Textual Situations: Three Medieval Manuscripts and Their Readers

By Andrew Taylor

Generations of scholars have meditated upon the literary devices and cultural meanings of The Song of Roland. But according to Andrew Taylor not enough attention has been given to the physical context of the manuscript itself. The original copy of The Song of Roland is actually bound with a Latin translation of the Timaeus.

Textual Situations looks at this bound volume along with two other similarly bound medieval volumes to explore the manuscripts and marginalia that have been cast into shadow by the fame of adjacent texts, some of the most read medieval works. In addition to the bound volume that contains The Song of Roland, Taylor examines the volume that binds the well-known poem "Sumer is icumen in" with the Lais of Marie de France, and a volume containing the legal Decretals of Gregory IX with marginal illustrations of wayfaring life decorating its borders.

Approaching the manuscript as artifact, Textual Situations suggests that medieval texts must be examined in terms of their material support—that is, literal interpretation must take into consideration the physical manuscript itself in addition to the social conventions that surround its compilation. Taylor reconstructs the circumstances of the creation of these medieval bound volumes, the settings in which they were read, inscribed, and shared, and the social and intellectual conventions surrounding them.

• Language: English
• Publisher: University of Pennsylvania Press
• Release date: Aug 31, 2015
• ISBN: 9781512808001

Andrew Taylor

Andrew Taylor is the author of a number of crime novels, including the ground-breaking Roth Trilogy, which was adapted into the acclaimed drama Fallen Angel, and the historical crime novels The Ashes of London, The Silent Boy, and The American Boy, a No.1 Sunday Times bestseller and a 2005 Richard & Judy Book Club Choice. He has won many awards, including the CWA Ellis Peters Historical Award (the only author to win it three times) and the CWA’s prestigious Diamond Dagger.

Book preview

Textual Situations

MATERIAL TEXTS

Series Editors

A complete list of books in the series is available from the publisher.

Textual Situations

Three Medieval Manuscripts and Their Readers

Andrew Taylor

UNIVERSITY OF PENNSYLVANIA PRESS

Philadelphia

Copyright © 2002 University of Pennsylvania Press

All rights reserved

Printed on acid-free paper

10  9  8  7  6  5  4  3  2  1

Published by

University of Pennsylvania Press

Philadelphia, Pennsylvania 19104

Library of Congress Cataloging-in-Publication Data

Taylor, Andrew, 1958–

Textual situations : three medieval manuscripts and their readers / Andrew Taylor.

p. cm. — (Material texts)

ISBN 0-8122-3642-4 (alk. paper)

Includes bibliographical references and index.

1. Bodleian Library. Manuscript. Digby 23. 2. British Library. Manuscript. Harley 978. 3. Smithfield decretals. 4. Manuscripts, Medieval—England. 5. Books and reading—England—History—To 1500. 6. Literature, Medieval—Criticism, Textual. 7. Transmission of texts. 8. England—Intellectual life—1066–1485.

Z106.5.G7 T39 2002

Contents

1: MEDIEVAL MATERIALS

2: BODLEIAN MS DIGBY 23

INTERSTICE: THE MINSTREL AND THE BOOK

3: BRITISH LIBRARY MS HARLEY 978

4: BRITISH LIBRARY MS ROYAL 10.E.4

5: THE MANUSCRIPT AS FETISH

LIST OF ABBREVIATIONS

NOTES

DISCOGRAPHY / BIBLIOGRAPHY

INDEX

ACKNOWLEDGMENTS

Chapter 1

Medieval Materials

In contrast to the representation of the ideal, abstract text—which is stable because it is detached from all materiality, a representation elaborated by literature itself—it is essential to remember that no text exists outside of the support that enables it to be read.

—ROGER CHARTIER

Pour engager d’autres savants à faire des recherches de ce genre, en les étendant à tous les siècles et à toutes les variétés de sujets, il convient de parler à l’esprit et aux yeux, de décrire et montrer en même temps les objets chantés et dessinés.

