Roughly three centuries lie between our youngest ancient manuscript and oldest medieval manuscript of Sophocles:Footnote ¹ Laurentianus 32.9 (abbreviated as L), from the middle of the tenth century, which also contains our earliest extensive text of the surviving plays of Aeschylus, and of Apollonius of Rhodes's Argonautica. This manuscript provides several unique readings likely to be true, and overall the best text of the dramas. However, it is not the case that all other medieval manuscripts were descended from L, and can thus be set aside by editors. Many manuscripts, or manuscript groups, are unique witnesses to parts of the tradition. A manuscript contemporary with L, Leiden B.P.G. (Λ), survives only as a palimpsest – it was reused by having another text written on top of the Sophoclean text, which can now be read only partially. Its readings are especially useful where L's are obscure or illegible. The next-oldest, Laurentianus 31.10 (K), dates from the last third of the twelfth century; it contains some unique correct readings and in general a good text, usually close to that of L(Λ).Footnote ²

Relying on the three oldest complete manuscripts of Sophocles would still leave us frequently adopting an unsatisfactory text. For the triad, the next most important are a group of five manuscripts dating from the fourteenth to the early fifteenth century, collectively known as a. A further group of three, known as r, can be identified, though many of their readings look like later attempts at correction rather than genuine preservations. Then there are at least a score of other manuscripts, some as late as the fifteenth century, which contribute several good readings. For the other four plays, which are represented in far fewer manuscripts, a different, smaller range contains most of the key information, but here too we cannot simply rely on L(Λ)K. This reflects the complex process whereby these seven plays made it from antiquity into the Middle Ages. We are not dealing with the survival of a single ancient manuscript from which all our medieval manuscripts descend; rather, these plays were preserved by multiple routes.

Some manuscripts reflect the activity of Byzantine scholars, in particular Maximus Planudes (c. 1250/5–c. 1305), Manuel Moschopoulos (flourished 1305/6), Thomas Magister (c. 1280–after 1347/8), and Demetrius Triclinius (flourished 1308–1325/30).Footnote ³ The most important of these was Triclinius, who made several changes accepted today as restorations of Sophocles’ original words. He could do this because of his rediscovery of the principle of responsion – of how most tragic lyrics are organized around pairs of stanzas with the same metrical shape. Significant deviations from the metrical shape within a pair of stanzas therefore provide evidence for textual corruption. Triclinius’ solutions to individual problems are not always to be accepted, and our understanding of metre has improved dramatically since his day – yet no-one in this field has ever made a discovery as important as his.

It is easy to forget how tenuous the transmission of Sophocles was during the medieval period. His plays were no longer read in the West, such Latin translations as had been completed in antiquity having disappeared; his seven surviving plays were read only in the Greek-speaking Byzantine Empire. They were not translated into Arabic or Syriac, as other classical texts were; the sudden collapse of the Byzantine state, or a reform of its teaching methods involving the removal of poetic texts from the syllabus, could easily have led to their total loss. As a result, our debt to the profound conservativism of that empire, and its veneration of its Hellenic heritage, is truly incalculable. Not until the fifteenth century do manuscripts of Sophocles reach libraries in the West. The first arrived in 1413, brought by the humanist Giovanni Aurispa.Footnote ⁴ The manuscript L came ten years later, from Constantinople, in the hands of the same man. One of these manuscripts would form the basis for the first printed edition of Sophocles less than a century later.

The coming of the printing press was of decisive importance for the preservation of the seven plays of Sophocles which had survived the vicissitudes of more than eighteen centuries since the playwright's death. Ancient Greek authors did not receive the benefit of this new technology as quickly as Latin ones, and Sophocles had to wait until 1502 before the first printed edition of the seven plays, produced by the Venetian scholar-printer Aldus Manutius. That edition made no scholarly contribution to the correction of the errors caused by centuries of manual copying (though see below, pp. 23–4, for one change that it did introduce which scholars argue over even today). Rather, the Aldine gave Sophocles’ surviving plays a far wider circulation. The number of Sophoclean texts could now be counted, not in the dozens, but in the hundreds; and people interested in what he wrote could buy a copy of his writings rather than having to track down and consult a manuscript.

