Richard H. Armstrong
Pro captu lectoris habent sua fata libelli.
“According to the reader’s grasp, books have their destiny.”
—(Terentianus Maurus, De litteris: 1286)
If you open a Greek text today, say a work in the Oxford Classical Text or Teubner series, you are the beneficiary of centuries of effort that have made this text appear before you, imparting a sense of immediate contact with the ancient world. However, there is no unmediated access to this ancient text or the world that produced it, and this awareness is the first step in developing a philological consciousness.1 Countless people have worked hard over centuries to give you this sense of access, i.e. that you have “Homer” or “Sophocles” in your hands. While the emotion of communing directly with antiquity may be sublime, it is in essence an illusion. Grasping this is important, because every reader of an ancient text both benefits and suffers from decisions made by a great many other people. Copyists and editors have decided to prefer one version of the text over another – or whether the text will come down to us at all. Outright mistakes were sometimes made by copyists that have led to permanent glitches down the line, which may have been magnified when someone else tried to fix the problem with very little to go on. Occasionally, marginal comments or interlinear glosses have been sucked into the main text and become a permanent addition. In some instances, there might even have been active interventions to make the text say something very different from what it did originally. The Testimonium Flavianum is an extreme but instructive case. This is a short passage in Josephus’ Jewish Antiquities (18.3.3), which contains the only reference to Jesus independent of early Christian writings. Why would its author, a Judean Pharisee of Roman-leaning sympathies, have written positively about Jesus, a Galilean troublemaker, to the extent of calling him the Messiah (Khristós) and attesting to his miracles and resurrection? We can easily imagine that Josephus wrote the opposite of this, and that Christian scribes decided to “convert” the text into a friendly witness. But we are stuck with the Greek text as it stands.2
So we begin with a simple truth: texts are not organisms and do not reproduce themselves. Their physical survival through centuries of political upheavals, ideological and linguistic changes, literary fashions, economic catastrophes, floods, fires – not to mention mice and bookworms – is far from guaranteed. Before the age of printing, their survival was tied to a) being valued by someone who would keep them from harm and b) being manually copied by someone educated enough to make appropriate sense and use of them. Generally when we speak of the “survival” of texts as distinct from their transmission, we are talking about their persistence as physical objects. It is rare for complete manuscripts (mss hereafter) to survive intact from earlier than the tenth century CE, though there are exceptions. There are three substantial mss of Vergil from the fifth/sixth centuries CE and fragments of four others, including one as old as the late fourth century CE (Reynolds 1983, 433–4). There are four great uncial mss of the Greek version of the Bible (Septuagint + New Testament) from the fourth and fifth centuries CE (in contrast, the earliest complete ms of the Hebrew Bible, the Codex Leningradensis, dates from 1008 CE). We have over half of an impressive New Testament in the Gothic language, written in gold and silver ink on purple-dyed vellum for Theoderic the Great in the sixth century CE. There are remnants of an Iliad ms of the late fifth to early sixth century CE, mostly comprising illustrations cut out from the original book (Bianchi-Bandinelli 1955). Such early mss are more the exception than the rule, and are notably of the most valued works from antiquity: Vergil, Homer, and the Bible. Far less care was extended to problematic, pedantic, and polemical works or refuted heresies; and we know of many major works that vanished for no good reason.
Aside from papyrus fragments, most of the textual evidence for ancient works comes from mss later than the tenth century CE. That evidence is as distant from Periclean Athens as our age is from the time of Mohammed. Which is to say, antiquity survives largely as an endeavor of medieval civilization – a fact not all classicists readily appreciate, since the general notion of the Classical archive is tied to modern print editions (more on this issue below). A relatively small percentage of Classical scholars today have actually worked directly with a medieval ms or a papyrus fragment. While the age of printing greatly democratized access to ancient texts, it also obscured their material origins by causing scholars to imagine the Classical text as an entity floating somehow above its extant manuscripts. It is therefore quite instructive to take a moment to consider how our knowledge of antiquity is heavily reliant upon the vagaries of material survival. So as a first mental exercise, think this: antiquity is a wound, not a world you can visit. We need to be truly open to feeling the pain of our losses.
The best way to discuss ancient literature’s survival is to contemplate the dark side of the moon first: i.e., what we know for certain did not survive, and how this affects our understanding of ancient culture. Perhaps the most dramatic example (no pun intended) is ancient drama, since we can quantify lost plays as units analogous to existing ones, which produces a simple ratio. We have seven of Aeschylus’ estimated 70 to 90 plays, and short quotations and fragments of others; seven of Sophocles’ 120-plus plays, again with sound bites and bits of others, including extensive fragments of the satyr play The Trackers (only made public in 1912); and 18 (19 with the disputed Rhesus) of Euripides’ estimated 90, with two volumes of intriguing shards ( TrGF 5.1–2). The survival rates of complete works for these pillars of Western drama are then roughly: 10 per cent to 7 per cent (Aeschylus), 5.8 per cent (Sophocles), and 20 per cent (Euripides). Compare this to Shakespeare: we have 38 plays in the great folio edition, with perhaps 2 lost; then there are a couple of plays extant of which he was perhaps partial author. Had the Bard’s works suffered at the Sophoclean survival rate, we would have instead something like two complete works and one-third of another. Just think of any two Shakespeare plays and imagine trying to create a picture of the Elizabethan stage out of, say, Titus Andronicus and Pericles, Prince of Tyre. We would like to think what remains of the great tragic trinity is their best work, but that is a problematic assumption. We would certainly understand a lot more about the development of myth into drama if we could read Aeschylus’ Laius, Oedipus and Sphinx (the missing parts from the tetralogy that included Seven Against Thebes) or Sophocles’ Phaedra or Palamedes (TrGF 4, F 677–93 and F 478–81). We would get more of Aristophanes’ jokes if we could read Euripides’ Telephus or Andromeda (TrGF 5.1, 10 and 5.2, 67), and we could understand better that tragedian’s capacity for scandal had his Aeolus (a play about brother–sister incest – TrGF 5.1, 2) or Chrysippus (concerning Laius’ pederastic passion for the bastard son of Pelops – TrGF 5.2, 78) come down to us – though the scandal perhaps explains their absence. In sum, more evidence for these authors would doubtless enhance and nuance our understanding of Athenian drama considerably.
