¹ This paper had its genesis as a talk first delivered at Swarthmore College, in November 1984, and since delivered, in various forms, at a meeting of the California Classical Association in Los Angeles, at Hobart College, at the University of Iowa, and at Cornell University; and I am grateful for questions and criticisms from many people, too numerous to mention individually, on these various occasions. I am particularly indebted to the encouragement and careful criticism of Professors Gordon M. Kirkwood and Phillip Mitsis, without whose help I should never have dared to submit it in written form to the sphere of larger scholarly debate; the usual disclaimer is in order, that the final form and details of argumentation are entirely my own responsibility.

² The standard bibliography, tallying the votes of every scholar who has treated the question, is Hester, D. A., ‘Sophocles the Unphilosophical’, Mnem. 24 (1971), 11–59CrossRefGoogle Scholar; see Murnaghan, Sheila, ‘Antigone 904–920 and the Institution of Marriage’, AJP 107 (1986), 192–207Google Scholar, for additional citations (since 1971). The most recent condemnation of the lines known to me is Andrew Brown's commentary on the play (Warminster, 1987). R. D. Dawe, in his new Teubner edition (1979), prints the lines without brackets, and I am informed by the editors of CQ that Hugh Lloyd-Jones and N. G. Wilson will do likewise in the forthcoming OCT. It is interesting that, whereas it has been many decades since a standard edition has dared to bracket the lines, commentators (who can back up their printed text with lengthy verbal justification) have had no such hesitation.

³ The traditional points of philological argumentation pro and con are catalogued, and their respective values weighed, by Szlezák, T. A., ‘Bemerkungen zur Diskussion um Sophokles, Antigone 904–920’, RhM 124 (1981), 108–12.Google Scholar

⁴ As Gerhard Müller puts it, the lines are regarded as ‘schlecht, aber echt’ (Sophokles: Antigone [Heidelberg, 1987], p. 200).Google Scholar

⁵ See Barthes, Roland, ‘Criticism as Language’ (TLS [27 09 1963], 739f.).Google Scholar As Barthes says, the job of criticism is to evolve the linguistic tools for rendering a systematic and coherent account of the language of the author under consideration; criticism's aim is one of ‘reconstituting the rules and compulsions which governed the elaboration’ of the work under consideration–to reconstitute, that is, (not its message but) its system. As will emerge, it is just this that I feel that previous defences of the lines in question have generally failed to do. The same Barthian approach shows why, theoretically, it might be possible, despite the long history of discussion of the lines, to make a substantially new contribution to the question at all. Criticism ‘is the ordering of that which is intelligible in our own time’; this is why criticism remains viable at all. I will argue that both those who have excised and those who have defended the lines have found them unintelligible because of a flawed representation of the poet's dramatic system; whereas in recent decades we have been provided with new tools for systematising Sophocles' methods so as to find the lines intelligible.

⁶ Eckermann, Johann Peter, Gespräche mit Goethe (Leipzig, 1836–1848)Google Scholar; conversation of 28 March 1827. Observe that Goethe then goes on to say, ‘I would give a great deal for an apt philologist to prove that it is spurious and interpolated’, already revealing the subordination of the philological prong of the attack to the prior belief that the lines are contextually unacceptable. For the cultural context of Goethe's remark, as well the original attack upon 905–13 of Jacob, August, Sophocleae Quaestiones (Warsaw, 1821)Google Scholar, and the defence of August Boeckh, ‘Über die Antigone des Sophokles’, Königlich-preussiche Ak. Wiss., Hist.-Phil. K1. (1824), 41–88, 225–37Google Scholar, along with other nineteenth-century readings of the play, both scholarly and literary, see now Steiner, George, Antigones (Oxford, 1984), Chapter 1.Google Scholar

⁷ I assume largely on grounds of scene-structure (though the point is not crucial to this paper) that the MSS. are right and 572 is spoken by Ismene. The scene is dramatically more effective and Sophoclean if it is constructed in blocks of paired speakers: first Creon and the guard, then Creon and Antigone, then Antigone and Ismene, then Ismene and Creon, and finally Creon and the Chorus. In this way the scene winds down with first Antigone at 560, then Ismene at 572, uttering a final significant line, then falling silent. Cf. most recently Davies, M., ‘Who Speaks at Sophocles Antigone 572?’, Prometheus 12 (1986), 19–24.Google Scholar

⁸ The Greek text is that of the old OCT of Pearson; it is meant to be non-controversial, and does not diner significantly, for example, from the text printed by Dawe in his new Teubner edition.

