The Cambridge Companion to Textual Scholarship 052151410X, 9780521514101

Citation preview

THE CAMBRIDGE COMPANION TO

TEXTUAL SCHOLARSHIP EDITED BY

NEIL FRAISTAT and JULIA FLANDERS

cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao ˜ Paulo, Delhi, Mexico City Cambridge University Press The Edinburgh Building, Cambridge cb2 8ru, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521730297  c Cambridge University Press 2013

This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2013 Printed and bound in the United Kingdom by the MPG Books Group A catalog record for this publication is available from the British Library Library of Congress Cataloging in Publication data Cambridge companion to textual scholarship / edited by Neil Fraistat and Julia Flanders. p. cm. Includes bibliographical references and index. isbn 978-0-521-51410-1 (hardback) – isbn 978-0-521-73029-7 (pbk.) 1. Criticism, Textual. 2. Intertextuality. I. Fraistat, Neil, 1952– II. Flanders, Julia. p47.c36 2012 2012027122 801′ .959 – dc23 isbn 978-0-521-51410-1 Hardback isbn 978-0-521-73029-7 Paperback

Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

CONTENTS

List of illustrations List of contributors Acknowledgments

page vii x xi

Introduction: Textual scholarship in the age of media consciousness neil fraistat and julia flanders

1

1 A history of textual scholarship david greetham

16

2 Anglo-American editorial theory kathryn sutherland

42

3 Continental editorial theory geert lernout

61

4 Late twentieth-century Shakespeares hans walter gabler

79

5 Apparatus, text, interface: how to read a printed critical edition paul eggert

97

6 The politics of textual scholarship michelle r. warren

119

7 Fearful asymmetry random cloud (randall m cleod)

135

8 What is a book? roger chartier and peter stallybrass

188

v

vi

contents

9 Orality john d. niles

205

10 Manuscript textuality michael g. sargent

224

11 Picture criticism: textual studies and the image kari kraus

236

12 Tracking the changes: textual scholarship and the challenge of the born digital matthew g. kirschenbaum and doug reside

257

Coda: Why digital textual scholarship matters; or, philology in a new key jerome mcgann

274

Guide to further reading Index

289 301

ILLUSTRATIONS

Chapter 1 1.1 Typical stemma or “family tree” of textual transmission, showing the putative archetype, inferred (Greek sigla) and extant (Roman sigla) witnesses. Diagram by author.

page 30

Chapter 7* r John Milton, Paradise Lost, ed. Richard Bentley (London, 1732).

r

r

r

r

Courtesy of the Thomas Fisher Rare Book Library, University of Toronto (e–10 02647, c. 1). Line drawings (passim throughout chapter) illustrating various aspects and consequences of book binding: dog-earing, quiring, sewing quires to cords, knotting of thread, folding of sheets for binding, offsets within a book and between books, cancellation, stacking and rearranging of quires, placement of setting-off sheets, and kettle stitches seen spine-side; designed by random cloud, produced by Brandon Besharah. Denis Diderot, Encyclop´edie (Paris, 1751–65). Courtesy of the Fisher Rare Book Library, University of Toronto (rous.E57 1751 ovs; Plates, Vol. viii, 1771, “Relieur,” Planche 1). John Milton, Paradise Lost, ed. Richard Bentley (London, 1732), pp. 199 and 202 mutually setting off. Courtesy of The William Andrews Clark Memorial Library, University of California, Los Angeles (fpr3560.a1 1732*). Denis Diderot, Encyclop´edie (Paris, 1751–65). Courtesy of the Fisher Rare Book Library, University of Toronto (rous.e57 1751 ovs; Plates, Vol. viii, 1771, “Relieur,” Planche 1).

134

141

149

158

* The illustrations in Chapter 7 are unnumbered. This list includes those illustrations requiring acknowledgment; all other diagrams and illustrations appearing in the chapter are the author’s own.

vii

viii

list of illustrations r John Milton, Paradise Lost, ed. Richard Bentley (London, 1732),

r

r

r

r

r

r

r

r

r

r

r

r

S3v, showing offset from Torquato Tasso, Gerusalemme Liberata (London, 1724), Vol. i, n1r. Courtesy of the British Library Board (83.k.23). Torquato Tasso, Gerusalemme Liberata (London, 1724), Vol. i, Vita, pp. 28–9 (d2v–3r) setting off on the poem, pp. 300–1 (2q2v–3r). Courtesy of the British Library Board (80.h.17). Ouroboros modeled on an illustration in Codex Marcianus 299, Biblioteca Marciana, Venice, 188v. Designed by random cloud, produced by Brandon Besharah. Ouroboros modeled on an illustration in Codex Marcianus 299, Biblioteca Marciana, Venice, 188v. Designed by random cloud, produced by Brandon Besharah. John Milton, Paradise Lost, ed. Richard Bentley (London, 1732), showing untrimmed parts of bifolium 2Z2v–3r. Courtesy of the Folger Shakespeare Library (214–16f). John Milton, Paradise Lost, ed. Richard Bentley (London, 1732), showing details of the sewing of 2G2v–3r. Courtesy of the Folger Shakespeare Library (214–16f). Sewing and beating books, Ren´e Martin Dudin, L’art du relieur: doreur de livres (Paris, 1772), Plate viii. Courtesy of Toronto Public Library (655.7 d77). Sewing and beating books, Ren´e Martin Dudin, L’art du relieur: doreur de livres (Paris, 1772), Plate viii. Courtesy of Toronto Public Library (655.7 d77). Noel Rooke’s illustration of the bookbinder’s posture in relation to the sewing frame, from Douglas Cockerell, Bookbinding and the Care of Books (New York: Appleton, 1903), Fig. 29. Courtesy of the Fisher Rare Book Library, University of Toronto (duff 03611). John Milton, Paradise Lost, ed. Richard Bentley (London, 1732), showing details of sewing on C2v–3r. Courtesy of the Folger Shakespeare Library (214–16f). Dog-ears in Domenico Delfino’s Sommario di tutte le Scienze (Venice, 1621). With permission of the Centre for Reformation and Renaissance Studies, University of Toronto (b 785 d43S6 1621). Detail, Ren´e Martin Dudin, L’art du relieur: doreur de livres (Paris, 1772), Plate viii. Courtesy of Toronto Public Library (655.7 d77). Kettle stitches in Domenico Delfino’s Sommario di tutte le Scienze (Venice, 1621). With permission of the Centre for Reformation

158

160–61

164

166

168

170

171

171

172

172

173

174

list of illustrations

r

r

r

r

r

r

and Renaissance Studies, University of Toronto (b 785 d43S6 1621). A meeting in the gutter of the felt side of one sheet and the mold side of another, illustrated from English Experience facsimile no. 95 (Amsterdam: Theatrum Orbis Terrarum; New York: Da Capo Press, 1969): John Florio, Florio his first fruites (London, 1578), I2v–3r. Courtesy of the Robarts Library, University of Toronto (pc 1121 .f5 1578aa). Undulating cross-section of the fore-edges of two copies of Ian MacEwan’s On Chesil Beach, 2007, 2008. Author’s copies. Offsets on the title page of Valerius Maximus, Factorum et dictorum memorabilium (Lyon, c. 1503). Courtesy of Department of Special Collections, Charles E. Young Research Library, UCLA (z 233 .a41 v235 1503 [i.e., c. 1]). Denis Diderot, Encyclop´edie (Paris, 1751–65). Courtesy of the Fisher Rare Book Library, University of Toronto (rous.e57 1751 ovs; Plates, Vol. viii, 1771, “Relieur,” Planche 1). Offset of an unidentified musical psalter in Torquato Tasso, Gerusalemme Liberata (London, 1724), 2L1v. Courtesy of the British Library Board (80.h.17). Ouroboros modeled on an illustration in Codex Marcianus 299, Biblioteca Marciana, Venice, 188v. Designed by random cloud, produced by Brandon Besharah.

ix

175

180 182

184

186

186

187

Chapter 11 11.1 Rembrandt, Faust (c. 1652). Ink on paper/Dry-point, etching, and engraving, 21 × 16 cm. Public domain image. 11.2 Frederic Bartlett, experiment in serial reproduction (1932). Copyright Cambridge University Press. ¨ 11.3 Albrecht Durer, selections from Vier Bucher von menschlicher ¨ Proportion (1528). Public domain image courtesy of the US National Library of Medicine. 11.4 D’Arcy Wentworth Thompson. A cartesian grid transformation of Argyropelecus olfersi into Sternoptyx diaphana (two related species of fish), from On Growth and Form (Cambridge University Press, 1961). Copyright Cambridge University Press.

243 249

251

252

LIST OF CONTRIBUTORS

´ ´ des Hautes Etudes r o g e r c h a r t i e r , University of Pennsylvania and Ecole en Sciences Sociales, Paris p a u l e g g e r t , University of New South Wales j u l i a f l a n d e r s , Brown University n e i l f r a i s t a t , University of Maryland h a n s w a l t e r g a b l e r , Ludwig-Maximilians-Universitat ¨ Munchen ¨ d a v i d g r e e t h a m , CUNY Graduate Center m a t t h e w g . k i r s c h e n b a u m , University of Maryland k a r i k r a u s , University of Maryland g e e r t l e r n o u t , University of Antwerp j e r o m e j . m c g a n n , University of Virginia r a n d a l l m c l e o d ( “ r a n d o m c l o u d ” ) , University of Toronto j o h n n i l e s , University of California, Berkeley and University of

Wisconsin – Madison d o u g r e s i d e , New York Public Library m i c h a e l g . s a r g e n t , CUNY Graduate Center and Queens

College, CUNY p e t e r s t a l l y b r a s s , University of Pennsylvania k a t h r y n s u t h e r l a n d , St. Anne’s College, Oxford University m i c h e l l e r . w a r r e n , Dartmouth College

x

ACKNOWLEDGMENTS

Our greatest debt is to Linda Bree, our editor at Cambridge University Press, who commissioned this book in the belief that Textual Scholarship should have a Companion of its own. We are grateful to her continued support and advice as the collection has evolved and that of her colleague at the Press, Maartje Scheltens, who has graciously and speedily replied to the multitude of questions we have sent her way. We are also deeply indebted, of course, to our contributors for their enthusiastic participation in this collection and for the inspiring intellectual depth those contributions display. Neil Fraistat would like to thank additionally his colleagues at the Maryland Institute for Technology and the Humanities (MITH), who keep him on his textual toes (as it were), particularly Matt Kirschenbaum, Doug Reside (now at the New York Public Library), and Dave Lester, who on an almost daily basis have helped him better understand the rapidly changing nature of textuality in a media conscious age. As ever, Neil is also deeply grateful for the unfailing support and encouragement of his family, especially of his marvelous wife Pam Wessling and those award-winning playwrights, Shawn and Ann Fraistat. Julia Flanders would also like to thank her colleagues, past and present, at the Center for Digital Scholarship, whose myriad talents and interests are a constant inspiration, and especially Allen Renear, who first pointed her to McKerrow, Bowers, and Tanselle as crucial forebears for the digital text theorist.

xi

NEIL FRAISTAT AND JULIA FLANDERS

Introduction Textual scholarship in the age of media consciousness We are entering a great age of editing. – Jerome McGann, 1997 We see before us a great age [of editing] – indeed, a heroic age, one filled with triumphs and false starts, messy, destabilized and destabilizing, and above all, dynamic. – Greg Crane, “Give us Editors! Re-inventing the Edition and Re-thinking the Humanities,” 2010

The role of the scholarly editor is too often assumed to be best performed when least visible: when the editorial work has produced a seamless artifact from which its own traces have been effaced. But as Greg Crane suggests, we can situate this work differently if we understand it as constitutive – rather than merely curatorial – of textual culture. If we are indeed standing at the threshold of a “great age of editing,” it is because we now understand the crucial importance of precisely how texts get mediated, and because the importance of textual scholarship to the humanities disciplines has never been clearer. The growth of interest in digital media over the past two decades has of course contributed substantively to our heightened awareness of medium as a methodological question. The process of theorizing the digital representation of objects of study has also required us to attend to the meanings of their original medium. But other influences have been crucial as well: the renewed attention to textual materiality spurred by Jerome McGann’s and D. F. McKenzie’s groundbreaking work, accompanied by interest in the history of the book and the politics of textual production, all of which have rendered our understanding of authorship and textuality immeasurably more complex.1 Crane’s evocation of both the heroism and the dynamic, risky messiness of editorial work responds to both this significance and this complexity: if the terrain seems to be populated chiefly by challenges and problems, it is because there is so much at stake. If the New Bibliographers at the turn of the twentieth century opened up a set of novel and pressing questions which appeared to be substantially resolved by 1

2

neil fraistat and julia flanders

the comprehensive and magisterial work of McKerrow, Greg, Bowers, and Tanselle, the turn of the twenty-first century has brought us a fresh agenda that, as McGann and Crane both suggest, will take decades to address. Our primary goal in this Companion to Textual Scholarship is both to suggest the scope of that agenda and to provide a grounding in its foundational issues for those who will be carrying out these new inquiries. The essays in this volume provide a clear, engrossing, and accessible introduction to the central topics animating the field of textual scholarship. Through their multiple perspectives they demonstrate the centrality of textual scholarship to current literary studies of all kinds, and express for those new to the field the intellectual excitement of a crucial scholarly discipline. There have been excellent introductions to the field by David Greetham (Textual Scholarship: An Introduction, 1992) and by William Proctor Williams and Craig S. Abbott (An Introduction to Bibliographical and Textual Studies, 1999).2 Both of these provide exemplary overviews of such traditional subfields as analytic bibliography, descriptive bibliography, historical bibliography, and textual criticism, which they consider at a level of comprehensiveness and practical detail that make them obligatory reading for the new textual scholar. Our complementary goal in this collection is to present more focused and critically probing explorations of the history, significant debates, and important conceptual dimensions of the field, charting not just the past and present, but also its future directions. For reasons of space we focus chiefly on the Anglo-American traditions of textual scholarship with an emphasis on textual criticism, but we situate these traditions in a broader context with respect both to geography and domain. This broader view reflects the importance of the field within humanities scholarship more generally. If, during the height of New Bibliography and thus for most of the twentieth century, textual scholarship was widely perceived as being merely a precondition for the more serious work of critical interpretation, it has now itself fostered new modes of critical reading and interpretation that focus upon the materiality, production, transmission, and reception of texts. Pages and screens are also now being read as quantum fields in which the meaning of words are interdependent with the graphic elements in which they are embodied, surrounded, and displayed. Textual scholarship, that is to say, is no longer best considered as a specialized field of concern only to scholarly editors and bibliographers, but rather as a domain fundamental to literary and media studies of all kinds, one that is not separate from contemporary theoretical concerns but inextricably connected with them. We are now, moreover, in the midst of a golden age of editorial theory. Both new and familiar forms of editions – eclectic, versioning, genetic, variorum – are finding fresh expression under the influence of the electronic

Introduction

3

medium, which is transforming editorial theory and practice. Similarly, the emergence of new objects of editing – such as images, born digital literature, film, performance art, and digital games – is prompting a reexamination of editorial methods that are adapted to more purely textual materials. These new “texts” are both the products and instruments of change; the emerging techniques they demand highlight the fact that editions are themselves theoretical constructs, foundational forms of scholarly argument. For such reasons, Jerome McGann and Dino Buzzetti can legitimately claim in a recently published volume on electronic scholarly editing: “Scholarly editing is the source and end and test of every type of investigative and interpretational activity that critical minds may choose to undertake.”3 Written by an international group of prominent textual scholars, the essays presented in this volume attempt to do justice to the history, range, complexity, and vitality of the field. They are organized around two overarching sets of concerns, each of which unfolds in a set of six chapters, framed by this introduction and a coda. The first half of the volume explores the history and culture of textual scholarship, moving from large foundational issues about the nature of texts, textuality, and the disciplinary evolution of the field to more focused examinations of the key concepts, methods, and debates within and among the primary “schools” of editorial theory and practice. It concludes by considering how textual scholarship produces itself through the rhetoric of reading surfaces in editions, the construction of specific types of readership, and the political investments that underlie its work. The second half of the volume builds on this history by exploring specific textual modalities and the methods by which textual scholarship works to denaturalize them and reawaken our awareness of them as modes of signification. By considering the ways in which we read and study books, oral texts, manuscripts, images, and digital inscriptions, the latter half of the volume both demonstrates and reflects upon textual practice as deliberate interpretive strategy. As a group, then, these essays consider how textual scholarship has been shaped as a field over several centuries, how it may develop further in the future, and what it means, intellectually and politically, to engage in such work. The opening essay by David Greetham, “A history of textual scholarship,” establishes both the significance and the scope of textual scholarship through a historical overview stressing the cultural centrality of the field. Greetham tackles the question – crucial to orienting the reader of this volume – of what textual scholarship is and what it includes: how its boundaries are drawn and how they have come to take their present contours. For Greetham, textual scholarship is, broadly put, “the history of history.” Moving from the classical to the postmodern, he traces a field whose main paradigms were established in the opposition between the Platonic textual

4

neil fraistat and julia flanders

idealism of Alexandria – with its attempts through the concept of analogy to construct from versions known to be “corrupt” an “ideal text” with no known prior physical incarnation – and the philologically inflected Aristotelianism of Pergamon, which attempted through the concept of anomaly to reproduce the “best text” among necessarily flawed but historically incarnate versions. The tensions between these two positions have reverberated throughout the centuries, playing themselves out in the reproductions of classical and biblical texts that were the main focus of textual scholars until the beginning of the nineteenth century, as well as in the subsequent turn to vernacular literature and national canons that has characterized the past 200 years. If the twentieth century was dominated first by the attempt to make textual scholarship more rigorously “scientific,” and then by the debate over the “sociology of the text,” as launched by McKenzie and McGann, it ended very much consumed with the problems and possibilities of the digital, according to Greetham, who is perhaps more reserved about these possibilities than McGann. He situates the debate in its most extreme form as between “liberation technologists,” who proclaim the revolutionary nature of the digital medium, and textual conservatives, such at G. Thomas Tanselle, who strongly rebut such claims. For Greetham, the jury is still out. While Greetham’s essay seeks to establish the broad cultural significance of textual scholarship, the next two chapters by Kathryn Sutherland and Geert Lernout offer a detailed account of how the field has been shaped over the past two centuries in both the Anglo-American and Continental traditions of editorial theory. Sutherland’s account in Chapter 2, taking as its point of departure the New Bibliography of the early twentieth century, traces a sustained debate concerning the role of the editor in relation to history, authority, and meaning. While at the beginning of the century Anglo-American editors favored the kind of scrupulous belletrism displayed in R. W. Chapman’s 1923 Oxford edition of Jane Austen’s novels, a vigorous and successful challenge to this method was launched by such New Bibliographers as McKerrow, Pollard, and Greg, whom Sutherland describes as championing historical recension over humanistic interpretation, with an “unwavering concentration on physical evidence and the facts of transmission.” They thereby situated textual scholarship as “objective” and “scientific,” professionalizing the discipline. For Sutherland, a crucial distinction between British and American editorial theory and practice emerged in the 1960s with the idealizing methods of the Americans Fredson Bowers and later Thomas Tanselle, who transformed Greg’s “pioneering enquiry into historical processes” into a “forensic investigation into the text as the scene of a crime.” Fundamental differences in understanding the ontology of texts and their relation to history underlie these traditions: “On the British side of the divide, history is a persistent and messy intermediary between author

Introduction

5

and reader. On the American side, and less equivocally than Greg, Bowers was in the business of ideal authorial reconstruction” in which history was to be transcended. These philosophical distinctions take opposing textual form in the “clean look of the edited American page with critical apparatus tidied away at the back of the volume (the encounter of author and reader unsullied by history)” as opposed to continued “standard British editorial practice, where the record of variant readings, deposited like archaeological layers at the foot of each page of text, . . . providing both a historically compromised and a more Classical look” (see pp. 45–6). The differences between American and British theory and practice thus hark back to those between Alexandria and Pergamon as articulated by Greetham in the previous essay and include debates within both countries over such issues as the meaning and use of copy-text, the value and interpretation of authorial intention, and the relations between “text” and “work.” Sutherland views the New Bibliographical drive for producing an ideal composite text freed from errors as being complicated and challenged in the late twentieth century by Anglo-American debates touched off by McGann and McKenzie about the text as a collaborative and social product; by editorial interest in “versioning” rather than reproducing a single state of a text; and by the editorial possibilities offered by digital media, reminding us, ultimately, that “All editorial theories imply the authority to represent or speak as the text; and all are ultimately revealed as temporal and temporary protocols for interpretation” (p. 48). Lernout’s contrasting account of the variety of Continental traditions shows them as concerned much more with textual scholarship as a way of investigating national history: philology as a way of reconstructing a “national past” in the origins of language and culture. The editorial practices arising from this history emphasize textual transmission and revision, and the creation of stemma through which textual transmission maps a detailed historical narrative, which is to say that its foundation rests principally on the methodology pioneered by the German scholar Karl Lachmann in his exemplary nineteenth century edition of Lucretius’s De rerum natura and later systematized into what has been called the “Lachmann method.” For Lernout, because the philological practices of Lachmann and his followers did not distinguish in terms of editorial treatment between classical and contemporary authors, they led to “one of the most important German export products, the historical-critical edition of texts in the vernacular” (p. 67), a product imported, of course, into the Anglo-American tradition. Two key later developments in German editorial theory, however, are in marked contrast to the Anglo-American tradition: a rejection of the kind of eclectic mix of readings from different versions of a text that is fundamental to copy-text editing and a resistance to treating versions of a text

6

neil fraistat and julia flanders

teleologically as eventuating in a final privileged version. Both of these positions emphasize the integrity and value of each version of a text and found resonance in France, where they were preceded and influenced by Joseph B´edier’s attack on the Lachmann method and his call for medievalists to produce, in the words of Lernout, “a faithful edition of the ‘best text’ of a work, which had at least the advantage of being historically consistent, . . . written by one scribe, in one moment in time” (p. 70). Critique g´en´etique, or genetic criticism, which arose in France in the mid-1960s through the work of Louis Hay and Jean Bellemin-No¨el on Heine manuscripts, positions itself even more radically. Aligning with French literary theory, it is interested in e´ criture, writing as process, not product. It therefore eschews the traditional critical edition for a “genetic dossier” of a written work that includes all surviving documents of the writing process, each retaining its own integrity without any teleological privileging of “final” form. While compatible with anti-teleological forms of German editorial theory, French genetic criticism differs from them, as Lernout points out, in its lack of interest in editorial matters and in edited texts. While German and French theory and practice have played a dominant role in Continental textual scholarship, there is now burgeoning activity in the field throughout Europe in a variety of languages and countries that, while difficult at this point to summarize, amounts to a new widespread return to philology and an intensive engagement with digital media. Concluding the first half of the volume, three essays by Hans Walter Gabler, Paul Eggert, and Michelle Warren explore the professional and political meaning produced through the editorial theories and practices discussed in the preceding essays. While the texts of Shakespeare were the laboratory through which the New Bibliographers largely developed their theories and methods at the beginning of the twentieth century, towards the close of the century they were the focal point for key challenges to New Bibliography. Focusing on the debates in the 1980s over the text of King Lear and the controversial one-volume Oxford edition of the Works of William Shakespeare, edited by Stanley Wells and Gary Taylor, Hans Gabler in “Late twentieth-century Shakespeares” (Chapter 4) shows just how high the stakes of such arguments can be. The radical and often contradictory differences between the 1608 Quarto and 1623 First Folio texts of King Lear had long been reconciled by editorial conflations of the two into a single work, whose only theoretical warrant for New Bibliographers was that these differences were caused by corruptions in each text of a lost original upon which they were both based. As Gabler explains, “to admit that King Lear could have come down to us in two versions would have rocked the foundations of a school of textual criticism whose newly developed bibliographical methodology was seen as a triumphant re-enforcement of the stemmatic thinking

Introduction

7

of old in the discipline” (pp. 82–3). Those foundations were indeed rocked in 1983 with the publication of The Division of the Kingdoms, a volume of essays edited by Gary Taylor and Michael Warren, which forcefully argued that each of the two versions of King Lear had its own distinct integrity as a performance text and reflected revisions made by Shakespeare over time.4 As Gabler remarks, “By assuming that Shakespeare was a revising author like other authors, and recognizing that the material transmission of King Lear held substantial evidence to prove the contention, . . . the proponents jointly succeeded in bringing Shakespeare textual criticism in line with the discipline of textual criticism as internationally and trans-nationally conceived and practiced” (p. 84). At stake in The Division of the Kingdoms and in the Oxford Shakespeare published three years later, which edits the plays in versions that approximate performance texts, is a crucial argument about the very ontology of Shakespeare’s texts that emphasizes their theatricality and places variability at their core. That this argument emerged in significant dialogue with a concomitant turn in Shakespeare criticism towards exploring the plays as performance highlights the interdependence between textual and literary criticism, long denied or effaced earlier in the twentieth century. In Gabler’s view, textual editions are increasingly inviting critical engagement with themselves “as products of critical scholarship.” For him, the “prime question to be asked, be it in editorial or in critical terms, is no longer simply the normative one of ‘what is the work’s text?’, but rather – and in far greater complexity – the critical one of: ‘what is the range of variability of the work’s text(s), from which we may gain analytic advantage toward interpretation?’” (p. 95). As the number of Shakespeare editions proliferate for classroom use in all of their textual variability, they bring this lesson home in a fundamental way, inextricably linking textual scholarship with literary criticism. In discussing the difference between editions that through their architecture and layout present readers with a “bookish” Shakespeare as opposed to a “theatrical” Shakespeare, Gabler engages an important issue developed by Paul Eggert in “Apparatus, text, interface: how to read a critical edition” (Chapter 5). Scholarly editions face the perpetual challenge of how to present the text both legibly and informatively; by the same token, readers of editions face the challenge of how to interpret the often complex and recondite detail of scholarly apparatus. Debates over the proper presentation of apparatus make it clear that this textual surface is far from transparent, but is most significant for what it omits or condenses, and for how it persuades. While introducing readers to the often complex and sometimes befuddling machinery of the critical edition as it has evolved over time, Eggert views the reading interface created through an edition’s “internal organization, declared conventions, and page layout” as “the systematic embodiment

8

neil fraistat and julia flanders

of the editorial argument.” The perspective gained through a generation of electronic editions, he argues, has denaturalized the conventions of the print edition and allowed us to understand “text, apparatuses, tabulations, appendixes” not simply as interface elements, but also as expressions of “various beliefs about the nature of the work” (p. 99). Eggert’s comparison of the architecture of Anglo-American and German critical editions of the 1960s exposes the extent to which edition design is ultimately ideological, dependent on assumptions about the ontology of literary works. Whereas the Anglo-American editions were geared toward presenting the work as such, German editions were designed to emphasize textual variations between the various manuscript and print versions of the work, replete with elaborate apparatus strategies highlighting the integrity of these versions. Eggert’s turn to the strategies through which recent editions of Hamlet have been embodied demonstrates how the ideological commitments and implicit argument of a critical edition is materialized in its interface and enacted on its page, while at the same time returning us with a different vantage point to the key debates about Shakespearian textuality upon which Gabler focused. The use of Shakespeare by both Gabler and Eggert for focusing on the ideologies implicit in the construction of textual editions highlights the ability of textual scholarship to explain in specific, material ways the constructedness of all texts. It also enables textual scholars, as Michelle Warren argues in “The politics of textual scholarship,” to identify with “historical and cultural precision the power dynamics (both oppressive and resistant) that sustain language and text systems” (p. 119). Warren thus views textual scholarship as belonging within the larger ambit of politically inflected literary and cultural studies, “from the postcolonial to the queer, the ecological to the philosophical.” For her, the politics of textual scholarship broadly construed involves not only the production of texts, but their dissemination and consumption as objects of cultural analysis, encompassing “the national contexts of professionalized textual scholarship, power relations embedded in editorial forms, and ideological implications of the language used to describe editorial practice” (p. 119). Her discussion of the French adoption of The Song of Roland as both a national epic and anchor for national identity illuminates many of these issues, placing B´edier’s development of “best text” theory as part of this larger narrative, including the French need to overtrump the predominant German editions of Roland that were based on Lachmannian recension. Her attention to the various relations of power embedded in the material forms of editions not only echoes Gabler and Eggert, but highlights the importance of treating editions as objects that in themselves require critical interpretation. As she explains, even so-called unreliable editions “have important stories to tell about ideological appropriations and political histories” – including, for example, editions of the Early English Text Society

Introduction

9

circulated for colonial purposes in British India, or Mary-Lafon’s edition of the French epic Fierabras as an explicit aid to North African colonization. Warren also usefully directs our attention to the power relations encoded within the rhetoric of textual scholarship itself, which is founded on metaphors of “corruption,” “variation,” “correction,” and “normalization.” But what does not get said, nor done, nor archived is as powerfully political as what does. Warren’s questions about what we might learn about the methods and theoretical underpinnings of textual scholarship from editions with explicit political commitments to feminist, postcolonial, or gender theory serve as an important reminder at the conclusion of the first half of the volume, which details the history of the field, about how relatively absent such editions have been from that history. The second half of the volume begins with two very different investigations of the material “bookness” of books and the interpenetration of materiality and textual meaning. Randall McLeod’s essay “Fearful asymmetry” (Chapter 7) offers a detailed and materially grounded performance of analytical bibliography, working through minutely observed physical idiosyncrasies such as dog-earing and stitching patterns to trace what these reveal both about the production processes through which the book is constructed, and the kinds of intentions and meanings that we can uncover as a result. His understanding of book production as a stack of tightly interrelated activities, of which authorship is only one instance, establishes a revealing dialogue with Chapter 8 by Peter Stallybrass and Roger Chartier, in which this same set of concerns is explored from a historical perspective. In answering their opening question, “What is a book?,” Stallybrass and Chartier situate the physical and textual book in a complex web of legal, philosophical, literary, and political debates. Our definition of “book” emerges as a constructed, contested, and shifting notion which carries considerable cultural freight: for example, in the lawsuits over copyright in the early eighteenth century that created “new relations between notions of individual authorship, originality, and copyright” in which the author came to have, for the first time, ownership of his own work. This legal construction of authorship as property has proven culturally decisive, but as McLeod’s very different anatomization of the book suggests, there are other ways in which a much more attenuated sense of authorship penetrates the material artifact we handle, animating while not actually accounting for the significant physical configurations the analytical bibliographer reveals. If what these two essays, in their different ways, establish for us is in effect a denaturalization of the printed book as a cultural form and as a surface through which meaning is transmitted, the next two essays probe the cultural centrality of print from another angle by examining the alterity – from the standpoint of textual scholarship – of manuscripts and oral texts. Both

10

neil fraistat and julia flanders

kinds of texts challenge traditional methodologies and require, as John Niles in Chapter 9 says of orality, “alternative ways of thinking about such central editorial issues as transmission, variation, and authorship” (p. 220). Michael Sargent in Chapter 10, “Manuscript textuality,” discusses the history of early modern literary manuscripts, finding that the main constant of such texts is variance itself, what Paul Zumthor has termed “mouvance.” For Sargent, although text of any kind “has always had a centrifugal tendency toward variation,” manuscript text “is variable in ways that we do not commonly think of printed text as being: no two manuscripts of the same work are ever identical” (p. 228). Sargent finds that print editions have been overtaxed in attempting to capture the complex “mouvance” of manuscripts, a problem further exacerbated in the textual remediation of oral performances, in which the limits of print and even video to represent the complex, embodied, and interactive performances of oral events are all too apparent, as John Niles discusses in Chapter 9: “Orality.” With few issues about oral texts theorized by textual scholars and no agreed upon methodological procedures for editing them having been established, modern editions of oral texts rarely explain or rationalize their own editorial procedures, which often vary dramatically and, Niles claims, foster misconceptions involving “an overestimation of the inherent capacity of either script or print to simulate oral performance, together with an undervaluing of the impact of scribes, collectors, and editors on what is widely received as ‘oral literature’” (p. 205). As an example of some of these problems, Niles explores the only recently understood dialogic interplay between oral and written transmission of the various songs comprising The Song of Roland, in which any “variant text might display the signs of both scribal degradation and oral/aural modification.” Neither Lachmannian recension nor B´edier’s “best text” method are sufficient to address this level of hybrid variability. Like Sargent, Niles hopes that electronic editions will provide adequate means for addressing the kind of textual alterity they discuss, but that medium has yet to produce a panacea. Animating their essays is the question of how textual scholarship constructs its object as a text, and how the methods of textual scholarship may be applied to documents (using this term broadly) that in various ways push at the boundaries of our traditional understanding of “text.” The final two essays in the volume, Chapters 11 and 12, take this issue even further to consider how images and digital inscriptions may be usefully studied and interpreted using (or adapting) techniques arising from textual scholarship. Although the term “textual scholarship” seems to point as explicitly as possible towards textual artifacts and text-oriented methods, Kari Kraus’s “Picture criticism: textual studies and the image” (Chapter 11) argues for the crucial role images have to play in the textual ecology, and suggests how

Introduction

11

“picture criticism” can both draw on and provocatively extend techniques that emerge from textual study. As she observes at the outset, the “pictorial turn” that W. J. T. Mitchell noted in cultural criticism has also been evident in the editorial milieu, where the importance of figurality and of the visual and pictorial qualities of text and page are now common concerns. The question, though, if we now wish to think rigorously about text as image – or about images with the same kinds of rigor we commonly apply to texts – is how to adapt our techniques of comparison and collation, and our theories of variation so that they help us make sense of images as well. In exploring these questions, Kraus first examines the ways these techniques operate in their original textual environs, and then draws on theories of deformation, caricature, and evolution to suggest how variation between image versions might be understood and subjected to formal analysis. In “Tracking the changes,” Chapter 12, Matthew Kirschenbaum and Doug Reside point to one of the newest and most urgent challenges for textual scholarship, the proliferation of born digital literature and art. As they remark, “all literature today is born digital in the sense that it is composed on the screen, saved on a hard drive (or other electronic storage media), and accessed as part of a computer operating system,” the consequences of which are large: “this means that writers working today will not and cannot be studied in the future in the same way as writers of the past, because the basic material evidence of their authorial activity – manuscripts and drafts, working notes, correspondence, journals – is, like all textual production, increasingly migrating to the electronic realm” (p. 260). The changing materiality of the archive will present new opportunities for textual scholars, who might now, for instance, be able to study a writer’s computer itself as a creative environment, with its own specific hardware and software affordances that condition the production of a text. Study of the computer’s files can through timestamps provide the entire genetic development of a text, or situate that evolving text in the light of relevant email correspondence, browser history, iTunes library, and related use of social media. Such opportunities pose new technical challenges and ethical quandaries for textual scholars, who, as Kirschenbaum and Reside argue, will need to develop new sets of skills, tools, and practices to be equal to them, blending textual criticism with archival science and computer forensics. But, as they also argue, details “regarding file formats and media types are no more or less exotic from the purview of textual scholarship than the minute particulars of letterpress, paper making, and book binding” (p. 273). The new agenda for textual scholarship for which Kirschenbaum and Reside call is placed in larger historical context by Jerome McGann in the Coda to this volume. McGann argues in “Why digital textual scholarship matters” that we are now in the midst of a globalized turn from a “textual

12

neil fraistat and julia flanders

condition” to a “digital condition” in which our entire inherited cultural archive is being digitized and will require re-editing “within a network of digital storage, access, and dissemination.” In such an era of media transition and migration, where text appears in altered, multimodal forms and with radically different vectors of significance, textual scholarship is essential and needs to adapt rapidly while retaining its deep-seated philological concern with text as a materially and temporally situated carrier of meaning. McGann thus views the emergence of digital media in the late twentieth century as requiring a fundamental shift in scholarly perspective “back to the view of traditional philology, where textual scholarship was understood as the foundation of every aspect of literary and cultural studies” (p. 275). And nowhere is that shift more evident than in the changing forms of textual editions themselves, which, as McGann points out, have always served as models or theoretical instantiations “of the vast and distributed textual network in which we have come to embody our knowledge” (p. 283). For McGann, then, we stand at a moment when textual editions are both the products and mediators of sweeping changes in the very nature of textuality, the Necessary Angels of our Digital Condition, where the very instability of “text” highlights the crucial importance of textual scholarship. As McGann makes clear, our new Digital Condition requires the resources of textual scholars more than ever, and it will require from them ever more resourcefulness. Is the history sketched in these essays one of steady change or cyclical oscillation over a set of familiar questions? Certainly some of the oldest issues – for instance, concerning the relationship between material witnesses and some ideal or hypothesized original – are as vibrant and troubling as ever after centuries of debate. And some of the newest questions, such as how to exploit information about a writer’s digital working environment, reflect interests that have long animated textual study. At the same time, there are several points on which particular traction has been gained in recent years. The first of these is in an overt understanding of the politics of textual scholarship, and a kind of candor on the part of textual editors concerning their own interests and editorial commitments. The trend in early electronic editing towards a kind of editorial agnosticism – according to which an ideal “edition” would be a collection of unedited witnesses framed within an interface that would support readerly exploration – was an attempt to make the editor’s role both overt and minimal: to make readers more aware of the editorial process and to give them a more active role therein. This sensitivity to the politics of editing – politics which Michelle Warren illuminates at length in her essay in this volume – has continued even as electronic scholarly editing has swung back towards more editor-centric

Introduction

13

approaches. We can also see its broader effects in the growing interest in collaborative editing environments and in interfaces through which reader input can take the form not simply of private experimentation but of public contribution. Another newly emerging paradigm in the era of electronic editing is the practice of treating editorial debate (whether with oneself or with other editors) as something that can itself be formalized and incorporated into the edition’s data. There is of course nothing new in annotating difficult textual cruxes so as to describe the reasoning process behind an editorial decision. But the use of text markup (for instance, the TEI Guidelines) to create electronic editions makes it possible to represent the editorial work itself – decisions, uncertainties, forms of evidence, alternative possibilities and viewpoints – in a comprehensive way as part of the formal data of the edition. As a result, it can be incorporated into the edition’s interface and made visible to the reader: for instance, revealing passages about which there is significant debate, or decisions for which the evidence is contested. In principle, such an approach permits a single edition to represent the disparate views of many editors, allowing the reader to experience the very debatedness of the text as part of its presentation. This sense of textual plurality, of seeking proliferation rather than convergence of textual possibilities, is very much in keeping with the “n-dimensionality” that McGann observes in textual documents. What is at stake here is an increased transparency and an interest in what we might almost term “negative capability”: a desire to treat the edition not as a final resolved statement but as a field of intelligently structured possibility, whether the edition is produced in print or electronically. Important editions will continue to appear in both forms, of course, and new editorial developments in each are likely to interrogate and influence the other. If we follow the logic of editorial “negative capability” and project it onto the emerging landscape of digital editions, we arrive at a set of desiderata for the digital tools and information of twenty-first-century textual scholarship. They should be: 1 interoperable with each other and with other texts, objects, tools, maps, user-driven annotation, exhibits, mashups; to make this possible, they should be based on scholarly and technical standards. 2 layered and modular, so that content is separate from interface and so that interfaces can be designed and redesigned over time for multiple purposes and audiences 3 multimodal, providing strong analysis not only of text but of spatial, temporal, topical, and material information

14

neil fraistat and julia flanders

4 dynamic, designed to encourage interaction and manipulation 5 scalable, to permit meaningful inquiry at both macroscopic and microscopic levels 6 everted and interconnected, so that data from scholarly resources and editions can be remixed in other contexts, in both the scholarly community and by the general public 7 sustainable, through publishers, repositories, community will, community access to content, and community expertise. The essays in this volume, which share a broad commitment to readerly engagement, interrogation of material culture, and editorial transparency despite their widely varying subject matter, attest to the desirability of many of these directives. In framing an introductory volume of this sort for a large and varied field, one necessarily proceeds by constructing a coherent center that can serve as a manageable starting point – and hence by creating an implied peripheral space of complexity that escapes description. The consensus marked above is thus both real and a simplification, and the subject areas covered by this collection represent both a true and an incomplete statement of the contours of the field: of its current center of gravity but not its potential boundaries. However, readers who engage seriously with the field will want to explore the complexities that occupy what in this volume is constructed as periphery. How, for instance, might we approach the editing of performance, especially in cases where a material or informational record does not exist, or of performative textual forms such as sign language? What are the special challenges of time-based media such as audio, video, and film? At the other end of the continuum, how can we extend ideas of editing to material artifacts such as statues, buildings, or complex built spaces like towns or archaeological digs? Virtual environments and digital games, too, pose distinctive challenges and are now subjects of intense critical scrutiny for the challenges they pose to techniques of preservation, analysis, and study. And finally, there is much fuller justice to be done to the more specific, distinct politics of empire, gender, race, and class. We hope that this volume will orient readers to pursue these and other challenging issues in a crucial field that continues to offer us the record of our cultural memory and the future conditions of its intelligibility.

NOTES 1 See, for example, McKenzie’s Bibliography and the Sociology of Texts, The Panizzi Lectures (London: British Library, 1986) and McGann’s, A Critique of Modern Textual Criticism (University of Chicago Press, 1983).

Introduction

15

2 Greentham, Textual Scholarship: An Introduction (New York: Garland, 1992) and Williams and Abbott, An Introduction to Bibliographical and Textual Studies (New York: MLA, 1999). 3 L. Burnard, K. O’Brien O’Keeffe, and J. Unsworth, eds., Electronic Textual Editing (New York: MLA, 2006), 54. 4 Gary Taylor and Michael Warren, eds., The Division of the Kingdoms (Oxford University Press, 1983).

1 DAVID GREETHAM

A history of textual scholarship

The role of textual scholarship Before outlining the history of textual scholarship, it is necessary to define its role in scholarship and criticism as a whole. At one point, it used to be thought that the work of textual scholars was primarily editorial (or at least proto-editorial). That is, the examination of extant documents and their lost but inferrable sources was supposed to determine accurate and reliable texts for the use of critics, historians, and all others interested in having at their disposal such a text or texts that could then be used in other, loftier pursuits. This distinction was the basis of the traditional division between the lower criticism, whose responsibilities were confined to the nuts and bolts of determining reliable readings (according to what sort of text was thought necessary or desirable) and the higher criticism, which then took these texts and subjected them to hermeneutics or “interpretation,” including historical and aesthetic commentary, critical annotation, and so on. This basic distinction still holds in some disciplines, especially classical and biblical scholarship, which were the first two areas in which textual analysis was practiced. But in the last 100 years and more, the division between higher and lower criticism, especially in vernacular literatures, has been challenged, and textual scholarship has been increasingly regarded as inevitably having a critical component, most evidently in that sub-division known as textual criticism, but also in such apparently technical (and thus supposedly “noncritical”) activities as selecting a document as the basis for a facsimile edition. The intersection and interrelationship of “science” and “criticism” in textual scholarship forms one of the major themes of the history of the discipline in the body of this essay. Nonetheless, it is still an institutional, academic, and popular reality that for many critics and readers, textual scholarship is seen at best as a “service” activity, providing the rest of the scholarly community with an authoritative text or texts, in single or multiple forms, for use in criticism and in the classroom. This is by no means an insignificant

16

A history of textual scholarship

17

responsibility, but perhaps especially in the electronic age, when textual “archives” are sometimes dismissively seen as simply the uncritical uploading of print or “born digital” texts on to a website, the essential critical components of selection, evaluation, emendation, and annotation of texts still need to be emphasized. All too often, even in graduate seminars, an instructor may declare that “any text will do,” even in a seminar featuring popular works with a long (and variant) history of editing. And typical reviews of major editions in such publications as The New York Review of Books or The New York Times Book Review (and even The Times Literary Supplement) will have little or nothing to say about the actual editing, preferring to concentrate on the juicier details of the author’s biography rather than how (and with what reliability) the text has been established. This situation could be regarded as an act of faith that the textual scholarship is unimpeachable, but a more perceptive understanding that editing is above all critical would make for a more judicious evaluation of both text and author. To this end, the following history seeks to record the interplay of science and art in the various areas of textual scholarship as it has evolved over the last two millennia and more. Text as history The history of textual scholarship is the history of history. At the very moment in each culture that documents begin to preserve the records of that culture, the issues familiar to textual scholars will appear: inscription, graphic representation, transmission, error/variant, authenticity, reception. All of these important matters are raised once the first text-producers, in whatever medium, begin to put chisel to stone, wedge to clay, stylus to papyrus, pen to paper, fingers to keyboard. To cover textual scholarship as an historical phenomenon, this account should thus properly include the identification and “editing” of potsherds (ostraka), cuneiform tablets (clay incised with “wedges”), graffiti (or “scratches,” typically on walls), inscriptions, monumental or quotidian (epigraphy), coinage (numismatics), runes (an angular Nordic alphabet, part based on Latin, part with special symbols), pictographs (pictorial symbols representing words or phrases), and many other media containing historical evidence, possibly even the bark of certain trees (notably beech, from which the word book may descend, although the etymology is now questioned). All of this in addition to the more ubiquitous diplomatics, paleography, codicology (the bibliography of manuscripts), print (by single-piece blocks or, following Gutenberg, movable type), in the three main categories of relief (letterpress, in which the areas to be inked stand out above the surface, as in woodcuts), intaglio

18

david greetham

(its opposite, as in etchings, in which the incised sections are inked), and planographic (which has a flat surface, with the parts to be inked treated in a chemical, photographic, or xerographic process as in lithography). When digitization and electronic text-production are added to this list, involving different systems of “mark-up” language and “text-encoding” as well as a huge range of graphics coding (e.g., TIFFs, GIFs, JPEGs) and audio-visual (e.g., AVI, MPEG, MP3), it becomes all-too-evident that the history of textual scholarship may include any system of marking on any surface: of such materials is “history” made. And, given the recognition of the importance of oral transmission (with a separate chapter in this collection, Chapter 9), it could well be argued that many of these issues can occur even in a non-literate society. What parent has not told (or read) a familiar story to a young child to be peremptorily informed “that’s not how it goes; it should be . . . ” – in an early emphasis on the “correct” or “authoritative” text, the textus receptus (“received text”). In their famous study of the “lore and language of schoolchildren,” Iona and Peter Opie record that one of the most persistent cultural phenomena encountered was the confidence that a child, or group of children, would assert about the “correct” text of a counting-out rhyme or other oral transmission that depended on a mini-society’s recognition of the authentic (even when it could be demonstrated that there were multiple historical variants in the development of the rhyme).1 The young oral transmitters had already encountered the major issues in textual scholarship – and had taken a stand on historical authority. Given that oral transmission is also a part of the history of textual scholarship, that history is not just the history of (literate) history, but the history of (preliterate) oral history as well. And anyone who has played the game of “Telephone” (with a message whispered to the first in a group and then whispered to the next until the “last” version of the message can be compared with the beginning) will know that variance is an inevitable part of any transmission – for good or ill. Greek origins The oft-cited decision of Peisistratus (560–527 bce) to have an “official” text of Homer compiled for the Panathenea festival might be regarded as the earliest record in the West to address the variance produced by the oral recitation of the epics by the professional declaimers of verse, or rhapsodes. While no trace of this “official” text remains, the act of depositing a single written text in one location to be used as reference can obviously be seen as the beginning of a sense of mistrust as the appropriate practice of textual criticism, a suspicion of any claims to authenticity no matter what those arbiters in the playground might believe.

A history of textual scholarship

19

But paradoxically, while founded on a mistrust of the variance in the texts of the rhapsodes, the deposit by Peisistratus simply transferred to a written medium the same insecurities that had existed in the previous oral transmission. Similarly, when the Athenian orator Lycurgus (c. 390– 324 bce) did the same thing for the major Greek tragedians (Aeschylus, Sophocles, and Euripides) by placing a single copy of the works of each author in the Athenian public archives (with, again, no subsequent trace), the archived versions were not necessarily any more “authentic” than others in circulation (and performance). These attempts to address (and arrest) variance could therefore at best arbitrarily favor one material, documentary state of text and by default to declare it to be authoritative. But what did (and does) “authoritative” mean? The inevitable connection with “author” has frequently suggested that the text to be striven for is the one that can be shown to be as close as possible to the form and expression in the author’s “fair copy” – that material version last worked on by the author and intended for release into social transmission. But all of these terms and concepts (“form,” “expression,” “material,” “intention,” “version,” and so on) are themselves not as transparent as we might hope, and all beg the question to one degree or another. The history of textual scholarship can therefore be seen as a series of arguments – often resulting in intellectual and scholarly and personal conflicts, even feuds – over the meaning and significance of its most important terms, from the classical period to the electronic environment of the twenty-first century. After these early Greek gestures toward authority, the most important conceptual and methodological shift occurred in the foundation of the Alexandrian library in c. 284 bce by Ptolemy Soter, who appointed Zenodotus of Ephesus as Chief Librarian. From the beginning, the Alexandrian library worked in exactly the opposite procedure and textual philosophy from that exemplified by Peisistratus and Lycurgus. Instead of promoting a single copy to the status of arbiter, Zenodotus and his successors began the process of the rejection of readings in particular documents. Initially, this rejection was inevitably founded on a speculative, conjectural interrogation of specific lines, phrases, and words that the librarian suspected of being inauthentic. Obviously, the recognition of the “inauthentic” would depend upon the reader/librarian’s having a preconception of what was likely to be authorial and what not; and this preconception would inevitably derive from an examination of previous documents of the same work, an examination that underwrote the identification of the spurious in other documents. Clearly, this was circular reasoning, and it is a charge that, in one way or another, textual scholarship has never entirely escaped. Despite the technological positivism of the analytical bibliography of the

20

david greetham

nineteenth–twentieth centuries; despite the construction of “family trees” of putative documentary relationships; despite the methodological and evidentiary rigors of the “New Bibliography” and editorial eclecticism; despite computer analyses of authorial “stylometrics”; despite the attempts to refine textual conditions to algebraic formulae, finally the textual criticism aspect of textual scholarship will to some extent depend on individual judgment, on the critic–editor’s conviction that this reading in this moment is suspect and should be replaced. It may be regarded as suspect because it does not represent what the critic–editor believes is original intention, or final intention (or any stage between these two hypothetical instances), or because it does not reflect that form in which the reading became part of the textus receptus, the “received text” accepted by a culture as available for commentary and further transmission, whether or not it could be demonstrated that the reading had bibliographical or authorial imprimatur. It must be a rare performance of Shakespeare’s Henry V that does not have Mistress Quickly recount how Falstaff “babbl’d o’ green fields” on his death bed, even though the line is a speculative invention by the eighteenth-century editor Lewis Theobald and has no documentary authority: it just makes sense. Analogy and anomaly Given the circularity of argument involved in all of these editorial/textual maneuvers, how then did the Alexandrians deal with this circularity at the beginning of the history of textual scholarship in the West? The easy answer is – by more circularity – but we do have to give credit to what was a major step forward, dependent as much on the international economics of the third century bce as on a shift in philosophy. By this time, Alexandria had supplanted any of the original Greek cities as the dominant near eastern ˆ and the formidable trading position of the city meant it became entrepot, a “hub” (much like the airport “hubs” of today) through which the great bulk of shipping would pass. And the textual result of this power was the insistence that all ships had to “declare” any manuscripts they might have on board, these manuscripts then to be copied and returned. (In fact, as we might have expected, the Alexandrians seem most often to have kept the original and given back the copy.) In this way, the library began to acquire multiple copies of what claimed to be the same works, under a rationale that is the precise opposite of the expectations in a modern library’s having multiple copies. That is, the familiar “c1,” “c2” etc. on the spines of books in our libraries reflect not just the evident popularity and circulation of a work but also the assumption that each of these “c”s is identical to all the others (although many bibliographers would offer the demurral that no two copies,

A history of textual scholarship

21

even of the same print of the same edition, are truly identical, part of the rationale for analytical bibliography). For the Alexandrians, the gathering of different copies was because they were different, and could therefore be used to prepare lists of variant readings, in a process called collation (from “placing/laying side by side”) for comparison. Such collation, now often done by computer, though requiring some human input and control, is still done today. So similitude (or at least the assumption of similitude) lies behind modern duplication in libraries; difference – and variance – behind the Alexandrians. Despite this access to multiplicity and variant evidence, the Alexandrians did not then attempt to designate a specific copy as the most authentic, but on the contrary, tried to use the admittedly corrupt remaniements (the “remains” of a text) in the extant documents to reach beyond the concrete and the actual into an “ideal” form not available in any individual state. Usually called analogy, this process clearly had affinities with the (neo)platonism that was to dominate Alexandrian thought in the third to fifth centuries ce, and can still be seen in the textual theories of some modern critics. Thus, when G. Thomas Tanselle plangently laments that all documents are “alien, damaged here and there through the intractability of the physical” and reaches after the stasis of a lost “inhuman tranquility,” he is articulating a latter-day platonic nostalgia that the Alexandrians would certainly have recognized.2 In practical terms, Alexandrian reconstruction according to analogy was characterized by such devices as Zenodotus’ marking of spurious lines with an obelisk (or his transposition and telescoping of problematic Homeric lines), and the attempts by Aristarchus of Samothrace (c. 220–145 bce) to remove the previous layers of conjecture by trying to isolate “good” (or perhaps “better”) manuscripts for such authors as Homer, Hesiod, Pindar, Anacreon, Archilochus, and Alcaeus. But throughout, the same problems of circularity occur, as when Aristarchus tried to identify such features as consistency and decorum (on the assumption that good authors would be both consistent and decorous), and having labeled specific readings (or witnesses) as meeting these criteria, used this documentary “evidence” to adjudicate readings found elsewhere. The texts (re)constructed thus fulfilled the aesthetic predispositions of the editor. If Alexandrian idealism and a suspicion of the concrete is still with us in the attitudes of Tanselle and others, so too is the Stoicism of the great city’s rival, Pergamum in Asia Minor. The story goes that, in a trade war with Pergamum, Alexandria forbade the export of papyrus (which, it was claimed, could be made only from Nile water), thus drying up the raw materials for Pergamanian textual production. This event would not have

22

david greetham

been remembered as bibliographically significant, were it not for the fairly well attested sequel in which Pergamum turned to animal skins (sheep, cows, and goats) as an alternative writing material, thus producing the parchment (i.e., “material from Pergamum”) that was to dominate text-production for the next fifteen hundred years and more. So the trade war changed the material, but what of the philosophy? Working against Alexandrian analogy and instead promoting the concept of anomaly (dependent on a Stoic acceptance of the unavoidable corruption of all worldly phenomena), the linguistically based textual scholars of Pergamum favored a careful analysis of the provenance, philological features, even the paleography and grammar, of each surviving witness. This analysis was then used to select a “best text” (warts and all) that could at least represent an actual historical moment rather than veering off into an idealism for which no concrete demonstration could be made. While none of these “best text” editions produced by Pergamanian scholarship survive intact (as do none of the Alexandrians’), but only in citations in later works, these later references and quotations in medieval scholia (commentaries) do make the procedural distinctions fairly clear (and record that the Pergamanians also extended the canon of editable literature into prose as well as poetry). The shift of medium from papyrus to parchment is usually associated with another shift – from scroll to codex, although the latter format was more specifically identified with early Christian writings (of the patristic commentators, or “Fathers” of the Church). While relatively cheap and easy to produce, the scroll was an inherently difficult medium to consult, since a specific passage had to be found by rolling up one end while unrolling the other. The parchment codex, however, the folded, stitched book still familiar in most cultures, while comparatively expensive (it did, after all, require the death of an animal, reared until its skin was available) could be easily consulted, easily scraped and written over – to produce a palimpsest – and was peculiarly hospitable to the text-citation methods of early Christian polemic. This introductory concentration on the Alexandrian versus Pergamanian split may be a “long preamble to a tale” but it lays out the basic features of much of the history of textual scholarship to follow. For example, Vinaver’s “mistrust of texts” was to be exemplified most productively under the mid-twentieth-century Anglo-American production of “eclectic” editions not representing any specific historical document and the Pergamanian selection of, and reliance on, a “best text” reappeared under the early twentieth-century theories of the French critic Joseph B´edier. It is even possible to see the now-dominant textual discourse of “social textual criticism,” associated primarily with the work of Jerome J. McGann and

A history of textual scholarship

23

D. F. McKenzie, as a further rejection of Alexandrian idealism in favor of social/historical “reality.” Indeed, the intervening history of textual scholarship between the classical and postmodern periods can in several ways be viewed through the lens of the Alexandria/Pergamum, platonic/aristotelian dialectic.

Enumerative bibliography The other component of classical textual scholarship that is still important to any modern scholar is the attempt to codify, enumerate, and describe the primary witnesses, that part of the textual enterprise usually called enumerative or systematic or descriptive bibliography. Given the enormous riches of the Alexandrian library (later destroyed in circumstances that are still in dispute3 ), with somewhere in the region of 750,000 rolls accessible in pigeon-holes rather than on library shelves, this counting and describing was a formidable task. Again, we have only the references surviving in later records, as for example, the quotations recorded in Athenaeus’ “conversations” in The Learned Banqueters.4 But the comprehensive Pinakes or “tablets” of Callimachus (c. 305–240 bce) aimed at including the totality of Greek documents, from the heroic and dramatic and lyrical verse to the recipes of a famed Alexandrian courtesan, compiled in a “manual” written for a daughter following in the same trade. The sole criterion was that to be recognized as an enumerable document, its language had to be Greek. The Pinakes was thus the forerunner of the various “national” bibliographies (for Britain, Germany, France, and the US) compiled from the sixteenth century onward and culminating in the various Short Title Catalogues, the National Union Catalogue, WorldCat, and the library catalogues of such national archives as the British Library, the Biblioth`eque Nationale, and the Library of Congress. The linguistic restriction imposed by Callimachus would have meant, for example, that although the existence of the Hebrew Bible must have been well known to Alexandrian scholars, its text did not become an authentic “document” until after the creation of the Greek Septuagint (c. 300–100 bce). It is a useful corrective to the preservation and codification efforts of Callimachus and others to note that, of the roughly 330 works of the four major Greek dramatists (Aeschylus, Sophocles, Euripides, and Aristophanes), only 43 survive in anything like a “complete” form. While some of the riches of classical, medieval, and modern literature have escaped fire and flood, to be occasionally “augmented” by the later discoveries of, for example, the Nag Hammadi gnostic gospels and the Dead Sea Scrolls, much more has been lost, in a process of accidental or deliberate destruction that has continued until the present day, as the devastations of Sarajevo, Iraq,

24

david greetham

and Louvain (whose library was deliberately destroyed by the Germans in both World Wars) demonstrate all too well. Late classical and biblical But despite such losses, the concepts, methods, disputes, and media established by classical Greek textual scholarship have been carried over into medieval and modern textual activities, partly through the Roman heirs and partly through the continued Greek traditions of the Eastern Roman Empire, ending with the conquest of Constantinople in 1453. Indeed, together with the work of biblical scholars, the inheritance of classical texts and their editors informed the terms of the following vernacular scholarship. As in several other aspects of culture, Roman scholarship imitated Greek, but with a different emphasis. Initially lacking a national poet like Homer, Latin scholarship was first characterized by linguistic rather than textual analysis, with such works as Varro’s De lingua latina, Nigidius Figulus’s Commentarii grammatici, Verrius Flaccus’s De orthographia and De verborum significatu, and Quintilian’s Ars grammatica being typical. However, with the founding of the Palatine Library in 28 bce and the appointment of Julius Hyginus as librarian, together with the acknowledgment of Virgil as the national poet, Latin scholarship also began to produce textual commentaries, notably Servius Honoratus on Virgil. The learning of the classical world was then passed on to medieval readers in such compendia as the seventh-century Etymologiae of St. Isidore of Seville, the fifth-century Nuptiae Philologiae et Mercurii of Martianus Capella (which became the model of the study of “liberal arts” for the next millennium), and the sixth-century Institutiones grammaticae of Priscian, the basic text for later work on Latin grammar. In the meantime, the attempts at establishing the texts of the Hebrew and Greek Bibles provided another model for textual scholarship. This is a very complex history, and only the basic outline can be given here. The initial problem for the text of the Hebrew Bible is in the (editorial) vocalization, accentuation, and word-division of the consonantal Hebrew. To this day, the “standard” version of this editorial process is the so-called Masoretic text (compiled by a group of Jewish textual scholars known as Masoretes), established between the sixth and eighth centuries ce, and differing substantially from the previous Greek Septuagint translation. This divergence has been further complicated by the testimony of the Dead Sea Scrolls, which have preserved much earlier texts than those in the Masoretic tradition. This multiple evidence can have major cultural significance: for example, the four books of Maccabees (the source for the Jewish celebration of

A history of textual scholarship

25

Hanukah) were not written in Hebrew and thus do not appear in the orthodox Masoretic text, but are included in the Roman Catholic and Eastern Orthodox canon. But the Hebrew Bible (Old Testament) is a paragon of consistency and uniformity compared to the transmission of the Greek New Testament. The work of such modern scholars as Bruce Metzger and Bart Ehrmann has demonstrated that the variance in the rival documentary traditions of even the canonical books can, for example, cast doubt on the authenticity of the ending of the gospel of Mark (and thus on the “pseudo”-Mark’s account of the resurrection, the descent of the Holy Ghost and the speaking in tongues, foundational elements in many Christian churches). Even within a single book, there may be major divergences. In Acts, there are two completely different accounts of Paul’s conversion on the road to Damascus, one in the first person (9.4) and one in the third (26.14ff), and only the second of these contains the mandate to go out and convert the Gentiles. Similarly, Paul’s supposed strident rejection of the role of women has been shown to be a later interpolation (and does not reflect the important contributions of women in the early Church). Given the sectarian conflicts of early Christianity, this degree of “determined variation” is hardly surprising, but the force of the textus receptus (the “received text”) has meant that many of these inconsistencies and inauthentic readings are still preserved in modern Bibles. With this Christian sectarianism and proliferation of witnesses, the goal of producing a single authoritative text became particularly challenging and was not comprehensively attempted until Jerome’s early fifth-century Latin Vulgate. Jerome (ce 331–420) already had the model of the Hexapla of Origen (d. 255 ce) before him, with its “parallel text” presentation of six different versions of the Old Testament, a method still used in modern editions of, for example, the 1805 and 1850 versions of Wordsworth’s Prelude or the F and G versions of Chaucer’s Prologue to the Legend of Good Women. Faced with a multiplicity that cannot yield a single authoritative text, Origen and some modern editors decided simply to lay out all the evidence in different columns. But the unwieldy testimony of the Christian Bible did not allow of such textual largesse. Instead, Jerome worked from multiple sources and in several stages, with the New Testament taken from Old Latin versions in consultation with the oldest Greek manuscripts available. His work on Psalms is exemplary of his eclectic and revisionary method: he produced three different redactions, the first based on the Old Latin, the second on Origen, and the third on the Hebrew instead of the Septuagint, a process that he then undertook for the rest of the Old Testament. The success of Jerome’s Vulgate is also its liability, for with 8,000 manuscripts

26

david greetham

the range of variance that came down through the Middle Ages and beyond was enormous. Medieval That Pope Damasus I appointed Jerome to produce an official Latin Bible (rather than Greek or Hebrew) was further evidence of the hegemony of Latin in the West. But what this meant was that, in an age when Latin (at least in the West) was the lingua franca and when very few of even the small educated class could read Hebrew or Aramaic or Greek, the task of such textual scholars as Lupus of Ferri`ere (c. 805–62) was primarily conservational rather than restorative, by trying to build up the holdings of Ferri`ere by loans from Tours, York, and Rome. He may indeed have used the Alexandrian system of collation (presumably without knowing it), but unlike the Alexandrians he would rarely risk a conjectural emendation, preferring to leave a blank space. This limitation was of course compounded by the monastic concentration on the copying of religious rather than pagan works. While the great task of the high Middle Ages might have been to reconcile pagan works with the revealed truths of Christianity, the fact remains that even such a learned figure as Bede (c. 673–735) probably derived his knowledge of the classics through Macrobius and Isidore, and the references that Alcuin of York (c. 730–804) makes to the classical authors (Virgil, Lucan, Pliny, Statius) in his account of the library at York were not the norm for the time. That the only extant version of Cicero’s De Republica survives as the lower (erased) text of a palimpsest, a manuscript “overwritten” by Augustine’s In psalmos, is vivid testimony to the relative cultural and transmissional significance of classical and Christian texts. Despite these constraints, it would be a mistake simply to fall into the later “renaissance” critique of medieval scholarship as uncouth and barbarian. Through such movements as the Caroline reformation of handwriting (overseen by Alcuin at the invitation of Charlemagne), which created a clear yet efficient script out of the corrupt and largely illegible pre-Carolingian “national” scripts, and the programs of copying at Corbie, Li`ege, St. Gall, Monte Cassino, and other monastic centers, the Carolingian “renaissance” did produce the great bulk of our earliest surviving manuscripts of classical literature. Similarly, while the standards of Alexandrian Greek scholarship slipped in the Eastern Empire, Constantinople remained an important transmitter of classical culture in Greek until the fall of the Eastern Empire to the Turks provided a counterweight to the hegemony of the Latin West through the work of such scholars as Michael Psellus (1018–78), Anna Commena (1083–1153), Eustathius of Thessalonica (c. 1160–92), and John Tzetzes

A history of textual scholarship

27

(c. 1110–80). While much is made of the deleterious effects on the archive of Greek manuscripts by the Turkish conquest in 1453, it is probable that the devastation of the Constantinople libraries by the Christian Fourth Crusade in 1204 was even greater, and the post-Crusade scholarship of such figures as Maximus Planudes (c. 1255–1305), Demetrius Triclinus (fl. 1305–20) and the Paleolologi is characterized by a shortage of the raw materials – even the parchment – on which textual scholarship is based. The last (and lasting) contribution of Byzantine textual scholarship was to provide the early Italian humanists with what manuscript riches remained. In 1398 Jacopo Angelo was sent by Coluccio Salutati (1331–1406) and the Florentine signoria to go book-hunting in Constantinople for texts of Homer, Plato, and Plutarch; and in the early fifteenth century, just before the Turkish conquest, Giovanni Aurispa came back to Italy with hundreds of Greek manuscripts. Although Salutati did arrange for the Greek scholar Manuel Chrysoloras (1355–1415) to leave Constantinople for Florence (and thus to begin the first serious study of Greek in the West in 700 years), the irony is that, until the print editions of the Aldine press, these Greek manuscripts seem to have been used primarily as the basis for Latin translations rather than as independent witnesses for Greek editions. Nonetheless, Constantinople’s function as a repository of Greek learning long after Greek had been lost in the West did mean that the work of the Alexandrians did find its way into late medieval Europe.

