. 433), sometimes he stays for a farewell (Hcld. 573, 1A 678). Orders issued by gods ex machina are carried out when the actors leave the stage.

Three passages in Euripides are less straightforward. At Med. 89–105, the nurse orders the children inside (89) but goes on speaking to the and before he can leave the stage with them Medea is heard moaning inside the palace (96–7); the nurse thereupon repeats the order with more urgency (100–5), and in they go. At Phoen. 1682, Creon orders.Antigone out of the country and promptly leaves the stage himself while Antigone remains; but the absurdity must be laid at the door of an interpolator (see most recently Fraenkel, , Sitzungsberichte der bayerischen Akademie 1963, Heft 1, 112–13).Google Scholar At El. 358, the bids Orestes and Pylades enter his cottage, but before they comply in 393 Orestes holds forth about at what length, ediion disagree (369–72 del. Vitelli, 373–9 del. Wilamowitz, 383–90 ‘suspecti’ Murray;. 386–90 del. Wilamowitz; it can all go, and along with it 368 and 396–400, as will argued elsewhere). At line 360 of the same passage, the tells the Orestes to take the into the cottage, but they are still outside at 393–4, presumably awaiting an order from Oresta himself (another possibility, suggested by Mr. Barrett, will be discussed in connection with Orestes’ speech).

^{page 54 note 2} Editors all prefer a comma at the end al this line, but Ҽєíρ σҼόμα(whatever that is), and make a poor subjed both logically and grammatically for No one has ever printed comma at the end of 1075.

^{page 54 note 3} (rather than more like a note on 1068, but the matter trivial. Commentators all agree that Meda means

^{page 55 note 1} No other commentator even raises the question.

^{page 55 note 2} The difficulty seems first to have been noticed by Stadtmüller: ‘Eine Ungereimtheit welche den Herausgebern entgangen zu sein scheint, findet sich zum Schluss dieses Monologs, wenn Medea das Resultat des Kampfes, zusammenfasst, den sich die siderstreitenden Empfindungen geliefert, Der erste beiden Verse soll bedeuten: due Leidenschaft ist stärker als die bessere tünsicht, er bedeutet aber: die L. triumphirt ober meine Entwürfe, sie zerstört also ihr agences Werk. Denn in Medeas Mund die Worte nichts bedeuten als den von ihr ersonnenen Man der Rache; dies beweisen die beiden Verse v. 769: und v. 772: dies zeigt in unwider-legbarer Weise die übereinstimmende Bezie-hung der Worte in v. 1044 und v. 1048, man müsste denn eine amphibolische Ausdrucks-weise der Art, für zulässig halten, dass dieselbe Person in einer Rede den nämlichen Ausdruck setzc zur Bezeichnung der entgegengesetzten Motive, welche für und wider eine That sprechen.’ Stadtmüller's solution is to read for

^{page 55 note 3} ‘Zuerst denkt sie an die Möglichkeit, ihre Kinder durch die Flucht der Verfolgung der Feinde zu entziehen, nachher ist yon diesem Auswege, der so nahe lag, nicht die Rede. Medea stellt die Lage so dar, als müßten die Kinder nothwendig schmachvoll den rachsüchtigen Gegnern erliegen, und deshalb sei es besser die Kinder selbst zu tödten.’ This inconsistency is essentially the same as the one observed by Hermann in 1056–61.

^{page 56 note 1} If he had come down firmly against 1056–80, he might have decided that they were meant to replace not 1021–55 but 1040–55, or that they were simply an addition composed by someone who shared the taste of modern critics for frequent changes of mind.

^{page 56 note 2} His reason was the one mentioned above on 1067–9.

^{page 56 note 3} He had also gone so far as to say ‘als Schluß einer Medearede kann ich mir 1049–55 etwa vorstellen, aber nicht mitten in dieser Rede’ (p. 14).