⁹ That the borrowing runs this way even if Ant. 904–20 are genuine, i.e. that Herodotus did not take the lines from Sophocles (or from a common source), is universally assumed, and not to be contested here; see Szlezák, op. cit. (n. 3). The usual publication date for Herodotus' work is 430–425 or so, whereas the Antigone is dated to roughly 440, give or take a couple of years (see, most recently on the question, Lewis, R. G., ‘An Alternative Date for Sophocles’ Antigone', GRBS 29 [1988], 35–50)Google Scholar; but ‘publish’ did not mean in the ancient world what it does to us. Herodotus is generally taken to have spent some years before 441 at Athens, to have ‘lectured’ there, and to have been a member of the Periclean circle and a friend of Sophocles; and it is accepted that there is other Herodotean verbiage in Sophocles. See Immerwahr, Henry R.'s chapter on Herodotus in the Cambridge History of Classical Literature (Cambridge, 1985), i.426Google Scholar, and, for the definitive discussion of the relationship between Herodotus and Sophocles and the best list of probable borrowings, the article on Herodotus by Felix Jacoby in the 1913 supplement (II) to Pauly-Wissowa. There is thus no objection a priori to the notion that Sophocles could have borrowed these lines from Herodotus; it is to be admitted, however, that no other such borrowing is anywhere near as long or as verbally close as this one.

¹⁰ Waldock, A. J. A., Sophocles the Dramatist (Cambridge, 1951).Google Scholar

¹¹ These are catalogued by Kakridis, Ioannis Th., Homeric Researches (Lund, 1949)Google Scholar, Appendix III, and in the sources cited by Bowra, C. M., Sophoclean Tragedy (Oxford, 1944), p. 94.Google Scholar The motif is no. 253.3 in Thompson, Stith's catalogue, Motif Index of Folk-literature (Bloomington, 1932–1926), where also see further bibliography.Google Scholar

¹² Müller, , op. cit. (n. 4), p. 199Google Scholar, though strongly condemning the Sophoclean lines, is really making the same admission as Waldock when he says of the Herodotean sentiment that ‘die Argumentation…macht dem König, wie dem Herodot, Spaβ.’ If we add ‘wie auch dem Sophokles’, this ceases to be a reason for impugning the Sophoclean lines. Indeed, a view of the Herodotean context very similar to Waldock's appears in Tycho's discussion of the Sophoclean lines, as a reason for accepting them as genuine (see below); Tycho's argument, however, falls short in what I see as an essential regard, for, as we shall see, it gives no other reason, apart from their memorability, as to why Sophocles would have wanted to use the lines here. One could also argue that some of Sophocles' other Herodotean borrowings are likewise more memorable than relevant to their contexts: the use at O.C. 339, for example, of the Egyptian inversion of Greek sexual roles in the management of the household–apparently a glance at Hdt. 2.35 (the men stay home and weave, while the women go out and do the marketing)–seems little more than a learned tour de force.

¹³ No one, for example, calls the closing lines of this speech, 925–8, sophistic or inappropriately rational, though they are structured and balanced to the point of high artifice. They consist of two pairs of two lines, the pairs being introduced by εἰ μ⋯ν τ⋯δε respectively; line 1 ends with εἰ δ⋯ οἵδε, line 3 with its antithesis καλ⋯; line 2 begins with κακ⋯, line 4 with παθ⋯ντες; even the use of an optative main verb in both lines 2 and 4 is an artifice, since the verb need not have been optative in line 2.

¹⁴ This, after all, is why Aristotle cites the passage with approval: it provides a rationale (not just a motive, but a reasoned argument) for something which otherwise is hard to understand. See also in general Reinhardt, Karl, Sophokles ⁴ (Frankfurt, 1976)Google Scholar, who argues in essence that the enunciation of such reasoned rules is the very business of Sophoclean tragedy.

¹⁵ This is implied not only by 904 τος φρονοσιν ε (itself one of the impugned lines), but also by her repeated assertions (903, 921, 924, 942f.) that her punishment is undeserved and improper.

¹⁶ Besides, if the lines are (genuine but) borrowed from Herodotus, then the form of 909–12, and perhaps that of 908 (if it is based on the last clause of the Herodotean passage), are a feature of the original; and it is then absurd to second-guess Sophocles by insisting that if he wanted to borrow the lines he should (or would) have borrowed them unfaithfully. But this leaves only 905–7 to object to; yet these are merely the minimum required in order to work 909–12 into the speech, especially since 905–7 are needed to render the existence of husband and child, which was real in the Herodotean original, hypothetical in Antigone's case. Thus the charge of ‘sophistry’ is to some extent just another way of saying that the lines are borrowed from Herodotus, and also repeats (4).

