surviving physical object considered in isolation. Obviously, in many cases it will be very difficult to approach a manuscript in any other way, because most provide few clues of how they might have been performed. As we shall see, however, even when these clues are abundant, as is the case with Harley 978, they are often ignored.
The Voice as Material Support
Modern bibliography has fought to call attention to the overlooked, the apparently trivial or insignificant details of a text’s physical form that turn out to play a crucial role in defining a literary work and its readership. But among scholars working with printed materials, this physical form is most often taken to refer to the book as a tangible object and to its visual appearance. This is not because social bibliographers are indifferent to the myriad ways in which a book can be performed. On the contrary, the “history of the book” that has been written during the last few decades has been equally a history of reading. For those who work in these later centuries, there has been no shortage of material. The proliferation of petits papiers and the Romantic autobiographical impulse have meant that details of daily life survive in ever greater abundance and that early modern texts can be located within a plausible and detailed history of reading and performance practices. No history of eighteenth-or nineteenth-century reading forgets the oral dimension, the importance of reading circles, salons, or young couples linked by a shared pleasure in illicit books.54 Nevertheless, private and silent reading increasingly became the norm. Reading aloud, whether in the family circle, the salon, or a theater seating two thousand, was structured around widely available texts. At all social levels, people gathered together to read books that were already bestsellers and enjoy “the public acknowledgment of a shared private experience,” Helen Small’s characterization of the immensely popular readings offered by Dickens.55 Although the readings accounted for roughly half of Dickens’s fortune, they remained, in his eyes and those of his friend and biographer Charles Kent, a slightly disreputable supplement to his printed works.56
The assumption that a book’s public performance is never more than a supplement to its private reading is fundamental to modern publishing. Dickens’s powerful dramatic readings, for example, depended on the widespread availability of standard printed editions to forge the sensibility of his listeners and provide them with common referents. These assumptions are reflected in a book’s printed form. Any edition of Dickens, from the first serial installments on, serves in the first instance as a text that one person can read silently and then, and only occasionally and often after some physical preparation, as a script for public reading.
The conventions governing medieval manuscripts are very different. First, silent reading cannot be assumed. The habit of silent reading was rare even in monastic communities before about A.D. 1000 and only gradually spread outward to clerics and then lay people, and from Latin texts to vernacular ones.57 As late as the fourteenth century, there seems an element of novelty in Chaucer’s depiction of himself sitting “as dumb as any stone” when he retires to read. To use Didron’s phrase, a manuscript must be recognized as a sung object, and singing covers a wide range of activities from solemn chanting to private mumbling. Furthermore, the conventions that now permit us to distinguish between a play script, a novel, and a piece of sheet music were yet to be defined. Musical notation was only partially developed, and few could read it, so medieval songs were not necessarily distinguished in manuscript from lyric poems. This means that a large body of medieval poetry, including the lais, romances, and chansons de geste, as well as short forms like the ballade, rondeau, or virelai, now exist in limbo as far as performance history is concerned, and in their own day may well have been presented in a variety of ways as manuscripts passed from one group of users to another.
These claims for the importance of the oral aspect of medieval works are scarcely new. Paul Zumthor asserts that a medieval text is only the occasion for a vocal act.58 J. A. Burrow compares medieval books to a modern musical score.59 Walter Ong argues that, in comparison to print culture, “manuscript culture felt works of verbal art to be more in touch with the oral plenum, and never very effectively distinguished between poetry and rhetoric.”60 Nevertheless, the challenges of addressing the sound of a manuscript are extreme, and the editorial and critical treatment of medieval texts has often failed to meet them, so that the vocalization of medieval texts has all too often been ignored, normalized, or consigned to unexamined stereotypes. Here too the mental habits induced by print have been harder to shake than is generally realized.
There are numerous difficulties, but the most obvious and insoluble is the ephemeral nature of vocalization. It is not just that we have no audio recordings of medieval singers or storytellers; we have very few detailed contemporary reports either. The culture of the book provided few models for detailed accounts of popular oral performance. Developments in plainsong and polyphony, patterns of monastic lectio, the pious reading habits of saintly aristocrats, the power of a mendicant preacher or a court’s designated reader—these are described in some detail. But for minstrel performance we have little to go on beyond the occasional allusion in a popular sermon or the highly conventionalized references in the lais, romances, or chansons de geste themselves, one of the trickiest of sources. Reconstructing minstrel performance involves us in speculation, generalization from a handful of examples, and a literalistic reading of literary texts as if they were social reportage. Such approaches are characteristic of the great antiquarians of the eighteenth century, Joseph Ritson and Bishop Percy prominent among them, who initiated the history of minstrelsy. On many points they did the job about as well as it can be done. Since then we have culled further references, but our methods for reading them remain much the same.
This methodological crudity, which will offend the modern professional, whether historian or literary critic, may partially account for the cool reception accorded performance history. Edmond Faral’s Les jongleurs en France au moyen âge of 1910 still serves as a standard authority on minstrel performance, while more recent work has made surprisingly little impact on literary studies, at least in the field of Old French.61 In Chapter 2, I will examine some of the evidence of performance practice, asking whether it was at all likely that a full-length chanson de geste was ever performed by a minstrel, and in particular what evidence we have for the existence of sustained recitation or what Léon Gautier termed the “séance épique,” in which a minstrel held an entire hall in his sway. For the moment, I wish merely to acknowledge the difficulty of reconstructing medieval performance, while insisting on the absolute necessity of making the attempt. There is very little we will ever be able to claim we actually know about any medieval performance, but overt speculation is better than unexamined assumption.
The influence of print and its dominant mode, silent reading, may encourage us to think that questions of performance can be ignored and that it is possible to avoid the dirty work of speculative reconstruction and approach a medieval text in a neutral fashion without prejudging the way in which it was performed. The editorial history of the Song of Roland provides a particularly forceful example of why this is not so, showing how much is at stake in classifying this poem as a song. Here Chartier’s formulation, with one slight modification, once again makes the point: no medieval text existed in its day outside the material support that enabled it to be read or heard. A medieval text might have existed as a monk’s slow mumbling, as an ongoing courtly flirtation, as a regular daily ritual in a monastery or great household, or as a few snatches from a familiar story sung on street corners—but it never simply existed. Just as an eighteenth-century poem existed in some specific edition, so a medieval poem existed in some specific performance, and this performance was no less fundamental in determining what the text was.
The Edge of the Book
“The idea of the book, which always refers to a natural totality, is profoundly alien to the sense of writing.”62 So Jacques Derrida argues in the opening pages of his famous De la grammatologie. While his use of the terms “book” and “writing” continues to perplex, he suggests the extent to which our familiar habits of thought, founded upon long-standing traditions of written authority, predetermine our understanding. The “idea of the book” provides an all-encompassing frame of reference, and efforts to think outside it will inevitably falter. Whatever form of proliferating