To incite other scholars to conduct research of this kind, reaching out to them across the centuries and in all manner of subjects, we must appeal to the spirit and to the eyes, describing and showing at the same time these sung and drawn objects.

—ADOLPHE NAPOLEON DIDRON

The core of this book is devoted to an examination of three medieval manuscripts, the support that enabled a variety of texts to be read and performed. One is now in the Bodleian; the other two are now in the British Library. All three have been in southern England for centuries, but apart from this proximity they have little in common. The earliest, Bodleian MS Digby 23, is a small double volume consisting of two separate booklets, each dating from the twelfth century. One booklet contains Calcidius’s translation of the Timaeus; the other the best-known version of the poem now called La Chanson de Roland. The second manuscript, British Library MS Harley 978, is a slightly thicker volume, a miscellany dating from the mid-thirteenth century that contains an amazing range of material, including the Middle English lyric Sumer Is Icumen In, Latin satires and drinking songs, a long celebration of Simon de Montfort’s victory over Henry III at the Battle of Lewes in 1264, and works by the twelfth-century Anglo-Norman poet Marie de France—her translation of Aesop’s fables and her short romances, or lais, which explore the themes of amour courtois. The third manuscript, British Library MS Royal 10.E.4, copied in the early fourteenth century in Italy but later acquired by St. Bartholomew’s Priory in Smithfield, on what was then the outskirts of London, is by far the largest and most ornate, a handsome folio-sized copy of the Decretals of Pope Gregory IX, also known as the Liber extra, with an astonishing series of marginal illustrations that were added later in the century in England.

Many of the works these manuscripts preserve are well known; some, such as the Song of Roland, the lais of Marie de France, or Sumer Is Icumen In, are famous. The three manuscripts, however, have remained all but invisible, and it is part of my purpose to inquire why this should be so. Of course, editors may always choose to ignore one manuscript witness and concentrate on another, and the vast majority of medieval manuscripts languish unread for years. But this is not the problem here. Sumer Is Icumen In survives only in Harley 978, which also provides the base text for almost all editions of Marie’s lais and most editions of her fables, while it is the version of the Roland in Digby 23 that has come to represent the poem. Why then has the editorial construction of the Song of Roland and the Lais of Marie de France been conducted with such comprehensive disregard for the manuscripts from which these poems were extracted? Or, to turn to the last manuscript, why do the marginal drawings from the Smithfield copy of the Decretals of Gregory IX, which crop up again and again as illustrations of medieval wayfaring life, circulate without any connection to the legal text whose borders they decorate? In examining this curious invisibility, I hope to suggest something of the way in which the material support of the medieval text, which is not just the manuscript but also the social conventions that surround it, differs from that of the printed book.

It seems oddly fitting that the second epigraph that opens this chapter should come from such a peripheral position, the final sentence of a note by the editor, Adolphe Napoléon Didron, to an article by Viollet-Leduc, père, in the second volume of the Annales archéologiques, published in 1845—scarcely the place for a manifesto. In setting out his program for iconographie studies of medieval sculpture, painting, and poetry, Didron makes three points that seem to me especially valuable. First, when he describes manuscripts as drawn objects, he appeals to what will become a crucial principle of textual materialism well over a century later, that texts only exist in precise physical forms, whose design, script, and accompanying apparatus are all integral parts of the texts’ meaning. This line of argument has been extensively developed in more recent years and expanded to cover modern printed editions, whose exact bibliographical format is now seen as a crucial component of a text’s meaning. But Didron suggests a second line of inquiry as well, and one that has not been so widely pursued, when he invites us to conceive of manuscripts as sung objects, stressing their acoustic as well as just their visual materiality. Finally, Didron recognizes that the study of medieval manuscripts is a cumulative and collaborative venture, one that reaches across centuries. It is a generous vision, and I can only hope that I have managed to do it justice, acknowledging my innumerable debts—to those who have maintained the tradition of painstaking scholarship that is needed to read medieval manuscripts, to those who have opened up medieval studies to the bracing winds of contemporary theoretical and cultural debate, and to those who have done a little of both.

