One problem common to the translation of ancient and medieval epic is that the status of the source text itself is not always clear and may even change drastically over time. In the cases of the Epic of Gilgamesh and the Homeric poems, the concept of the source text remains in a state of flux as new evidence arises and new theories of textual crystallization and transmission alter the conceptual horizon. This being the case, a scholarly translation may well have to dispense entirely with the notion of a unitary source text, as is done for example in Andrew George’s translation of the Gilgamesh material, which might be described as a translated textual repertory rather than a rendering of a single ‘‘original text’’ (1999). His inclusive approach effectively dispenses with the notion that there is a single definitive form which we must posit, construct, assume, or select before translating, and in this regard his work is very different from the scholarly rendering of only the ‘‘standard version’’ of the epic (the so-called Sin-leqi-unninni version), as was done by Gardner and Maier (1984), Kovacs (1989), and a great many others. He includes along with the standard version the remains of the earlier Old Babylonian version as well as the Sumerian tales of ‘‘Bilgames’’ which form the nucleus of the legend. George’s approach thus refuses to synthesize the disparate pieces into a single whole for the modern reader, as in modern retellings like the ‘‘verse narrative’’ version by Herbert Mason (2003) or the Penguin edition by Nancy Sandars (1972). In his introduction to the paperback edition, George reminds us dramatically of the instability of his source: ‘‘New pieces of Gilgamesh continue to appear. This paperback edition differs from its hardback predecessor in being able to use on p. 90 a fragment ofTablet XI that came to light only in June 1999’’ (1999: xi).
A similar problem faced the translator/editor of the Greek medieval epic, Digenis Akritis, which exists in two linguistically and stylistically distinct forms. The Grottaferrata Version is in a higher linguistic register and was championed as ‘‘the original’’ by modern adherents to katharevousa (the ‘‘purified’’ version of Modern Greek, only officially abandoned in 1974), while proponents of the demotic upheld instead the Escorial Version, a less polished piece of poetry couched in a markedly more popular idiom. The translator/ editor Elizabeth Jeffreys, convinced that both versions are drawn from a common but missing source and refusing to take sides on the language issue, had no choice but to translate both versions and present them seriatim as traces of a missing ancestor (1998). We are thus not presented with an ‘‘original’’ and a translation, but with two witnesses to an hypothesized original. Even in the comparatively simpler case of Virgil’s Aeneid, the translator must address the problem of a text that was reportedly left unfinished, includes seemingly unfinished verses, and whose very beginning lines vary in the textual tradition. The Virgilian translator thus faces the task of in a sense completing the text (see section 4), and still must grapple with the matter of textual variation, even if it is on a smaller scale.
The problem of the source can be greatly complicated by the idea that an epic text may represent the residue of an oral tradition; indeed, the most radical view of oral epic challenges all assumptions about the singularity and originality of any ‘‘text’’ (on these issues, see especially Nagy 1996b and 1996c). A team of scholars at the Center for Hellenic Studies in Washington, DC is undertaking a translation of the Homeric textual repertory in conjunction with their development of a Greek multitext, which will include ‘‘plus verses’’ from the papyri not represented in the medieval manuscripts and as many textual variants as the technology will allow. This would represent a decisive change in Homeric translation, one which in conceptual terms moves well beyond even the most literal translations of the past by leaving the source text open to different philological interpretations. The age of hypertext, which can facilitate such complex and layered approaches to a fluid textual tradition, may well permanently alter the possibilities and parameters of translation as we know it at the same time it moves us beyond the Platonic idea of a fixed recension (i. e., a singular “reconstructed original’’).
We might also say that the notion of the source as a ‘‘naked text’’ encountered without mediation is a widespread myth, quite often propagated by translators themselves, who may be quick to deny their reliance upon others. Besides the crucial interventions of copyists and editors on the source, we must consider the influence of scholia, commentaries, dictionaries, and other ancillary works on the translator’s understanding. It is also rare in the case of ancient epic that a translator is truly the first to render the source text, which leads us to ask: Did the translator work via an intermediary translation or translations? Did these influence the target text’s style or approach, either as models to follow or as bad examples to be avoided? Was the intermediary translation in the target language, or in another (e. g., a Latin translation of Homer used by an English translator)? One must seriously consider the possibility that the ‘‘source text’’ is not in fact the ‘‘original’’ at all, but rather that the source comprises a whole series of previous translations with perhaps only some input fTom the ‘‘original,’’ or even none at all. In sum, the hieratic primacy of the ‘‘original’’ is itself an ideology or position to be taken, and it has not always been the position taken in practice by every translator for a variety of reasons. A tenuous connection to the source often has no effect on the success of the translation in its target environment. Indeed, we know fTom the Ossianic ‘‘translations’’ of James Macpherson (1736-96) that a highly successful target text can even lack an identifiable source (see Chapter 29, by Due). There are also known literary texts that trade on the outright fiction of their translation, like the History of the Destruction of Troy attributed to Dares the Phrygian. This Latin work sports a bogus letter of introduction supposedly from the Roman historian Cornelius Nepos to his colleague Sallust, which attests to the scrupulous translation of a Greek original Nepos found in Athens. While the Latin may in fact be a translation of a Greek original, the point is: the story of its translation by Cornelius Nepos is definitely a fiction used to authenticate it for the target audience.
Considerations stemming from the source text go well beyond the basic outlay of the textual tradition - i. e., its manuscripts, tablets, or papyri (on which media see Chapter 10, by Haslam) - and its intermediaries. How one will translate a text in point of style, diction, and even narrative approach will differ according to the kind of text one assumes the source to be. For example, if one assumes that the epic to be translated reflects the poetics of an oral tradition, then features like formulaic diction, repeated verses and epithets take on a new importance (see Chapter 4, by Jensen). If one assumes rather that the poem was composed by a court poet in a high and solemn style, then this will condition the style and language of the target text accordingly. If one assumes the ancient poem is just one version of a tale available for retelling, then the translator feels understandably freer to render the plot with significant lexical, stylistic, and even narrative divergence according to the tastes of the target culture. If one assumes instead that the poet’s ipsissima verba contain arcane wisdom of great importance, the tendency to take a word for word approach (perhaps with exegetical digressions) is more likely. In each of these cases, then, we can see that the assumptions that frame the source text for the translator are the point of departure for the target text which will emerge. Since assumptions about the source text are rooted in the target culture, there is understandably a certain hermeneutical circularity at work. By way of illustration, we shall now examine more closely the four sets of assumptions mentioned in the scenarios above.
World History