Renaissance If the humanist movement beginning in Italy was to produce such scholarly editions as the Greek and Latin classics of the Aldine press, then there had to be a period in which the documentary output (in both preservation and copying) of the medieval repositories could be accessed and examined with some sort of bibliographical rigor. The output had been greatly increased as a result of the gradual passing of clerical (i.e., monastic) responsibility and control to commercial scriptoria, employing professional scribes, often associated with the new centers of learning in the universities. While such figures as Petrarch and Boccaccio are perhaps the best known of the humanist collectors, the earlier work of, for example, Lovato Lovati (1241– 1309) provided the initial momentum. In fact, Lovato was already familiar with Catullus, Propertius, Tibullus, and Lucretius before the later humanists’ claims to have recovered such authors for the first time. Similarly, Giovanni de Matociis (fl. 1306–20) was the first to make a successful distinction between Pliny the Elder and Younger, and Geremia da Montagnore (c. 1255–1321) can be credited with uncovering the tragedies of Seneca.

28

david greetham

But Petrarch’s (partial) construction of the histories of Livy showed how a careful examination of the fragments surviving in medieval repositories, through collation and correction in a traveling notebook, could begin the long editorial task of reconstruction. The exilic papal court at Avignon may have been an embarrassment and liability to the Roman Catholic church, but the documentary riches accumulated at Avignon (together with such depositories as the Chapter Library of Verona) furnished Petrarch and his compatriots Boccaccio (at Monte Cassino) and Salutati (with agents throughout southern Italy and Constantinople) with the manuscript resources that provided the raw materials for the textual study of classical authors. This peripatetic re-discovery of the classical inheritance is best exemplified in the work of Poggio Bracciolini (1380–1459), covering several different scholarly fields. In his role as a papal secretary, he traveled all over Europe (Konstanz, Cluny, St. Gall, Cologne) and found manuscripts (of Cicero, Quintilian, Valerius Flaccus, Lucretius, Manilius, Ammianus Marcellinus) wherever he went. He also invented one of the new humanist scripts, breaking with the dense lattice work of textura (what we usually think of as gothic), to reform scribal usage toward clarity and legibility, necessary if the further transmission of the discovered texts was to proceed with anything like accuracy. Moreover, he described his various activities in a lively and engaging series of letters (which still survive) to his fellow scholars and patrons. The activities of Petrarch, Boccaccio, Poggio and others demonstrate several important themes in renaissance textual scholarship. The triumph of humanist (roman and italic) script over textura, everywhere except in Germany and in the practice of law, showed how the evolution of inscription technologies could promote consistency, legibility, and broad international standards for transmission. The migration of witnesses from the Greek East to the Latin West showed how the fortunes of war, institutional collection practices, and the sharing of textual resources could affect the relative significance and influence of centers of learning. And the promotion of pagan literature, philosophy, science, and language studies over theological texts showed the practical, documentary effects of the “humanist” agenda of the Italian city-states (particularly Florence and its signoria). In these and other related areas, politics, military power, financial advancement, and civic pride could produce identifiable textual results. With all of this manuscript collecting and collation underway, it fell to such figures as Lorenzo Valla (1407–57) and Angelo Poliziano (known as Politian, 1454–94) to undertake what we might recognize as a philological method. Valla is best known (one might say notorious) for having exposed the Donation of Constantine – a document supposedly sent by the Christian Emperor Constantine to the pope and thus conferring secular power on the

A history of textual scholarship

29

papacy – as a forgery. Through an examination of its bibliographical and textual characteristics, Valla showed that the Donation had been written in the eighth or ninth century, not in the fourth, and the claims of the papacy crumbled before this evidence. He brought the same interrogative philology to his exposure of the “letters” between Seneca and St. Paul – another forgery – and undertook the emendation of Jerome’s Vulgate, by reference to Greek as well as patristic texts, in his Adnotationes in Novum Testamentum, which Erasmus published in 1505. Politian took Valla’s scholarly principles to another level in proposing what later became known as the “genealogical” method, with a less authoritative witness descending from an inferred (or sometimes extant) version. This search for “origins” was later to fall into disrepute because of its almost mystical veneration of the archetype (conventionally referred to as O′ –, i.e., O prime), the earliest recoverable state of transmission; but the work of both Valla and Politian was the necessary first stage in the formulation of what developed into the scholarly discipline of Altertumswissenschaft (the “science” or “study” of ancient times), whereby textual, linguistic, and bibliographical features could be arranged along a linear path of historicity. The same principle was later to be used by Mabillon and the Maurists in the arrangement of scripts in an historical line of development. Politian’s devolutionary method (and rejection of the readings of later manuscripts) was later codified into an editorial principle: the eliminatio codicum descriptorum (the elimination of “descriptive,” i.e., derived, copies as witnesses to an authorial text), and more fully developed in the genealogical models of Lachmann and Maas. The motto has been severely questioned in later periods – from those who would argue that later witnesses may still preserve or even reinvent authentic readings and those who would claim that all witnesses show a text in social negotiation and all are therefore potentially valuable, if not all for the same reasons. But the basic concept remains a foundational ideal in much contemporary scholarship, and is the raison d’ˆetre behind the genealogical arrangement of stemmatics (the stemma codicum or “family tree” of witnesses): see Figure 1.1, where the archetype is represented by , the “inferred,” i.e., non-extant, witnesses by Greek-letter sigla, and the textual “remaniements” by the Roman letters. In this stemma, B and C are presumed to share errors different from those in E, F, and G; and the witness D is a “descriptive” manuscript with no independent authority. Because it is closer to the archetype, A will supposedly be less corrupt, and thus more authoritative, than other witnesses. Valla’s interrogation of the Greek New Testament proved a stimulus to the decision of Erasmus (1466–1536) to undertake his very influential edition of both Greek and Latin versions. He was rigorous in collecting manuscripts in both traditions, but where tradition was deficient he had no

30

david greetham

1.1 Typical stemma or “family tree” of textual transmission, showing the putative archetype, inferred (Greek sigla) and extant (Roman sigla) witnesses. Diagram by author.

qualms about providing a reading from the other. Thus, his primary Greek manuscript lacked the conclusion of Revelation, so he simply translated the Latin back into Greek. And he similarly did not hesitate from “improving” his sources where he thought intelligibility (or just good style) required. Finding some of the Latin in the textus receptus of Romans inelegant, he claimed “it is only fair that Paul should address the Romans in somewhat better Latin,” (when, of course, Paul would in fact have written in Greek).5 Nonetheless, Erasmus did bring a scholarly skepticism to the editing of the biblical texts, as is shown in the single passage for which the Erasmus edition encountered the most criticism: an omission that struck at the heart of Christian orthodoxy – the so-called comma iohanneum on the Trinity. The sentence familiar to us in the King James version and beyond, reads, at 1 John v. 7 “For there are three that bear record in heaven, the Father, the Word, and the Holy Ghost: and these three are one.” This is the only biblical passage that refers unambiguously to the Trinity: the problem was that it was not in any Greek manuscript, and so Erasmus left it out, and brought down the wrath of orthodox theologians, especially Stunica, one of the editors of the Complutensian polyglot Bible, produced at the university in Alcala´ de Henares (Latin Complutum). Erasmus did agree that if a Greek manuscript with the Trinity passage was discovered, he would reinsert it in a future edition. And, of course, in one of the greatest archival “discoveries” of the time, a Greek manuscript (especially written for the purpose in 1520 by a Franciscan friar in Oxford) was produced. Despite demurrals about the authenticity of this “discovery,” Erasmus made good on his promise and reinserted the Trinity in future editions, where, with some exceptions, it remains still, a dogma based on a forgery. Indeed, no less a textual critic

A history of textual scholarship

31

than Pope Leo XIII declared in 1897 that it would be “not safe” to “deny that this verse is an authentic part of St. John’s Epistle,” though the papacy backtracked on this assertion in 1927.6 A measure of the success of Valla, Politian, and Erasmus can be seen in the confidence with which Petrus Ramus, polyglot and polymath, could assert in his famous oration of 1546 that modern scholarship had overturned medieval Scholasticism (particularly the reliance on Aristotle) and that in his time, all of the major classical authors were now available in reliable editions. It is true that the Aldine editions beginning at the very end of the previous century had often been overseen by reputable scholars (Erasmus among them), but the optimism of Ramus was more a reflection of the familiar “renaissance” rejection of the medieval than of the growth of a genuinely philological method. Rise of philology and modern textuality That this method was slow in coming can be measured by the later comments of A. E. Housman on the continued work of Joseph Justus Scaliger (1540– 1609) on the text of Manilius, the first-century author usually credited with composition of the anonymous Astronomica. When the first edition appeared (1579), no reliable witnesses had yet appeared, and Housman thus notes that “the transformation which first made Manilius a legible author was the work of Scaliger’s own unaided wits”;7 whereas, by the time of the second (1600) edition, the Gemblacensis collation (a manuscript “G” from the monastery of Gembloux in Brabant) had appeared, and the judgment of Scaliger’s “wits” could be checked, and confirmed, against a bibliographical authority. The results were, in Housman’s words, that “no critic ever effected so great and permanent a change in any author’s text as Scaliger on Manilius.”8 It is precisely this combination of an identification and evaluation of bibliographical resources with a critically acute sensitivity that has been the goal of textual scholarship in its various manifestations since the Renaissance. Again, Housman can be an appropriate guide: he defined textual criticism as the “science of discovering errors in texts, and the art of removing them,” although most textuists might today argue that science and art are present in both stages.9 The range of activity in the last four centuries is so various that it would be foolhardy to try to encapsulate this history into a few paragraphs. But there are several emblematic moments in this history that can serve to illustrate the critical issues at stake. Take, for example, the textual career of Karl Lachmann (1793–1851), generally recognized as the progenitor of the genealogical method still used in much editing today. Lachmann’s reliance on very rigid bibliographical principles can be shown in his edition of Lucretius

32

david greetham

(1850), where he claimed he could prove not only that the three major extant manuscripts all descended from a single archetype, but that this (lost) archetype was composed of exactly 302 pages, each with 26 lines. He also determined that the archetype had been copied from a minuscule original and that this minuscule (a style using both what we would now call “capital” and “lower case” letters, though these terms are more properly applied only to print and typefaces) was itself a copy of a rustic capital text (a slightly less formal version of monumental “square capitals”) from the fourth–fifth centuries. In other words, Lachmann’s bibliographical research could, he claimed, show exactly where the page breaks occurred in the no longer extant archetype. But beyond the archetype he would not go, a reining in of speculation that also appears in the very influential work of his disciple Paul Maas, whose stemmatic models proceed upwards (from extant but corrupt witnesses) to the putative source(s) – until we reach the archetype, at which point reconstitution of the text is impossible. This argument is based on the assumption that later witnesses will contain more errors as the text is disseminated and that errors held in common by two or more witnesses will show that they descend in the same line. As already noted, this procedure may facilitate the “elimination” of “derivative witnesses.” The irony may be that Lachmannian stemmatics may lead to an overvaluation of the codex optimus, or best manuscript, and provide a respectable rationalization for editorial promotion of a manuscript already favored for other reasons. The irony is that this codex optimus, or best text, procedure was later advocated by Joseph B´edier as an alternative to what he saw as the inherent falsification of the Lachmann method, since, as B´edier observed, stemmatics could all too often (and implausibly) lead to a binary opposition between two texts (and no more). However, this apparently cautious approach is not observed in Lachmann’s edition of the vernacular Nibelunge Not und die Klage (1826), for which documentary sources, almost by definition, could not be marshalled in the same way. That is, in editing a classical author, Lachmann could be bibliographically restrained, whereas he felt that the search for the Germanic Geist putatively to be found in the Nibelungenlied justified greater speculation. And this dichotomy could previously be seen in the work of Richard Bentley (1662–1742). In his classical scholarship (for example, his 1726 edition of the Roman author Terence) he could practice a rigorous philological method, but in his notorious edition of Milton’s Paradise Lost he convinced himself that the infelicities he observed in the printed text were the result of Milton’s blindness, and his thus having to use an amanuensis, on whom all these “unmiltonic” readings – including even the ending of the

A history of textual scholarship

33

poem – could be foisted. As with Lachmann and the Nibelungenlied, Bentley thus gave himself carte blanche for conjectural emendation. The challenge of possibly reaching beyond bibliographical evidence can be seen in two almost contemporaneous moments: in biblical scholarship, the “quest for origins” was to all intents and purposes formally abandoned when the Einleitung in das Alte Testament (1780–83) of Johann Gottfried Eichhorn (1753–1827) took the position that the biblical texts were no different from secular and that, because of the numerous layers of copying, it was impossible to achieve a “transcendent” or “original” text. Similarly, Friedrich August Wolf (1759–1824) argued in the Prolegomena ad Homerum of 1795 that the original versions of the Homeric epics were irrecoverable, again because of the multiple generations of copying. In essence, Wolf was declaring that the conservational or reconstructive agenda of Peisistratus and even of the Alexandrian librarians was untenable. National bibliographies Such counsels of despair (if that is what they were) did not mean that the search for “origins” (or at least for witnesses who testified to an otherwise unrecorded state of the text) did not continue, and during the nineteenth century and on into the twentieth, perhaps the greatest contributions to textual scholarship were to be found in the collecting, sorting, describing, and transcribing of a documentary history that would reflect a national patrimony. This process had already begun in the founding of what became national libraries, to replace the monastic institutions of the Middle Ages. The most famous examples are Duke Humphrey’s library at Oxford, refounded by Bodley in 1610; the Lambeth library of Archbishop Bancroft (1610); Charles IV’s academic library in Prague (1348); the Louvre library of Charles V (1368), combined with the Fontainebleau library of Francis I, reorganized by Guillaume Bud´e, to become the basis for the Biblioth`eque Nationale; Pope Nicholas V’s refounding of the Vatican library (1448); the institution of the Ambrosian library in Milan by Cardinal Borromeo (1609); the Medici’s Laurentian library (1571); and the British Museum Library, based on the collections of Arundel, Sloane, and Cotton (with the manuscripts in what is now the British Library still bearing the names of these original collectors, just as the Rawlinson, Laud, and Douce manuscripts in Bodley carry their original identities), complemented by the Public Records Office (1838) for official government and court records. In a later period, the United States saw the founding of the New York Public Library (1895, based on the Astor, Lenox, Arents, and Berg collections); the Huntington library in San Marino, California (1919); the Pierpont Morgan in New York (1924); the Houghton

34

david greetham

(1942) and Widener (1915) libraries at Harvard (1942); and Lilly at Indiana (1960); plus the enormous holdings of the Library of Congress (1800, destroyed 1812, refounded 1815). The concentration of documentary records in such centers facilitated the production of national bibliographies, particularly when union catalogues of the collections were compiled, leading to the publication of national historical records. The chief examples include the Rerum britannicorum medii aevi scriptores (the Rolls Series, 1858–91) for Britain; the Rerum gallicarum et francicarum scriptores (1738–1904) for France; and the Monumenta Germaniae historica (1819 onwards) for Germany. More comprehensive enumerative bibliographies of specific value to textual scholars in English-language studies began with the so-called Short Title Catalogues (Pollard and Redgrave for 1475–1640, revised; Wing 1640–1700, revised; Eighteenth-Century Short Title Catalogue for 1701– 1800), now all combined into the English Short Title Catalogue (ESTC) for 1473–1800.10 The two other types of basic materials that characterize nineteenth-century scholarship can be illustrated by the Early English Text Society and the New (later Oxford) English Dictionary. In fact, the EETS was founded by Frederick Furnivall in 1864 specifically to provide textual readings for the historically organized OED, which was also initiated by Furnivall, but did not begin the lengthy process of compiling records until the appointment of James Murray in 1878, and did not finish its first edition until 1928, with a second edition in 1989 and a third in preparation. EETS produced the first print editions of, for example, the Cotton Nero A.x. manuscript, containing the sole witness to the texts of Pearl and Sir Gawain and the Green Knight and is still publishing valuable volumes of medieval material. While some of the EETS publications can be regarded as “critical,” with an editorial examination and evaluation of all significant witnesses, because of its role as producer of raw materials for the OED, EETS published many editions that are “diplomatic” transcriptions of specific manuscripts. Vernaculars This interest in, and promotion of, collecting, editing, and publication of vernacular materials was a testimony to the major shift in textual scholarship in the nineteenth century. As this brief history has shown, for over 2,000 years the great developments in textual expertise had been concentrated on transmitting the classical and biblical heritage to later periods. It is true that the activities of a printer/publisher like Caxton were responsible for producing print texts of the major English works of Chaucer, Lydgate, Malory and so on, but the great strides in editorial theory were still being

A history of textual scholarship

35

made in classical and biblical texts. The career of Lachmann embodies this distinction, given his rigorous genealogical method in Lucretius as opposed to the idiosyncratic editing of the Nibelungenlied. But with the combination of the growth of national libraries, national bibliographies, and the eventual introduction of vernacular studies in the curricula of research universities, the balance began to shift away from the classics and towards vernacular works. It is undeniable that, in the early years, the study of the vernacular in the universities was modeled on classics. Indeed, there was still a suspicion of vernacular study as being mere “chatter about Shelley,” so that the courses of study had to prove that they were as “difficult” as those in classics; and the major result of this inferiority complex was a concentration (in British, American, German, and French universities) on philology in the sense of historical linguistics, grammar, laws of sound changes, and so on. The linguistic bias thus meant that there was very little place for what we might recognize as literary criticism; but it also meant that the typical university student, especially at the graduate level, was expected to produce a scholarly edition as his (and later her) final exercise and admission to the academy. As current textual scholars have now recognized, this sort of project is increasingly rare, although newer aspects of textual work (for example, the growth of interest in the history of the book) have begun to emerge. Recent Anglo-American and Continental textual scholarship is covered in the essays by Sutherland and Lernout respectively, so this account will cover only those aspects of this later period and these fields that relate to the general history of textual scholarship already laid out. For example, the “strict and pure” bibliography of Fredson Bowers, following upon the copytext theory of W. W. Greg, can be seen in part as a development of both the “difficulty” deliberately espoused by the still-new vernacular courses of study at the universities and in part as a reflection of a general late nineteenth- and early twentieth-century confidence in positivism, as both a scientific and a philosophical agenda. There are several ironies in this belief – or hope. One example of this positivism is the oft-cited address by Lord Kelvin to the British Association for the Advancement of Science in 1900 in which he stated, “There is nothing new to be discovered in physics now. All that remains is more and more precise measurement.”11 This was just shortly before Einstein’s 1905 publication in Annalen der Physik of the four “miracle year” essays that assigned Newtonian physics to the history rather than the present of the discipline. And this assumption that all had been achieved can be paralleled by the entry for “textual criticism” in the famed eleventh edition of Encyclopaedia Britannica that “[a]s time goes on, textual criticism will have less and less to do. In the old texts its work will have

36

david greetham

been performed so far as it is performable. What is left will be an obstinate remainder of difficulties, for which there is no solution or too many. In the newer texts, on the other hand, as experience has already shown, it will have from the outset but a very contracted field.”12 This confidence that there is “rien a` faire” was a mark of optimism rather than Beckettian despair, but at the beginning of a century that witnessed a series of revolutions in textual studies as well as the inauguration of what has often been seen as a “great age of editing” it seems singularly misplaced, though of a piece with its time. It is a product of a belief that, given enough facts, any problem can be solved with the application of a “scientific” rigor. Indeed, despite the disdain leveled by A. E. Housman against those German scholars who had mistaken textual criticism for mathematics, and despite the insistence of even reputable analytical bibliographers like Tanselle that textual scholarship was not a science, the desire for scientific surety continued to be a grail followed by many textuists during this period. For example, Tanselle quotes an address before the Bibliographical Society of America claiming that “Bibliography, as taught and practiced in the circle to which I address myself, ranks now equal to, if not among, the exact sciences.”13 The temptation to align textual scholarship with the physical sciences can perhaps be understood against the increasingly technological research of analytical bibliography, as for example, in the cyclotron analysis of the inks used in the printing of the Gutenberg Bible, or the stylometric analyses of Shakespeare plays. Such studies appeared to be so technical (often using a vocabulary that was comprehensible only to the initiate) that textual scholarship became associated in the minds of many academics and serious readers with a dryasdust, incomprehensible, and eminently ignorable discipline that had little to contribute to the major intellectual debates of the later twentieth century. This isolation of textual scholarship was unfortunately abetted by some of the more doctrinaire pronouncements of Anglo-American bibliographers against what is often regarded as the “monstrous regiment” of (mostly French and German) exponents of structuralist or new historicist or even post-structuralist textuality. Thus, when Hugh Amory reviewed the proceedings of a conference on La bibliographie mat´erielle (responding to Roger Laufer’s question “La Bibliographie mat´erielle: pourquoi faire?”, with its unacceptable answer, for “communication”), he then launched into an attack on everything that such “foreign” textual study stood for: “[t]he object of bibliography is just to describe books: not texts, not ideas or states of mind, not semeiotic [sic] systems or social relations, not even the Book; but books – assemblies of paper and ink and thread and glue and cloth and leather, remember?”14 Similarly, in David Shaw’s account of bibliologie, he made it clear that the French lack of a concern with the scientific certainties of analytical bibliography was a result of the “theoretical interests of French

A history of textual scholarship

37

literary scholars [who] notoriously tended to structuralist and its various offshoots.”15 Intention and sociology To balance such partisan positions, it must be admitted that analytical bibliography, when handled by a sensitive, even quirky and entertaining scholar, can be both technically rigorous and provocatively presented, as the work of Randall McLeod illustrates so well (see, for example, Chapter 7 in this volume). But the demurrals against the positivism of analytical bibliography registered by, for example, D. F. McKenzie, emphasized the importance of the “social, economic, and political motivations of publishing, the reasons why texts were written and read as they were, why they were rewritten and redesigned, or allowed to die.”16 In so doing, McKenzie charged that analytical bibliography was an incompletely historicized procedure, since the focus on a book as opposed to the book depended on “a virtuosity in discerning patterns in evidence which is entirely internal, if not wholly fictional.” The enlarging of the socio-cultural context of bibliography (to include, for McKenzie, even landscape and topography), was further promoted by the work of Jerome J. McGann and others (especially in A Critique of Modern Textual Criticism17 ), whereby the previous concentration on authorial intention by Fredson Bowers and Tanselle was enlarged to include later socialized states of the text as it appeared and reappeared at different historical moments and by various agents: author’s friends and relatives, publishers, editors, readers. McGann did not (as he was sometimes charged) remove authorial intention, but he regarded this so-called “originary moment” as just one in a sequence of textual negotiations. The major shift that the McKenzie–McGann revolution in textual scholarship brought to the discipline extended much beyond the Romantic and Victorian authors that McGann had edited or the work of McKenzie in the seventeenth century. The tacit (and sometimes overt) assumption had long been that the goals of textual criticism were to restore the text to an “original,” “authorial” form. Thus, the editing of medieval texts during the 1960s and 1970s was still primarily (perhaps even unthinkingly) dominated by a desire to uncover lost authorial intention in the face of what looked like the garbling of authorial composition by unintelligent and meddlesome scribes. A notorious example of the rejection of scribal involvement was the highly contentious Kane–Donaldson editions of Piers Plowman, which justified often highly speculative conjectures on the assumption that the author’s mind and expression were so original that mere scribes would mangle the text in the act of copying – an editorial position based on the dictum that the lectio difficilior probior est (the more “difficult” reading is more likely

38

david greetham

to be authorial than the “easier” one adopted by scribes).18 During the same time, I contributed to an edition of John Trevisa’s late fourteenthcentury On the Properties of Things,19 in which a similar attempt at the restoration of a lost authorial intention was the guiding principle, when the only “remaniements” were later scribal copies. And for that reason, just as George Kane and E. T. Donaldson rejected the evidence of two manuscripts that were felt to have been too heavily worked over by scribes, so the Trevisa editors relegated the readings found in the only paper manuscript – and one full of idiosyncratic variants, interlineations, and other marks of individual intervention – as unreliable. But since that quirky manuscript showed the work in its social circulation (in fact, was described as probably having been made by a person for his own use, not by a professional scribe), if the Trevisa text were now to be re-edited, that formerly rejected witness would probably be accorded a textual significance of a particularly valuable kind, the socialized version of an authorial text. For similar reasons, manuscripts that were thought to be contaminated or conflated (again, showing social negotiation of a text) were usually rejected as unauthoritative during the heyday of intentionalist editing – in medieval studies as much as in any other period. As Tarrant points out, both Lachmann and Maas had been unable to deal with contaminated witnesses, an unease expressed in the formula “No specific has yet been discovered against contamination,” which, as Tarrant observes, is much “starker” in the original German: “Gegen die Kontamination ist noch keine Kraut.”20 But a combination of an interest in the genetic formation and transformation of texts, together with the “sociology” of texts, meant that, in the latter twentieth century, a book like Bernard Cerquiglini’s In Praise of the Variant: A Critical History of Philology could be seen as representing the concern among medievalists for “whatever is unstable, multiple, and precarious,” which Cerquiglini, like many other textual scholars, specifically links to electronic text production, characterized as “mobile, various, and fluctuating.”21 In fact, Cerquiglini is best known for his assertion that what he calls mouvance (variance) is not just an accident of medieval textuality, it is its major feature: “[v]ariance is the main characteristic of a work in the medieval vernacular.”22 The instability, multiplicity, and precariousness that Cerquiglini found in the electronic medium was perhaps the most debated feature of late twentieth- and early twenty-first-century textual scholarship. This topic is covered more fully in the essay by Kirschenbaum and Reside, but it may be useful to place the debate within this longer-range history. In brief, there were those whom Paul Duguid tellingly referred to as “liberation technologists,”23 who saw the apocalypse in the move from print to digital transmission, and found that the very ontology of text had been changed.

A history of textual scholarship

39

These critics were all too frequently carried away by their own enthusiasms, claiming, for example, that the “revolutionary” goal of hypertext was “freeing the writing from the frozen structure of the page” and “liberating the text,”24 which will “disempower . . . the force of linear print” and “blow wide open” the social limits of the codex book to “create that genuine social self which America has discouraged from the beginning.”25 While the rhetoric of the liberation technologists might have been overblown, even risible, it is undeniable that the shift from print culture to digital was as significant a change as the previous moves from manuscript to movable type, from roll to codex, and from oral transmission to written. But, given the platonism that underlay much of Tanselle’s agenda, it should not be surprising that when he was co-opted to write the “Foreword” to an MLA collection on Electronic Textual Editing, he discounted the “hyperbolic writing and speaking about the computer age, as if the computer age were basically discontinuous with what went before,” and warned that “when the excitement leads to the idea that the computer alters the ontology of texts and makes possible new kinds of reading and analysis, it has gone too far.”26 In the second decade of the twenty-first century, it may be too soon to determine which of these two extreme views is likely to prevail (though Matthew G. Kirschenbaum has provided a good antidote, using traditional textual criticism with a technological sophistication).27 There were equally extravagant claims made for the revolutionary nature of print, and equally passionate dismissals of its value (just as when Cardinal Giuliano della Rovere had ordered the copying back into manuscript of the offensive Wendelin Spire print edition of Appian’s Civil Wars). But as we now know, print did not simply and suddenly replace manuscript: manuscript transmission of literature was still a favored medium long into the seventeenth and early eighteenth centuries, and calligraphic handwriting went through several important revivals during the 400+ years of supposed print hegemony. It is true that we no longer have much use for rolls (although the survival of “legal size” paper and the still-common binding of legal papers along the top, not the side, of a document bears the traces of roll inscription). And it may be that, for both price and portability, electronic readers like the Kindle, with the capacity of hundreds of print volumes in a quarter-inch thick tablet, may complement the reading of hardbound books with a new generation to whom the e-book may become the preferred vehicle. But Tanselle was surely right to at least caution that, for example, the sort of problems typified by the concepts of analogy and anomaly, first encountered by the Alexandrians and Pergamanians, will not disappear from the long history of textual scholarship. The history of history as preserved in documents of various types is thus always a balance between an acknowledgment that new materials, new methods of storage and transmission, and new critical principles will

40

david greetham

have to be part of the textual scholar’s armory, while also recognizing that certain basic principles (the role of the author, the function of the audience, the variance in texts) will continue to guide both procedures and judgments of the practice of textual scholarship. The revolutions of the codex, movable type, and electronic media simply exemplify the continual revaluation of theories and practice that textual scholarship has had to confront, so that at this point in the evolution of the discipline it is very unlikely that anyone will be able to repeat the confidence of the 1911 Britannica that there is almost nothing left to do.

NOTES 1 Iona and Peter Opie, The Lore and Language of Schoolchildren (Oxford University Press, 1959). 2 G. Thomas Tanselle, A Rationale of Textual Criticism (University of Pennsylvania Press, 1989), 65, 93. 3 See Luciano Canfora, The Vanished Library: A Wonder of the Ancient World, trans. Martin Ryle (University of California Press, 1990). 4 Athenaeus, The Learned Banqueters, ed. and trans. S. Douglas Olson (Loeb/Harvard University Press, 2007–8). 5 Epistle 337 in Collected Works of Erasmus (University of Toronto Press, 1974 onwards), 15:167. 6 Metzger, 102. 7 A. E. Housman, Selected Prose, ed. John Carter (Cambridge University Press, 1961), 23. 8 Ibid, 23. 9 Ibid, 131. 10 Available in simplified electronic form at http://estc.bl.uk. 11 I am grateful to Jeffrey Drouin for having provided the details of this citation. See http://scienceworld.wolfram.com/biography/Kelvin.html. 12 Encyclopaedia Britannica, 11th edn. (Cambridge University Press, 1911), xxvi:715. (J. P. Postgate). 13 J. Christian Bay, qtd. in Tanselle, “Bibliography and Science,” Studies in Bibliography, 27 (1974), 55–89; at 61. 14 Hugh Amory, “Physical Bibliography, Cultural History, and the Disappearance of the Book,” Papers of the Bibliographical Society of America, 78 (1984), 341– 8; at 343. 15 David Shaw, “La bibliologie in France,” in The Book Encompassed, ed. Peter Davison (Cambridge University Press, 1992), 211. 16 D. F. McKenzie, Bibliography and the Sociology of Texts, The Panizzi Lectures (London: British Library, 1986), 5. 17 Ibid., 7. McGann, A Critique of Modern Textual Criticism (University of Chicago Press, 1983; repr. University Press of Virginia, 1992). 18 Piers Plowman: The B Version, ed. George Kane and E. Talbot Donaldson (London: Athlone, 1975). 19 John Trevisa, De Proprietatibus Rerum, ed. M. C. Seymour et al. (Oxford: Clarendon Press, 1975), 2 v.

A history of textual scholarship

41

20 R. J. Tarrant, “Latin Classical Literature,” in Scholarly Editing: A Guide to Research, ed. D. C. Greetham (New York: MLA, 1995), 109. Tarrant cites Paul Maas, Textual Criticism, trans. Barbara Flower (Oxford: Clarendon Press, 1958), 31. 21 Bernard Cerquiglini, In Praise of the Variant: A Critical History of Philology, trans. Betsy Wing (Johns Hopkins University Press, 1999), xiii. 22 Ibid., 37. 23 Paul Duguid, “Material Matters: The Past and Futurology of the Book,” The Future of the Book (University of California Press, 1996). 24 J. David Bolter, Writing Space: The Computer, Hypertext, and the History of Writing (London: Erlbaum, 1991), qtd. in Duguid, “Material Matters,” 73. 25 Richard Lanham, The Electronic Word: Democracy, Technology, and the Arts (University of Chicago Press, 1993), n. 18. 26 Tanselle, “Foreword,” in Electronic Textual Editing, ed. Lou Barnard, Katherine O’Brien O’Keefe, and John Unsworth (MLA, 2006), 2. 27 Matthew G. Kirschenbaum, Mechanisms: New Media and the Forensic Imagination (MIT Press, 2008).

2 KATHRYN SUTHERLAND

Anglo-American editorial theory

From gradual beginnings in the late nineteenth century, the critical study of the composition and transmission of works of modern English literature was put on a professional footing and developed through the twentieth century as sets of procedures for editing major works. Before this, and with relatively rare exceptions, like Samuel Johnson’s “conjectural criticism” (as he called it) practiced on the texts of Shakespeare, and Richard Bentley’s notorious incursions into Milton’s Paradise Lost, textual criticism referred primarily to the establishment of early texts, usually those of classical or medieval works, from manuscript sources; but from the late nineteenth century it came to apply to editorial work in a more general sense and across a wider literary range. Profoundly implicated from the start in this shift was the growing pride in promoting the study of vernacular literature from all periods of history as, in the founding statement of the English Association, “an essential element in the national education.”1 An academic discipline steadily overtaking the classics, at least in provincial British universities in the early years of the twentieth century, English studies at this time appeared to depend for its authority (as distinct from its democratizing reach) upon the same standards for accuracy informing the recently commissioned volumes of series like Oxford Classical Texts. So much so that, while the study of works of English literature came to be associated with progressive thinking and improvement of several kinds, the terms of their institutional presentation and canonical incorporation were firmly rooted in the distant past. Some of the credit can be laid at the door of influential scholars like A. W. Verrall, Arthur Platt, and A. E. Housman, who brought their classical training to bear on a textually conscious reading of modern literature. In a series of essays written between 1916 and 1918 for the Times Literary Supplement from war-torn Macedonia where he was on active service as an artillery officer, R. W. Chapman argued the merits of the following propositions: “There is no humaner science than grammar, and few more exciting pursuits than textual criticism”; “the history of a text is the gradual 42

Anglo-American editorial theory

43

accretion of error”; “The first edition is the very image that first flattered its author’s eye”; “The Shakespeare quartos and the folio of 1623 are the title deeds of a national inheritance”; “The causes to which it is due that the text of Shakespeare is less certain than that of Sophocles are well known”; “It is generally accepted that the most authoritative edition is the last published in the author’s lifetime”; “emendation is only a small part of [modern editors’] duties. The chief is restoration”; “to restore, and maintain in its integrity, the text of our great writers is a pious duty, and it is a surprisingly difficult task.”2 For their time, none of these statements was remarkable; all were part of the common discourse of modern textual criticism as it was then being formulated. The mix of antiquarian and aesthetic interest and of wealthy amateurism that had dominated the rare-book market through much of the nineteenth century was being leavened by a new academic specialism in the historical transmission of texts and, in particular, in the technologies of print; at the same time and by the same process, editing was becoming less opportunistic and more regulated. In Chapman’s phrases we hear intimations of the new laid upon the old. Back in England after the war, Chapman soon held the influential position of Secretary to the Delegates (in effect, publisher) of the Clarendon Press at Oxford University where he formed policy for a series of Oxford English Texts, companions to the Oxford Classical Texts. He himself went on to become a distinguished editor of Samuel Johnson and Jane Austen; his 1923 edition of the Austen text inaugurating the modern critical engagement with Austen but also the scholarly investigation of the English novel as a literary form. His scrutiny of the text of the Austen novels, after the careless reprint history of the later nineteenth century, is the first such treatment of any English novelist. Specifically, Chapman undertook the first full collation of those novels for which we have more than one edition published in Austen’s lifetime and whose revision she was known to have overseen (that is, Sense and Sensibility and Mansfield Park). In keeping with the general theory of his day – that final intentions, as represented by the latest authorized version, are best – he enshrined for both Sense and Sensibility and Mansfield Park the second-edition text as the basis of all modern editions. His analysis recognized two things: the potential for deterioration in relatively modern texts; and Austen’s significance as a painstaking and intelligent rewriter, a novelist exercising extensive oversight of her text. It seems likely that Chapman’s model for the great English novelist was Henry James, the great American novelist; the massive and elaborately retrospective twenty-four volumes of the New York edition of The Novels of Henry James, over which James had labored so hard, had been issued by Macmillan in Britain in 1909; it was a supreme moment in the author’s control of his text. Chapman’s Austen served until the end of the twentieth century as the most authoritative

44

kathryn sutherland

and the basis for every other edition. Yet even as he persuaded the specialist and general reader of the unlikely truth that there were important issues of accuracy and reliability at stake in preparing the text of a mere modern novelist, his method, a Johnsonian mix of scrupulousness and belletrism, was contested. A vigorous challenge was launched by means of an association of bibliography and science. For W. W. (later Sir Walter) Greg, R. B. McKerrow, and A. W. Pollard, the most influential English-speaking textual scholars of the first half of the twentieth century, the new emphasis on the history of texts meant systematic analysis, technical history rather than humanist enquiry, and an unwavering concentration on physical evidence and the facts of transmission – historical recension rather than interpretative emendation. Accordingly, the scientific base of enquiry was not textual criticism but bibliography: the dispassionate analysis of every aspect of book production. All three were Renaissance scholars, specializing in works from the first age of print. In the common absence of manuscript evidence from this period, in the form of authorial working papers or printer’s copy, texts were best understood in light of the early modern operations of the Elizabethan and Jacobean printing-house and book trades, a complexly subdivided work force representing accountable stages in text production: paper-makers, scribes, printers, binders, booksellers, and the guilds and regulations by which these agents were constrained. As Greg summarized the new method towards the end of his career, this “fresh approach to textual problems” places “stress upon the material processes of book-production, concerning itself primarily with the fortunes of the actual pieces of paper on which the texts were written or printed, and the vagaries of scribes and compositors.”3 Greg himself undertook pioneering research into the practices of Elizabethan printers in such matters as casting-off (the term for the printing-house procedure for calculating the length of a manuscript before setting it in type) and proof correction. He investigated the business of dramatic companies and their relations with playwrights and censors; and he distinguished the hands of particular dramatists and playhouse scribes. Earlier, in his 1928 “Principles of Emendation in Shakespeare,” Greg set aside a whole tradition of nineteenth-century scholarly editing when he argued that “No emendation can or ought to be considered in vacuo, but criticism must always proceed in relation to what we know, or what we surmise, respecting the history of the text.”4 His 1932 pronouncement that a text is “a living organism which in its descent through the ages, while it departs more and more from the form impressed upon it by its original author, exerts, through its imperfections as much as through its perfections, its own influence upon its surroundings” makes it clear that history is not the enemy but the maker of meaning. Textual analysis must

Anglo-American editorial theory

45

proceed skeptically, but history is substance and consequence as well as the vehicle of transmission, ensuring that “[a]t each stage of its descent a literary work is in some sense a new creation, something different from what it was to an earlier generation, something still more different from what it was when it came from the author’s hand.”5 The late twentiethcentury idea that the discernment of authorial intention should not govern scholarly investigation into relatively modern works of literature was at this stage virtually unthinkable. Yet Greg’s search for evidence of authorship was always constrained by an imaginative understanding of history: technology may come first to hand in weighing the facts of any literary case, but a far denser web of history informs every action and every artifact. His sense that neither literary works nor their authors and editors can stand outside history makes him a pragmatist rather than an idealist and grants to all textual witnesses of a work a degree of authority and meaning. The immediate intellectual origins of this historical enquiry, what subsequently was labeled the New Bibliography,6 lay in late nineteenth- and early twentieth-century technological and legal developments – in new ways of thinking about intellectual property associated with new reproductive technologies and adjustments to English copyright law. Contemporary change goes far to explain the twin preoccupations of its practitioners to detect and protect the original authority or impression of the author’s mind, and also to trace the processes of reproduction and transmission by which (and only by which) texts live.7 At the same time, an intense dedication to the print forms of Shakespeare and other Renaissance play texts, suggests something of the challenge or paradox the first generation of New Bibliographers faced in scrutinizing the evidence of transmission in the author’s interests: not only do play texts have other lives than print where print is only one possible medium of reproduction, but, crucially, without the tangible evidence of authorial agency that the draft manuscript provides in later periods, print and the means of transmission do not form a solution from which authorial intent can reliably be precipitated. Only later, however, from the 1960s, and in association with the idealizing methods of the American theorists Fredson Bowers and G. Thomas Tanselle, does Greg’s pioneering enquiry into historical processes take on the settled tone of a forensic investigation into the text as the scene of a crime. The distinction is immensely important and not voiced regularly enough, for it represents the dividing line between a whole tradition of British and American editorial theory and practice down to the end of the twentieth century. On the British side of the divide, history is a persistent and messy intermediary between author and reader. On the American side, and less equivocally than Greg, Bowers was in the business of ideal authorial reconstruction. His ideal text, determined, after Greg’s method, from scientific consultation of multiple material copies and specifically from

46

kathryn sutherland

an intimate knowledge of the processes of production, represented the purified text as a distillate. Even the clean look of the edited American page with critical apparatus tidied away at the back of the volume (the encounter of author and reader unsullied by history) signaled an important departure from what would remain standard British editorial practice, where the record of variant readings, deposited like archaeological layers at the foot of each page of text, continue to provide both a historically compromised and a more classical look. Between them, Greg, McKerrow, and Pollard set the agenda for bibliography and textual criticism for at least half a century. Pollard directed the work of the London-based Bibliographical Society for over forty years and was long associated with its journal, The Library. Chiefly he contributed to the bibliography and textual criticism of Shakespeare. His Shakespeare Folios and Quartos: A Study in the Bibliography of Shakespeare’s Plays, 1594–1685 (1909) argued, against views accepted since Edmond Malone at the end of the eighteenth century, that some of the quartos may have been set from autograph copies, and it remains a landmark in Shakespearian criticism. Compiled with R. G. Redgrave, the Short Title Catalogue of Books Printed in England, Scotland, and Ireland, and of English Books Printed Abroad, 1475–1640 (1926) was a major achievement of virtual construction whose catalogue numbers and entries, now online as the E[lectronic]STC, underlie the records of the texts we regularly consult on EEBO. McKerrow’s edition of the works of the Elizabethan prose writer Thomas Nashe appeared in five volumes between 1904 and 1910, and was at once recognized as setting a new standard in English editing. In 1925 he founded the ongoing Oxford journal Review of English Studies, which he continued to edit until his death. His Introduction to Bibliography for Literary Students (1927) provided the standard university textbook on the subject in Britain until it was replaced in 1972 by Philip Gaskell’s A New Introduction to Bibliography. McKerrow’s last major project was a critical edition of Shakespeare for the Clarendon Press. But only his work on the early plays was finished and only his Prolegomena for the Oxford Shakespeare (1939), laying out his principles and methods of text construction, was published before his death. Greg’s editorial reach was more flexible than McKerrow’s, extending to questions of manuscript as well as printer’s copy. Although he produced few editions with critical text and commentary, his probing textual studies and essays shed light on specific editorial issues. Less successful was the thesis that grew out of his work on the miracle plays, The Calculus of Variants (1927), an attempt to expand on the suggestion that mathematics might provide a model for weighing with greater certainty the material evidence for variance in literary texts – a “calculus” for generating a predictive scheme of textual deviance from an ideal norm (and thereby redressing it).

Anglo-American editorial theory

47

But his late essay “The Rationale of Copy-Text” (1950) expanded McKerrow’s restrictive definition of that concept and in this form became an almost universal law within later twentieth-century Anglo-American editorial theory and practice (see below). His parallel-text edition Marlowe’s Doctor Faustus, 1604–1616 (1950), his Editorial Problem in Shakespeare (3rd edition, 1954), and The Shakespeare First Folio: Its Bibliographical and Textual History (1955) remain major achievements. Greg was a subtle and a supple thinker. At the same time as he advocated a “mature science,” a “science of the material transmission of literary texts,” with the technologies of bookmaking as the key to that process, he also insisted that such methods be at the service of something more. “[B]ooks are of importance only as the vehicle by which . . . contents reach us.” In weighing the price of history and the usefulness of scientific method, he registered the value of a “self-conscious” bibliography. For Greg, there is an essential link between bibliography, textual criticism, and literary criticism: bibliography and textual criticism make of literary criticism an articulate activity able to communicate its insights at a profounder level. “Thus in spite of my interest in bibliography,” he explained, “it is as the handmaid of literature that I still regard it.” The phrase he borrowed, not without controversy, to describe the infusion is “bibliography is the grammar of literary investigation.”8 Greg refined much of his thinking in the 1930s, when English literary studies had little professional edge – little to distinguish their insights from an untutored impressionistic and sensational response. He drew attention to the instructive opposition and friction between bibliography and literary criticism. Only in the last twenty years, and only embodied in a direct critique of the New Bibliographical method, has the link between textual and literary critical studies again been vigorously urged; now, however, as the inevitable conjunction of a socialized view of texts as documentary forms and as the situated acts of reading rather than authorship. Greg considered an interest in the printed page (as distinct from early, usually Classical, manuscripts) as “mainly a development of the twentieth century, largely even of the post-war years.”9 He offered this view in a 1933 lecture, and it is worth considering how far world war – its remembrance and, in the 1930s, its growing anticipation – conditioned the brand of textual ideology with which his name became associated. Not only was this the case during the 1920s in the aftermath of the Great War, but even more so in the years following the Second World War, when Greg’s American disciple Fredson Bowers was establishing in Charlottesville, Virginia, the influential periodical Studies in Bibliography, dedicated to serving the interests of the restored text. In weighing the price of history in the preservation and transmission of literary texts, twentieth-century British and American textuists

48

kathryn sutherland

not surprisingly registered the burden of their own historical moment. Editing as the preparation of serviceable texts, like the principles or theories that underlie its practices, confers authority and that authority, however explained or defended, is subject to its own historical present. In the words of the textual theorist Louis Hay “editing has always embodied the main ideological and cultural concerns of its day.”10 This is its strength and its weakness: the terms and the limits of its power. All editorial theories imply the authority to represent or speak as the text; and all are ultimately revealed as temporal and temporary protocols for interpretation. The New Bibliographic project now appears both elegiac and impossibly optimistic. As James Thorpe summed it up in his relatively late handbook, Principles of Textual Criticism, “[d]ecay inhabits the works of the past”; and, more expansively, “A text is never self-correcting or self-rejuvenating, and the ordinary history of the transmission of a text, without the intervention of author or editor, is one of progressive degeneration.”11 Left to themselves, or rather to the processes of history, texts deteriorate; but authors and their agents (editors) can arrest this process. In this way of thinking, editing is restorative, healing, and its practitioners stem erosion. By contrast, in the more recent, opposing, point of view, any activity that banishes signs of historical passage from a text “obliterates (rather than preserves),” and even represents “a form of institutionally sanctioned forgetting.”12 The rhetoric of romance in which the scholarly editor regularly cloaked his otherwise invisible activities through much of the twentieth century should be seen in the light of the challenge it issued to the forces of time as they operate on the texts of literary works. The history of textual criticism and editorial theory for most of the twentieth century was a history of progressive eclecticism born of an increasing distrust of (or, perhaps, a growing confidence to reverse) the conditions under which texts are transmitted. But if from its beginnings a sense of intense loss precipitated the New Bibliographic textual project, it is also true that the note of alarm and moral urgency to reverse the “degeneration” is a late and peculiarly North American reading of text in history. For Greg, history is not the enemy but the maker of meaning; nevertheless, its processes are detectable as errors that compromise authorial intention, and it is the task of the critical bibliographer to intercept them. But only towards the end of his career, in his suggestion for a divided authority in copy-text, did he endorse the practice of eclectic editing. Like McKerrow, whose defining statement on copy-text (“the copy-text, by which . . . I mean the text used in each particular case as the basis of mine . . . followed exactly except as regards evident misprints”13 ) he had hitherto adhered to, Greg was not impressed by the older idea that a critical text emerges from active editorial choices among variant readings. The procedure smacked too much

Anglo-American editorial theory

49

of subjectivism. Similarly, it was far from the view of either that editing is openly constructive or creative of a new text as opposed to narrowly emendatory. Both worked by the rule of the latest lifetime edition likely to have been overseen by the author as the basis, or copy-text, for a new edition whose readings would only be altered at obviously erroneous places and on the strength of bibliographical evidence. Greg’s aim, as he described it in his Clark Lectures for 1939, was “to present the text . . . in the form in which we may suppose that it would have stood in a fair copy, made by the author himself, of the work as he finally intended it.”14 But partly as a result of his study of Shakespeare and of Marlowe’s complexly versioned Dr Faustus, it became clear to Greg that in this period, with the obvious exception of Ben Jonson, authors (and certainly dramatists) could not be assumed to have seen their work through the press. To recognize the actual circumstances of Elizabethan and Jacobean printing practices (and the limits on an authorial concept of intention derived from non-autograph witnesses) he suggested a distinction between the “substantive” readings of the text (the words, which “affect the author’s meaning or the essence of his expression”) and its “accidentals” (“such in general as spelling, punctuation, word-division . . . its formal presentation”). Upon this distinction between essential and formal features, into which all texts can be divided, he built a double “rationale” to guide the editor: one text, usually (though not invariably) the earliest complete state, of a work should be chosen as copy-text, and that should be followed in matters of accidentals, while later states should be consulted for substantive alterations wherever it could be shown that these later changes were authorial. The critical edition resulting from such a procedure would no longer be a single corrected witness to the work but potentially a composite form built from several witnesses and incorporating both early features (accidentals) and late revisions (word changes). It would be eclectic, but an eclecticism underwritten by informed investigation into the work’s textual recension and not by a reliance on personal preference.15 Greg was nowhere more supple than in his refinement of McKerrow’s original formulation of copy-text editing. For what sustains his reconsideration is a practical division of authority between author (words) and history (formal presentation) in realizing the dynamic of the text, with history only appearing to gain the ascendancy. The control or copy-text is chosen for its historical features, on the well-corroborated assumption that no copyist, of either manuscript or print, feels bound to follow an author’s preferred punctuation; individual readings of words, by contrast, are selected on the basis of probable authorial re-entry into the text. Authorial revision does not override history but operates within its bounds; history does not override authorial intervention but accedes to it. Or rather, history accedes to the

50

kathryn sutherland

editor’s inferred understanding of the author’s likely behavior with regard to his text, and as Greg admits, in a characteristic turn around, the basis for this can be “frankly subjective,” especially when it comes to judging the claims of particular readings: should an “inferior” late authorial reading be admitted to the text?16 Only taken to its extreme, however, in the writings of Bowers and Tanselle does Greg’s theory lose its poise and his eclectic history of the text become an editorially inferred intention for the text. When Greg writes, in “The Rationale of Copy-Text,” of the “author’s original” the empirical evidence provided by physical documents appears to be in question; for Bowers, ten years later, the evidence itself and not just its interpretation has become “inferential.” His essay “Textual Criticism,” circulated as part of a Modern Language Association of America pamphlet setting out an agenda for future American scholarship, may not have been issued as an official statement on editing, but it gained that status. It showed the usefulness of Greg’s procedures to editing post-Renaissance authors (Fielding, Sheridan, Shelley, Hawthorne, Whitman, F. Scott Fitzgerald, Sinclair Lewis), and it extended the limits of the editor’s task to include “an attempt to approximate as nearly as possible an inferential authorial fair copy.”17 For Bowers, with his focus already on modern (nineteenth-century) texts, the regularly surviving evidence of authorial manuscripts overturns Greg’s dual authority, shifting the eclectic text of editorial construction towards a more exclusive witness to authorial practice, distinct from even the regular operations of history upon it. So, in the case of The House of the Seven Gables, The Blithedale Romance, and The Marble Faun, despite Hawthorne’s approval of thousands of accidental variants between his manuscript and its first printing, in Bowers’s view, the editor should restore manuscript punctuation because “the real flavor of Hawthorne . . . can be found only in manuscript.”18 The author set free, in some paradoxical way, from the constraints of print publication by the intervention of his (print publishing) editor. For Bowers, analytical bibliography, with its emphasis on mechanical explanation, appeared to set standards of proof that argued editorial activity as a pre-critical act, a neutral or “pure” form of textual restoration. At its extremest, his melding of scientism and moralism turned Greg’s detection of printing-house errors and irregularities in transmission into the rooting out of sin. “One can no more permit ‘just a little corruption’ to pass unheeded [in a text] than ‘just a little sin’ was possible in Eden.”19 Bowers’s important study Principles of Bibliographical Description (1949), which he dedicated to Greg, had already signaled the decisive dehistoricizing shift in the status of material evidence by establishing the authoritative notion of the “ideal copy,” the control for determining the physical make-up of the book as it

Anglo-American editorial theory

51

was intended as distinct from realized in the act of printing.20 In his subsequent transference, to the editing of relatively modern nineteenth-century American works, of procedures established by McKerrow and Greg for correcting the texts of sixteenth- and seventeenth-century literature, there can be detected a prescriptive cultural agenda. The functions Bowers famously set for textual criticism, the “recovery of the initial purity of an author’s text . . . and the preservation of this purity,”21 procedures which he argued lie “at the base of all intellectual endeavor in our cultural heritage,” were soon instituted and monitored, in the case of American editions, through the agency of the Center for Editions of American Authors (CEAA), set up in 1962–3 and in 1976 restyled the Committee on Scholarly Editions (CSE). Greg died in 1959 but already the textual critical axis had tilted westwards, and developments for the rest of the twentieth century would be dominated by American thinking – much of it in contention with Bowers’s model in so far as it implied impartiality, rigorous method, and uniformity. These were the years when academic presses in Britain and America began to commission ambitious critical editions of post-Renaissance literature, extending principles established for dealing with the patchy evidence from Elizabethan printing houses to the more heavily documented practices of modern authors and later literary genres. Also in 1959, Bruce Harkness published an important article in Bowers’s Studies in Bibliography, in which he argued the existence of what he called “the novelistic fallacy”: the fallacy being that to critics and students, dedicated to the close reading of literary works, the source of the text of a novel does not matter. Unlike a play or poem, for which a well-edited and well-printed text were by then held to be the standard starting point for critical investigation, the cheap reprint of an error-strewn novel-text was still widely considered good enough for all specialist purposes. According to Harkness, “People . . . commit bibliographical nonsense when handed a novel.” But as he also conceded, “A false word in a sonnet may change a fifth of its meaning; the punctuation at the end of the “Ode on a Grecian Urn” can be considered crucial to the meaning of the whole poem; but who . . . can stand the prospect of collating 700 pages of Dickens to find a few dozen misplaced commas?”22 Summarized thus, the essay’s vigorous counter-argument for the importance of inspecting the print history of novels might have struck an implausible note, even in the ears of the committed textual critic, its assumed reader. What does it mean to say that the modern literature we buy, borrow, and read might be in need of correction? Are we wrong to take texts on trust? Fifty years on much has changed. Within a few years of Harkness’s call for a critical revaluation of the novel (which is what his plea for accurate texts amounted to), the CEAA enshrined rigorous procedures for editing nineteenth-century American fiction – at this point, according to eclectic

52

kathryn sutherland

and idealist principles.23 About the same time in Britain, Kathleen Tillotson’s 1966 Oxford University Press Clarendon edition of Oliver Twist launched the first full critical edition of any major English novelist since Chapman’s Novels of Jane Austen in 1923. Despite local differences of approach, editors on either side of the Atlantic were joined in believing it might be possible to make critical editions, accessible to specialist, student, and general reader, and of such an accuracy and comprehensiveness as never to need editing again. Bowers’s own edition of The Works of Stephen Crane, completed in ten volumes between 1969 and 1975, is an excellent witness to the painstaking nature of editorial labor so defined. Issued with the stamp of approval of the CEAA, it presents a clean-page eclectic text, described in the case of Maggie as a synthesis of the ideal forms of two historical witnesses, followed by separate registers of substantive and accidental emendations and a historical collation of the editions consulted, all listed as a tabular backof-book apparatus from which readers might (in theory) reconstruct the work’s complete textual history and, if they wished, weigh or question at every point the reading established by the editor from multiple variants. In Britain, a concern for evidentiary display resulted in divided reading zones for the text-page and variants set as footnotes. In each case, the complexity of apparatus and protocols substantiated in some profound if ultimately unclear way the cultural value of the object in hand – in these instances, no longer mere novels. The faith of later twentieth-century practitioners lay in what we might call ideal stasis: the establishment, through emendation and the eclectic compression or more discrete variant presentation of a range of textual states, of as complete a witness to the author’s intentions for and labor upon the work as the inevitably corrosive laws of history and the material limits of the book would allow. Tillotson, for example, held as an important element of her edition the recovery of early periodical versions of Dickens’s novels wherever they had been cropped in production. At the same time, she believed his texts were only fully expressed in late lifetime book editions, and that it was equally relevant to include, as a feature of each Clarendon edition, Dickens’s number plans and working notes where they survived. Accordingly, the Clarendon Dickens texts represent in their compressed textual display his earliest and latest published versions; at the same time, they extend authorial influence beyond the text (any text) prepared for publication to include its earliest genesis. For Oliver Twist Tillotson detected “five substantive editions,” representing “five stages of revision.” The novel had begun to appear in monthly installments in 1837 but by Tillotson’s account “the text achieves stability only in 1846. Until then, Dickens was still writing his novel.”24 The characteristic Clarendon apparatus provides a dense fabric of variant readings from the manuscript through to the 1867

Anglo-American editorial theory

53

Charles Dickens edition (eleven different texts), intruding upon the 1846 copy-text seemingly settled on the page above it.25 A criticism leveled at Bowers was that he had created of the text of Crane “muddles of editorial synthesis and intervention . . . an unwillingness to leave Crane as he was,” with the consequence that “All the texts are battered by the intrusion of editorial conformity.”26 The contrast between the historically burdened Clarendon page and the timelessness of the Virginia Crane, cut loose from its apparatus, may appear to suggest an intellectual distance between two conceptions of the edited text: the one parading its casualties; the other a step closer to forgetting them. In fact, both trade in similar illusions. The Herbert Rosengarten and Margaret Smith edition (1984) of Charlotte Bront¨e’s Villette, for example, has the regular Clarendon page layout but its intricate textual weaving of early and late states provides an equivalent ideal synthesis. Its copy-text, the first published edition of 1853, is corrected in two directions: backwards, with readings restored from the manuscript and from a copy made up from first author proofs, and forwards, with readings introduced from the Nicholls copy (an imperfect copy of the first edition annotated by Bront¨e). The method transforms the text at various moments into ideal conformity with what it was (and is no longer) and what it might become (but is not yet), effectively unpicking, in the author’s interests, a production process she clearly endorsed. The compression, undertaken in the name of authorial intention, sits oddly with the Clarendon policy of open gestures to history – not least, its distinctive practice of prefacing all its critical editions with photo-facsimiles of firstedition title pages. History – under ideal conditions. Tanselle’s Rationale of Textual Criticism (1989) pushes elements in this late twentieth-century approach to one logical if extreme conclusion by arguing an essential distinction between the literary work and any and every documentary support that may carry it. Do literary works exist apart from attempts to reproduce them? His conclusion is that they do, and that this essential intangibility confers special status on critical editions: “for the work can only be reconstituted through the application of critical judgment to each element of every surviving text (even if there is only one, and it is the only one there has ever been).” Granted the profound indeterminacy of works of literature (as distinct from paintings or sculptures), “there is no way that we can talk about such works without assuming the responsibility of deciding just what we are talking about – without in effect becoming editors.” This being so, the best editors are expert editors, for “the reconstructed texts created by scholarly editors represent exemplary acts of reading by persons with specialized historical knowledge.” Further, the best editions are eclectic editions, and “Those editors who maintain that they cannot take readings from different editions because they do not wish to mix versions

54

kathryn sutherland

together – and there are such editors – have failed to understand how the texts of documents are different from the texts of works.” This is so because, finally and emphatically, “Verbal works, being immaterial, cannot be damaged as a painting or a sculpture can; but we shall never know with certainty what their undamaged forms consist of, for in their passage to us they are subjected to the hazards of the physical.”27 Tanselle implies a shared ground between his textual “rationale” and Greg’s of almost forty years before; but the reconstructive role he claims for editorial activity marks a distinction between his nature of texts and Greg’s history of texts. Greg’s concern was with the failures of historical process, discernible from material evidence; Tanselle’s is the gap between a Platonic notion of the literary work and the conditions of expert investigation under which its shape might most closely (never fully) be identified. It is the gap between a phenomenology and an ontology of text. A feature of debate from the 1970s on was the implied distinction, even opposition, between the terms and concepts of “work” and “text”: between, on the one hand, an authorially intended idea of the work, and on the other, its realization as a text, where the text can seem no more than a local manifestation of something that will always elude full expression. That has been Tanselle’s view; but the same “work” “text” distinction was simultaneously argued in precisely opposed terms, and some would say far more influentially, as far as cultural theory was concerned, by Roland Barthes in his 1971 essay “From Work to Text,” with its allocation of material contingency to the work and of irreducibility to text. For textual criticism, a kind of compromise position can be found in Jerome McGann’s later contention that the only mode of existence of a work is as a text – that all we ever have are physical expressions (“Literary works do not know themselves, and cannot be known, apart from their specific material modes of existence/resistance.”)28 By contrast with such pragmatism, Tanselle’s suspicion of all socially mediated forms ennobles the editor’s bold recourse to intention rather than circumstance and makes of the text he reconstructs a hypostasization – text given substance as concession to the physical, a compensatory resource. Such extravagant thinking marks the end of the line for New Bibliography. Yet the questions its practitioners raise concerning our encounters with literary works – questions to do with history and authority, circumstances of transmission and the technologies of reproduction – have not gone away; only our ordering of the counters in the argument has shifted, and the range of technologies at our disposal. As a means of establishing text the New Bibliographic model was only seriously challenged from the 1980s and (not coincidentally) around the time when technology in the guise of the computer appeared to offer the opportunity to simulate the contingency and hybridity of all textual forms, as celebrated in contemporary post-structuralist

Anglo-American editorial theory

55

theories. Only now, and thanks to the computer’s massive storage facilities, do we appear to be in a position to reclaim literary works by other than synoptic and eclectic means – currently, under the sign of their rich textual histories and their multiply contextualized forms. Literary works can now seem as dauntingly uncontained as previously they were defined or limited by the choices of their expert editors. In some senses this is to do no more than turn the old method inside out – a scaled-up or macro-eclecticism which owes its existence to celebrating the conditions of history rather than defying them. Thus far, it might be said, technology and ideology remain complicit. In fact, another way of representing the New Bibliographer’s thinking about the nature and condition of text is to invoke print technology itself, the thing its practitioners came to distrust. The concept of the perfected, ideal text, like that of the text as a mirror in which we find the image of the author, is a Romantic notion and was born of the kind of uniformity in print only mechanically possible from the later nineteenth century. Up to then, it was generally difficult if not impossible to guarantee the identity of printed copies – variation, rather than a disease or imperfection or corruption of print text, was its normal condition and a function of the technology that transmitted it. In the course of the nineteenth century the commercial use of stereotyping made it easier to reduce the play of typographic variation between copies and between impressions and to render the appearance of the printed page uniform and seemingly neutral. This material change in the appearance of text eventually informed a concept of text at the immaterial level and legitimated an intellectual turn to perfection and purity. In embracing what was new in the technologies at their disposal, the New Bibliographers put faith in their analytic enquiry into earlier print technologies as a means to reverse historical change, creating from their eclectic procedures unified texts which would be general objects of cultural agreement and by extension the material exemplars of a literary canon. Only in recent years and as a consequence of the electronic storage of digital surrogates of print and manuscript objects have we recovered an appreciation of the historically expressive modes of existence of literary works. As with the New Bibliographic agenda, whose purpose was the stabilization in print of the error-strewn processes of print, the new faith in electronic technology is built on paradox: in this case, the digital fetishization of print objects. According to the one paradox we must trust the print editor to eliminate rather than repeat the errors endemic to print; and in the other, we put faith in what has been called screen essentialism – an uncritical endorsement of the computer’s power to reproduce the real, uncompromised and undistorted by its own materiality. Both ways of thinking need to be recognized for their timeliness rather than their essential truth; both are responses to the

56

kathryn sutherland

burden of media consciousness, informed by new technologies, by which the materiality of our traditional literary culture (the functions of type, paper, format, book structures, etc.) comes into sharper focus. Already in 1972, James Thorpe, an eloquent exponent of the dominant eclectic tradition underlying projects like Tillotson’s Dickens and Bowers’s Crane, had questioned its universal applicability when he argued that “works created by authorial revision” may be in each version, “either potentially or actually, another work of art.” Turning to the notion of final intentions, enshrined since before Greg’s “Rationale of Copy-Text,” he suggested with devastating simplicity: “It is a bit puzzling to know why this dictum should for so long have passed unchallenged. For it is much like saying that an author’s last poem (or novel, or play) is, as a general rule, his best: it may be, and it may not be.”29 The point was driven home when in 1988 Stephen Parrish, one of the general editors of the Cornell Wordsworth series, declared his “dissent from Whig interpretations of a literary text, with their notions of an inner logic of inexorable growth toward what could have been foreseen from the start as the author’s final intention.” On the contrary, with the manuscript and print evidence of the actual practices of a range of Romantic and post-Romantic poets in mind (in Parrish’s case, Wordsworth and Yeats), it appeared more correct to “plead the autonomy and the validity of each steady state of the text as it changes in a confused, unpredictable way, through patterns which the author may never have foreseen, let alone ‘intended’.”30 The sheer weight of evidence available for Wordsworth’s poetry argued against an ideal organic unity and in favor of textual layers or directions – now this way, now that. In a bold move, the Cornell Wordsworth provides comprehensive facsimile images of manuscript pages that effectively decompose its editorially inferred “Reading Texts.” Still ongoing, the Cornell Wordsworth remains a landmark in its self-conscious questioning of the limits of print-based editing and the theories that underpin it. Further attacks from within the intentionalist camp came from Jack Stillinger who promoted the authority of version-editing or “versioning” by drawing attention to the multiple authorized forms of specific Romantic poems and the facts of joint authorship for many literary works – Eliot’s Waste Land, for which we have no obvious single authoritative text, being only one example.31 More recently still, Joseph Grigely revived the extreme language of Tanselle – in this case, in an argument for the prosecution – when he denounced Bowers’s purist eclecticism on the grounds that it is informed “by the residue of eugenic ideology surviving in the discourse of midcentury idealism.”32 The charge hardly seems warranted in its shocking and distorted conjunction of eclectic textual bibliography with some of the most monstrous twentieth-century political movements for control and

Anglo-American editorial theory

57

“social betterment.” And yet, the train of thought it induces helps expose something curiously contentious in the language of so arcane a study as textual criticism: its rhetoric (of error, castigation, purification) extends back to classical times, its metaphors are as dead as can be – except that they hold the potential for relevance, for life, and the possibility that our engagements with texts might, after all, be complicit with more apparently urgent social concerns.33 Where once the ideal, composite text, freed from the mishaps of transmission, delivered what could be defended as the author’s intended appearance before a readership, now the dependency of the transmission process itself determines the course of most discussions, including how we think about authors. Recent moves among editors and theorists have been in two directions: towards text-based theories grounded in the literary work’s collaborative status as a socially constructed product, and towards a reconsideration of authorial intention as promoting unstable versions. The former, social model, with its roots in reception theory (or in American varieties of readerresponse criticism) is associated with Jerome McGann and D. F. McKenzie; the latter, with its directly opposed presentation of text as an act (or acts) of writing (not reading), is represented in English-speaking textual circles by the European-based “genetic” theory and practice of Hans Walter Gabler, best known for his edition of James Joyce’s Ulysses (1984).34 In both reconceptions, what is in question are the nature and limits of intention. Whose intentions matter in regard to a literary work? In what sense can we say that authorial intention establishes a literary work?35 Which embodiments and aspects of the work: the holograph manuscript only? or the printed text also? and what about the reader’s text? How passive or active is the reader in producing meaning? Then again, what is an authorial intention? – an intention to mean or to do? How fixable is intention? – do authors intend singly, serially, multiply? What such questions and their variously inflected answers undoubtedly establish is the status of editorial theory and practice as culturally embedded critical endeavors. As such, they at last openly acknowledge that they can no longer deal definitively with text, but interpretatively, subjectively, and, of course, temporarily. Further concentration on the post-1800 literary tradition with its plethora of forms (autograph and non-autograph, pre-print and print) threatens to deconstruct not only old-style editorial activity but the editorial project in general. As one recent electronic bibliographer puts it, and in conscious deformation of Bowers, “every reader his own bibliographer.”36 Paradoxically (we are never far from paradox in thinking about theories of text), there is a real problem in recognizing the powerful critical work of the eclectic printed edition. Its confident determinacy can make it appear disengaged from or beyond the critical acts performed upon it. By contrast,

58

kathryn sutherland

the recent suggestion that works only exist under the sign of change and that such change is what signifies has been used to bring the bibliographical methods of textual criticism to bear more widely on what previously seemed exclusively matters of literary interpretation. Accordingly, much editorial activity has been reformulated as distributed and provisional, a matter of the informed selection of particular readings for particular purposes – that is, as contingent critical acts. But if by this we are all now textual critics, where does that leave the peculiar textual activity of the editor (neither intending author nor openly participating critical reader), whose work traditionally lies in distinguishing among and reassembling documentary forms? By what criteria do we now measure a “good” edition or recognize the value of its contribution – to the life of the literary work, to the community of readers, to the wider culture into which it inserts itself? The question is compounded by the new limits set for our ambitions by the computer – more texts in more forms and faster and faster means to manipulate and search them. But we have yet to formulate an electronic equivalent of a theory of textual editing as sophisticated as those of Greg and Bowers. Paul Eggert’s recent call for such an electronic theory is timely. In a related move, Matthew Kirschenbaum has set out to confront some of the shibboleths current in the critique of electronic textuality. Crucial to his argument is a Greg-like insistence on the fundamental materiality of digital writing technologies, and therefore of electronic texts.37 The contention that electronic texts are artifacts or mechanisms which, like the older technology of the book, are amenable to material and historical investigation, challenges editors and textual critics to respond to the new medium in terms of its own materiality, architecture, and functioning, as distinct from those of print. The implication is that the contours of a truly electronic-sensitive theory will only become apparent as we shed assumptions (still too prevalent) of how digitization and encoding enhance or betray the operations and purposes of print. Electronic editions on these terms are still in an early stage of design; our theorizing, though vigorous, is still in its infancy.