Before proceeding any further, it will be useful to say a little more about these three manuscripts and the kinds of problems they present. All three juxtapose remarkably divergent material, and this was one of the reasons I chose them. Digby 23, as already mentioned, consists of two parts, both copied in the twelfth century. The first, Calcidius’s fourth-century translation of Plato’s Timaeus, was one of the most important philosophical texts of the high Middle Ages. The Digby version was probably copied by a Norman or northern French scribe, and it includes numerous and substantial interlinear and marginal glosses from the twelfth and thirteenth centuries. Later and more informal marginalia show that this copy was still in use in the fourteenth century, by which time some would have considered it something closer to a literary classic than a work of rigorous contemporary philosophy. The second part of Digby 23, the Roland, was copied either a little earlier or about the same time; most paleographers favor the second quarter of the twelfth century, although some would prefer a date as late as the 1170s. It was copied by an Anglo-Norman scribe, perhaps one working in the household of a bishop or a great magnate. There are some signs that the two parts were gathered together (that is, either bound together or kept together in the same parchment wrapper) by the thirteenth century. The first identifiable owner of Digby 23 is an Oxford scholar, Master Henry Langley, known to have been alive in 1263, who donated the book, or one part of it, to the Augustinian canons at Oseney Abbey, on the edge of town. Henry might have owned both booklets, but it seems at least as likely that the Roland was added later. However, the two booklets do seem to have been gathered together as a single person’s private collection within at most a few decades of Henry’s death, because someone has added the word Chalcidius to the Roland section in what appears to be a thirteenth-century hand. This would mean that the first reader of the Roland who can be even partially identified would be an anonymous canon of Oseney, possibly one of Henry’s friends. Whatever the arrangements by which the booklet came into the abbey’s possession, one thing seems clear: by the end of the thirteenth century, the Oxford Roland had become reading matter for English clerics.

This codicological information has been long known. But it has been largely, one might almost say systematically, ignored. Scholars have been remarkably slow to abandon the notion (for which there is not the slightest supporting evidence) that the Roland booklet belonged to a minstrel. Others have mounted a desperate rearguard action, assuring us that, even though the Roland booklet is found in the library of English canons, they did not actually read it and only kept it out of pious respect for their chivalric benefactors. And even now, when the idea of the minstrel manuscript has finally been laid to rest, nobody has shown more than the most passing interest in these thirteenth-century English readers. Henry Langley, it turns out, is more than just a name. There is a good deal we know about him, or at least about his father, arguably the most hated man in England in his day. But the world of thirteenth-century English clerics seems irrelevant to the prevailing understanding of what the real poem must be, an eleventh-century French sung epic. What has displaced the history of the manuscript is the modern editorial construction of La Chanson de Roland. The title of this work, which we inherit from its first editor, Francisque Michel, sums up the vision of the poem as a minstrel’s song. I have dwelt on this editorial construction at some length because it furnishes a powerful example of the way a manuscript can be effectively ignored while the words in it, or in some part of it, are treated with scrupulous care.

Each of the three manuscripts has presented different challenges, and for each I have taken a different approach. While I have grouped them in chronological order, I have also found that by good fortune they fall into a methodological order, so that the problems raised by the first are illustrated more forcefully by the second and more forcefully still by the third. In the case of Digby 23, I have from time to time indulged a certain empiricist hubris, pitting hard facts and concrete objects against the free-floating fantasies of the old philology as I call into question the very existence of the Song of Roland before its publication by Michel in 1837. My encounter with the glosses in the Timaeus, however, marks the beginning of a long erosion of that certainty, as the diversity of the manuscripts reveals the inadequacies of my knowledge. I have cast some lines from the world of Old French epic to that of the Anglo-Norman schools, but my treatment of the Timaeus remains limited and my reader must not hope for balanced coverage. If my account draws attention to some of the work that is being done in this area, work that will not be familiar to all literary scholars, it will have served its purpose.