More printed editions followed. A decisive point came in 1552, when Adrianus Turnebus published in Paris his edition of the seven plays based on a manuscript which had belonged to Triclinius. As we have seen (p. 18), Triclinius was a great scholar, and his work on the text includes contributions of permanent value. But no scholar, however eminent, can be relied on to such an extent that his text can be regarded as the default option, to be rejected only when it can positively be shown to be false. Yet that is what now happened with Triclinius’ recension, giving it an authority which it could not deserve. The idea of assembling a variety of manuscripts, examining the interrelations between them, and working out which reading (if any) was most likely to be what Sophocles actually wrote – all that lay in the future. Nor was there any apparatus criticus at the bottom of the page to inform readers when they were encountering conjectures made by Triclinius rather than text from the manuscripts of Sophocles.Footnote ⁵

Some of the best scholarly criticism on the text of Sophocles was never published, such as the writings of Ludovicus Valckenaer (1715–85) and his Dutch contemporaries in the eighteenth century, or those of the reclusive English scholar Jeremiah Markland (1693–1776).Footnote ⁶ The great contribution during this period which did appear in print, in 1786, was the monumental Sophoclean edition (of the complete plays and the fragments) by R. F. P. Brunck (1729–1803). He based his edition not on the Triclinian tradition which had held sway for over two centuries, but on other manuscripts which he collated himself, a shift which meant that Triclinius’s conjectures were no longer accorded special importance. Brunck's edition won him a pension from the king of France – perhaps not at the most auspicious moment to receive such an accolade.

In the nineteenth century Peter Elmsley's edition of the complete plays was published, in 1803 or 1804, but swiftly retracted by its editor, and only one copy is known to survive;Footnote ⁷ subsequent editions with commentaries of Oedipus the King and Oedipus at Colonus did remain in public view, and were particularly important for our grammatical understanding of Sophocles’ language. From now on the number of editions increases dramatically. Some of the most important were those of Gottfried Hermann (1772–1848), whose understanding of metre was unsurpassed; of Friedrich Schneidewin (1810–56), who was the first to accompany his text with a commentary explaining not just textual points, but matters of literary criticism and staging even when these were not tied to establishment of the text; and of Lewis Campbell (1830–1908) and Richard Jebb (1841–1905), who reacted against the excesses of textual criticism, seeking, often successfully, to provide a more nuanced understanding of Sophocles’ language (pp. 60–1 below).

The Oxford Classical Text by A. C. Pearson (1924) was long generally used as the standard edition of Sophocles;Footnote ⁸ it grew out of date over the next half-century as papyri were discovered (see below, pp. 21–2) and our understanding of language and metre increased. The two most recent complete editions are those by Roger Dawe in the Teubner series, and by Hugh Lloyd-Jones and Nigel Wilson in the new Oxford Classical Text, replacing Pearson.Footnote ⁹ Dawe's edition was a by-product of his three-volume monograph Studies on the Text of Sophocles, the first major effort since Campbell to assess a large group of manuscripts.Footnote ¹⁰ No slave to the manuscripts which he collated, Dawe happily adopted many rather bold conjectures; moreover, he never explains why he cites the particular manuscripts that he does. As a result, his edition fell short.Footnote ¹¹

The Oxford Classical Text that followed Dawe's edition made virtually no advance in terms of the manuscripts, but did revert many questionable textual choices; in all, it differed from the Teubner in more than a thousand places.Footnote ¹² Nevertheless, it contained questionable decisions of its own, not least a fondness for Lloyd-Jones's conjectures; the ‘pigeon-holing of the Teubner as “radical” and the OCT as “conservative” may therefore be misleading’.Footnote ¹³ The OCT text formed the basis for Lloyd-Jones's two-volume Loeb edition, which also includes a volume dedicated to the fragments.Footnote ¹⁴ Since the OCT and Teubner, editions of individual plays have continued to appear: in the Cambridge Greek and Latin Classics series,Footnote ¹⁵ the De Gruyter series ‘Griechische Dramen’,Footnote ¹⁶ and the Mondadori series ‘Sofocle: tragedie e frammenti’.Footnote ¹⁷

The most important development during this period was the discovery and publication of ancient manuscripts from Egypt containing fragments of the lost plays. This began in 1897, with the appearance in print of a papyrus of Sophocles’ Niobe;Footnote ¹⁸ a further fragment of the same papyrus was published nine years later.Footnote ¹⁹ Then in 1912 papyri of Eurypylus Footnote ²⁰ and Ichneutae appeared;Footnote ²¹ and over subsequent decades there followed papyri of further lost plays: Scyrians,Footnote ²² two of Inachus,Footnote ²³ an unknown play,Footnote ²⁴ The Prophets,Footnote ²⁵ a second papyrus of Niobe,Footnote ²⁶ Ajax the Locrian,Footnote ²⁷ Epigoni,Footnote ²⁸ and Tereus.Footnote ²⁹ (Another papyrus containing ‘the Achilles of Sophocles’ should be ascribed to Sophocles’ homonymous grandson.Footnote ³⁰) Of these, Ichneutae is by far the most extensive, amounting to fragments of hundreds of lines; but Eurypylus, Inachus, Niobe, and Tereus too are substantial texts offering significant insights. So although Sophocles has not been favoured as much as Euripides by the sands of Egypt, our knowledge of his poetry has substantially increased; and for as long as Greek literature continues to be studied, there will always be the prospect of more.