¹⁷ We might, of course, dismiss this objection as an instance of wanting to eat one's cake and have it too: if the lines are clumsy and illogical, then they are not sophistic and rationalistic, and insofar as Antigone seems not to have thought out her ‘sophism’ clearly, but is speaking confusedly, as one does in moments of crisis, Sophocles is rendering the sentiment less ‘sophistic’, and more dramatically believable, just what the attackers of the lines would have him do. Besides, the search for formal validity in Antigone's words is probably anachronistic and misplaced; the notion of a formally valid syllogism, that is, the expectation that all valid inferences can be reduced to recognisable syllogistic form, is purely Aristotelian in origin (see Barnes, Jonathan, ‘Proof and the Syllogism’, in Berti, Enrico, ed., Aristotle on Science [Padua, 1981], pp. 17–59)Google Scholar, and even for Aristotle the irreducibility of a given argument to a valid syllogism, especially an argument in the world of human affairs, does not mean that the argument is not respectable (see Burnyeat, Myles F., ‘The Origins of Non-deductive Inference’, in Barnes, Jonathan et al. , edd., Science and Speculation [Cambridge, 1982], pp. 193–238).Google Scholar

¹⁸ So SirJebb, Richard, Sophocles: The Plays and Fragments, Part III: The Antigone, 3rd edn. (Cambridge, 1900), ad loc.Google Scholar

¹⁹ Take, as a random example, 897 ⋯ν ⋯λπ⋯σιν τρ⋯φω, just above. Jebb states that this is equivalent to ⋯λπ⋯ζω, but there is no parallel whatever for this particular expression, nor for any of its features: neither for τρ⋯φω taking indirect discourse (in case ⋯ν ⋯λπ⋯σιν is to be taken adverbially of manner), nor for τρ⋯φω taking ⋯ν + dat. as an object rather than the simple acc. (in case ⋯ν ⋯λπ⋯σιν is supposed to mean ⋯λπ⋯δας or ⋯ν ⋯λπ⋯σιν τοτο, the whole expression thus taking an infinitive, like Ajax 606 ⋯λπ⋯δ᾽ ἔχων), nor for τρ⋯φω used intransitively (in case the whole expression is supposed to mean ‘I feed upon hopes’, or simply ‘I am in a state of hope’, and again take an infinitive; this must be Jebb's view, since he compares ⋯ν ⋯λπ⋯δι εἰμ⋯). Yet editors have not felt called upon to print anything else, despite Müller's strenuous objections and his perfectly reasonable conjecture στρ⋯φω. Apparently we are supposed to accept the expression as a unique but perfectly possible poeticism. Very well, I accept it, and perhaps it is not really particularly difficult; but I personally find it much harder to understand than τοδε and κατθαν⋯ντος, which I would expect any third-year student to be able to figure out.

²⁰ A number of prose examples of headless genitive absolutes are listed in Kühner-Gerth, such as Xen. An. 5.4.16 οἱ πολ⋯μιοι προσι⋯ντων ⋯σνχ⋯αζον, where ‘the Greeks’ must be understood from the previous sentence. If it comes to that, indeed, 455 θνητ⋯ν ⋯ντα is ‘headless’, yet no one objects to it for that reason, even though σ⋯ has to be supplied from τ⋯ σ⋯, two lines earlier, and after a listener might have assumed that κηρ⋯γματα was the subject of the ὥστε-clause. As for Jebb's argument that the word for that subject, π⋯σεως, does not exist, one can scarcely keep from laughing; Jebb knew better how language works, than to suppose seriously that a hearer would have been prevented, just because the specific word π⋯σεως was unusual, from being able easily to supply the concept of a husband! And besides, if π⋯σεως was truly an unimaginable form, which I am not convinced it was, all the more reason for the genitive absolute to remain, headless! (Also, the ‘interpolator’ must have operated in the fourth century, as the citation of these lines by Aristotle proves; is he supposed not to have known Attic?)