NOTES 1 From the “Constitution of the English Association,” cited in C. Nowell Smith, The Origin and History of the Association (English Association Pamphlet, London: Oxford University Press, 1942), 5. 2 R. W. Chapman, The Portrait of a Scholar and Other Essays (London: Oxford University Press, 1920), 10, 49, 51, 66, 68, 77, 79. 3 W. W. Greg, The Editorial Problem in Shakespeare: A Survey of the Foundations of the Text (Oxford: Clarendon Press, 1942; 3rd edn, 1954), 3. 4 W. W. Greg, “Principles of Emendation in Shakespeare,” Proceedings of the British Academy, 14 (1928), 152.

Anglo-American editorial theory

59

5 W. W. Greg, “Bibliography – an Apologia,” in Collected Papers, ed. J. C. Maxwell (Oxford: Clarendon Press, 1966), 259. 6 By F. P. Wilson, “Shakespeare and the New Bibliography,” in The Bibliographical Society 1892–1942: Studies in Retrospect, ed. F. C. Francis (London, 1945); revised as Shakespeare and the New Bibliography, ed. Helen Gardener (Oxford: Clarendon Press, 1970), p. 97. 7 See Joseph F. Loewenstein, “Authentic Reproductions: The Material Origins of the New Bibliography,” in Textual Formations and Reformations, ed. Laurie Maguire and Thomas L. Berger (New York: University of Delaware Press, 1988), pp. 23–44. 8 Greg, “What is Bibliography?” and “The Present Position of Bibliography,” in Collected Papers, 82–3 and 215. 9 Greg, “The Function of Bibliography in Literary Criticism Illustrated in a Study of the Text of King Lear,” in Collected Papers, 274. 10 Louis Hay, “Genetic Editing, Past and Future: A Few Reflections by a User,” TEXT, 3 (1987), 117. 11 James Thorpe, Principles of Textual Criticism (San Marino, CA: Huntington Library, 1972), 51–2. 12 The first phrase is from Jack Stillinger, Multiple Authorship and the Myth of Solitary Genius (New York: Oxford University Press, 1991), 199; the second from Joseph Grigely, Textualterity: Art, Theory, and Textual Criticism (Ann Arbor, University of Michigan Press, 1995), 30, where he references Stillinger’s objection. 13 R. B. McKerrow, in The Works of Thomas Nashe, 5 vols., (London: A. H. Bullen and Sidgwick and Jackson, 1904–10), i: xi. 14 Greg, “Prolegomena” to The Editorial Problem in Shakespeare, x. 15 Greg, “The Rationale of Copy-Text,” in Collected Papers, 374–91. 16 See ibid., 387–8. 17 Fredson Bowers, “Textual Criticism,” in The Aims and Methods of Scholarship in Modern Languages and Literatures, ed. James Thorpe (MLA, 1963), 33. 18 Fredson Bowers, “Some Principles for Scholarly Editions of Nineteenth-Century American Authors,” (1962), reprinted in Art and Error: Modern Textual Editing, ed. Ronald Gottesman and Scott Bennett (London: Methuen, 1970), 59. 19 Fredson Bowers, Textual and Literary Criticism (Cambridge University Press, 1966), 8. 20 Fredson Bowers, Principles of Bibliographical Description (Princeton University Press, 1949), 8, 113–15 (“Ideal copy cannot always be rigidly defined, and with some books may indeed be quite hypothetical”). See the major challenge to Bowers’s approach, generously published by Bowers himself, from D. F. McKenzie, “Printers of the Mind: Some Notes on Bibliographical Theories and PrintingHouse Practices,” Studies in Bibliography, 22 (1969), 1–75. 21 Fredson Bowers, “Textual Criticism,” in Aims and Methods, 24. 22 Bruce Harkness, “Bibliography and the Novelistic Fallacy,” Studies in Bibliography, 12 (1959), 60, 64. 23 The CEAA issued its Statement of Editorial Principles: A Working Manual for Editing Nineteenth-century American Texts in 1967 (based on Bowers’s “Principles” of 1962). The current version of the MLA CSE-approved guidelines includes advice for editing in the electronic medium and are online at www.mla.org/cse guidelines.

60

kathryn sutherland

24 Oliver Twist, ed. Kathleen Tillotson (Oxford: Clarendon Press, 1966), xi. 25 See ibid., 91–2, where an extended passage from the periodical version, recorded in the critical apparatus, risks unbalancing the proportions and authority of the page. 26 David J. Nordloh, “On Crane Now Edited: The University of Virginia Edition of The Works of Stephen Crane,” Studies in the Novel, 10 (1978), 103–4. 27 G. Thomas Tanselle, A Rationale of Textual Criticism (Philadelphia: University of Pennsylvania Press, 1989), 27, 33, 38, 80, 93. 28 Roland Barthes, “From Work to Text” in Image–Music–Text, trans. Stephen Heath (London: Fontana, 1977), 155–64; Jerome J. McGann, The Textual Condition (Princeton University Press, 1991), 11. 29 Thorpe, Principles of Textual Criticism, 37, 47. 30 Stephen Parrish, “The Whig Interpretation of Literature,” TEXT, 4 (1988), 349. 31 Jack Stillinger, “A Practical Theory of Versions,” Coleridge and Textual Instability: The Multiple Versions of the Major Poems (New York and Oxford: Oxford University Press, 1994), 118–40. 32 Grigely, Textualterity, 26. 33 On the rhetoric of textual criticism, see Stephanie H. Jed, Chaste Thinking: The Rape of Lucretia and the Birth of Humanism (Bloomington and Indianapolis: Indiana University Press, 1989); and Gary Taylor, “The Rhetoric of Textual Criticism,” TEXT, 4 (1988), 39–57. 34 For summary early statements, see Jerome J. McGann, A Critique of Modern Textual Criticism (University of Chicago Press, 1983); D. F. McKenzie, Bibliography and the Sociology of Texts (British Library, 1986); Hans Walter Gabler, “The Synchrony and Diachrony of Texts: Practice and Theory of the Critical edition of James Joyce’s Ulysses,” TEXT, 1 (1984), 305–26. 35 The subject is a minefield, but see James McLaverty, “The Concept of Authorial Intention in Textual Criticism,” The Library, 6th series, 6 (1984), 121–38. 36 See Edward Vanhoutte, “Every Reader his own Bibliographer – An Absurdity” in Text Editing, Print and the Digital World, ed. Marilyn Deegan and Kathryn Sutherland (Farnham: Ashgate, 2009), 99, n. 1. 37 See Paul Eggert, “The Book, the E-text, and the Work-Site,” in Text Editing, Print, and the Digital World, ed. Marilyn Deegan and Kathryn Sutherland (Farnham: Ashgate, 2009), 63–82; and Matthew G. Kuschenbaum, Mechanisms: New Media and the Forensic Imagination (Cambridge, MA: MIT Press, 2008).

3 GEERT LERNOUT

Continental editorial theory

The title of this chapter may give the impression that there is such a thing as a single Continental editorial theory, preferably different in interesting ways from an equally monolithic Anglo-American theory. Kathryn Sutherland has demonstrated the variety of perspectives among British and American textual scholars, the tradition that most of the readers of this Companion will know best, but at least all these theorists and practitioners have a common language that enables them to discuss their different and evolving opinions. This is a luxury that Continental textual scholars do not have. A national literature (or in countries like my own, each of three different national literatures) is studied by scholars who have chosen to work in their native tongue and not in a foreign language, which is often taught and studied in an entirely different department. More importantly, these textual scholars tend to discuss the details of their work primarily with colleagues active in that same language. In some cases this language barrier means that one’s peers consist of a very select group and then there is no reason to cross national boundaries to look for colleagues confronted with similar problems. As a result, these scholars’ work is little known outside their own tradition and the individuals themselves often remain unaware of developments abroad. Despite attempts by professional organizations such as the European Society for Textual Scholarship (established as a forum to encourage discussion of methodological issues across linguistic and national borders), the level of cross-fertilization remains low. Even if they do look at other intellectual traditions, scholars tend to rely on a single editorial tradition that for the most part has been borrowed from one of the neighboring cultures, with just a few cultures that have been and continue to be more important. These will be discussed in more detail here, with the German and French traditions as the most influential centers of editorial theory and practice. In the case of the ESTS, the working language is English, but despite the seeming ubiquity of this new lingua franca, many older colleagues working 61

62

geert lernout

in this field do not function usefully in English and thus many potential members of the Society may not even know of its existence. As a result, these scholars’ work is unknown outside their own tradition and the individuals themselves often remain unaware of developments abroad. Yet all of these textual scholars have something in common with other textual scholars in Europe and elsewhere. In the scholarship on literary texts in all the different European languages, textual studies play the same role. In many cultural traditions we find the same tension between scholars who engage in textual scholarship and those critics who concentrate on an already established text, interpret it, place it in its historical or critical context, or deconstruct it in terms of philosophical, psychoanalytical or other theoretical frameworks. Just as in the English speaking world, European exponents of the textual approach often engage in theoretical debates with their fellow editors. Yet the details of these discussions and the precise differences of opinion are almost never the same and when they do appear to be related, this impression may be based on the fact that often similar terms are used: when confronted with seeming parallels we should never forget that words such as work, text, variant, fragment, or version, can denote many different things. There is little to be gained by translating these discussions all too hastily into terms that make sense in the current debates among AngloAmerican textual scholars. Each of these cultural or national traditions may have its own unique history, a history that we cannot even begin to do justice here. In the last decades two factors have been important for a new climate that can often be observed among younger scholars: the opening up of Eastern Europe and the increased mobility of researchers, with students more often studying abroad. Graduates from other European or American universities bring home new ideas that are now beginning to influence local traditions, a process that is not always without its own difficulties, especially in academic contexts that can still be very hierarchical, leading to conflicts that tend to be reported, just as on CNN, only by those participants who are able to communicate in English but who are not necessarily more accurate (or even sufficiently fair) than their opponents who do not master English. We can therefore only try to be fair in our account of developments in European Continental theory and practice, but certainly not exhaustive. In fact the great majority of national and cultural traditions in Europe will not even be mentioned, a decision that is not just due to lack of time and space, but also to the writer’s inability to read Portuguese, Romanian, Russian, Swedish, Polish, and so many other European languages. My ambition can merely be to present a first introduction to the often quite disparate ways in which textual scholarship has developed in some parts of Europe.

Continental editorial theory

63

Early traditions of textual scholarship When we discuss the study of national literary traditions, we should keep in mind that in most European countries the general field of textual study has its own particular history that often but not always follows the same general course. As a separate discipline, textual study dates from Hellenistic Greece: the Alexandrian critic Eratosthenes was the first person to call himself philologos. Anthony Grafton, the most recent historian of the earliest tradition of textual scholarship, has shown that textual study began with the philologists of Alexandria while he and earlier historians of the discipline such as J. E. Sandys and Rudolf Pfeiffer, have traced the path of these first beginnings to the Romans and then to Western Europe. After a period of relative neglect in the Middle Ages, textual criticism was rediscovered in the fifteenth century by humanist scholars such as Lorenzo Valla of Rome and Desiderius Erasmus of Rotterdam in the careful study of Latin and Greek classical texts. It was then further developed, in the contentious climate of the reformation, in the study and edition of patristic literature and even of the biblical text, on which both Valla and Erasmus had worked. Central in the development of this first phase of modern textual scholarship was the invention of the printing press and the resulting publication of an editio princeps of each of the major Greek and Latin works, a first printed version of a text that had hitherto existed only in manuscript but that in its printed form could serve as reference text. In itself it was useful to have a single text against which variants could be attested, because most of these early printed texts had been edited on the basis of a single or at most two manuscripts, with various emendations unsystematically applied. E. J. Kenney has explained that before the nineteenth century “editors of texts were largely engaged in a piecemeal and haphazard attempt to undo the damage that had been inflicted in the period between the ninth century and the Renaissance.”1 Also important was the fact that the findings in this field were reported in Latin, the language of science and lingua franca of what came to be called the republic of letters, an international network of scholars and scientists that included not a few textual editors. Although political conflicts and religious wars in the centuries following the reformation created sharp divisions between nations and religions, the members of this republic still managed to communicate their ideas. Protestant Dutch and English scholars traveled to France and Italy to hunt for manuscripts and they corresponded widely with like-minded but Catholic colleagues. Occasionally Protestants and Catholics would even contemplate collaborating on projects that were as contentious as a new translation of the Bible.2

64

geert lernout

By the seventeenth century the critical and historical study of texts, which for a pioneer like the Dutchman Baruch Spinoza included the text of the Bible, was a central part of a literary and scientific culture that did not yet know the difference between what C. P. Snow famously called the two cultures. Isaac Newton spent considerable time on an attempt to coordinate biblical and other chronologies and in that process the role of the accuracy of the text was an important issue. In 1690 Newton wrote an essay on “notable corruptions of scripture” to demonstrate that, as Erasmus had shown earlier, the comma johanneum in the first Epistle of John was a late intrusion into the Latin text and could not constitute biblical evidence for the Trinity. Versatile scholars like Richard Bentley (who corresponded with Newton) applied their erudition and critical skills to a wide variety of texts. On the basis of a close study of Homeric meter, Bentley even correctly postulated the existence of the digamma, a Greek letter that was still present in some words in Homer, but not represented.3 Philology and the natural sciences became part of the core tasks of the new type of research university that had been developed and pioneered in Berlin by Wilhelm von Humboldt in the early part of the nineteenth century. Most of the early philologists had been amateurs, either independently wealthy or else working under royal or aristocratic patronage. To some extent the figure of the literary and historical scholar survived into the nineteenth century (especially in its comic guise as the pedant), but in Germany textual scholarship was one of the cornerstones of the new university that had the close study of classical texts in their proper historical context as an integral part of its curriculum and that would become the model of the new kind and non-confessional type of higher education in the rest of the world. In a parallel development it was in the late eighteenth and early nineteenth century that, partly as a result of the French Revolution, most of the manuscripts that had earlier moved between great private collections, at last found a public resting place in university and national libraries, especially in those nations that had been occupied by the revolutionary French armies who closed the monasteries and appropriated the religious, royal, and aristocratic libraries. It was in these new public institutions that the manuscripts became available for the kind of serious study that could only now begin. Sebastiano Timpanaro has described how this new learning developed in Germany, on the basis of insights derived from the study of the complex transmission of the biblical text,4 while E. J. Kenney comments that Bentley’s aborted plan for a critical edition of the New Testament makes him the chief founder of the science of historical criticism.5 Although both Timpanaro and Kenney claim that Karl Lachmann had been preceded in many respects by other critics, the German scholar in

Continental editorial theory

65

his edition of the Roman poet Lucretius is considered to have been the first to formulate fully the main principles of modern textual criticism. This set of principles would grow, especially in the work of his students and followers, into the so-called Lachmannian method, a careful examination of the manuscript record to define the relationship between the different documents. Lachmann saw that a systematic study of all the available manuscripts of a text would reveal complex relationships that could be represented in a stemma, a tree-structure. This preliminary work, the recensio, leads us to conclude which of the existing manuscripts is the closest to the original text, selectio; the text of that manuscript is then carefully studied, the examinatio, in order to correct the errors with the help of a fixed set of procedures. In the exemplary case of the edition of Lucretius’ De rerum natura, Lachmann was able to demonstrate that the three different manuscript traditions all derive from a single archetype with 302 pages of 26 lines each. He also showed that this conjectured manuscript of Lucretius’ work was itself copied from a manuscript written in minuscule letters, in its turn copied from one in rustic capitals. Although each of the methodological steps of the method had been anticipated and used by earlier critics, Lachmann seems to have been the first to employ all of them systematically and in any case his “method” became identified with his name in the decades after his death. Both Kenney and Timpanaro stress that the method has remained influential in classical studies although not uncontroversial. In the same period another and more important shift took place: under the influence of romantic ideas about the national past, especially in a politically divided Germany that badly needed a unified national past, medieval texts in the vernacular were rediscovered, collected, edited, and presented as proof of the fundamental historical continuity of the Volk, the ethnic group, the nation, which found one of its identificatory moments in the study of languages and cultures. This scholarship led in its turn to the insights of comparative linguistics about the historical relations among the different language groups. In the second half of the nineteenth century the German philological movement spread rapidly abroad, again in league with the new idea of a research university, first in the rest of Europe and then in the United States. In many European countries, most of them transforming themselves into the new nation states, some of the initial research had been done by enterprising German scholars, like August Heinrich Hoffmann von Fallersleben in Holland and Flanders, and others in France and in Italy.6 The new philology was a discipline that consisted of the study of the history and structure of classical and of modern languages, as well as the literature written in these languages.

66

geert lernout

The historical-critical edition was an integral part of this new philology, but it was first and foremost a practice, not a coherent set of theories. The aim of textual scholarship was to offer the best editions of the texts of the past and most of the scholars involved thought that the aim of the exercise was to publish a text that best represented the intentions of the original author. This seems to have been so self-evident that the aim was only rarely made explicit and then often only in terms of finding in the manuscript record the original version: philology was supposed to go ad fontes, back to the original source, in order to recapture the pure waters that had been muddied by the passage of time. Until recently the Lachmann method has been in the background of all editorial discussions in Europe, obviously in the case of classical, biblical, and medieval texts from the manuscript era (Lachmann had been active in all three fields), but even in the publication of literary texts of the print era. Lachmann himself had developed his principles on the basis of the traditions of textual criticism of classical and biblical texts but he applied his method not just to literary works in Latin, but also to the texts of the medieval writers in the vernacular such as Walther von der Vogelweide and Wolfram von Eschenbach and even of the eighteenth-century writer and philosopher Gotthold Ephraim Lessing. Whereas Bentley had only made elaborate plans for an edition of the New Testament, Lachmann was responsible for the first critically edited version of the New Testament independent of the textus receptus, the version that had been the base text for most Protestant Bible editions and translations. The historical-critical edition Following Lachmann, the early history of textual editing in the modern period has been predominantly German, with scholars from that country responsible for most of the innovative work on the Bible and on Greek and Latin classical texts, but the move to texts in German was also crucial in establishing the legitimacy of the study of vernacular literature as a new academic discipline. The first university chairs in that subject were only established in the first decades of the nineteenth century and the critical editions of mostly medieval texts played a crucial role in demonstrating the scientific legitimacy of the new field of study. Rudiger Nutt-Kofoth ¨ writes that “the name of Lachmann is closely bound up with the emergence of German scholarly editing and the recognition of German studies as an academic discipline.”7 As a result, the philological sub-discipline is still called Editionswissenschaft, the science of editing. But the legitimacy of the new discipline was based so self-evidently on the methodology of classical philology that Lachmann and his early followers do not even seem to have

Continental editorial theory

67

distinguished in principle between editing works from the manuscript era and works dating from after the invention of printing, treating for example Lessing’s works no differently than a classical or biblical text, much in the same way that Richard Bentley had edited (and heavy-handedly emended) Milton’s Paradise Lost. The end result of this process would become one of the most important German cultural export products, the historical-critical edition of texts in the vernacular. The first project that identified itself as historical-critical was an edition of Schiller’s works by Karl Goedeke, published between 1867 and 1876. It was followed by the massive 133-volume Weimar Goethe Edition, between 1887 and 1919. These two ambitious projects have been seen to represent the two main ways of editing the two classic authors of the modern period: the Weimar edition was ostensibly based on Goethe’s own edition or “Ausgabe letzter Hand” in which at the end of his life the old master had organized sometimes heavily revised versions of all his work according to genres. In contrast, Goedeke’s Schiller edition presented the author’s works in their historical sequence, with the variants printed at the bottom of the page, as used to be the case in editions of Greek and Latin texts. All of this information was designed to give the reader at least some sense of how the author had actually composed his works. Although the editors of the Weimar Goethe dismissed Goedeke’s procedures as a luxury they could not afford in the case of the much more prolific writer Goethe,8 they did give the variants in a separate section of each volume, which made it in fact possible to have a better sense of what these various documents contained. Klaus Hurlebusch has identified the tension between concentrating on the production of the literary works or on their reception as one of the most important concerns in the history of German textual editing.9 After the oeuvre of Goethe and Schiller, the work of Friedrich Holderlin ¨ has been central in the history of German editing, following a major rediscovery of that poet in the first years of the twentieth century, based on a set of hymnic poems that had only survived in the form of unpublished and sometimes fragmentary working drafts. In his so-called “Great Stuttgart Edition,” the first volume of which was published in 1943, Friedrich Beißner edited these later works and he gave them a final form, on the basis of transcriptions of earlier versions that were also included in the edition. In his work on Holderlin, Beißner for the first time made the extremely ¨ important distinction in the study of modern manuscripts between genetic and transmissional variants, differences between versions that were the result of a conscious choice on the part of the author and those that were merely copying errors or other mistakes. Although he later admitted that these early versions of Holderlin’s poems (mostly two early versions out of which ¨ the editor then created a single final form) did not represent these works’

68

geert lernout

real genesis but only what he called their “ideal growth,” this radical step towards an interest in the development of the literary work was extremely influential among the immediate post-war generation of textual editors in Germany. It was only a decade or so after the publication of the first volume of the Stuttgart edition that in the theoretical discussions about the new developments, the emphasis shifted towards the genetic study of the work and towards final authorial intentions. Ever since the Weimar Goethe Edition, German editors had seen themselves as executors of the last wishes of the author, agreeing with Goethe himself on privileging the aesthetic superiority of the final versions created by the author. But now this bias came under attack, especially when Ernst Grumach in his “prolegomena” to a new Goethe edition was able to prove that the final revision of many of the poet’s late texts had in fact not been made by the poet himself, but by an assistant.10 The editors of the new Akademie-Ausgabe of Goethe formulated a set of principles that would be extremely influential in Germany and abroad. They identified a text’s version as one stage in its development, not necessarily a step towards a “better” final text, but just substantially different from earlier or later versions and, on principle, of equal value. Siegfried Scheibe, one of the Goethe editors, even claimed that the real text was not the final result, but the sum of all the different versions.11 According to Scheibe it is the editor’s most important task to choose one of these versions of the text and to explain the reasoning behind that choice; but once the choice has been made, editors cannot go back to other versions in order to emend. In strong contrast to Anglo-American copytext editing, German editorial practice avoids the contamination of one version of the text by (elements of) other versions. Concretely the principle of the equivalence of all versions is problematic because it makes all editions practically impossible, since in most cases the different versions (which may differ only in a word or even in a single letter) cannot all be represented in full and even when they are, the crucial critical work is then left to the reader of the historical-critical edition. Yet despite the principles, editions continued to produce privileged versions of an author’s work and since the deathbed decisions of the author had been discredited, there was a greater interest in earlier authorial versions. Scheibe’s principle of equivalence among versions made it impossible to really ground such a choice for one version, which was necessary if only to distinguish between text and apparatus. The representation of text and apparatus had also been the basis of the critic Beda Allemann’s accusation against Beißner that the Große Stuttgarter Ausgabe only gave the results of the editorial work in providing the reader with earlier versions of the poems,

Continental editorial theory

69

but did not give the edition’s user the means to check the editorial decisions against the manuscript record.12 It was the editor of a new Frankfurt Holderlin edition who gave Alle¨ mann what he wanted. In this new complete edition, D. E. Sattler includes reproductions of all the manuscripts, sometimes even in the form of facsimiles that mimic the form in which they were transmitted (as in the case of the so-called Homburg Folioheft, a gathering of folios that contains all the late unfinished poems), then typographically faithful transcriptions of these manuscripts and only then his suggestions for the final form these poems were supposed to have: the editor’s active role in this process is stressed by calling the latter forms “constituted texts”). Typically, in the more limited paperback version of the edition, only the reproductions are missing: both the transcriptions and the constituted texts are included. Since 1975, when Sattler’s first volume was published, his example has been followed by the editors of the work of Georg Buchner and most importantly of Franz ¨ Kafka, whose works, with the exception of just a few short stories, had not been authorized for publication: the only texts we have are fragmentary and unfinished and this is the form they are given in this edition, not the versions that were later finished by Max Brod. Philology had been a German science from the beginning and it was so well-organized and institutionally embedded that outside influences on German editing theory and practice were minimal: most of the challenges came from within. It is only in the last few decades that some of the younger generation of critics have begun to look across the borders, first at French developments and later also at Anglo-American editorial theory. It is probably not a coincidence that quite a few of these critics do not work on modern German literature but in medieval studies or in English literature. The most prominent is Hans Walter Gabler, whose ground-breaking 1984 edition of James Joyce’s Ulysses was the result of a sophisticated but not uncontroversial mixture of French, German, and Anglo-American developments in editorial theory and practice. It is in this final stage of development of thinking about how to represent literary texts that German editors have been most open to French literary theory and genetic criticism; French and German editorial theorists have attempted to stay in touch with the evolutions across the border. Bilingual conferences and especially the efforts of a few bridge builders have managed to keep some kind of contact between the two major Continental traditions. In the nineteenth century Victor Cousin introduced German-inspired educational reforms in France, and it is probably not a coincidence that he was also one of the first intellectuals to stress the importance of modern manuscripts. By the end of the nineteenth century the French universities,

70

geert lernout

like universities all over the world, began to adopt the German model of scientific research (before 1914 more than half of the library budget of the Ecole Normale Sup´erieure was spent on German publications). The new philology was an important part of this legacy. The first chair in medieval French literature at the prestigious Coll`ege de France was established in 1852, but the study of medieval vernacular literature remained a contentious subject for much longer than in Germany. ˆ Especially after the d´ebacle of the Franco-Prussian war in 1870 (which the French writer and philosopher Ernest Renan called a victory for German science), the long reform of the old Napoleonic university system was politically contested. Especially controversial in that reform was the role of German philology, most notably in the decade before 1914, during the controversy following the university reforms of 1902 in which the role of Latin had been diminished in favor of the sciences and of modern languages (which included French). This crisis represented a belated version of the late seventeenth-century Quarrel of the Ancients and the Moderns in which the loss of the central role of Latin was seen as a betrayal of the essential values of the nation. These discussions should also be seen in the context of the political struggle between conservatives and democrats that marked the period between 1870 and 1914. Quite a few traditionalist literature specialists felt that they had to defend Latin-French values against an aggressive German positivism that in France was first advocated by Gaston Paris and later by Gustave Lanson. It is interesting that the reformists at the beginning of the twentieth century were accused of the same crimes as the French genetic critics 100 years later. Most influential against the Lachmann method in medieval studies was Joseph B´edier who had noticed that most concrete stemmas for medieval texts seemed to consist of only two traditions, a situation that enabled the editors to construct a text of their own devising. He ended up rejecting what he saw as an ultimately subjective methodology and began to argue for the faithful edition of the “best text” of a work, which had at least the advantage of being historically consistent, having been written by one scribe, at a particular moment in time.13 At the other end of the spectrum, Lanson’s ideas, formalized by his student Gustave Rudler, concentrated on literary history and did not exclude an interest in the genesis of literary works, but the efforts did not lead to French historical-critical editions. Even the prestigious Pl´eiade editions of classic works, which began to appear in 1931, were often incomplete, sometimes lacking any kind of apparatus: the genetic critic Almuth Gr´esillon correctly calls these editions at most “half-critical.”14 Although a few scholars wrote on the manner in which a number of modern literary works came into being, both the interest in editions and in the textual development were silenced

Continental editorial theory

71

by what Gr´esillon calls the “structuralist wave” of the 1960s, which came to prominence with a debate between the Sorbonne professor Raymond Picard and Roland Barthes in which the former represented the philological tradition dominant at the universities and the latter defended the right of the new critics to introduce structuralist methodologies. While this French form of New Criticism, like its American cousin, concentrated its work on the published form of the literary text, in the early 1970s a number of young critics began to develop an interest in the careful study of modern manuscripts. The history of this genetic criticism is welldocumented: in 1966 the Biblioth`eque Nationale acquired a collection of Heinrich Heine manuscripts and Louis Hay became the head of a team of scholars that would study this rich hoard. From the beginning both Hay and his colleague Jean Bellemin-No¨el stressed their general agreement with structuralism; they used structuralist terminology and downplayed references to the author’s intentions. Important for the survival of genetic criticism as a discipline was the institutional support the early group of researchers received, especially when the original Heine group joined forces with critics working on the genetic study of complex and interesting works by Gustave Flaubert, James Joyce, Marcel Proust, and Emile Zola. The initial Centre d’Analyse des Manuscrits became ITEM (Institut des Textes et Manuscrits Modernes), a “laboratory” of the French National Center for Scientific Research (CNRS). Although some of the first French geneticists were specialists of German literature (Louis Hay, Gr´esillon) or of James Joyce (Daniel Ferrer), genetic criticism grew out of French structuralism and post-structuralism. Most French historians of the field downplay the importance of French manuscript work before the seventies, for reasons that have more to do with contemporary academic politics in France than with historical realities. Genetic criticism, a discipline that depends on larger research groups always in need of adequate funding, has been under constant attack by representatives of other scholarly approaches, who are competing for the same, ever diminishing funds. As a result the first generation of g´en´eticiens tended to stress the continuity in their thinking with then dominant (post-)structuralist ideas and their fundamental differences with the old-fashioned criticism, an emphasis that has recently been criticized.15 Most of the French genetic critics are active both in practical genetic work and in textual theory: the members of ITEM have found a balance between work on the chosen set of authors and advanced work on the material study of manuscripts, closely related to what is called “analytical bibliography” in the Anglo-American tradition (where the emphasis tends to be on books instead of manuscripts). At the same time there has been a tendency to reflect on the theory of editing as well as a lively confrontation of ideas with other

72

geert lernout

members of the team, and increasingly over the years, with people working in related fields (such as music, architecture, art) and with genetic critics in other countries, in Europe and America. ITEM’s emphasis on theoretical reflection and the exchange of ideas among its members has not resulted in a single theory of literary manuscripts or of the genesis of literary texts, although the more central figures of the group seem to agree on a number of defining characteristics, one of which is certainly the initial dependence on (post-) structuralist thought and its insistence on the fundamental openness to interpretation of the literary work. This emphasis was based on a psychological model where the writing subject is believed not to be in full control of the writing process but instead to be controlled by e´ criture: poems, just like other works of literature, write themselves. The particular history of a given work, to the extent that it survives in the form of manuscripts, was called the avant-texte by Jean Bellemin-No¨el, although Gr´esillon noted that the term itself implies an identifiable text that seems to have a definite enough form to become the telos of the process that precedes it. But genetic critics reject this kind of teleology and Gr´esillon favors the more neutral term “genetic dossier” which she defines as “the sum total of the written documents that one can afterwards consider to have been part of a writing project of which it does not matter whether it resulted or did not result in a published text.”16 The anti-teleological bias is crucial in the self-understanding of genetic critics who routinely stress that they are not interested in editing, in the constitution or emendation of final texts. In fact, for most genetic critics the final or published version of a text is neither more nor less interesting than any other stage of development: all versions of a text have equal value and what interests the genetic critic is the process of writing, not its result. Genetic criticism thus feels itself closest to those of the recent German editors who stress the temporal dynamics of writing and the fundamental equivalence of all versions of a work, but it differs from the German tradition by its lack of interest in editorial matters and in edited texts. In their introduction to a volume of translations of genetic essays, Daniel Ferrer and Michael Groden even come to the conclusion that “genetic criticism is not concerned with texts at all but only with the writing processes that engender them.”17 But this does not mean that all genetic critics disdain edition. Some g´en´eticiens have identified “genetic edition” as one of their tasks: reproducing the complete genetic dossier, which of course first needs to be studied, transcribed, made sense of and ordered. Again, this does not mean that the end of the process then acquires a more special status than any of the intermediate stages. Ordering the documents in their temporal succession is only the first part of the work. The dossier should be read, both in reverse, to

Continental editorial theory

73

see where all the later versions come from, but also in the order in which the writer produced them, in order to observe how each choice necessarily implies the neglect of alternatives, the roads not taken. The result of some of the uncompromising genetic thinking about editions has led to criticism that such monster collections of all of a writer’s manuscripts simply do not have readers in the real world: in most modern editions, Flaubert’s story “Un cœur simple” takes up less than thirty pages, but the story’s genetic dossier is 700 pages long. Almost from the beginning the genetic critics were aware of this problem (if only because it was constantly repeated by critics of their work) and some of them became pioneers in the exploration of the possibilities of hypertext and the use of digital resources. This has resulted in a rich literature that merits more presence in the debates about the use of computers in editing in other traditions. Although some of the early work in this genre suffered from an exaggerated optimism, the metaphor of hypertext has helped the g´en´eticiens to formulate more efficiently what genetic criticism can mean to the study of texts and how it can create a new kind of reader. This type of work has also had practical results, for example in the digitization of many books by the project Gallica of the Biblioth`eque Nationale, an initiative that can now be recognized as having been a precursor to Google Books. From the very beginning the French national library worked closely with ITEM, more specifically with the Zola research group which in 2003 brought out an exemplary genetic dossier of the novel Le Rˆeve.18 At the same time it may be relevant that the announcement of the birth of a new genetic reader as prophesied by some geneticists,19 seems to have been premature: of all the thematic dossiers available on Gallica, the Rˆeve dossier is the one that has been least popular with browsers.20 Developments in the twentieth century and beyond In Italy, German editorial thinking was introduced at the end of the nineteenth century, although the new method was probably mediated for the Italian philologists through the work of Gaston Paris on medieval manuscripts. By the turn of the century, there were already editions of Italian texts, among them those of Dante, based on the new principles. The relative success of the new approach was hindered by the enormous influence of Benedetto Croce’s idealist aesthetics in the study of Italian literature in the first half of the twentieth century. Croce rejected textual studies as a positivist aberration and with other conservative critics he protested against what was considered as a rationalist attack on the sovereign nature of literary genius. As late as 1947 he explicitly rejected the idea that one could learn anything by looking into a writer’s wastepaper basket.21

74

geert lernout

Between the world wars, editorial theory, especially in the field of classical Greek and Latin texts, was still largely the domain of German critics. Paul Maas’s concise description of the basic principles of the edition of classical texts in his book Textkritik was the occasion for a reaction by Giorgio Pasquali, first in a review and then in an ambitious and still immensely readable book. Pasquali did criticize the rigidity of the method, but his critique remained well within the parameters set by Lachmann. In the wake of Pasquali’s study and in a book that has been translated into German and English, Sebastiano Timpanaro was one of the first scholars to identify the exact role of Lachmann in the development of the methodology that carries his name. In the late 1930s Pasquali’s conclusions became the basis upon which Italian scholars such as Michele Barbi, Gianfranco Contini and others built what Barbi called the nuova filologia or variantistica: a careful study of the specific textual tradition of the text which has been described as “neolachmannian” and even “translachmannian.” If in England the work of Shakespeare has been at the center of editorial discussion and in Germany the writings of Goethe, Schiller, and Holderlin, a ¨ similar role is played in Italy by the rich and complex history of the works of Dante. Barbi edited the Vita nuova, but the 600 manuscripts of the Divina commedia continue to baffle the scholars: there is still no genuine critical edition of the text. Here, as in the case where we can be sure that certain writers produced different authorial versions of the same text, Italian critics have given priority to the historical reality of each of the variants over a rigorous methodology. Every national culture in Europe, including regional cultures such as the Catalan or the Flemish, have sought legitimacy as nations in the publication of texts of their literary past. In most of these cases, the editorial principles for such initiatives were borrowed ultimately from a Lachmannian tradition, sometimes via quite circuitous routes. In the case of Spain, for example, editors hesitated a long time before they adopted the principles that had first been applied by Italian editors, whose nuova filologia, itself inspired by Gaston Paris’s edition of medieval manuscripts, was developed in the wake and the spirit of Lachmann. For their part, Dutch editors have worked mainly in the German tradition, whereas a small group of Belgian editors and geneticists (first in the research group Genese and later in the Center for Manuscript Genetics in Antwerp) have attempted to compare and reconcile the different scholarly approaches (Dirk Van Hulle) or have done pioneering work in electronic editions (Marcel De Smedt, Edward Vanhoutte, and Van Hulle again). In Denmark too German editorial theory was more influential, whereas Swedish editors, at least since the early 1990s, show a tendency to follow the developments in Anglo-American theory.

Continental editorial theory

75

For political reasons the study and edition of literary texts in Russian could only develop in the brief period between the revolution and before the reestablishment of censorship in the Soviet Union. A new discipline was created as part of Russian philology and this Russian “textology” had its own theory of artistic creation. It is interesting to see that some of these textologists (whose genetic work has just been translated into French) turn out to be the same formalists that (post-)structuralists had considered as their precursors: Boris Eikhenbaum and Boris Tomashevskii.22 In most European countries, editorial practice and theory began to receive institutional support in the 1970s and 1980s, with the establishment of the French ITEM and the German Arbeitsgemeinschaft fur ¨ germanistische Edition. Most of the Scandinavian editors have been active in a multilingual ¨ ¨ Editionsfilologer, founded in 1995. It is only in the Nordiskt Natverk for last thirty years that some of the members of these larger groups have begun to look across national and institutional borders, and initially it was the German and the French scholars who most closely studied each other’s theory and practice, beginning with a colloquium they organized in Paris in 1977. Since then there have been frequent contacts between the different groups, usually in the form of participation at conferences or surveys of the developments in the other traditions. In the last decade it has become difficult in this field to find collections of essays in German or French or issues of specialist journals without at least one article referring to the “other” tradition.23 The Anglo-American tradition was much slower to make an impact in France and Germany, and it was Hans Walter Gabler’s edition of Ulysses that for the first time demonstrated that the three traditions could be made relevant, a three-way confrontation that has also been studied by Dirk Van Hulle in his book on editions of Joyce, Proust, and Thomas Mann. Since the 1990s, mostly through the efforts of a few critics like Peter Shillingsburg, a dialogue has been established between Continental editors and their American and British colleagues that led directly to the establishment of the European Society for Textual Scholarship, which in its turn fostered much closer contacts among some of the Continental textual scholars. To some extent, English has replaced Latin as a lingua franca among the younger generation of editors and the much closer contact has resulted in international conferences, joint publications and research projects. As the controversy about Gabler’s Ulysses has demonstrated, it might be hazardous to combine principles from different editorial traditions,24 but at the same time such rare cases also show that it is not always useful to argue about textual problems and editorial decisions on the basis of purely theoretical principles. Although the reference to post-structuralist theory

76

geert lernout

may have had some salutary effect in editorial theory, some of the work in this area does not go beyond a sophisticated discussion of the metaphors used by editors when they describe their work. One of the most exciting recent developments in Continental editorial theory has been the increased interest in the related fields of classical and biblical textual scholarship and in disciplines such as paleography and book history, all of this despite the obvious difficulties created by the increased specialization in each of these fields. Such interest may be due to a growing awareness of the origins of textual scholarship, an interest that should not be surprising in such a preeminently historical discipline. Few disciplines have been so ruthless in their interest in the history of their own tradition and in their critique of the political relevance of some of its work (especially in Italy where Giovanni Fiesoli has written a detailed study on the genesis and the reception of Lachmann’s work). There have been extremely critical studies of the ideas of some of the early philologists by Luciano Canfora, Edward Said, Hans Ulrich Gumbrecht, and Bernard Cerquiglini. In 1989 ´ Cerquiglini’s Eloge de la variante was one of the inaugural statements of yet another “new philology,” this time a call for a return in the study of medieval vernacular literature to the uniqueness of the single manuscript, against a philology that was accused of homogenizing, with its belief in authorial intentions (the similar movement of “radical philology” in classical studies does not seem to have had many European followers). Few critics have noted how the rhetoric and the thrust of this pamphlet signal a return to B´edier’s editorial practice and to other rejections of rational enquiry by post-structuralism that I have elsewhere described as romantic.25 In many European cultures, textual studies have become much more central in recent years than they were during the height of post-structuralism. There has been a return to the archives and a new interest in the materiality and the genesis of texts, old and new. It is interesting to see that at least some of this new work returns to the interdisciplinary roots of philology and of historical inquiry in general. We also note that it is in the course of some of this work that young scholars of different backgrounds (national, disciplinary, and theoretical) have discovered a common basis of scholarly work on texts. Not surprisingly, some of this new work is in digital editing, because it is there that technical developments have made not just a quantitative but also a qualitative difference with what was possible in the print era. But computer resources too have had an important impact on the development of digital techniques of manuscript comparison, with a strong input from information theory and, appropriately perhaps, from evolutionary genetics. For this new group of textual scholars, the object of study is no longer just the major works of a select group of modernist writers; it now includes classical and medieval works, as well as biblical texts.

Continental editorial theory

77

Despite the many names it has acquired over the years, textual study has not only survived the twentieth century in Continental Europe, but both its theory and its practice seem to be thriving. In its intense study of its own past, its interdisciplinary and increasingly international interest in developments in related philological, literary, biblical, and historical scholarship, it is no longer just a preliminary and according to some critics even pre-scientific auxiliary discipline; it has established itself as one of the more interesting and coherent ways of studying our written heritage. NOTES 1 E. J. Kenney, The Classical Text (University of California Press, 1974), 4. 2 Paul Auvray, Richard Simon: 1638–1712. Etude bio-bibliographique avec des texts in´edits (Presses Universitaires de France, 1974), 36–8. 3 James Henry Monk, The Life of Richard Bentley, D. D. (London: Rivington, 1834), 361–7. 4 Sebastiano Timpanaro, La genesi del metodo del Lachmann (Florence: Le Monnier, 1963; rev. edn. Liviana, 1985), 100–1. 5 Kenney, Classical Text, 100–1. 6 Joep Leerssen. “Literary Historicism: Romanticism, Philologists, and the Presence of the Past.” Modern Language Quarterly, 65.2 (2004), 221–44. 7 Rudiger Nutt-Kofoth, “Two Paradigms in Nineteenth Century German Editing: ¨ Goedeke’s Schiller Edition and the Weimar Goethe Edition as Different Steps towards a Particular Concept of Editing Modern Authors,” Variants, 5 (2006), 319. 8 Ibid., 326. 9 Klaus Hurlebusch. “Conceptualizations for Procedures of Authorship.” Studies in Bibliography, 41 (1988), 100–35. 10 Ernst Grumach, “Prolegomena zu einer Goethe-Ausgabe,” Goethe-Jahrbuch, 12 (1950), 60–88. 11 Siegfried Scheibe, “Zu einigen Grundprinzipien einer historisch-kritischen Ausgabe,” in Gunter Martens and Hans Zeller, eds., Texte und Varianten: Probleme ihrer Edition und Interpretation (Munich: Beck, 1971), 1–44. 12 Beda Allemann, “Review of Holderlin-Ausgabe, Vols. i and ii, ed. Friedrich ¨ das deutsche Altertum, 69 (1956), 75–91. Beißner,” Anzeiger fur 13 Joseph B´edier, “La tradition manuscrite du Lai de l’ombre: R´eflexions sur l’art d’´editer les anciens textes,” Romania, 54 (1928), 161–96, 321–56. 14 Almuth Gr´esillon, El´ements de critique g´en´etique: Lire les manuscrits modernes (Paris: PUF, 1994), 184. 15 See Geert Lernout, “‘Critique g´en´etique’ und Philologie” in Rudiger Nutt-Kofoth, ¨ Bodo Plachta, H.T.M. Van Vliet, and Hermann Zwerschina, eds., Text und Edition: Positionen und Perspektiven (Berlin: Erich Schmidt, 2000), 121–42. 16 Gr´esillon, El´ements, 109. 17 Daniel Ferrer and Michael Groden, “Introduction: A Genesis of French Genetic Criticism,” in Jed Deppman, Daniel Ferrer, and Michael Groden, eds., Genetic Criticism: Texts and Avant-textes (Philadelphia: University of Pennsylvania Press, 2004), 11. 18 See http://gallica.bnf.fr/Zola/

78

geert lernout

19 Jean-Michel Rabat´e as quoted in Ferrer and Groden, “Introduction,” 12. 20 Catherine Lupovici, “La politique de num´erisation des manuscripts a` la Biblioth`eque de France,” in Aur`ele Crasson, ed., L’edition du manuscrit: de l’archive de creation au scriptorium e´ lectronique (Louvaine-la-Neuve: Bruylant-Academia, 2008), 111. 21 Croce quoted in Erica Durante, “Sous la rature, la litt´erature. L’Exp´erience de la philologie italienne au service de la litt´erature compar´ee,” TRANS-, Revue de Litt´erature Compar´ee de l’Universit´e Paris III, 2 (June 2006), www.trans. univ-paris.fr 22 See the French translation of a number of these essays in Andr´e Mikhailov, Andr´e ´ Grichounine and Daniel Ferrer, eds., La textologie russe (Paris: CNRS Editions, 2007). 23 See e.g. Michel Espagne, “Quelque tendances de la philologie allemande,” Genesis, 3 (1993), 31–44 and Peter Shillingsburg, “Anglo-Amerikanischer Editions¨ wissenschaft: Ein knapper Uberblick,” in Rudiger Nutt-Kofoth, Bodo Plachta, ¨ H.T.M. Van Vliet and Hermann Zwerschina, eds., Text und Edition: Positionen und Perspektiven (Berlin: Erich Schmidt, 2000), 143–64. 24 See Geert Lernout, “La critique textuelle anglo-am´ericaine et le cas de Ulysses e´ dit´e par Hans Walter Gabler,” Genesis, 9 (1996), 45–65. 25 Geert Lernout, The French Joyce (University of Michigan Press, 1990), 237–44.

4 HANS WALTER GABLER

Late twentieth-century Shakespeares

Textual scholarship and criticism: the case of King Lear Shakespearian textual scholarship was driven, for the larger part of the twentieth century, by twin forces of select methodology. One was the transfer of the ways of text-critical treatment of the medieval dispersion of texts in manuscripts to the early post-Gutenberg transmissions of texts through books. The other, closely related, was its submission to analytical and textual bibliography. In this doubly determined field of force, Shakespearian textual scholarship rose to ever greater heights of sophistication, and indeed achievement. Yet, in the final quarter of the century, its dogmatisms (unhinged ultimately, too, by wider understandings of textual criticism such as those offered in Jerome J. McGann’s A Critique of Modern Textual Criticism (1983) and debated in the pages of the journal TEXT) were shaken by pragmatic challenges from literary as well as theater-oriented Shakespeare criticism. The moment for retrenchment arrived with the re-opening of a long-standing debate over the nature of the differences between the Quarto (1608) and First Folio (1623) texts of King Lear. The proposition raised was simple and seemed, from the point of view of an analytically refined bibliography, strangely unsophisticated. Why, it was asked, could it not be that these texts, respectively, represented two distinct versions of the play from Shakespeare’s own hand?1 Why not, indeed. The two-version hypothesis had been a minority opinion at least since Dr. Johnson mooted it in the eighteenth century. In 1931, it was resurrected by the young American scholar, Madeleine Doran.2 But she retracted, submitting to (since she felt unable to resist) the contrary verdict of W. W. Greg, most eminent among the eminences of the New Bibliographers of the first half of the twentieth century. Greg’s authority was grounded in the application of an all but unrivalled faculty of analytic logic to a rich archival observation and experience. His personal strengths, in turn, were rooted in classical and medievalist methodologies of textual criticism, pivoted on the organicist historicism of stemmatics. He saw the extant earliest printings 79

80

hans walter gabler

of Shakespeare’s texts as derivative of lost manuscripts (which of course they are), and perceived them thus in analogy to the late derivations, as they survive in scribal manuscripts, of long-lost originals of classical or medieval texts. What is problematic in this is less the material analogy than the submission of both sets of materials to analogous methodological treatment. A basic assumption about post-classical and medieval transmissions in the era before printing is that the causes for the textual differences between extant manuscripts lie in scribal error (unintentional slips or intentional changes in copying). With no possibility of authorial correction or revision long after their authors’ death, texts inexorably deteriorate through progressive error. The application of analytical thought to error as the one and only category of change permits grouping errors into patterns, tabulating the groupings in the matrix of a family tree or stemma, and using the stemma, in its turn, as a graph to support the rationale for a procedure of re-constructive editing based on the given transmission. While the original at the root of a family tree (by hypothetical definition) is lost, it is posited to harbor the pristine error-free text. This may not always be recoverable; yet the stemma, as a logical construct, will always lead back as far as the evidence permits to one nodal point of presumptive origin. This, in terms of textual criticism as applied to medieval manuscripts, is the archetype. The stemmatic method’s predication on error may, however, and indeed must be seen as the congenital flaw of the textual criticism that W. W. Greg and his fellow New Bibliographers applied to the textual situation of the plays of William Shakespeare. Admittedly, pressing the Shakespearian transmission into this mould of text-critical thought yields, at first sight, exhilarating results. How clean-cut and lean, after all, are the family trees of texts in books! No complex branchings-off, here, into manifold directions: the twig we hold in hand, the text, say, of a play in the First Folio, may have sprouted from one extant antecedent in a quarto; else, it stemmed from a printer’s copy in manuscript (it must have done so, even though that manuscript is materially lost); the trunk out of which this manuscript grew was perhaps a scribal transcript directly from an autograph (if the printer’s copy wasn’t itself such a transcript); and at the root was the author’s original – if not, indeed, the author’s most original original, his so-called “foul papers,” were even themselves the printer’s copy. By systemic definition, then, within the stemmatic matrix, a (lost) printer’s copy could thus already be the archetype, that is: the document whose text was by logical differentiation of readings recoverable; and in cases where it could be reasoned that the printer’s copy was an original in the author’s hand, archetype and original became conceptually and logically one.

Late twentieth-century Shakespeares

81

Archetype-directed text-critical and editorial thinking thus possessed the distinct virtue of rendering a maximum, with luck even an optimum, of original Shakespeare text logically recoverable from the derivative witnesses-inprint to these texts. True with respect to the one-witness (Folio-only) play texts anyhow, the approach was considered applicable, too, in all cases where two surviving witnesses (Quarto and Folio) related linearly to one another because the Folio text (with only the Folio compositors’ venial slips) demonstrably reprinted the Quarto text. A difficulty arose, however, when, as in the case of King Lear, surviving Quarto and Folio differed in their texts in so complex a manner that a linear derivation of the Folio from the Quarto could not be posited. It was here that habits of stemmatic reconstruction hit back with a vengeance. They failed to be recognized as deeply misleading, since they brought with them, instead, the jouissance of inventive hypothesizing. The challenge recognized was to recover, even against the odds of the bewildering and often enough contradictory variation between Quarto and Folio Lear, Shakespeare’s one and only original text of the play – in other words, by means of text-critical analysis and reasoning to recover the pristine archetype-cum-original. The alternative hypothesis was not admitted that there could have been two originals of competitively authoritative and authorial standing. Translated into categories of text-critical analysis, in other words, it was ruled out a priori that the differences between the two printings of the play could reflect acts of revision: based on the one hand on mind-sets of textual criticism derived from pre-Gutenberg situations of transmission, and on the other hand, on the assertion of the First Folio editors that Shakespeare’s papers had come into their hands with “scarse a blot” in them. Instead, the variation was considered to have been caused exclusively by errors in transmission. W. W. Greg and the New Bibliographers wedded bibliography-centered textual criticism to the logical procedures of retrospection towards archetype and original grafted onto the matrix of the stemma. Within this framework was established a hierarchy of reasoning. Among the methods to identify the variation as corruption, bibliographical demonstration or inference took absolute priority. Only where the application of analytical and textual bibliography remained inconclusive were lines of argument admitted that explained differences between texts on critical grounds and considered variation, for example, as originating in the process of composition, resulting from working over and revising a text, or indicating corrections. The author, and authorial writing, in other words, were taken into account only where bibliography failed, and conceded only as a fall-back cause, so to speak, for textual variation. This, as we should not fail to note, is a stance exclusive to Anglo-American, and Shakespearian, textual criticism. Other methodologies

82

hans walter gabler

assume composition and revision as causes for textual variation by default, and concede corruption only when critical explanation fails. The empiric truth that all transmission corrupts, elevated to an axiom of bibliographyoriented textual criticism, admittedly coexists easily with the assumption of Shakespeare’s (lost, but presumptively recoverable) one and only original text of, say, King Lear behind all (by definition corrupt) variation. By contrast, against the strength of the critical axiom that any two or more texts of a work are different because (regardless of interspersed corruption and error) they are meant to be different, the one-and-only-text situation will be posited only when the critical premise remains stubbornly inapplicable to the case in hand. Hence, from the vantage point of critically oriented textual criticism, King Lear conforms to a norm defined through methodology not because presumptively reducible to one archetypal original, but instead because it survives in two printed shapes distinguishably retraceable to two separate authorial versions. The assumption that Quarto and Folio King Lear represented two differently corrupted derivations from one authorial text, and one only, did not originate with W. W. Greg and the New Bibliographers. They merely reinforced through analytic logic a majority consensus going back to the earliest days of Shakespeare editing in the eighteenth century. Dr. Johnson’s gesture of critically intuiting two versions3 went, even then, against the grain of a text-critical awareness and editorial practice trained on the analysis of transmissions of classical texts, and their editing. What this entailed when, in the face of the transmission of King Lear, it was assumed that the divergent Quarto and Folio derived from one common lost original, was an eclectic conflation of the transmitted texts into one edited text. This editorial stance inherited from the eighteenth-century gentlemen editors did not change when, in the twentieth century, the New Bibliographers, with W. W. Greg foremost among them, established “bibliography and textual criticism” as an autonomous discipline (and went so far even as to seek to re-define it as a science). It was in Shakespearian textual criticism and editing, above all, that the methodologies of this would-be science were developed. They were dominantly formalist, and thus distinctly not hermeneutically critical. Under the aegis of bibliography-centered textual criticism, consequently, the problem of editing King Lear was not critically re-assessed. On the contrary: the model of eclectically conflated editing that it represented was raised from an intuitive to a would-be scientific practice. More, therefore, was at stake in Greg’s silencing Madeleine Doran than a male-chauvinist (or avuncular) saving the neophyte from errors of judgment through the initiate’s higher insight. For at bottom, to admit that King Lear could have come down to us in two versions would have rocked the foundations of a school of textual criticism whose newly developed bibliographical methodology was

Late twentieth-century Shakespeares

83

seen as a triumphant re-enforcement of the stemmatic thinking of old in the discipline. As a result of the twentieth-century division of the kingdoms of textual criticism and literary criticism – the scholarship versus criticism schism – literary criticism progressively lost awareness of its roots in the scrutiny of the authenticity and transmission of the texts it analyzed and interpreted. The new critics, and the multiple theorists who followed in their wake, even the new historicists, ceased to integrate textual study into their thinking or their teaching, and so largely ceded the field to the scholars. Paradoxically, however, this development liberated the literary and theater-oriented Shakespeare critic of the latter half of the twentieth century, too, from constraints ingrained in the traditions of textual criticism and editing. It was thus that Steven Urkowitz, for one, found sufficient scope for the critical enquiries leading to his book Shakespeare’s Revision of King Lear, of 1980. He took a hard look at Kent’s encounter with the Gentleman in the current conflated editions (King Lear, iii.i). By critically looking at the progression of the argument as well as of the dramaturgy, he found the scene in practice unplayable. So, slighting all mediation via editorial rationale, edited text, apparatus, and explanatory notes, he turned as a critic, not as an editor, straight to the Quarto (Scene 11) and Folio (Act iii, Scene i) source texts – only to find that the scene’s difficulties vanished. In each text, he discovered, the relation of the characters, the dialogue line, and the overall dramatic situation were in themselves consistent. They were simply different. Needless to say, this one demonstration would not by itself prove revision, let alone sustain a claim that the revision was Shakespeare’s. Yet it provided a seminal example from which to generate a two-version hypothesis for the King Lear transmission. Urkowitz’s book is consequently concerned throughout with recurrent patterns of variant theatrical realization, and with variant character conception. It does not initially urge its title. Instead, it at first rather defuses the issue of authorial origin of the textual and stage-play variation by not insisting that the evident differences were of Shakespeare’s own making. It is only when brought to full awareness of the quality and internal consistency-in-heterogeneity of the textual and theatrical variants that the reader is induced into also considering that they are most likely, on the whole, to have originated with William Shakespeare. This overall rhetorical gesture highlights the two-fold argument necessary to establish a renewed understanding of the Shakespeare transmission. By one of its strands, it maintains that there are demonstrable cases comprised within that transmission where plays have come down to us not merely in differently corrupt derivations from one (lost) original, but indeed in genuinely different versions. The second strand of the argument is intent on making out the versional differences (those, for instance, represented by the Quarto and

84

hans walter gabler

Folio texts of King Lear, respectively) as most easily – which means: by the most economic hypothesis – accountable for as Shakespeare’s own. This double-pronged attack amounted to a shifting of the ground for Shakespeare criticism and textual criticism alike. The debate it elicited was controversial when at its height in the 1980s. For King Lear, it culminated in the volume of essays The Division of the Kingdoms, edited by Gary Taylor and Michael Warren in 1983. The book starts out from a historical framing of the whole King Lear question, followed by cameo pieces of genetic criticism simply taking the two-version hypothesis for granted. Thence it progresses, via essays reflecting on method so as to safeguard against a question-begging circularity of interpretative argument, to, finally, suggesting a place (a copy of the Quarto of 1608) and a time (between 1608 and 1610) for the Folio revisions. For all the shifting positions taken over the past quarter-century on many specific issues raised in The Division of the Kingdoms, what may safely be stated in looking back is that the book has turned tables in Shakespeare criticism and textual criticism. The onus of proof was soon, and is meanwhile altogether, no longer on the proponents of the two-version hypothesis, but rather on those who reject it outright, or (as one occasionally finds) still remain uneasy about it. By assuming that Shakespeare was a revising author like other authors, and recognizing that the material transmission of King Lear held substantial evidence to prove the contention (notwithstanding the encomium of his first editors that they had “scarse receiued from him a blot in his papers”), the proponents jointly succeeded in bringing Shakespeare textual criticism in line with the discipline of textual criticism as internationally and trans-nationally conceived and practiced. The fresh light cast initially on King Lear has in the meantime been rewardingly focused, too, on the transmission of several other plays, with Hamlet foremost among them. With this, the scene has been set for a re-integration of textual criticism and literary criticism. A recent volume of essays on King Lear shows this process happily in action.4 The volume opens, after a comprehensive introduction by the volume editor, on a subtle double-take on the two-texts versus one-text question. The two essays, R. A. Foakes, “The reshaping of King Lear”, and Richard Knowles, “The evolution of the texts of Lear” constitute a seeming acknowledgment that both hypotheses still represent genuine options for understanding the Lear situation.5 But in truth, their juxtaposition subverts that assumption. Richard Knowles’s contribution, soundly in terms of old-school bibliographic scholarship, insistently quantifies its evidence and hence, by dint of its own method, neither seeks nor reaches critical conclusions, but instead throws out a wide net of ex negativo reasoning to avoid having qualitatively to associate the differences between quartoattested and folio-attested text with William Shakespeare. R. A. Foakes’s

Late twentieth-century Shakespeares

85

essay, though likewise in substance a text-critical investigation, operates, by contrast, from stringently hermeneutical premises. It concludes that holding Shakespeare responsible for the essential differences between the texts as transmitted – that is, to assume that Shakespeare revised King Lear – represents the likeliest, because it constitutes the most economical hypothesis to govern the two-text situation. Here, over and above a factual (quantitative) assembly of the variation, in other words, a qualitative leap is made to its critical assessment. This is an opening toward re-integrating textual and literary scholarship in a mode of critical scholarship capable of leaving behind, in our twenty-first century, modes of positivism grounded in nineteenth-century stances of learning. Critical procedure trained on approaches of intertextual study may fruitfully be adapted to intratextual exploration, often known, too, as genetic criticism. In terms of textual criticism, accepting the two-version hypothesis for King Lear has mainly, perhaps, brought with it a soberer view of conditions and facts determining the composition and transmission of texts. Still more importantly, surely, it has generated insight in genuinely critical terms. From the debate over the two-version hypothesis for King Lear, we now know more than could previously be known about the work’s problems of structure, be it in terms of theme, plotting or timing. The play’s two versions – that is: the work’s distinct texts – offer variant solutions to these problems. The theme of war, for instance, is differently treated in the versions: the plotting of the play as attested in the quarto culminates explicitly in a situation of war with France, whereas the folio-attested version plots the action resulting from Lear’s abdication as a chronic state of civil war.6 Collaterally, this altered thematic perspective has the effect of speeding up the action that at times, through too much incident, threatens to stagnate through the play’s second half in the quarto-attested version. The artistic will to change appears to have been quite ruthless. For, most spectacularly – and to many a critic, disturbingly – one of the work’s outstandingly memorable scenes fell victim to the replotting. This is the mock trial of Lear’s daughters held in the hovel on the heath.7 Its absence from the folio-attested text, admittedly, is critically hard to swallow as an act of revision. But if it is so accepted, as it must be under the two-version hypothesis as a whole, it is most easily acceptable as removed on highest revisional authority, which would mean: by Shakespeare, the revising author, himself. Shakespeare’s creatively renewed composition in revision may indeed be inferred, too, from many a textual detail reinforcing the thematic, dramatic, and poetic overall shape of the play in its First Folio version. Gary Taylor, for instance, in The Division of the Kingdoms, assumes that Shakespeare, in order to revise the play, re-read King Lear in a copy of the Quarto of 1608.8 The assumption implies that he had the entire text as achieved in 1605/1606

86

hans walter gabler

freshly before him when he undertook to alter it. This proposition, in turn, puts the critic in a position to assess, in some significant instances, where revising meant working over the text retrospectively so as to render it more consciously prospective in its ultimate unfolding. For instance, when Lear in the hovel on the heath asks to be assisted in stripping off his clothes, the quarto-attested version reads: “off off you lendings, come on.” Only the folio-attested version has: “Off, off you Lendings: Come, vnbutton here.” Interpretatively, one would suggest that it was out of Lear’s dying words from the very end of the play, “Pray you undo this Button,” that the revision at the play’s mid-point was generated. Similarly, it is the quarto-attested version that, in the first act’s first encounter between Lear, the Fool, and Kent in disguise, already chimes insistently on the word “nothing” (“Kent: This is nothing Foole. . . . Foole: . . . (to Lear) can you make no vse of nothing Nuncle? Lear: Why no Boy, nothing can be made out of nothing.”). In terms of the play’s progressive composition by way of revision, consequently, it is a pre-echo of this passage that gets worked into the folio-attested text at the all-important moment in the opening scene when Cordelia refuses to say / to speak what Lear expects to hear from her: “Lear: Nothing? Cordelia: Nothing.” These two lines, inserted, are to be found in the First Folio text only. The abdication scene reveals altogether, and most richly, a creatively revisional energy lavished on it. The scene is reinforced in expositional precision (all acting characters, for example, are now named in the spoken dialogue, not merely indicated in stage directions and speech prefixes). Thematic oppositions such as age and infirmity versus youth and strength, or nature versus merit, gain explicitness in language and dialogue. Through ampler pacing, the scene is most sensuously set swinging. The dialogue, and above all the King’s speeches, expand in reverberations of metaphor, simile, and imagery. The images of clothing, divestiture, and the unburdened crawling to death, by which we so vividly experience the drama of the abdicating king, appear, on comparative analysis of the play’s two versions, for the first time in its folio-attested text. That emotive cluster of images, all engendered from language and imagery elaborated in the play’s later progression as already shaped in the quarto-attested version, represents quintessentially Shakespeare’s creatively retrospective revision for prospective effects to the poetic and dramatic progression of King Lear.