The challenge of diversity is even greater with the second manuscript, Harley 978, a trilingual, multidisciplinary miscellany whose separate sections have almost never been considered together. Harley 978 is a small portable collection, or manual, to use a term current in thirteenth-century Anglo-Norman circles. Like other manuals of the period, the choice of items that were included reflects the social, intellectual, and spiritual ambitions of its owner, who would select them personally. This is a book to fashion an identity. From this single collection one reader might learn the language of amour courtois, the technicalities of hawking, various treatments for imbalances of the body’s four humors, and the arguments used against the king at the time of the Baron’s War. Harley 978 is also an important manuscript for several modern fields of study. With the exception of the lais, the fables, and a few of the Goliardic poems, most of the works in Harley 978 survive nowhere else. This is true of both Sumer Is Icumen In, one of the earliest and most famous of Middle English lyrics, and the Song of Lewes, the long encomium for the baronial leader Simon de Montfort, which sets out a theory on the limits of royal power, making it an important document in English constitutional history. Harley 978 is no less important for French literature. While there are numerous later copies of Marie’s lais and fables, Harley 978 is the earliest surviving manuscript of either and begins to define her canon. There are actually a number of Anglo-Norman texts attributed to a woman or women identified only as Marie and other manuscripts that contain anonymous lais in a style at least somewhat similar to those in Harley 978. The first editor, Jean-Baptiste Bonaventure de Roquefort, for example, used a thirteenth-century Picardian manuscript, BN fr. 2168, and made up a somewhat different collection.¹ And only one of the Harley lais, Guigemar, actually mentions Marie by name, the eleven others being anonymous. Yet the elegant, early Harley 978 has carried the day. For modern readers, the Harley lais have taken on the stability of an authorized collection; they have become the Lais of Marie de France. Some scholars go so far as to claim the order of the lais is the very order Marie imposed in a final reworking or assembling of her work.² Works that might be attributed to Marie but appear in other manuscripts receive far less attention.³ Despite the manuscript’s preeminence, however, there has been a staggering indifference to its full contents or to the history of its readers.

The debate on the identity of Marie de France offers a somewhat embarrassing illustration. Most scholars now accept that the author of the Lais, who at the beginning of Guigemar calls herself Marie, ki en sun tens pas ne s’oblie (Marie, who in her day should not be forgotten), is also the author of the translations of Aesop’s fables, who tells us in the epilogue, Marie ai nun, si sui de France (Marie is my name and I am from France), and is also the author of the tale of a knight’s descent into the otherworld, the Espurgatoire Seint Patriz, who tells us at the end, Jo, Marie, ai mis en memoire/ le livre de l’Espurgatoire (I, Marie, have recorded for memory the book of the Purgatory).⁴ Marie obviously must have been born in France but later lived in England for her name to make any sense. It is generally assumed that she was an aristocrat, someone elevated enough to know a Count William, for whom she translated the fables, and the noble king, probably Henry II, for whom she wrote the lais. Her career spanned several decades. The poet Denis Piramus, believed to be writing in the 1170s or even earlier, refers to her lais and their popularity scornfully, and a reference in the Espurgatoire to Saint Malachais, canonized in 1189, provides a terminus a quo for her later work. Her lais and fables suggest that she was well educated and apparently literate in Latin, at home in the court milieu, not at times unworldly in her attitudes (but also interested in the life of a convent), and proud of her success as an author. This description might fit, among others, the abbess of Shaftesbury, the abbess of Ramsey, the countess of Boulogne, and the daughter of Waleran II, count of Meulan (now often considered the most likely candidate).⁵ But it has also been suggested, precisely because of the manuscript’s provenance, that Marie may have been a nun at Reading, perhaps even its abbess. If Marie had been at Reading in the late twelfth century, it would be quite in order that about half a century later the abbey should acquire or make another copy of her lais, either because its members happened to have an earlier copy to work from, perhaps even her autograph, or because they were proud of the connection, or both. This would mean that the historical reader of Harley 978 was very close indeed to Marie and might even have been her successor. The images in the lais of educated women, independent of spirit but often painfully immured, would be matched in the real world, where writer and reader walked the same cloister. Admittedly, most scholars of Marie de France seem uneasy at the identification of Marie as nun or abbess of Reading, but no one, to the best of my knowledge, has dismissed it out of hand.