²¹ I say ‘if εἰ τ⋯κνων μ⋯τηρ ἔφυν is supposed to imply that Antigone has also hypothetically lost one of her children’, because perhaps in fact we ought to interpret the lines just as they stand. (It turns out, though I did not know it when I conceived this argument, that I am somewhat anticipated here by the remarkable discussion of Kaibel, G., De Sophoclis Antigona [Göttingen, 1897]Google Scholar; this defence of the lines, though Kamerbeek calls it the best ever, is too little known.) Lines 905 and 910, which are both impugned, have in common that the faultiness of expression in the one, and of reasoning in the other, have both to d o with the hypothesis that Antigone's child had died: neither line states the hypothesis or its implications properly. But then that hypothesis is nowhere stated; we are simply assuming it, presumably because of the Herodotus passage. Suppose, then, that Antigone is not putting forward two separate hypotheses as to who has died, but only one, a husband, and is discussing this with regard to its implications for her ever having a complete husband-children family unit. Then she might be saying: (905) I wouldn't have buried a dead husband if I were already a mother, nor (906) even if I were not; since (909) in the first case I could have another husband, thus restoring the whole family unit, and (910) if my old husband had died while I was still childless, I could still have children from the new husband. The advantages of this reading are: (i) 905 and 910 are now correctly stated, with no contamination of hypotheses; (ii) 910 τοδε now has a referent, namely the dead husband in 909; (iii) the order of discussion is now parallel, with 909 answering 905, and 910 answering 906, rather than chiastic as on the usual reading. But there is also a new disadvantage: we have to suppose that the force of 905f. οὔτε … οὔτε is such as to imply that 906 includes the negation of 905, i.e. that in this case she did not have children (a new unstated concept), while 905 also includes the stated part of 906, i.e. her husband is the one who has died in both cases; and it isn't clear that οὔτε … οὔτε can do this. AH the same, this is perhaps a small obscurity in relation to the obscurities of the usual reading, and the illogicalities of the usual reading are almost completely eliminated: that is, except for the strained οὔτε, every word of the passage now means exactly what it says. For a somewhat differently but just as drastically strained οὔτε, likewise after a sentence-initial negative, cf. Ant. 4 (and see the explanation of Mazon, P., ‘Notes sur Sophocle’, RPh 25 [1951], 7–17Google Scholar). Of course one could always try to emend, though (as with line 4) no convincing correction occurs to me.

²² See Blumenthal, H. J., ‘Euripides, Alcestis 282ff., and the Authenticity of Antigone 905ff.’, CR 24 (1974), 174–5.Google Scholar Alcestis uses reasoning which runs, in effect: ‘I choose to die for Admetus, even though had he died I might have had another husband; indeed, his parents should have died for him, since he being an only son and they being no longer able to have children, they could not have had another son.’

²³ Highly Sophoclean are 916 δι⋯ χερν λαβών (see Jebb ad loc., and Moorhouse, A. C., The Syntax of Sophocles [Leiden, 1982], Ch. 7, §6)Google Scholar; 917 ἄλεκτρον ⋯νυμ⋯ναιον, cf. 876; 917f., cf. 813f.; 919 ⋯ρμος πρ⋯ς φ⋯λων, see Moorhouse §16; 920 κατασκαφ⋯ς, cf. 891 (as well as for the lamentation over the loss of marriage), and, for the sentiment (‘I am dead among the living’), cf. 308f., 559f., 567, and esp. 821ff., 852. This sort of repetition of phrase and theme is a mark of Sophocles’ style, and indeed the list could be greatly extended; it is clear why Jacob was at pains not to involve 916–20 in the excision (and perhaps they might never have been condemned had not Dindorf opted to sweep away the entire end of the speech, all the way to 928, compared to which Lehrs' cut 905–20 seems gentle). For a splendid argument insisting that 905–20 are demanded by the sense (and not merely the syntax) of 921–8, see Kaibel, op. cit. (n. 21).

²⁴ As Kaibel puts it, ‘callidum adgnosces interpolatorem, qui sui operis fines vaferrima arte eflecit ut nemo certa ratione terminare possit’!

²⁵ I say nothing about the a priori arguments based on the overall length of 883–928, which are indecisive (see Szlezák, op. cit. [n. 3]): it is true that the speech, if nothing is cut, is twice as long as any other continuous rhesis of Antigone's, but then, if the speech is cut, it seems far too short under the dramatic circumstances.

²⁶ Gerhard Müller, op. cit. (n. 4); another very fine catalogue of defences, with powerful refutations, is that of Waldock, op. cit. (n. 10).