What of the other great dramatists of the Classical age? We would certainly like to read more of early tragedy to understand the evolution of the genre, but we have none of Choerilus’ alleged 160 plays or of Pratinas’ 50, nor are we sure who voted them off the island of posterity. I would love a look at Phrynichus’ Fall of Miletus, which so upset the Athenians they fined him and outlawed its further production – setting a lasting precedent against tragedies written from contemporary events (Herodotus, Histories 6.21). Agathon was a significant enough poet to be portrayed in works by both Plato (Protagoras and Symposium) and Aristophanes (Thesmophoriazusae), and later gossip claimed Euripides wrote the Chrysippus out of infatuation for him (Aelian, Var. Hist. 2.21); but of his work we have less than 50 lines. While we get by on an appallingly low percentage of known tragedies, we had up to 1912 a single satyr play – Euripides’ Cyclops – to represent that whole genre, one that was a part of every tragedian’s output at the dramatic festivals (the normal entry was three tragedies and one satyr play). Now we have a fair amount of one by Sophocles, and much less from two by Aeschylus; this is still far from enough to get a feel for the genre and its relation to comedy and tragedy. Similarly scanty are the remains of the dithyramb, a choral genre that was a major part of several Athenian festivals, with wide participation from among the men and boys of Athens’ tribes. It is largely represented by fragments of Pindar, Bacchylides and Timotheus, in spite of centuries of competitive output. Some 20 dithyrambic compositions a year were produced for the City Dionysia alone, not to mention the other festivals in Athens and beyond (Hordern 2002, 22). All of Old Comedy is represented to us by 11 plays of Aristophanes (we know of an additional 32 titles for him); perhaps he was the best choice, but how do we know, really? He falls rather late in the chronology of Old Comedy (he was born between 460 and 450 BCE, while comedy was first entered in the competitions in 488 or 487 BCE), and we have precious little from early comic poets like Chionides and Magnes (Mensching 1964). We could learn much about both comedy and contemporary Athenian politics if we also had at least as many comedies by Eupolis and Cratinus, who along with Aristophanes made up a kind of comic trinity for Horace (Sat. 1.4.1). Most of Middle Comedy is in shreds, though we know the names of 50 working poets and Athenaeus attributes more than 800 plays to the period. For New Comedy, we know of nearly 80 playwrights active between 325 and 200 BCE, and 50 working beyond then.
Things are not better if we think of other genres and disciplines. A look at the many names and fragments in Felix Jacoby’s Die Fragmente der griechischen Historiker can be quite sobering. Hecataeus of Miletus (fl. 500 BCE) was the most important early prose writer in Ionia, but we have only fragments of his writings on geography and mythography. The mythographer and chronicler Hellanicus of Lesbos (c. 480–395 BCE) put a lot of ancient myth and history into an order that was extremely influential in antiquity, but we have little directly from his works.3 Fifty-three of the 58 books of Theopompus’ swollen history of the reign of Philip II of Macedon, published after 324 BCE, survived intact into the ninth century CE; but now it lives in mere snippets in the works of others. We have to write the whole history of Alexander the Great without direct access to any of the contemporary sources: the court history of Callisthenes, the memoirs of Ptolemy, Aristobulus, Nearchus and Onesicritus, or the highly popular and entertaining history of Cleitarchus, full of stories, it seems, culled from surviving veterans. We lack the well-informed and informative histories of the statesman Hieronymus of Cardia (late fourth/early third century BCE), which were an authoritative source for other ancient historians on the period of Alexander’s successors in the Hellenistic kingdoms. Nor can we thrill to the gripping but missing Histories of Phylarchus (third century BCE), Hieronymus’ successor, which Polybius felt came too close to tragedy (Polybius, Histories 2.56.6–13). Timaeus of Tauromenium (c. 350–260 BCE) was the most important historian of the western Greek world, who wrote about the colonial history of Magna Graecia, the early Romans, and the Carthaginians; his work lives on in 164 fragments and the background radiation he still projects through Diodorus Siculus’ first-century BCE Bibliothēkē and the critical animus of Polybius.4 Similarly, Ephorus of Cyme (c. 405–330 BCE) wrote a 30-book magnum opus that Polybius deemed the first universal history and was widely quoted; but we have to get at it mostly by viewing it as an éminence grise behind Diodorus. Of Diodorus’ own Bibliothēkē —which is the most extensively preserved history by a Greek author—we have only 15 of the original 40 books. Historians would doubtless pay a king’s ransom for the lost works of Polybius (c. 200–c. 118 BCE), one of the most articulate voices in ancient historiography. Only five of his 40 books of the Histories survive intact, with abridgements and excerpts making up the rest of his literary remains. In sum, though harder to calculate, our losses in historiography are clearly vast; Hermann Strasburger ventured a figure of 2.5 per cent for the total of historical works from Classical and Hellenistic antiquity to survive into modern times, or a ratio of 1 : 40 (1977 / 1990, 180–81).