Somebody should have. There is a very simple reason why Marie cannot have been the abbess of Reading: by her day there was none. Before the Conquest there had been at least one and possibly two nunneries in Reading, but when the abbey was reestablished by Henry I in 1121 it was as an all-male house. There were no nuns and therefore no abbess.⁶

If this were simply a matter of a single critic advancing an ill-founded argument, it would be less alarming. But the failure to challenge this hypothesis reflects a general lack of historical information among literary critics about the texts’ circulation. We are confronting a disciplinary gap. One group of scholars reads Anglo-Norman lais and another reads English ecclesiastical history, and the two remain in splendid isolation. As a result, modern scholars are a long way from understanding anything about the milieu of one particular reader, or group of readers, of Marie’s work. Critics who do not know that there were no nuns at Reading after the Conquest probably do not know very much about Reading at all. They will not know that it was very nearly dissolved for bankruptcy in the 1280s or that the bishop chastized one of its dependent priories for keeping hunting dogs and birds of prey or that it was a center for avant-garde music or that one monk ran away from the abbey and joined a gang of brigands. The word Reading will not conjure up a detailed vision of a specific place for them, as it did for Jamieson Hurry, whose popular histories are always well illustrated (see fig. 10 below). In the commentary on Marie de France, Reading is merely a tag on which to project stereotypes of monasticism.

So far, I might feel warranted to write in a tone of moral indignation. But as we pursue the variety of Harley 978, it will become apparent that blunders of this kind will be very difficult to avoid. Doubtless I have made many, just as I have in my efforts to transcribe glosses from the Digby Timaeus or Royal 10. E.4. The sheer range of material, from medical texts to hawking manuals and from musical pieces to political satires, will defeat any single scholar. The lesson to be drawn, it seems to me, is that as medievalists we need to establish protocols for much more extensive collaboration, a question I shall return to in the final chapter.

The challenge of these manuscripts is not just the range of the materials, however, but the tensions or hostilities they evoke. The gap between the hawking manual, a guide to a particular form of erotically charged conspicuous consumption, and the Song of Lewes, a panegyric for a saintly Christian warrior, is but one example. The problem becomes most acute with Royal 10. E.4. This massive legal compendium demands a significant study of canon law from anyone who hopes to read it. My own rudimentary effort, as I piece my way through a single passage, draws heavily on the assistance of friends and on preliminary course work that I did years ago as part of a conservative training offered by the Pontifical Institute for Mediaeval Studies in Toronto, whose scholars expended much time on students such as myself who must often have seemed mere dilettantes. The Institute’s commitment to meticulous and traditional scholarship, to Latinity, and to a vision of the Middle Ages as a coherent period defined by certain well-recognized intellectual traditions was a good match for the demands of texts like those in the Royal manuscript. For many, the study of canon law is sustained by a vision of rational, and ultimately benevolent, order imposed upon human chaos or, to quote the title of a study by one of its most distinguished scholars, a vision of harmony from dissonance. The marginal images in Royal 10.E.4 are, however, fundamentally at odds with these values. The bottom margin in particular offers a marvelous comic strip that runs the length of the manuscript, switching from one story to another and drawing on romances, fabliaux, saints’ lives, and miracles of the Virgin. These images seem to celebrate dissonance and the resistance to higher authority, as do many of the best modern readers of marginalia. It would be far too reductionist simply to equate the legal text of the Decretals with authority and rational order and the images with populist resistance and the unconscious, but the tensions between the two do run somewhat along such lines, and scholars committed to one have so far, for the most part, had little to say about the other. If a proper investigation of Digby 23 would require at least two scholars, one specializing in medieval philosophy and one in the chansons de geste, a proper investigation of Harley 978 would require a team, and a proper investigation of Royal 10.E.4 would require at least two scholars who were in fundamental disagreement on matters of principle. There may, then, be an important sense in which Royal 10.E.4 is unreadable in the modern world, for no single modern person will be able to embrace the contradictions it contains.