²⁷ Kitto, H. D. F., Form and Meaning in Drama (London, 1964), p. 171.Google Scholar Elsewhere (Greek Tragedy ² [London, 1952], p. 127Google Scholar) he calls the lines ‘the finest borrowing in literature’; but few could see this, especially on his own explanation of them, as anything but rhetorical hyperbole. Müller includes in this category the arguments of Boeckh, who appeals to Antigone's use of a ‘sophistic of doubt’, as she tries to make sense of her current situation, in order to explain not only her inconsistency but the rationalistic nature of her words; Schadewaldt, W., ‘Sophokles und Aias’, Neue Wege zur Antike, iv (Leipzig, 1929), 59–117Google Scholar, for whom the lines are an appeal to special human circumstances to explain her opposition to the state, because she now doubts her ‘absolute’ correctness; and Kirkwood, G. M., A Study of Sophoclean Drama (Ithaca, 1958)Google Scholar, who characterises the lines as ‘a momentary loss of certainty about the wisdom of what she has done’. Here, too, one should probably class Gellie, G. H., Sophocles: A Reading (Melbourne, 1972), pp. 47fGoogle Scholar: ‘She has been robbed of the security of [her earlier] moral and emotional imperatives…. She is driven back to a rationalization.’

²⁸ Certainly Antigone shows signs of feeling that her field of opponents has been enlarged from consisting of Creon merely to including also the citizens in general (so e.g. the 2nd strophe of the lyric scene, as well as 907 β⋯ᾳ πολιτν), and perhaps, in their neglect of her, even the gods (so 922f.); but then her insistence upon the injustice and impiety of her punishment (e.g. 924), in the face of such opposition, is a sign, if anything, of even greater conviction than before.

²⁹ Under this heading Müller lists the approaches of Pohlenz, M., Die griechische Tragödie ² (Göttingen, 1954)Google Scholar, and Lesky, A., Die tragische Dichtung der Hellenen ² (Göttingen, 1964).Google Scholar

³⁰ Adams, S. M., Sophocles the Playwright (Toronto, 1957), p. 54.Google Scholar

³¹ Knox, B. M. W., The Heroic Temper (Berkeley, 1964), pp. 106f.Google Scholar Add now under this category of defence Heath, Malcolm, The Poetics of Greek Tragedy (London, 1987), p. 74, n. 63Google Scholar: ‘In the former passage she explains…what justified her act; in the latter passage (a more personal context, full of pathos: note especially the framing apostrophe) she speaks of the pressures which impelled her to do it. ‘

³² Müller also speaks of a third class of defence, involving subordination of the details to a larger conception of the play. Here he includes Diller, H., Göttliches und menschliches Wissen bei Sophokles (Kiel, 1950)Google Scholar, and also Reinhardt, on whose view, he says, the sister vs. wife/mother opposition is part of, not opposed to, the divine command: nor is it a dictum she can accept or decline, but something that operates on her feelings directly. This, as Müller rightly says, is neither what the lines say nor all they say; but it will be seen that I, too, believe that only an altered larger conception of the play will encompass the lines.

³³ The impugners of the lines, indeed, must suppose that maintaining the image of this ‘unified psychological life’ overrides all other considerations in Sophocles' construction of drama, so as to entitle us to conclude, if any lines fail to maintain it, that they are suspect. As Hugh Lloyd-Jones puts it (‘Tycho von Wilamowitz-Moellendorf on the Dramatic Technique of Sophocles’, CQ 22 [1972], 214–28Google Scholar), the traditional approach has been ‘to look upon character portrayal as one of the main elements, if not the main element, in dramatic art; and by character portrayal [critics] usually meant a minute psychological analysis’.

³⁴ The methodological error involved is like the one neatly parodied by Koestler, Arthur, The Ghost in the Machine (New York, 1967), p. 17Google Scholar: ‘One might as well tell a team of land surveyors that for the purpose of mapping a limited area they could treat the earth as flat–and then subtly instil the dogma that the whole earth is flat.’

³⁵ Wilamowitz-Moellendorf, Tychovon, Die dramatische Technik des Sophokles (Berlin, 1917)Google Scholar, Phil. Unters., vol. 22.Google Scholar As Lloyd-Jones says, op. cit. (n. 33), Tycho ‘ruthlessly flings out the accumulated rubbish of over-subtle psychologising interpretation’. But he evidently has not flung it out decisively enough, for just such interpretation is what is called upon in both attacks upon and defences of the lines in question. For a standard criticism of Tycho, see Easterling, P. E., ‘Character in Sophocles’, G & R 24 [1977], 121–9.Google Scholar A recent revival of strong Tychoism (applied to Aeschylus) is Dawe, R. D., ‘Inconsistency of Plot and Character in Aeschylus’, PCPhS 9 (1963), 21–62Google Scholar; see also Heath, op. cit. (n. 31). Karl Reinhardt is often classified as a moderate Tychoist (so for example by Lloyd-Jones, op. cit. [n. 33]); but his explanation of 905–20 is actually psychological, and similar to that of Knox. According to Reinhardt (op. cit. [n. 14]), Antigone is actually describing a feeling (the feeling that one's closest relation is one's brother), but promoting it into a rationalised generation (a ν⋯μος) because Sophoclean tragedy is concerned with such ν⋯μοι, not with psychological feelings. He then uses this reasoning to shield Antigone's argument from the charge that replaceability is irrelevant, by asserting that it is only this ν⋯μος, not Antigone's particular action, which is based on the irreplaceability notion. But then Antigone feels one thing but says another; since what she does say is incomprehensible without a knowledge of what she is feeling, Reinhardt is just supplying us with a psychological story for Antigone, no differently from Kitto or Knox. Besides, if irreplaceability is not relevant to Antigone's actions, why does she adduce a ν⋯μος based upon it?