The resurgence of criticism in editing: the Oxford edition What after more than two decades meanwhile stands out clearly, is that the King Lear debate importantly paved the way, as well, for the arrival –

Late twentieth-century Shakespeares

87

meaning, too: the controversial reception – of the one-volume Oxford edition of the Works of William Shakespeare, edited by Stanley Wells, Gary Taylor, and their associates, and published in 1986.9 It is the latest to date in the line of ground-breaking editions of Shakespeare’s plays, or works, that have had a habit of appearing at roughly century-long intervals: 1623, the First Folio; 1709, Nicholas Rowe’s multi-volume, play-by-play edition (the first in the rapid sequence of editions by named editors in the course of the eighteenth century); 1803, Samuel Johnson’s and Isaac Reed’s Variorum Edition (textually based on George Steevens’s and Edmund Malone’s historically founded reassessments of Shakespeare’s texts); and 1863–6 the (Old) Cambridge Edition, again edited from the ground up, and textually providing the standard for Shakespeare editions through most of the twentieth century. The edition of 1986 – having, with some modifications to its editorial solutions, meanwhile also become a Norton Shakespeare10 – may be expected to be the edition for the twenty-first century. It could not have been edited without the encompassing achievements of Shakespearian textual scholarship as engendered through analytical and textual bibliography throughout the twentieth century. Nor, however, could it have been conceived and realized without the new openings, both editorial and critical, implicit in the debate over King Lear. Among the multiplicity of Shakespeare editions produced and published in the twentieth, and continuing to appear in the twenty-first century, the Oxford edition of 1986 is the one edition of the present age that reassesses the textual situation radically and comprehensively from the surviving sources of the transmission, the quarto and folio original printings. If herein it renews the bids of comparable scope made at the end of the eighteenth and nineteenth centuries, it is singular in its ambition to transcend the strictures of methodology-bound editorial scholarship. In establishing its edition text, rather, it welds the critical to the text-critical perspective. Therein, it extends the critical scholarship exercised in the Lear debate to encompass the Shakespearian oeuvre as a whole. It is thus the Lear debate and the Oxford edition together that, in terms of the main argument of the present essay, exemplify the resurgence of textual criticism as criticism in late twentieth-century Shakespeare studies. The Oxford edition departs radically from the main critical, textual, and editorial traditions going all the way back to the First Folio of 1623. The First Folio canonized the Shakespearian transmission by the force of its disposition that originated with the volume’s editors. These were, on the one hand, members of Shakespeare’s own theater company. On the other hand, the editorial syndicate comprised the several printers sharing the entrepreneurial risk of posthumously monumentalizing Shakespeare. As likely as not, it was mainly the publishers who urged a bookish disposition of the texts

88

hans walter gabler

themselves at the “interface” level of the book pages. An instance of this disposition is to be seen in the fact that the plays, in the First Folio, are (with varying insistence) divided into acts and scenes. The introduction of act-and-scene divisions goes together with the literary (and ultimately “classical”) claim made for the Shakespearian oeuvre by compartmentalizing his plays into three generic categories: Comedies, Histories, and Tragedies. Over the centuries, this division has become engraved ever more deeply in next-to-innumerable Shakespeare editions, and consequently adopted, too, as an encompassing matrix for Shakespeare scholarship and criticism. By the agency of the first collected edition of his works, Shakespeare, the playwright, became an author under a literary rather than a theatrical dispensation. Whereas his fellow-actors John Heminge and Henry Condell, as editors, underscore how carefully the texts that they offer “To the great Variety of Readers” were written, they say nothing about how they were played. They fashioned Shakespeare into a writer containable in a book, and in book conventions of textual representation. “Shakespeare the book,” moreover, was designed to be read: divided into its three main sections, as well as carefully patterned in the layout of its pages, the First Folio favors a reading encounter with Shakespeare’s plays. The one-volume Oxford edition revitalizes in one significant respect the First Folio’s mode of presentation. Its pages are text pages only, wholly without annotation, and arranged in two columns to the page. Its “firstfolio” reading effect has both an aesthetic and a didactic dimension. Page size and double-column printing perceptibly shape a play in its large structural outlines. The text gathers pace in the reading from its unbroken sweep over the pages as from a sweep of action on the stage. The Norton edition, by contrast, does not emulate this gesture of staging in the book design. It is consciously a book, and in this a study edition with a careful layering of editorial mediation, both through introductory essays and on-the-page elucidations of the plays’ language. While adhering in terms of its page design to the book tradition instituted by the First Folio, the Oxford edition in other respects breaks trebly with that first edition’s fashioning of William Shakespeare. First, it abandons the generic in favor of a chronological ordering of the plays. Secondly, while the First Folio contains thirty-six plays only, the Oxford edition utilizes its break with the generic ordering to incorporate, with these, not only plays of Shakespeare’s authorship or part-authorship excluded from the First Folio, but his short epics and lyric poetry as well. As against the edition of 1623, the edition of 1986 expands into an edition of Shakespeare’s complete works. Thirdly, and perhaps most significantly, the Oxford edition endeavors to edit Shakespeare’s plays in versions. Its greatest momentum inheres in its principle of establishing, wherever possible, the text of a given play closest

Late twentieth-century Shakespeares

89

to performance (or, to overstate the principle: of giving a company text in preference to a pure author text). A performance text, where identifiable, is a version text. To think editorially in terms of versions follows, as we have already argued, from assuming a revising author. Evidently, the edition thus intimately conjoins editorial and critical parameters in its design. Over the four centuries of Shakespeare transmission that originated in the First Folio, its generic ordering has provided a formidable a priori matrix, regardless of inconsistencies. The “Tragedies” section in the First Folio, for instance, is strangely framed by the apparent non-tragedies Troilus and Cressida and Cymbeline (and of these, moreover, Troilus and Cressida is even missing from the list of contents, the typographic show-piece headed “CATALOGVE of the seuerall Comedies, Histories, and Tragedies contained in this Volume”). But this never led to manifest counter-measures in the course of the later editorial tradition. Rather, on the grounds that the Folio editors must have erred, these plays were critically either reasoned out of any “tragedy” status, or, conversely, they were, off and on, actually reasoned into such a status by the pull of the Folio’s generic grouping. As to the “Histories” it explicitly introduces, the First Folio is, doubtless, genuinely innovative: “Histories” is not a classic, but a specifically Elizabethan dramatic genre. The First Folio’s grouping of the plays it considers belonging to the genre has distinctly furthered a sense, in later traditions of Shakespeare criticism, of a comprehensive design, for instance, of two tetralogies on British history, regardless of the circumstance (as it has become increasingly researched) that Shakespeare constantly switched genres as he moved in his writing from one play to the next. The First Folio’s strict compartmentalizing of his oeuvre into genres goes totally against the grain, too, of that overarching quality of Shakespeare’s art, its highly effectful mixing of genres in next to every individual play. Nonetheless, nobody had ever, before the Oxford edition, attempted to replace the generic by a chronological ordering of the oeuvre. This could simply not have been done without the historical and documentary research into Shakespeare and the English Renaissance accomplished in the twentieth century. Naturally, though, the Oxford editors needed to justify their procedure. The edition’s Textual Companion volume contains a chronological essay that, as it were, plots the sequential make-up of the text volumes of the Works of William Shakespeare. Between its modern-spelling and original-spelling text volumes, however, as well as meanwhile in its successive printings, the edition has entered into dissent even with itself. Owing to adjustments in the reasoning for the chronology, the order of the plays has repeatedly shifted. Blocks of pages have physically been moved so as to keep the book structure abreast with the reconsidered chronology.

90

hans walter gabler

The one-volume Oxford edition, then, comprises the entire Shakespearian oeuvre in all its genres. In its overall conception, it turns to editorial account the range of debate of Shakespearian scholarship and criticism of the twentieth century. The way this edition now images Shakespeare differs fundamentally from the image of the author conveyed by Shakespeare’s own contemporaries by means of the First Folio. This is borne out above all by its most controversial feature, namely, to present the plays in versions, wherever the evidence of the transmission so permits. Versions in the sense of the Oxford edition may be authorial versions, if they can be ascertained, as in the case of King Lear, for which the edition expressly establishes and presents two separate texts: The History of King Lear, and The Tragedy of King Lear. Yet theater versions are equally aimed at, shaped from authorial text by the company of players. Procedurally, the transmissions of the individual play texts have been closely scrutinized for evidence pointing to revision either in the course of composition or, significantly too, in the theater. The resulting edition texts represent endeavors to capture individual plays in a shape as close as possible to how they were communally texted and staged by the Lord Chamberlain’s / King’s Men, rather than exclusively as they were, in origin, drafted and written by William Shakespeare alone. In terms of presentation, transmitted text overflowing the bounds of the critically determined and editorially established versions, is in each case given as appended matter. More comprehensively than in the individual reassessment of King Lear, this conception for an edition, and concomitant editorial policy, is an outcome of the interest in the theatrical nature of Shakespeare’s plays resurgent in the latter half of the twentieth century. It takes measures against the received format of Shakespeare editions, which do not, on the whole, support explorations of the plays’ theatricality. By standard precepts of textual criticism and procedures of editing, instead, play texts too, as a rule, are and have been editorially seen and presented as book texts. Far too often, the cumulation of all text felt to have the true Shakespearian ring has been considered to result in the best, since most comprehensive, editions. But these have not, generally, yielded playable texts. Admittedly, it is through books that dramatic texts in general come down to us, and Shakespeare’s texts do so without exception. Yet this should not free the editor from taking account of their specific, performance-related textuality. Not to do so vitiates the theatrical potential inherent in the transmitted textual substratum of plays – demonstrably so, as we have seen, in the case of King Lear, where dramatic structures were effectively destroyed by the bookishly edited conflated text. As for theatrical and stage matters in bookish editions, these have commonly found their way at most into annotation, so as to render the edited texts more determinate as reading texts. For it is with readers far

Late twentieth-century Shakespeares

91

more than with theater practitioners in mind that editors have customarily marked act and scene divisions, localities, dramatic situations, or even entrances and exits, gestures, props, and general stage business. Often, the book-edited texts have, on the one hand, been over-determined as reading texts, in ways alien to Elizabethan practices of performance; on the other hand, they have seldom been editorially shaped out of a searching regard for the larger dramatic and theatrical sweep of performance texts. The 1986 Oxford edition, by contrast, adumbrates, where the evidence from the textual transmission permits, the shape of a Shakespearian play on the stage of Shakespeare’s theater company. It so endeavors to provide texts that are as close as possible to the theater. As a theatrically aware edition, however, it is erected according to the sophisticated standards of professionalism reached in Shakespeare editing by the end of the twentieth century. These standards it has itself, in fact, helped to advance. This is true not only with respect to the computer assistance it calls upon even just to keep the wealth of editorial material under control, or, say, to conduct research into questions of authorship or chronology. The advancement of editing standards is to be observed also in the mode the edition adopts to mediate its rationale and decisions to readers and users. The listings in the Textual Companion volume of variants for the individual plays, in particular, liberally employ natural language rather than the semiotics of editorial symbols to explain textual decisions, and so go a long way towards making the less-than-expert user of the edition understand textual problems and their solution. But by so discoursing its own procedures and decisions, the edition not only insists – and rightly so – on demonstrating that textual criticism and editing, by virtue of being critical activities, are always areas of debate; and that to consider them as merely legislative of “right” or “wrong,” were to mistake their nature. A central tenet of the Oxford edition is that “no edition of Shakespeare can or should be definitive.” Quite naturally by such a stance, it also opens itself up to genuine controversy. Notoriously (in terms of reactions provoked), for example, the text established for 1 Henry IV names the character Oldcastle who subsequently, in The Merry Wives of Windsor, 2 Henry IV and Henry V, carries the name of Falstaff – which, in fact, he already does, too, in the earliest extant textual witness to 1 Henry IV, namely the Quarto of 1598. The Oxford edition emends, in other words, the material text on which it is based; or, in terms of what can be inferred about the state of the text in composition, the edition restores what Shakespeare originally wrote. There can be little doubt that Shakespeare did first give his stage character the historical name. The combined external and internal evidence adduced seems convincing enough. It comes, on the one hand, from records of protest from the historical Oldcastle’s descendants

92

hans walter gabler

against the portraiture of their ancestor by name in Shakespeare’s play; and, on the other hand, from a minute metrical analysis revealing that, where the character is named in versified contexts in the play, the manifest reading “Falstaff” of two syllables disturbs the metrical expectation of a word of three syllables, such as “Oldcastle.” Yet if so much may be analytically ascertained – what options does an editor, and does an edition, have to act? Say, the editorial procedure is governed by the rules of copy-text editing: then the copy-text – in the nomenclature of the Oxford edition, the “control text” – will be followed unless the editor critically determines that the copy-text reading results from an error of transmission. In the case in point, it is clear from the metrical disturbance that a change has occurred in the transmission behind the Oxford edition’s control text; and what is more, the surrounding circumstances safely suggest that the change has been from “Oldcastle” to “Falstaff.” The question then remains, was this an error of transmission, or was it, say, a revision? Or should a third category be invoked, namely coercion through censorship? This is what the Oxford editors opted for when, countermanding the coercion, they decided to restore in their edited text the name that Shakespeare originally gave the character. In view of the evidence, it was, editorially, not “wrong” to do so; nor, however, was it unambiguously “right.” It was an editorial decision by judgment call – which is how editors’ decisions in establishing texts should largely be recognized. In the case in hand, were the critical assessment of the variant to shift, the drift of the editorial decision would alter. The Oxford editors need not be followed in their rejecting the reading they impute to censorship; for it is simply moot, in terms of editorial procedure, whether it should be considered obligatory to eliminate interventions of censorship in critically establishing an edition’s text. The simpler question remaining is then whether the variant, “Falstaff” (testified) for “Oldcastle” (inferred), be an error of transmission or not. If not, the change (critically posited) will be assessed as a willed revision – actually, an ingenious one! – which, by force of editorial method, must be accepted as offered by the control text. “Falstaff,” when granted the critical weight of a revision, gains collateral support, moreover, from the consistent linking of character and name in all four of the Henry IV / Merry Wives of Windsor / Henry V plays as they have materially come down to us. Errors of transmission aside, which by common rules of editing need to be eliminated, it is in the forms in which they have been preserved, ultimately, rather than in the undocumented and therefore mostly indeterminate variant shapes Shakespeare first wrote them in, that his plays have established their theatrical tradition, as well as their textual one. The theatrical shape of Shakespeare’s plays, as we have already emphasized, provided a dominant focus for the Oxford edition. As a name, “Falstaff” is inscribed in the

Late twentieth-century Shakespeares

93

theatrical tradition. For this reason, too, it was a trifle inconsistent, in terms of the edition’s sense of versions, to have called the antagonist of 1 Henry IV “Oldcastle” in the edited text, notwithstanding the fact that this was nearto-demonstrably Shakespeare’s original name for him. As the flaw this retronaming is, however, it shows up in a nutshell, too, into what complexities editorial reasoning may at times be running, and it consequently moves into focus how needful it is that editors as well as users and readers of editions engage critically with transmitted and edited texts. Looking at the alternatives of adjudication in this one instance: is “Falstaff” an error of transmission; the coerced response to censorship; a revision; or authenticated by tradition; one begins to realize that larger issues are at stake. Of the options for assessment, it is especially the last that touches not only on the enterprise of the Oxford edition of the Works of William Shakespeare as a whole, but impinges equally on theorizable assumptions by which we understand and categorize authorship, oeuvres, works, texts and editions; as does, similarly, the King Lear issue discussed above. The focus on “Oldcastle” versus “Falstaff” is author-centered and concerned with authority. In this one instance of an alternative, as with the two whole versions of King Lear, we are faced with initial and revised authorial text competing for authority in editorial terms. For King Lear, it is possible to differentiate two whole texts and to edit each on its own textual authority. Between “Oldcastle” and “Falstaff,” on the other hand, one must decide; it is impossible to uphold both names in editorial, and thus textual, parallel or simultaneity. In terms of textual criticism, and Anglo-American editorial rationale in particular, one has customarily at this point taken recourse to the tranquillizer of authorial intention: “Falstaff,” as the revised reading, is declared to overwrite, on the strength of later authorial intention, the reading “Oldcastle” that precedes it. This is likely to be true, factually. Yet it remains a hermetically author-centered reasoning. It has little regard for the autonomy a text attains in its accomplishment. The progression of Falstaff through 1 Henry IV, The Merry Wives of Windsor, 2 Henry IV, and Henry V is the progression of an autonomous fictional character through texts originally conceived and written by Shakespeare. Yet the fictional character acquires autonomous agency, too, in that he contributes directly, in his turn, to constituting these texts as accomplished works, as well as to linking these works into a quadruple sub-unit of the dramatic oeuvre. In the larger perspective, furthermore, the oeuvre must then be seen as determined not solely by the author, Shakespeare, in complicity with his fictional characters, but together with a company of players as well, the Lord Chamberlain’s / King’s Men. Indeed, it is on the joint realization of the oeuvre in writing, on the stage and in an edition, the First Folio, that the afterlife of “Shakespeare,” and with him Falstaff, depends.

94

hans walter gabler

Such might be the reasoning for the Oxford edition’s choice to edit “Shakespeares,” that is, wherever possible to edit theater versions from the earliest surviving evidence to Shakespeare’s plays in quartos and First Folio. The implications are momentous. For all its endeavor of meticulously preserving and restoring Shakespearian text in detail, the edition’s focus, when taken in as a whole, is no longer exclusively on one author’s (Shakespeare’s) authentic text. More widely seen, it is on an inherited body of texts as oeuvre that, while it has the general authority of Shakespeare’s name behind it, equally carries marks of an authority ramified by transmission through editions, as well as by each fresh performance. The edition is therefore trebly centered: centered on authorship, similarly on text and textual transmission, and on theatrical realization as well. The authenticity, if granted, of the versions of Shakespeare’s plays that the edition establishes, rests therefore jointly on foundations of the texts’ composition, of their life and after-life on the stage and in their first editions, and, further yet, of their manifold rootedness in the cultural heritage. Such thinking, when pursued to its conclusions, is likely to confine scholarly editing no longer to routines and formalisms trained solely on authorship, authority, intention, or judgmental distinctions between good or bad, right or wrong, readings. Instead, it provides renewed openings for literary criticism to re-integrate textual criticism as one of the fields of its overall hermeneutical endeavor. Such new openings in Shakespearian editing call upon literary criticism, in turn, to reconsider its bearings. Through the advances in presentation formats of Shakespeare editions over the past couple of decades we are meanwhile provided with paradigms for engaging innovatively with works and their texts. As published in 1986, the one-volume Oxford edition of the Works of William Shakespeare offers, as said, two Lears: The History of King Lear, and The Tragedy of King Lear; and it constructs from the transmission, wherever it can, theater (company) versions of the plays. The Norton Shakespeare adoption of 1997 of the Oxford edition goes further yet in the case of King Lear: firstly, it prints the History and the Tragedy not successively, but on facing pages. Secondly, it provides a third rendering of the play in a freshly established conflated version. In so doing, however, it nonetheless does not uphold the claims made for conflated Lears for over two centuries, from Alexander Pope to Kenneth Muir. The Norton conflation is no longer given out as the ideal realization closest to William Shakespeare’s own authoritative text. Instead, it stages, wholly on the editor’s responsibility, a text analogous to any text adaptation that, say, a theater company would feel free, on its artistic responsibility, to assemble for performance. The stance, in other words, is not one of editorially decreeing an authoritative text. Instead, it is reception-oriented. It allies itself with us at the receiving end where we engage with the play – and where in fact Barbara Lewalski, editing the new conflated version, equally engages with

Late twentieth-century Shakespeares

95

it, as she mediates the work by way of critically merging the texts from the documents of transmission. Texts become what they are, and what they mean, through critical engagement with them. This is a home truth in literary criticism. The advances in Shakespearian editing through the latter years of the twentieth and meanwhile into the twenty-first century contribute to recognizing that it is valid, all the more strongly, in face of the material evidence as it survives for Shakespeare’s texts. Editions are taking a development toward inviting critical engagement with themselves as products of critical scholarship. The prime question to be asked, be it in editorial or in critical terms, is no longer simply the normative one of “what is the work’s text?,” but rather – and in far greater complexity – the critical one of: “what is the range of variability of the work’s text(s), from which we may gain analytic advantage toward interpretation?” Moreover, this is not a recondite question for the scholar in her closet. First and foremost, it has vital repercussions in the classroom. Shakespeare editions are becoming available at an increasingly variegated range: the Oxford and the Norton editions, the growing numbers of quartotext editions alongside the standard edited-text editions in the several series of one-volume-per-play editions such as the Arden Shakespeare, the facsimile editions, the electronic editions expandable beyond traditional formats to becoming learning sites.11 What brackets them nonetheless is an encompassing perspective. They establish before our very eyes that it is fundamentally in the nature of writing, indeed of language, that texts are variable and can always also be otherwise. Their united challenge is therefore to use to critical advantage the variability in the textual condition that they, as editions, give access to. To acknowledge the nature of the textual condition as one of fundamental variability contributes to securing the theoretical grounding for a re-integration of textual and literary criticism. Perceiving them as integral is, in truth, anything but innovative in an historical perspective. Yet it is, in our day, still a fresh conceptual departure and should affect both sides of the divide between criticism and scholarship that we have perhaps as yet not wholly overcome. Its gains are most likely to be secured through intense debate between orthodox attitudes and new beginnings. (Bibliographical addendum: Since completing this contribution to the present volume, I have discussed at greater length the issue it takes its departure from, namely the transition from stemmatological to authorand intention-oriented scholarly editing, and W. W. Greg’s role therein in “Beyond Author-Centricity in Scholarly Editing.” A further recent publication has bearings on the medial shift from book to digital of today that plays a distinct role throughout this volume: “Theorizing the Digital Scholarly Edition.” See the “Guide to further reading,” at the end of this volume for full bibliographical citations of these essays.)

96

hans walter gabler NOTES

1 For the main contributions in support of this position see in particular Michael Warren, “Quarto and Folio King Lear and the Interpretation of Albany and Edgar,” in Shakespeare, Pattern of Excelling Nature, ed. David Bevington and Jay L. Halio (University of Delaware Press, 1978), 95–107; Gary Taylor, “The War in King Lear,” Shakespeare Survey, 33 (1980), 27–34; Steven Urkowitz, Shakespeare’s Revision of King Lear (Princeton University Press, 1980); Peter W. M. Blayney, The Texts of King Lear and their Origins, Vol. i, (Cambridge University Press, 1982); and, most diversified in its approaches, The Division of the Kingdoms, ed. Gary Taylor and Michael Warren (Oxford University Press, 1983; paperback 1987). Throughout this essay, I draw on three earlier publications of mine: Hans ¨ Walter Gabler, “The Two Versions of King Lear: A Review Article,” Archiv fur das Studium der neueren Sprachen und Literaturen, 225 (1988), 137–44; Klaus Bartenschlager and Hans Walter Gabler, “Die zwei Fassungen von Shakespeares King Lear: Zum neuen Verhaltnis von Textkritik und Literaturkritik,” Deutsche ¨ Shakespeare-Gesellschaft West: Jahrbuch 1988, 163–86; Hans Walter Gabler, “Assumptions Fruitfully Disturbed: The One-Volume Oxford Shakespeares,” Deutsche Shakespeare-Gesellschaft West: Jahrbuch 1989, 344–50. 2 Madeleine Doran, The Text of King Lear (Stanford University Press, 1931). 3 See the discussion in Steven Urkowitz, “The Base Shall to th’ Legitimate: The Growth of an Editorial Tradition,” in Taylor and Warren, eds., Division, 23–43, esp. 31–3, with n. 13. 4 Jeffrey Kahan, ed., King Lear: New Critical Essays (New York and Abingdon: Routledge, 2008). 5 Ibid., 104–23; 124–54. 6 See Taylor, “The War.” 7 Roger Warren, “The Folio Omission of the Mock Trial: Motives and Consequences” in Taylor and Warren, eds., Division, 45–57. 8 Gary Taylor, “King Lear: The Date and Authorship of the Folio Version” in Taylor and Warren, eds., Division, 351–468. 9 William Shakespeare, The Complete Works, gen. eds. Stanley Wells and Gary Taylor, The Oxford Shakespeare (Oxford: Clarendon Press, 1986); The Complete Works. Original-Spelling Edition. (Oxford: Clarendon Press, 1986 [1987]); Stanley Wells and Gary Taylor, with John Jowett and William Montgomery, William Shakespeare: A Textual Companion (Oxford: Clarendon Press, 1987 [1988]). 10 Stephen Greenblatt et al., ed., The Norton Shakespeare, Based on the Oxford Edition (New York: Norton, 1997). 11 See for instance a model for an electronic edition of King Lear, dynamically relating the versions within the larger framework of an edition as a learning site, devised by Alexandra Braun-Rau: William Shakespeare’s King Lear in seinen Fassungen. Beihefte zu editio 20 (Tubingen: Niemeyer, 2004). Electronically, the prototype may be accessed at www.textkritik.uni-muenchen.de/abraun-rau/lear/ index.html.

5 PAUL EGGERT

Apparatus, text, interface How to read a printed critical edition

Encountering a printed scholarly edition for the first time All editions of literary works interpose between the reader and earlier documents that present the text of the work. Most new trade editions of literary classics take the easy option and typeset or reprint from the copy lying nearest to hand, perhaps with the application of a little extra copy-editing polish. Whether reliable or not, the source document, those others not considered, and now the new edition, form part of the work’s transmission over time. Scholarly editions differ from ordinary trade editions in that they typically enlarge the field to take account of all or many earlier forms of the work. The ambitiousness of the survey introduces complexities that have always made such editions hard to use at first. Now that digital-electronic forms of presentation have become more fully naturalized, the otherness of the printed scholarly edition is perhaps the first thing that is likely to strike us as we turn the pages of one for the first time. We see straight away, as our forebears tended not to, that we need to make a deliberate effort to understand how it works. This in turn means that we need to appreciate the series of problems the editor has had to resolve. These mainly bibliographical problems turn on the relationships of the various forms of the text to one another. The analysis of the history of the text’s transmission over time, from version to version in document to document, printing to printing, is called textual criticism (as opposed to literary criticism). It leads to a defense of the editor’s decision about how to treat editorially the textual and documentary forms that have been surveyed. What principles might justify the editor in establishing a single reading text of the work from among its variant forms? A self-evident need to cater to the expectations of the reader? A moral commitment to respect the author’s intentions? Or an ideological one to the work as aesthetic object, which needs only to be cleaned of the textual detritus acquired over the decades and centuries in order to stand now, in effect, outside of history? Alternatively, if the editor believes that the work ought to be understood 97

98

paul eggert

as being constituted by its history of textual variation, then the goal of establishing a single text to present the work makes far less sense. More emphasis will fall on the so-called apparatus that records the work’s textual history over time. The editor’s study of that history will have revealed its own peculiar problems, to which both kinds of edition must respond in some coherent way. Given this situation, there can be no simple, one-to-one relationship between the work and the edition. As the texts of works never come down to us in unmediated form, every scholarly edition is necessarily an embodied argument about the text or texts of a work. Definitiveness is thus ruled out as a feasible outcome, though reliability for a specified purpose, or to respect a specified goal, is certainly not. The editor is always aware, in every decision about the text, that he or she is by nature an interloper, a traverser of the textual field of the work. The editor’s travel guide, as it were, is one or other of the competing traditions of editorial endeavor. Its conventions will have to be adapted to respond to the discoveries of that text-critical journey and later to usher the reader editorially along the same route. The edition’s condition as text-critical argument soon unfolds, for the editor, into the strategic question of the architecture and interface of the volume. The edition’s internal organization, declared conventions, and page layout become the systematic embodiment of the editorial argument. How to understand that system better is the theme of this chapter. As we learn the ropes, we turn from being na¨ıve and captive first-time readers of editions into informed and properly skeptical ones. We learn to address ourselves to the question that all editions implicitly raise: whether or not we ought to be convinced by the argument that the edition constitutes. This chapter mainly restricts itself to printed editions simply because, at time of writing, electronic editions remain in active development and have not yet settled down into typical formats. Printed scholarly editions will in any case remain in general use as reference books for a very long time, if only because of the quantum of scholarly labor, and thus the unrivalled resource, that they represent. Coming as we do to them with a twenty-first-century mindset, the effort of adjustment means that we cannot but be aware of their material interface. These editions encode things differently from a normal book. Like a big website or an electronic archive, they have much ground to cover, much information to record and to link via cross-references. These latter, analogue versions of the hypertext link must be precise and economical, especially when notating the great many differences between the versions. In the absence of supplementary volumes of facsimiles, a scholarly edition can rarely display within its covers the original materials. Unlike a

Apparatus, text, interface

99

website, its interface must needs be highly mediated, not accumulative. The labor of comparing, connecting, and commenting must already have been done. Situated as we are at the point of changeover in the modes of delivery and analysis of text, we have a windfall opportunity that only our very alert forebears managed, more painfully, to acquire: to understand the printed interface of the edition (text, apparatuses, tabulations, appendices) as an expression not only of technology but also, as I will go on to demonstrate, of various beliefs about the nature of the work. The accommodation can take fascinatingly different forms, both in the Anglo-American editorial tradition (with which I am primarily concerned here) and in the German. But the economics of publishing mean that relatively few works are ever accorded the luxury of being presented in competing scholarly editions. Shakespeare is an exception. For this reason I have chosen, in the second half of this essay to look at Hamlet as a case study. Because it is one of the most deeply canonical, culturally ineradicable works in English, when it comes to weighing up the claims of the competing texts, the editorial stakes are very high indeed. Which text should be presented to readers? And how? As we shall see, the question of interface, and the editorial engagement it expresses, cannot be separated from the substantive bibliographic questions of the work’s production history. Whether the edition is analogue or digital, such difficult questions as these remain unavoidable for the reader. Indeed, they are the very conundrum that the edition offers to resolve. At first though, to simplify the matter, I will deal with scholarly editions of modern works and further pursue the question: how to read a printed scholarly edition? The central matter In the Anglo-American editorial tradition the central matter has always been the reading text. At first, the provision of it by the editor seems like a nonissue since a literary work has obviously got to have one. And sure enough (here is the primal scene), when we open the edition, we soon see that there is a text waiting and we settle in to read it. The design of the volume will usually encourage us in this preconception, towards which our prior experience as book readers predisposes us. Indeed, many editions launch us directly into the clear text and leave all the editorial matter for the end of the volume. As we start to read, it is as though we have entered directly into the work, as if it is transparently available to us and self-identical. We move along nicely, reading in the normal way until we hit a snag. What can this line mean?, we ask ourselves. Or: this wording feels like a political

100

paul eggert

allusion, but to what is it alluding, and how does that affect the intended meaning? We are used to reading for intention. It is one of human culture’s enduring and most rewarding games. We practice it in every conversation we have. In our daily reading, because we come across many ambiguous signals, we have learned to sort out whether the text-fragment we are trying to decode is, as the computer scientists say, (malformed) signal or merely noise. Is it an accidental typo or some deliberately unusual verbal formation? We get irritated if we can’t sort this out. We go unrewarded if we get it wrong. Having hit the snag in the reading text of the edition, we may encounter a discreet signal – an asterisk or a tiny superscript note number – pointing us to an annotation. Forcing ourselves to stop reading the text, we consult the volume’s Contents page, discover where the annotations are to be found, and noting the page and maybe also the line number in the text we have just come from, we track down the explanation we are looking for. Darkly muttering to ourselves that it would have been easier to have had it clickable or (sorry, wrong medium!) at the foot of the page, but glad nevertheless to have had the meaning clarified, we push on. This happens a few times. We start to think we are getting the hang of the thing. The reference book’s structure seems to have a logical “grammar,” rather like computer files do, so that they can be processed by the computer’s operating system and software. The notes are attendant lords; like T. S. Eliot’s Prufrock, they may swell the royal progress of the text we are reading but they remain strictly subservient to it. They are meant to make its old-fashioned or other-worldly decorums legible to us. Sometimes, however, they seem to have a will of their own, one that is apt to upset our old, instinctive way of reading. Maybe our next such sortie to the annotations lands us in this new kind of trouble. On this occasion, we find the volume’s editor sending us off to a different part of the volume to compare other information there, and perhaps outside it altogether. The volume we are reading, in this scenario, is the edition of Women in Love in the Cambridge University Press Letters and Works of D. H. Lawrence. There is a note on the oddly intense character Hermione Roddice. She wants to gather the novel’s principal character Rupert Birkin, an inspector of schools, more securely into her social, intellectual, and sexual orbit. She is a sophisticate, and literally out of his class; but they have been having an affair. Now both attracted and repelled, he strikes back at her unrelenting insistence with merciless psychoanalyses of her. Later in the novel, standing behind him, she cracks him over the skull with a lapis lazuli ornament. He moves at the last instant, injured but not killed. Where is this coming from?, we feel ourselves asking. And why is this novel so frequently, so nearly unbearably, on the edge? We turn to the

Apparatus, text, interface

101

note.1 It refers to the surname Roddice being common in Eastwood where Lawrence grew up, to his mixing up the name in one of the typescripts of the novel with another name he had elsewhere used for fictional versions of his childhood sweetheart Jessie Chambers (the Miriam of Sons and Lovers), to references in Lawrence’s letters indicating that the character owes something to the high-society hostess Lady Ottoline Morrell, to cross-references to other notes that confirm and extend this correspondence, to Lady Ottoline’s own view of Hermione and Lawrence’s response, as documented in one of his letters. The note finally quotes another letter where Lawrence contests as an over-simplification the idea that Hermione is Ottoline. All this, and more besides, is done in seventeen lines of type. No overtly literary-critical opinion or ready-made interpretation is offered, and every statement is linked to a printed or documentary reference – even if the language, so packed with information, is less than elegant. It is tempting as a reader to pause and follow up some of these references. If we do so, we find an ever-widening circle. The note unfolds “Hermione” into the politics of Lawrence’s friendships, both personal and professional, and into some of his other writings and his views about the sources of inspiration for all of his work. So the note’s very effort of explanation, which we supposed was there to clarify and undergird a difficulty we had in reading the text, ends up working with equal strength in the opposite direction. Looking to have our uncertainty about the meaning or sources of what we have just read closed off with authoritative information, we simultaneously find that the text-fragment “Hermione” has become the temporary focus of a widening contextual and biographical web. Teetering on the edge of our own reading experience and forced temporarily to adjust the bookish instinct of how-to-read that we brought to it, we realize that we are not first-time readers any more. This edition has challenged us – as we really pay attention now – to think about what reading-the-work consists in and therefore about what the work is.2 What the interface is telling us Having got to this point, it is not long before we realize that the text of the work we are reading is adhering to more than one web. The crucial moment is the one when we realize that we can either read the text that the editor has established or we can bypass it, using it only as a point of reference in order to, as it were, choose our own text. Rather as for a densely hyperlinked website, we can choose to read in the work, among its versions, comparing the differences. If we do, we will soon find ourselves prompted to wonder why this variant was rejected for that one and by whom. Then perhaps we begin to see, or to fancy we see, patterns of correction, or deliberate revision,

102

paul eggert

or bowdlerization. We begin to appreciate the history of the work’s coming into being through its successive textual forms. All of this, a good edition makes possible, even as it records the editor’s adjudication of these matters in the reading text. The adjudication is inevitable, for if there is to be a readable text, established according to some defensible criterion, then some variants will have to be accepted in favor of the very many more that are rejected. So, mixed in with the annotation, or perhaps in a separate section of the edition called Commentary, we find many notes that discuss the verbal make-up of the text: how the editor has treated the author’s changes of mind, the alterations of the text made by copyists, the bowdlerizing or regularizing done by in-house editors fearful of prosecution or loss of sales, and the multiplying alterations of the text as one edition followed another over the years and as the new typesetters’ attention wandered. It is not a widely known fact that, for a novel, there can be many thousands of changes. The bulk of them will be matters of detail – punctuation, spelling, capitalization, and word-division – but, even so, the effect of them adds up. If we want to know how the author heard the sentence rhythms and cadences as he or she wrote, or how far the author’s spelling habits differed from the typesetters’ practices of the day, or if we begin to ask what text, exactly, those first readers had access to, then we realize that we need access to those variant forms and wordings. We begin to sense, as we pore over the editor’s recording of the variants, that this work has had a life of its own, a process of coming into being, and then a further process of being socialized through taking public forms in print. Its text did not remain stable over time: which is the opposite of the expectation with which we have naively come to literary works – or at least to the books that embody them. This dawning awareness complicates things for us. We used to think that if texts are unstable then that is because, as countless critics after Foucault have shown, they are the site of discursive inscription. Thus they are open to varying kinds of appropriation by readers, who, metaphorically at least, become “writers” in the act of reading. But, counter to this conventional understanding, the new awareness I have been describing seems to be pointing at a literally physical kind of textual instability: a different sense in which a work is never identical with itself. Reading “in” the work makes us realize that the text of the edition is, as it were, always already looking over its shoulder at its alternative versions. Its apparently flat bibliographic surface is in fact textually porous; and the textual apparatus is telling us where it is and how. This new awareness, into which the edition is drawing us, is fundamental. If all works have variant texts, and if at least some of these variant texts are retrievable by editors, then the nature of the questions we can ask of

Apparatus, text, interface

103

literary works alters. The differences revealed by the edition always lend us a certain leverage on the text, since every moment of its copying is a crisis point in its history. Each change in the text, or in the way it looks on the newly designed or illustrated page, raises a question as to its origin, its agency (who did it?), their motivation and its likely effects. This shift does not negate discursive analysis. Every change of mind about the text of the work may point us off in new directions altogether, or suggest newly clarified, more precise questions that hook the text back onto its successive moments of production, distribution, and reception. This factor of variance, and the questions it raises for us as we attend more closely to it, forms the second (textual) web that the edition of the work must reveal to us. But it is a potentially sticky web; there are traps for the editor who must weave skillfully and tread lightly if we are to be able to follow where he or she has gone before us. The text, apparatus, and appendices – the edition’s materialized interface – must carry the results of the editor’s research, while also allowing the reader room to move and air to breathe. The editor has normally studied all the extant manuscript and printed material that witnesses the passage of the work through time and has compared at least the early ones word-for-word and comma-for-comma. This usually computer- and machine-assisted process is called “collation” and it forms the basis of an extended report in the edition, usually called the Historical Collation. For practicality, it is often restricted to variant wordings (“substantives”), leaving the now-orphaned commas and other “accidentals” to look after themselves. The editor has worked out – bibliographically – how each of the versions relate to one another: which served as copy for which, and thus what the line or lines of descent of the text was or were. (This can usually be achieved for modern works, but for medieval and earlier works the answer will often at best be probabilistic since, typically, the original manuscript is lost and some early scribal copyings will not have survived.) Provided the archive of surviving versions and letters is rich enough, the editor will, in addition, have been able to appreciate and to communicate a sense of how the text has emerged from the biographical–contextual weave of the author’s life. How to allow us to traverse those textual and contextual webs so that we can ask our own questions of the work is the problem that the editor has to solve. The editor knows that whatever form the presentation takes, our initial desire to read as we normally do – in a linear fashion, uninterrupted – will have to be infringed in one way or another. What forms, then, should the editor’s record of variant readings – the textual apparatus – take? How should the text or texts be physically presented? Surely, we are entitled to ask, this is merely a technical problem that ought to have been solved for reference books once and for all, long before this?

104

paul eggert

Forms of apparatus By the late 1960s it looked like it had been: or, at least, the Anglo-American school of editors assumed as much as their evolving practice became institutionally entrenched in American universities, with its methodologies validated by the Center for Editions of American Authors (CEAA) or of its successor body the Modern Language Association’s Center (later Committee) for Scholarly Editions (CSE). The edition would have at its core a clear reading text of the work that the editor would have to “establish” from the several versions that might be extant and from any others whose existence and make-up could be inferred. Usually taking the first completed form of the text as a basis (the “copy-text”), the editor would incorporate into it whatever changes the author had made: in proof, say, or for subsequent editions. If, for instance, a manuscript in the author’s hand was extant but its proofs were not, and yet verbal changes in the first edition had all the hallmarks of the author revising, then the editor would feel justified in incorporating those changes into the copy-text. In being an amalgam of two or more sources, the reading text that resulted was said to be “eclectic.” The editorial effort was to more nearly approach an ideal: the text that the author would have wished to see published, had his or her control over it been perfect. This is the reason for the term “critical” edition. Over the last several pages, that is the kind of edition I have been describing. The business of determining what the author intended is inevitably a critical act, though in a critical edition its subjectivity is traditionally hedged in at every point by whatever can be ascertained or inferred about the history of the work’s writing and early production. It is worth noting, however, that the term “scholarly edition” is a wider one. It also embraces editions that exclude critical determination in favor of reproducing uncritically the texts of existing documents. Furthermore, some scholarly editions, such as the Cornell editions of Wordsworth and Yeats, provide, both in facsimile and transcription, the raw materials for a study of the work’s evolution. The emphasis is on recording the compositional genesis of each of the works. The Cornell Yeats, for instance, presents facsimiles and transcriptions of the extant documents, but no critically established reading text as such. This appears to matter little at first since, for most of his poetry collections, Yeats was able to supervise and sanction the final forms of his poems. But he died before the texts of his Last Poems were finalized, and many of the poems remained in multiple manuscripts or heavily revised typescripts. So the critical question remains.3 While many variations on the textual situation I have been describing emerged over the years, the critical edition proved flexible enough to deal

Apparatus, text, interface

105

with them. The clear reading text was supplemented, usually at the back of the book, with a list of the editor’s emendations (this was the primary listing). If a manuscript existed, its deletions and additions would be recorded separately. Then would follow the historical collation. The peculiar virtue of the chosen copy-text was that it either was, or was the closest available version to, the author’s own manuscript. It therefore gave access to his or her habits of spelling, punctuation, and all the other aspects of presentation of the words of a text. As W. W. Greg famously argued in 1950, these so-called accidentals were the things that typesetters of new editions were more likely to change as part of their ordinary business of preparing a text for printing. They were far less likely to change the wording on their own authority.4 Because all the editor’s changes to the copy-text were recorded in the primary list, and because the historical collation gave all the later variants in wording, the reader had the resources to accept or challenge the editor’s critical decisions when establishing the reading text. If in fact the author had played no part in those later editions the record of them would be mainly a record of typesetters’ and others’ unauthorized changes (“corruptions”) of the author’s text. But, even so, the list would be a reminder of the unsatisfactory nature of the texts of the work currently available in print – since all of them would, by one route or another, have descended from those early ones. Book logic, and editorial dispute about the nature of the work To be intelligent readers of printed critical editions we need to understand what notion of the work the edition is, in its material procedures, enacting: in other words (looking at the book from a digital vantage-point) what its analogue interface implies. The date of the edition, the orientation of the series within which it falls (usually spelled out in a general editor’s preface), and the intellectual–editorial allegiances that the editor declares or assumes, will help us orient ourselves in relation to it. Hitherto, the naturalizing capacity of the book, and in a more sophisticated form, of the printed critical edition, hindered active perception of this. But our experience of the electronic domain means that the signals line up differently than before. If we think of an electronic edition or “worksite” as providing all the materials and tools we need to build our own edition, then we see instantly that all printed scholarly editions, no matter their prestige, can only be an intervention in the life of the work – though in most cases an exceedingly well-provisioned intervention, to be sure. Only lazy reviewers, in awe of the sheer amount of effort that it has taken to bring the printed critical edition to fruition, claim any longer that such editions are “definitive,” that the text of the work has

106

paul eggert

been settled once and for all. Editors themselves never say this, for they are all too aware of their own intervention, and of the incompleteness of the evidence on which they have had to adjudicate, repeatedly balancing one set of reasoning or inference against another. They know all too well the epistemological abyss over which their decision-making so continuously hovers. They know, too, that there is no inherently right approach. There are only arguable approaches: either well argued or ill. But editors will not necessarily concede that the critical edition reflects, at the fundamental level of edition design, an ideological commitment: that it expresses an aesthetic–philosophical assumption about what the literary work is. However, comparison of the text-display and apparatuses of AngloAmerican critical editions with German historical-critical editions readily bears this observation out. With the latter, especially from the 1960s, far more effort was put into devising ingenious forms of apparatus to capture textual variation in manuscripts and printed versions than on establishing the reading text. The latter was selected from one of the authorized versions essentially to help reveal the textual history of the work: whereas, in the Anglo-American edition, it presented the work. That was the central aim of the edition, even if critically based and even though the editor knew it could not be definitively achieved. Amongst editors of the same post-war period, the influence of New Criticism helped justify the common-sensical assumption that it was the business of editors to give readers what publishers assumed they wanted: a single, not multiple, reading text that represented an ideal form of the work, to which all other documentation in the volume would be ancillary. This manner of proceeding had the happy consequence of allowing that reading text to be lifted from the critical editions and reproduced alone by photolithography in cheap, student, and general-reader editions. In other words, an empirical book logic prevailed.5 The main difference of opinion in post-war critical editing was not here. Rather, it revolved around the question of when the work being edited can be said to have come into existence: in draft, in final manuscript form, or upon its first public appearance in print. If the author’s intention all along was to write a work for public consumption, how could a pre-print form of it be chosen as copy-text – as the doyen of the American post-war movement, Fredson Bowers, advised? Would not the author have anticipated that his or her punctuation and spelling would be tidied up by the typesetters? Did not that expectation passively authorize all changes made in the first edition, especially if the author was known to have corrected proofs? Did it not therefore make sense to take the first edition as one’s basis, correcting only its typographical errors (which the author cannot have intended)? Those

Apparatus, text, interface

107

corrections could be sourced from the manuscript or elsewhere, allowing the editor then to incorporate whatever revisions for later editions the author was or could have been responsible for. This latter editorial approach, employing copy-texts drawn from a first or a later edition extensively revised by the author, was associated with James Thorpe in the USA and Philip Gaskell in England. It retained adherents and had its effect on many critical editions, especially on the textual apparatus. It minimized the size of the apparatus (and therefore the costs of the printing) since, conventionally, pre-copy-text variants did not have to be reported. But the general argument and this practical consideration struck the unavoidable Greg–Bowers rock: that the earliest extant form of the completed text would always be the closest one to the author before the collaborators in book production had become involved. This went to the basis of what it was felt underwrote the existence of the work. If it was authorial intention (as all the participants in the debate seemed to agree), and if the work was understood to exist in an aesthetic realm (as opposed to that of a historical document, another point of agreement), then clearly its text was ideal and only the critical editor could recover it. This was because its text would inevitably have deviated from its author’s intent, both in initial production and over time. Only the works of such self-producer poets as William Blake and Emily Dickinson evaded the general fate. (He illustrated and printed his own poetic books, and she copied out her poems for an immediate domestic readership.) Editors of this Greg–Bowers persuasion had to accept an expanded responsibility for recording variants. In order to cope, the duty led them to record only the substantive (but not the accidental) differences between the variant versions. The interface, in other words, expressed the editorial commitment; and the editor’s Note on the Text summarized the compromises (the “silent categories”) that had been implemented. Shifts in the 1980s and 1990s From the 1980s the general understanding amongst editors of what a work is began to shift. The significance of the shift went little noticed at the time because of the rise and rise of poststructuralist-based forms of literary theory, with the redefinitions of text, as described above, and the attendant marginalizing of the concept of the work. Editors began understanding textual authority differently. How could it derive solely from authorial intention when readerships participated in the life of the work – as they realized it in the act of reading? If that were the case, then would not the text and its material form (the original volume design and page layout) have entered into that realization? Alternatively (as in the thinking of German editors

108

paul eggert

from the 1970s), if text is better defined as a system of semiotic relationships then how could an eclectic text of the Anglo-American kind, based on the recovery of final authorial intention, do other than misrepresent those relationships? They operated synchronically (in a text as published and read) whereas the eclectic procedures of critical editing operated diachronically. They gathered into the one presentation readings from versions separated from one another in time. There were notable experiments with apparatus in the 1980s, inviting new forms of reading. Like earlier experiments with non-critical “diplomatic” editing (where the editor deliberately recorded manuscript alterations in-line, within the edition’s reading text, rather than in separate apparatus), they encountered too much resistance at the time to give other editors the heart to follow suit. Longstanding Complete Works editions, which could not readily change their methods mid-stream, had a conservative effect, discouraging editorial innovation. It is unlikely in any case, that their publishers, who were investing large sums of money in the printing, would have allowed such experimentation. Editors began to look forward to the liberation promised by personal computers, and then after 1992 by the newly invented internet. Meanwhile, by the mid-1980s some anglophone editors, chafing against the chain of pragmatic book logic, began to report in conference papers on the importance of the textual variation in the works they were editing. For instance, the production of works published after the promulgation of new US copyright laws in the 1890s required that separate copy be provided to publishers in London and New York to allow for simultaneous publication. Authors, revising these separate copies, seemed perfectly willing to allow variant texts to go into production. Yet the critical editor was obliged to reduce this variation to a single text and relegate the rest to the apparatus of the edition. Distinctions between the edition’s privileging of textual “product” (the clear reading text) as against the textual processes of composition, revision and production began to be made. Mark Kinkead-Weekes, whose work on his edition of The Rainbow had begun in the mid-1970s, would comment in its Introduction in 1989: “there are vital ways in which a work of Lawrence’s is its process of becoming, which an edition should try to preserve in some fashion, however necessary it may be to choose between variants in order to produce a readable text.”6 In 2002, John Bryant’s book The Fluid Text belatedly named, as it brought together, the threads of discontent that had been growing for the previous fifteen years. For some editors, the text–apparatus interface too tightly constricted the new understanding of what works are. The allure of the flexible interface of the electronic domain can thus be readily understood. I return to it at the end of the chapter.

Apparatus, text, interface

109

Case study: Some recent editions of Hamlet I began with the question: How should we, in this age of electronic text, read a printed edition? The argument so far, that we must first understand what conception of the work underlies the edition, implies, in turn, that critical editions don’t just advance an argument about the work’s textual transmission over time. They also materialize that argument in the edition’s interface. How does this happen in practice? How, in other words, does the argument enact itself on the page? So far, the references have been to works of the last century or so. We move now to a more complex case. Rather like the contrast between German and Anglo-American editions touched on above, it demonstrates how the substantive editorial argument ramifies through the interface, which of course, being that, is what we encounter first. Nearly all educated readers will, during their schooling, have come across an edition of Hamlet. It will have taken some adjusting to. We learned what to ignore for our present purpose. So we headed for the great speech or scene, ignoring all temptations offered by line numbering to go off to the one or two sets of footnoting. Or we tried to engage to some extent with the resources in the edition at a deeper level. We perhaps then recognized that the markers we considered firm ground – such as Act–Scene numbering or characters’ names or stage directions or even spelling – were not firm at all, but had only become so because of the editorial tradition. Pursuing this latter tack more deliberately now is to understand the relationship between the edition’s presentation and the reported documents (the textual “witnesses” of the work) that it processes. This understanding has become almost unavoidable in recent years with new editions of Hamlet that draw attention to their provisionality or that refuse to finalize a single reading text. If we make the effort to appreciate the editorial problem that the various editions of Hamlet, in their texts and apparatuses, are trying to answer, a more knowing reading of the editions then becomes feasible. And once with Hamlet editions, then with all . . . The editorial problem with Hamlet The problem that all editors of Hamlet face is simply put. There is no manuscript extant. All that editors have to go on are the printed versions. They are conventionally named after their original (folded) sheet sizes: the First Quarto (Q1) of 1603, the Second Quarto (1604–5) and the collected works First Folio (F) of 1623. The text of Q1 is only half as long as the version in Q2, which is itself somewhat longer than F and is generally agreed to be closest to the author’s (lost) manuscript. The latter is traditionally called

110

paul eggert

the “foul papers” – as opposed to the fair copy, conventionally prepared by a scribe and which functioned as the prompt book or playscript. (Doubt has been cast on how robustly these distinctions would actually have applied in Shakespeare’s time; nevertheless, they help to sketch the problem before having to problematize it.) The Elizabethans simply called it “the Booke.” It was usually under the control of the acting company’s book-keeper, and it would have been much simpler in its stage directions than prompt books of today. F is now generally taken to have been based on that prompt book, or on a transcript of it. The text in F seems to be a theatrical text, assembling as much material as possible to afford players some choice when cutting for performance but not playable as it stands.7 The Q1 version can be successfully staged and probably represents – well or badly, but probably the latter – an adaptation or adaptations “latelie Acted by the Lo: Chamberleyn his servants” by July 26, 1602 when the play was entered in the Stationers’ Register. The likelihood is that Q1’s text was memorially reconstructed: that is, prepared for sale to a publisher (Nicholas Ling) by performers in the play. If they remembered their own parts tolerably well, they could cobble together the rest from memory so as to produce a saleable manuscript. Probably only the longer parts would have been copied for the actors originally; the minor roles could be learnt in rehearsal. This was economical and minimized the chance of competitors getting hold of the whole script.8 Many of the great speeches of Q2 and F appear in Q1 only in truncated form, and the poetic quality of the script is generally low. Q1’s Hamlet does not, as a result, cut an especially philosophical figure, so that Q1 becomes a revenge play full of action that drives towards a conclusion. In Act iii scene i, the first nine, magnificent lines of the “To be, or not to be” speech, for instance, are reduced in Q1 to “To be, or not to be, I there’s the point, / To Die, to sleepe, is that all? I all” (TLN 17109 ); and the speech is transposed to the previous scene, a position where it makes good stage sense. When Q1 was published some time after May 19, 1603 (when the Chamberlain’s Men – or servants – became the King’s Men), its title page attributed it to “William Shake-speare” and described it “As it hath beene diuerse times acted by his Highnesse seruants in the Cittie of London: as also in the two Vniversities of Cambridge and Oxford, and else-where.” While no evidence to verify this claim has been found, it is assumed to be true: the claim presumably served as a selling point for the book. Q2 is similarly attributed but is described differently as “Newly imprinted and enlarged to almost as much againe as it was, according to the true and perfect Coppie.” One explanation of Q2’s appearance (from the same printer in the year following Q1’s) is that copies of Q1 were selling, that Shakespeare understandably desired to gain

Apparatus, text, interface

111

some commercial benefit and perhaps also wished to enhance his reputation by supplying the much longer version that he himself had written. On the basis of a great deal of book-trade and other evidence, Lukas Erne argued in 2003 that the Shakespearian plays with longer texts (those over 2,800 lines, including Hamlet Q2 at 3,668 lines) may represent literary versions consciously intended for reading rather than for the stage. In other words, Shakespeare may not have been as careless of his literary reputation as was assumed by most editors in the 1980s and 1990s.10 Erne’s argument about literary versions also applies in principle to F Hamlet as it contains 3,537 lines. And, of course, the luxury First Folio was a consciously memorializing venture – “Published according to the True Originall Copies” – seven years after Shakespeare’s death in 1616. Its theatrical properties are nevertheless plain. F Hamlet has more consistent speech prefixes, more precise stage directions, heavier punctuation and more capitalization than Q2, and F incorporates a partial attempt to introduce act and scene divisions. Shakespeare would probably not have given thought to these formal divisions when writing Hamlet. Scene divisions corresponded to a cleared stage anyway; and act divisions were not adopted by the King’s Men until 1608. While some of this perfecting probably reflects revision for the theatre (thus, incidentally, suggesting that F’s text came after Q2), it is impossible to know how much of it was introduced by the publishers of F for their collectors’ edition, that is, for reading purposes. F Hamlet’s theatrical function is more strongly suggested by the fact that the text of Q1 (which is almost certainly theatrical) and the text of F are more closely related than those of Q1 and Q2. Q1 and F both omit material present in Q2; some of it is repetitive and is deleted with judgment and precision.11 In addition, there are examples of Q1 apparently adapting while misremembering the reading in Q2, before F goes on to resolve or further adapt it. In Act i scene ii, Q2’s “O that this too too sallied flesh would melt, / Thaw and resolue it selfe into a dewe” becomes in Q1: “O that this too much grieu’d and sallied flesh / Would melt to nothing.” Thus Q1 interprets “sallied,” not as a spelling variant for “sullied” as some editors have done, but as meaning assailed or besieged. F decisively clarifies: “Oh that this too too solid Flesh, would melt, / Thaw, and resolue it selfe into a Dew” (TLN 313–14). F does not always clarify in this way, but the example is suggestive. Again, Laertes, who needs the court’s permission to leave Denmark, is described in Q2 by his father as having “wroung from me my slowe leaue / By laboursome petition, and at last / Vpon his will I seald my hard consent” (Act i scene ii; Q2 TLN 240–240+2). Q1 abbreviates to “wrung from me a forced graunt”; and F deletes altogether. Other variants confirm this new intention in F to portray Laertes as more admirably his own man rather

112

paul eggert

than, as here in Q2, supplicant to his father. When Laertes sees the mad Ophelia in Act iv scene v, Q2 has: “O heauens, ist possible a young maids wits / Should be as mortall as a poore mans life.” Q1 mangles this to “I’st possible a yong maides life, / Should be as mortall as an olde mans sawe?” [i.e. a maxim or a commonplace]. F clarifies and extends the speech, thus confirming Laertes’s mature sensitivity to the meaning of the disaster: Oh Heauens, is’t possible, a yong Maids wits, Should be as mortall as an old mans life? Nature is fine in Loue, and where ’tis fine, It sends some precious instance of it selfe After the thing it loues. (TLN 2912–16)

Two other additions in F that have no counterpart in Q1 or Q2 also tend to strengthen Laertes’s character, so that finally, as Paul Werstine argues, he is “more worthy, in his own right, of the respect that he eventually wins from Hamlet in [Act v scene ii] than the Laertes that Q2 depicts.”12 Editorial interfaces respond to the problem What effects do such attempts to clarify the direction of and motivation behind the textual transmission have upon the editing? In the past, the most common approach was to produce a conflated text of Q2 and F, based on one and emended from the other. The aim was to best approximate the text of the original foul papers – that is, Shakespeare’s text. Most of the errors deemed to be scribal or compositorial could be corrected by conflating Q2 and F, and the method ensured, incidentally but crucially, that nothing of Shakespeare’s was left out. As Philip Edwards would comment in his New Cambridge edition of the play in 1985: “it is sadly true that the nearer we get to the stage [via Q1 and F], the further we are getting from Shakespeare,” since this textual journey had necessarily involved Shakespeare’s colleagues in the Chamberlain’s Men.13 Part of the argument revolves around who prepared the transcript of the foul papers for the acting company. Edwards assumes it was a scribe, though with some consultation with Shakespeare. In the Oxford Shakespeare edition (1986), Gary Taylor argues it was Shakespeare himself, partly because of the frequency of changed wordings. It has been calculated that 2,879 of Q2’s words are not “found in substantively the same form and place in F, and 1,889 of F’s words are not found in Q2.”14 In addition, many F readings anticipate or echo other new F readings. Taylor argues that there is no need to postulate, as Edwards does, an excessively interfering scribe or book-keeper to account for the variation. The simplest explanation is that Shakespeare himself made the first theatrical transcription of the foul

Apparatus, text, interface

113

papers, and in doing so also made the deletions and clarifications that most obviously differentiate F from Q2. In this way, the Oxford Shakespeare contrived to marry its dual editorial impulses: to identify cases where Shakespeare revised the text of his plays, creating more than a slightly different work of art – which is as far as Edwards was prepared to go in relation to Hamlet – changes made simultaneously with a theatrical orientation in mind.15 In Chapter 4, Hans Walter Gabler discusses the case of King Lear, where the Oxford editors found the courage to provide two texts rather than the traditional (single) reading text. It was a shock for most readers of Shakespeare’s great tragedy to have the received text multiplied in this way. But with Hamlet, the editors were not so brave, choosing F as what they call their “control text” (i.e. the source of wordings) but emending it rather heavily (293 cases), mainly from Q2.16 They present the extra Q2 passages separately. (The Oxford Shakespeare also provides an old-spelling edition without the usual, modernized spelling and punctuation; its copy-text (for spelling and punctuation only) is Q2 as it was set directly from foul papers, while F remains the “control-text” for wordings.) This reorientation of the editorial endeavor towards the stage did not satisfy some other editors who, in the early 1990s, began to publish Shakespeare texts that were completely unemended and were supposed to be interpreted by the reader on their own evidence, without being editorially construed in advance as versions of one another.17 Rather, they were to be taken as material objects produced for the stage. This went with seeking to rehabilitate the authority of Q1 Hamlet, which, from 1978, had begun to enjoy a revived life on the modern stage. The new, stronger emphasis on the idea of performance sought to supersede the old idealist one (i.e. recovering the text of the work that Shakespeare intended) by claiming to be more firmly historical. “The primacy of performance,” Thomas Clayton could write by 1992, is “so firmly entrenched at present that there is scarcely a black sheep’s bleat to be heard in dissent.”18 However, in 1994 Janette Dillon objected that the new editorial rationale proceeds by demonizing authorship and relying instead on “an abstract and unitary conception of performance as a characteristic of text.”19 This, she observed, was merely to replace one piece of idealizing (authorial intention) with another (performance), when all that we have to go on in either case are the printed texts. Applied to Hamlet, the competing claims of Q1, Q2, and F would, in other words, have to continue to be subject to bibliographical reasoning. Jesus ´ Tronch-P´erez’s critical-synoptic edition of 2002 was innovative; it emerged from a doctoral dissertation of the mid-1990s in which he adapted and advanced Bernice Kliman’s edition “The Enfolded Hamlet” (originally

114

paul eggert

produced as a study tool only but subsequently published in print and online).20 In the edition, he gives the bibliographical evidence, and the existing interpretations of it, a rigorous analysis.21 He lays great stress on what he argues to be the ultimately subjective nature of emendation: that is, the capacity of the evidence from the three printings to be weighed one way or the other to suit an overriding aesthetic judgment on the part of the editor. Instead he presents both Q2 and F, emended only where no sense can be made of the questioned reading, rather than being judged, as conventionally, in relation to one another’s readings. This extreme editorial conservatism is rendered more acceptable by the presentation of the texts in a synoptic form. Here, unmistakably, we see the editorial interface carrying the argument that no final text can responsibly be settled on. Where Q2 and F coincide in wording (which is most of the time), TronchP´erez’s text reads in the normal way. Where they differ, Q2’s text is placed immediately above that of F; and where a line appears only in one of them the other is represented by a row of dots. In contrast to the old spelling in quotations above, note the modernizing here, as well as the Through Line Numbering that facilitates comparison of Q2 and F:22 2912

O heavens, is’t possible, a young maid’s wits,

2913

poor Should be as mortal as a an old man’s life?