The material has imposed certain demands that may sometimes burden or irritate my readers, and I ask in advance for their patience. It has seemed to me important to capture as much information as possible about specific human beings known to have used these books. I do not wish to imply that the ones I have managed to discover, William of Winchester, who owned Harley 978, or Henry of Langley’s friend, who owned Digby 23, hold the ultimate clue to what these works really mean or that they are in any sense definitive readers. But they were readers. And if more information survived and we could find other readers, even if we could find earlier readers from the right century or the right social group—twelfth-century knights and ladies for twelfth-century chivalric poems, for example—they too, if we knew them one by one, would prove no less idiosyncratic and elusive. It is the contact between these messy people and the more rarified order offered to them in books that I wish to explore. So I have pursued my Williams and my Henrys. My reader must suffer through a good deal of biographical minutiae, labored efforts to reconstruct lost chronologies, and a litter of words like probably, maybe and perhaps and may still think at the end that the links between the books and the people are tenuous, the description of their reading patterns alarmingly speculative.

Second, this study is painfully incomplete. I offer a good deal of information about each manuscript and explore some of the material at fair length, but my treatment is partial, in both senses of the word. I have but a little to say about the history of glossed copies of the Timaeus. I only touch on the medical and Goliardic texts in Harley 978. I cover only some of the marvelous marginal stories in Royal 10. E.4. Nor, in general, do I offer the full, detailed textual analysis that is the glory of modern literary criticism. What I have tried to suggest is how a given collection of texts might have taken meaning in the mind of a particular reader, a real person, at a given moment. As part of this approach, I have explored the different modes of reception that might have been available and most readily brought to bear upon each manuscript: minstrel recitation, chant, or refectory reading for Digby 23(2), silent reading and fantasization for Harley 978, scholarly consultation for Digby 23(1) and Royal 10. E.4. Such an exercise is, I think, a useful contribution to cultural history and one that has considerable bearing on how we choose to read medieval texts today. But it is not a substitute for sustained close readings; it is perhaps at best a powerful disruption.

The internal diversity of these manuscripts also creates stylistic problems. Moving from one genre to another, and from one discipline to another, I have shifted tone and acknowledged different levels of proof. The voice used to discuss whether William of Winchester commissioned all of Harley 978 from the booksellers of Oxford and the voice used to discuss how he might have read one of Marie’s lais cannot really be the same. Once more, the conclusion I draw from this is that for some kinds of scholarly project, including most of those that might wish to be considered historicist, single authorship has severe limitations.

The field of manuscript studies has often been seen as an intensely conservative one, not least by its practitioners, who are much given to presenting it as a bastion of certainty against the rages of modernity and the over ingenuity of literary critics. But this is not how I would choose to justify my interest. The three manuscripts I examine offer not some absolute origin but rather a testimony to the complexity of textual production and a measure of the difference between our cultural categories and those of earlier times. They offer us, too, a measure of the gulf between the lives of medieval people and the roles their culture assigned them, whether as those who fought, those who worked, or those who prayed. By preserving traces of the activities of actual readers, who often did a little of all three, the manuscripts take us back to the complexities of human behavior and human desire, bringing us not firm answers but new questions.