³⁶ Humphreys, S. C., Anthropology and the Greeks (London, 1978), pp. 203f.Google Scholar

³⁷ Sorum, Christina Elliot, ‘The Family in Sophocles' Antigone and Electro’, CW 75 (1982), 201–11.Google Scholar An alternative but related view, represented, for example, by Goldhill, Simon, Reading Greek Tragedy (Cambridge, 1986)CrossRefGoogle Scholar, is that individuality such as Antigone's is regarded by the playwright as an old and strange thing, that is, as a throwback to a Homeric value system which is at odds with contemporary civic values; for the Athenian moral crisis in the large described in these terms, see MacIntyre, Alasdair, After Virtue ² (Notre Dame, 1984), esp. Chapters 10 and 11.Google Scholar

³⁸ Cf. Murnaghan, , op. cit. (n. 2), p. 195Google Scholar: ‘Antigone's adoption of this argument emerges from her preoccupation at this point with the subject of marriage….Antigone draws an important distinction between ties of marriage and ties of blood’ [my italics]. Murnaghan's approach is still fundamentally psychological: for her, Antigone's understanding changes under pressure of events. Nevertheless, she rightly points out that marriage-ties are essentially artificial, and that this fact is thematically relevant; for, as she says, it was contemporarily the nature of the new polis social institutions generally, that they treated people as interchangeable; such institutions are at the heart of the play throughout, and in particular are just the sort of thing to which Antigone is opposed from the start. Thus Murnaghan would say that the ‘replaceability’ argument is really an extension of the stance that Antigone has taken all along; and this, as we shall see, is surely right.

³⁹ Humphreys, , op. cit. (n. 36), pp. 202f.Google Scholar

⁴⁰ Burkert, Walter, Greek Religion (Cambridge, MA, 1985), esp. pp. 260–4Google Scholar; see also his Homo Necans (Berkeley, 1983).Google Scholar The crisis is well exemplified, and its intimate connection with child-bearing made manifest, by something so simple as the fact that girls at marriage transferred their ‘allegiance’ from Artemis, patron of virgins, to Hera, the patron of married women–while the same Artemis who threatens virgins with death for violating that virginity and becoming pregnant is nevertheless a goddess of childbirth. This conflict in the nature of Artemis, that is, of the very essence of being female, is the subject of the Callisto myth.

⁴¹ The evidence is well summarised by Pomeroy, Sarah B., Goddesses, Whores, Wives, and Slaves (New York, 1975), Chapter 4Google Scholar; see also Cantarella, Eva, Pandora's Daughters (Baltimore, 1987), Chapter 3.Google Scholar

⁴² These quotations are collected by Lefkowitz, Mary R. and Fant, Maureen B., Women's Life in Greece and Rome (Baltimore, 1982), whose translations are used here.Google Scholar

⁴³ See, for an excellent statement of this point, Sorum (op. cit. [n. 37], p. 202): not only was it true that ‘marriage and childbirth defined womanhood’, but also this definition was contemporarily in crisis, since by Sophocles' time, ‘family structure was incorporated into state institutions, for familial descent determined the valuable rights of citizenship…. Thus the significance of the blood relationship was appropriated by the state…and the worth of strong internal bonds of familial loyalty was called into question.’ See also Lacey, W. K., The Family in Classical Greece (Ithaca, 1984).Google Scholar For tragedy in general as portraying contemporary moral and social dilemmas, see Winkler, John J., ‘The Ephebes’ Song: tragôidia and polis’, Representations 11 (1985), 26–62.CrossRefGoogle Scholar

⁴⁴ So most recently, Brown, op. cit. (n. 2), counters Bowra's argument that Antigone's being cheated of marriage is exactly the issue, by asserting that the lines permit us to conclude that had Antigone's parents been living, she would not have buried Polyneices; the argument is an old one, occurring for example in the edition of Schneidewin and Nauck (Berlin, 1880).