2914

. . . . . . . . . . . . . Nature is fine in love, and where ’tis fine,

The appearance of the alternative readings here and elsewhere helps justify Tronch-P´erez’s decision to reduce the level of emendation to near vanishingpoint: there are in total only thirty-five emendations of Q2’s dialogue wordings. The policy here can be seen as an editorial inflection of the spirit of the multicultural 1990s. The intention is to liberate textual diversity – “offering readers as many variations as possible,” Tronch-P´erez states – rather than preferring one and relegating the other to a half-life of neglect in the apparatus.23 He takes this approach one (controversial) step further when dealing with invariant readings shared by Q2 and F that are nevertheless ambiguous, usually because of variable spellings in Shakespeare’s time. For instance, in Act iii scene ii, Hamlet, having just exposed the King’s complicity in his father’s murder via the play within the play, buoyantly claims his right to a partnership with the company of actors. Horatio responds with “Half a share.”; and Hamlet replies identically in both Q2 and F: “A whole one I” (TLN 2151–2). Because “I” could have been intended to mean “ay,” and because this is a modernized edition in which the editor is expected

Apparatus, text, interface

115

to disambiguate old spellings by making a decision about which (modern) word is intended, Tronch-P´erez presents Q2 as reading “I” and F as “ay,” with a note to explain. Thus he creates textual diversity when, in the original texts as printed, there was none. Ann Thompson and Neil Taylor, in their Hamlet edition in the third series of the Arden Shakespeare (2006), turn the bibliographic and other evidence over one more time, with exemplary thoroughness and clarity.24 They describe Tronch-P´erez’s new ideal of textual diversity with respect, but do not follow suit. Neither are they fully won over to Gary Taylor’s arguments about Shakespearian revision underlying F – which they see as very defensible, indeed likely, but not proven. Although they do not fully share Tronch-P´erez’s skepticism about the fundamentally subjective nature of emendation, they confess to having no strong theory of the text’s transmission and certainly not one that would justify thoroughgoing conflation of Q2 and F. In this situation, they offer a multi-texted edition in two volumes. Once again, the edition’s material interface instantiates the textual argument. The main one is based on Q2, sparingly emended from F and occasionally from Q1 (128 emendations to dialogue wording in total). Folioonly passages are presented separately. The second volume contains Q1 and F, edited independently. Variant readings appear at the foot of the reading page in the traditional (“lemmatized”) format. The reading in the text (the “lemma”), based on their control-text Q2, appears before the square bracket and variants (here, from F) follow it, with their source symbol: a poor] an old F

If bibliographic reasoning leads ineluctably to a multiplying of Hamlet’s texts to three in so venerable an enterprise as the Arden Shakespeare – whose first-series Hamlet appeared in 1899 – then, as far as the editors are concerned, so be it. This is the editorial implication of their conclusion that “each of Q1, Q2 and F records a distinct Hamlet.”25 If there is an air of stalemate around their conclusion, it is nevertheless one that is difficult to see around and all the more sobering for the commonsense clarity that the editors exhibit throughout – an empirical approach to their evidence that seems to say that inspiration for editors from the Literary Theory movement of the 1980s and 1990s has finally dried up.

Future Hamlets and the electronic critical archive Both the Tronch-P´erez and Arden editions of Hamlet represent phases in an editorial journey that will never end, since conceptions of the Shakespearian work change from generation to generation. Predicting the future is a

116

paul eggert

hazardous occupation but one thing seems certain: reasoning from bibliography, book history, and stage history will go on generating a spider’s web of delicate inference about the textual transmission of Hamlet’s texts. A mild push in one place causes the whole network of reasoning to spring into a different shape of changed interconnections. Tronch-P´erez’s edition enacts a break with the self-naturalizing transparency of book logic; and the Arden 3 edition only avoids the appearance of a break by tripling the editorial load. Each in its own way points towards and half-demands electronic forms of presentation. Even as each engages in conversation with an editorial tradition stretching back three centuries, a tradition in which the two editions are notable interventions, they break from that tradition at the level of interface. In each case, interface encodes and enacts the editorial position of the editors. In the critical edition, there is no avoiding this fate. The electronic edition, as it more fully emerges, will be subject to the same one, even though editors will, of course, inflect things differently there. Exactly how they will do so is not yet clear. Many editorial experiments have been undertaken since the early 1990s, a period in which the digital environment itself has been rapidly changing. The form that electronic editions will typically take, though it has not yet settled down, will almost certainly have to respect the logic of their new environment. An electronic counterpart of the printed scholarly edition will presumably therefore take advantage of the visualization capacities of the medium to deliver highquality scanned images of, say, the original manuscripts and printings of the work. It will also provide transcriptions of those originals, at least those that were difficult to read on screen. It will preferably offer an automatic method of collating and displaying the variant readings. Unlike the electronic editions of the 1990s and 2000s, it will probably open its files up to collaborative interpretation, including perhaps by readers. If it did, and to the extent that they wished to do so, readers would be transforming their previously passive relationship to the printed book. The interface of the digital edition could, in this case, be described as a “worksite,” with all the materials either immediately available for text-construction or consultation, or able to be connected to. Automatic searches of archival documents (e.g. relevant newspaper reports, images, adaptations) would increase the size of the relevant archive that users were able to work on anew. Thus the electronic edition would not parallel the printed one by consisting of a finalized, unchanging set of files located on an authoritative server: between one set of electronic covers, as it were. Such an electronic edition (perhaps better termed a “critical archive”) would be designed to remain permanently unfinished. This is because it would be committed to recording the successive sets of interaction carried

Apparatus, text, interface

117

out and recorded by its users. Their work would become part of the always unraveling life of the work that is the subject of the edition. To this extent, the editor and the reader would soon recognize themselves to be interveners, appropriators, re-shapers: but always in the name of the work. The work emerges, in this scenario, as no longer an aesthetic object – a marble statue whose delicate but hitherto blurred or damaged lines the conservator–editor cleans and refines, but then walks away from, leaving it for the contemplation of others. Rather, the electronic work-site becomes the locale of active processes, the record of a journey towards never-completed realization. The substantive issues of textual criticism, such as we saw with the Hamlet edition will not subside – by no means. But we will engage with them differently than before.

NOTES

1 2

3 4 5

6

7 8

9

For their helpful and often shrewd comments on earlier versions of this essay, I am grateful to the volume editors, Ann Thompson, Jesus ´ Tronch-P´erez, and my former colleague Katherine Barnes. D. H. Lawrence, Women in Love, ed. David Farmer, Lindeth Vasey, and John Worthen (Cambridge University Press, 1987), 531. These last two paragraphs are adapted from Paul Eggert, “Reading a Critical Edition With the Grain and Against,” in Charles L. Ross and Dennis Jackson, eds., Editing D. H. Lawrence: New Versions of a Modern Author (University of Michigan Press, 1995), 27–40, at 37–8. W. B. Yeats, Last Poems: Manuscript Materials, ed. James Pethica (Cornell University Press, 1997). W. W. Greg, “The Rationale of Copy-Text,” Studies in Bibliography, 3 (1950–1), 19–36. For the idea of book logic, as opposed to the logic of the electronic digital domain, see further, Paul Eggert, “The Book, E-Text and the ‘Work-Site’,” in Text Editing, Print, and the Digital World, ed. Marilyn Deegan and Kathryn Sutherland (Farnham: Ashgate, 2009), 63–82. D. H. Lawrence, The Rainbow, ed. Mark Kinkead-Weekes (Cambridge University Press, 1989), lxiv. For the general shift in interest, see Philip Cohen, “Textual Instability, Literary Studies and Recent Developments in Textual Scholarship” in Texts and Textuality: Textual Instability, Theory and Interpretation, ed. Philip Cohen (New York: Garland, 1997), xi–xxxiv. See Andrew Gurr, “Maximal and Minimal Texts: Shakespeare v. the Globe,” Shakespeare Survey, 52 (1999), 68–87. Trevor Howard-Hill, “‘Nor Stage, nor Stationers Stall can Showe’: The Circulation of Plays in Manuscript in the Early Seventeenth Century,” Book History, 2 (1999), 28–41, at 38 and n. 31. Citations are by Through Line Numbers (TLN), based on F but with additional (+) numbers as necessary for Q2 passages not in F. The TLN also allows Q1’s text to be ascertained, wherever it matches, via the convenient, unemended oldspelling Three-Text Hamlet: Parallel Texts of the First and Second Quartos and

118

10 11 12 13 14 15

16 17

18

19 20

21

22

23 24 25

paul eggert

First Folio, ed. Paul Bertram and Bernice Kliman (Brooklyn, NY: AMS Press, 1991). See also n. 20 below. Lukas Erne, Shakespeare as Literary Dramatist (Cambridge University Press, 2003), chap. 6. See, e.g., Q2 TLN 2455+1 – 2455+5 and 2456+1 – 2456+4. Paul Werstine, “The Textual Mystery of Hamlet,” Shakespeare Quarterly, 39 (1988), 1–26, at 6. Hamlet, Prince of Denmark, ed. Philip Edwards (Cambridge University Press, 1985), 32. Hamlet, ed. Ann Thompson and Neil Taylor (Arden Shakespeare, 2006), 486. Hamlet, ed. Edwards, 30; cf. Gary Taylor, “Hamlet” in William Shakespeare: A Textual Companion, ed. Stanley Wells and Gary Taylor with John Jowett and William Montgomery (Oxford University Press, 1987), 396–420, at 400. The count of emendations here (and below in the Tronch-P´erez and Arden editions) are from Hamlet, ed. Thompson and Taylor, 512. Shakespearean Originals: First Editions, ed. Graham Holderness and Bryan Loughrey (Hemel Hempstead: Harvester Wheatsheaf, 1992): discussed in Hamlet, ed. Thompson and Taylor, 505–6. Thomas Clayton, “Introduction: Hamlet’s Ghost” in The Hamlet First Published (Q1, 1603): Origins, Form, Intertextualities, ed. Clayton (University of Delaware Press, 1992), 32. Janette Dillon, “Is There a Performance in This Text?,” Shakespeare Quarterly, 45 (1994), 74–86, at 77. Shakespeare Newsletter, Extra issue (1996), 1–44; The Enfolded Hamlets: Parallel Texts of and {Q2} Each with Unique Elements Bracketed, ed. Kliman (New York: AMS Press, 2004), also available at http://global-language.com/ enfolded.html, accessed September 21, 2009. A Synoptic Hamlet: A Critical-Synoptic Edition of the Second Quarto and First Folio Texts of Hamlet, ed. Jesus ´ Tronch-P´erez (Universitat de Val`encia; and SEDERI, 2002). See n. 9 above. For a more visually pleasing adaptation of synoptic text to a twoversioned play, see William Akhurst’s pantomime and its anonymous Sydney localization The House that Jack Built in Australian Plays for the Colonial Stage 1834–1899, ed. Richard Fotheringham (University of Queensland Press, 2006), 217–314. Synoptic Hamlet, ed. Tronch-P´erez, 64. Hamlet, ed. Thompson and Taylor; and Hamlet: The Texts of 1603 and 1623, ed. Ann Thompson and Neil Taylor (London: Arden Shakespeare 3rd edn., 2006). Hamlet, ed. Thompson and Taylor, 509.

6 MICHELLE R. WARREN

The politics of textual scholarship

A political view of textual scholarship locates editorial practice and theory within the broadest conceptions of literary and cultural study. To the extent that textual scholarship highlights the constructed nature of all texts, it opens perspectives on the various political forces that traverse those constructions – from historical circumstances to textual content, from intentional manipulation to the more diffuse effects of ideology, from assertions of power to its resistance. Politically oriented textual scholarship can identify with historical and cultural precision the power dynamics (both oppressive and resistant) that sustain language and text systems. Textual scholarship thus has important alliances to forge with any number of politically inflected modes of literary and cultural study – from the postcolonial to the queer, the ecological to the philosophical. The idea of politics in textual scholarship encompasses both the production of texts and their use – especially their use for partisan purposes. In other words, a political view means that textual scholarship belongs as much to its readers as to its creators. In this essay, I write quite self-consciously as one of those readers – as someone who has not formally published an edition but who avidly seeks multiple layers of meaning in the material and formal properties of texts (including editions). My examples of how politics impinges on textual scholarship come primarily (although not exclusively) from the field I know best – European medieval studies. I have organized them to illustrate several different understandings of the “political”: national contexts of professionalized textual scholarship, power relations embedded in editorial forms, and ideological implications of the language used to describe editorial practice. While hardly exhausting the possibilities for politically inflected approaches to textual scholarship, these topics suggest some of the ways in which attention to politics can orient both the creation and reception of edited texts. A fully rigorous account of the politics of textual scholarship should really address definitions of the political. What does editing have to do with community, the polis? What are the ethics of transhistorical and cross-cultural 119

120

michelle r. warren

engagements? How does ideology, variously defined, traverse institutional sites of editorial production? What do different conceptions of power illuminate about textual processes and forms? What sorts of social dynamics surround the disciplinary formations and receptions of textual scholarship? While I cannot delve deeply into these fundamentals here (instead drawing on an eclectic and shifting array of definitions in the spirit of exploration and provocation), the very term “politics” requires serious discussion at every stage of textual scholarship. Disciplinary histories The development of textual scholarship in nineteenth-century Europe coincided with the consolidation of national identities and, in many cases, overseas imperialism. Editorial production also coincided with the increasingly structured division of disciplinary inquiry in universities: edited texts made many other kinds of inquiry possible, from literary criticism to history to linguistics. Editorial methods thus shaped the basic materials for disciplinary development. Arguments over method, however, owed a great deal to national and imperial rivalries. Textual scholarship thus participated actively in the broader politics of scholarship per se, both on the diffuse level of “power relations” as described by Michel Foucault (The Order of Things; The Archeology of Knowledge) and on the pragmatic level of the strategic deployment of texts to support social authority. The very development of editorial methods occurred within complex political frames, often leading to identifiably national schools of critical method (each with its own defining figures). In many national contexts in nineteenthcentury Europe, the medieval epic became a privileged target of editorial politics, since the genre was understood to legitimate the ethnic identities of modern nation-states. Thus in Britain, politically inflected scholarship developed around Beowulf, while in Spain the Cid inspired analogous attentions. Some of the most overtly political disputes took place between French and German scholars around the Song of Roland. Before the rediscovery and publication of the poem in 1837, France had long been judged to lack the “epic spirit” considered fundamental to a fully developed national consciousness. The Roland was thus immediately hailed in France as a national poem, its publication a triumph over both German culture and German editorial control. If the publication of Roland could prove that France had produced an epic to rival those of other countries, its popular diffusion could contribute to a profound re-imagining of the national character.1 A close look at the development of textual scholarship around the Roland illustrates well the politics of editing. For Roland could not legitimate French national origins without some inventive textual scholarship: the

The politics of textual scholarship

121

oldest manuscript resided not in France but in Oxford (Bodleian Library, Digby 23), it was written in an Anglo-Norman dialect (that is, not directly connected to Continental France); the poem’s heroic Franks could function as the ethnic predecessors of the Germans as well as the French (depending on how one dated the poem). These problems concerned French scholars from the beginning, since they were acutely aware that German scholars were publishing editions of “French” texts at a faster rate and with more “rigorous” methods than the French themselves. The French recovery of Roland, however, took on greater urgency after the successful Prussian invasion of France in 1870. Already during the siege of Paris in December 1870, Roland became a lightning rod for patriotic reflection: in the space of six days, two professors opened their courses with lectures on the poem, and another signed his introduction to a new edition (calling urgently for the “repatriation” of the Oxford manuscript). Techniques of textual scholarship were soon directly implicated in both the diagnosis of weaknesses that led to France’s defeat and the prognosis for future strength. Several scholars, for example, linked German superiority on the battle field with German superiority in editing. And for the politically sensitive Roland, the lack of a French-produced edition of high scholarly caliber represented something of a national embarrassment. Even though the first post-war edition appeared in 1872, followed soon by a number of highly portable popular editions and translations, scholars relied primarily on German editions (Theodor Muller 1851, 1863, 1878; Eduard ¨ Bohmer 1872; Edmund Stengel 1878). In 1894, Joseph B´edier declared it ¨ “humiliating” that Germans were recovering French literature before the French; again in 1913, he complained that it was “deplorable” to have to read the Roland in “detestable” German editions. In terms of textual scholarship, then, France needed not only a Roland produced by a Frenchman but one produced according to methods both “scientific” and properly “French.” A close comparison of the edition that B´edier himself ultimately produced in 1922 with those of his predecessor L´eon Gautier illustrates some of the very detailed ways in which politics can inform textual scholarship. Gautier edited Roland almost incessantly from the 1860s until his death in 1897. Largely as a result of his efforts, the poem became obligatory reading in French secondary schools in 1880 (for girls in 1882); the first translations for young children appeared in 1885. One of the early recipients of Gautier’s epic patriotism was B´edier: in 1878, as a youth living in France’s Indian Ocean colony, R´eunion Island, B´edier received a copy of one of Gautier’s editions as a school prize. While B´edier mentions this fact in non-scholarly contexts, his scholarship reveals a concerted effort to disavow the influence of both Gautier and the German editorial methods that

122

michelle r. warren

Gautier ultimately embraced. This disavowal derives from the nationalist politics of textual scholarship as well as from the imperial politics of colonial memory. The classification of Roland manuscripts lies at the center of these political conflicts. The German editions that B´edier so reviled followed the stemmatic method of critical editing first practiced by Karl Lachmann. Intended to reconstruct the “original” text by comparing versions, Lachmann’s method reflected the political value of linear national development. For Roland, the German editors Wendelin Foerster and Edmund Stengel gave the Oxford manuscript a “brother” and granted equal weight to two other groups of manuscripts; together, the three groups provided evidence for the epic’s oldest form. Citing Foerster, Gautier abandoned his initial privileging of the Oxford manuscript. As of 1880, he followed the Germans in giving equal weight to three manuscript families, producing the composite text that shaped students’ experience of Roland for generations. Displacing this tradition, B´edier in the 1920s argued in favor of the Oxford manuscript’s unique, dignified authority. In lieu of Gautier and the Germans, B´edier made Francisque Michel, who first transcribed the Oxford text in 1837, the heroic originator of a purely French tradition of textual scholarship devoted to editing the one “best manuscript” rather than critically combining different sources. B´edier’s version of editorial history, however, simplifies rather tendentiously the interactions between French and German practices. B´edier, for example, overlooks Gautier’s early editorial theory and practice, which largely resemble B´edier’s own. And even Stengel, the champion of the composite method, once edited the Oxford manuscript – in diplomatic transcription no less (1878). B´edier’s focus on Michel sidelines another German editor, Theodor Muller, whose 1851 edition is actually based on Michel’s of ¨ 1837 – and whom Gautier followed in his editorial work of the 1860s and 1870s. B´edier’s editorial history, in other words, excises both Muller and ¨ Gautier in favor of Michel. By overlooking the formative role of a Gautier edition that followed Muller, B´edier purifies his own editorial genealogy, ¨ reconnecting with an origin prior to the attentions of German philology. This nationalist argument simultaneously elides and recovers colonial memory: B´edier ignores the Gautier edition he read on R´eunion, while producing a text actually quite similar to Gautier’s. The dialect of the Oxford manuscript posed another kind of political problem: how could a poem in Anglo-Norman represent French national origins? Nineteenth-century editors addressed this embarrassment according to their respective theories of origin: Gaston Paris amended the text to reflect the Francien dialect, while Gautier sought to restore the “purity” of the Norman dialect (liberating it from the “vices” and “dishonor” of

The politics of textual scholarship

123

Anglo-Norman “dust”). B´edier rejected these kinds of interventions while remaining as committed as Gautier and Paris to the Roland’s French origins. He “purifies” the poem’s seemingly contaminated form in several ways. First, he argues, Roland’s basic unity derives from the singularity of its original French author. The poem’s universality and “classical” style (akin to Racine’s, he says) transcend local particularities: the manuscript’s graphic forms do not in themselves orient the poem toward Britain. B´edier concedes that the manuscript offers a “pure” representation of the French spoken in England around 1170. This language, however, shares a number of dialectical features with the Continental “French” of 1100, and so does not actually differ very much from the poem’s original language. B´edier reasons further that the poet himself could have used inconsistent forms: since we do not know what rules he followed, we cannot make the copyist responsible for everything that looks like an error; correcting anything could change the poet’s own words. In effect, B´edier argues that the manuscript’s AngloNorman may as well be French, for all anyone knows about the nature of French in 1100. While scrupulously preserving the “foreign” dialect, B´edier renders it utterly meaningless. His edition thereby claims a double continuity – simultaneously faithful to the lost original poem (in French) and the surviving manuscript (in Anglo-Norman). B´edier’s editorial arguments reinforced the national recuperation of Roland begun by his predecessors who had lived through the FrancoPrussian war of 1870. After the more recent war of 1914–18, B´edier’s Roland reanimated the longstanding political dynamics of nationalist philology. He gives the politics of textual recovery a decidedly colonial twist when he dedicates his completed edition to his island homeland, “Bourbon Island” (R´eunion). B´edier’s methodological conclusion – to edit from a single manuscript considered the “best,” with a minimum of intervention – became a new disciplinary norm in French textual scholarship. In its requirement to prioritize a single source over all others, “B´edierism” subtly sustains colonial ideology. Stemmatic reconstructions also rely on hierarchies that result in one “true” text. This similarity, though, was often obscured by the nationalist politics of editorial practitioners. B´edier’s Roland oriented scholarly and popular understanding of the epic for much of the twentieth century. The edition remained widely available into the 1980s, serving as a durable reference despite the many other editions produced since 1922. Even when students and scholars have used other editions, they have felt B´edier’s influence: most editors have focused exclusively on the Oxford manuscript, with ever greater respect for its exact forms. Arguably, even Cesare Segre’s influential revival of the stemmatic approach to the manuscript corpus remains fundamentally B´edierist: Segre relies on the Muller stemma that B´edier defended, changing the Oxford text ¨

124

michelle r. warren

much less than the stemmatic method authorizes; in his second edition he relegates interpolated passages to the notes.2 The prevalence of B´edier’s ideas in Roland’s twentieth-century editorial history reveals the ways in which his unspoken political motivations have shaped textual scholarship. B´edier’s theories of authorship and language support not only the French nation, but general conceptions of cohesion that underlie most forms of national belonging. In B´edier’s case, these conceptions owe a great deal to colonial experience and implicit theories of racial identity. Today, the Oxford Roland seems somewhat liberated from the strictures of B´edier’s conclusions. In France, the French translation of Segre’s critical edition has become the new authority (required for French national exams in 2004); students not trained in Old French are likely to read Ian Short’s version in the Lettres gothiques series, which introduces numerous changes to the Oxford text in an effort to account for scribal knowledge of oral composition; scholars now have access to the full corpus of French-language Roland manuscripts through the new editions overseen by Joseph Duggan. Duggan’s project dramatically shifts editorial focus from questions of origins and authorship to the dynamics of reception and rewriting: the new editions encourage attention to the Oxford manuscript’s twelfth-century context, to the significance of manuscript artifacts themselves, and to later medieval thinking on the renewable significance of Roland’s story. New politics, of course, shadow this project – published in English by the Belgian press Brepols, edited entirely by scholars working outside of France. The internationalization of textual scholarship in this case signals a shift, but not an end, to editorial politics. Textual politics The politics of editorial method that surround Roland are closely bound up with the politics of textual selection and dissemination: since the epic genre served a broader function as a literary witness to national origins, the production of editions responded to politically inflected desires while their distribution aimed to foster those desires in new readers. In terms of textual selection, politics conditions the very materials that draw the attentions of individual editors and the institutions in which they work. Preferences for particular textual sources may also break down along national, and class, lines (with materials considered elite receiving greater attention in one context while those of more modest origins are prioritized in another). In terms of dissemination, editions’ material forms encode power relations and political aspirations of varying degrees of explicitness. Selection and dissemination thus signify within broader cultural dynamics, embedding textual politics within a general sociology of textuality.

The politics of textual scholarship

125

The editing of ideologically charged material can have profound consequences for practical politics. Any project involving the legacy of political figures encounters editorial questions with direct political consequences. The editing of colonial histories illustrates these consequences clearly: various interpretative fallacies have arisen from tendentious and selective editions produced from the manuscripts associated with Christopher Columbus, manuscripts that are themselves filled with heavily politicized marginal commentary.3 Approaches to editing obscenities, or other kinds of “objectionable” content, also raise questions about the politics of speech itself. The very availability of certain texts, and not others, conditions the formation of textual canons that in turn have decisive impacts on education, popular culture, and scholarship. Textbooks and other educational materials are also the products of editing and textual scholarship. Everything from the selections to the framing narratives to the introductory headnotes of anthologies bespeaks both political design and ideological effect. Politics here encompasses both the direct politicization of educational content (the ways that textbooks frame colonialism, for example) to the broader politics of knowledge formation itself (the ways that any anthology marks limits in a given field).4 Indeed, the current phenomenon of subscription-based textual databases entails a number of legal and economic issues that bear political scrutiny. Material forms also structure a number of different relations of “power” – between editor and text, edited text and source materials, edited text and readers, different parts of the edition itself. Layouts, for example, can emphasize similarities among texts (uniform shapes for disparate sources, familiar modern forms for ancient sources) or differences. Practices around annotation establish more direct networks of power over and around textuality, marking both authority and resistance, laws and their transgression, completion and its failure.5 The proportionality of “apparatus” to “text” can physically signal the power plays of editorial practice: when the apparatus far exceeds the text, it signals a kind of interpretative exhaustion even before the reader reaches the text itself; when reduced to a bare minimum or even eliminated, the silence signals the text’s self-sufficiency, naturalizing editorial intervention. The very “blankness” of a page devoid of notes or other apparatus suggests a transparent autonomy associated with disembodied authority. Even the seemingly open-ended structures of digital editions encode hierarchies in the architecture of both software and hardware, insinuating power relations into both the production and reception of electronic forms. (Many of these issues have been explored at length in the critical work of writers such as David Greetham, John Bryant, and Jerome McGann.) The ways in which form can attest to politics suggests that “old” and “inadequate” editions may have important stories to tell as witnesses to

126

michelle r. warren

histories of power relations. Early Spanish poetry, for example, became “monolingual” through distinctly ideological editorial interventions, which suppressed the presence of Hebrew and Arabic. In Britain, in the nineteenth century, editions of medieval literature produced by the Early English Text Society circulated the ideology of “English culture” in British India; colonial pedagogy included a specific role for compilers of “useful” works. Meanwhile, the production of the Monumenta Germaniae Historica and other philological efforts in Germany responded to a myriad of shifting and contradictory political ideals.6 And in France, the epic Fierabras was conceived by its translator, Jean-Bernard Mary-Lafon, as a literal aid to North African colonization: Why wouldn’t we show in this chivalric epic the great figure of Charlemagne to the Arabs of Africa, whose hearts would leap with joy at the narrative of the great blows given by Fierabras, and who would probably see, with their fatalist belief, a premature decree of God and the finger of Allah in the submission and baptism of the most brilliant of their heros? . . . In this fact, so strangely remarkable, of the Fierabras of legend kneeling down at Charlemagne’s feet and of the Abd-el-Kader of history kneeling before Napoleon, isn’t there enough to strike imaginations less impressionable than those of the Arabs?

Mary-Lafon concludes with the hope that his French translation be translated into Arabic, “to repeat in the Orient and in Africa, under the tent and under the kiosk, the heroic epics of our fathers.” Mary-Lafon’s translation, along with the one he imagines, witness the agency of textual scholarship in colonial ambition. Mary-Lafon’s own resistance to the ideology of French unity, due to his defence of Provenc¸al autonomy, places his expansionist rhetoric within an especially complex logic of internal and external colonialism. And since a new edition of Fierabras wasn’t published until 2003, nationalist and colonialist pasts have probably been circulating in scholarship in various unexamined ways. Even “unreliable” editions, in other words, have important stories to tell about ideological appropriations and political histories. Material forms also lend themselves to eco-political interpretation. What happens to the aesthetics of the medieval manuscript, for example, when we think of the object – made of animal skins – from the perspective of the politics of animal rights?7 The contrast between a “worthless” carcass and a “priceless” ancient book can provoke critical thinking on the intersections of economics and aesthetics. On what grounds, in turn, are collections of such animal objects superseded by printed editions, or by electronic bits whose processing and storage depend on the mining of irreplaceable precious metals (often from lands formerly subject to European colonialism)? Similar

The politics of textual scholarship

127

interrogations can be brought to bear on relations between editions and any number of materials – from birch bark to acid paper to stone. We can also ask pro-actively political questions about editing. Feminist editing, for example, includes a focus on the gendered inflections of textual cruxes, the “recovery” of texts overlooked in patriarchal canon formation, and efforts to break with patriarchal methods. Rather than reproduce masculine notions of authorship and audience, how can editors favor “fluidity” and challenge the numerous “essentialisms” that have often informed editorial thought? How can they avoid complicity with misogynist materials through the rhetoric and content of annotation?8 On one level, feminist approaches commit to the radically provisional, resisting closure and the finality of definitive authority. What happens when editors explicitly “politicize” their approach in this way? What are the implicit politics of those who don’t? How might commitments to postcolonial or gender theory shape editorial practice? What might the categories of “colonial mimicry” or “queer” do to conceptions of relations between editions and the materials that inspire them? Both of these categories unsettle the traditionally normative ethos of editorial ideology. Explicitly politicized approaches ultimately underscore the power plays that inhere in every project of editorial appropriation, whatever its form, content, or method. Editorial discourse The very ways in which the makers and users of editions talk about the editorial process and its products also inflect relations of power. Many different metaphors conceptualize textual scholarship in terms that both imply and exert ideological force. Every descriptive term carries prescriptive potential; whether we speak of “corruption” or “variance,” diplomatic “neutrality” or critical “intervention,” “readability” and “clarity” or the “clutter” of zealous erudition, we find ourselves entangled in a plethora of connotations that condition our textual experiences. The very bitterness with which editors criticize each others’ methods testifies not only to the high stakes of editorial action but to an oppositional structure of scholarship that has its own consequences. Questions of method are often presented in a zero-sum game in which the truth-value of one method is predicated on demonstrating the worthlessness of another.9 Progressivists rhetorics of openness, accessibility, and collaboration operate just as politically. Institutional contexts loom large in the politics of method, as professional organizations, publishers, universities, archives, and governments all set some of the terms of usefulness. In seeking to understand definitions of editorial success and viability, it can be useful to think of both editorial projects and methodological arguments

128

michelle r. warren

as “plots” – narratives that follow certain prescribed structures, much like the historiographical plots identified by Hayden White in Metahistory (e.g. romance, comedy, tragedy, satire). Do happy endings prevail? How do the modes of metonymy, metaphor, and irony structure relations among texts? How do seemingly plot-resistant terms like “browsing” map out divergent but nonetheless prescribed experiences? In articulating the editorial plots that we perceive, and in consciously recognizing the emplotment of descriptive narratives, we can account more fully for the varied stakes of textual scholarship. One of the more powerful structuring plots of textual scholarship is evolutionary biology, partly a legacy of nineteenth-century institutional structures but also a sign of the enduring explanatory claims of the biological sciences (from the Human Genome Project to the Innocence Project). Scientists themselves posit biology as a model for all of the humanities. In textual scholarship, metaphors of descent (phylogeny), taxonomy, botany (rhizome), and genetics “naturalize” textual relations; ideas about a given individual’s “likely” actions in a given textual situation rely on psychological and biological assumptions. These metaphors cast textual scholarship as “scientific” rather than “humanistic.” The maintenance of this divide, Marjorie Garber has shown, relies on normative ideas about “human nature.” What happens to the politics of textual scholarship if, following Garber’s suggestion, we treat “human nature” as a question rather than an answer?10 To the extent that textual scholarship relies on the science of human nature, it remains caught in an interdisciplinary feedback loop, in which science has claimed literary texts as the proof of human nature and the methods for publishing those texts have drawn on scientific models. Garber’s examples of how scientists cite Shakespeare typify the problem, since Shakespeare also serves an iconic function in Anglophone textual scholarship (due to the cultural stakes attached to his name and works): while scientific literature ignores the complexity of Shakespeare’s texts, textual scholarship often ignores the complexity of biology. In this loop, mutually reinforcing fantasies of “human nature” both expose and hide the politics of metaphor in disciplinary reasoning. Scientific discourse also underwrites the conceptualization of “progress” in textual scholarship – the desirable path to a “definitive” edition of the “best” materials using the highest “standards.” The quest for definitive forms encourages editors, publishers, and readers alike to overlook the constructed nature of their textual models – to discard everything that precedes or exceeds the clear printed text. We should therefore be wary of the “myth of progress” by which a “good” edition renders its predecessors obsolete. Similarly, the idea of selecting the “best” manuscript for editing implies hierarchies of value that imbue certain phenomena with greater authority than others. Editions themselves are often assigned lesser value than the

The politics of textual scholarship

129

“original” and “unique” materials that inspire them. The politics of singularity opposes multiplicity and contingency. As William Robins points out, privileging the “best” often focuses on “central” rather than “peripheral” witnesses to a given textual tradition – a structural orientation that works metaphorically on a number of a different political levels. Robins explores the effects of constructing “worst” text editions: what do we learn from the “periphery” of the corpus? This approach has the potential not only to overturn traditional hierarchies but to expose the ideological underpinnings of hierarchy itself. The sciences have provided attractive metaphors for textual scholarship precisely because they imply legitimate authority and rigorous technique. Expertise itself, then, functions as a master trope in disciplinary politics. Critical interpretation often depends on repressed power relations with experts and the products of their expertise. What is left to the “experts” can easily be objectified as admirable or irrelevant. The institutional consecration of expertise can also affect editorial method, contributing for example to the privileging of “difficult” vocabulary over quotidian usages.11 Textual scholarship of course cannot do without method, training, and practice. Nonetheless, generalized awareness of the implications of “consuming” the products of editorial scholarship can facilitate an analytic engagement with those same materials, making their construction and form as much an object of analysis as their seemingly neutral “content.” Textual scholarship itself can endeavor to account for the powers of technique. A willingness to proceed sometimes as “amateurs,” rather than as experts, might open some surprising new possibilities. By considering authoritative expertise in theoretical and practical terms, critics of all kinds (“textual,” “interpretative,” “cultural”) can find ways to overcome the divisions of labor that too often constrict our view. Labor itself, in quite material terms, constitutes an important aspect of the politics of textual scholarship. Approaches to textual scholarship attendant to class or historical materialism can include analysis of both the cultural work of editions within ideological formations and editorial labor itself within the academic market system. In class terms, editing has sometimes boasted of a solid middle-class usefulness of higher value than literary criticism (sometimes cast as the dilettant privilege of the leisured classes).12 Conversely, editorial labor can look like the textual “proletariat” – a means of production upon which rests the “superstructure” of literary criticism. Sophisticated Marxist cultural analysis, of course, underscores the recursive mutual actions of ideas upon materials and vice versa, opening possibilities for new ways of conceiving the interdependence of editions, their readers, their sources, and their makers. How do divisions of labor parallel other kinds of differences?

130

michelle r. warren

One politicized difference that has received significant attention in textual scholarship is gender. When conceived as a process of “engendering,” editing implicitly casts its objects in erotic terms, pointing toward the polymorphous circulation of desire throughout the editorial process. The traditional stemma, for example, follows a parthenogenetic model, in which the legitimate text resists “relations” of any kind: seen as a sexual system, the stemma defines the “other sex” as a contaminating intruder.13 The stemma itself, though, can sustain other metaphors. B´edier, for example, mocked the stemmatic quest for family relations by imagining that one of the “brothers” in the group donned make-up.14 In today’s terms, B´edier suggests here that “gendered performance” rather than sexual biology lies at the heart of stemmatic logic: in mistaking gender for sex, we assume essential differences that may in fact turn out to be provisional variables; surface differences between textual witnesses (which could lead to classifying them in different “families”) may merely mask deeper similarities (justifying classification in the same family). What happens to the textual scene if we take the idea of a “transvestite” text seriously? B´edier in fact posited the transvestite as the norm rather than the exception: precisely because any text could be in drag, the logic of the stemma falls apart. Current theoretical reflections on crossdressing could help elaborate on how imitation and “passing” can operate in textual scholarship. These same reflections prompt reconceptualization of the idea of the body. Analogies have often been drawn between the human body and the physical text. But the metaphor shifts significantly with the image of the “cyborg” body posited by Donna Haraway: the cyborg challenges naturalized genealogies of (textual) transmission from generation to generation, underscoring the body’s construction through purposeful interventions. Current thinking around what Haraway terms the “posthuman” follows on the cyborg precedent, intervening between gender and biology to disrupt prevailing models of naturalized human bodies.15 Politicized interventions in textual scholarship cannot help but take place in the form of new metaphors. I would like to close with one inspired by Mireille Rosello, who has identified a “stuttering” narrative structure as a metaphor for “queer editing.” In analyzing a Qu´ebecois novel (Daniel Sernine’s Chronoreg), Rosello argues that when a particular sentence repeats, narrating an event in a slightly different form, the repetition performs a narrative stutter that creates queer time: discontinuous, non-linear, never complete (“coming-out” never ends). Rosello’s treatment of the stutter describes textual scholarship marvelously well: “The stutterer is both repeating and not repeating, because no repetition is complete, and because no statement is unique or ever completed;” “no real ‘truth’ ever replaces an old ‘lie;’”

The politics of textual scholarship

131

“[e]ach new version both includes the earlier ones and moves away from them, refusing to become the absolute reality of the narrative, building on the past while suggesting that it is always possible to go back, to have to start from scratch again.”16 Editing turns the editor into a stutterer – repeating but not completing a text that, in the first instance, was also not entirely complete. By creating the second (or third or fourth or . . . ) repetition, the edited text marks the source material as incomplete. Relations between originals and copies become blurry, as editions take the place of their sources. The edition becomes a move both backward (repetition) and forward (toward completion) – a “stuttered” utterance, following and preceding other stutters. The graphic representation of stuttering incompletion comes in the format of Chronoreg’s final pages – double columns familiar to editors and their readers: “The last pages of the text are split in two separate and schizoid columns that follow two different directions without giving us any indication that one version is more plausible than the other.”17 Chronoreg here enacts its own parallel-text edition of two sources, a double column stutter that remains unresolved. The point of attending to metaphors and discourse in textual scholarship, of course, is not to eliminate metaphors or the ideologies they engage. As Edward Said notes in Orientalism, the exposure of orientalist discourse does nothing to diminish its powers to organize knowledge and structure social relations. Awareness of how discourse works can lead to cynicism (damaging repetition can seem inevitable), but it can also incite intervention at the level of rhetoric itself. Discourse analysis brings ideology into the editorial conversation, embedding politics within even the seemingly “objective” components of editorial activity. Discourse analysis thus underscores the pervasive action of interpretation – the lack of any pre-hermeneutic moment in our textual relations. Conclusion Editions shape the conditions of possibility for interpretative engagements. Editions are thus diffusely “political” regardless of their content or form simply because they disseminate certain forms of knowledge and not others. As Said observed in Orientalism and elsewhere, textual scholarship actively shapes discourse – and the idea that “true” knowledge resides outside of politics in fact obscures the pervasive roles of politics in knowledge production. Said’s examples illustrate powerfully the myriad politics of editing, translating, and generally transmuting knowledge from one form to another. Political approaches to textual scholarship challenge us to account for the implicit and explicit politics that inform editorial valuations, decisions, and

132

michelle r. warren

products. Politics also challenge us to recognize the elusive presence of missing archives – pieces of textual and human history rendered invisible by the very techniques used to create textual scholarship. NOTES 1 Michelle R. Warren, Creole Medievalism: Colonial France and Joseph B´edier’s Middle Ages (University of Minnesota Press, 2011), 1–25, 150–63 (source of the examples in this section). 2 La Chanson de Roland, ed. Cesare Segre (Milan: Riccardo Ricciardi, 1971); 2nd edn. (Geneva: Droz, 1989). 3 This example and others in this section from Warren, “Post-Philology,” in Postcolonial Moves: Medieval Through Modern, ed. Patricia Clare Ingham and Michelle R. Warren (New York: Palgrave Macmillan, 2003), 19–45. 4 On Anthologies: Politics and Pedagogy, ed. Jeffrey R. Di Leo (University of Nebraska Press, 2004). 5 Annotation and Its Texts, ed. Stephen A. Barney (New York: Oxford University Press, 1991); Anthony Grafton, The Footnote: A Curious History (Cambridge, MA: Harvard University Press, 1997). 6 Kathleen Biddick, The Shock of Medievalism (Durham, NC: Duke University Press, 1998), 92–6; David Townsend, “Alcuin’s Willibord, Wilhelm Levison, and the MGH,” in The Politics of Editing Medieval Texts, ed. Roberta Frank (New York: AMS Press, 1993), 107–30. 7 Bruce Holsinger, “Of Pigs and Parchment: Medieval Studies and the Coming of the Animal,” PMLA 124 (2009), 616–23. 8 Betty T. Bennett, “Feminism and Editing Mary Wollstonecraft Shelley: The Editor And?/Or? the Text,” in Palimpsest: Editorial Theory in the Humanities, ed. George Bornstein and Ralph G. Williams (Ann Arbor, MI: University of Michigan Press, 1993), 67–96; Valerie Wayne, “The Sexual Politics of Textual Transmission,” in Textual Formations and Reformations, ed. Laurie E. Maguire and Thomas L. Berger (University of Delaware Press, 1998), 179–210. 9 William Robins, “Toward a Disjunctive Philology,” The Book Unbound: Editing and Reading Medieval Manuscripts and Texts, ed. Sian ˆ Echard and Stephen Partridge (University of Toronto Press, 2004), 144–58. 10 Marjorie Garber, “Who Owns ‘Human Nature?’” Quotation Marks (London: Routledge, 2003). 11 Bill Readings and Bennet Schaber, Postmodernism Across the Ages: Essays For a Postmodernity That Wasn’t Born Yesterday (Syracuse, NY: Syracuse University Press, 1993), 19–23; David F. Hult, “Reading It Right: The Ideology of Text Editing,” in The New Medievalism, ed. Marina S. Brownlee, Kevin Brownlee, and Stephen G. Nichols (Baltimore: The Johns Hopkins Press, 1991), 121. 12 Alan Roper, cited in Gerald Maclean, “What’s Class Got to Do with It?” in The Margins of the Text, ed. D. C. Greetham (Ann Arbor, MI: University of Michigan Press, 1997), 31. 13 Mary Ann Caws, “The Conception of Engendering the Erotics of Editing,” in The Poetics of Gender, ed. Nancy K. Miller (New York: Columbia University Press, 1986), 42–62; D. C. Greetham, “Phylum-Tree-Rhizome,” in Reading from the Margins: Textual Studies, Chaucer, and Medieval Literature, ed. Seth Lerer (San Marino, CA: Huntington Library, 1996), 100–2.

The politics of textual scholarship

133

14 B´edier, “La tradition manuscrite du Lai de l’ombre: r´eflexions sur l’art d’´editer les anciens textes,” Romania, 54 (1928), 161–96, 321–56, at 339. 15 Joseph Grigely, “Editing Bodies,” in Reimagining Textuality: Textual Studies in the Late Age of Print, ed. Elizabeth Bergmann Loizeaux and Neil Fraistat (University of Wisconsin Press, 2002), 60–84; Donna J. Haraway, “A Manifesto for Cyborgs: Science, Technology, and Socialist Feminism in the 1980s,” Socialist Review, 80 (1985), 65–108; and Haraway, When Species Meet (University of Minnesota Press, 2007). 16 Mireille Rosello, “The National-Sexual: From the Fear of Ghettos to the Banalization of Queer Practices,” in Dominique D. Fisher and Lawrence R. Schehr, eds., Articulations of Difference: Gender Studies and Writing in French (Stanford University Press, 1997), 246–71, at 268–9. 17 Ibid., 269.

7 RANDOM CLOUD

Fearful Asymmetry for Jerome McGann at 70 on the upper or lower outer corners of the leaves of many eighteenth-century books, raking light reveals systematic dog-earing. Opposite, on pp. 238 and 240 of the University of Toronto’s Fisher Library copy 1 of Richard Bentley’s quarto edition of Paradise Lost, 1732, you can barely discern these creases now the ears stand upright.1 Dog-ears can be detected here in 51 of 222 leaves. The 75 such occurrences in the Folger Library copy are more obviously methodical, occurring in two consecutive quires, but not the third: in C & D, but not E (here the rhythm begins); then in F & G, but not H; and so on, almost regularly, through the penultimate 3H (where the Index quires are merely half-sheets).

. . .

C F

D G

E H

Above left, I depict unopened the upper folds (or bolts) in C & D, two full-sheet Folger quires. Their four consecutive dog-ears bend at a single angle. As the angle of the corresponding fold in the next pair, F & G, is different, each pair must have been folded as a unit – the last two bolted leaves in C with the first two in D, the last two in F with the first two in G – finally to 3F & 3G (where only single leaves pertain). Such dog-earing was a preparation for sewing, after which binders would have ploughed off the bolts in a single operation or left the task to the first reader, who, knife in hand, sliced them open one by one.2 The next two openings of this essay display all the quires of the Folger textblock with bolts still intact; but the wings are open now, to reveal sewing in the central gutters. A strict correlation pertains between the presence or absence of dog-ears and the configurations of (usually) two stitches (among six possible). I’ve numbered stitches in the order sewn – up ↑the gutter, in quires A (stitches 6, 4, & 2) and a (5, 3, & 1), or down ↓, in quires b (1 & 5) and B (2, 3, 4, & 6).3 135

136

random cloud

Α

a

C

D

E

I

K

L

P

Q

R

U

X

Y

Fearful Asymmetry

137

b

B

F

G

H

M

N

O

S

T

Z

138

random cloud

Fearful Asymmetry

139

140

random cloud

dog-eared: aft

stitch

fore

aft

fore

cord

station

mm

a b c d e

C

D

E

F

G

H

With its systematic symmetries and asymmetries, the above diagram condenses information from Row 2 on pp. 136−7, a typical six-quire sewing unit. The vertical lines represent the central gutters of quires C–H, identified below by signature. The tilted tops of C, D, F, and G signal which of them have dog-ears (aft or fore, as labelled). The far-left column counts six stitches, the near right seven sewing stations. Dimensions in millimetres appear on the far right. Lettered on the near left are five cords, invisible to readers (but detected as bulges on the spine); three of these cords (a, c, and e) lace into the covers. As we’ll soon see in detail, cords dictate the locations of all sewing stations except the first and last: there, kettle stitches link one quire to another; at stations 2–6, however, the thread binds the quires by circling the cords.4 In the cradle days of printing, sewing normally linked all stations in a quire: hence, six stitches per quire for a five-cord binding like this. But by the eighteenth century, sewing quires three on was common: in the Folger volume, only one stitch was sewn initially in each of the first two quires, C and D; then came two consecutive stitches in E; and finally the second and final stitch in each of the first two quires – all three gutters finally exhibiting merely two stitches each, not six. (Be patient: my next two openings will walk you step by step through the sewing of these quires three on.) It compromised strength, but this corner-cutting required only a third of the thread and a third of the binder’s time. It also reduced the swell caused by thread and folds, which threaten to make the spine of a book wider than its fore-edge.

Fearful Asymmetry

141

Having dragged you from an outer corner of the leaf down into the gutter, I owe you an explanation. The obscure evidence here in the underworld of Paradise Lost is textile. Now to understand its textual implications. It is no accident that these words sound alike: both derive from the Italic for “weaving” and have Hellenic and Germanic cognates (like “technique”, “technology”, and “thatch”). The Indo-European root also blossoms in Hittite and Indic, all derivations pertaining to fashioning, fabrication, construction, architecture – in short, to artifice. In the wide-spreading etymology, we sense the fecundity of cognate associations – a sisterhood of arts. This essay will address what is deracinated in our understanding of the codex and seek to explore and integrate its complexities. “Complexities” – the root of this Italic word is cognate with Germanic “fold” – and with folding we began. I am referring not only to the objective textile, but also to our subjective reading of text. We need to fold ourselves body and mind into this mix. What parts of text do we not discover when we open a volume and delve into its depths? What parts of literature hath reading not yet raised to understanding and made pregnant? Venture with me down into the textual unconscious, while I pursue things unattempted yet in prose or diagram.5 From L’Encyclopédie, Philip Gaskell’s A New Introduction to Bibliography illustrates Bindery work in eighteenth-century France without comment on Diderot’s error: his quires open toward “b”, Madame Couſeuſe, on her side of the cords, but they should lie on ours and open toward us. (There is no table-room between sewer and frame.) The forward thrust of shoulder and elbow is correct, but the forearm is wrong: it should arc forward, so that her hand reaches beyond the cords and into the gutter of the quire (not rests on top of it), there to receive the needle thrust in through the spine by her right hand and dispatch it back through the gutter. In sum, she needs one hand at the spines of the quires on her side of the cords and the other serially in their gutters on ours, the needle shuttling between them (as in badminton, say, but with cords for net, needle for cock, and hands for rackets). It is the work of the gutter-hand as the engraver should have shown it (hidden it, rather), that requires this forward thrust.6 But isn’t weaving a metaphor more apt for this woman’s work? In Old English (spoken when Eve span), “woman” was “wifman” – and wife and weave are cognates.

IE*tekth-

IE *plek-

Their snaky folds

b

142

random cloud

The gradual piling up and sewing of quires on the frame can start from back or front; and either hand can work the gutters. Sewing of the Folger copy began, I deduce, at the front; and I have arbitrarily chosen the right hand to ferry the quires, open them, and sew in their gutters. Below, to display the quires as a binder would see them, I have rotated the previous diagram and flipped it mirror-image. We are leaving that missionary position of the book known to all good readers: for quires are sewn lying down and from behind. The gutter stitches, shown black before, are grey now (as the sewer cannot see them) and only the cross- and kettle-stitches, which connect them on the spine side (which alone are visible to her at the frame), now appear black. (The cords are vivid for the same reason.) But understand: black and grey comprise a single thread − actually six lengths knotted together. The first knot, in quire C, stitch 6, is shown here (as well as on pp. 136, 140, and 172).7 e

d

c

b

a

H G

fore

F

aft

E D

fore

C

aft

To begin her sewing of C, D, and E three on, the binder takes, from the collated pile at the back of her table, quire C, already dog-eared (so I deduce), turns it over, and places it snug against the cords on top of what had already been sewn – the (blank) endpaper and the (printed) quires A, a, b, and B.8 Perhaps D now also accompanies C, as these two quires were dog-eared together. In any case, once C is on the frame, the right hand slides down its central opening – which alone is accessible from the head of the quire (the bolts there being still intact) – almost to the tail, stopping at what will become sewing station 7 when her left hand thrusts the needle through the spine there. The right hand takes the needle, draws it and the (newly-knotted) thread 3cm rightwards (towards the head of the quire), and thrusts them out at station 6, on the right side of cord e, to complete stitch 6 (see the top of the next page).

Fearful Asymmetry

143

e

C

aft

stitch 6

The right hand then withdraws from C to fetch dog-eared D to the frame (if it had not come earlier, along with C) and slides almost as far down its central gutter, to catch the needle being thrust in at sewing station 6, now, however, on the left side of cord e. (Around e, therefore, the binder’s left hand has in the meantime essentially looped the thread, upward one quire.) Having received the needle at station 6, her right hand draws it further up the gutter, and passes it out of D at station 5, just to the right of cord d, where the binder’s left hand receives it, to complete stitch 5 – as shown in this diagram. e

d

D

fore

C

aft

stitch 5

The right hand then brings quire E to the frame (it surely comes alone), and slides down its central opening to catch the needle thrust into it at station 5, left of cord d, and (as expected now) draws it along to station 4, and thrusts it out through the gutter on the right side of cord c, so completing stitch 4. e

d

c

E

stitch 4

D

fore

C

aft

But now the pattern alters (and this change will explain why E has no dog-ears). Before bringing new quires to the frame, the binder must make these quires stable by sewing the second and final stitch in each of E, C, and D.

144

random cloud

First, she thrusts the needle into the spine of the very quire from which it just emerged and at the same station, though into a new hole now, slightly to the left of cord c (and also somewhat off the gutter, so as to avoid piercing the previous stitch). Having waited all the while at this station, the right hand receives the needle, draws it rightward to station 3, and thrusts it out through the gutter, right of b, to the waiting left hand. And so concludes stitch 3. e

stitches 4&3

d

c

b

E D

fore

C

aft

Withdrawing now from E, the right hand must find its way back into C, for the sewing of the second stitch there (stitch 2) – following which it will leave C to sew the second stitch in D (stitch 1). When the needle emerges again from E, the left hand draws it down two quires to station 3 on quire C. Locating the spine side of this station is easy, for the spine is always in view. e

d

c

b

a

E

stitches 6&2

He barks.

D

fore

C

aft

But turned as her body is to access both it and the gutter, the binder cannot see down to the right hand’s points of access to the gutter of C – or of D: her view of them is obstructed from above. With many gutters to choose from in the dark – A, a, b, B, C, D – her hand will grope, blind as Milton, unless led by It must be Guided by this Woman’s Best Friend, her trusty Seeing Eye dog -ears. (Atop the next page, survey the scene of guided entry in the very thick of the action – from a vantage point just off Gladys Brochure’s right elbow.)9

Fearful Asymmetry

145 entry to E D

entry to D

e d c b

C

a

C

B b a A endpaper

}

D

E

entry to C

D C B b a A

This illustration shows how easily dog-ears can lead a blinder’s right hand under them into the central gutter of C (its first destination), or above them into that of D (its final destination). No dog-ears are required for E, of course; having once entered and exited this quire, Gladys’s right hand never returns. How different is C.12.h.17,18, a British Library copy of Spenser’s Fairie Queene, 1590 & 1596: all of its quires are dog-eared, for the binder had to enter each three times. The format of these volumes, by the way, is quarto-in8s and the sewing – on five cords again – is two-on (not three-).10

M Q

N R

O S

She never returns.

P T

One more stitch in Paradise Lost brings needle and thread to the right of the frame and to the head of the gutter. Exiting D at station 1, they will pass in a kettle stitch through the loop of thread linking the endpaper and quire a at station 1, where the tail of the first stitch will be knotted (see above and p. 142) – now, by the way, is time to unfold the dog-ears in C and D, to facilitate the functioning of those in F and G – then needle and thread will rise to sew F, G, H (see p. 142), just as they sewed C, D, E (but mirror image, now), then conclude with another kettle stitch. This whole pattern will be repeated in the next six quires (I, K, L, & M, N, O), in the next five (!), in the next six . . . .

Unfold the dog-ears.

146

antic understanding

random cloud

Sewing from station 7 to 1 and back again (with some 60 cm of thread), performs a reading, for the order of quires depends on letter signatures present at the base of the first two rectos of each quire. In Italian, the connection is more obvious: literature is letteratura. For many, to read labelling and ordinal functions of the alphabet as “literary” will be a stretch ; but toward that antique understanding I’m now heading.11 To some, this signing alphabet may look foreign: there is no “J”, for example, in the run of quires just listed overleaf. In the early eighteenth century, English did not separate “J” from “I” or “V” from “U” as letters, though all four shapes – I, J, U, and V – had been in the English upper case since the 1630s, when J and U first appeared there. A century later, English still had only 24 upper-case letters (though it did have our 26 shapes). Thus, it is not really the absence of “J” and “V” from the signatures that makes this signing alphabet unEnglish. It is, rather, the lack of venerable Germanic “W”.12 Our book is signed with the 23-letter Latin alphabet! − and stutstutteringly so: “A, a, b, B, C . . .”, with a partial lower-case sequence worked into an upper-case one. In an age when biology conceives of “libraries” of genetic codes, literary criticism may well speak of the splicing of signing alphabets as examples of intertextual genetics. Reading these neglected signatures will now help lead us to novel insights into literary form. Below, I have started a numbered list (to be concluded on p. 154) of this volume’s texts and paratexts – with indications of sequence, signatures, quire sizes, and the numbers of quires originally planned. (“Originally” allows for any unforeseen structures that may have – that did arise during production.) 1 2 3 4... last

A2 a4 b4 B–3E4 3F–3I2

half title and title Bentley’s Preface; Milton’s The Verse Milton’s Argument; Bentley’s Errata Paradise Lost Index

1 half-sheet quire 1 full-sheet quire 1 full-sheet quire 50 full-sheet quires 4 half-sheet quires

From col. 2 of this list, we can devise a codicological formula for the book: 4º: A2 a–b4 B–3E4 3F–3I2 The format (4º) is quarto: a sheet is folded twice, to make a four-leaf quire.13 (The Faerie Queene just referred to is also quarto, but it requires superscript “8”, not “4” because it has two sheets per quire.) Signing of a book’s main text commonly begins with “B” and proceeds through as many alphabets as necessary – “B, C, D . . . 2A, 2B, 2C . . . 3A, 3B, 3C . . .”, as happens here. Signatures of paratexts are often non-ordinal symbols: ¶, †, ‡, ∗, §, ☞, , and [, for example. Like “a” and “b” in Paradise Lost, they are often syntactically ambiguous with regard to “B, C, D . . .”, the main sequence.14

Fearful Asymmetry

147

Sew what? How can knowledge of stitching, dog-earing, quiring, and signing impinge on literary interpretation? Well, the repeated sequence “two quires dog-eared, one not” imparts an anapestic rhythm to this volume. I cannot think of a literary meaning to ascribe to the repetition of this beat, but I certainly can to its interruption – to arrhythmia. If, from U,15 say, we project down the line, Folger quires 2C and 2D ought to be dog-eared (respectively aft and fore); and 2E should not. And shouldn’t they be sewn like this?

segue

A glance ahead shows that 2C is indeed dog-eared, aft – so far so good –

but 2D is not dog-eared, and it is the only quire to be signed in the middle. Where we expect a plain quire, 2E is dog-eared – aft again. And these last two quires exchange their pattern of stitches. This pattern is not actually impossible, but our logic of folding, sewing, and signing certainly does not explain it. Though 2E is sewn unexpectedly, the anapestic rhythm is apparent thereafter and continues regular to the end of Paradise Lost – but now out of synch. That’s something. Our obsession with the relation of dog-ears and stitches pays off in a vision of the inflected body language of this book: some structural deformity exists in or near 2C–2D. As it involves letters, it is (need I repeat?) a matter of literary organization. To higher criticism, rhythmic irregularity in mere signatures may be no more communicative than a stutter, hiccough, or sneeze during speech. To frame the matter loftily, then, consider coordinate bibliographic evidence in copies at other libraries: Bayley/Howe (University of Vermont), Clark (UCLA), and Houghton (Harvard). I’ll focus now on just five consecutive openings in these copies, 2C2v–2D3r, within which must lie the deformity. (By sublimation gradual, I’ll work bodily functions up to spirit.) First, let me read the texture of the paper – another cognate! Made from rag fibres in a wire mold, a sheet of laid paper is conspicuously marked by the wires on its under side. The mold gives its name to this surface of the sheet.

IE*tekth-

148

random cloud

When a sheet leaves the mold, its upper side is turned onto felt, whence the name for that surface. In raking light, corrugated mold and smooth felt sides have distinct textures, as the frontispiece shows: more vivid chain lines on all but p. 238 indicate mold sides.16 (See n. 4 for an even clearer mold example.)

felt

mold

mold

felt

felt

Suppose we arbitrarily print the inner forme on the felt side of a sheet and the outer on the mold side. The first (of two) folds to make a quarto quire of a sheet is across its longer axis. This fold creates the bolt and also moves the felt side to the interior (hence the name “inner forme”). The second fold creates the spine and gutter (and moves half of the mold side to a new interior). After the quire is sewn and its bolts are cut, readers experience regular alternation of felt and mold sides from one book opening to the next. Thus, in the first opening here, the facing pages 1 verso & 2 recto are both felt; 2v & 3r (the central opening) are both mold; and 3v & 4r are both felt. (At the outsides of this sheet, the mold sides can, of course, variously face mold or felt sides of the flanking sheets.) In the interiors of single-sheet quires, therefore, one always expects symmetry of mold or felt across the gutter. (Between quires, however, where 4v and 1r confront each other, the odds for symmetry are 50/50.) For the Vermont copy, the diagram opposite above displays the mold and felt locations for each leaf of the chosen range (see “m” and “f” atop the gutters). The first and last openings here should be at the centres of consecutive quires. Their “mold” and “felt” designations are paired as expected: in each example of internal openings, both pages should have the same value, “f” or “m” – and both happen to be “m”. Both pages of the middle opening here (the face-off between these quires) should therefore be “f” − and so they are.