^{page 56 note 4} As regards the rhetorical structure of the two Jachmann is right, but there is more poetry in Medea 1056–80 (1069–75). Admittedly it may be an accident of quotation that Medea in Neophron speaks no words to her children except the order to withdraw from her sight; but her hesitation is caused not by their looks and smiles but by the drab reflection that she is in danger of doing wrong and estranging her dearest ones (2–3).

‘Daß Stob. den Monolog als ganzen ausgehoben hat, beweist v. I mit der Anrede an den θ03C5;μός und der Schluß’, according to Diehl, RE s.v. ‘Neophron’ (1935) 2433; but the quotation comes from the chapter ’Oργñς, and the assertion that the addrecu to her must have begun the speech a unfounded—and incompatible, moreover, with Diehl's own view of Medea 1056 (Euripides Medea mit Scholien [kleine Texte 89, Bonn, 1911Google Scholar]).

Von Arnim's onslaught on Neophron (pp. ix–x) could have been directed equally well at Medea 1056–80.

^{page 56 note 5} The Drama of Euripides (London, 1941). 160–2.Google Scholar

^{page 56 note 6} Humanitas, iv (1952), 14–15.Google Scholar This change makes the staging clear, but is unduly peremptory (contrast 894–5 Hec. 171–3 Phoen, 1264 and Or. 112 and in spite of Dodds's remarb is no word to import into tragedy by emendation. At the possibility considered above (p. 54), that Medea in 1069 raised her voice and shouts into the house, Dodd's levels the fair objection that ‘the word are no instruction to invisible slaves, but an integral part d her soliloquy’.

^{page 56 note 7} Euripide i (Budé, Paris, 1947), 162 n. 3.Google Scholar

^{page 57 note 1} In fact he opts for a different escape p. 60 n. 3).

^{page 57 note 2} Eranos xlviii (1950), 44–5.Google Scholar

^{page 57 note 3} SIFC xxv (1951), 65–82.Google Scholar

^{page 57 note 4} e.g. that 1078–9 introduce a moral content:l out of keeping with Medea's attitude in the rut of the play (pp. 70–2) and that they . have been written after Plato (pp.13.4). As Müller is aware (p. 74 n. 1), they cannot have been written much more than a century after Plato, because they were into service by Chrysippus (Galen 5. e-5–6 7, 408 Kühn; cf. Diog. 7. 180).

^{page 57 note 5} Philologus, ci (1957) 217–37.Google Scholar

^{page 57 note 6} p 74.

^{page 57 note 7} Voigtländer admits as much:'Die Frage Medeas Tragik spitzt sich am Ende Tat auf die Interpretation der Verse 1979–80 zu’(p. 219).

^{page 57 note 8} Pohlenz too maintains that Müller Medeas Tragik’ (Die griechische [Göttingen, 1954], ii. 107).Google Scholar According however, he ‘verkennt die Besonderless dieser in den erhaltenen Tragödien einzigartigen Partie’ (Die tragische Dichtung der Hellenen [Göttingen, 1956], 164 n. 2).Google Scholar

^{page 57 note 9} To make this point is not to deny that the authenticity of some passages in tragedy depends on their relationship to the rest of the play. Antigone 904–20 is a notorious instance, except that some scholars are loath to ascribe one or two clumsy expressions to Sophocles.

^{page 57 note 10} pp. 238–37.

^{page 57 note 11} If Voigtländer can appeal to Pohlenz in support of his doctrine that emotion is entitled to waive the laws of logic, appeal may also be made to Regenbogen: ‘&vermag ich nicht einzushen., wie man einen solchen Widerspruch mit der Formel von den changes of temper [Page on 1058] abtun kann’ (p. 47).