Moving now to philosophy, most of what we know about the pre-Socratics comes from quotations by authors who are trying to refute them. We can be grateful the later philosophers were kind enough to quote their predecessors to some extent, but the bulk of pre-Platonic speculation can be fitted out in a single volume, if one were to dispense with all the apparatus necessary to put the fragments into context.5 Thus if you want to peek into Parmenides’ provocative poem On Nature, you’ll find only 160 of some estimated 3,000 verses (5 per cent). Only slightly better results await you for Empedocles’ On Nature and Purifications; we have 550 lines out of perhaps 5,000 (11 per cent). Democritus of Abdera, the “laughing philosopher,” is credited with 70 book titles by Diogenes Laertius (9.46–9): we have exactly none of these. What we know of the fifth century BCE sophists is largely through Platonic characterization – if not caricature. While we have the flashy Encomium of Helen by Gorgias, those wishing to read his On Nature or the Non-Existent will have to do with tendentious paraphrases of a much later date. For Protagoras, we don’t even have that much, though probably not because of a rampant Athenian book-burning caused by his frank agnosticism, as a later story has it (Diog. Laer. 9.52). Sheer quantity of production was no guarantee of survival to judge from cases like the Stoic philosophers Chrysippus, who reportedly wrote 705 books, routinely cranking out 500 lines a day (Diog. Laer. 7.180–81), and Posidonius, whose work reached far beyond the conventional topics in philosophy to matters ranging from anthropology and geography to seismology and zoology. We know these two thinkers were widely influential, but scholars have to reconstruct their works with considerable effort and much disagreement. Who wouldn’t want to read Chrysippus’ outrageous interpretation of Zeus and Hera in his lost work On the Ancient Natural Philosophers, which used language so indecent it was considered unmentionable by ancient bibliographers (Diog. Laer. 7.187–8)?
Even for well-represented philosophers like Aristotle, there are key parts missing in the extant corpus. Political historians lament the loss of his works on the constitutions of 158 states; though the recovered Athenian Constitution may repair that loss for Athens, we clearly have but a tiny fraction of the whole. His On the Forms torments us in its absence, since it is a critical discussion of Plato’s theories that could reveal a lot about the development of Aristotle’s later philosophy.6 The Protrepticus was widely influential in antiquity, moving many to devote themselves to philosophy; how is it this work does not survive, while we still have the pseudo-Aristotelian Physiognomics? Cicero’s dialogue modeled after the Protrepticus, the also lost Hortensius, still had the power to turn the young Augustine into a budding philosopher four centuries after it was written (Conf. 3.4.7). In fact, we generally lack textual evidence of the elegant Aristotelian dialogues (of which there were something like 18) that Cicero claimed were like a “golden river of discourse” (flumen orationis aureum, Acad. pr. 38.119); mostly in the Aristotelian corpus we have revised lecture notes, the style and organization of which leave much to be desired at times. While we do have part of his Poetics – arguably one of the most influential aesthetic texts from antiquity and of all time – there appears to be a significant portion missing that treats of comedy, and perhaps discloses the secret of that mysterious term catharsis, about which we are too free to speculate in that book’s absence.7
Our canonical view of Greek literature is the product of such enormous absences. We read Homer with little else to compare it to, since so much of the epic cycle is missing, as is the Hesiodic Catalogue of Women. We concentrate heavily on 44 extant plays, but we are unable to reconstruct the complete line-up for a single year of the City Dionysia. We accept Plato’s caricatures of the sophists as valid (hence the pejorative meaning of “sophistry” and “sophistic”), and may assume all philosophy is a footnote to him because what came before is so scanty. Our entire view of Greek culture is decidedly skewed toward Athens. We don’t get to read the 50 odd plays of Theodectes of Lycia, whose Lynceus and Tydeus were admired by Aristotle (Poet. 1452a27, 1455a9, b29), or the comedies of Epicharmus, who wrote in Sicilian Doric (PCG 1, 9–137). We don’t get to read five whole books of Tyrtaeus of Sparta or nine of Sappho of Lesbos, or five of Corinna of Tanagra, as the ancients reportedly did. We sadly lack most of the mimes of Sophron of Syracuse, which, we are told, Plato first brought to Athens and actively imitated in his dialogues, loving them so much he kept a copy under his pillow (Diog. Laer. 3.18). Whether we believe such stories or not, the mimes are a loss (remains in PCG 1, 187–253). We would gain a good deal of information on Athens’ famous rival had the works of Sosibius of Sparta (fl. mid-third century BCE) survived. We typically think about the intense production of the poets and scholars of Alexandria only in relation to Roman literature, for which Hellenistic models loomed tremendously large, or in relation to the earlier Greek works they helped to preserve. What remains of the great commentaries of Aristarchus or the alleged 3,500-plus works of Didymus must be gleaned from the scholia in the margins of medieval manuscripts of Homer. Imagine what we would at least know about our losses had Callimachus’ vast bibliographical work, the Pinakes, survived, which detailed the holdings of the library of Alexandria? The work’s full title, Lists of Those Eminent in All Areas of Learning and Their Writings, shows the amplitude of his bibliographical ambitions, and its reported 120 bookrolls attest to the scope of his achievement. The fragmentary remains of 58 entries can only serve to torment us about its loss.