Fearful Asymmetry

149

But in the two other openings, each internal to a different quire, mold and felt are asymmetrical across the gutter. Something is fallen in Bentley’s Paradise Lost: neither 2C nor 2D can be as originally created. Having read texture closely enough to differentiate felt and mold, we may also have noticed an uncanny multiplication of text. (I will represent it with arrows in a recasting of this diagram, below.) Before the ink of Paradise Lost cured, it set off, and created a mirror-image .tsoL esidaraP At the top of the diagram, for example, the texts of pp. 196–7 set off symmetrically (as we might expect), verso onto recto, recto onto verso – and so at the bottom, on pp. 204–5. Our book is revealed as self-reflexive. Try as we might to dismiss this obscure dimension of literature as mere echolalia of the text we have been conscious of reading thus far, we really can’t ignore it, for offsets in the three other openings are asymmetrical. We behold “991” on p. 202 of the Clark Library copy (not “302”) and “202” (not “891”) appears on p. 199. Such remote offseting adumbrates a brave new wor d of literary form.17 Vermont M

M

M

C v F

M

C v F

C r

IE *plek-

F

r M

F

r M

M

r

Vermont

M

991

C v F

M

C v F

C r F

’Tis new to thee.

r M

F

202

r M

M

r

150

random cloud

In a metasubtextual perception such as this, consciousness floods back to the medium, from which reading has abstracted it. But no more abstraction now. The same v configuration of arcs in Fisher and Clark copies as in the Vermont copy helps convince us that this asymmetry is F F M not a fluke. The Folger copy (opposite) may seem to offer a symmetrical map of offsets; but the mutual offsetting of C v C r 2C4v and 2D1r is darker than normal. (I have registered the differences in offset density by varying the size of F F arrowheads.) The offsets of 2C4r onto 2C3v and of 2D1v onto 2D2r are just as dark: thus, all four pages of 2C4 r and 2D1 set off more vividly than the other six pages in this range. 2C3v and 2D2r do talk back to 2C4r and M M 2D1v, their dominant mates, but softly. In these diagrams, asymmetrical densities and arcs tell us exactly where to r focus within 2C and 2D: on 2C4 and 2D1. But the clincher is the presence of stubs in the gutters M M of the Fisher and Folger copies. We can read them for felt and mold to determine their conjugacies through the r gutter: the stub between 2C3 and 2C4 is conjugate with 2C1, the one between 2D1 and 2D2 with 2D4. All this Fisher evidence of conjugacy, remote offset, density of offset, and stubs points unequivocally to cancellation of the original 2C4 and 2D1 and their replacement by a single bifolium (as diagrammed below).18 The lower model on p. 147 was deficient: it represented neither the stubs present in many copies, nor the actual sewing in this range. If we take into account the stitching of the cancelling bifolium in this copy, the anapestic rhythm of dog-earing proves absolutely regular, quire by quire. What is irregular, and what derailed my proleptic reading of signatures, is the placing of the lone signature “D d” on the second recto of the cancel, for no other quire is signed in its centre − or only there; and no other signature M

M

M

appears on more than one sheet: the fifty different signing letters thus do not disclose this belated fifty-first quire. (We already knew in life to read between the lines. From now on, we will also have to read between the signatures.)

Fearful Asymmetry

151

We certainly can read the conjugacy of 2C4 and 2D1 in the gutter between them. Amid debris there, one often discerns the continuous surface of the facing pages, or at least the collinearity of their chain lines. We can also deduce the conjugacy of these pages from the very presence of sewing in this gutter and from the two pages’ always being both felt (as in the Vermont and Fisher copies) or both mold (as in Folger). Our earlier notion that quire 2C had dog-earing but 2D did not was right only misleadingly: those quires are no longer as originally printed, each now has only three leaves (and sometimes stubs of a fourth), and a cancelling bifolium has replaced their two adjacent leaves. What I had formerly presented as dog-earing on the last leaf of 2C was wrong; dog-earing pertains instead to the first leaf of the cancel. The collation advanced on p. 146 was adequate for the book as planned, but not for the book as produced. And so, I offer to improve it herewith: the signing alphabet has grown in subtlety and complexity – for 2C is not quite 2C, nor 2D 2D. (Such alphabetic reformulation is a matter of literary criticism.) 2

4

4

4

4

196

FF

C v Cc2v 198

M M

C v Cc3v 200

M M

Cc4v 202

C Cc3r M

FF

FF

Dd2v

199

C r Cc4r M M

201

Dd1rr

M FF M M

Dd1v 204

197

FF

203

Dd2rr FF

205

Dd3rr

Folger

2

4º: A a–b B–2C (±2C4) 2D (±2D1) 2E–3E 3F–3I [cancels 2C4 and 2D1 comprise a bifolium: 2C4.2D1]

Other pathways can lead to this structural awareness. First readers of a Harvard copy experienced a gap between 2C3 and 2D2: for the intervening cancelling bifolium had been bound after the remains of 2D. This copy is now in narrative order, but notes on 2C3v and 4r disclose the original disarray. My reconstruction certainly makes sense of the latter note, “turn two leaves for p. 209”. But the note on 2C3v contradicts, saying that the cancel is “inverted”. If this is not a mistake for “inserted”, it may mean upside down or folded inside out. The latter was a common condition of the cancel, as offsets show; but if the bifolium had been Ee so folded, 2C4r would have been no [ dC D r]d more than two pages from p. 209, not (as I show it) two leaves. Was “misplaced. turn two leaves “leaves” a mistake for “pages”? (If it was, a bifolium both inverted and “ The two next leaves are misplaced, inside out could also be said to have 19 begun two pages from p. 209.)

152

random cloud A r canceled

iii

A r xciii

xcii

cxii

f

K

K v

K

r

Fisher

Traditionally, a pile of booked sheets was consolidated for beating, pressing, or storage by being folded roughly in half across its longer axis, in anticipation of the first precise fold of each sheet (as diagrammed on p. 148). In such a structure, the six half-sheets (A, 3F–3I, and cancel 2C4·2D1) could have remained flat. The frequent misfolding of the cancel thereafter served to bring its interior signature to the front of the quire for the routine alphabetical collating of quires prior to sewing: what more conventional (absent instructions from the printer) than to fold each half-sheet signature-out (as all but one of them indeed required), so that the first signatures in this range read “C c”, “D d”, “D d 2”, “E e”? Such retrofolding of the cancel allows us to reconstruct a common literary misinterpretation from 1732, consequent, I suppose, on the revising compositor’s literal response to “D d” (the only signature in the cancelans), who could have saved the day by signing “C c 4” instead on the cancel’s other leaf. Perhaps some of the cancel’s other paratextual features (its catchwords, or page- or versenumbers) eventually provided Gladys with accurate, less equivocal clues of narrative sequence; but she and her sisters must initially, at least, have read only signatures, and so folded the cancel inside out. In scores of copies seen, only Harvard gives evidence of ever having been wrongly bound so. So, be not hard on Gladys (or on Amaryllis or Neæra). Who at first does read (or understand) everything printed in Paradise Lost or ?tsoL We now know that the original printing of this edition was altered before publication. Was an error removed? a clever idea added? Alas, the sixty copies of this work I have seen reveal no precancelled state, either as it was first stamped in the printing press or as it subsequently set off. Many copies of this work are yet to see, however, and chances are that one of them (or some printer’s waste in the binding of an early eighteenth-century book) will turn up these states of 2C4 and 2D1. I am optimistic because of what I know about Bentley’s masterpiece, A Dissertation upon the Epistles of Phalaris, 1699. This octavo contains four cancel leaves – A1, e7, F3, and K5: A8(±A1) a–e8(±e7) f 8 B–F8(±F3) G–K8(±K5) L–2M8 2N6

Fearful Asymmetry

A is a full octavo quire, comprising the title page and the start of the preface. Thus, a–f, concluding the preface, is not a self-contained literary unit – in contrast, say, to each of the quires a and b in Bentley’s Paradise Lost: if a–f are not spliced into the main, uppercase alphabetic sequence, A–2N, as “A a–f B–2N”, chaos ensues. In Bentley’s Dissertation, a monument of Greek philology and a stalwart in the Battle of the Books, the guerdon lies near the surface. Offsets in the Fisher copy (mapped opposite) and the Brotherton Library copy, Leeds University (on this page) reveal traces of the cancelled states of e7 and of A1, the title page.20 A comparison of the text of the offsets in these (and many other) copies with those of the cancels now bound where the leaves that set off once stood allows for a detailed reconstruction of an earlier plan: this volume was to have contained not only Bentley’s text, but also one by William Lloyd. However, Bentley’s A Dissertation and the bishop’s A Chronological Account of the Life of Pythagoras parted ways at the eleventh hour. With the exception of an unsung copy at the University of Alberta, they everywhere seem to be bound separately, each with cancels smoothing over the separation. In the Alberta copy, Bentley’s text does have the usual four cancels; but Lloyd’s text is, uniquely, without any cancels at all. In A Dissertation, cancels A1 and e7 adapt the very settings of the pages they cancelled. Thus, by the end of the runs of the relevant formes, the printer must have known he needed to replace these two leaves, and so did not immediately distribute the type of the (three) relevant pages. Every offsetting copy of this edition proves an anarchive of its own production, no two copies quite the same in their patterns of remote offset. We take for granted that each manuscript copy of a text is unique; but the same often holds true − praise the Lord − for early printed books. Mechanical reproduction did not stabilize their texts. There is much more to say about remote offsets in the 1732 Paradise Lost, even if there are no more cancels to generate them. Half the copies I have seen exhibit unique patterns of such kiss-and-tell. Let us go first to Toronto, to the Fisher Library, for mirror-image offsets; then, to London, to the British Library for right-image examples.

153

ii

iii

A r xciii canceled

xcii

xciii

xciv

xcv

v xcvi

xcvii

K

with pasteon cancel paste-on cancel removed

N6v

Brotherton

154

random cloud

In the Fisher copy, remote offsets usually connect 1r and 4v pages. Such evidence points to the stacking of folded quires before sewing. Consider (left) the four arcs in the 50 (now 51) poem quires (B1r–3E4v). Their being parallel hints that they are systematic: evidently, all these quires were once divided into five groups of ten, as follows – to each of which I will assign an ordinal, continuing the count from the “4 . . . ” that I began to define on p. 146:

[blank]

4: B–L 5: M–X 6: Y–2H 7: 2I–2S 8: 2T–3E

Y

r

T [blank]

– after which, the concluding Index, can now be numbered “9”. In the switchboard atop the next page, the final ideal order of all nine sections stands on the right. On the left is the binder’s prior practical arrangement, attested to by the map on the present page. The four paratexts crowned the pile (see opposite, below) in the order 2–9–1–3 (as they also did, I see, in the McGill University copy).21 Beneath it lay the five groupings of the poem quires in a striking arrangement 8–7–6–5–4 (see opposite, mid page). The offsets internal to each of these five sections reflect narrative order, but the sections themselves ranged in the other direction. Arrows between the columns reveal the eventual chiastic reordering of these five sections of poem quires, as 4–5–6–7–8. The early configuration of the pile and its parts was alphabetically structured, no more random than the bound (or, in Harvard’s case, re-bound) book. Forget your “A B C”: here B touched X (not A); M, 2H (not L); Y, 2S (not X); 2I, 3E (not 2H). These contacts cast a spell, set off in the book as text. Whenever text is legible, read it. Read it − so you’ll know who slept with whom. For this is literature X B C letterature is this. Perhaps these nine units were gathered by genres? We can be read the stratiform pile geologically for a history of production: perhaps the sheets of paratext

Fearful Asymmetry

155

were printed after the poem. (Refering to a printing error in Bk. 12, quire b must have been.) Such a reading reflects literary criteria at play in the manufracture of the book. There is no departure from the text of this copy in an anarcheological reconstruction. Text can now be read not only abstractly, as a linear string of words, The Verses of Milton, say (unhappy though in Heaven), but also concretely, as disjunct articulations of the codex itself. And when

text echoes so within the volume, the very tissue of these locations (that root again!) signifies, enchaunting the book’s Song of Itself, if you will − Even if you won’t. Who gives a damn that No One Up There is watching? Stand on Mother Earth. Open a book of leaves, hands dirty with printer’s ink. Lo, a star falls on your left sandal. Your barbaric yawn arcs o’er the rooftops of Mississauga. Turn a new leaf. Stride forth into Eternity. Read. Read the medium.

IE*tekth-

They yawp.

156 Return Credit, return: the d ead voice is past.

random cloud

Return, Credit. Remote offsets in the Folger copy that have not yet been discussed are mapped opposite; they point to different prior structures. Three runs of quires seem clear: 2B ↔ 3I ↔ A; 3H ↔2C; and cancel 2C4·2D1 →Y. It is not clear yet how to link them. But we are on somewhat familiar ground: the first and last pages touch again. The text block was divided at roughly its midpoint, between 2B and 2C, and the lower-right corner of 2C1 was dog-eared forward. Its projecting beyond the spine must have facilitated reordering.22 When this structure came under pressure, in a standing press, say, the fold was remotely debossed (as indicated in the diagram by the circle-heads instead of arrowheads). Such folding at text-block divisions is another kind of dog-earing, not uncommon. Witness the University of Virginia Black copy (left). Its lone remote offsetting links first and last pages (why are we not surprised?), but the folded corners of R1 and 3I1 deboss 2H4 and Q4, respectively, hinting at the fœtal postures of this copy. The two halves of the Folger copy (right) may once have been reversed – or if not all of the first half, perhaps only through quire X, for offsetting shows that the cancel once stood before Y1r. In the eighteenth century, when the run of a sheet was finished and all the copies of it were dry enough to handle, they would have been stacked in a heap and warehoused until ready for collation. So, there would have been fifty-seven heaps of sheets of Paradise Lost (eventually fiftyeight) when gathering began in earnest. It was impractical to gather one sheet from each pile in order to comprise a single copy: over 25 metres of table top are required to lay out that many heaps, and gathering from them all at once is too cumbersome. The earliest printer’s manual, Moxon’s, is mum on the practical issue, but a later one, Caleb Stower’s The Printer’s Grammar, 1808, advises a man to gather a Virginia large book in stages, ten or eleven sheets at a time, “beginning with the first signature of the body . . . which is sometimes marked A, but in general B . . .; he then follows with C, D, &c . . . till he has laid down a sufficient number of sheets, which is commonly from B to M [B–L is the corresponding unit in the Fisher copy] . . . where a volume runs through the alphabet two or three times, several gatherings must be made. [Stower says nothing about where one stores the first gathering when the next is begun. We shall return to this question shortly – for that is where our eureka is hiding.] In such cases, eleven or twelve sheets

Fearful Asymmetry

157

in a gathering is enough. The title, little a, b, c, cancels, &c if any, should be left till the last, and placed at the end of the gathering, so that, when [it is] folded, they may be found withinside the gathering”. To gather Paradise Lost by elevens would see the fifth gathering starting at 2Z. It could be the last, for the remaining bulk of paper, 2Z−3E4, 3F−3I2 A2 a−b4, is just under that of eleven sheets, even though it includes fourteen signatures (for five of them are on half sheets). Such a final gathering, with its full sheets folded in half and its half sheets left unfolded, would bring 2Z to the A1r outside – and A could be “withinside”, cheek by jowl with 3I, thus creating the striking contact we have often noted of the first and last pages.23 The next stage is a “booking” of these gatherings. Whether it merely bundles or integrates them could be revealed by offsets: does 2Z1r set off on 2M4v, for example, or on 2Y4v? In Paradise Lost, I have not seen unambiguous evidence of setting off in gatherings or bookings (as opposed to setting off in full sheets or in quires); but offsetting certainly can take place within and between bookings. I have seen it frequently in early-sixteenth-century quartos and octavos printed in Venice and Lyons – as I will document in a note.24 It is not clear whether, in the case of the Fisher copy, an integrated booking (if it ever existed) was subdivided for folding into units of ten, or the original gatherings were in units of that size. But after folding of the first ten sheets made ten quires in alphabetical order, we can suppose they were set aside together, B-up; and when the next ten, M–X, joined them, they were placed on top, M-up and X-down, so that extreme pages X4v and B1r touched and, with sufficient pressure, set off on each other. (There is our Εὕρηκα.) According to Moxon, this pressure would have come simultaneously in a standing press; it could have applied to the sewn text-block (as Folger illustrated in “d” in the plate on p. 186), or, as is relevant here, to the gatherings, each now quired, but before narrative order prevailed among them. Thus such stacking of alphabetically-arranged quires from five of the Fisher gatherings can explain the backwards sequence revealed in that copy by the five leap-frogging arrows in the remote-offset diagram on p. 154. But there is another source of pressure, which Moxon does not mention. In the diagram for the Clark Library copy (on p. 159), contrast the leapfrogging arrows with the non-overlapping ones that directly link pages

69

random cloud

b a

Eve Adam

Which two great Sexes adambreak the Word (VIII, 151)

Tasso

Bentley Milton Paradise Lost, 1732 (S3v, p. 134)

Gerusalemme Liberata, 1724 (vol. 1, n1r, p. 97) –

Fearful Asymmetry

159

R1r & 2L4v (19 quires apart) and 2M1r & 3E4v (17 apart). (The rest, 15 quires of poem and 7 of paratexts comprise a third group, which I will come to in a moment.) The absence of leap-frogging here demands a different explanation: whereas in Fisher all the remote offsets could have come about simultaneously in a standing press, the remote offsets in Clark must have occurred over time, during which each of the three quire groups was briefly reconfigured. Consider the first group, R1r–2L4v. Imagine dividing it and interchanging its halves, so that its extremes, 2L4v and R1r, kissed each other in the middle. “H ammered ”, a recent study by R. MacGeddon, offers a cogent rationale for such an astonishing tryst.25 To flatten it for binding, a text block was routinely divided into equal units (three in our case) and a man wielding a broad-faced hammer, Diderot’s “a” (see him at work, opposite) repeatedly struck each “beating” (aptly named) on both sides. Next, dividing it and interchanging its halves, he beat it again, front and back, then restored its original order.26 Remote offsets thus arise in the second phase under the hammer, when the beating’s outer faces randy vous briefly in the interior; and local offsets arise in both phases – and also in the standing press (where even remoteoffset pages can acquire local offsets). Beating 3 consists of B–Q, the remaining poem quires, plus all three paratext quires from the front of the book and all four from the back. As in the prebound order of merely the poem quires in Fisher, that of the mixed quires of beating 3 shows that internally the subsections were basically in narrative order at the time they set off, but they themselves ranged backwards.

B –Q

a–b

A

I

portrait young

A1r

blank

portrait old

blank

blank

blank

F– H

But all this is tame, don’t you think? Why not come upstairs with me now and see my princely collection of intertextual right offsets. They’ll leave you in stitches.

Clark

random cloud

Fearful Assymetry

162 emerging from under ground at King’s Cross, I had not thought the book had saved so many

HCE

Word gets around.

RNA

random cloud

Here in London, the King’s Library Milton bears remote offsets from both itself, in mirror image (the kind we have been dealing with so far), and, in right image – not from itself, but rather from Tasso’s Gerusalemme Liberata, “Printed . . . Upon Royal Paper in Quarto”, dedicated to the monarch, and published in two volumes in Londra in 1724 by Giovanni Watts and Giacob Tonson, the very Tonson who, as mere “Jacob”, had published Bentley’s edition of Paradise Lost in London in 1732. (Don’t ask me to explain how ink in Tasso could have been so alive as to set off in Milton up to nine years later!) The Offsetting Saga does not end here. Besides setting off remotely on itself, the King’s Library Tasso receives offset, mirror-image, from a yet-to-be identified 32mo French psalter and also from something in our own language.27 In these volumes we overhear a mutlitude of tongues – English, Italian, and French, the last translating Hebrew. (“All Europe contributed to the making of Volume 1 begins with Tasso’s Vita (* *a–*b a–u) and concludes with the beginning of the epic (B–Z 2A–2U). I have not yet identified any of the dozens of psalter pages, but I have charted the sources of a hundred offsets of the poem onto itself and of the Vita onto both itself and the poem. Opposite, I arbitrarily bend the two alphabetical sequences of these signatures, as if they were flexible docking proteins, to display their transfer of information efficiently, the poem on the left and the Vita on the right. Now, whereas, in the Fisher copy of Paradise Lost, remote offsets are mainly from and to the outsides of quires (i.e., 1r or 4v pages), any of the eight pages of a quire in this Tasso can give or receive offset from any page in another. Offset originating in Vita p and projecting into Vita p points to a fact (confirmed by other evidence) that the source of remote offsets in Tasso came from another exemplar of the same edition. (So don’t be taken in by the shorthand of my diagram: the royal paper of George III’s crazy copy of vol. 1 was talking not merely to itself.) Offsets on the two pages of an opening (within a quire or between quires) usually reproduce the two pages of another opening: overleaf, pp. 160–1 show Vita d2v–3r (pp. 28–9) – plus sixteen psalter pages – setting off on 2Q2v–3r (pp. 300–1) in the poem. (The psalter tilts 5°, the Vita 8°.) These offsets point to slip-sheets as the means of textual transmission. They would have been inserted into an opening prophylactically, to keep adjacent pages from setting off onto each other during beating or pressing; but then (and here is the madness of it all), they were recycled to other openings before the ink they had sucked up was dry, where they served not only the intended function, to be set-off upon again (again in mirror image) – to blot, but also, contrary to intention, to set off themselves (now in right image) – to blot in quite another sense.

←B–Q ↔ a–b ↔ A ← 3I ↔ 3F–3H →

163

Opposite, below, see such a quire of Tasso with two setting-off sheets. Suppose the one on the right, unsized (for maximum absorbency), has here received offset en miroir from Tasso. With this ink still wet, it will soon set off that text, rightimage, elsewhere in this work. The left sheet, still wet from having been stamped earlier with psalter text, has just set it off mirror-image in this opening of Tasso and, simultaneously received the local Tasso text, mirror-image. The next time round, this slip sheet will set off psalms again, mirror-image (more faintly now), and (strongly) the new Tasso, right-image. And there you have it, dear viewer: these offsets were created by the very means of preventing them. Vita slip-sheets were recycled in both the Vita and the epic, as arrows between these parts indicate. Startlingly, those from the range Vita d–l migrated not only to 2P–2U at the end of the poem, but also, with offset from m, to B–C, at the start. Of course, in bound copies of Tasso, beginning and end are as far apart as can be; during blotting, however (which took place after the collation and folding of the sheets, but before sewing), they were as close as could be. Tasso Unbound, this libro libero, was configured like an uroboros, that ancient serpent with its tail in its mouth, a mystical figure of undifferentiated fullness. Before it awoke, the line was a circle. B A

C D

Z Y X

E F G

U T S R Q P

H I K

O N M L K I H G F E D C B

M N O P Q R S T U

* *a *b a b c d e f g h i k

u t

l m

s r q

n o

p

Word gets round.

no bar of membrane or joint (VIII, 620−29)

164

random cloud

Now, James Joyce conceived of Finnegans Wake as a circe, a simultanus short-circuit of all myth (Every Thing Equally and Immediately Remote), an indefinitely wyrm-edened book − in a word, an apocalapse. But its practicale mbodiment 70 years ago by Fabbro & Eliot in no way challenged thel inearity of the codex. In these pre-bound quires of our Paradise Lost and Gerusalemme Liberata, however, we do witnes sexamples of a navel (if eserotic) structure, each a book in them aking, with its head up its asp. The diagram opposite spells out what I have divined of such literature in a hundred examples of right-image offset from vol. 1 of Gerusalemme Liberata into Paradise Lost. The middle column is an alphabetical ordering of the quires of the 1724 Tasso as bound. The right column lists quires in the 1732 Milton almost as bound, but commencing abruptly in E. Quires B–D, starting the epic, begin in the lower left of the chart, with the preliminary quires above them. (To establish continuity, I have snaked a grey line along the circumference to join these parts of Lost Paradise.) Many arrows of intertextual offsetting link these epics in sustained and intimate conversation across the years. (Sample it three openings back, p. 69, in the union, tête-bêche, of Tasso n1r and Milton S3v). The alphabets of the epics run parallel, in two basic orders: witness the more or less horizontal arrows of the conversation, above, and the gracefully sloping ones, below – with an overlapping weave of the two styles of line beside Paradise Lost quires O−S. When, in utero, Tasso addressed Milton, each text seems to have been in narrative order, but variously out of phase. Quick now, find it. Our pot of gold lies low in the central column: here pages in Tasso 2C–2U set off, not only as expected, into the very end of Milton, 3C–3H (dexter), but also (sinister) into the very beginning, A, a, b, B. Why, this is our familiar epic Hero, that Serpent olde with its tale in its myth. Imagine that, when this setting-off occurred, Tasso was in its bound sequence, but Milton was not: the Paradise Lost quires would have begun in the midst (Horace take note),28 passed through end directly into beginning, concluding where they startled, in the mi st of things. Imagine instead that Milton Unbound ran as presently stitched. Now Projecting Tasso would have started midway, returning thither, having passed through end into beginning. Imagine, ultimately, both texts – or more (recall the psalter) – undressing each other as uroboroi, projecting tane into tother through narrative cycles that could have begun and ended any where in the muddle of things? envoi Jerry, man of three-score years & ten, tell me if you know: this shuffling of epics like decks of cards – will we ever more closely map the Sex of Angels?

l

o

ad

Ghostlier Demarcations

165

Tasso Vita

*, *a –*b a b c d e f g–h i k l m n o–p q r–u Tasso Poem A–C D E F–G H I K L–O P Q R S–U X Y Z

Milton Prelim A a b Milton Poem B C–D

E F G H I K L M–N O P Q R S T U X Y Z

see p. 69

Milton Index

166

random cloud

CODA: verbum et reverberatio

σπαραγμός

Not echolalia, but echolocation. In radar, the interest lies not in the megawatt, coherent, crafted signal, but in its faint, fragmented, and unscripted reflections read against the original whole: such a dialectic of pulses intact and broken echoes maps the directions, distances and surfaces of remote objects, and even (Enter Doppler) their s p e e d s. The practice of reading printed books in this fearful asymmetry is similar, but with an awful difference, of course, for literary signals are not monotonous electronic beeps. They are instead, as it were, the precious life-blood of master spirits – King David, Tasso, Milton, Dr. Bentley, Bishop Lloyd – embalmed and treasured up on purpose to a life beyond life. The gambit of my textual criticism has sacrificed the desired encounter with sublime literature to attend to “reverberature”, scattered traces of textual fragments dimly reconfigured, egami-rorrim, perhaps, upside down , seemingly unintelligible, each changed into a thing, like a brain-dead stroke victim or a dead god. Despite appearances, however, reflected letters are not mere dust and noise; nor are they hovering near death. Like echoes in radar, these remote reflections fragmend – bearing precise new messages, rich and strange, cohering in a fresh narrative, like fingerprints at a crime scene, treasured up without purpose to a meaning beyond meaning. Often they inscribe the inscape of the books they inhabit, dealing out that being indoors each one dwells. No Webster’s defines these echoing songs: what they amount to is scarcely to be generalized, at least in the dawn of our study. As I have endeavoured to show, however, their meanings may come clear, not only in their own locales, the structures of which they articulate, but also remotely, as in offset from the cancelled title page of Bentley’s Dissertation, which (as in a comic anagnorisis) brings word of a Lost Twin, separated at birth – not alien, but a second self. A book’s Song of Itself, as I have called it, spells out a diachronic codicology of the book in hand – of this book in this hand. But a books like Paradise Liberata sings more, anticipating that apocalyptic Librarynth, where ( Amen) euerie booke shall lie open to one another

Fearful Asymmetry

167

NOTES My thanks foremost to Moira Fogarty, who, in 2001 directed my attention to remote offsetting in the Fisher Library copy of Bentley’s 1732 edition of Paradise Lost and provoked the decade of research that has led to this essay; to Betsy Palmer Eldridge of the Canadian Bookbinders and Book Artists Guild, for information on sewing as I started out; to Brandon Besharah for help with diagrams and design; to Anthony Ossa-Richardson for research at the British Library; to the conservators of the Folger Shakespeare Library, Stephen Galbraith, Frank Mowery, and Renata Mesmer for their help, understanding, and teaching during my excavations, and also for allowing photography of the Folger copy of this edition before it was rebound, near the end of my study (partly because of it). Thanks also to those with whom I discussed this research and to readers and auditors of drafts for their encoeuragement and corrections: Alex Brett, Rebecca Bullard, Michael Cahn, Joshua Calhoun, Bradin Cormack, Yuri Cowan, Joseph Dane, Gabriel Egan, Jeremy Ehrlich, Susan Green, Pamela Harris, Charles Lock, Scott Mandelbrot, Steve McCaffrey and Karen Mac Cormack, Raimonda Modiano (who supplied the phrase “textual unconscious” on p. 107), Rebecca Niles, Philip Oldfield, Thomas O’Reilly, Mark Owens, Stephen Pender, Varun Raj, Annie Russell, Matt Schneider, Scott Schofield, Jyotsna Singh, Sean Starke, Michael Suarez, Steve Tabor, Elisa Tersigni, Elizabeth Bernath Walker, Michael Warren, and Marta Werner. Thanks ultimately to Bill, my father, a Repairer of Books. This is his cenotaph. Elements of this research were first presented April 20, 2006 at the Inaugural International Conference of the Danish Book History Forum, University of Copenhagen, at the invitation of Charles Lock. “Fearful Asymmetry” first appeared as a keynote address at the inaugural biennial meeting of the College Book Arts Association at The University of Iowa Center for the Book, January 9, 2009 at the invitation of Matt Brown; then as a lecture in the English Department at SUNY Buffalo, at the invitation of Steve McCaffery, October 21, 2009; and in the Textual Studies Program at the University of Washington, February 26, 2010, at the invitation of Raimonda Modiano. My title is a hommage to my teacher, Northrop Frye, and a bookage to his study of his own, William Blake. For financial support, I am grateful to the Mellon Foundation (for a period of study at the Folger Shakespeare Library), to the Social Sciences and Humanities Research Council of Canada, and to all those who wrote on my behalf. 1. Thomas Fisher Rare Book Library (E–10 02647, c. 1), by whose courtesy the photos on p. 134, 141, and 186; and Folger Shakespeare Library (214–216f). The crease on p. 238 is fainter than that on p. 240: the latter leaf cushioned the former when they were folded forward together. (See n. 16 for more on the relevant quires, 2H and 2I.) The Folger’s four leaves routinely folded together also show progressive softening of the crease leaf by leaf outward from the centre of the fold. A raking light is one that skims along a page, brightening the face of whatever obstructs it and casting a shadow thereafter and also into depressions. In n. 4, for example, raking light effectively reveals the slightly recessed chainlines in the paper (every five lines of text apart), and three depressions, of increasing size, centred on ll. 350, 355, and 360. It may be no accident that these latter contours of the leaf coincide with these locations of type, for it is the bite of typeface that creates them.

The filial Pow’r

Systolic successon

168

random cloud

In this essay, I focus on dog-ears at the head of the leaf. But binders also dog-eared the tail: these diagrams show folio-in-fours format (left) and octavo (right). (I have not found a book with both kinds; but see an apparent exception in n. 22.) Dog-eared at bottom are the Fisher copies of Dryden’s Fables Ancient and Modern, 1699 and Diderot’s Encyclopédie, both folios-in-fours. In folio format, such folding not only locates a quire’s central gutter during sewing, but also consolidates its loose bifolia until they are bound. The latter explanation does not pertain, however, to the similarly dog-eared Valerius Maximus discussed in n. 24, because of the presence in octavo format of the bolt across the top of the quire, usually not removed until after sewing. 2. In quarto format, point holes for registering a sheet during printing pierce the bolt. They are somewhat obscured if the bolt is cut open (see Fisher c. 2), and disappear entirely if it is ploughed. (See p. 186 for an illustration of ploughing by workman “c”.) We must allow for the elimination of small dog-ears during ploughing: thus, some books once dog-eared may now give no evidence of that fact. 3. The diagrams on pp. 136−9 and the collation begun on p. 146 (and revised on p. 151) are for the text-block only, not the whole book-block, which includes endpapers: at front and back of each of the Fisher and Folger copies are marbled bifolia (paste-down and fly-leaf), sewn all along. Also, both copies have the two engravings often, but not always, found in this edition (and not in it alone), of Milton Young and Milton Old. (In the illustrations opposite, they are depicted grey.) Before the Folger copy was rebound, in the fall of 2009, the direction of sewing from Row 1 to 2 as diagrammed on pp. 136–7 could be deduced from a difference in thread colour before and after the knot in quire C. All the stitches in B were dull brown, as was the bottom half (up to the knot) of the lower stitch in C. The top half of this stitch, after the knot, as well as the upper stitch in C and all the stitches in D and E, were copper coloured. The sewing therefore joined the bottom of B and C – in my analysis, down B and up C. Consequently, the sewing of Row 1 began at the tail of A. In the diagram, the numbers “6, 4, 2” under quire A and “5, 3, 1” under quire a assert that the first stitch (only somewhat arbitrarily numbered “6”, as I’ll make clear below) appeared in the gutter close to the tail of A; the second (no. 5), slightly higher, in a; the third (no. 4), higher still, in A; and so on, as the needle and thread shuttled back and forth from one quire to the next, circling the cords en route, advancing stitch by stitch toward the head of the quire. After the needle and thread left the gutter at the head of quire a (at the top of stitch 1), they would have made a kettle stitch on the spine side. (That stitch is thus not visible in this diagram, though it is in those on pp. 142 and 145.) Needle and thread appear next in the gutter at the head of quire b (for the first stitch there); then lower down, in B, in three adjacent stitches (nos. 2–4); then lower still, in b (no. 5); and finally near the tail of B (no. 6), in preparation for the next kettle stitch. So, Row 1 (on p. 136−7) documents the usual tally of twelve gutter stitches in a complete sewing cycle, in this case, from tail to head and back.

Fearful Asymmetry

A

a

169

B

In a book like this, sewn mainly “three-on”, terminal text-block quires are commonly sewn “two-on” for extra strength, as was done here, or even “all along”. (See pp. 140−5 for more on these terms.) Terminal Row 11 (p. 138) offers the sole example in this chart of sewing a single quire all along, though it does not employ every station). Surviving holes and two adjacent stitches in each bifolium endpaper suggested that they were originally sewn all-along, i.e., with the full six stitches. (This reconstruction of the sewing of the endpapers with stitches 1 through 6 explains why I began Row 1 of my chart (on p. 136) with stitch 6 rather than 1.) With one exception, the sewing holes exist only at the stations where there is thread: holes were made by the needle in the act of sewing, not collectively beforehand, by a saw cutting a notch or “kerf” across the spine of the book block. Frequently, especially in cols. 3 and 6 (on pp. 136−9), a sewing-station hole stands eccentric, and the stitch passing through it therefore runs to the next station not merely along, but also off or across, the gutter. Whenever there are adjacent stitches, I diagram this eccentricity, for it shows the binder’s care not to run the needle through a stitch just made. When stitches are isolated, however, as in quire a of the Folger copy, I ignore the occasional random eccentricity of the holes. (Contrast the more-accurate representations of three off-centre holes in quire a in the above diagram with the pro forma holes in the depiction of this same quire on p. 136.) Also shown above are three other quires in which irregular sewing is of particular interest. The engraved portraits (in grey) are tipped in in A and sewn in in B. In A, the binder sewed not in the original gutter, but in the new one created once the plate had been tipped in – obviously, then, before sewing. The dog-earing malfunctioned, steering the binder (see pp. 144–5) to the wrong opening! This new gutter is displaced some 5mm toward the fore-edge. Consequently, thread is conspicuous six times on the verso side of A2, running horizontally for this distance between the (normal) gutter separating A2 and a1 and the back side of the (displaced) gutter on A2. In B, the presence of four stitches rather than the two expected for a quire so deep into the text block may have been the binder’s tactic for securing an engraving printed on heavier stock than the rest of quire B. Note that, although this plate, standing between coordinate dog-earings (in b and B), is not itself dog-eared: the plate must have taken up its present position after the letter-press quires had been dog-eared. It seems that this dog-earing did not take place on the frame. (Dog-earing there is a binding option, however, as the discussion in n. 9 will reveal.) As shown above, there is an empty hole in 2B just under the hole of the top sewing station, on the lower side of cord b, which runs behind these two holes. The empty hole suggests a false start for the third stitch in the run Z−2A−2B. This lone empty hole reminds us that stations in this volume did not predate sewing. 4. The 286mm height is of the text-block after it was ploughed head, tail, and fore-edge. But if part of a leaf folds away from an edge, it can escape the plough, and so preserve earlier dimensions. The photo opposite (by permission of the Folger

170

random cloud

Shakespeare Library, as are those on pp. 170 and 172) shows a deckle-edged fold-out from the gutter of 2Z, where it had slept for almost three centuries until I woke it. An earlier height was at least 301 mm; but even this calculation is not final, for the head was also ploughed. (Page height in the unploughed Fisher c. 2 averages 304 mm.) The sloped lines extending from the gutter to near the present bottom (trimmed) edge of the leaves and lines running parallel to that edge reveal the posture of this fold-out before it opened. (A trace of thread of the lost tail band visible at the top of this photo is not part of stitch 6; it occupies a higher position, as the map on p. 139 shows.) In many books, a sewing station is a single hole through which thread enters or exits the gutter or spine of a quire. This is the case for books in which the stations are created with a saw before sewing. In the present volume, the station is indeed single at the top and bottom holes in the gutter, where there is no cord, and also for the top and bottom of any isolated stitch (except for the doubled top station on 2B just referred to). With adjacent stitches, however, we should conceive of a station as a pair of holes: when the needle punches an exit hole in the gutter on its way to wrap the thread around the cord, it returns to the same quire but re-enters the gutter about 4mm behind the first hole, as measured against the local direction of sewing. This distance between holes is a function of the thickness of the cord around which the thread passes, on the spine side of the quire. The final distance between chad, hinged at the these two holes is often less right of the hole, than the actual diameter of the cord, however, for tension in the thread can tear the holes towards each other. To illustrate this point, see the photograph at the top of this page; it shows two adjacent stitches in quire 2G, heading to the right. Stitch 3 overshoots cord c, then, formed in the having circled it, emerges

Fearful Asymmetry

171

about 4 mm back (and 5 mm out from the gutter, to avoid sewing through the previous stitch, an act that would have prevented stitch 3 from being pulled taut). The bright nipple below station 4 on 2G2v – it is shown enlarged, opposite, below – corresponds to the dark hole above it where stitch 4 emerges. When text-block, or book-block, or bound book, or all three, were pressed, as Moxon describes in Mechanick Exercises (see p. 157 and n. 24), 2G2 met no resistance from the hole on 2G3, and welled up into it. Such great pressure causes paper to conform even to the slight recesses of chain lines on adjacent leaves, a phenomenon seen in many copies of this edition of Paradise Lost. Note also that the holes in stations 3 and 4 are wider than the thread passing through them. At the eye of the needle, the thread would have been doubled at the time of sewing. More importantly, however, when the thread was pulled taut, it tore a small channel for itself down the gutter (often creating a chad), until it was tight against the cord. (Such chads are dragged in the direction of sewing. The face of this one reveals a small part of the other, verso, side of leaf 2G3.) 5. The IE root is *tekth-. Surprisingly, “testament” and thus “testes” are not cognate with “text”, but “tête” and “test” are. See Eric Partridge, Origins: A Short Etymological Dictionary of Modern English (New York: MacMillan, 1958). Grimm’s Law formulates the relation of Hellenic and Italic t- and p- to Germanic th- and f-. Interchanging of a vowel and a liquid [l] (see Italic pli-and Germanic fol-) is metathesis; the unmetathetic form appears in English “flax” (< IE *plek-, “to intertwine”). (Flax is the source of both paper and ink! See Joshua Calhoun, “The Word Made Flax”, PMLA, 126:2, March 2011, pp. 327–44.) 6. Philip Gaskell, A New Introduction to Bibliography (New York and Oxford: Oxford University Press, 1972), p. 150. Diderot’s Encyclopédie appeared

172

random cloud

between 1751 and 1780. “1771” is the date of Vol. 8 of the plates, where “Relieur” Planche 1 appears. It misrepresents the block of quires – flush against the cords, its left edge absurdly not parallel to the sides of the table. (As the table is represented foreshortened, I am not speaking of Euclidean parallels, but of ones that should meet in the evoked depth of the engraving – but the vanishing point for this book is absurdly in the foreground!) Also, although, at the cords, the quires’ height appears to be several centimetres, their far edge is collinear with the edge of the sewing table, so that the quires appear to taper there to no height at all. Of course, from some angles, the edges of the quires and the table could indeed align; but so to relate them in an expository perspectival engraving was, at that time, to evoke Hogarth’s 1754 Satire on False Perspective, or, in ours, the surreal alignments of Magritte’s 1933 La Condition Humaine. The draping of the binder’s skirts implies that her legs extend directly under the table. But at the sewing frame, the “forward thrust” of the shoulder, as I called it on p. 141, is actually achieved largely by the binder’s facing along the front edge of the sewing table. The thrust is rather to the side, therefore, as is apparent in Noel Rooke’s drawing two centuries later (see above) in Douglas Cockerell’s Bookbinding and the Care of Books (New York: Appleton, 1903 [c. 1901]), p. 104, fig. 29. Dudin’s L’art du Relieur: Doreur de Livres (Paris, 1772) (see overleaf, reproduction courtesy of the Toronto Public Library) is also spectacularly wrong. This time, the quires do lie correctly across the cords from the binder, but both of her hands now appear there as well, in defiance of the cords, which should block the way of at least her right arm. As shown in the blow-up, also overleaf, she unconvincingly grips an oversized needle in the palm of her hand, as if it were a gouge (Why, it’s as thick as her thumb!) and although the other hand looks poised to receive a needle, no leaves stand between her hands to be pierced by it. (See n. 9 for more on the needle.) As in Diderot, the vanishing point of the quires her hands rest upon is absurdly in the foreground, and the quires seem well away from and not attached to the cords. Charmingly, however, the thread exiting the needle mm appears to become the screw-thread of the sewing-frame upright! (This pun works in French too. Vt poesis pictura.) But the artist does correctly depict the end of the table towards us as a workspace: scissors and thread are there, for example. But it would have been more representative if the sine qua non, the stack of quires about to be sewn,

Fearful Asymmetry

173

was so near to hand, for madame must take a row quire from this pile at least every few stitches. 7. A weaver’s knot, with one loop and two Z ends, is commonly made on the spine and drawn Y through to the gutter, so as not to mar the leathX er backing. See Edith Diehl, Bookbinding: Its U Background and Technique (New York: Reinhart & Company, 1946), 2 vols, 1: pp. 125–6. T S (One-volume Dover reprint, 1980). R Opposite is the knot in Folger C before it was disQ bound; its two ends point to 4 o’clock. The lower, P blond, one is totally frayed; don’t mistake it for a shadow. A former, 5-o’clock, position of the intact O end is debossed into the verso, at 7-o’clock. (Together N with its faint stain on the recto, it makes a wishbone M L shape.) Sometimes the former presence of a stitch can K be inferred from such a stain, as in mutilated quire I 2D2–4, where the lower, knotted, stitch remained, but the upper did not. The loop and two ends of the knot H in C totaled about 50mm. (The distribution of knots G in the Folger copy, in C, Q, 2D2–4, 2Q, and 3D, looks F calculated, as there is never more than one in any of E D the six stitch positions. The Fisher copy, by contrast, C appears haphazard, with two knots in the second stitch, two in the fourth, one each in the first, fifth, B and sixth, and none in the third.) b A reliable test of whether a book was sewn a from front or back is the structure of the “kettle” A or chain stitches. From Latin catena, “chain”, endpaper comes Old High German ketina, then Modern German Kette, and its diminutive, Kettel. (Kessel, the German word for a “water kettle”, comes from Latin catillus, diminutive of catinus.) In the above diagram, the appearance of the kettle stitches is imagined at the tail of the front of the Folger copy: the rounded part of the kettle stitch always points in the direction where sewing began. As such stitches are covered by the backing of the spine, they are not visible until disintegration sets in. When conservators rebound the Folger copy (with reference to the diagrams laid out on pp. 136−9 and in nn. 3 and 8 for the end papers), they concluded that the original sewing had indeed been from the front. No record was kept, however, of the chain stitches.

174

random cloud

The image overleaf, below, shows triple dog-earing in Toronto’s Centre for Reformation and Renaissance Studies copy of Domenico Delfino’s Sommario di Tutte le Scienze X (Venice, 1621) (B 785 D43S6 1621). The sewing here is twoU on on three cords from the front of the volume, which collates T 8o: a4 b−c8 d4 A−Y8 Z4. There are no endpapers. The beginning S of the sewing, at the tail of a, is knotted to the thread exiting at the tail of b (I have not shown the knot in the diagram, left); R that loop received the first kettle-stitch, passing from A to B. Q There are but three kettle-stitches at the tail, for every second P opportunity was missed: A to B, I to K, and R to S employ this O stitch, but E to F, N to O, and X to Y do not. For this reason, designating every head or tail stitch a “kettle stitch” is wrong. N Often, only an archeological examination of an uncovered or M dissected book can identify such a stitch. L The blow-up opposite, below, shows the exposed foot of K the spine of this copy of Sommario. Its cords are recessed; and the thread merely passes over and does not circle them, I as it does in the Folger Paradise Lost. A kerf at the foot slopes H gently down right, 2 mm below the chain stitches. It may be G that the tail stitches originally terminated in this kerf, but that, F with tension on the thread, they tore upward. But the systematic double and occasionally triple dog-earing in this copy of E the aft half of each quire argues that the present stitching is D not original: it may not have used this kerf at all. C Mirjam M. Foot reports that German and Dutch books B were sewn from back to front, whereas the French and English were sewn from front to back (Bookbinders at Work: A Their Roles and Methods (London: The British Library and d New Castle, DE: Oak Knoll Press, 2006), p. 51). But surely c individual binders would suit themselves in this matter, as they b would on which hand to employ in the gutter. (I conclude that the direction of sewing the Fisher copy was from back to front. a A detailed English description of sewing in this direction can be found in n. 9.) For engravings of a binder with the right hand in the gutter, see Foot, pp. 39, 138, and 144. As engravings print their image en miroir, one cannot be sure they report right and left without reversing them. However, in the image of a Dutch bindery on p. 138, the sewer has his left hand at the spine and a beater hammers with his right. 8. Most stitches in the endpapers were missing, but indications were that the sewing was all along. A, a, b, and B were sewn in two units of two-on, as shown in Row 1, on pp. 136−7 and as discussed in n. 3. 9. The reach of Gladys’s arm can be guessed from the length of thread between knots, the longest of which I calculate at approximately 115 cm, Y

Fearful Asymmetry

175

almost four feet, or (mindful of her e elbow) about one ell, a cubit. When she started a d new thread, the needle could c sit in the middle of its doubled length; therefore, the maximum b length to pull through the first hole a need not have exceeded 27 inches or so. (In all, about twenty feet of thread were required to sew the Folger copy.) The sewing of C, D, and E is better appreciated if just the top part of the illustration on E p. 145 is considered, as you see it here, enlarged. On that page, I represented the needle as straight, D as is common in modern sewing. Ephraim Chambers, however, reports “a long Needle a little crook’d” C (Cyclopædia, 1728, “Bookbinding”, 1: p. 116, col. 2), a detail originating with him, it seems, not his source, Jacques Savary des Bruslons, Dictionaire Universel de Commerce, d’Historie Naturelle, d’Arts et Métiers (Paris: Jacques Estienne, 1723), as identified by Graham Pollard (cont. by Westher Potter), Early Bookbinding Manuals, Occasional Publication No. 18 (Oxford: Oxford Bibliographic Society, 1984), items 70 and 41. Dudin speaks of the needle as curved: “une aiguille d’acier de la grosseur d’une ligne, de trois pouces de longueur . . . & courbée en forme d’arc; sa courbure est de huit lignes ou environ” (p. 23). He depicts it as shown opposite, lying on the sewing table beside a ball of thread. (A curl of thread approaches the eye of the needle, but seems not to enter it.) If, for contrast, we imagine the heads of the quires lying to the left and with the left hand, say, accessing their central gutters (with the aid of dog-ears), we can appreciate the relative difficulty in sewing the Folger pattern of three-on from the back of the book kettle stitch from R to S loops forward around that from I to K at K

kettle stitch from I to K loops forward around that from A to B at B

back

front

kerf

S

K

kerf

B

176

random cloud

to the front. Let the rear endpaper be sewn all along, in six stitches, and 3I all along, in four (as on p. 138). For neither is dog-earing required. Then 3F, 3G, and 3H all need to be brought to the sewing frame together, for sewing has to begin, awkwardly, with the uppermost of these, 3F – and with no kerfs to stabilize the other two, lower, quires against the cords as we begin to sew. The left hand’s entry into the central gutters of the top two quires, 3F & 3G (for the first of two stitches in each, stitches 1 and 2) is facilitated by their dog-earing. But entry into the central gutter of the third and lowest quire, 3H (for stitches 3 and 4) has no such aid. Dog-earing the third quire as well would not solve this problem, however, for three dog-eared quires in a row would present ambiguous choices for a groping hand. (By contrast, when sewing starts from the front – as I have represented it in the body of the essay – the third quire of three to be sewn is brought to the frame separately by the very hand that immediately enters its central gutter for a single stitch or two consecutive ones, after which it leaves for good. In this procedure, the third quire of three is unambiguously identified for the gutter hand without a need for dog-earing. Sewing of the Folger copy was certainly from the front. Modern binding manuals sometimes advocate the use of something like a paper clip to yoke adjacent aft- and fore-quires. See one illustrated in Paula Rosati, Bookbinding Basics (New York: Sterling Publishing, Co., 2001), p. 45 (originally Riligatura: Tecnica, Idee, Progetti (Milan: RCS Libri, 1998). If the second had been temporarily so joined to the third, no trace of this technique would have survived in the bound book. (But if a clip had indeed joined the second and third quires, one wonders why another would not also have joined the other wing of the second to the first quire, and so have spared the book the disfigurement of any dog-ears.) In her discussion of two-on sewing, Edith Diehl (2: p. 128) has the binder keep her place in the centres of a quire by inserting a bone folder. [G. Martin], The Bookbinder’s Complete Instructor in All the Branches of Binding (Peterhead: P. Buchanan, 1823), p. 12 uses cardboard instead to guide the hand back to the centre of a quire in a book sewn two- or three-on. (I will underline his references to cardboard.) to sew on two or three sheets in once passing from head to tail, by taking one stitch in the first sheet laid down, and placing a bit of card in the middle [i.e., in the central gutter] of the sheet before the left hand is withdrawn in order

head

tail

Fearful Asymmetry

177

to find it again with readiness; then lay down a second sheet and make the second stitch in this, just as if the first had been continued, withdraw the hand from this, and place a card as before; then return to the first sheet. Margaret Lock illustrates sewing in Bookbinding Materials and Techniques 1700– 1920 (Toronto: The Canadian Bookbinders and Book Artists Guild, 2003), with something like a bean bag sitting in the middle of a quire, able to prop it open. Of course, all such means of facilitating repeated access to the appropriate gutter – folder, card, or bean bag – leave no signs in the bound book of having been employed. I am grateful to Nicholas Pickwoad for drawing my attention to the following historical reference to dog-earing in The Bookbinder’s Manual[, 1828], pp. 13–4 by the printer and supposed author, George Cowie. He describes sewing two-on on two chords from back to front, with the left hand working the gutters, two stitches in the first quire (sewn at tail and head) and one in the second (sewn at the middle). In the following excerpt and its footnote, I have underlined his cryptic references to folding down and turning down: dog-earing occurs on the frame in this instance, in alternate quires (or “sections”) after the first of the two stitches in each (not as, on pp. 142–3, I imagined it for the Folger copy – all its quires dog-eared before sewing, and tipping in of the plate, began). As Cowie’s account is difficult to visualize, I follow the excerpt with an expansion of it. My additions there are in italics; the numbers refer to my diagram (opposite), which shows the completed path of the thread for the first two quires. Cowie: [After the sewing of the fly-leaf all along from head to tail,] the first section of the work is then placed on the bands, ready for sewing ; the needle is first put through the ketch-stitch mark with the right hand, and drawn through with the left, and by the same hand put through close to the band on the side next the tail of the page; and, being drawn through with the right hand, with the left the corner of the top half of the section is folded down;* the next section is then taken up, and the band being placed in the saw-marks, the needle is put through on that side the band next the head of the book; it is then brought out on the left-hand side of the band, and passed through the section turned down on the right-hand side of the band, and brought out on the ketch-stitch mark of the same section . . . . ________________ *It is better, perhaps, to sew the first and last sections of a work entirely through [i.e., all along], and turn down the third and following. Cloud: “The first (kerfed) section or quire of the work, drawn from the end of the book, is then placed on the two bands, ready for sewing, this time from tail to head; (1) from the spine side, the needle is first put through the ketch-stitch mark at the tail of the page with the right hand, and drawn through into the gutter with the left, (2) and by the same left hand it is put back through to the outside, close to the first band on the side next the tail of the page; and, being drawn through to the outside with the right hand, and (3) with the left hand now leaving the gutter, the corner of the top half of this the first section is folded down (i.e., the fore-quire is dog-eared at its head toward the back of the book); the next (penultimate) section is then taken up, and the two bands being placed in the two middle saw-marks, (4) the needle is put by the right hand through the new section on that side of the first band next the head of the book (the thread having

178

random cloud

a

[A] E

B F

C G

D H ...

passed around the spine side of this band); (5) it is then brought out of this second section by the left hand on the left-hand side of the second band, and (6) passed by the right hand back to the left through the spine of the first section (whose corner had been turned down for the very purpose of guiding the left hand thither), on the right-hand side of the second band, thus completely encircling it, and finally (7) the needle is brought out to the right hand on the ketch-stitch mark of the same section. (8) The turned-down corner of the first section can now be straightened out before the sewing of the next pair of sections, this time from head to tail.” (In some books, dog-earing seems to have been accomplished by pinching the leaf merely at one edge of the ear, rather than by running a bead of even pressure along it. Folds that peter out seem good candidates for having been made on the frame.) Note that the thread passes merely over the back of the first band, but completely encircles the second. Since Cowie does not explain this difference, it may well be a mistake. However, whether the book is stitched all one way at the bands or all the other − or, as here, with a combination mold mold of the two − would not inhibit the functioning of the book when bound. Cowie’s footnote says to “turn down the [corner of the top half of the] third [section] and following”. He meant the third, fifth, seventh, etc. He neglects to advise unfolding the third before turning down the fifth . . . . According to Bernard Middleton, two-on sewing dates from the early seventeenth century to the late nineteenth: A History of English Craft Bookbinding Technique, 4th ed. (New Castle, DE: Oak Knoll Press; London: The British Library, 1996), p. 20. I find no one but Cowie who refers to dog-ears. 10. The same dog-earing pertains to a vellum Aldine Greek Homer, 1504 (HRC, University of Texas, Uzielli 64). As its quires are now sewn all along in a 19th-century binding, the dog-ears may be survivals of an earlier (but not likely original) two-on sewing, like that in the Spenser, with its more modern covers – or in the Boswell shown above. I suppose the bolts in Spenser diagrammed on p. 145 and those in Homer (but not those in Boswell) would have been long gone when the present sewings occurred. Atop this and the next page, I compare two Fisher copies of Boswell’s The Journal of a Tour to the Hebrides: a) of the first edition (B–10 5242), and b) the second (B–10 5243). Stitches and dog-ears in a resemble those features of the Spenser volumes, but it is sewn two-on on three cords, and the format is octavo. The binder entered each quire twice, and dog-earing facilitated each entry. In example b, the sewing is also two-on, but as in the 1732 Paradise Lost,

Fearful Asymmetry

179

b

B F

C G

D H

E I ...

one quire being entered once only (for adjacent stitches) needed no dog-earing (as in the Cowie example, in n. 9). Example a collates: [A]4 B–M8 (±M4) N–2K8 2L8 (–L8); example b: [a]4–b8 (–b7  F5 cancel) B–F8 (±1F5), G–2M8. A Fisher Library copy of Francis Bacon’s Tvvoo Bookes, 1605 (bac. B33 A38 1605: STC 1164), mapped below, has a structure reminiscent of the three-on sewing of the Folger Paradise Lost. Here, however, each dog-eared quire is folded aft, and separately. Also, the sewing of the second stitch in the first two quires of each of the two groups of three in this sample does not have a constant order. (I have arbitrarily chosen the direction of sewing for numbering stitches: “6, 1” could be “1, 6”.) 11. Ancient letters have arithmetic as well as phonetic functions. In Greek, “α, β, γ” are “1, 2, 3”. In Latin, “i” is “1”, “v” is “5”, “x” is “10”, “l” is “50”, “c” is “100”, “d” is “500”, and “m” is “1,000”. Every word in Hebrew has a number made by adding together the numerical value of each of its letters: alef is “1”, bet is “2”, gimel is “3”. Words are thus opened to mystical interpretation, and can, as it were, rhyme by number. In felt modern practice, algebra continues to represent quantities with letters: felt “x” and “y”, for example, are variables, though “π” and “e” are constants. Literate Greece marshalled its oral inheritance by naming each book of The Iliad and The Odyssey with one of the twenty-four letters of its alphabet. Epic came to seem “the alphabet writ large”. (Compare Psalms 25, 34, 37, 119, and the Lamentations of Jeremiah.) 12. Signature “W” is rare. Rebecca Niles has drawn my attention to two early examples in A Collection of Poems: viz. The Temple of Death: by the Marquis of Normanby, &c., 1701 and 1702 (ESTC T116471 and T124630; Fisher Library B–10 02150 and B–12 06881). These books sign with “I” and “U” but not with “J” and “V”. And thanks to Philip Oldfield for pointing out B. J. McMullin’s “Gatherings Signed “W”: A Footnote to Sayce”, Bibliographical Society of Australia and New Zealand Bulletin, vol. 6 (1982), which adds to examples in R. A. Sayce, Compositorial Practices and the Localization of Printed Books 1530–1800, Occasional Publication no. 13 (Oxford: Oxford Bibliographical Society, Bodleian Library, 1979). 13. “Quarto” derives from Latin “quattuor”, the number 4, and is cognate with Italic reflexes “quart” and “quarter” in English. There seems to be something fundamental to book-making about quarto format: French and Italian cognates, “cahier” and “quaderno”, having

180

random cloud

felt

mold

both lost their specific numerical associations, now mean merely “notebook”. Few can be the native speakers of English who associate “quire” with “four”; but indeed, Germanic “four” is cognate with this Romance term, though seeming to violate Grimm’s Law, which derives Germanic “f-” from IE *p- (as “flax” from IE *plek-, in n. 4). IE *kwetwor- is supposed to have mutated into something like *pekwor- before Grimm’s Law came into effect. This mutated form, the ancestor of Germanic “four”, does obey Grimm’s Law. 14. The Aldine Caesar, 1576 collates as follows: 6 8 6 8 8 8 8 8 10 2 A–I8 a–e8 Θ–3Θ8 * a–e f †–3† H–S–2H–S [–2[ §–2§ A–2S 2T

No logical order pertains among *, †, [, and §. Some may when Θ follows e: the second Greek e, eta (letter 7), is followed by theta. But isn’t there solid allegory when three crosses precede three “H–S” ligatures? Isn’t “H–S” part of the Christian “IHS”, in hoc signo (as ligature “&” is part of “&c”)? But no, “H–S” (or “IIS” or “llS” crossed) denotes a Roman sestertius (< semis tertius), as Stephanie Treloar and John Grant have taught me. (So it does in Aldus’s 1521 Suetonius, on C1r, C2v, etc.) 15. For no discernable reason, the rhythm also goes off in quire T. (See p. 137, second last row.) Hence, in searching out arrhythmia, I have started my projection from quire U. If, before sewing began, an error had been made at T in the count during dog-earing, and it was not noticed until much later, it might have been easier to make an adjustment when one finally came to sew at that point, rather than to flatten out every dog-ear after T and recommence the anapestic rhythm of folding. 16. The four leaves shown in the frontispiece are 2H2, 2H3, 2H4, and 2I1. (Models of quires 2H and 2I are in the previous opening: 2H(o) is mold; 2I(o) is felt.) Before sheet 2H was folded, pp. 236 and 240 were coplanar, as could easily have been deduced even with the sheet folded, as long as its bolts were still intact. The texture of mold and felt sides can be differentiated in modern books too. Consider two University of Toronto copies of the English Experience facsimile no. 95 of Florio His Firste Fruites, 1578 (STC 11096). The outer formes of the Centre for Reformation and Renaissance Studies copy are always mold. And so with the Robarts Library copy, except for quire 6 (Florio G3r– I2v). The above photograph of the gutter for C D I2v–I3r (between quires 6 and 7), shows felt on the verso, mold on the recto. Partially

Fearful Asymmetry

181

visible in the gutter, where it is somewhat pulled apart by glue applied to the spine during binding (which penetrated slightly between quires) are the code “EE 95” and, below it, two black squares, each about 3 mm tall, with half a millimetre between F G them. These figures march down the spines, one to a quire. Just prior to sewing, they allow for easy detection of missing or missequenced quires. Even if such a system of signing the spines had existed in 1732, its pattern would have been broken by the introduction of the cancel discussed on p. 150. G 17. The photographs are reproduced courtesy of the William Andrews Clark Memorial Library. 18. Often, a fold crosses 2C4v–2D1r; it is concave in McGill, convex in Fisher (see the diagram, opposite). Such can exist only on conjugate leaves, of course, which are not expectG G ed here. If the sheet from which the bifolium comes was printed twice with identical content (by work-and-turn), then roughly folded for pressing, the offset onto 2D1v observed in the Rylands copy would have come from the 2D1r on the other half G G of the sheet, before it was cut in two (on a more precise fold). 19. There are good (but wrong) reasons to fold the cancel, but not any other part of the book, inside out. Remote offsets make clear that Sheet G of the Humanities Research Center copy was once inside out. The interlocking remote-offset patH tern in the diagram, right, is typical of such inside-out folding. 20. On pp. 68 and 71 in the Fisher diagram and xcvii in the Brotherton, a hollow-headed arrow arising from the tip of a dark-headed arrow shows a right-image local offset from an adjacent mirror-image remote offset. (Such ricocheting argues that these diagrams combine evidence from different times, as from multiple pressings or beatings.) The various interconnections of pages from F, K, and 2N in these copies argue that cancels F3 and K5 were printed on the same octavo sheet as 2N6. That K5r sets off on 2N3v in one copy and on 2N6v in another argues that the sheet was folded in two different ways, inside out in Fisher and outside in in Brotherton. Folding the final sheet in order to locate the edges of the cancels prior to extracting them makes good sense, though it may seem odd also to press the whole sheet beforehand. But this practice does explain copies, like that of Scripps College, that bind the cancels in the last quire: the binder must have received the last sheet intact. (For an extensive study of the Bentley-Lloyd project, see my forthcoming essay, “The Librarynth”.) 21. McGill also divides between quires X and Y and between 2S and 2T. 22. A dog-eared corner of a 1r leaf (see the illustration) may now be too short for its tip to reach beyond the spine, to serve as a flag for restoring narrative order. But recall, from n. 3, that planing of the fore-edge and tail after binding reduces the dog-ears’ dimensions. The diagram of Folger remote offsets, p. 157, excludes facts about the cancel previously offered, on p. 150. (Of interest, if the two diagrams were integrated, would be the setting off of 2D1v on both 2C4r and Y1r: as in n. 20, different pile structures over time are thereby implied.) The Clark copy diagram, p. 159, also excludes previous information, from p. 149.

random cloud

cc 4

r

23. Caleb Stower, The Printer’s Grammar (London, 1808), pp. 407−8. 11  12  23, the number of letters in the Latin alphabet. See pp. 416−7 for Stower’s use of “booking”. He is not specific about how a number of once-folded gatherings for a single copy relate to each other physically – whether they are stacked one on top of another to make the booking, or are integrated in a single, once-folded pile. (For dog-earing related to ordering or beating sheets, rather than to sewing them, see n. 22.) Note my use of “signature” strictly to mean “symbols that sign a sheet or leaf”, not to mean “quire”. Nor do I use “gathering” for “quire”. Like Stower, I use it to mean a consolidation initially of at least some of the unfolded sheets for a volume, generally in alphabetic order, as B–L, for example. (The term might well continue to apply after a gathering had been folded.) In a larger work, like the 1732 Paradise Lost, a booking might normally require six gatherings. 24. Moxon lays out heaps in “Signatural succession”, A, B, C, but gathers in the order Z, Y, X, setting no limit to the number gathered in one stint (Joseph Moxon, Mechanick Exercises: or, the Doctrine of Handy-Works. Applied to the Art of Printing. The Second Volumne, London: Joseph Moxon, 1683[–4] (Wing M 3014), § 25, ¶¶ 2−3 (Of Laying the Heaps” and “Of Gathering of Books”); or see the modern facsimile edition, Mechanick Exercises on the Whole Art of Printing, Herbert Davis and Harry Carter, eds. (London: Oxford University Press, 1958, 1962 (2nd ed.)), p. 315. Sheets would have been made to “stand in Press about a Day and a Night”. This pressure would help flatten both the cockling that afflicts sheets (when the water added to them before presswork eventually dries) and the debossing of the paper by typeface during printing. Consider offset patterns in a gathering. Let the inner-forme pages of one sheet set off on the outer-forme pages of a sheet piled on it, and the inner-forme pages of the second set off on the outer-forme pages of a third: K1v ↔ L1r | L1v ↔ M1r. And, if L alone rotates 180°: K1v ↔ L3r | L3v ↔ M1r. A vivid example of such offsetting appears in UCLA copy 1 of Factorum et dictorum memorabilium by Valerius Maximus (Z 233 .A41 V235 1503). It is a Lyonese contrefaction [c. 1503] of a Aldine octavo, dated “October 1502” in its colophon. The imitation collates: π4 A–Z8, aa–cc8. All its quires are dog-eared forward in the lower aft quire, as in the model in the upper right atop p. 168.) Its quires are sewn two-on on five raised cords, with three non-adjacent stitches each. Offsetting throughout the contrefaction shows that, when the ink set off, all sheets except the half-sheet preliminaries were in alphabetical order, and L and O were rotated 180º with respect to the others. (In UCLA copy 2, H, I, and Z cc 5v

v cc 1

3

π

v

π1

3v CC

v

7 cc

π4

v

CC

6r

r

cc 8r

π2

182

r

r cc 2

Fearful Asymmetry

183

CC5v

π1v

v cc 1

π4 r

6 cc

cc 8r

r

π 3v

CC

π 2r

4r

were rotated as well as L and O.) That both formes of the preliminary half-sheet πset off on full-sheet cc(i) at the end of the volume indicates that the gathered full-sheets were not flat, but folded in half as a unit, with the outer formes generally facing outwards and with half-quire π unfolded “withinside” and setting off as shown in these two diagrams. This is an example of a more integrated format than that Stower advocates. For him, I suppose, the booking of a vast work, like Holinshed’s Chronicles, would have bundled all the separate gatherings, each folded in half, and not integrated them in a single once-folded pile, which, for so massive a work, would have been unstable and unwieldy. But booking the Valerius Maximus would have required only two of Stower’s half-alphabet gatherings, or three at most. A once-folded integration of these gatherings would have been only 1/4 of the 15 mm-thickness of the sewn book-block − not at all unwieldy. Now, the width of a once-folded pile grows at the fore-edge as a function of the number of sheets and of their thickness. See what I mean by contrasting the bottom fore-edges of two copies of Ian MacEwan’s On Chesil Beach (printed on paper stock of comparable thickness). These so-called “perfect-bound” books have no sewing to divulge the original quire structure; nor do they have visible signatures. But the original quiring can be deduced. The upper copy (Alfred A. Knopf Canada, 2007), opposite, has six quires (1–416 58 616); the lower (Vintage Canada Edition, Random House, 2008) has four (1–324 416). Numbering beside the photograph locates the centre of each quire. As only head, tail, and spine were ploughed, the fore-edges of the original bifolia of each quire continue to protrude more so as they are more central to the folded quire. Quires of the upper book are two-thirds the thickness of those below; and, accordingly, their projections are in the same proportion. In the production of the Valerius Maximus, intertextual offsets among the outer faces, A(o), of the once-folded bookings show that they were stacked and pressed with other copies of the same edition in alternating orientations, as shown here to the right. The middle booking (for UCLA copy 1) received offsets (represented in grey) from the two flanking ones, with A1r, A4v, A5r, and A8v of the top copy setting off on the same, outer, face of the middle copy and with A3r of the bottom copy setting off on the corresponding cc

v cc 3

7v

cc 2

r

π1r

π 2v

π 3r

cc 3 v

A1 r

A4v

cc 6 r

A 5r

A8v

A4v

A1r

A8v

A5r

A2

A3

cc 5 v

v

A6v

A7r

π 2r

A 3r

A 2v

A7 r

A6v

π3v

r

π 2r

π1v

cc 4 r

π3v

184

random cloud in another copy

in the other copy show-through headline

face of copy 1. (See the offsets of A1r and A8v of one copy on A1r of another in the above photo (reproduced by permission of the Regents of the University of California). In octavo, such offsets shift sideways as a function of the combined fore-edge protrusions of the two adjacent bookings. Sideways for folio too, but lengthwise for quarto. (The diagram, overleaf now, also shows a slight rotation in the pressing pile.) In theory, if only the title page of one of these works survived, we could estimate the number of sheets in the volume from this displacement, at about 1mm per sheet. That is, a gathering would show an average protrusion of 0.5mm for each of its bifolia (depending on the thickness of the paper, of course) – but this number is to be doubled because the two gatherings protrude in opposite directions. (I have yet to see evidence of different titles pressed together in this way, but the offsets from Psalms in Tasso and from Tasso in Milton, with which this essay concludes, suggest that it is only a matter of time until I do.) [P.S. Just found one.] 25. R. MacGeddon, “Hammered”, in Pete Langman, ed., Negotiating Jacobean Print Culture (Farnham, Surrey and Burlington, VT: Ashgate, 2011), pp. 136–99. 26. Diderot’s “a” is dwarfed by the other male figures in the plate (see p. 186): the engraver did not scale local spaces to a common standard implied by the architectural lines of convergence. Diderot does not extensively describe beating and he incorrectly states in his description of “Planche I” that the relieure is beaten, not the feuilles – the binding, not the sheets. Dudin (see n. 6) describes beating in great detail, specifying that a batté of six quires of 12mo would be struck (incredibly) a total of 160 overlapping blows, with one rearrangement, in which the exterior faces would go to the middle for half of the blows. His illustration of beating shown on p. 171, however, is as unconvincing in its details as that of the sewer beside him. His hammer has the wrong shape and is too small; and the face of the book is held perpendicular to the stone rather than laid down upon it. (The workman’s slack blow looks destined for his hand or the stone, not the batté.) Would Gladys Brochure also have beaten books? A love sonnet by Gianfrancesco Maia Materdona (Rime, Bologna, 1629, p. 25) says of the “Bella Libraia” who binds his verses: “Martella i fogli & il mio cor martella” (“She hammers the leaves and my heart she hammers”). This poem is cited in Marianne Tidcombe, Women Bookbinders, 1880–1920 (London: The British Library and New Castle, DE: Oak Knoll Press, 1996), pp. 15–6. Modern binders do routinely use a small hammer, but for rounding the back of the book-block: such a procedure is hardly beating “i fogli”. (But sure and all is metafare in love and poetry.)