Occluding the Material

In the last two decades there has been a renewed attention on the part of philosophers, historians, and literary and cultural critics to the material state in which texts are preserved and disseminated.⁷ Once largely relegated to an ancillary discipline whose obscure calculations could be dispensed with the moment it had fulfilled its duty and produced an accurate version of the author’s final intention, editing is now widely recognized as a field in which the historical construction of a work is brought to light. Texts, it is argued, exist only in specific material forms, or, to borrow Chartier’s phrasing in the epigraph at the beginning of this chapter, they exist only in specific kinds of material support. Although the phrase material support is cumbersome, it has the advantage of leaving as open as possible the question of exactly what this support is. Material support can refer to a good deal more than just the physical book. It might, for example, be applied to the sounds of a text that is sung or to the singer’s voice. It is to the physical book, however, that the term has most often been applied, and it makes sense to begin here, with the argument that the precise physical form of a particular manuscript or edition is a vital part of any given text’s meaning and social function.

An early and influential statement in the field of print bibliography is that of D. F. McKenzie, who compares two early editions of William Congreve and argues that it is impossible to divorce the substance of the text on the one hand from the physical form of its presentation on the other.⁸ According to McKenzie, the physical format of the 1710 three-volume octavo collected works, which was printed under Congreve’s personal supervision, provided fundamental evidence of Congreve’s vision of himself as a respectable neo-classical author. When in the same article McKenzie explained that the original 1678 edition of Pilgrim’s Progress was a duodecimo, set in pica roman, to a measure of only 14 ems and therefore had a short prose line suited for the less literate reader, he demonstrated how the fine detail of textual criticism could feed into social history.⁹ Historians have advanced similar arguments for the social significance of particular formats.¹⁰ Robert Mandrou’s study of the bibliotheque bleue established a fundamental link between the material support—in this case, cheap pamphlets suitable for sale by peddlers—and its social dissemination.¹¹ Robert Darnton’s work on booksellers’ lists and indices of proscribed books in eighteenth-century France established a surprising connection between political radicalism and pornography, both falling under the heading of libertinage and frequently being sold and condemned together.¹² The effect of cheap printed forms such as serials or mass-market paperbacks on popular reading patterns during the last two centuries has been investigated extensively.¹³

These claims for the fundamental importance of the material support are neither trivial nor mere commonsense; indeed, they represent a major disruption of certain fundamental assumptions subtending much of the close reading of literature and even the very category literature itself. Social bibliography, history of the book, textual materialism—these overlapping approaches all call into question the self-contained, self-referential, and stable literary artifact, whether the well-wrought urn of New Criticism or the closed semiotic system of structuralism. Thus Jerome McGann objects to the contemporary fashion of calling literary works ‘texts’ on the grounds that it "suggests that poems and works of fiction possess their integrity as poems and works of fiction totally aside from the events and materials describable in their bibliographies This usage of the word text does not at all mean anything written or printed in an actual physical state; rather, it means the opposite: it points to an Ur-poem or meta-work whose existence is the Idea that can be abstracted out of all concrete and written texts which have ever existed or which ever will exist."¹⁴ Roger Chartier’s rejection of the ideal, abstract text cited above runs along similar lines.¹⁵

If the ideal text is stable and unique, the material text is multiple. So far the implications of this insight for literary criticism have had perhaps their greatest impact on the study of Shakespeare. Bibliographic minutiae, once valued as evidence from which one could reconstruct a stemma and recapture what Shakespeare actually wrote, have now become a mark of textual multiplicity. As Margreta de Grazia and Peter Stallybrass observe: "For over two hundred years, KING LEAR was one text; in 1986, with the Oxford Shakespeare, it became two; in 1989, with The Complete King Lear 1608–1623, it became four (at least). As a result of this multiplication, Shakespeare studies will never be the same."¹⁶ De Grazia’s and Stallybrass’s collaboration with Randall McLeod/Random Cloud/Random Clod demonstrates how apparently minor bibliographic details can problematize the categories of author, character, and work. If we return to the early printings of the folio and the quartos, we find no fixity but instead an almost scribal fluidity in which the famous weird sisters of Macbeth are more often wayward and the very identity of works such as Lear or Hamlet is in question (a fluidity McLeod extends to his own name). The bibliographic details of the early printings ultimately bring us back from the solitary genius immanent in the text and removed from the means of mechanical and theatrical reproduction to the complex social practices that shaped, and still shape, the absorbent surface of the Shakespearean text¹⁷ While strong claims for traditional recensional editing can still be made, the ultimate goal of recapturing a single authorial origin, whether of Shakespeare or anyone else, is increasingly recognized as a chimera. Fredson Bowers’s hope that in the case of Shakespeare in the end, one man may be able to digest the widely assorted technical and critical problems and unify them into a single great work of scholarship, leaving us a text as close as mortal man can come to the original truth, seems to belong to another age.¹⁸