We suffer greatly as well from the loss of Greek texts by non-Greek authors, who adopted the literary koine to relate important matters to the oikoumene, or by Greek authors on non-Greek matters. We have very little of Q. Fabius Pictor, a Roman writing one of the first Roman histories, but in Greek so the civilized world could read it. Lucius Cincius Alimentus also wrote in Greek of his captivity under Hannibal in the Second Punic war – but this eyewitness account is only known indirectly. We have nothing of the works of Silenus of Sicily and Sosylus of Sparta, whose histories of Hannibal’s campaigns were heavily pillaged by Polybius and Livy. Hannibal himself was also author of lost works in Greek, we are told—wouldn’t those make interesting reading? King Juba II of Numidia wrote extensively in Greek and Latin; his guide to Arabia was a best-seller in Rome, while the Athenians erected a monument in honor of his writings—yet we get few glimpses of this fascinating output, though it percolates Pliny’s Natural History (Roller 2003). Christian interest has preserved for us the works of Josephus, a Jewish witness to the Roman wars in Judea in the first century CE, but not the vast output of Nicolaus of Damascus (cf. the fragments collected in FGrHist 90), from whom Josephus drew a great deal. This court historian to Herod the Great and biographer of Augustus would hardly be an impartial source, yet his works would certainly fill out the picture of at least the official view of things. And our knowledge of the ancient world might look very different if we had complete copies of Manetho’s Aigyptiaka and Berossos’ Babyloniaka, Hellenistic works detailing the histories of Egypt and Babylon written by knowledgeable if fallible men; or Ctesias of Cnidus’ Persika, written in opposition to Herodotus’ Histories and still being read by Photius in the ninth century CE, who left us a synopsis (Bibl. 72). Imagine what we could learn from the 20 books in Greek about the Etruscans, penned by the future Emperor Claudius, who was among the very last people to know Etruscan (Suetonius, Divus Claudius 42.2). How different would later antiquity look to us if we could read all of Arrian’s 25-book opus on the Black Sea region, or his full account of the Alans, or his History of the Parthian Wars (17 books), or Priscus’ history treating extensively of the Huns in the fifth century?
These are examples of our known losses; we should also consider the real possibility of unknown losses (our “unknown unknowns,” as Donald Rumsfeld once put it). The texts enumerated here may be just the tip of the iceberg of all that is missing—and I have not even mentioned technical, rhetorical, novelistic, legal, scientific, musical, mathematical, and medical works, or Hellenistic Jewish or early Christian and patristic authors. The contemplation of such deficits is important, as it humbles our attempts to create totalizing reconstructions of antiquity or to effect a mind-meld with an ancient author. It is crucial to distinguish the ontological plenitude known as The Ancient World from the scrappy realities of the ancient archive.
Only because our losses are so staggering are we greatly advanced by whatever scraps of ancient texts happen to come back to light. Thus, while it is healthy to contemplate seriously the state of our ignorance, it is equally important to be aware that there are still surprises to be had. If we look at antiquity as a whole, we can note with amazement the particular finds that have truly revolutionized our knowledge within the past 120 years: the Amarna letters and Hymn to the Aten (Akkadian and Egyptian texts revealing the lost history of the “heretic Pharaoh” Akhenaten); the Ugaritic texts (Canaanite tablets discovered in 1929, which have contributed substantially to understanding the Old Testament in its ancient context); the Dead Sea Scrolls (texts discovered 1946–1956 that have vastly improved our understanding of Second Temple Judaism and the textual tradition of the Hebrew bible); and the Nag Hammadi library (12 Coptic codices discovered in 1945 that revolutionized the study of gnosticism and early Christianity). We should also note that mere physical discovery only begins a long process that has its dramatic later developments. For example, Linear B, though discovered on artifacts in the late nineteenth century, was not actually deciphered and identified as a syllabary script for an early form of Greek until the 1950s. Or to take another famous example, the papyri found at Herculaneum were “discovered” in the 1750s, but their charred condition left most of them unreadable for decades. Various attempts to unroll them since their discovery often yielded disastrous results, until the introduction of new methods in the mid-1960s and 1970s secured better outcomes; but the process of reading these works remains ongoing, as do the site excavations (Sider 2005).