Papyrological discoveries have included paraliterary texts (fragments of summaries, commentaries, and the like) which shed light on other plays: Thyestes,Footnote ³¹ perhaps Tereus (a hypothesis),Footnote ³² Niobe (another hypothesis).Footnote ³³ And about twenty ancient manuscripts of plays that survived into the medieval transmission, although lacking the celebrity of new discoveries, constitute the earliest evidence for the text of Sophocles. They offer several unique readings likely to be true, suggesting (as we have already inferred from Athenaeus, pp. 13–14 above) that the Sophoclean text circulating in the early Roman Empire was closer than the medieval manuscripts to the text actually written by Sophocles.Footnote ³⁴ They also offer vital evidence for the survival of Sophocles’ plays in antiquity (pp. 14–16 above).

The last person to edit all the works of Sophocles then known was A. C. Pearson; his commentary on the fragments remains the most recent one on the then entire corpus.Footnote ³⁵ A new collection of the fragments, badly needed thanks to the accretions of papyri, was supplied by Stefan Radt (1977, updated 1999) as part of Tragicorum Graecorum Fragmenta, one of the great monuments of recent classical scholarship. Papyri published since 1999 (Epigoni, Tereus) mean that even this edition is no longer fully up to date. Several fragmentary plays, such as Tereus and Polyidus, have received dedicated commentaries.Footnote ³⁶

The ancient scholia for Trachiniae, Electra, and Oedipus at Colonus have received a modern edition by Georgios Xenis; there is also one by Georgios Christodoulou for Ajax, and (unpublished) by Timothy Janz for Philoctetes.Footnote ³⁷ For Antigone and Oedipus the King we must still use Papageorgius’s edition, which relied solely on L, ignoring other useful material.Footnote ³⁸ The Byzantine scholia are even less well served. There is an edition for Oedipus the King by Oddone Longo,Footnote ³⁹ as well as a forthcoming one for Electra by Andrea Cuomo; but for the other plays the most recent resource is Dindorf's edition of 1852, a book inadequate to modern needs.

***

One might have thought that, after centuries of editing, consensus had now been reached on the text of Sophocles. On the contrary, profound disagreements remain; as a result, for anyone concerned with the literary and dramatic appreciation of Sophocles’ works, the choice of edition can matter a great deal. In the rest of this chapter, I look at several passages printed differently in different editions, where the choice of which text to read or produce has significant consequences for the interpretation of a play.

Anyone reading this passage from Jebb's commentary or Pearson's Oxford Classical Text would find line 572 attributed to Antigone.Footnote ⁴⁰ The manuscripts give the line to Ismene, but since Sophocles will have written merely a horizontal stroke to indicate speaker change, such information has no authority; the attribution, added by someone long after Sophocles’ death, may be right or wrong, but there is no initial presumption in its favour.Footnote ⁴¹ The suggestion that the line properly belongs to Antigone goes back to the Aldine edition of 1502 (pp. 18–19 above). According to Jebb,

But the fact that this would be a solitary reference should raise suspicions. Greek tragedy would not usually make such a sudden, glancing revelation of a character's innermost feelings, especially without ever returning to the point; from a formal point of view, too, an attribution to Antigone is problematic, since it would suddenly and briefly introduce a three-cornered dialogue into a play that lacks such dialogues elsewhere. Rather, the speaker of line 572 should be identical with the speaker of 574, since both lines, though different in expression, focus on the same point: the wrong done to Haemon through his father's action. Giving the line to Antigone provides a beautifully romantic insight into a passionate longing which elsewhere in the play receives no expression whatsoever; giving it to Ismene ensures that Antigone remains silent and uncommunicative as her sister pleads on her behalf, refusing even to acknowledge her betrothed in her determination to do what piety demands for her unburied brother. It is the latter which fits the presentation of Antigone within the drama as well as the norms of tragic usage; someone relying on Jebb's text here would end up with a striking but distinctly un-Sophoclean picture of Antigone and her relationship with Haemon.

