When Dan Simmons's 2003 novel Ilium begins, the Trojan War is in its ninth year, just as it is at the beginning of Homer's Iliad. The events that unfold upon the field of battle some time around 1200 BCE, however, are manipulated by gods, actually technologically enhanced post-humans, who live on Olympus Mons on Mars, and broadcast the war to the dwindling and degenerate surviving population of “old-style” humans who still inhabit the earth sometime in the late fourth millenium CE. The novel plays not with the well-established science fiction trope of the time paradox, as one might expect, but rather with the idea of narrative temporality itself. The text of the Iliad, often quoted, operates like destiny: the readers outside the text know how it ends, and so do some of the readers inside the texts, the “scholics” kidnapped from early twenty-first century Classics departments to report on the degree to which a given day's events do or do not map onto Homer's account. And when events begin to diverge from that account, when Achilles and Hector join forces to make war upon the meddling gods, all kinds of possibilities seem to open out: what if Hector didn't die? What if the city never fell? What if narrative time and destiny were not one and the same?

John Hollander's poem “The Altarpiece Finished” is an ekphrastic meditation on the impossibility of ekphrasis. The poet speaks in the voice of the painter Hubert van Eyck, elder brother of the more famous Jan, who together created the Ghent altarpiece. Hollander evokes the miraculous complexity of a work of art that contains the entire painted world, not only “a great biblical golden talent of invention” in the depiction of God and all his angels and saints, Adam and Eve, Cain and Abel, and a host of other figures from scripture and Church history, but also “a grey pond with five tiny figures gliding | over it” and “a pair of burghers discussing something | underneath the sign of the Ram's Head.” Whether these details can actually be detected within the universe of the van Eycks’ masterpiece matters not at all (looking at the extraordinary high-resolution image developed by the Getty Foundation, I cannot see the pond or the sign of the Ram's head, but I can certainly see pairs of tiny burghers in deep discussion); rather, it is the act of creative imagination involved in the description that is the point and also the problem. Of ekphrastic poems like this, Hollander remarks in an essay, “Because I could never draw or paint, and I loved and remembered and brooded over pictures even in childhood; the only way to make them was to describe (a) actual ones or (b) those which I invented.” The descriptive mode may thus be entirely fictional, even when it attaches itself to a real work of art. At the end of the poem, the speaker links artistic creation to both time and death: “But what is time? something paintings take to finish, or to rot or | To become the way things look in.

The primary subject of this book, its central time machine, is Benoît de St-Maure's Roman de Troie, which I will argue participates in the creation of a new kind of narrative time as the poem moves towards its inescapably tragic end. The Roman de Troie was one of the great medieval best-sellers. In its pages, Benoît invented the love story of Troilus and Briseida, which inspired, directly or indirectly, Boccaccio's Il Filostrato, Chaucer's Troilus and Criseyde, and Shakespeare's Troilus and Cressida. Indeed, the Roman de Troie generated the most familiar medieval version of the story of the Trojan War through its subsequent translation into Latin by Guido delle Colonne as the Historia destructionis Troiae, itself the source of Lydgate's Middle English Troy Book and of the Middle French text translated by Caxton into English as the Recuyell of the Historyes of Troye. Not until vernacular translations of the Aeneid, and eventually the Iliad itself, began to multiply in the early modern period would Benoît's version of the story lose its authority, but then it did so precipitously. From the seventeenth century onwards, it was little read. Although the text was edited by Leopold Constans between 1904 and 1912, only excerpts appeared in English, and the modern French facing page translation which appeared in 1998 was much abridged; the editors deemed its many battle scenes too much for modern sensibilities and cut most of them. There is something ironic about this version's dramatic reduction of Benoît's original, given the way he operated on his own source text, Dares of Phrygia's De Excidio Troiae, which he expanded almost beyond comprehension, turning a scant 52 pages of prose (in the Teubner version) into an octosyllabic poem over 30,000 lines long, twice as long as the Iliad.

This book explores the curious intersection of translation and temporality in the Roman de Troie and its sources and inspirations; it pays attention to the ways in which the act of translation from Late Latin into Old French, from Late Antiquity to the twelfth-century Renaissance, and from prose into verse also translates history into romance, reconfigures narrative time, and causes the appearance of unexpected textual objects, the time machines of the texts: characters, armour, clothing, and monuments.