For all the complications in handling and reading papyri, the gains since the late nineteenth century are still considerable. For Greek literature specifically, we have had significant discoveries for Aeschylus, Alcman, Archilochus, Aristotle, Bacchylides, Callimachus, Empedocles, Epicurus, Euripides, Herodas, Hyperides, Isocrates, Menander, Pindar, Sappho, Simonides, Sophocles and Stesichorus – just to list the better known names. Even recovered copies of things we already have, such as Homer’s Iliad, can still have great value for understanding ancient textuality, literacy, translation, schooling and even elements of performance. And archaeology has yielded some things beyond canonical authors that were truly unexpected (i.e., among the unknown unknowns), such as the Derveni papyrus. This dates from c. 340 BCE (or even earlier) and is perhaps the oldest surviving ms from Greek antiquity; it was discovered in 1962 and contains an allegorical commentary on an Orphic poem (Betegh 2004, Laks and Most 1997). Another case of unexpected surprise is the Posidippus roll, a papyrus roll of the third–second century BCE comprising 112 poems that surfaced on the market in 1992, which constitutes our first physical example of a book of poems from the ancient world and contains over 600 lines of previously unknown Hellenistic poetry (Gutzweiler 2005; cf. Höschele, ch. 12, pp. 194–6 in this volume).
There is certainly an understandable sense of triumph in recovering great works of the past; it is no small thing to overcome death and oblivion. The case of the comic poet Menander (344 / 3–292/1 BCE) will doubtless remain a prime example of the thrill of cliff-hanging survival. For centuries, one knew that Menander’s works were widely influential and imitated by Greek and later Roman authors. Over 900 quotations of various lengths preserved in ancient authors showed the esteem in which he was held, yet not a single play survived intact. But in the twentieth century, archaeological excavations recovered a repurposed codex of Menander (the “Cairo Codex”—used as a jar topper) from a house and another (the “Bodmer Codex”) from an abandoned monastery library. Further fragments were recovered from mummy cartonnage, from among the Oxyrhynchus papyri, and, quite recently, from a twice-palimpsested ms in the Vatican (Handley 2011). This means that the casual student of classics today has vastly more Menander to go on than the greatest scholars of the nineteenth century, and is thereby in a position to better understand the post-Classical Greek theater and the Roman stage to boot. Menander now has his own volume in the Penguin Classics series, and can be known widely alongside his canonical colleagues. One should never make light of such material advantages, even if what we possess is still a fraction of the whole (Handley 2011, 146 suggests perhaps 5 per cent).
Apart from such spectacular recoveries of major authors, we should also consider the importance of things we learn from less exalted papyrus finds of non-canonical, and unexpected works. The philosophical library uncovered in a villa at Herculaneum has been unkindly described as one filled with the works of “an obscure, verbose, inauthentic Epicurean from Cicero’s time” (thus Domenico Comparetti, cited in Sider 2005, 63). Yet from Philodemus’ works we learn a good deal about the Epicurean milieu in the Bay of Naples, a background that is quite significant for Roman poets of the generation of Vergil and Horace. As a philosopher–poet, his works reveal attempts made in later Epicureanism to bridge the gap between philosophy and poetry, and some of the poetic theory dealt with in his works clearly has its echoes in the Augustans. Moreover, his philosophical writings teach us many things about some aspects of Epicurus’ thought now lost to us – particularly his conception of the gods (Sider 2005, 84–7). He quotes extensively from Epicurus’ On Nature, a major lost work, portions of which are recoverable from other papyri from the same site. The collection of books taken as a whole provides an interesting picture of the Greco-Roman intellectual life at a crucial moment in Roman culture – and some, like David Sider, hope further excavation may yet reveal still greater treasures, both Greek and Roman (2005, 94–5). It is no longer absurd to hope for such a thing – even if at long odds.
Let us take a closer look at a modest example of what a found text can do to open up our view of antiquity. A papyrus containing a mere 14 lines of text was acquired by the Louvre in 1891, and waited over a century to be published. Annie Bélis first published it in 2004, and M. L. West subsequently published a corrective article on it in 2007 – again we see the long latency between the physical and scholarly phases of “discovery.” Two things make it of immediate interest in spite of its brevity. First, it seems to be a scene from the lost Medea of Carcinus the Younger (fl. 380–377 BCE), a major figure in his day who composed over 160 plays, one of the highest numbers on record for a tragedian, and is credited with 11 victories in tragic competitions. He was a third-generation tragedian; his grandfather Carcinus the Elder was victorious at the Dionysia in 446 BCE, and his father Xenocles beat out Euripides in the Dionysia of 415. Yet again, we wonder why this tragic dynasty seems cut out of posterity, particularly since Aristotle refers to Carcinus the Younger several times, and authors like Lysias, Timaeus, Menander and even Athenaeus (late second century CE) quote him – this last author’s citation implies Carcinus’ works were preserved into the common era. Second, the papyrus has musical notation, precious traces of ancient performance that take the significance of this fragment to a new dimension. From what Bélis and West have reconstructed, this version of the Medea story hinges on an interesting twist: Medea has not killed her children, but believes she has sent them away to safety. Yet because she cannot produce them, she is accused of having killed them. This is a delicious irony (which West further explores in his article), and a pointed divergence from the Euripidean version that effectively defined Medea as a horrific child-murderer for Western literature. As is often the case, we learn a lot about the creative options poets enjoyed when we have more than one version of a tragic plot. But the musical details are also intriguing: the melody recorded is not contemporary to the text, but one later added for a kind of virtuoso solo recital. This is not, in other words, a text of the whole play, but a performance text for a later soloist who, in Imperial times, is bringing out this now classic scene set to new music. So not only does the fragment give us something of intrinsic interest by a reputable post-Classical tragedian, it also gives us a snapshot of the play’s reception and re-performance (see Nervegna 2007). This fragment therefore affords us a view not just of textual survival, but cultural tradition – i.e., proof of what at least some ancients valued in their own literature, and how they repurposed it in a later milieu. Carcinus’ work clearly had an afterlife, even if it never made it through to the Valhalla of Penguin Classics.