Fearful Asymmetry

185

I assume Fisher and Clark copies were routinely beaten and pressed, but that remote offsets in Fisher resulted from pressing and in Clark from beating. The sequence of these means of flattening paper seems open to me: as pressing could occur several times in production, so too might have beating. Foot, p. 49, has folding of individual sheets followed by beating of a pile of them. In De Bray’s Onderwijs van ’t Boek-Binden (facsimile edition: Amsterdam: Nico Israel, 1977), B[1]r, beating and collation of a pile follow folding and pressing. (Unusually, De Bray signs the second leaf of his quires with the number “1”, as “B1”; the first is plain “B”.) The diagram below shows a rearrangement rightward to narrative order of just the subsections of beating 3 in the Clark copy. Contrast it with the diagram atop p. 155 of the reordering of the whole Fisher copy after it left the standing press. Both diagrams reveal that the paratext quires were clustered. (The sequence of Clark quires there anticipates the bound order more efficiently than does the Fisher sequence.) Below, I have added three dots in the left column of beating 3 to mark the end and beginning of the work as bound. Only the present diagram has a twoheaded arrow (on the left) connecting top and bottom: whereas the left side of the Fisher diagram represents the actual top and bottom of the pile, the two-headed arrow in the present diagram reminds us that, though Clark beating 3 must have had an original top and bottom (and also a subsequent temporary top and bottom, in the second phase of beating), their identity can now only be guessed at. My diagram below thus assigns these positions arbitrarily. In The Art of Bookbinding (London: G. Bell, 1880 [i.e., 1879]), pp. 9–10, the scrupulous William Zaensdorf warns that “it is advisable always to have a piece of paper at the top and bottom of the sections when beating, or the repeated concussion will glaze them”. But hammer-glazing is very frequently observed in old books. So, if a protective paper was not used for beating the Clark copy, I could possibly be less arbitrary: of the ten outer surfaces of the five subsections in beating 3, only four might appear glazed. These four would identify merely two candidate pairs for original top and bottom. If only one of their exterior recto (or verso) leaves should be dog-eared (as in n. 22), it would be a strong candidate for the original top (or bottom). Gaskell’s New Introduction appeared twice in 1972, from Clarendon and Oxford University Presses. It was a technical advance on McKerrow’s pioneering 1927 Introduction to Bibliography for Literary Students. As a simultanus reader of Gaskell’s two impressions on a collator, I learned of many variants, not just those listed in the “Preface to the Second Impression”. For example: all Gaskell says about beating is that after the folding and collating of sheets, “The folded book was then beaten flat with a hammer*, * several signatures at a time, in order to get it down to its proper thickness. Next the book was placed on the sewing frame” (pp. 147−8). We already know that Gaskell’s drew his illustration of bindery work uncritically from Diderot, but may not have realized (I certainly didn’t) that he excerpted only what, overleaf, I frame in white, presenting us with A a–b

B–Q

B–Q

a–b A

. . .

[R–

] –

–

186

random cloud

d

b a

c

merely the trochaic b–c, not the iambic a–b and c–d. His exclusion of hammer and press suggests why it has taken me four decades to think outside his tight box and begin to integrate the literary effects of a and d into textuality. Where, in the previous quotation (eight lines above), I have placed an asterisk, the second impression of A New Introduction silently added “and block”. That Gaskell states that one beats “with a hammer and block” (as he calls the stone) rather than “with a hammer on a block” suggests that, even in revision, he did not have beating in sharp focus as a practice in time and space. The same holds true for his statement that the quires were collated for beating, as beating also routinely entails rearrangement or “uncollating” under the hammer, as the remote offsets attest. Gaskell concludes, “next the book was placed on the sewing frame”. But next it was placed on the sewing table, whence it came to the frame only quire by quire. 27. Milton 83.k.23 and Tasso 80.h.17 are reproduced by permission of the British Library. (It has other copies of each work.) The reference to “Royal Paper” comes from an Advertisement of books “Lately Publish’d, printed by J. Tonson and J. Watts”

Fearful Asymmetry

187

that appears in the Library’s copy of Tonson’s 1724 printing of Longinus (837.1.35). Mere body language would suffice for identifications of the psalter pages, if only I knew the edition. Can you help me find it? ([email protected]) In the illustration opposite from Tasso 2L1v, one of the clearest, staves were evident; and the three lines of text seem to begin with “Que . . . un”, “plus”, and “Grand ROI, pour”. 28. Horace, Ars Poetica, ll. 148–9: “semper ad euentum festinat et in medias res / non secus ac notas auditorem rapit”. (This could be my epitaphI mean my epigraph The struggle for the text is the text.

© Random Cloud, 2012

STOP THE PRESS, JULIA. TELL NEIL. ¡Elephant ears (right)! Just spotted in Siena, in the Biblioteca Comunale degli Intronati copy of Bessarion’s 1516 In calumniatorem Platonis, an Aldine folio-in-4s (n.b., a format without bolts). They differentiate the wings of the quire and speed access to its centre. I don’t suppose these humungous folds are for the original binding . . . . But consider (below) this newer news, just in for a dog-eared 1514 quarto-in-8s Scriptores de re rustica in a contemporary binding – two centuries, therefore, before the Folger Milton. (Craig Kallendorf and Maria X. Wells report Anthony Hobson’s identification of this north-Italian binding in Aldine Press Books at the Harry Ransom Humanities Research Center, The University of Texas at Austin , Austin, 1998, item 109a.) Exposed kettle stitches show that sewing was from the front. There are three raised cords. And there is dog-earing of the upper outer corners of many aft quires toward the back of the book. In all the wicked world, this is the very earliest dog-damned intact original god-eared binding that I have ever found! (So far.) But consider this (curiouser and curioser), just in from Simon Fraser University: a copy of the 1515 Aldine Lucretius, an octavo. For the most part, alternate quires are dog-eared above AND below! − as I’ll indicate with these underlinings: *, a, b, c, d, e, f, g, h, i, k, l, m, n, o (below only), p, q, all such being safe to penetrate from the bottomSorry, Julia, no more changes.Promise.

8 ROGER CHARTIER AND PETER STALLYBRASS

What is a book?

Opus mechanicum and authorial text What is a book? This is the question that Kant raises in his 1797 The Metaphysics of Morals.1 He answers by distinguishing between the two natures of a book. On the one hand, a book is an “opus mechanicum,” a ¨ material (“korperlich”) object that has been produced by a mechanical art and that can be reproduced by anyone who is in rightful possession of the text. On the other hand, a book is a text addressed to the public by its author or by the publisher who has received a mandate from its author to speak in the author’s name. Without such an authorial “mandatum,” a book is unauthorized and constitutes an act of piracy. Participating in the eighteenth-century debate over the property rights of writers and publishers, Kant framed in legal and juridical terms the double nature of a book that had previously been expressed metaphorically. About 1680, Alonso V´ıctor de Paredes, a compositor and printer in Seville and Madrid, turned upside down the classical metaphor that described the human body or face as a book by conceptualizing the book as a human being: “I compare a book to the making of a man.” Both book and man have a rational soul (“anima racional”) and a body that should be elegant, handsome, and harmonious. The soul of the book is not only the text as it was imagined, written, or dictated by its author, the “buena doctrina”; it is ´ this text as it has been given a suitable presentation [“acertada disposicion”] by the publisher.2 But if the physical body of the book is the work of pressmen and bookbinders, its soul is not the author’s invention alone since it is co-created by the printers, compositors, and proof-readers who take care of the punctuation, spelling, and layout of the text. For Paredes, as later for Don McKenzie, “form affects meaning” and the “substantive essence” of a work cannot be separated from the “accidentals” of the printed text.3 If a book could be compared to a man, it was because God created human nature in the same manner that a printer makes a book. In 1675, Melchor de Cabrera Nunez ˜ de Guzman, a lawyer in the Royal Council of the King 188

What is a book?

189

of Spain, published a pamphlet which aimed to prove that printing was not a mechanical trade but a liberal art. It was because it was such an art that Cabrera recommended the renewal of the fiscal privileges and exemptions that had been granted to master printers, correctors, compositors, and pressmen.4 For Cabrera, mankind is one of the six books written by God. These books are the Heavens, compared to an immense chart of which the stars are the alphabet; the World, which is a universal library or “compendium” encompassing the entire Creation; the Book of Life in the form of a register containing all the names of Christ’s followers; Christ himself who is both the “exemplar” to be copied and an “exemplum” to be followed; and, the first book of all, the Virgin, whose creation was decided even before the creation of the World, in the Divine Mind. Man is the only book actually printed by God: “God put his image and seal on the press so that the copy would be true to its design” and “he desired to celebrate himself with a great number and variety of copies of his mysterious Original.” For Paredes describing his art, for Cabrera justifying printers’ privileges, and for Don Quixote visiting a printing-shop in Barcelona, textual production is a material process that involves specific places, machines, and workers. Between “the author’s genius and the capacity of the reader,” as Moxon wrote,5 a multiplicity of technical operations define a process in which the materiality of the text and the textuality of the material form cannot be separated.6 But even if many or most books that were printed were by dead authors, it was the problem of living writers that Paredes, Cabrera, and Moxon addressed. Paredes underlined the fact that the corrector “must understand the author’s intention in the text he sends to the ´ printing-house, not only for marking the right punctuation [la apuntuacion leg´ıtima], but also for checking if the author made some mistake [descuido] so as to advise him.” But for Cabrera, it was less that an author’s intentions had to be understood than that his careless errors required correction. Cabrera emphasizes that compositors “must take care of question marks, exclamation marks and parentheses because often the writer’s expression will become confused if these elements, which are necessary and important for the intelligibility and comprehension of what is written, are missing, and because if they are absent, the meaning is transformed.” In 1608, Hieronymus Hornschuch described a similar problem in his Orthotypographia, when he bitterly complained about the negligence of authors who give printers “faulty manuscripts, which cannot be read except with extreme difficulty.” “Therefore,” wrote Hornschuch, “I should like, not so much in the name of the correctors as of the printers, earnestly to advise and request all those who ever intend to publish anything in print to present it in such a way that the question need never be asked in the printer’s office which the comic slave [in Plautus’s Pseudolus] asked: ‘Do hens too have hands?’” Inverting the usual

190

roger chartier and peter stallybrass

distribution of roles, Hornschuch enjoins the author to imitate the printer’s attention to punctuation: “Almost most important of all, let [the author] punctuate his writing. For every day many mistakes are made by people on account of this; and in poetry nothing is more tiresome or blameworthy than the number of people who leave out punctuation marks . . . Moreover, correct punctuation produces true elegance and leads more than anything else to a clear understanding of the subject-matter, whereas inconsistent punctuation is the product of a disorderly mind.” The author is consequently told to send to the printing house not his “actual rough copy” but a copy “re-written as neatly as possible either by himself or his amanuensis on firm, non-absorbent paper, and checked again with the utmost care.”7 Such reciprocal relations between writer and printer are absent from Kant’s account of the authorial book. And in 1791, Johann Gottlieb Fichte redefined the book not on the basis of its materiality but, to the contrary, on its existence as a text. He framed a distinction, not only between the phys¨ ical (“korperlich”) and ideal (“geistig”) aspects of a book, but also, within the text itself, between the ideas and the form given to them by the author. The ideas are the material (“materiell”) aspect of the work, its content. Universal by their nature, ideas cannot be personal property. Ownership, he argued, can only derive from the form in which the ideas are expressed: “each individual has his own manner of thinking, his own way of forming concepts and connecting them . . . Hence, each writer must give his thoughts a specific form, and he can give them no other form than his own because he has no other. But neither should he be willing to hand over this form in making his thoughts public, for no one can appropriate his thoughts without thereby altering their form. The latter thus remains his exclusive property in perpetuity.”8 It was only by separating authorial style from any materiality, whether the materiality of the printed book or the “materiell” ideas, that a text could be conceptualized as individual property and owned as real estate. Changing definitions of the “authorial text” in the eighteenth and nineteenth centuries changed the material forms that scholars and editors privileged. In the early seventeenth century, the works of Shakespeare and Cervantes were defined by the collaborative practices of writers, scribes, compositors, and correctors. No one collected the “authorial” manuscripts of Hamlet or Don Quixote, because such drafts were only the starting place for a printed and bound book. It is a mere accident of history that Shakespeare’s authorial hand may survive in The Book of Sir Thomas More, a theatrical script containing at least six different hands. The hand of Shakespeare, if it is his, is the hand of a reviser, bringing an old play up to date. And why would one be interested in preserving a rough draft with lines such as the following:

What is a book?

191

all Shreiue moor moor more Shreue moore9

Such a form of writing required the intervention of other professionals, then as now, to add punctuation, regularize capitalization and, in the process, decide whether “more/more/moore” are different words or versions of the same word, whether the word[s] is a common or a proper noun (“more” or “[Sir Thomas] More”), and whether to set the word[s] in print as “more, more, more,” “more More, more,” “More, More, More” or any one of a number of possible formulations. In the Renaissance, whatever the appeals on title pages to the “True Originall Copies,” the printed book displaced the scribal copy, which itself displaced the writer’s rough draft.10 But in the new regime of authorship that was formulated by Kant and Fichte, the printed book lay at the furthest remove from the authority invested in the originality of a writer’s “style.” While that style was theoretically prior to any specific form of materialization, it was increasingly sought in the author’s own hand as it was preserved in his manuscripts. The archive of authorial manuscripts at the University of Texas would have been impossible – or perhaps we should rather say, pointless – before 1750. But during the eighteenth century, autograph drafts with erasures, corrections, alterations, and annotations began to be collected and preserved. In France, such manuscripts exist for Rousseau’s Nouvelle H´elo¨ıse, Diderot’s La religieuse, Laclos’ Les liaisons dangereuses, Bernardin de Saint-Pierre’s Paul et Virginie, and they include the twelve-meter roll of Sade’s Cent vingt journ´ees de Sodome. And authors self-consciously began to preserve their own manuscripts: Rousseau wanted to leave his Dialogues ou Rousseau juge de Jean Jacques to Notre-Dame but, finding the cathedral gates closed, he gave it to Condillac instead. He went on to make three other autograph copies of the Dialogues. Before the mid-eighteenth century, authorial manuscripts were preserved only for exceptional reasons. Brantome left the seven manuscript volumes ˆ of his Livre des dames to his heirs so that they could publish them after his death. The fragments of Pascal’s defense of Christianity were similarly gathered, transcribed, and put into order by the Messieurs de Port-Royal for a posthumous edition, the Pens´ees of 1669–70. And Montaigne’s autograph manuscript survives but only in the form of additions, and additions to additions, written in the wide margins of a luxury quarto of the 1588 edition of his own Essais (known today as the “exemplaire de Bordeaux”). Montaigne’s manuscript survives because the bound volume survives, whereas none of the manuscript drafts that preceded the first printed edition have been preserved.11 The preservation of authorial manuscripts in literary archives cannot be separated from the construction of philosophical, aesthetic, and juridical

192

roger chartier and peter stallybrass

categories that defined a new regime for the composition, publication, and appropriation of texts. The lawsuits in England following the 1710 Statute of Queen Anne created new relations between notions of individual authorship, originality, and copyright. Defenders of the traditional rights of London booksellers and printers, which had been undermined by the 1710 legislation limiting copyright to fourteen years, argued that ownership of the authorial manuscript implied a patrimonial right once the publisher had acquired it from the author. Hence, the author must now be attributed with a novel right: the ownership of his own composition. Defined by the fundamental and perpetual identity given to it by its author’s mind and hand, the literary work could now transcend any possible material embodiment. According to Blackstone, The identity of a literary composition consists intirely in the sentiment and the language; the same conceptions, cloathed in the same words, must necessarily be the same composition: and whatever method be taken of conveying that composition to the ear or the eye of another, by recital, by writing, or by printing, in any number of copies or at any period of time, it is always the identical work of the author which is so conveyed; and no other man can have a right to convey or transfer it without his consent, either tacitly or expressly given.12

For Diderot, every work is the legitimate property of its author because a work of literature is the irreducibly singular expression of that author’s thoughts and feelings. As he put it in his M´emoire sur le commerce de la librairie: “What property can a man own if a work of the mind – the unique fruit of his upbringing, his studies, his evenings, his age, his researches, his observations; if his finest hours, the most beautiful moments of his life; if his own thoughts, the feelings of his heart, the most precious part of himself, that which does not perish, which makes him immortal – does not belong to him?”13 The “property” was the work as composed by its author in its immaterial existence, “invisible and untangible [sic]” in the words of William Enfield in 1774. The nearest that one could come to materializing such an immaterial work was the trace left by the author’s hand.14 The autograph manuscript thus became the outward and visible sign of an inward, invisible state: the writer’s genius. And if the autograph manuscripts did not exist, as in the case of Shakespeare, it was increasingly necessary to invent them. As a result, forging autograph manuscripts became a major enterprise. In February 1795, William Henry Ireland exhibited several recently discovered Shakespeare manuscripts. They included the autograph manuscripts of King Lear and of two unknown plays, Henry II and Vortigern and Rowena (performed once at Drury Lane Theatre on April 2); the letters exchanged by the poet and his patron, Southampton; a very Protestant

What is a book?

193

Profession of faith; and a letter addressed to Shakespeare by Queen Elizabeth. When the documents were published in December 1795 under the title of an Authentic Account of the Shakespearian MSS., Edmond Malone was the first to expose Ireland’s forgeries by comparing the handwriting and spelling of the forged documents to authentic ones. His meticulous exposure of the imposture was published with the significant title of An Inquiry into the Authenticity of Certain Miscellanies Papers and Legal Instruments. Published Dec. 2 MDCCXCV, and Attributed to Shakespeare, Queen Elizabeth, and Henry, Earl of Southampton: Illustrated by Fac-similes of the Genuine Hand-writing of that Nobleman, and of Her Majesty; A New Fac-simile of the Handwriting of Shakespeare, Never before Exhibited; and Other Authentick Documents.15 “Genuine Hand-writing” was the necessary material trace of an immaterial genius. Where Dante had imagined the writer’s scattered traces finally materialized in the form of a bound book, the new regime saw the material book as a more or less corrupted form. The editor’s job was to look through or behind the book so as to discover the scattered traces prior to their “contamination” by previous editors, publishers, compositors, correctors, and binders. A unique hand, contained in unique manuscripts, could now displace the multiple scribal or printed books that are the only ways through which the works of Dante, Cervantes, and Shakespeare have survived.

Scattered and gathered leaves In the third book of the Aeneid, Virgil wrote about the failure of historical memory when, with nothing to bind them together, the leaves of the Sibyl’s verses were scattered by the wind: What she commits to Leafs, in order laid, Before the Caverns Entrance are display’d: Unmov’d they lie, but if a Blast of Wind Without, or Vapours issue from behind, The Leafs are born aloft in liquid Air, And she resumes no more her Museful Care: Nor gathers from the Rocks her scatter’d Verse; Nor sets in order what the Winds disperse.16

The Sibyl’s scattered leaves were a recurring topos from Late Antiquity. Petrarch imagined his Rime Sparse, or scattered rhymes, as similarly cast to the winds of fortune, even as he helped to establish new forms of the book in which they would be gathered together as the materials of immortality. But Virgil’s account of the Sibyl’s leaves had taken on a radically new meaning by the time that Petrarch was writing. For Virgil, the Sibyl’s leaves

194

roger chartier and peter stallybrass

could not be the leaves of a book for the simple reason that the “books” he read and on which his own poem was preserved were scrolls. Although scrolls are glued together in segments (membrana), they do not have leaves. Folium, a leaf, only took on the new meaning of the leaf of a book in the fourth century ce – by which time the codex was rapidly becoming the dominant material support for pagan texts, as it had been for Christians from the beginning of the second century, if not earlier. Far from being a dead metaphor, the Sibyl’s leaves could now be reinterpreted as the constituent elements of a book. The leaves of parchment or paper on which an author wrote might indeed be literally scattered. But the folia or leaves could be gathered together and preserved in what would later be called a folio – the form of book that would best preserve a work for posterity. Just as the limbs of Shakespeare’s scattered body were brought together in the First Folio, the textual leaves that time had scattered could be brought together in the material form of a book. Dante made the gathering of scattered leaves the telos of his Commedia, conceiving of Paradise itself as being organized like a book. Paradise so overwhelms him that his “memory fails”; indeed his “vision almost wholly fades away” just as, “in the wind, on the light leaves, the Sibyl’s oracles were lost” [cos`ı al vento ne le foglie levi/ si perdea la sentenza di Sibilla]. But all is not lost. For in Paradise, Dante sees “gathered together, bound by love into a single volume, leaves that lie scattered through the universe”: Nel suo profundo vidi che s’interna legato con amore in un volume, cio` che per l’universo si squaderna.17

Paradise is preserved, as Dante’s poem about it will be, because the “scattered” leaves have been “gathered together” and bound. According to Boccaccio, Dante published his poem in a series of pamphlets [quadernetti] of six to eight cantos but the final pamphlet was unpublished at his death. If Dante wrote about a gathering together of scattered fragments in Paradise, the gathering together of his own leaves was a dream that had not yet come true. Perhaps the very precariousness of his poetic gatherings led him to envisage Paradise all the more powerfully as a bound book. Dante’s word for binding together, legato, is derived from the Latin legere, which means both to gather and to read. Gathered together, the leaves that had previously been scattered become legible or readable. Equally specific is the word that Dante uses to describe the scattering that precedes the gathering: squaderna. The verb squaderna suggests the active process of ungathering the quires [quaderni] out of which a bound book [volume] is composed. A quaderno was literally the four sheets of parchment or paper that, folded together, make a folio gathering of eight

What is a book?

195

leaves or sixteen pages – either a small pamphlet or one of the units out of which a book is composed. The prefix “s” in squaderna negates the meaning of “quaderna,” like the prefix “un-” in English. “Unquadernoed,” not even held together by the provisional sewing that creates a pamphlet, the separate leaves will be scattered by the wind unless they are brought together in un volume, “in a single volume.” One way of defining an author is as someone who is bound with him or her self. Such binding of oneself with oneself was only occasionally negotiated by a living author in Medieval and Renaissance Europe. Petrarch undoubtedly thought of himself as an author, but his I Trionfi circulated at first in miscellanies as often as in a single volume. During the second half of the fourteenth century, however, 63 percent of the 248 surviving manuscripts of I Trionfi appeared in volumes containing only Petrarch’s work and the percentage increased to 74 percent of 78 manuscripts in the fifteenth century.18 The material “authoring” of Petrarch as a canonical figure was as much the work of readers, editors, and booksellers after his death as of his own literary project. In the manuscript publication of Dante, Petrarch, and Boccaccio in Italy and of Christine de Pizan and Ren´e d’Anjou in France, a select group of vernacular writers were canonized by bringing together in book form a text or series of texts under the proper name of a single author.19 When Antonio Gramsci named his prison notebooks Quaderni, he was registering the distance between the tenuous status of his writing – provisional, revisable, subject to additions and subtractions – and the authorial book.20 In the twentieth century, quaderni took on a new significance as the material form of what Deleuze and Guattari call a “minor literature”21 – a literature that resisted the monumentalizing pretensions of the classical and that took as its preferred forms the notebook, the diary, letters, essays, fragments – the material forms that Kafka left to posterity with the instructions that they should be burned. But by the twentieth century, even Kafka, whatever his hesitations and resistances, had already been bound in a volume (although not together with himself) before he died. For Dante, by contrast, the gathering of his work was not the preordained destiny of authorship. He had to work to establish the conditions in which his quaderni would be preserved from the winds of time that had scattered the Sibyl’s leaves. Books versus pamphlets In 1623, the editors of the First Folio of Shakespeare’s plays brought together the “remains” of Shakespeare in a single large volume so as to repair the author’s body, his corpus. Before they gathered up his scattered limbs into a bound book, Shakespeare had survived, they claimed, only in “diuerse stolne, and surreptitious copies, maimed, and deformed.” They were now

196

roger chartier and peter stallybrass

preserving Shakespeare’s “owne writings,” “offer’d to [the reader’s] view cur’d, and perfect of their limbes.”22 It was, of course, in Heminges’s and Condell’s interest to emphasize just how bad all previous editions of Shakespeare had been. But one thing they were certainly right about: the pamphlets that were the material form in which Shakespeare circulated during his lifetime were peculiarly subject to the ravages of time. Their leaves were literally scattered, title pages and final quires coming free from the rough stabstitching that only temporarily held them together. These pamphlets were indeed “maimed, and deformed.” The First Folio “cur’d” Shakespeare’s limbs by securely sewing the folded sheets of his plays, quire by quire, onto cords and protecting them with wooden boards, covered in calfskin. The book form has played a particularly important role in materializing authorship. It is difficult to think of a more famous author today than Shakespeare. One can buy his texts, collected or in individual volumes, in an extraordinary variety of bindings. Yet in his own lifetime, Shakespeare did not write a single text that was designed to be bound as a book. His poems were written as pamphlets, his plays for performance and secondarily for publication as pamphlets. Dedicating The Rape of Lucrece to the Earl of Southampton, Shakespeare wrote: “The love I dedicate to your Lordship is without end: wherof this Pamphlet without beginning is but a superfluous Moity.”23 Shakespeare’s love may be “without end”; his “Pamphlet” is not, as the fate of most of the copies (and even of whole editions) of his pamphlets shows. That his poems and plays were usually sold as pamphlets is shown both by the loss of all but a few copies and by the fact that many of those remaining copies are defective. But there is more concrete evidence that Shakespeare’s plays were only exceptionally sold as bound books. If one looks carefully at the surviving quartos, elegantly bound later when they had become a valuable collector’s item, one can find in copy after copy the tell-tale traces, often disguised by later repairers, of small holes, stabbed through the leaves a centimeter or two in from the gutter. These holes show that the plays were first sold stab-stitched as pamphlets. So rare are these unbound pamphlets now that we have only seen a single example of a Shakespeare quarto that survives in its “original” state. Even that pamphlet is a composite, made up of a 1609 copy of Pericles and a 1606 copy of Samuel Daniel’s The Queene’s Arcadia. Given that the plays have different publishers (and printers), the pamphlet was probably made up by a bookseller trying to get rid of an unsold play by Daniel by adding it three years later to the hugely popular Pericles. From the perspective of 1623, it seems inevitable that Shakespeare-the-dramatist is “not of an age, but for all time.” It is, after all, materially true. The Folger Library alone has more than eighty copies of the First Folio. Not surprisingly, given how many copies survive, the First Folio (and the millions of copies of the “works” that have succeeded it) conditions how we think about Shakespeare as an author.

What is a book?

197

By contrast, the first edition of Venus and Adonis (by far his most popular work in his lifetime) survives in a single copy in the Bodleian Library and the first edition of The Passionate Pilgrim only survives as a fragment, without a title page. Thus, it is only from the title page of the second edition that we learn that the poems in The Passionate Pilgrim are “by W. Shakespeare,” although only five of the twenty poems are now attributed to him. Unbound pamphlets are not the materials of immortality, whatever a writer’s claims. Shakespeare may have wanted to be a literary author, as Lukas Erne has argued, but he did not succeed in materially becoming one during his lifetime either in terms of published books or in terms of collectors’ compilations.24 It was only in 1619, three years after his death, that the first attempt was made to publish Shakespeare in the form of a bound volume. In that year, Thomas Pavier published ten “Shakespeare” quartos. Although each play had a separate title page (most of them with false information about printing and date of publication), the first three plays were printed with continuous signatures. It is clear that Pavier intended sell the ten plays bound together, even if he could also market them as separate titles.25 While the 1619 “Shakespeare” volume is radically different from the 1623 “Shakespeare” of the First Folio, they both have as their organizing principle a single author: the professional playwright, William Shakespeare. Pavier’s project was, indeed, the first serious attempt to materialize Shakespeare as an author in the form of a bound book. Prior to Pavier, Shakespeare’s writing was published in the perishable and disposable form of pamphlets. Inexpensive pamphlets, however, do not stand in any simple opposition to canonical books. Shakespeare’s canonization in the First Folio was by no means the end of Shakespeare the pamphleteer. To the contrary, publishers continued to produce Shakespearian pamphlets in the form of theatrical editions of individual plays. And while the First Folio was the beginning of an attempt to establish the authoritative works, single-play pamphlets and books produced radically new versions of Shakespeare, from Nahum Tate’s 1681 adaptation of King Lear to give it a happy ending to Richard Curtis’s Skinhead Hamlet, in which the hero’s dying words are “I’m fucked. The rest is fucking silence.”26 From 1600, Shakespeare’s plays were being mined for detachable lines and speeches that circulated in printed and manuscript commonplace books.27 The fragmentation of “works” into fragmentary selections was already common practice in the case of the canonical Bible. Throughout the late Middle Ages and the Renaissance, books of hours and psalters massively outsold Bibles. In 1643, the Bible was published in the form of a sixteenpage pamphlet: The Souldiers Pocket Bible. In addition to being cheap, this pamphlet had two other advantages, both of which were advertised on its title page: it “reduced [the Scriptures] to severall heads, and fitly applyed [them] to the souldiers severall occasions, and so may supply the want

198

roger chartier and peter stallybrass

of the whole Bible”; and it was small and light enough for a soldier to “conveniently carry about him.”28 In this case, it is the imagined reader, not any concept of a divine or inspired author, that determines both the textual selections and the material form of a canonical text. But it the case of the Bible and Shakespeare alike, the proliferating fragments repeatedly referred back to the canonical books that authorized them: “the whole Bible”; the First Folio or, later, The Works of William Shakespeare. But the material and ideological problems of constructing such canonical books is nowhere more manifest than in the case of the Bible. Making the canonical book How did the Bible become a canonical book in the first place? Most of the texts that compose the Hebrew Bible were recited orally long before they were written down and repeatedly redacted and rewritten. Yet “Bible,” the word that we now commonly use for both the Jewish and the Christian scriptures, first referred neither to the texts of the scriptures nor to their divine authorization. It is derived from the Greek biblia, the plural form of biblion (a diminutive form of biblos), meaning the inner bark of the papyrus reed. “Bible” thus referred to the material support on which any oral text could be transcribed. If biblia first referred to the sheets of papyrus, glued together to make up a scroll, the word was increasingly transferred from the material surface to the written scrolls. But it is important to stress that what we now call “the Bible” was, when first transliterated from Greek into Latin, a neuter plural: “scrolls.” Even in the fourth century ce, St. Jerome, whose biblia was in the form of codices, did not think of the scriptures in the singular. He referred to them as a bibliotecha, a word derived from the Greek for a container of scrolls and meaning, by extension, a library or collection of scrolls. But by the ninth century ce, biblia began to be used as a feminine singular to mean not “scrolls” or “books” but “the book.” The changes in the word’s meaning inscribe the history of a material technology. Before biblia was used in the singular, the Scriptures, previously written on multiple scrolls or divided up into small codices, had been gathered together into massive codices, known as pandects, single volumes that defined what did and, by exclusion, what did not count as divine scripture.29 If we move from the semantic transformations of “biblia” (from papyrus to scrolls to scriptures to “the Bible”) and turn to how the scriptural texts, once selected, were organized, we see yet again the importance of the material form in the shaping and defining of Scripture. Before the massive scroll of the Pentateuch was made, each “book” of the Scriptures usually corresponded to a single scroll. If the texts were too long to fit onto a single scroll, they would be written on two scrolls. Hence First Samuel and

What is a book?

199

Second Samuel; First Kings and Second Kings; First Chronicles and Second Chronicles. The division between the two scrolls, while it could be given a textual significance, was the result of a simple material necessity. Vice versa, if the texts were very short, it made sense to copy several onto a single scroll. Hence, the Twelve Minor (i.e. “short”) Prophets, contained on a single scroll. (The first such scroll of the Twelve Prophets at Qumran dates from the mid-second century bce.) Given that most of the biblical books of the Bible do not follow on from each other textually, as do, for instance, the consecutive books (i.e. scrolls) of the Iliad, the sequence of the scrolls is not at all clear, and there was considerable Rabbinic debate about the different ways in which the “Writings” could be ordered, long after the Pentateuch scroll had given a definitive order to the first five books of the Bible. In a codex that contains the whole Christian Bible, on the other hand, the sequence in which the books were ordered had striking ideological effects. Perhaps the most obvious example of this is the distinction between Catholic and Protestant Bibles. In the latter, fourteen books or parts of books that had been scattered through the Catholic Old Testament were extracted and either omitted altogether or cordoned off between the Old and the New Testaments as the Apocrypha. Even when it is not a question of what is included and excluded, the sequence of books is strikingly different in different versions of the Bible. St. Jerome’s Bible, like the King James Bible, separates out the “apocrypha.” But the sequence of books in the two Bibles is different: Esther is the seventeenth of thirty-nine Old Testament books in the King James Bible; it is the last book of the Old Testament in St. Jerome’s version. Far more striking is the different positions of the Apocalypse (Revelation) in the two Bibles. To Protestants (and, indeed, to most later Catholics), it was “obvious” that the Bible began with the beginning of the world (Genesis) and ended with the end of the world (Revelation). But although Jerome began with Genesis, his Bible ended with St. Paul’s Epistle to the Hebrews, since he put all the Pauline epistles after Revelation. Although the sequencing of the Jewish scriptures was a topic of debate long before the invention of the codex, the material form of the latter gave a new theological significance to the sequences of the biblical books, which, when written on multiple scrolls, were always potentially miscellaneous in their ordering. Above all, the pandect, whether as a massive sixth-century codex or as a thirteenthcentury complete pocket Bible, gave a new material sense of the Christian Bible as a single canonical work. Conclusion What is a book? In this essay, we have moved between analyzing the ways in which material books construct literary works and the ways in which

200

roger chartier and peter stallybrass

literary works are transformed into a variety of material forms. At the same time, we have emphasized that the book, understood as a codex, is related in multiple and complex ways to the formation of what Michel Foucault calls the “author function.” The cult of the author and of aesthetic “originality” presuppose that authors make books by writing their own works. So it is important to recall that the opposite is often the case: books create authors when scribes, editors, or publishers bring together a range of texts under an authorial name. Moses, once he had been posited as the author of the Pentateuch, was retrospectively defined by those very texts. On the other hand the name of an author (Cervantes) or a title (Don Quixote) can, as we have seen, be attached to an endless variety of different material forms. Genetic criticism is only made possible by the existence of the different stages of the authorial process: outlines and sketches of the work, notes and documents, drafts, corrected proofs. However, the literary archive is itself being transformed by digital photography and the internet. The unique authorial drafts of Walt Whitman’s Leaves of Grass, scattered in unique geographical locations from the Huntington Library in California to the University of Pennsylvania to the British Library in London are now gathered together electronically in the Whitman Archive, and can be seen and read by anyone who has access to a computer and the internet. At the same time, genetic critics have begun to show how such archives testify to the multiple hands and agencies that go into the composition of a text. Sometimes, different agencies can be the work of a single hand. The Whitman Archive reveals the poet as compositor, marking up one manuscript for the printer in red ink with instructions to “# leave a blank space,” “indent 3 ems,” use a “one em dash” and a “2 line A.”30 And three separate copies of the printed proofs of Whitman’s “Shakspere-Bacon’s Cipher” show the author as corrector, in each copy emending “anthor” to “author” in purple crayon. The Dickinson Electronic Archive shows the different hands that entered into the composition and correction of Dickinson’s poems, and it includes the critique of one of Emily’s poems by her sister-in-law Sue that prompted Emily to compose at least two completely new versions of a single stanza.31 If the literary archive emerged to enshrine the single hand of the author, the electronic archive is increasingly being used to explore the range of hands that transform a text as it moves from notes to rough draft, from rough draft to printer’s copy, from newspaper or magazine to book. As we have tried to show, the codex has always been only one of the material forms in which texts circulated. But digital technology opens up new possibilities of polyphonic and palimpsestic writing, freed from any proper name or any unified concept of “the work.” To understand the new digital technology, we need to interrogate or abandon the classical definitions of the literary work,

What is a book?

201

which were themselves the basis for specific forms of intellectual property. At the same time, the book, in the material form of a codex, is revealed as only one of the many material forms that structure how we write and how we read. David Kastan has recently characterized the view that a work transcends any possible material incarnation as “platonic,” in contrast to the “pragmatic” view that no text exists apart from the material forms and practices through which it is read or heard.32 These contradictory conceptions of the text divide both literary criticism and editorial practice, opposing those who attempt to restore the author’s immaterial intentions by repairing the “wounds” inflicted in the course of copying the work and setting it in type to those who argue that we must respect the irreducible diversity of a work’s historical states. A similar tension between the immateriality of texts and the materiality of books characterizes the relations between readers and the texts that they appropriate. In a lecture delivered in 1978 on “The Book,” Jorge Luis Borges said: “One day I thought I would write a history of the book.” But he added: “I am not interested in the physical aspect of books . . . but in the various ways in which books have been considered.”33 For Borges, then, the material form of a text is irrelevant; what counts is how a “book,” meaning the abstract literary work, regardless of its form, has been understood. Borges, in other words, took the “platonic” view. But when Borges dictated his autobiography to Norman Thomas di Giovanni, he spoke quite differently about one of the books that had shaped his life: Don Quixote. Now, he vividly recalls the specific object that he had held in his hands and read: I still remember the red bindings with gilt titles of the Garnier edition. Eventually, after my father’s library was broken up and I read Don Quixote in another edition, I had the feeling that it was not the real Don Quixote. Later, a friend obtained for me a Garnier copy with the same engravings, the same footnotes and the same errata. For me, all these things were part of the book; in my mind, this was the real Don Quixote.34

For Borges, the “real” Don Quixote took the irreducible material form of an illustrated edition that had been imported from France, the edition that he had read as a boy. The platonic view of the text is displaced by the visceral force of memory. Borges’s contradictory views suggest the problem with any single answer to the question, “what is a book?” A work can be read only in some specific material form. Yet in the West, Neo-Platonism, Kantian aesthetics, and a new definition of literary property have all contributed to the construction of an ideal authorial text that is prior to any of the forms in which it is materialized. This ideal text, however, is only made possible by the invention

202

roger chartier and peter stallybrass

of literary archives – archives that dissolve the literary work in the form of a bound book into a multiplicity of fragmentary notes and drafts. The question that remains is whether the proliferating materials that are now brought together and made available through electronic archives will be used to recreate the platonic ideal of the authorial book or to explore the multiple human agencies and material forms that have shaped the literary text and will continue to reshape it in new and unforeseeable ways. NOTES 1 Immanuel Kant, The Metaphysics of Morals, trans. Mary J. Gregor (Cambridge University Press, 1996). ´ y origen del arte de la imprenta y Regla 2 Alonso V´ıctor de Paredes, Institucion generales para los componedores, Edicion de Jaime Moll (Madrid: El ´ y prologo ´ Crotalon, ´ 1984) [re-ed. Biblioteca Litterae, 2002], f. 44v. 3 D. F. McKenzie, Making Meaning: ‘Printers of the Mind’ and Other Essays, ed. Peter D. McDonald and Michael F. Suarez, S. J. (University of Massachusetts Press, 2002), particularly “Typography and Meaning: the Case of William Congreve,” 198–236. ´ 4 Melchor de Cabrera Nunez y pol´ıtico en ˜ de Guzman, Discurso legal, historico prueba del origen, progressos, utilidad, nobleza y excelencias del Arte de la Imprenta (Madrid, 1675). 5 Joseph Moxon, Mechanick Exercises on the Whole Art of Printing (1683–4), ed. Herbert Davis and Harry Carter (Oxford University Press, 1958), 311–12. 6 Margreta de Grazia and Peter Stallybrass, “The Materiality of the Shakespearian Text,” Shakespeare Quarterly, 44.3 (1993), 255–83. 7 Orthotypographia, by Hieronymus Hornschuch: A Facsimile with a Parallel Translation of the Earliest Printers Manual. First published at Leipzig in 1608, ed. P. Gashell and P. Bradford. University Library Publications (Cambridge University Press, 1972), 29ff. ¨ ¨ Ein 8 Johann Gottlieb Fichte, Beweis der Unrechtmassigkeit der Buchernadrucks: ¨ Rasonnement und eine Parabel, 1791. Fragments of Fichte’s essay are translated and commented on by Martha Woodmansee, The Author, Art, and the Market: Rereading the History of Aesthetics (Columbia University Press, 1994), 51–3. 9 The Book of Sir Thomas More, ed. W. W. Greg (Oxford: The Malone Society, 1961), 75. 10 Scribal copies can include manuscripts made by an author for presentation or sale. For Thomas Middleton’s role as a scribe in the manuscript circulation of A Game at Chess, see T. H. Howard-Hill, “The Author as Scribe or Reviser?: Middleton’s Intentions in A Game at Chess,” TEXT, 3 (1987), 305–18. 11 George Hoffmann, Montaigne’s Career (Oxford: Clarendon Press, 1998), 97– 107. 12 William Blackstone, Commentaries on the Laws of England (Oxford, 1765–9), qtd. in Mark Rose, Authors and Owners: The Invention of Copyright (Harvard University Press, 1993), 89–90. 13 Denis Diderot, Oeuvres compl`etes, Vol. viii, Encyclop´edie iv (Lettres M–Z). Lettre sur le commerce de la librairie, ed. John Lough and Jacques Proust (Paris: Hermann, 1976), 509–10.

What is a book?

203

14 William Enfield, Observations on Literary Property (London: Joseph Johnson, 1774), 11. 15 Samuel Schoenbaum, Shakespeare’s Lives (Oxford University Press, 1970), 193–223, and Margreta de Grazia, Shakespeare Verbatim: The Reproduction of Authenticity and the 1790 Apparatus (Oxford: Clarendon Press, 1991), 107–9. 16 John Dryden’s translation of Virgil, Aeneid, 3.445–52 in The Works of Virgil (London: Jacob Tonson, 1697), 284. 17 Dante, Paradiso, in The Divine Comedy, ed. and trans. Charles S. Singleton (Princeton University Press, 1977), 33.55–66 and 33.85–7. Our account of this passage is deeply indebted to John Ahern, “Binding the Book: Hermeneutics and Manuscript Book Production in Paradiso 33,” PMLA, 97.5 (October 1982), 800–9. 18 Gemma Guerrini, “Il Sistema di Comunicazione di un Corpus di Manoscritti Quattrocenteschi: I Trionfi del Petrarca,” Scrittura e Civilta` 10 (1986), 121–97. 19 See Ahern, “Binding the Book,” 800–9. 20 See Joseph A. Buttigieg, “Philology and Politics: Returning to the Text of Antonio Gramsci’s Prison Notebooks,” boundary 2, 21.2 (Summer 1994), 98–138. 21 Gilles Deleuze and F´elix Guattari, Kafka: Toward a Minor Literature, trans. Dana Polan (University of Minnesota Press, 1985). 22 Mr. William Shakespeares Comedies, Histories, & Tragedies Published According to the True Originall Copies, (London: William Jaggard et al., 1623), sig. a3. 23 See Roger Chartier and Peter Stallybrass, “Reading and Authorship: The Circulation of Shakespeare 1590–1619,” in A Concise Companion to Shakespeare and the Text, ed. Andrew Murphy (Oxford: Blackwell, 2007), 35–56. 24 See Lukas Erne, Shakespeare as Literary Dramatist (Cambridge University Press, 2003). 25 Andrew Murphy, Shakespeare in Print: A History and Chronology of Shakespeare Publishing (Cambridge University Press, 2003), 255–8. 26 Richard Curtis, The Skinhead Hamlet, http://sub-zero.mit.edu/bakunin/hamlet. html, accessed February 10, 2009. 27 See Chartier and Stallybrass, “Reading and Authorship.” 28 The Souldiers Pocket Bible: containing the most (if not all) those places contained in holy Scripture, which doe shew the qualifications of his inner man, that is a fit souldier to fight the Lords battels, both before he fight, in the fight, and after the fight; which Scriptures are reduced to severall heads, and fitly applyed to the souldiers severall occasions, and so may supply the want of the whole Bible, which a souldier cannot conveniently carry about him: and may bee also usefull for any Christian to meditate upon, now in this miserable time of warre (London, G. C., 1643). 29 On the Bible as a codex, see Robert A. Kraft, “The Codex and Canon Consciousness,” in The Canon Debate, ed. Lee Martin McDonald and James A. Sanders (Peabody, MA: Hendrikson, 2002), 229–33; Karel van der Toorn, Scribal Culture and the Making of the Hebrew Bible (Harvard University Press, 2007); Harry Y. Gamble, Books and Readers in the Early Church: A History of Early Christian Texts (Yale University Press, 1997); Anthony Grafton and Megan Williams, Christianity and the Transformation of the Book: Origen, Eusebius, and the

204

30

31 32 33 34

roger chartier and peter stallybrass

Library of Caesarea (Harvard University Press, 2006). “Scripture” itself at first meant simply material forms of inscription: scriptura or “writing.” Walt Whitman, proof copies of “Shakspere-Bacon’s Cipher,” two at the Harry Ransom Humanities Research Center, University of Texas at Austin, and one at Temple University. See www.whitmanarchive.org/manuscripts/finding aids/U Texas and www.whitmanarchive.org/manuscripts/finding aids/Temple. Accessed April 22, 2009. See Martha Nell Smith and Lara Vetter, “Emily Dickinson Writing a Poem” at www.archive.emilydickinson.org/safe/index.html. Accessed April 22, 2009. David Scott Kastan, Shakespeare and the Book (Cambridge University Press, 2000), 117–18. Jorge Luis Borges, “El libro,” in Borges oral, (Madrid: Alianza Editorial, 1998) 9–23, at 10. Jorge Luis Borges with Norman Thomas de Giovanni, Autobiograph´ıa 1899– 1970 (Buenos Aires: El Ateneo, 1999), 26.

9 JOHN D. NILES

Orality

For at least the past 150 years, textual scholarship (often conceptualized as “textual criticism” or simply “philology”) has been a foundational aspect of literary studies, even if controversies have arisen as to what constitutes its best aims and practices. Is there such a thing, though, as “oral textual scholarship”? And if so, of what does it consist? Questions of this kind have not often been posed. Partly as a consequence, the presentation of “oral texts” in modern printed editions differs dramatically from one book to another, while the manner in which such texts are mediated to the public is often neither explained nor justified. Such opacity regarding basic matters of editorial policy can easily lead to misconceptions regarding the nature of “oral literature” itself, whether as presented in scholarly editions or in books with a popular appeal. Typically, as this chapter will suggest, these misconceptions involve an overestimation of the inherent capacity of either script or print to simulate oral performance, together with an undervaluing of the impact of scribes, collectors, and editors on what is widely received as “oral literature.” Such misconceptions can be diminished, however, through attention to the complex processes by which “oral texts” are made. Scholars writing from a postmodern perspective have recently put a great deal of pressure on the concept of editorial fidelity to an ideal, quasi-Platonic source. In a sense this is nothing new, for an attentiveness to the mouvance of literature – that is, to multiple locations of authority in the production and reception of texts – has long characterized the academic study of folklore and oral literature. This is because the essence of all oral art forms, from traditional ballads and epic poems to contemporary urban legends, is that they exist only in variants, not in a single authoritative ur-text of which other versions are corruptions. The problem of variance that has recently energized literary studies has been addressed in their own manner, since the late nineteenth century, by folklorists pursuing one or another variety of the historic–geographic method of comparative analysis. This method involves, 205

206

john d. niles

in principle, the documenting of an item of lore through the publication of every available witness to it.1 The cataloguing and publishing of variant texts does not make the task of theorizing variance go away, of course. This is one reason why in its older modes, the historic–geographic approach now seems somewhat pass´e. When one consults the monumental anthologies of English-language balladry that were published by Francis James Child at the end of the nineteenth century and Bertrand Bronson fifty years later2 – volumes that encompass a vast number of variant forms of ballads identified by Child as “popular” – it is easy to see that not every recorded version of an oral art form has equal ontological status. Present-day folklorists are therefore likely to want to problematize each and every variant as a historically determined cultural artifact. While item A might represent a faithful record of a unique oral transaction, item B might result from the editorial collation of texts. Yet another item might reflect an editor’s normalizing tendencies as regards dialect, meter, stanzaic structure, grammar, or other features of the text. While one variant might be littered with scribal errors or typological misprints, a different one might reflect an editor’s attempt to smooth out such anomalies, as well as to correct such unsatisfactory features as a fragmentary structure or “missing” lines. Each and every recorded variant of an oral art form, therefore, would seem to require analysis in its own right, with attention to the aims and assumptions underlying its textualization. Scripting the voice: two problematic examples What is involved then, first of all, when a folklorist sets out to capture an oral art form in the medium of print, making an effort to preserve not just the words of a performance but also some features of its oral style? And, second, what is involved when a medievalist or classicist sets out to edit an ancient text that is thought ultimately to derive from oral tradition, hoping to preserve traces of its maker’s voice? It will be useful, as a prelude to a more wide-ranging discussion of the practices and challenges of oral textual scholarship, to introduce an example of each of these two related activities. Example 1. “George Collins” in Tennessee During the years 1928–33, the New Jersey-based educator Mellinger Henry collected a number of songs of older British origin in the region of Cades Cove, Tennessee (now part of the Great Smoky Mountains National Park). Though not a professional folklorist or anthropologist, he was motivated by his former professor George Lyman Kittredge (Child’s successor at Harvard University) to follow the example of the influential English folksong collector

Orality

207

Cecil Sharp and record American songs and ballads in the field.3 In August of 1930, Henry had a stroke of luck, for Mr. and Mrs. Samuel Harmon, who had recently moved out of Cades Cove, happened to turn up at the log cabin where Henry and his wife were staying. The Harmons were a goldmine of Southern Appalachian regional lore. Over a period of two days, Henry’s wife wrote down twenty-four songs from their dictation. What makes this collection of special interest today is that Samuel and Polly Harmon, who were essentially illiterate, seem to have felt under no compulsion to produce texts that fulfilled conventional expectations as regards meter, stanzaic structure, and other formal qualities associated with printed folksong and balladry. When Henry published his collection eight years later, moreover, he evidently set the words out on the page much as his wife had written them down, even when the results look curious from a literary perspective. The song known as “George Collins” (Child no. 85) can serve as an example.4 “George Collins,” recorded by Mrs. Mellinger Henry from the singing of Mrs. Samuel Harmon, Cade’s Cove, Tennessee, August 12, 1930.

1. George Collins rode home One cold winter night And taken sick and died. 2. Little Nellie being in the other room A-sewing on her silk so fine, And when she heard her George was dead, She lay her silk aside. 3. She weeped, she moaned for her true-love. ....................................... “O daughter, O daughter, what makes you weep? There are more pretty boys than George.” 4. “O mother, O mother, I know that’s true, But he’s got this heart of mine.” ....................................... ....................................... 5. “Set down the coffin, unscrew the lid, And roll back the linen so fine, And let me kiss his cold clay lips. I’m sure he’ll never kiss mine.”

208

john d. niles

The lines of dots that appear in stanzas 3 and 4 of the song are editorial; they are supplied by Henry in an apparent attempt to accommodate Mrs. Harmon’s phrases to the conventional four-line ballad form. Henry also, of course, supplied the text’s punctuation and capitalization, lineated it, and organized the lines into five irregular stanzas. Importantly, the musical notation he provided, with its 4/4 time signature and its haphazard bar lines, is virtually unsingable. Henry had little musical training, it seems, a fact for which he apologizes in the introduction to his book. When Bronson reprinted this version of “George Collins” in Volume ii of The Traditional Tunes of the Child Ballads, he therefore felt free to revise Henry’s musical notation, converting the tune into 6/8 time and, accordingly, lengthening the duration of certain of its leading tones.5 Here, then, is what seems to be a naive record of the singing of an unschooled person from the Southern Appalachians in the year 1930. No strong editorial effort seems to have been made to smooth away the song’s oddities. A number of questions might arise when one tries to make sense of this text, however. How did Mrs. Harmon accommodate these words to her tune? Was her performance in fact stanzaic, as Henry’s transcription suggests, or did she sing the song in rhythmic and melodic phrases that succeeded one another in a manner not easily reduced to the conventional four-line ballad form? Did she think there were “gaps” in her performance, or not? Is the first “stanza” of the published version best lineated as a series of two 2-beat lines followed by one 3-beat line, as in the edited text, or should that passage be printed as a single 4-beat line followed by a 3-beat line? And what about the quotation marks introduced in stanzas 4 and 5? Why are there two sets of them, as if to denote either that two different persons are speaking or that there is some discontinuity in the young woman’s lament? These words might alternatively have been printed as an unbroken 6-line speech. One might ask, in addition, if we can accept with confidence Bronson’s emended notation of the tune. Certainly his version of the tune is coherent whereas Henry’s is not, but what were the rhythmic contours of Mrs. Harmon’s actual singing? In any event – to raise a problem that affects nearly every tune ever recorded from oral tradition – how can one discern the melodic variations that naturally occur from phrase to phrase in traditional singing if only the tune of one idealized stanza is printed? To generalize from this example, an edited text of a singing or storytelling event is emphatically the result of a collaborative process. It is a version of a performance that itself remains unknowable, though subsequent re-editings may strive for a more perfect representation of that inscrutable source. One can even argue that our language distorts reality as soon as we speak of “the Harmon variant” of “George Collins” rather than speaking of “the Harmon/Henry variant” or “the Harmon/Henry/Bronson” variant, phrases that acknowledge the inherent hybridity of oral texts.

Orality

209

Example 2. Re-animating an Old English charm The so-called Lacnunga manuscript (London, British Library MS Harley 585), a small, plain-looking manuscript written in Old English during the late tenth century, contains a miscellany of healing texts. Among these is a remarkable charm known as Wiþ Færstice (‘Against a Piercing Pain’, folios 175r–176r). This charm, though by no means a pagan document, is meant in quasi-shamanic fashion to heal an afflicted person by warding off an assault made by malevolent elves, witches, or old pagan-style divinities. This text has attracted considerable attention thanks not just to its sensational content but also to its vigorous style, which is suggestive of live performance. Although the main part of the charm is composed in alliterative verse that is obviously meant to be sounded out aloud, the text is written out from margin to margin in wrap-around style, as is typical with scribal copies of the vernacular verse of this period. In the standard collective edition “The Anglo-Saxon Poetic Records” (ASPR), the main body of the charm is presented in what is today the usual manner of publishing Old English verse: namely, in four-stress alliterative lines, with a caesura (marked by a visible space) separating each pair of half-lines. The basic rhythm of the verse is thus made clear to the eye of the modern reader:6 Gif ðu wære on fell scoten oððe wære on flæsc scoten oððe wære on blod scoten oððe wære on lið scoten næfre ne sy ðin lif atæsed.

(lines 20–2)

In plain modern prose, what this passage means is “Whether you were shot in the skin, or were shot in the flesh, or were shot in the blood, or were shot in a limb, may your life never be stricken.” In a radical departure from ASPR editorial practice, which (he argues) the scribal text “plainly resists,” the medievalist A. N. Doane has argued for a “performative edition” in which the words are distributed on the page more expressively. Influenced by the writings of the French critic Paul Zumthor on the voice as a dimension of the manuscript text, as well as by the ethnopoetics movement of the 1970s and 1980s, Doane treats the text as a series of short phrases, laying them out on the page so as to stand in isolation from one another in a manner reminiscent of open form verse. The following excerpt illustrates his method: on.fell secoten gif.ðu.wære oððe.wære on.flæsc scoten oððe.wære on.blod scoten oððe.wære on.lið scoten næfre.ne.sy ðin.lif atæsed7

210

john d. niles

In the manuscript, as can be seen from a photographic facsimile that Doane provides, the words are run together as follows: [ . . . ] gifðuwære onfell scoten oððewære onflæsc scoten oððe wære onblod scoten oððewære on lið scoten næfrenesy ðinlif atæsed [ . . . ]8

Doane’s aim in re-editing the text as he does is to suggest how (in his view) the scribe re-voiced the text in the process of writing it: namely, in short, emphatic syntactic units, with certain words run together quickly. (To clarify Doane’s punctuation system, when he inserts an editorial dot between two words, this is to signify that in the original manuscript, these words are run together without any division. Those same words should be run together quickly, he feels, when the text is voiced today.) The staggered spacing of the phrases on the page is meant to provide a visual counterpart to the rhythm of syntactically parallel phrases. Doane explains that, while leaving the words of the charm unchanged, he wishes to provide modern readers with a “prompt” or “score” that will promote “oral readings that respond to the movements of the [scribal] text.”9 This practice is called for, he argues, not because the charm was meant to be orally performed in its original medical context (though it undoubtedly was), but because residual features pertaining to the scribe’s manner of “re-vocalizing” the text are visible in the manuscript copy. Doane thus shifts attention away from a putative originary moment of poetic composition (or shamanic performance) to an observable act of scribal “re-performance.” Like Henry’s attempt to represent Polly Harmon’s words and tune in the medium of print, still, Doane’s editorial treatment of this Old English text raises as many problems as it resolves. By what authority (or by what intuitive powers) can one determine if there is any significance to the way the scribe spaced the words? Word divisions are notoriously erratic in the manuscript records of Old English, and experts do not agree if this fact has any significance. In any event, is a special mise en page of the text necessary for the charm’s rhythm to be grasped? Its parallel rhetorical structures would seem to be quite evident in any conceivable edition. More importantly, does Doane’s re-editing of the manuscript text embody a “desire for performance” that relies on the unspoken assumption that the recovery of a putative vocality is the editor’s primary task? Without trying to resolve those questions, one can see that any edited version of a medieval text is a crafted artifact that results from a process involving multiple centers of authority. Whether one prefers the standard “Dobbie/Harley” edition of Wiþ Færstice, the revisionist “Doane/Harley” edition, or an alternative setting of this text, what each printed version represents is a negotiation of the Anglo-Saxon scribe’s work, normalized

Orality

211

into modern orthography and spaced out on the page in such a manner as to indicate its ontological or generic status. The scribe too, of course, was a negotiator, registering an oral charm (that is, a real or potential speech act) in the medium of script. As for whatever live “shamanic” event the text may reference, it is as unknowable today as Polly Harmon’s manner of singing “George Collins” when no collector was present, though such events remain intriguing objects of speculation. What is “oral literature,” then, and what is an “oral text”? Each of the two preceding examples illustrates a form of “oral textual scholarship,” in that someone is using the medium of print to preserve traces of orality. Taken together, these two types of text-making go far toward defining the category of “oral literature” for our time. While, among most people, this term is commonly used with reference to the world of live performance itself, it might more accurately be used to refer to legible texts that have their origin (or are thought or are claimed to have their origin) in the spoken or sung word. Taken in this more specific sense, “oral literature” is what comes about when an oral form of art is textualized. Strictly speaking, of course, the term “oral text” is just as much of an absurdity as is “oral literature,” a term which Walter J. Ong has criticized as being an oxymoron.10 On the one hand, people watch and listen to live performances; on the other, readers read texts. These are two quite different types of experience, not just in terms of their acoustic properties. Whereas a text is fixed on the page and, in principle, is infinitely reproducible, the defining trait of oral art forms is that no song is ever sung (nor is any story ever told) in the same way twice. Moreover, the words of a work of verbal art may be only one element in an embodied event involving such factors as intonation, facial or bodily gesture, melody, instrumental accompaniment, dance, ritual actions, audience participation, and so forth. Since such elements as these are very difficult to reproduce on the printed page, the impression of oral art forms that is imparted by books is always a partial and distorted one, even if an editor has tried to document an item well while maintaining the utmost fidelity to a source. Regardless of that latter point, there is no question that the media of script and print have proven their utility as a means of displaying and disseminating oral-derived texts. It is hard to imagine a literary world in which the poems of Homer or the traditional folktales and ballads of Europe had no place. What is not always appreciated is that by their nature, the media of script and print are saturated by the culture of literacy, so that a performance that is fixed on the page tends, by that very fact, to be transformed into a new kind of aesthetic object. As I have argued elsewhere with particular reference to the Old English epic poem Beowulf, an oral-derived text is by its nature a

212

john d. niles

hybrid form, something that is neither fully literary nor fully oral in nature. Rather, it is a tertium quid (“third entity”) that combines features of the realms of literacy and orality.11 The relevance of that observation to our understanding of the great body of texts that have come down to us from the Middle Ages is worth exploring in greater detail.