A similar confluence of conservative editing (which values the unique qualities of each manuscript rather than attempting reconstruction of a lost original) with structuralist and poststructuralist critical theory has flourished in romance philology.¹⁹ Here the crucial early study is Paul Zumthor’s famous Essai de poétique médiévale of 1972. Zumthor argues that the high degree of variation between manuscript copies is an essential quality of the medieval vernacular tradition and that the differing versions of a medieval poem, whether minstrel recitations or manuscript copies, should be seen not as corruptions of one original true version but as part of a continual process of recreation and modification he terms mouvance.²⁰ This view has led to a new respect among manuscript scholars for the work of individual scribes, glossators, and correctors.²¹ Since manuscripts are inherently more open to alteration than printed books, they are also more likely to be polyvalent or dialogic, so that diverse forms of representation, both of text and image, may be enclosed within a single copy.²² Stephen Nichols thus sees the manuscript matrix as one that brings together heterogeneous or even conflicting systems of representation:

Recalling that almost all manuscripts postdate the life of the author by decades or even centuries, one recognizes the manuscript matrix as a place of radical contingencies: of chronology, of anachronism, of conflicting subjects, of representation. The multiple forms of representation on the manuscript page can often provoke rupture between perception and consciousness, so that what we actually perceive may differ markedly from what poet, artist, or artisan intended to express or from what the medieval audience expected to find.²³

The challenges posed by this conflicted multiplicity will be one of the recurring themes of this book.

The recognition that an early text exists diachronically in the different layers of its copying and glossing can also be extended to cover the text’s printed history. The tendency had long been to see the printed edition as a reproduction of the manuscript’s text, one that was either neutral or, as Didron argued, deficient. More recently we have come to recognize that the poem we read is in significant ways the product of its editorial history. By stabilizing the textual tradition and isolating literary texts from the diversity of their earlier circulation, traditional textual editing has produced an origin for vernacular literature. It has excerpted texts from their codices in accordance with generic categories that are central to Romantic philology, concentrating on those vernacular texts that most readily conform to the category of literature, secular poetry expressing the genius of a people and the creative imagination of the artist. Finally, it has grouped these works together around categories of authorship that often differ significantly from those of their original makers, whether poets, compilers, or scribes.

Unless we were to revert to the world of eighteenth-century antiquarians like Thomas Tyrwhitt, one of the first editors of Chaucer, who appears to have read the entire Roland in the Oxford manuscript just to cull information on medieval literary traditions, what we read when we read a medieval poem will be some form of printed edition—and the form matters. Medieval poetry has been shaped into modern literary canons through the visual design and interpretive apparatus of modern editions. Taking as his example the Vie de Saint Alexis and comparing various editions to the illustrated manuscripts, Michael Camille demonstrates how nineteenth-century philologists erased all aspects of enactment—sound, sight, and sense from poems: Carefully classified blocks of print and their footnoted apparatus, together with clearly demarcated beginnings and endings, remade texts written in the twelfth and thirteenth centuries into nineteenth century intellectual commodities.²⁴ The choice of titles, the connotations of different fonts, the treatment of illustrations and musical notation, as well as the layout—all these details of print bibliography are therefore of concern