As we have just seen, a papyrus can be discovered and still languish in a drawer (or a cigar box) for decades until a scholar makes it public to the world. This simple fact reveals that mere physical survival means little without the conscious intention of someone to transmit this text to a reading public. In the pre-print era, transmission refers to the laborious process of copying out texts by hand; but we have not been in that situation for centuries. We have long since moved from the flux of manuscript culture to the fixity of print, and, save for those texts that come to us directly from the sands of Egypt, such as the Carcinus fragment above, every student of Greek literature should understand the layers of mediation that print culture has created, especially because many of the conventions and limitations of print culture are rapidly changing in the digital age. First, one should know that the advent of printing created a watershed in Western historical consciousness in many ways. The printing of Greek grammars, lexicons, and texts allowed Hellenic culture once again to be a broad possession internationally, this time well north of its homeland in the Mediterranean basin (Geanakoplos 1962; Wilson 1992). However, transforming Greek into an idiom of print culture required the creation of new Greek styles of typography, which gradually left behind the Byzantine manuscript hands and scribal culture (Proctor 1900). A new textual regime also came into play, with standardized accentuation and punctuation, the introduction of line numbers, indexes, and conventions for the layout of poetry and prose. But the precision technology of printing identical copies also brought with it another possibility: the ideal of widely circulating and constantly improved texts, i.e. texts that aspire to be much more than a copy of a copy. The cultural shift from “a sequence of corrupted copies to a sequence of improved editions” is a general one affecting all fields of knowledge in the age of printing, when “large-scale data collection did become subject to new forms of feedback which had not been possible in the age of scribes” (Eisenstein 1993, 76–7). The production of such improved editions of ancient texts is known as textual criticism (not to be confused with literary criticism, though the two overlap), which reached its greatest development as a field in the nineteenth and twentieth centuries. The development of modern textual criticism was a slow process and it had predecessors going back to the great scholars of Alexandria and Pergamon in antiquity. From the earliest days the desideratum has been to make public the “best text” possible; but just what this entails in terms of both methodology and risk has to be understood in some detail.
We should first distinguish between the words copy, recension, and edition in their technical senses. A copy or transcription is a mechanical attempt to make a new iteration of an original, and we have many late examples of handwritten copies made of much older mss; classed as recentiores, these later copies are often assumed to be less interesting than their older originals. A recension, however, is a careful collation of available mss copies (and other witnesses, like passages quoted in other works), showing broadly all the evidence for the text in question, known as the textual tradition. A recension’s success is clearly reliant on the amount of available evidence – but, paradoxically, can be greatly complicated by a large tradition. Inevitably, such a recension will disclose variant readings, or places where the various mss say different things, and sometimes mss can be classified neatly into groups or families based on these differences (often mistakes they hold in common are the best criteria for classification). Ideally – or rather, idealistically – the mss can then be arranged and classified in a genealogical chart known as a stemma, showing which mss are derivative of others (their archetypes), and this stemmatic method would then lead us back to the earliest form of the text (often best represented, one might assume, by the oldest mss) to which we have access. This form of the text might just be late antique or early medieval, and potentially still a good distance from a putative “original” of the fifth century BCE or earlier, depending on the author’s date. Most of us can live with that distance, but there are some scholars with unsettling certainty of mind who are convinced they can get us back to a vision of the autograph ms of the ancient author himself. Fundamentally, the classic aspiration of textual criticism is, in the formulation of Paul Maas, “to produce a text as close as possible to the original” (1958, 1). Upon examination of the textual tradition, “if it proves not to give the original, we must try to reconstruct the original by conjecture (divinatio) or at least try to isolate the corruption” (Maas 1958, 1).
Thus the traditional textual editor may either be seen as a harmless drudge reporting on the many mss he has either seen in microfilm, examined personally, or learned about from trustworthy persons; or as a daring interventionist and diviner, who makes very real judgments about emending or fixing up the text through conjecture. (It is conceivable that he may be both in the course of editing a text over many years.) But to be clear: the textual critic makes use of the recension to produce an edition, which he might feel is the newly restored text of the original author or just the best version possible, but which is always a scholarly construct, a modern by-product of an ancient and medieval textual process. Whether or not the modern edition misrepresents “the text” depends entirely on the judgment of the editor and the scholarly community to which he responds – and it may take generations for the jury to render its verdict fully. But the problem is: the modern critical text – through the authority of its editor and availability in libraries and bookstores – displaces the ancient or medieval mss in importance. If you believe the text has been restored or improved, then you will find this unproblematic. But if you are deeply unhappy about the choices the editor made, you will find it vexatious.