In the last chapter, I argued that the Roman de Troie asks to be read as a history, and also that history was, in the twelfth century, an extremely capacious genre embracing fiction as well as fact, poetry as well as prose. Perhaps the only element that is truly common to texts as varied as Orderic's Historia Ecclesiastica, Henry of Huntingdon's Historia Anglorum, Wace's Roman de Rou, and Benoît's Chronique des Ducs de Normandie is that all are concerned with the past. But what happens when a history, a tale of things past, looks suddenly to the present and even on to the future? In this chapter, I intend to argue that while Benoît cites Dares and Dictys and often draws on Ovid and other classical writers, one of his most significant influences is actually Geoffrey of Monmouth's De Gestis Britonum. The De Gestis Britonum, like the Roman de Troie, makes its claim to be history early, in the opening line, in fact: “While my mind was often pondering many things in many ways, my thoughts turned to the history of the kings of Britain.” What follows certainly looks like one kind of history: beginning with Brutus, the grandson of Aeneas, Geoffrey traces the lineage of the Kings of Britain down through the ages until 682 CE. The work constructs a linear chronology, reign by reign, father to son (or, in some cases, daughter), until the end of the sixth book when something peculiar happens and the prophecies of Merlin erupt. Regardless of whether they are true or false, a question much debated by Geoffrey's contemporaries, the prophecies of Merlin change the temporal orientation of the text; they not only predict the future from the perspective of those within the text, they carry on to predict a future occurring after the text ends with the fall of Britain to the Anglo-Saxons, and then further still to a future that extends through and even beyond the lifetime of the author. Geoffrey, the most widely read of all twelfth-century historians, models for Benoît first how to create authenticity for a long and fantastic narrative by claiming to translate an ancient and unread work;

In the generation after Geoffrey of Monmouth, the authors who retold the Troy story and invented the genre of romance took advantage of his central intuition about time: that it was profoundly malleable. In the De Gestis Britonum, as we have seen, time runs in two directions, pressing simultaneously backwards and forwards from the present of the text's composition. The great originality of the next phase of the medieval evolution of the Troy story in the Roman d’Eneas and the Roman de Troie lies in the invention of a new kind of narrative time, a romance time fundamentally at odds both with the baldly teleological historico-fictive time that dominates the narrative of the De Excidio Troiae and with the bi-directional genealogical/prophetic time that shapes the De Gestis. Both of these earlier temporalities are essentially objective, impersonal, and public. As Ricoeur puts it, “in history, death, as the end of every individual life, is only dealt with by allusion to the profit of those entities that outlast the cadavers – a people, nation, state, class, civilization.” The De Gestis is not, except almost coincidentally, the story of any particular individual, even Brutus or Arthur, but of the Kings of Britain as a category (tribal, ethnic, or national); the importance of each king lies only in his relationship, more or less important, to those who come before and after. By contrast, the Roman d’Eneas, the immediate precursor to the Roman de Troie, is not only part of the initiation of a new genre but also helps to invent the temporality that will come to characterize that genre. This temporality, which Bakhtin in “Forms of Time and the Chronotope in the Novel” calls romance time, appears to be subjective, individual, and rooted in a heterosexual erotics that embodies dynastic claims upon the future.3 Romance time, in fact, initially looks like women's time, since it is in the romance that women suddenly come to occupy centre stage. In spite of its association with the feminine, however, romance time in the Eneas will prove uniquely oriented towards the reformation of the male hero and the furtherance of his dynastic interests.

Turning to the manuscript tradition of the Roman de Troie, especially in this Plague Year of 2020, it is difficult not to feel like the weary narrator of Borges's “Library of Babel.” True, the manuscripts of the Troie are not infinite; thirty complete witnesses survive, and thirty-one fragments. Yet the potential relationships between these manuscripts generate a virtually infinite number of possible texts, a fact which has not yet been sufficiently explored, at least not in scholarship in English. In these few pages, I will attempt to give a bird's eye view of the situation, with special attention to the questions of which texts were grouped with the Troie and what such groupings might tell us about the reception of the text, especially in the thirteenth and fourteenth centuries. Increasingly, the manuscripts (especially those in Paris) are available online, which is a great boon to scholars;3 because the pandemic curtailed travel to see others for myself, I am deeply indebted to Marc-René Jung's encyclopedic La légende de Troie en France au moyen âge, which attempts nothing less than a complete catalogue of the more than 350 French manuscripts containing one version or another of the Troy story: the Roman de Troie and its mises en prose, French translations of Dares (usually part of the Histoire ancienne jusqu’à César), French translations and adaptations of Guido delle Colonne's Latin mise en prose of Benoît, and a few other related texts.

