The manuscripts of Sophocles’ plays were hand-copied for about a thousand years, and then neglected for much of the next nine hundred, before they were securely put in print at the end of the fifteenth century. It is not surprising that there is a scatter of miscopyings and errors in our few important medieval manuscripts. Nevertheless, the text seems to be largely intact, and it is mostly agreed upon among scholars; the places where a corrupt or disputed reading is serious enough to have a significant impact on a translation like this are fairly few and far between. I have generally followed the Oxford Classical Text (see Select Bibliography), but have felt free to disagree with it; and I have raised what I regard as the most important textual questions in the notes to the lines in question.
The kind of possible corruption of the text that makes the most impact on a translation is post-Sophoclean additions to the text, usually known as ‘interpolations’—so far as they can be detected. Their extent is much disputed, and some scholars believe they are widespread—even as much as one line in every fifty on average—while others insist that there are very few, if any. I am one of those who believe that there are quite a lot, mostly very few lines in length, scattered through our texts. The great majority of these will have been added by actors in the fourth century, in order to elaborate their parts. This is not at all an implausible scenario because the plays went through a century and more of being widely performed professionally throughout the Greek world. And we know that the actors changed and adapted the texts because the Athenians eventually attempted to establish an ‘authorized version’.
The problem is that detecting these interpolations inevitably involves a high degree of subjective judgement. The main justification is bound to be the detection of alleged problems or anomalies in the text as it is transmitted. But this method tends to imply that everything that is authentic Sophocles is perfect, and that everything that has been added is somehow intrusive or incompetent. This is obviously a highly dubious business, but there is, nonetheless, no avoiding the challenge. I have attempted to make up my own mind on such issues, and wherever I think there is a serious possibility of interpolation, I have raised that in the notes to the passage in question.1 Where I believe that interpolation is highly probable (certainty is impossible), I have put square brackets (i.e. [. . .]) around the lines in question. I would encourage readers simply to jump over those lines.
Omissions of bits of Sophocles from the transmitted text (as opposed to additions) are far less widespread; editors have tended to be too ready to suppose that a line or lines have been lost merely in order to account for some minor problem or other. It remains true, however, that there are a few places where this is the most likely explanation of something that has evidently gone wrong with the text as handed down to us. In those places where I take this to be the best solution, I have added some words of my own to give the gist of the missing bit, and have put them between angled brackets (i.e. <. . .>). There are also occasionally places where some lines have very probably become displaced and need to be transposed back to where they originally belonged.
In addition to such textual issues, there are some recurrent problems which have to be faced by any translator, whatever the nature and style of the version. A particularly prominent one is the spelling of proper names. It might seem at first glance that the names should simply be transliterated from the Greek, since this would be closest to the original. But in practice this is over-alienating, because many of the names have become familiar in adapted spellings. This is most obvious with place-names, such as Thebes for Thebai, Troy for Troia, etc. English has also traditionally used Latin spellings for other proper names. In many cases it would be easily acceptable if these were returned to their Greek spellings, e.g. Elektra, Haimon, Olympos, etc. The trouble is that there are others which would seem obtrusively strange, and some that would strike most readers as outlandish—some blatant examples are Oidipous, Klytaimestra, and Mykenai, long familiar as Oedipus, Clytemnestra, and Mycene. So I have (with some reluctance) taken the easy way out and used the standard Latin spellings in nearly all cases.
There are some other relatively minor recurrent features which pose particular problems for translation. One is insults and abuse: it is very hard to find equivalents across languages, which carry the same kind of tone and level of aggression without being too crude or too stilted. If only English still used ‘villain’! Another is interjections, i.e. places where characters utter sounds that are non-verbal: cries of pain or grief, and also of surprise or joy. Modern English is largely lacking in conventional interjections of grief (‘alas’ and ‘ai me’ have become archaic); and the most common interjections in Greek—such as oimoi, papai, aiai, iou, e e—do not generally have phonetic matches in English. Some translators resort to the solution of simply transliterating, but to my ear the resulting sounds are over-intrusive. In many places I have simply attempted to use one of the interjection-like phrases still current in English today, such as ‘ah’, ‘oh no!’; in others I have thought it better to replace the Greek interjection with a stage-direction, such as ‘(cries of pain)’.
Finally there is the whole issue of what textual aids the translator should add to help the modern reader. The most prominent is stage-directions. I have followed the usual convention of including them in italics. It would be positively unfriendly to the reader not to do so; also it would dampen the awareness that this is a text made for performance. At the same time it needs to be stated clearly that our Greek text has no explicit stage-directions as such; in other words, all those included in this version are my own editorial additions. These are not proposals for a modern performance: all I have tried to do is to work out the likely stagings of Sophocles’ own first performance, and have envisaged them as set within the general space of the ancient Greek theatre, as briefly laid out on pp. xxi–xxii above.
I have applied the basic working assumption that when stage-action is important in a Greek tragedy it is clearly implied and signalled in the text. This thesis (which I argued for in Greek Tragedy in Action and elsewhere) is not, however, above challenge. And even assuming it is justified, modern productions do not—and should not—regard themselves as bound by it. This means that I have limited printed stage-directions to those which I feel are pretty securely justified.2 It should not need saying that stage-directions can have a very significant bearing on interpretation, and I have discussed these, especially those that are open to serious dispute, in the notes.
I have broken with the usual editorial convention by adding scene-headings and numbers (e.g. ‘Scene 4’), and similar indications for choral songs and lyric dialogues (e.g. ‘Second Choral Song’). Most modern editions and translations have no scene-headings at all; some, however, follow an arcane analysis and numbering which obscure more than they help.3 The scene-divisions added here may, I hope, help readers and performers to have a clearer idea of the play’s construction. I should emphasize, though, that they are the product of my own structural analysis, and have no higher authority.
Finally, the marginal line-numbers are another editorial addition. It should also be made clear that the numeration is that of the Greek text, as is uniform in modern editions. This is not by any means consistently matched by the number of the lines in this translation.