This is not the only place where different characters speak different lines depending on what edition one happens to pick up. The following passage from Oedipus the King, in which Oedipus presses home his accusation that Jocasta's brother Creon is attempting to dethrone him, reads like this in the manuscripts:

A clear break in the sense between 623 and 624 indicates a lacuna – i.e. that at least one line has dropped out. Moreover, the question in 625 suits Creon, not Oedipus: Creon is begging Oedipus to believe him and to yield from his purpose (as do the chorus and Jocasta: cf. 646, 651, 673). And 624 suits Oedipus, who is obsessed with envy and the threat that it apparently poses to him (380–6), not Creon. Having reassigned 625 to Creon, we must posit a lacuna between 625 and 626, since Oedipus must have said something between Creon's obviously separate utterances in 625 and 626. That leaves 624–5 with lacunae fore and aft; whether a further lacuna in this evidently mutilated passage stood between 624 and 625 we cannot tell. This leaves us with:

Yet the edition of Sophocles most in use, the Oxford Classical Text, gives 624 to Creon, and 625 to Oedipus, as if Oedipus were pleading with Creon rather than the other way around. A reader of that, or any, edition must constantly be on their guard for wayward decisions that can give a shaping to a passage which is untrue to Sophocles.

The decision over the printing of a single word can have considerable implications for the meaning of a whole play:

The Messenger's reply to the chorus's question in 1583–4 contains difficult Greek; what does τὸν ἀεὶ βίοτον (literally ‘the always life’, or even ‘eternal life’) actually mean? Perhaps λελοιπότα (‘has left, departed from’) is a scribal error for λελογχότα (‘has obtained as his portion’), first suggested by Zachariah Mudge in the eighteenth century and adopted by many editors since. The emendation posits a small change, of just two letters; and since ‘leaving’ life is a more familiar concept than ‘obtaining’ it as one's portion, a scribe may have ‘corrected’ the unusual language to something more readily comprehensible. But the editorial choice could scarcely be more significant for the audience. The dead Oedipus retains power to help friends and harm enemies; but does he also enjoy eternal life, a profound reward from the gods in return for his suffering? The decision is bound up with the interpretation of the whole play, and either option could be right. Editors have to choose, but their choice should be the beginning of debate among readers, not its end.Footnote ⁴²

An even smaller textual point, in terms of letters changed if not in overall significance, from a choral song in Antigone is important for our appreciation of Sophoclean imagery:

‘Bloody dust’ predicated of ‘mowing’ has caused consternation, and J. J. Reiske suggested changing κόνις (‘dust’) to κοπίς (‘chopper’), remarking ‘As if dust could mow! We would put up with that in Euripides; but such trivialities do not suit Sophocles.’Footnote ⁴³ The tiny change has proved popular among editors, but the manuscript reading is probably right. ‘Bloody dust’ is a vivid image, the dust being literally bloody because it was cast on a putrefying corpse, and metaphorically bloody because it will cause the death of the corpse's sister. Reiske objects to its predication with ‘mow’, but ‘it is by no means clear that mixed metaphor should be grounds for emendation’, as Douglas Cairns has recently noted; he teases out allusions in the phrase ‘bloody dust’ to earlier texts, as well as analysing the intricacies of Sophocles’ imagery through the passage.Footnote ⁴⁴ All this is lost to readers using many of the standard texts, however – unless they are keeping an eye on the critical apparatus at the foot of the page and are prepared to challenge the decisions of the editor.

Papyrological discoveries continue to reveal the shaky basis of our critical editions. The latest papyrus coincides with the following passage:

Thus the medieval manuscripts; but an ancient manuscript published in 2017 gives οὗ δή, τέκνον, σὺ ποικίλως (‘from whom, my boy, you intricately…’) in line 130. As its editors note, ‘the new reading συ…puts more emphasis on the double game Odysseus wants Neoptolemus to play, while also asserting distance between the latter and Philoctetes, and thus seems to provide a slightly stronger text’; they compare 878–9 τέκνον, | σύ μ᾽ αὐτὸς ἆρον, σύ με κατάστησον, τέκνον, where τέκνον is also followed by σύ (although the passage is far from a perfect match, given that the repetition of σύ there gives a different emphasis).Footnote ⁴⁵ Long ago, F. H. M. Blaydes challenged the medieval manuscript text, suggesting οὗ δή σύ, τέκνον, ποικίλως; the new papyrus suggests that he was on the right lines.

These and other such papyri suggest that many Sophoclean passages would look quite different if we had better and older manuscript evidence. The recognition of such profound textual indeterminacy – the consequence of the centuries-long process of transmission which has conveyed some of Sophocles’ plays down to our own time – needs to be at the heart of serious engagement with Sophocles’ dramas.