The view of stemmatic editing given above is admittedly oversimplified. But at its most extreme, we can see that the stemmatic approach betrays a certain Platonic penchant: one must get through the detritus of the material many (i.e., the various mss) to a noetic union with the immaterial one (the notional archetype, the Urtext, or the autograph ms representing infallibly the author’s thought and intentions). The truth is, things are far messier in the typical textual tradition, and they get much worse in cases where we have a great many mss instead of a very few (Homer being a particular case in point). For one thing, medieval Byzantine scholars and scribes did not simply copy from one text mechanically and uncritically. Although every instance of copying from ms to ms could introduce errors (due to ignorance, inattention, or problems with the master copy), it was also an opportunity to collate and correct the new copy from other mss (which the stemmatic enthusiast might impatiently denounce as “horizontal contamination” since he wants things to be neatly arranged in a vertical order of ascending authenticity). We know for a fact that the Byzantine Greeks had access to far more mss than we do, so what they add may be something of great value and antiquity, and not “horizontal” tomfoolery at all. Thus against the initial prejudice in favor of the most ancient mss, Byzantinists have raised the cry recentiores non deteriores (more recent copies are not worse ones – Browning 1960). Once one removes the simplistic genealogy of texts from the process, it becomes harder to know what to do with all the discordant readings.
Moreover, the idea of grounding textual criticism on the certitude of an Urtext – stable in meaning, clearly intended by its maker, accessible by hard labor and divination – has been greatly challenged by our evolving sense of ancient textuality’s relationship to orality (cf. Reece, ch. 3, pp. 55–6 in this volume), oral poetics, (re)performance (in the cases of epic and drama in particular), multitextuality, and multilocality. In medieval studies, the Platonic stemma lost its attraction long ago in favor of publishing the best ms (Leithandschrift) or something more reflective of the typical medieval reader, not the demiurgic author (McGann 1992; Cerquiglini 1999; Jeffreys 2008,). Outside of Classical studies, in other words, scholars have opened themselves to the charms of the many over the one, being often more interested in chasing real readerly contexts than notional Urtexts. Pockets of Urtext-hunting textual criticism remain strong in classics – see especially M. L. West’s Teubner edition of the Iliad (1998–2000) – though as Bolter points out, there remains the irony “that for classical authors the Urtexts are irretrievable, and the texts now established through textual criticism will always be unstable, vacillating among various readings offered by manuscripts and editors” (1993, 161).
But now that the fixity of print is giving way to the new flux and multidirectionality of the digital age, the old print culture consensus may be melting away. For one thing, mss formerly languished in restricted access collections, usually far from the reader of the printed text. The reader had to make do with whatever limited report the editor cared to make of the mss in the apparatus. But now the digitization of major mss, such as those being undertaken by the Homer Multitext Project, effectively – and for the first time, really – democratizes access to the material foundations of ancient texts. While such mss images can be hard to read – one has to have some real paleographical skills – the experience is eye-opening, for the scholar will come to see just how far the print editions are from the textual tradition not only in variant readings, but also in format, orthography, and purpose. Digitization of this textual tradition will also lead to powerful new ways of tracing the complex relationships between texts, doing away with simple Platonic stemmas. And a multitext edition can make available – in ways unthinkable in the fixed medium of print – a wide range of evidence for the text, and not just in the crabbed, apparatus crunching manner of our OCT and Teubner critical texts of yore.
Let us take a concrete example once again in order to understand the kind of problem an editor – and a reader of a critical edition – face. I will use a problem in the text of Homer, since Homeric textual editing presents extremely rich, complex challenges. In the Iliad, old Nestor intervenes at a point in book 1 when Achilles and Agamemnon are wrangling shamefully in front of the Greek army. He reminds them of his own past association with other heroes in order to whip them into behaving themselves. Lines 260–61 read in certain manuscripts: ἤδη γάρ ποτ’ ἐγὼ καὶ ἀρείοσιν ἠέ περ ὑμῖν / ἀνδράσιν ὡμίλησα (“for back in the day I kept company with men much better than you”). Now in line 260 we find other mss reporting ἡμῖν, which changes the meaning to “men much better than us.” This is a difference of a single letter in the Greek (upsilon vs. eta: which is it?), but the overall change in tone is clear: either Nestor shames them by saying his past associates were superior to these two peevish fellows, or humbles himself by saying it has been his honor to have consorted with heroes better than himself – and by implication, them too. When confronted by variants, an editor in traditional print editions had to pick one and only one for pride of place in the main text, while relegating the other(s) to the marginal existence of the apparatus criticus at the foot of the page. So Thomas Allen’s OCT text prints ὑμῖν in the main text, with ἡμῖν reported in much smaller print way at the foot of the page, followed by a short and highly abbreviated list of mss that give this variant.