The general question is of considerable importance, affecting as it does a fair number of passages in a variety of authors. That Voigtländer's doctrine has a long history is shown by a wry comment of Ribbeck's on Aen. 4. 548–9: ‘Scilicet acerrimo maeroris impetu corrcptam modo hue modo illuc animo fluctuare docebunt elegantiores intcrpretes [e.g. Austin: ‘Dido has forgotten all logic now’], atque id ipsum, quod sine ullis orationis tamquam compagibus et coagmentis singula membra sibi invicem ut Cyclopum saxa superfunduntur atque insiliunt, animi furentis et omnibus sanae rationis vinculis liberi tumultus ingeniosissime exprimi [read ‘exprimcre’ or ‘eo ipso’; above] admirabuntur' (Emendationes Vergilianae [Berne, 1858], 5). The modern world is full of elegantiores interpretes, one whom will perhaps favour the sceptical some day with a demonstration that calculated illogicality was a recognized device in ancient poetry; or were the poets themselves thrown off balance by the emotion of their characters?

^{page 58 note 1} Rivier, , Enlretiens VI Fondation Hardt (Geneva, 1960), 60–8, 84–5;Google Scholar

Lesley, , ib. 83, 139–42, and Gymnasium, lxvii (1960), 18–19;Google Scholar

Christmann, , Bemtrkungen van Text der Medea des Euripides (Heidelberg, 1962), 125–45;Google Scholar

Schwinge, , Die Stellung der Trachinierinnen im Werk des Sophocles (Göttingen, 1962), 28CrossRefGoogle Scholar

Schlesinger, Hermes xciv (1966), 28–32

Diller, ibid. 267–75

Steidle, , Sludien zum anliken Drama (Munich, 1968), 157–65Google Scholar

Baumert, , ΕΝΙΟΙ ΑΘΕΤΟίΣΙΝ(Tubagen, 1968), 161–7,Google Scholar

Seeck, , GRBS ix (1968), 291–307.Google Scholar Voigtländer's article is invoked by Lesky and Schlesinger, and Diller is warm in its praise: ‘vor allem hat H.-D. VOIGTLANDER eira durchweg überzeugende Widerlegung voe MÜLLERS Gründen gegeben’ (p. 268).

Schwinge is out of touch with the problem, Schlesinger commits the same error a method as Voigtländer (see p. 57): the monologue ‘wird erst dann voll verständlich wenn man ihn aus dem Stück als Ganzen und aus seiner Stellung in der Dann verschwinden auch gewise Anstöße der letzten zwanzig Verse’ (p. 32) Steidle fancies he can circumvent the illogicality of 1056–66 by reading two gratuitous implications into 1060–1 (pp 158–61). Baumert discovers a different implication in 1060–1 (p. 190 n. 2: ‘Medea hat keinerlei Anspruch auf die Kinder; be einer Scheidung fallen sie immer dem Vate zu … Da Medea so fremdes Eigentum no sich führt, sind die Kinder in Athen tatsäel lich gefährdet, wenn sich Aigeus auf eince formal-rechtlichen Standpunkt stellt. Vermutlich waren diese Rechtsverhältnisse der Athenern geläufig’ and can also supply two orders in defence of 1053–1069 986, no exception to the principle down above in p. 54 n. I, and Ion 1266, order of another kind but disobeyed for equally good reason).

^{page 59 note 1} In the lacuna stood something like ich nicht auf der Flucht mit meinen ergriffen werden, wenn die Braut meine stirbt?’ (pp. 133–6 This idea is the most intelligent yet put forward by any champion of 1056–80, but it requires a connecting particle in 1064. Christmann's treatment of the illogicality in 1056–66 (pp. 125–33) is lucid and exhaustive; other things he shows that emending in 1058 Hermann, adopted with different articulation by Pohlenz, op. cit. u. 106) merely transfers tlie inconsistehcy to 1045 (cf. p. 60 n. 3, and see also p. 55 n. 3).

^{page 59 note 2} ‘Medea stellt fest, daß die Leidenschaft Herr über ihre Pläne ist’ (p. 274). Diller's cose rests on Med. 443–5 where he accepts Elmsley's in prefence to Porson's and takes to mean (‘eine andere, eine Königin kam als Herrin über dein Ehebett ins Haus’); but even if is right, there is nothing to stop the phrase meaning ‘another queen superior to your bed’ can stand for as indeed Diller makes it, because σοίt begins the sentence). In general, anyone who is a thing is logically susceptible of being the same thing, which is' not true of Diller's Fr. 718 cited by Steidle (p. 165 n. 180) in connection with Diller's suggestion, is another argument against it if any more were needed.