⁴⁵ For example, Creon says at a number of points that he will not back down from his condemnation of Antigone, because she is a woman; but this does not justify us in constructing an imaginary play, in which Creon would back down if the burier had been a man, inconsistently with the principles of his opening rhesis: the lines speak to this situation, in which (as we learned from Ismene, 62) the opposition between the respective stances of Creon and Antigone is reinforced by and intertwined with a sexual opposition. Similarly, as Seth Benardete says (‘A Reading of Sophocles' Antigone’, Interpretation 5 [1975], 148–84Google Scholar), no one would take 450 to mean that Antigone seriously considers the possibility that Zeus might have told her not to bury her brother! Even the presence of the word ‘if’ (in the present passage) does not negate this point: as Müller points out, for example, 925 εἰ μ⋯ν does not imply that Antigone really thinks there is a possibility she may be wrong; the structure merely expresses the fairness of her wish that Creon should suffer. The intimate connection between this modern raising of irrelevant hypotheses for Antigone and the psychologising misconstrual of Sophoclean character is brilliantly outlined by Jones, John, On Aristotle and Greek Tragedy (London, 1962), p. 196Google Scholar: ‘We [mistakenly] picture an inward debate, perhaps protracted and painful, which is determined in favour of burial and which results in a victory for the divine law being registered at the abiding seat of consciousness. This is our image; and it now requires us, when the question arises as to what Antigone would have done if Polyneices had not been her last brother, to answer that she would have looked within…, would have been true to herself and would have buried him. Our reading of the case is not concerned to note that Antigone was not in fact faced with the burial of anyone other than her last brother…. [T]he omission is crucial in that the presupposition which renders vain the distinction between actual and hypothetical circumstances is entirely unSophoclean–the presupposition of that abiding seat of consciousness.’

⁴⁶ Recently, Simon Goldhill, op. cit. (n. 37), has given an admirable description of the first part of the play in these same terms; this is well worth seeing, and I regard it as support for my view. My discussion, however, was conceived and written in complete independence from his; and he does not recognise the importance of the shift away from this theme in the second part of the play (so that his view of the role of Haemon [see below] is incomplete with respect to my own). Nor does my characterisation of the ‘key theme’ mean to rule out or neglect other characterisations of the same thematic material (though I do think that all such other characterisations ultimately boil down to mine). In particular, Hester's discussion of the role of τιμ⋯ in Antigone's behaviour (as well as the entire accompanying polemic against the traditional question of who, Antigone or Creon, is ‘right’) is extremely salutary: op. cit. (n. 2), and ‘Law and Piety in the Antigone’, WS 14 (1980), 5–11.Google Scholar But I cannot agree with Hester's overall psychological reading of the play, and especially of the scene in question: ‘Why should she talk of divine laws, when the gods behave this way? Surely in her last moments, when she has nothing to hope for, it is her true feelings we hear…’ This view is no improvement over that of Knox, or Kitto.

⁴⁷ Well put by Heath, Malcolm, op. cit. (n. 31), p. 74Google Scholar: ‘philia is not, at root, a subjective bond of affection and emotional warmth, but the entirely objective bond of reciprocal obligation; one's philos is the man one is obliged to help, and on whom one can (or ought to be able to) rely for help when oneself is in need.’ This is a real problem for translations, which almost universally mislead the reader into supposing that Antigone and Ismene are talking about a feeling (family affection) during this scene. Later in the play, of course, the notion of ἔρως figures prominently; but this has nothing to do with Antigone's conception of her duty and connection to her blood-family.

⁴⁸ 72f., θανε / φ⋯λη 461–4 with 466–7 (to die is a κ⋯ρδος given what her life now is, it is no grief to minister to τ⋯ν ⋯ξς μητρ⋯ς); 524–5 (Creon: so die, since that's where your φ⋯λοι are); 559–60 (Antigone ‘already dead’); and the like.

⁴⁹ Again, as Humphreys has shown, this theme represents a contemporary reality of social roles in crisis (op. cit. [n. 36], p. 201): ‘The development of the city meant–especially in democratic Athens - a sharp distinction between public and private life, between the impersonal, egalitarian interaction of the open agora and the enclosed, intimate, hierarchic relationships of the oikos. A feeling of conflict between the norms of citizen behaviour and the personal loyalties of the family is obvious in the debate between Antigone and Creon in Sophocles' Antigone.'