To decide which is better, one has to come up with stories to explain why the variants exist at all, and two easily suggest themselves: either scribes confused the letters eta and upsilon for some reason (making it a problem for the paleographer to explain), or the linguistic shift known as itacism is responsible, whereby the pronunciation of those two formerly distinct vowels became identical, so us and you sounded alike: /i: mi:n/. Of the two explanations, the linguistic is more convincing than the paleographical in this case. But explaining why variants exist does not help us to decide which variant to prefer, so one then must investigate further. The modern editor has help in the case of Homer from his ancient predecessors. We find that this discrepancy has been argued about since the time of the Alexandrian editors, to judge from things reported in the scholia written in the margins of Homeric mss. The first critical editor of Homer was also the first librarian at the great library of Alexandria, Zenodotus of Ephesus (fl. 280 BCE). He appears to have written ὑμῖν in his edition of the text, which was cobbled together from a variety of mss collected at the library from places around the Greek world.8 (It is possible that Zenodotus, besides making such judgment calls on variants, was the first to divide the Iliad and Odyssey into 24 books.) In this instance, however, Zenodotus only succeeded in irritating his editorial successor, Aristarchus of Samothrace (c. 220–c. 143 BCE), whose substantial role in shaping Homeric textuality as we encounter it today is being recognized more and more (see especially Nagy 2004). He found Zenodotus’ version of line 260 “insolent,” (ἐφύβριστος), and argued that by saying “better than us” instead, Nestor would be reckoning himself among the lesser men dwarfed by the heroes of yore. Later editors and commentators have agreed with Aristarchus, such as the Byzantine archbishop Eustathius of Thessalonica (c. 1115–1195 / 6 CE), who thought “better than us” was more in keeping with the sentiment “such mortals as we are now” (cf. Il. 1.272; Van der Valk 1971, 1:156–7), and G. S. Kirk, who felt Aristarchus might well be right given what follows in the speech (1985, 80). We see “better than us” in certain rather important mss, like the splendid one in Venice known as Venetus A (Dué 2009).
However, other editors side with old Zenodotus, such as Walter Leaf (1852–1927), an assiduous editor of the Iliad who quipped that with “better than us” Aristarchus “wished to import into heroic language the conventional mock modesty of the Alexandrian Court. The whole meaning of Nestor’s speech is that he himself is the peer of better men than those he is advising” (1900, 1:23). Leaf may have made the assumption here that Aristarchus was a naughty interventionist critic, who willfully “corrected” the text according to his own aesthetic conjectures. Increasingly, however, Gregory Nagy and other scholars are asserting that Aristarchus was not rewriting the text of Homer, but rather putting forward variants that already existed in a textual tradition full of such options.9 But with the traditional editorial mindset, the variants still seem to leave us to decide this as a matter of taste: how tactful can we imagine the indignant Nestor to be here? If textual criticism is “the art and science of balancing historical probabilities,” (Kenney 1974, 146), we seem to be at a standstill in a case like this, when ancient variants have equal probabilities. So which one gets to ride in the front seat? Should we flip a coin?
We might well be asking the wrong question here. In fact, Homeric scholarship has recently called into question the posing of such a dilemma, deeming it a pseudo-dilemma. This has to do with a scholarly revolution at both ends. On the one end, Homeric textuality is now being understood very differently by the Oralist school, for whom variation is not a sign of textual corruption, but of living variants of an oral poetic tradition (Nagy 2005, 59). Authenticity, in other words, does not lie with the one, the Urtext, but with the many, the real functional variants. Therefore, at the other end, i.e. the modern critical text of Homer, we should not work towards a text made up of arbitrary binary choices that yield an apparent “unitext” of a great poet (unius munus […] maximi poetae: West 1998, v), but a multitext reflecting a repertory of reported versions, options, processes. The fixed regime of print and the paged book could hardly pull this off in any economical manner, but an online multitext edition is currently under development at the Center for Hellenic Studies, fruit of a more radical philology (http://www.homermultitext.org). This is just one example of the great Digital Renaissance under way in our lifetime – and one very good reason for the student of Greek literature today to be very excited about the future of the past. Our voracious technical sophistication may finally help us to get much closer to the manuscript culture of the middle ages and, though in more fragmentary form, the papyrus book roll culture of antiquity. We might even hope, conceptually at least, to glimpse the distant outline of vastly ancient oral traditions from the dawn of Greek civilization. There is much work to be done – and much work to be undone. But there has never been a better time to be a scholar and transmitter of Greek literature; the auspices are good for tending the wound that is the ancient archive.
Transmission
Pöhlmann 2009 is a great place to start to understand fully how texts come to us from antiquity; Reynolds and Wilson 1991 is perhaps still the best and fullest introduction to textual transmission for the undergraduate or graduate. Gastgeber 2010 is a straightforward article on Byzantine transmission of Greek texts.
Dué 2009 offers a very nice illustrated introduction to one of the most important manuscripts of Homer. Dué and Ebbott 2010 is an exemplary edition of a “multitext” of Homer, focused on the most controversial book of the Iliad from the editorial perspective. Gurd 2005 gives a provocative examination of textual philology centered on a vexed text of Euripides.
Cerquiglini 1999 is a good introduction to a medievalist’s view of how textual variation can be seen beyond “corruption” of the text. Jeffreys 2008 offers a short but pugnacious introduction to textual criticism from the Byzantinist’s perspective. Eisenstein 1993 is a classic statement of the epistemological shifts afforded by the print revolution. Kenney 1974 is a solid, learned discussion of how printing changed the status of the Classical text. McGann 1982 / 1992 is a short, readable monograph that challenges settled notions of what an editor must accomplish. Dickey 2007 is a very useful guide to help navigate complexities of vital ancillary works.
Sandys 1903/1998 is now dated, but still a standard work in multiple volumes; Wilamowitz-Moellendorff 1927/1982 is a briefer monograph on mostly modern Classical scholarship by an icon of Classical studies. Pfeiffer 1968 and Pfeiffer 1976 are standard histories of Classical scholarship that go into considerable detail.