There are two major schools of textual criticism. The first, named for Karl Lachmann, attempts to reconstruct a lost original by classifying manuscripts into families according to shared errors and their treatment of difficult passages, thus trying to establish what the author actually wrote.

“Marriage, which has been the bourne of so many narratives, is still a great beginning,” remarks George Eliot at the end of her grand narrative of many marriages, Middlemarch. The word bourne is a curious one, archaic already when Eliot was writing. It can be a terminus, a limit, a destination, and indeed this is precisely how marriage functions in the happily-ever-after ending of the Roman d’Eneas: the courtly hero has won his lady and what comes after the wedding – the birth of children, the establishment of dynasties – is essentially non-narratable, at least within romance time. But what about those for whom marriage isn't a happy ending? Or those for whom, more simply, marriage just isn’t? A bourne can also be a boundary, the line between one plot of land and the next, a sign establishing difference. Looking from the Roman d’Eneas to the Roman de Troie, we need to ask what happens when marriage is not a great beginning. For Troy there can be no happily-ever-after. The Horse delivers its treacherous progeny, towers burn, streets run with blood, Priam is cut down upon the altar, Polixena on the grave of Achilles, and Hecuba stoned to death. Even for the victors, things don't go well. Benoît provides a summary of the Greek homecomings, derived from Dictys: Agammemnon is murdered by Clytemnestra, Pirrus dies at the hands of Orestes because of his abduction of Helen's daughter, Ulysses is haunted by nightmares and eventually killed by Telegonus, his unrecognized son by Circe. For the heroes of Troy, it turns out, romance time, associated with the love of women, the contract of marriage, and the establishment of heterosexual bonds, is not all that it's cracked up to be. The return to domesticity and the embrace of family is no great new beginning but only a return to death. Both romances recognize this phenomenon, although while the Roman d’Eneas, as we have seen, works hard to produce a positive and secure vision of the heterosexual bonds of marriage and courtly love, the Roman de Troie, ultimately, doesn't bother. Each text, moreover, proposes alternatives to romance time and the imperatives of heteronormativity, even if only to foreclose them.

Sometimes history and romance, truth and fiction, seem more closely entangled than usual. Writing in December of 1792, after five months of extraordinary events from prison massacres to the abolition of the monarchy, Maximilien de Robespierre remarked, “our revolution has made me fully feel the force of the saying that history is a romance.” As I will argue in the pages that follow, the Anglo-Norman regnum in the mid-twelfth century forms another nodal point where the boundaries between history and fiction reveal themselves as particularly fluid. Benoît de Sainte-Maure's Roman de Troie, composed around 1165, both emerges from and participates in this fluidity, not least for the simple reason that it is a poem in octosyllabic rhyming couplets which declares itself to be a translation of a late antique prose text not even a tenth of its length. Translation as practiced by Benoît is a complex temporal phenomenon that reconfigures the past and produces specific effects upon the present and thus, potentially, upon the future. The use of translation to erode the borderline between history and fiction in the Roman de Troie reveals the essence of a peculiar kind of narrative temporality that would prove to animate the medieval imagination for centuries, a temporality we might call Trojan time.

Trojan time as it emerges in the twelfth century (there are other Trojan times, of course, notably the Homeric and the Virgilian) proves to be an erratic and brilliant manifestation of what Paul Ricoeur calls the “interweaving of history and fiction” according to which the two modes of narration use each other to refigure time itself. This refiguration is necessary because of the fundamental incompatibility between cosmological time – the movement of stars and planets, creating days and nights and years – and phenomenological time – lived, human time. The first volume of Time and Narrative begins with a reading of the eleventh book of Augustine's Confessions, in which the Bishop of Hippo confronts the impossibility of thinking about time: the past is always already past, the future is not yet, and by the time a thought can be articulated about the present, it has become the past.