⁵⁰ Another nail in the coffin of arguments (1), (2), and (3). Observe that we are not simply saying, like Knox, that the later passage is an emotional re-expression of the earlier: it is crucial to my argument that Antigone be seen at 904–20 as speaking to a completely different set of issues from those at 450ff.

⁵¹ This technique of introducing a new theme while the old theme is still, as it were, standing on the stage, reminds me of the common technique now often used in movies, where we are still watching one scene, but suddenly hear the sound-track of the next scene, a few seconds before the visual changes over to show us that new scene and reveal the meaning of the sounds we are bearing.

⁵² It is also at this point that the ‘bride of Hades’ motif is first raised–by Creon, 575 (and again later at 654)–which will figure so heavily later, when Antigone laments her death before marriage.

⁵³ Note also 750 and 761, where Creon is explicit about his intention that Antigone's death should ‘replace’ her marriage to Haemon. We are now steadily pounding away at argument (1).

⁵⁴ The suddenness of the revelation of Haemon's existence is also the subject of an investigation by Jost, L. J., ‘Antigone's Engagement: a Theme Delayed’, LCM 8 (1983), 134–6.Google Scholar Jost takes a psychological, not a structural and thematic approach, but rightly points out that the matter of Haemon does not emerge until ‘Antigone has done all she can for her brother, and her trial and sentence at Kreon's hands has been accomplished.’ I think this is right, but it should be taken structurally: the first part of the play details the nature of the opposition between Creon and Antigone, leading up to Antigone's condemnation to death; once this is enacted, Haemon is introduced, and the rest of the play is material leading to the punishment of Creon.

⁵⁵ When I speak of ‘typical Sophoclean economy’, I am thinking of Sophocles' frequent trick of using one character to fill more than one function. The best example is in the double one in the OT, when the messenger who comes from Corinth to inform Oedipus of Polybus' death also happens to be the man who knows that Oedipus is not Polybus' son, who in fact received him from the Theban shepherd, and when that shepherd turns out to be the very man who witnessed the death of Laius, and whom Oedipus has already summoned.

⁵⁶ 737–9 brings out this point nicely, as well as rendering it a marvellous mirror of what Creon has done to Antigone: by trying to stand for the π⋯λις without paying attention to the will of the other πολται, Creon is a π⋯λις unto himself, the ruler of an ⋯ρ⋯μη γ (739), just as he has rendered Antigone ⋯ρμος of all living φ⋯λοι (919). Since Creon has earlier tried to equate φ⋯λοι with (good) πολται, this linkage is particularly pointed.

⁵⁷ That Creon's fate will be somehow appropriate to his deed is foreshadowed by Antigone's 927 (‘Let them suffer precisely what they unjustly do to me’) and by Teiresias' warning 1076; it is thus all the more important that the equivalency of his deeds and his punishment be clear and precise.

⁵⁸ One might say that, in a sense, Creon is punished not for what he has done to Antigone, so much as for what he has done to the universe of social roles and ties around which the play revolves: he has rendered them problematic (or, perhaps, brought their already problematic nature into high relief), and it takes three deaths to bring them back into equilibrium. Creon is left alive, but destroyed, a man whose world has blown up in his face because he tried to play games with it, like a child who misuses a familiar toy and breaks it forever, or, better perhaps, like a man who grabs the tail of a tiger, seeing only the tail, and is then surprised by having the rest of the tiger–the universe, really–turn around and bite him.

⁵⁹ In fact, Antigone's two ‘reasons’ for burying her brother participate, as we can now see, in an interesting structural parallelism. (This point is suggested to me by my reading of Benardete, op. cit. [n. 45].) Earlier, Antigone's point was that death was better than disobeying the divine laws in favour of mere mortal laws; now, her point is that joining her family (her φ⋯λοι) is better than dishonouring her blood-family in favour of the marriage-family. But since her φ⋯λοι are all dead, so that to join her family is to die, the two evaluations collapse into one: they are two sides of the same coin, not two inconsistent stances at all. The first structure expresses the notion in terms of life and death: she is willing to die, because the laws that never die are more important than the laws of people who can die. The second structure expresses the notion in terms of family: she will join the only family she now has, because the ties of the blood-family are more important than the ties of the marriage-family. This parallelism also enables us to see that the characterisation of both structures as ν⋯μοι is not problematic: the claims of the world of the dead upon her are called a ν⋯μος, just as the claims of the blood-family upon her are called a ν⋯μος; but this doesn't make the two ν⋯μοι mutually contradictory: each ν⋯μος, each ‘reason’, enunciates the essence of a different thematic aspect of the play. In short, what 450ff. does for the first part of the play, 904–20 does for the second, and so is equally valid and necessary.