^{page 59 note 3} Seeck will have no truck with psychological defences of 1056–64 (see especially p. 294 a. 12), but his own approach is an extraordinary mixture of distinction-drawing (pp. 297–304) and distinction-blurring (p, 306.-n. 27). So far from solving the problems in 1064, the deletion of 1060–3 aggravates them.

^{page 59 note 4} They cannot take it to mean ‘do not be so cruel as to leave the children here’, because according to 1021–39 the pain of separation is Medea's and not theirs.

^{page 60 note 1} ‘Von dem, was sie vorhat, darf sie freilich nichts verraten, und unwillkürlich gibt sie sich selbst der Vorstellung hin, als gälte es nur eine Trennung hier auf Erden, als würden die Kinder fern der verbannten Mutter in Korinth weitcrleben. Verständ-nislos [i.e. presumably ‘not understanding the true situation’] lächeln sie die Kinder an' (Pohlenz, op. cit. i. 2553 cf, ii. 106). He does not explain how the phrases dis-cussed above square with this view. Verrall too no more than touches on the problem when he says: ‘Medea in this speech, where the children are present, uses no such unmistak-able language [as in 1062–3].’ Grube goes a little further: ‘Clearly, neither the children, nor any one besides the chorus, must hear these words [1062–3]. Both lines occur again at 1240–1, and are, therefore, usually omitted here as an actor's interpolatibn. That seems quite probable, but it does not solve the problem, for the whole passage (1056–69) is almost as clear' (160 n. I).

^{page 60 note 2} Elmsley apart, only Grube and Steidle seem to have considered this possibility, Grube rejects it (p. 160 n. 1) and presses it further than 1053 permits (p. 163 n. 67: ‘Von den Kindern ist von V. 1043 bis V. 1069 nur in der dritten Penson Rede'; cf. a later statement on the same page: ‘Die Aufforderung an die Kinder, ins Haus zu gehen, kann … nur gesagt sein, ohne oder höchstens mit einem kurzen Blick auf sie’).

^{page 60 note 3} The only lines that editors have removed from this section are 1045 ( Kvíčala, Nauck, Wecklein, Méridier) and 1048 (del. Kvíčala), but 1046–8 or 1045–8 could go if necessary. The deletion of 1045 is usually coupled with an alteration in ( Hermann, Barthold, v. dd. Kvíčala) so that Medea considers no except killing the children or leaving there in Corinth; but if 1056–80 are deleted, 1045 can stand (Wecklein's objection to is captious). Cf. above, p. 59 n. I.

^{page 60 note 4} Die Darstellung des Menschen im des Euripides (Basel, 1947), 62.Google Scholar

^{page 60 note 5} p.52.

^{page 60 note 6} The retention or deletion of 1240–1 p. 53 n. 1) makes no difference. Seeck, op. cit 304, anxious to eject 1062–3, correctly takes 1240–1 with the preceding lines (for a different view see Meissner, , Hermes, xcvi [1968], 157Google Scholar) but is mistaken in holding that they are integral.

^{page 61 note 1} Most of it has been used on 1056–80, next to none on 1021–55. Why?

^{page 61 note 2} Particular mention may be made of the philosophical debate between Euripides and Socrates that Snell has manufactured out of Medea 1078–80, Hipp. 377–83, and Protagoras 352 d (Philologus, xcvii [1948], 125–34,Google Scholar superseded by Scenes from Greek Drama [Berkeley, 1964], 47–69;Google Scholar in the more recent version Snell tries to meet the objections of Barrett, , Hippolytos [Oxford, 1964], p. 229,Google Scholar but ignores Müller's assault on the foundations of his theory).