Between orality and textuality: editing medieval manuscript texts As is well known, eighteenth- and nineteenth-century literary scholarship was marked by a strong turn toward a taste for the “primitive,” the “natural,” and the “gothic” in literature, as well as by the emergence of philology as an analytical practice and the study of folklore as an academic discipline. A general fascination with the Middle Ages as a period of national origins led naturally to controversies concerning how best to edit medieval texts – those texts in particular, such as anonymous epic poems, romances, and ballads, that showed signs of having emerged at the interface of writing and oral tradition. Ever since the time of Thomas Percy, whose influential (yet highly mediated) anthology Reliques of Ancient English Poetry was first published in 1763, scholars have generally acknowledged that medieval culture was permeated by the culture of orality, though the implications of that fact have only lately come in for nuanced attention. Even among the clergy (who represented only a small minority of the population), knowledge was transmitted through oral dictation as well as through books. For most other people in society, education took place not through the schools but rather through face-to-face verbal encounters in the home, the workplace, and elsewhere. The entertainments of this period are thought to have included verbal arts cultivated by professional or semi-professional singers and storytellers who practiced their arts at fairs and festivals, along pilgrimage routes, in courts, and wherever else people congregated. It is only through the work of scribes, of course, that we have access to medieval literature today. In former years, it was widely assumed that the advent of literacy in a society destroyed its oral culture. Modern concepts of authorship, together with modern attitudes toward textual scholarship, were therefore extended back into the Middle Ages. Since editors tended to assume that the variant texts of a work were the result of a degenerative process of scribal transmission, late nineteenth- and early twentieth-century critical editions generally aimed to restore a more perfect text than had survived in manuscript form, ideally by restoring an authorial archetype (even if the identity of the author remained unknown). During the past quarter century, thanks in part to the research of such scholars as the classicist Eric Havelock, the social anthropologist Jack

Orality

213

Goody, and Brian V. Street (an expert in literacy practices), the interrelations of oral culture and literacy in a number of societies of the world have been discovered to be much more complex than used to be supposed. Oral art forms and elite literary works are thus no longer conceived of as dwelling in separate domains; rather, they are now generally thought to have existed side by side with one another during earlier historical periods, as indeed they do today, constantly exerting reciprocal influences on one another. The relevance of the concept of “authorship” to works dating from the Middle Ages has therefore come under intense scrutiny. The medievalist Tim Machan, for example, has found it useful first of all to distinguish between works written in Latin, a language of relatively high prestige, versus works written in the vernacular, which was of lower status and was in a state of constant flux. While a work of Latin literature can often be traced to a known author (auctor) whose text was thereby crowned with authority (auctoritas), vernacular literature tended to be refashioned in the act of its copying. Scribes seldom hesitated to change the dialect of a vernacular text into something closer to their own; moreover, they often introduced substantive changes to a text, adding to it or deleting elements at will. Developing a useful analogy, Machan likens the alterations effected by scribes as they “copied” texts to the changes made by oral poets as they re-create song. He suggests that rather than trying to making a dualistic distinction between the authors who composed texts and the scribes who corrupted them, we might more profitably think of each extant vernacular text as the product of a unique moment of composition and interpretation.12 From such a perspective, the question of the ultimate oral versus written origins of a text can perhaps best be left open since it is likely to be unanswerable. What interests Machan is “the illusion of orality”13 cultivated by scribes who drew on the rhetoric of oral performance. This approach does not rule out the possibility that a given work of medieval or ancient literature was in fact orally composed. It has been argued, as well, that certain medieval poems, like the poems of Homer at the starting point of their written existence, may have been recorded through acts of oral dictation. Moreover, the living oral traditions of Central Asia have been shown to display patterns of variation that closely resemble what is found among the manuscript records of medieval Europe. To complicate matters further, medieval “hack writers,” too, may have drawn on the rhetoric of oral composition when reworking older poems with a popular appeal. This may account in part for the fecundity of the genre of Middle English popular metrical romance, a genre that exists in a notorious flux of variants.14 To see how scholars have responded to the challenge of editing the oralderived literature of the Middle Ages, it will be instructive to focus attention on one of the most famous works of them all.

214

john d. niles

Editing the Song of Roland The Song of Roland is preserved in six independent versions composed in several different dialects of Old French. In addition, a few Old French fragments have survived, along with versions in other European languages. The earliest and the most influential version is preserved in Oxford, Bodleian Library ms Digby 23, a twelfth-century manuscript showing Anglo-Norman dialect traits (O).15 The first scholars to edit this text according to the developing principles of nineteenth-century German philology, Theodor Muller ¨ (1851, 1863, 1878) and Edmund Stengel (1900), attempted to establish a stemma deriving O and the other known versions from a lost archetype (), drawing on their concept of the stemmatic relations of the texts to correct many of O’s readings, thus eliminating most of its Anglo-Norman dialect traits. By collating versions, Stengel added some 600 lines to O in an attempt to come closer to a postulated archetype. Similarly, the great French medievalist L´eon Gautier (1872 and subsequent editions) edited away many of the dialect features of O and included 28 laisses that are not found in that version, supplying them from the collateral tradition. A crucial methodological shift occurred when Joseph B´edier (1921), challenging Lachmannian forms of textual criticism, rejected the interventionist policies of previous editors and offered an edition of the Roland that strictly reproduced the readings of O, correcting only its obvious copying errors. Taking the Oxford version as an example of the “best text” on which (in his view) a critical edition ought to be based, B´edier denied claims that oral tradition had a role in the poem’s transmission, viewing O instead as the work of a learned late eleventh-century author. B´edier’s “individualism,” together with his respect for the minute particularities of the transmitted text, fit in well with the New Critical discourses that become dominant during the middle decades of the twentieth century. When the American medievalist Gerald J. Brault offered a critical edition of the poem in 1978, for example, he did so under the assumption that O was the work of a single poet of genius, whom Brault took to be a learned person who drew freely on the medieval Christian exegetical tradition. Individualism of this kind encountered resistance from two different quarters, however. On the one hand, the Spanish medievalist Ramon Men´endez Pidal, championing “neotraditionalism,” emphasized the oral roots of the Roland story-complex and spurred renewed interest in the independent value of each variant. Concurrently, influenced by the model of oral-formulaic composition that Milman Parry and Albert Lord had developed through the comparative study of Homer and modern oral epic traditions, some critics, including Jean Rychner and Joseph J. Duggan, argued that certain features of O (such as its highly formulaic diction) were direct reflexes of its oral

Orality

215

origins: the true “author” of the Song of Roland was thus no longer a person, but the tradition. In a brilliant return to Lachmannian stemmatics, on the other hand, the Italian scholar Cesare Segre (1971 and later editions) assumed and, in part, constructed a new archetype for the Roland complex based on a strict theory of literate transmission. The textual criticism of the Song of Roland had thus reached an impasse when Paul Zumthor, offering a more general critique of medieval orality and literacy, established new ground for these debates through the concept of textual mouvance. In his view, each stage in the evolution of the Roland song-complex was potentially subject to both oral and written influences. Any variant text might display the signs of both scribal degradation and oral/aural modification. Zumthor’s critique prepared the way for the most recent scholarly edition of O, the one by Ian Short that forms the first part of a splendid collaborative edition of all the Old French versions of the Song of Roland. Here, in recognition of its independent value, each variant text is presented in a fully developed critical edition. While the Oxford text holds pride of place among these texts, it is presented as primus inter pares rather than as a unique work of individual genius. Adopting a basic premise of the neotraditionalists and the oral theorists, Short accepts that multiple “original versions” of the Song of Roland may have existed, thus undercutting the need to construct an archetype. Moreover, he accepts that a phase of oral transmission might have preceded any of the written versions that survive. As for the individual scribes who wrote down the different versions, he conceives of them not as mere copyists, but as participants in the tradition, “negotiating between oral recollection and the written transmission mode” (i: 48). In the end, rather like Tim Machan in his view of later Middle English literature, Short sees the “inherent orality” of O as being accommodated to its written mode, which in turn was subject to scribal degradation. He therefore – somewhat surprisingly, in the current conservative editorial climate – approaches O with greater latitude than any other editor has done for many years, emending the text rather freely so as to eliminate its orthographic, grammatical, and metrical irregularities. Whether or not medieval scribes negotiated with the text while copying it, then, each new critical edition of the Song of Roland that has appeared in print since the mid-nineteenth century, culminating (for now) with Short’s 2005 edition, represents a negotiation not just with “the work itself” (however that work is conceptualized), but also with the scholars and critics who have previously brought their intelligence to bear on understanding both that work and the culture that produced it. As for the splendid collaborative edition of the Songs of Roland (as they might just as well now be called) in which Short’s volume has a key place, it still invites questions. While each full-length Old French version of the Song is awarded equal editorial

216

john d. niles

attention, and while the category of authorship is not pressed, how are we to conceptualize the nature of variance in the Roland tradition as a whole, other than through a rather sweeping acknowledgment of mouvance? As for Short’s emendations, do they eliminate anomalies that, whether we like them or not today, were a substantive feature of the Song of Roland as it was transmitted and received during the Middle Ages? Are we then left with a “dream poem” whose only existence is between the covers of the Brepols edition? Moreover, supposing we accept the hypothesis that O (or any other text in the Roland song-complex) does derive ultimately from oral tradition, as some have maintained, should we not still ask the question, “Through what process were the works of oral poets captured in script during the Middle Ages, and how might they have been transformed as a result of that process?”

The influence of collection events on “oral texts” This last question brings us back to an aspect of orality whose crucial importance has come into focus only during the past fifty years or so. This is that no “oral text” is an unmediated example of the art of a singer or teller of tales. Rather, the textualization of oral art forms generally presupposes a meeting of two worlds, that of the tradition-bearer and that of a lettered elite, in the course of which the words of singers or storytellers are transmuted into new forms. An instructive example of this process is Black Elk Speaks, perhaps the most celebrated Native American oral autobiography yet to have appeared in print.16 The making of this work involved at least six stages. They can be summarized as follows: r Nick Black Elk (Hehaka Sapa), a member of the Oglala tribe of Lakota Sioux, told elements of his life story to John G. Neihardt, a poet and a historian of the American West, speaking in his native tongue in a series of interviews. r Black Elk’s son, Ben, provided a simultaneous paraphrase of Black Elk’s speech into the English language (for Neihardt knew little Sioux). r Neihardt’s oldest daughter, Enid, took down a stenographic record of Ben’s words, entering additional comments in her notebooks as she did so. r Enid made a typed transcript of those notes, adding some handwritten corrections. r Neihardt wrote the book Black Elk Speaks as we know it today, transforming his daughter’s transcription of these interviews into a carefully shaped text. During that compositional process, Neihardt infused the

Orality

217

narrative with sentiments of his own, created biblical-sounding cadences, made other stylistic improvements, and added certain passages of his own invention (such as Black Elk’s closing words), which were meant to express Neihardt’s sense of “what Black Elk would have said if he had said it.” In composing the book, Neihardt drew on memories of other encounters he had had with Black Elk, plus his familiarity with a much wider body of knowledge involving the history of the West. r Neihardt’s manuscript, with its many handwritten corrections, was converted into book form at the press through the normal process of copyediting, printing, and proofing. Since this complex process was not spelled out in the first edition of Black Elk Speaks (1932), the extent of Neihardt’s role in it was effaced. In the preface to the 1961 edition, however – responding perhaps to critics who had questioned the work’s authenticity – Neihardt added an account of how the book was produced.17 Moreover, included as appendices to the 1961 edition (and most subsequent editions) are photographic facsimiles exemplifying Enid’s handwritten notes, her typewritten transcriptions, and Neihardt’s manuscript draft of the book. Since, in addition, Neihardt’s papers are available for study at the Western Historical Manuscript Collection at Columbia, Missouri, there is no question of a hoax here, only of an act of “unique collaboration,” as has been aptly remarked.18 To generalize from this example, what an “oral text” is very likely to represent is a hybrid production that would not have come into being were it not for the fusion of two cultural worlds. Moreover, this example illustrates changing attitudes toward “authenticity” in the representation of orality from the 1930s to the 1960s in the realm of Native American studies – a field where, by the 1960s, “natives” began to speak quite distinctly for themselves. Does the book silence Black Elk’s voice in the act of disseminating it? Certainly it is true that Neihardt effaces such matters as Black Elk’s conversion to Catholicism, which must have mattered much to the man himself. Was Neihardt’s whole project tainted by cultural colonialism, and should a qualified member of the Sioux nation now attempt to create a new “oral autobiography” of Black Elk on the basis of a fresh reading of Neihardt’s daughter’s stenographic notes? If so, then what added authority or authenticity will that text have? Should we even care, if we admire the book for its outstanding literary qualities, that the sweet illusion has been broken that the relation between Black Elk and the reader is almost unmediated? Such questions as these are possible to pose only now, not at the time of the book’s first release, and they will perhaps never be answered in a way that is universally satisfying.

218

john d. niles

Ethnopoetics and after The example of Black Elk Speaks throws into relief certain large questions that can arise when published works claim to represent the oral culture of either non-Western peoples or marginalized minorities. Other questions that focus more narrowly on typography can arise in this same context, as well. Through the ethnopoetics movement of the late twentieth century, a shift has occurred in the manner in which traditional Native American literature – and, by extension, any literature grounded in oral performance – is represented on the printed page. Such scholars as the sociolinguist Dell Hymes and the anthropologist Dennis Tedlock, in keeping with a post-1970s tectonic shift toward the study of performance and context in folkloric research, have argued for arranging “oral texts” on the page in a manner that uses visual cues to call attention to their rhythmic and acoustic features. Working with native Northwest Pacific Coast stories as written down and published by late nineteenth- and early twentieth-century linguistic anthropologists, Hymes has re-edited such texts with the aim of making visible their rhetorical structures and their rhythmic pulse through the grouping of words on the page. “Lines” of text are grouped into “stanzas” of varying length, so that the resulting text resembles open-form verse. Tedlock and others, relying on tape recordings that can be played back again and again, have used additional visual cues to mark extra-verbal aspects of performance, including the acoustics of the voice. Words that are spoken loudly may be set out on the page in strings of capital letters, while those spoken softly may be set in small font. Italics and/or acute accents may mark various kinds of emphasis. Vocables (like hi – hi – hiyo!) may be represented as such, at considerable length. Some key native terms may be left untranslated, thus defamiliarizing the English text in an effort to restore the native voice. To cite just one fairly restrained example of such methods, the widely respected folklorist Barre Toelken (former President of the American Folklore Society) has retranslated parts of the Tsimshian story “The Grizzly Bear,” taking as his base text a prose paraphrase of that tale published by the pioneering anthropological linguist Franz Boas in 1902. Supplementing the initial publication of the text was Boas’s linguistically precise phonetic transcription of the tale in its original language, accompanied by an interlinear phrase-by-phrase English translation.19 The following excerpt starts in at that point in the story when a she-bear (the “Grizzly wife”), after having cohabited peacefully with a human husband for a considerable period of time, is scolded by a villager for spending more than her share of time at the salmon trap (as, after all, bears tend to do!). Boas’s paraphrase of this part reads as follows:

Orality

219

She became angry, ran out, and rushed up to the man who was scolding her. She rushed into the house, took him, and killed him. She tore his flesh to pieces and broke his bones.

In his re-translation of this passage, which is based in part on Boaz’s linguistically precise transcription and translation, Toelken accents its rhythm by grouping verse-like “lines” into thematically-based “stanzas”: Then she came, being sick in heart. Then quickly she ran out at him, greatly angry. Then she went where the man was who scolded. Then into that place she stood. Then she took the man. Then all over she killed him. It was dead, the man. All over was finished his flesh. Then were broken all his bones.

While granting that his ethnopoetic retranslation involves “rough guesswork,” Toelken maintains that “even so, a more engaging – I think one may say exciting – dimension becomes available to us: the immediacy of dramatic presence.” Still, one may ask, whose dramatic presence do we witness through this kind of textual refashioning? Is it that of the Tsimshian storyteller, who remains an ideal and disembodied personage, the object of readerly imagination? Or is it that of Boas, a remarkable German-American Jewish intellectual whose chief interests were linguistic and ethnographic, not aesthetic? Or is it that of the contemporary folklorist Toelken, who, wishing to restore a native voice to Native American literature, produces deliberately primitivist effects (“It was dead, the man”)? Or do all three of these figures merge in an application of oral textual scholarship that, skeptics might say, is difficult to disentangle from the technology of empire? While ethnopoetic transcriptions or translations constitute an act of resistance to all that is lost when a storyteller’s voice is flattened to the lowest common denominator of prose, then, a case can be made that what they tend to highlight is the editor’s sensibilities even more than those of the native speaker, since the editor inevitably takes on the role of performer of the text on the printed page. When a textual “libretto” is typographically overdetermined, as it can be in some instances, a problem of readerly

220

john d. niles

resistance can arise, as well. On the other hand, many specialists agree that ethnopoetic representations, particularly when accompanied by nuanced analyses of texts in their performance context, tend to bring out aspects of a work (including moments of humor) that might otherwise be lost in transcription. Conclusions and future prospects In sum, the realm of orality constitutes an important foil to the disciplines of textual scholarship, challenging conceptions of textual authority and fixity and suggesting alternative ways of thinking about such central editorial issues as transmission, variation, and authorship. This chapter has focused on the challenge of editing at the boundary between the oral and the textual. It has explored the process whereby the voice of the singer or teller of tales is silenced, in the act of its being recorded, at the same time as other voices (the collector’s, the transcriber’s, the editor’s) take its place. In many ways, despite all that has been done during recent centuries to make oral art forms available to the reading public, oral textual scholarship remains curiously under-theorized. My own approach to the problem of the textualization of oral art forms is evident through a number of keywords used in the present chapter: these include transmutation, mediation, collaboration, negotiation, fusion, and hybrid production. Looking at historical trends in oral textual scholarship, it is easy to trace a revolution in sensibility from the era of Percy and Sir Walter Scott, when ballad scholars made uninhibited changes to the texts they ascribed to oral tradition, to the middle years of the twentieth century, when fieldworkers like Parry and Lord set out to preserve the exact words of oral performance just as captured via audio technology. The current respect that is generally given to the precise linguistic features of orality has its antecedents not just in twentieth-century anthropological linguistics, but also in the practices of nineteenth-century philologists who established the principle (now universally recognized) that the exact original wording of a manuscript text should be registered in the apparatus of a scholarly edition in all instances when the text is changed through emendation. Intermediate steps in the shift from “cavalier” to “scientifically philological” attitudes toward the representation of orality can be traced throughout the nineteenth century, when, following the lead of the Brothers Grimm, folklorists working in all parts of Europe took down texts from the lips of singers and storytellers but hesitated to publish them without major aesthetic “improvements.” During the later decades of the twentieth century, once the potential of new technologies such as the portable reel-to-reel tape recorder was tapped, some scholars came to advocate “authenticity with a vengeance” in the representation of nearly

Orality

221

every conceivable feature of an oral performance on the page. Analogous attempts have been made to “restore the voice” of the medieval scribe or poet, sometimes through hyperconservative approaches to editing. By the end of the twentieth century, postmodern critical impulses had gone far to undermine positivist assumptions that underlie the concept of authenticity, thereby problematizing virtually any attempt to remain faithful to an imagined source. Whether late twentieth-century postmodernism is now ripe for analysis as a fin de si`ecle trend, who can say? As “oral texts” are produced in the years ahead, debates are likely to continue as to how much or how little mediation they require, whether as regards the typographical choice of verse or prose on the page, the signaling of the acoustic properties of a text, or the representation of a speaker’s dialect, non-standard grammar, verbal fillers, obscenities, obvious mistakes, or pauses for a cup of tea. More fundamentally, the question of who has authority over the shaping and disseminating of “oral texts” will continue to be posed, with answers that are sure to be in the plural. In the meantime, as the World Wide Web provides greater access to archival recordings housed in such depositories as the Archive of Folk Culture at the Library of Congress, Washington, DC and the Archive of the School of Scottish Studies, Edinburgh, the transmutation of audio recordings into legible texts may seem a less pressing need than before. In future years, to the extent that direct access to archival recordings is provided via the Web, the need for either hyperconservative printed editions or elaborate ethnopoetic displays may be found to diminish. If one can hear a recording of a person’s voice, then there is correspondingly less need to torture the medium of print so as to represent every feature of oral performance. Still, it would be naive to suppose that an electronic edition of an “oral text” offers a magic platform from which to transcend the problems of representation – and of the politics of representation – that are endemic in print culture. Digital artifacts can provide yet another mediated simulation of performance, but never performance itself. Moreover, it is unlikely that direct access to voice recordings – the sine qua non of much recent scholarship in this field – can provide an adequate substitute for the integrative syntheses offered by critical editions. As Karl Reichl, the distinguished German medievalist and collector of Turkic oral epics, has observed, “even the most complete replica” of oral performance, “catching and preserving sound and sight to the fullest technologically possible degree,” is “radically different from the event itself,” so that in the end we must define the editor’s task not as complete documentation but as critical selection and evaluation.20 Moreover, as the American medievalist and folklorist Carl Lindahl has observed on the basis of long hours of experience working with sound recordings housed at the Library of Congress, the technical quality of some archival

222

john d. niles

recordings is so wretched as to render them almost unusable.21 Moreover, some audio recordings contain sensitive personal materials that, from an ethical perspective, ought not to be made available indiscriminately. In this light, yet another contemporary folklorist has argued that certain recordings of Native American sacral experience should be destroyed, rather than disseminated, as a gesture of respect to native cultures, and has acted (with regret) on that principle.22 In sum, it is reasonable to expect that when oral textual scholarship is practiced in future years, emergent technologies will matter, but so will a good ear, a supply of patience, socially responsible ethics, and integrative, critically alert modes of editing, always attuned to the special conditions through which “oral texts” are made.

NOTES 1 L. Virtanen, “Historic–Geographic Method,” in Folklore: An Encyclopedia of Beliefs, Customs, Tales, Music, and Art, ed. T. A. Green, 2 vols. (Santa-Barbara, CA: ABC-CLIO, 1997), Vol. i, 442–7. 2 F. J. Child, ed., The English and Scottish Popular Ballads, 5 vols. (Houghton Mifflin, 1882–98); B. H. Bronson, ed., The Traditional Tunes of the Child Ballads, 4 vols. (Princeton University Press, 1954–72). 3 The “ballad hunting” movement of the 1930s is contextualized by D. E. Whisnant, All That is Native and Fine: The Politics of Culture in an American Region (Chapel Hill: University of North Carolina Press, 1983). For a revisionist view of Child, Sharp, and other ballad scholars from a Marxist standpoint, see D. Harker, Fakesong: The Manufacture of British “Folksong” 1700 to the Present Day (Milton Keynes: Open University Press, 1985). 4 M. E. Henry, Folk Songs from the Southern Highlands (New York: Augustin, 1938), 89. 5 Bronson, The Traditional Tunes, ii: 399. Bronson remarks that the timing should “pretty certainly” be 6/8, thus leaving room for other possibilities. 6 E. V. K. Dobbie, ed., The Anglo-Saxon Minor Poems, The Anglo-Saxon Poetic Records 6 (Columbia University Press, 1942), 122. Since line 21 of the charm appears to be metrically defective, many editors have supplied what they believe to be a “missing” half-line at that point, though Dobbie in his conservative edition does not emend. 7 A. N. Doane, “Editing Old English Oral/Written Texts: Problems of Method (with an Illustrative Edition of Charm 4, Wiþ Færstice),” in D. G. Scragg and P. E. Szarmach, ed., The Editing Of Old English (Woodbridge: D. S. Brewer, 1994), 125–45, at 139. S. Glosecki, Shamanism and Old English Poetry (Rochester, NY: Garland, 1989), 111–12, makes a different effort to animate the text. 8 Doane, “Editing Old English,” 135–7. 9 Ibid., 140. 10 W. J. Ong, Orality and Literacy: The Technologizing of the Word (New York: Methuen, 1982), 11–14.

Orality

223

11 J. D. Niles, “Understanding Beowulf: Oral Poetry Acts,” Journal of American Folklore, 106 (1993), 131–55, revised as chapter 4 of Homo Narrans: The Poetics and Anthropology of Oral Literature (Philadelphia: University of Pennsylvania Press), 89–119. 12 T. W. Machan, “Editing, Orality, and Late Middle English Texts,” in A. N. Doane and C. B. Pasternack, eds., Vox Intexta: Orality and Textuality in the Middle Ages (University of Wisconsin Press, 1991), 229–45. 13 Ibid., 236. 14 A. B. Lord, “Homer’s Originality: Oral Dictated Texts,” in his Epic Singers and Oral Tradition (Ithaca, Cornell University Press, 1991), 38–48; K. Reichl, “Orality and Performance,” in R. L. Radulescu and C. J. Rushton, eds., A Companion to Medieval Popular Romance (Cambridge: D. S. Brewer, 2009), 132–49; and Reichl, “The Middle English Popular Romance: Minstrel versus Hack Writer,” in J. Harris, ed., The Ballad and Oral Literature (Cambridge, MA: Harvard University Press, 1991), 243–68. 15 The present discussion is indebted to J. Duggan, gen. ed., La Chanson de Roland / The Song of Roland: The French Corpus, 3 vols. (Turnhout: Brepols, 2005), especially Duggan’s “General Introduction: Editing the Song of Roland,” i: 5– 38, and to Ian Short’s edition of O, which occupies part 1 of Volume i of the collected edition (i: 1 to i: 338). For full bibliographical references to works cited only in passing in the present chapter, see Duggan, i:125–86. 16 J. Neihardt and Black Elk, Black Elk Speaks (Lincoln, University of Nebraska Press, 2004; originally published 1932). 17 Ibid., xv–xix. 18 B. Holloway, Interpreting the Legacy: John Neihardt and Black Elk Speaks (Boulder: University Press of Colorado, 2003), 2. 19 B. Toelken and T. Scott, “Poetic Retranslation and the ‘Pretty Languages’ of Yellowman,” in K. Kroeber, ed., Traditional Literatures of the American Indian: Texts and Interpretations, 2nd edn. (Lincoln: University of Nebraska Press, 1997), 88–134, at 90. 20 K. Reichl, “Silencing the Voice of the Singer: Problems and Strategies in the Editing of Turkic Oral Epics,” in L. Honko, ed., Textualization of Oral Epics (Berlin: Mouton de Gruyter, 2000), 103–27, at 107–8. 21 C. Lindahl, American Folktales from the Collections of the Library of Congress, 2 vols. (Armonk, NY: Sharpe, 2004), i: 34. 22 B. Toelken, “The Yellowman Tapes, 1966–1997,” Journal of American Folklore, 111 (1998), 381–91.

10 MICHAEL G. SARGENT

Manuscript textuality

Around the world, people have been writing – producing visual records of spoken language and ideas – for about five thousand years. The media of representation have ranged from drawing or pressing with a stylus on a clay tablet to scratching on shells, bone, stone, tree bark, or wood, to painting or inscription with a brush, a pencil or a pen on stone, on wood panels, bamboo strips, tree leaves, papyrus, vellum, paper, or cloth, to the impression of characters on an inked surface onto a medium such as paper (printing) or the impression of a medium onto a clean surface into which characters or images have been mechanically or chemically inscribed (etching, lithography), to the electronic production of text on a screen. All forms of writing that depend on manual production of characters – the actual copying out of texts by hand – share a number of characteristics of production and transmission that we think of as the field of manuscript textuality. Writing systems Types of writing include pictograms (simplified images of actual objects), ideograms (images that represent ideas irrespective of the words in any given language, like a cigarette with a line through it, meaning “No Smoking” in English and “D´efense de Fumer” in French), logograms (images representing individual words or morphemes), syllabaries (writing systems where each character represents an entire syllable, like the “ka”, “ki”, “ku”, “ke”, “ko” of the Japanese kana scripts), consonantal systems that represent all of the consonants of a language, but not the vowels (languages like Hebrew in which the pattern of occurrence of vowels is predictable, and thus need not be recorded), as well as alphabets. Writing may have been independently invented as many as half a dozen times in different cultures, and have been adapted in some cases to record languages wholly unrelated to those in which the writing system originated. At least since the French philosopher Jean-Jacques Rousseau wrote his Essay on the Origin of Languages in 1781, these phenomena – language, 224

Manuscript textuality

225

pictograms, ideograms, logograms, syllabaries, and alphabets – have been viewed as representing hierarchically ordered stages or phases in the development of human culture. Languages that use alphabets (like English) have thus traditionally been considered more highly developed than those that use syllabaries or consonantal systems (like Hebrew or Japanese), which are more highly developed than those that use logograms (Chinese), which are in turn more highly developed than those that use ideograms (Egyptian and Maya hieroglyphs); and the least developed of all are those that have never developed writing systems at all. None of this is true. First of all, the characterizations referred to above (for example, that Egyptian hieroglyphs are ideographic) are gross simplifications, often made by people with little actual experience of the languages they are describing. Second, no writing system is a pure representative of any of these “stages” (the term is a misnomer) of cultural development, nor did the various types of writing occur one after another in chronological order in any actual, historical writing system; all writing systems incorporate elements of different types of representation. Third, the factors that make it useful or necessary to move from one type of system to another (the development of the Greek alphabet out of the Semitic consonantal system, for example) have more to do with the way that a particular language organizes sound patterns into words than with any abstract, absolute hierarchy of cultural development. Finally, the late-twentieth-century French philosopher Jacques Derrida’s deconstruction of traditional hierarchies has demonstrated that even the fundamental observation that writing is a development consequent upon the prior invention of vocal speech is itself prejudiced, for speech depends upon the diff´erance of meaning – the deferral of meaning from a particular object (the “signified”) to a particular word, a humanly produced sound (the “signifier”) whose meaning itself is only known through its difference from other words – that, as Derrida notes, is in itself characteristic of reading and writing. Marks scratched on stone, bone, shell, or wood, and signifying various kinds of objects or numbers of objects (e.g. “cow” or “twelve cows”) have been found in archeological sites predating by hundreds of years the rise of the major writing systems of the world, and are often referred to as proto-writing. The earliest known form of writing was that developed in the Sumerian civilization of southern Iraq in the thirty-fourth century bce. Earlier non-systematized forms of linear markings made on clay tokens (proto-Sumerian) had been in use for the keeping of records for centuries earlier, but it was not until the Early Bronze Age that a recognizable system of approximately 1,200 logograms was developed. The earliest form of Sumerian script was made by drawing recognizable pictograms with a reed stylus on a clay tablet; this system later developed into a more stylized form in which the wedge-shaped end of the stylus was pressed into the clay to

226

michael g. sargent

make a simplified version of the pictogram – cuneiform writing (from Latin “cuneus” – “wedge”). The earliest Sumerian literary texts (including a creation myth and a version of Gilgamesh) date from approximately the twenty-fifth century bce. It is also possible that the earliest form of Elamite script (the language of the Persian Empire) was a parallel development from proto-Sumerian. Sumerian cuneiform was adapted for a number of other languages, including Akkadian, by the addition of logographic and syllabic features, and influenced the Ugaritic and Old Persian alphabets. The transition from logographic to syllabic or alphabetic script tends to occur particularly when a writing system is transferred to a language unrelated to that in which it arose, in which words or other morphemes that cannot be pictured may pun with words in the original language that can be pictured: all of these things happened in the adaptation of Sumerian cuneiform to other languages in the region over a period of centuries. Although substantial literatures survive in all of these languages, cuneiform became extinct in the second century bce, replaced by writing systems that developed from the Egyptian. Like Sumerian, the Egyptian writing system had a long prehistory of “proto-writing”: line-drawings, inscriptions, and pictograms used for tallies and other records, but which did not necessarily represent spoken words, are found in civilizations to both the west and the east of ancient Babylon – in the Indus valley between the twenty-sixth and the twentieth centuries bce, and in Egypt from before the thirty-third century bce. If either was influenced by Sumerian, it was probably only the idea of using marks on some recording medium that was borrowed: the actual markings show no sign that they are derivative in any way. The logographic character of Egyptian hieroglyphs was obscured by the fact that throughout their history, they remained highly pictorial, unlike Sumerian cuneiform, which became so stylized that the element of depiction in any given character can often only be discerned by comparison with its ideographic precursors. The use of standard Egyptian hieroglyphs continued until the late fourth century ce. The most important derivative of Egyptian hieroglyphic writing for the history of Western culture was the proto-Canaanite alphabet, a purely consonantal writing system that came into existence in the fifteenth century bce. Forms of proto-Canaanite script spread throughout the region, and their descendants eventually came to replace all other cuneiform- and hieroglyphic-derived scripts. A particularly important development in the eleventh century bce was Phoenician writing, used for the Semitic language spoken in the area of modern Lebanon and Syria, and parts of Israel and north Africa. From Phoenician descended a number of more specialized forms, including Aramaic, Hebrew, and Arabic. All of these languages share the characteristic

Manuscript textuality

227

that every word is made up of a two- or three-consonant “root”, with different forms derived by alternation of vowels and the addition of prefixes and suffixes. In the Indo-European languages, on the other hand – like Greek, Latin, and English – the placement of vowels is not automatic, and they must be represented in the spelling system. When Greek speakers adopted writing from the Phoenicians in the late ninth or early eighth century bce, they adapted and altered the use of certain characters to represent the vowel sounds; all of the later alphabets of Europe, including the Latin alphabet that we use in English and the Cyrillic alphabet used to write the eastern Slavic languages like Russian were developed from Greek. Similarly, the logographic system of writing that was already used in China during the Shang dynasty (1200–1045 bce) has survived, although much altered in form, to the present day, and formed the basis of writing of other east Asian languages. Japanese, on the other hand, while it has continued to use Chinese characters (kanji), has also adapted simplified forms of some forty characters – in two separate series – chosen to represent single syllables according to Japanese pronunciation. The Mesoamerican writing systems, possibly dating back as far as 900 bce, and best known from the Maya glyphs of Yucatan and adjacent areas are also logographic, and – like the Egyptian hieroglyphs – were long misunderstood as being simply pictographic.1 Thus the traditional narrative of the logical development of writing systems from pictograms to ideograms to logograms to syllabic scripts to alphabets fails: the earliest systems of writing that have been invented across the globe seem already to be logographic, and the development of syllabic and alphabetic scripts seems to have been driven primarily by the constraints of adapting an already developed logographic script appropriate to one language, its grammar and pronunciation, to an unrelated language with a different grammar and pronunciation. Writing materials Many different kinds of materials have been used in the production of manuscript texts. Sumerian and other cuneiform writings, as mentioned above, were made by the impression of a reed stylus on the surface of a clay tablet. Proto-writing in Egypt, China, and the Indus valley all took the form of scratching on stone, wood, bone, or shells. The earliest Egyptian and Maya hieroglyphs were incised into and/or painted on stone or wood. Some early Buddhist literature from south Asia was written on palm leaves; early Chinese writing was done on tablets or strips of wood or bamboo (by the fifth century bce), and even occasionally on silk. Different kinds of paper were separately invented in Egypt, China, and Mesoamerica. Egyptian papyrus dates back as far as the third millennium bce, and continued in

228

michael g. sargent

use until it was replaced by parchment and paper between the eighth and eleventh centuries ce. The Chinese invention of paper, composed of plant fiber (normally from the bark of the mulberry tree), hemp and other rag cloth, occurred perhaps as early as the second century bce; the technology spread to Japan in the early seventh century, to the Islamic west in the eighth century, and thence to Europe in the twelfth. Asian paper (“rice paper”) is usually thin and translucent, and only written on one side; it is either rolled into scrolls or accordion-folded into pages to form books. Thick, modern European-style paper was invented by Arab Muslims in the world’s first true paper-mill, in Baghdad, c. 794 ce. The amatl paper used in Mesoamerican (Maya and Aztec) codices was made from the bark of the fig tree, covered with a lime gesso, and folded into pages. Parchment (“pergamenum,” from Pergamon, where it was reputedly first made) is made by soaking the hides of animals to loosen the fat and flesh adhering to the inside and the hair follicles on the outside, and scraping them to remove any residue. The sheets are then stretched out to dry, and rubbed with chalk to whiten them, and pumice to roughen the surface slightly so that the ink will adhere. Sheets of parchment are then sewn into scrolls, or folded into gatherings of separate folios. The earliest books in the Indo-European tradition were written as scrolls; folded pages were only used by merchants for their records; the Christian scriptures are the first literary texts to be written down as books. Manuscripts and textual authority Manuscript text is variable in ways that we do not commonly think of printed text as being: no two manuscripts of the same work are ever identical. Intentionally or unintentionally, every scribe introduces variations in his or her copy of a text. In the Western tradition, this has resulted in a constant concern for the accurate production of the original, “authentic” text. As long ago as the third century bce, scholars were arguing over the shape and content of Homer’s Odyssey, and commenting on the probability of authenticity of the variant readings of individual lines as they had been preserved in manuscripts of the first four centuries or so of the poem’s transmission. Even in the most careful and conservative of scribal traditions, such as the copying of the Torah by Jewish professional scribes (sofrim), individual variations still occur. From classical antiquity until the Italian Renaissance, the primary tool of the textual critic was his knowledge of the history and grammar of the language, applied to the collation of divergent manuscript witnesses. This was the method employed, for example, by the greatest of medieval critics, Lupus of Ferri`eres (c. 805–62). To this, Renaissance philologists like

Manuscript textuality

229

Colluccio Salutati (1331–1406), Poggio Bracciolini (1380–1459) and Lorenzo Valla (1407–57) added a conscious seeking out of older exemplars, free, as they saw it, from the “gothic” corruptions of medieval Latin. Textual criticism of the scriptures developed in parallel with that of the classical pagan authors. The first translation of the Torah into Greek is traditionally attributed to six scholars from each of the twelve tribes of Israel (the “seventy,” or “Septuaginta”), and produced for the library of Alexandria in the third century bce. This form of the text was used by Jews in the Hellenic diaspora, and adopted by early Christians as their form of the Jewish scriptures. Other translations in Greek were produced in the rabbinic tradition, including two second-century ce versions attributed to Aquila and Symmachus and one in the second century ce by Theodotion. The Christian theologian Origen (c. 185–254), interested in the divergence of the versions, produced the “hexapla,” a six-column comparative version of the Hebrew text, the Hebrew transliterated into Greek, Aquila, Symmachus, the Septuagint, and Theodotion. The Septuagint comprises the Torah and expanded versions of the Nevi’im (Prophets) and Ketuvim (Hagiographic writings); the Hebrew canon of the Tanakh (an acronym for Torah, Nevi’im, Ketuvim) as established by the rabbinic tradition between the second century bce and the second century ce lacks such apocryphal books as 1–2 Esdras, Tobit, Judith, Wisdom, Ecclesiasticus (Sirach), and Baruch. Manuscript versions of the Septuagint, the Tanakh, and the Greek New Testament, as well as a number of independent Latin translations of the Greek Bible, circulated for centuries, with occasional efforts at standardization. The Latin Vulgate, a translation undertaken by St. Jerome in the early fifth century ce to replace the various Vetus Latina (“old Latin”) versions, became standard in the Western European Christian church, although manuscripts containing at least parts of the Vetus Latina continued to be produced into the thirteenth century. The Masoretic text of the Tanakh was produced between the seventh and tenth centuries ce and became standard. The text of the Torah remained remarkably stable in its manuscript transmission, with few differences between the Masoretic text and the textus receptus (received text) as evidenced by manuscript fragments dating back to the second century ce and from the Qumran texts, although greater divergence can be seen between the Hebrew text and that which must have underlain the Septuagint translation. Textual variation in the other books of the Hebrew Bible, although greater than in the Torah, is still relatively slight, as compared with the Christian traditions. The impulse to textual standardization and uniformity, particularly in copies of the Bible and books used for religious services was equally important in the Western Christian tradition, but the emphasis was not so much on the development of a tradition of scribes who (at least theoretically) didn’t

230

michael g. sargent

make mistakes, but rather on the correction of manuscripts and the (idealized) restoration of texts to their pristine, original form. This was accomplished, in a continuation of the pagan classical tradition of textual criticism, by the comparison of manuscripts against other copies whose authenticity (as demonstrated by their age, the clarity of their language and writing, or the prestige of the churches or monasteries that owned them) was considered greater. Another impulse toward standardization in manuscript texts came from the growth of the universities in the thirteenth and fourteenth centuries: the University of Paris famously developed the pecia system under which licensed scribes would rent approved copies of set texts from the university stationer one eight-folio gathering at a time in order to produce identical copies for masters and students alike. Johannes Gutenberg’s invention of printing from movable type in the mid-fifteenth century offered the massive advantage of being able to produce large numbers of cheap, identical copies, which was seized upon by various authorities almost immediately.2 It is often remarked that Gutenberg’s choice of the Latin Bible for his opus princeps was a master-stroke of marketing because the demand for copies would have been so great, but we must also remember that without some other advantage, he was selling into an already relatively full market: Western Europe in the fifteenth century was hardly poor in manuscript Bibles. The other advantage that print offered was the ability to produce a large number of identical, authorized copies of a text. We ought to remember here that an “auctor” in Latin was not just a person who wrote a text. Coming from the verb “augeo” – to increase, to found – the noun “auctor” applied to any number of foundational sources: Romulus and Remus were the “auctores” of Rome; the Senate the “auctor” of its laws. “Auctoritas” did not derive from the person who originated an opinion, but from the body that sanctioned that opinion; the word “authorize” not from the writing, but from the licensing of a book. “Authority” and “authenticity” thus have more to do with the authorities than with the author. This is not to argue, as Roland Barthes does in “The Death of the Author,” that the concept of the author is a fragment of the false consciousness of capitalist ideology, but rather that we need to remain aware of the socio-cultural work that is being done by the authorization and publication of a text. The publication of books is the earliest case of the “technological reproduction” of an art form, although it does not fit Walter Benjamin’s Marxist paradigm because books (in the Western tradition, at any rate) do not meet his primary criterion for consideration, the physical uniqueness of the original work.3 It is of the nature of texts that they have always been reproduced, although the “artistic” aspect of the individual copy (with the exception of illuminated masterworks that have become available on the mass market in the twentieth

Manuscript textuality

231

century in precisely the way that Benjamin describes) has never played the prominent role in the Western scribal tradition that it has in the calligraphic production of texts in the Chinese, Korean, and Japanese tradition. Manuscript and print, manuscript and electronic text The difference that print has made in the Western tradition of the production of text is that the industrial revolution occurred in the production of books long before it did in other arts: printing – considered as an industry requiring investment of capital in the means of production, resulting in multiple, uniform mass product, with the profit going directly to the owner of the means of production – is in fact the earliest form of industrial capitalistic enterprise. And authority has belonged, since the invention of the printing press, not to the author of the text, but to the printer and the patrons who supply him with texts and subventions to print them, and to the compositors, editors, and correctors whom the printer employs. It was not until John Dryden succeeded in suing publishers of pirated editions of his works in the early eighteenth century that the text was seen as the legal possession of the one who wrote it. We must also note that the tendency of print to create a category of “authentic” or “authorized” texts – particularly after the establishment of governmental approval and licensing of the right to print texts in the sixteenth and seventeenth centuries – resulted in the back-formation of a new category of textual production: “coterie” or “scribal” publication.4 These terms are used to describe the transmission of texts in manuscript, passed hand to hand or scribe to scribe (which, before print, was the only way to transmit texts) as a conscious choice on the part of the author. The production of corrected editions of classical Latin authors and scriptures in the Italian Renaissance, among northern European humanists like Erasmus, and in the sixteenth-century Reformation, brought the appeal to antiquity and to authority as defined by propinquity to authorial originality into play as a major, if not the major, critical criterion. This heuristic shift was part of the self-definition of the Renaissance and the Reformation as a re-birth of a higher culture that had been obfuscated by the intervening, “gothic” Middle Ages. The criterion of the categorical superiority of older manuscript witnesses was invoked explicitly by Angelo Poliziano (1454–94) in his work on the establishment of the text of Cicero’s Epistolae ad familiares: observing that one of the surviving manuscripts derived directly from another, and that a further set of manuscripts descended from that derivative copy, he formulated the principle of the eliminatio codicum descriptorum (the elimination of derivative manuscripts). Succeeding centuries of scholarship shifted the theoretical basis of textual argument away from aesthetic,

232

michael g. sargent

grammatical, and rhetorical criteria (with an inherent tendency to reflect personal taste) to more “scientific,” evidential criteria (although even these tended to have a personal, aesthetic quality). Modern textual criticism began with the idea of deducing the original form of the text by applying a genetic model to account for textual variation in the manuscript culture. The first time that the genetic model was applied in textual criticism appears to have been by the Lutheran pastor and scholar Johann Albrecht Bengel, whose 1725 Prodromus and 1734 edition of the Greek New Testament (which included a critical apparatus in which he divided the surviving manuscripts into a debased “Asiatic” [Byzantine] and a pristine “African” [Alexandrian] family of texts), proposed the eventual possibility of constructing a tabula geneologica as a tool for the establishment of the authentic text. In 1831, the German philologist Carl Zumpt published the first known genealogical tree of texts in his edition of Cicero’s Verrine orations, and gave it the name by which it has been known ever since – a stemma codicum. The fullest formulation of the evolutionary tendency in textual criticism is in Karl Lachmann’s 1850 edition of Lucretius, which posited that the erroneous readings that various manuscripts of a given text shared with each other could be used to trace their descent from exemplars in which the errors had first been made, and that by tracing the family tree of errors in reverse, a stemma could be constructed, on the basis of which the editor could establish the original text, freed from all its errors: this method is known as recension. Recension became the standard method for the production of critical editions of text both in the classical and vernacular languages throughout Western Europe in the later nineteenth and early twentieth century. In French and Italian, however, critics noticed that Lachmann’s method was flawed in that it assumed that the errors in manuscript copies of a text were always genetically related (that they descend from the same exemplars, or mastercopies), while it is an observable fact of manuscript transmission that scribes can come up with the same error independently, and that in the history of the transmission of a text, manuscripts often turn out to have been compared and “corrected” against one another, with the result that variant readings characteristic of one manuscript family can appear in manuscripts that do not descend from the same exemplars. In 1928, Joseph B´edier registered a very different objection to the logic of the stemma codicum: he noted that textual histories constructed by the use of recension tended – improbably – always to have two branches, each of which was further divisible into two branches, and so on. For B´edier, the answer was the “best text” edition – an edition based on a particular manuscript that for historical reasons, seemed to be the best representative of the author’s original. This tendency to base textual research on the particular, rather than

Manuscript textuality

233

upon a reconstructed ideal version, has been characteristic of editorial work in the Romance language tradition throughout the twentieth century. In his influential 1972 essay Toward a Medieval Poetics, Paul Zumthor argued that the anonymous character of early medieval French verse – particularly lyric verse – meant that every manuscript version of a poem was like an oral performance by a jongleur. For Zumthor, the work (l’oeuvre, e.g. La Chanson de Roland), being not the creation of an individual, identifiable poetic voice, but the product of recognizable social and cultural forces and relationships, is instantiated equally in each manuscript in which it survives: les textes (les Chansons de Roland). The constant tendency to variation among manuscript versions of early medieval vernacular poetry, as opposed to the more stable forms in which later French poetry survives, he termed mouvance. This idea was carried even further by Bernard Cerquiglini, who, in his enthusiastic paean In Praise of the Variant, observes that variance (his preferred term) was not the product of anonymity or orality in medieval poetry, but was the very condition of medieval textual culture. Far from being the result of anonymity, variance itself produced an attitude toward textual authority, according to Cerquiglini, that was different from that of either classical or modern culture. It would be truer to say that text has always had a centrifugal tendency toward variation, against which there has always been a greater or lesser centripetal cultural impulse toward stability, and that these tendencies play out differently in different media of textual transmission. Important work on textual-critical theory in the field of medieval and modern English literature has been done by W. W. Greg, Fredson Bowers, and G. Thomas Tanselle; but the most influential late twentieth-century position on the textual criticism of medieval English literature was that of the editors of the Athlone Piers Plowman, George Kane, Talbot Donaldson, and George Russell. This edition completely rejected recension as an editorial method because, first, recension – as described by Kane – is based upon a fundamental begging of the question: the editor must first decide which readings are erroneous (i.e. non-authorial) in order to construct a stemma, the purpose of which is to identify non-authorial (i.e. erroneous) readings. Second, Kane’s work is characterized by a fundamental disdain for the intelligence of medieval scribes; and where earlier critics had produced guidelines for the identification of the unconscious errors made in the process of copying a text, Kane produced a list of the kinds of errors introduced by scribes who thought that they could improve on the words of the original author. The intertextual and relational capabilities of electronic media offer an intriguing set of possibilities for the future of textual criticism. An electronic critical edition would offer the reader the ability to see the text in the multiplicity of form and variation characteristic of manuscript culture, without

234

michael g. sargent

the (unspoken) assumption of authoritative textual uniformity that a print edition creates by the very fact of its existence. This need not mean that the function of the editor as the arbiter of the text would disappear, but rather that it would be differently placed. It would be possible, for example, to create an electronic critical edition in which the default form of the text would be the editor’s critical reconstruction of the author’s original – or at least of the archetypal form of the text from which the surviving manuscript witnesses derive. Lemma by lemma (word by word, phrase by phrase, line by line), this text could be linked to transcripts and facsimiles of the manuscripts themselves, to critical editions of particular textual groups, and to a variety of ancillary materials and tools: manuscript descriptions, identification and comparison of scribal hands, dialectal analysis of the surviving copies of a given work, bibliographic studies of the history of ownership and reception of the manuscripts in which it survives and its collocation in the surviving manuscripts with various other works, stylistic and metrical analysis of a given author’s oeuvre as it is represented in the surviving manuscripts, and concordances of the vocabulary of a given author or work – all suggest themselves as possibilities. The postmodern French school of “genetic criticism,” which is primarily concerned with printed texts, is of coincidental interest here.5 Its primary aim is to create a “genetic dossier” recording all of the manuscript materials (the “avant-texte”) leading up to the initial published form of any given text (e.g. the oeuvre of Heinrich Heine, or of Gustave Flaubert). It shares with B´edierist textual criticism – as well as with the postmodern French attitude to manuscript textuality that we can see in Zumthor and Cerquiglini – a focus on the individual documents, rather than the construction of a standard form of the text. The “genetic criticism” differs from the tradition of criticism of manuscript textuality in that, of course, in the case of printed literature, one is reading against the authority of a “standard form” (albeit not nearly as univocal or monolithic as is popularly assumed) that already exists. Electronic media offer the possibility for the production of a form of text that is truly a textus – a fabric, in the original Latin, as well as a page that has been written upon – in a sense that transcends both manuscript and print. Like a manuscript, an electronic edition can present a heteroglossic text: but where the textual polyglossia of manuscripts lies in the fact that each speaks the text with a different voice, an electronic edition can present all of those voices simultaneously. Like a printed critical edition, an electronic edition can also present the editor’s construct of his or her closest approximation to what the author originally wrote, linked word by word and line by line with the manuscripts from which that critical edition derives – but (because the reader can revise the editor’s decisions in his own copy, and pass these

Manuscript textuality

235

revisions along as suggestions to other readers) without the hegemony of an “authorized” text, established once and for all. NOTES 1 Michael D. Coe (London: Breaking the Maya Code (London: Thames and Hudson, 1999). 2 Elizabeth Eisenstein, The Printing Press as an Agent of Change: Communications and Cultural Transformations in Early-modern Europe (Cambridge University Press, 1979); Lotte Hellinga, “Printing”, in The Cambridge History of the Book in Britain, Vol. iii, 1400–1557, ed. Lotte Hellinga and J. B. Trapp (Cambridge University Press, 2000), 65–108. 3 Walter Benjamin, “The Work of Art in the Age of its Technological Reproducibility”, in The Work of Art in the Age of its Technological Reproducibility and Other Writings on Media, ed. Michael W. Jennings, Brigid Doherty and Thomas Y. Levin; trans. Edmund Jephcott, Rodney Livingston, Howard Eiland, et al. (Harvard University Press, 2008), 19–55. 4 Harold Love, Scribal Publication in Seventeenth-Century England (Oxford: Clarendon Press, 1993). 5 Jed Deppman, Daniel Ferrer, and Michael Groden, eds. Genetic Criticism: Texts and Avant-textes (University of Pennsylvania Press, 2004).

11 KARI KRAUS

Picture criticism Textual studies and the image Pictures are special cases. Pictures are problems. Morris Eaves1

In his classic work Prints and Visual Communication (1953), William Ivins vividly describes early efforts to transmit botanical information through the medium of illustration based on testimony dating back to the first century ce. Generalizing from these accounts, he identifies a technological limitation that was to profoundly impede the circulation of knowledge for hundreds of years: the inevitable distortion of subject matter introduced by the manual copying of pictures led many scribes to fall back on the resources of language as their exclusive tool of representation.2 The purpose of this chapter is to inquire more deeply into the nature of this distortion: Is there anything predictable about the way pictures change after several cycles of copying? Have we evolved any methods for notating variants between two or more versions of a picture? Is it possible to try to recover a prototype of an earlier version of a picture from later iterations of it? Such questions are at the heart of a fledgling discipline of “picture criticism,” as the Princeton art historian Kurt Weitzmann called it,3 that does for images what textual criticism has traditionally done for words, namely to provide an adequate scholarly framework for studying their reproduction, transmission, comparison, and – more controversially – their reconstruction. Unlike textual scholarship, however, picture criticism doesn’t yet exist as a self-conscious field or distinct community of practice. For this reason, the approach of this chapter differs somewhat from that of other chapters: it is more speculative – more heavily invested in establishing working assumptions, however provisional, and in proposing an intellectual apparatus and set of analytical strategies that can help readers make sense of prior scholarship in picture criticism and shape an active research program around future efforts. It also attempts to clarify the extent to which textual criticism can reasonably accommodate images, and, conversely, the degree to which they fall outside its explanatory grasp. A mature discipline of picture criticism 236

Picture criticism

237

will need to venture beyond textual studies, drawing not only on art history, but also on cognitive science and mathematics. The comparative method in textual criticism functions as the armature on which to drape a set of ideas and conversations about picture criticism. It includes techniques first developed in the nineteenth century for comparing variants between two or more related texts with the purpose of reconstructing their genealogy and archetype. The comparative method operates through a network of procedures that collectively constitute the field of textual scholarship: replication, collation, reconstruction, and so forth. Those media most susceptible to the comparative method tend to be “allographic,”4 a term coined (or repurposed) by Nelson Goodman in Languages of Art. Allographic systems, unlike autographic ones, resolve into discrete, abstract units of information – into what might be more loosely called digital units – that can be systematically copied, transmitted, added to, subtracted from, transposed, substituted, and otherwise manipulated: examples include alphabetic script, pixels, and numerical notation. Where words are conventionally allographic, images are typically thought to be autographic: they operate through what we now think of as “analog” representational methods, with smoothly continuous rather than discrete and stepwise units of information. The marks through which they are constituted often shade into one another and don’t appear to organize into abstract types whose individual members can be freely exchanged with one another, in the way that the letter “p” represented in Courier font, for example, might be substituted for the letter “p” rendered in Times New Roman without compromising meaning. The last two decades of textual criticism have witnessed a wealth of scholarship contesting the validity of these distinctions and exploring the text’s bibliographic or iconic codes. However, despite the virtues of such visual approaches to textuality (and there are many), a number of the traditional functions of textual scholarship require a different semiotic framework to make them intelligible from a historical and technological perspective. Each of these textual operations thrives in an allographic environment, and is more apt to falter in an autographic one. The next section is devoted to examining the feasibility and consequences of applying the comparative method to images. Focusing on the challenges of image collation, we will look at the artifice involved in anatomizing pictures for the purposes of isolating and comparing variants. The exercise is another version of Zeno’s paradox, in which an arrow can never reach its target because it must always first cross half the distance, then half the remaining distance, and so on. We can never properly reach the goal of comparing two pictorial objects because each variant must first be broken down into sub-variants, and those in turn into sub-sub-variants, and

238

kari kraus

those into sub-sub-sub-variants, and so on, through a never-ending recession. But even this analogy is grossly inadequate, for it suggests that the problem lies in determining the depth of tokenization, when in fact it is tokenization itself that is in question. Each attempt to dissect the image into discrete components is liable to feel arbitrary, subjective, and provisional, because of the lack of a cultural consensus concerning what constitutes an essential picture element. As a result, higher-level procedures of textual scholarship such as comparison and conjecture may prove quite difficult to arrive at. However, rather than leaving the reader with the impression that picture criticism is just textual criticism manqu´e, the second half of the chapter posits a set of ideas and hypotheses that might be further tested, developed, and evaluated at a later date. One promising direction arises from the study of topology, a branch of mathematics that deals with how shapes can be transformed through twisting or deformation. In topological terms, two shapes are equivalent if one can be transformed into another by what Henri Poincar´e calls a “continuous deformation.” If we can examine the transformation of images as shapes that undergo a process of topological deformation, rather than as systems of differential units subject to a small set of formal operations (e.g., substitution, transposition, deletion, and insertion), we may be able to identify new procedures of analysis that carry the spirit of textual scholarship more effectively into the domain of images. It is my hope that these hypotheses are promising enough to steer a nascent science of picture criticism in new directions. One final note by way of prologue. It has been well over a decade since W. J. T. Mitchell first documented the “pictorial turn” in cultural criticism and even longer since Jerome McGann coined the phrase “bibliographic codes” to denote the signifying visual properties of texts.5 Bibliographers, textual scholars, book artists, media theorists, and cultural critics have been shaping an active research agenda around the semiotics of the page for over a quarter century. This agenda is reflected in the contemporary idiom of textual scholarship, where we routinely speak of “visible language,” “the iconic page,” “the word as image,” “the visible word,” or the “the image-text.” Randall McLeod documents the shift more provocatively: “photography killed editing. Period. (Someone has to tell the editors.)”6 But what the camera has taken away, the computer gives back: if McLeod is right that editing was the corpse of the photographic age, then it has become the Lazarus of the digital age. Unediting has, or will soon, give way to deep editing, whose possibilities have increased by many orders of magnitude in our e-milieu. The pervasiveness of advanced information technologies provides an opportunity to revaluate and reassert the digital rather than analogue properties of texts. The objective of this chapter, therefore, is not to organize knowledge

Picture criticism

239

about texts in terms of images, but rather to go against the grain by mapping correspondences in the other direction: from text to image. Comparison The types of images that are relevant to picture criticism are those for which multiple representations exist. Our concern is less with original paintings and drawings than with manual or mechanical reproductions of them, including engravings, slides, photographs, digital copies, or hand-drawn illustrations: pictorial artifacts that persist in time and space by virtue of being iterable rather than durable. It is not the reproductions per se that are of interest so much as the relationships among them; their points of agreement and disagreement; their manner of production, mode of dissemination, and connection to the original. In an influential paper delivered before a digital humanities audience in May 2000, John Unsworth proposed seven “scholarly primitives,” which he glossed as “self-understood” activities and behaviors that are pervasive and seemingly universal across disciplines.7 The list, which is not meant to be exhaustive, includes annotating, discovering, illustrating, and representing. According to Unsworth, collectively these primitives “form the basis for higher-level scholarly projects, arguments, statements, [and] interpretations.” One of the primitives identified in the paper is comparison, an operation so central to textual scholarship that the discipline ceases to be meaningful or functional in its absence. Indeed, if we were to devise a set of primitives that taken together define and structure the field, comparison would preside over all of them. Attempting to distill textual scholarship down to its primitives, I have come up with the following list: r r r r

comparing copying or representing relating (through stemma or family trees) reconstructing

We evaluate the accuracy or fidelity of a copy by comparing it to the original; we determine the genealogical relationships among copies by comparing them and isolating their variants; and we reconstruct an archetype by comparing the surviving manuscripts and inferring their most recent common ancestor. The comparative method thus organizes, enables, and sequences all of the other primitives. Comparison, though, is not a monolithic act that operates indifferently across media. There are two distinct comparative modes, one of which is highly problematic in picture criticism. These two modes are collation

240

kari kraus

and collocation. Because it is common to both fields, collocation will be briefly covered first, followed by collation, a form of comparison whose methods have been largely impracticable in art history, despite interesting and ongoing efforts to adopt them. Collocation To collocate two objects is to place them side by side. Within the context of humanities scholarship, collocation can be defined as arranging texts, images, or other artifacts in such a way so as to facilitate direct comparison. Collocation circumscribes a space for comparison and organizes twoor three-dimensional objects within it. A researcher can purposefully and selectively (or, for that matter, vaguely and indeterminately) move her eyes back and forth between collocated images by virtue of their being adjacent to one another. Despite an intuitive grasp of this process, however, we can’t yet fully explain it from an optical or cognitive perspective, nor can we say how predictable it is. How does the viewer prioritize visual content across pictures? What factors condition this process? How do we perceptually compare regions of interest? Collation To collate two or more documents or objects is to systematically compare them for the purpose of identifying, analyzing, and – perhaps most importantly – notating their points of agreement and disagreement. The ability to express the results of comparison in a concise manner using commonly agreed upon symbols, procedures, and principles is essential to the definition. The field of textual studies, for example, has evolved clear conventions for reporting textual variants in an apparatus: a bracketed lemma followed by readings in other extant witnesses, as in this example from the Arden edition of Shakespeare’s King Lear, which collates early quarto editions (“Q”) against the so-called “First Folio” (“F”): contentious] F; crulentious Q uncorr., Q2–3; tempestious Q corr.8

What traditionally sets textual collation apart from pictorial collation is that the unit of variance is isolable and determinable. In the case of literary works, it takes the form of a letter, a punctuation mark, a word, a finite textual omission or extension. We are able to attend to differences judged more salient (contentious versus crulentious) only by ignoring others judged less so (distinct typefaces, say, in the Penguin and Arden editions of King Lear). Higher-level information is contrastive only insofar as lower-level information is not.

Picture criticism

241

In the visual arts, potential sources for collated information include illustrated descriptive catalogs and catalogues raisonn´es, which are authoritative listings of a single artist’s work. The compiler of a catalogue raisonn´e for a printmaker will usually attempt to identify and document multiple states of an etching, engraving, or woodcut. “State” is the technical term given to each of the successive stages of a print. When an artist has finished working his or her plate and pulls the first proofs, these are said to constitute its first state. Thereafter, each time the plate is retouched and impressions from it pulled, the resulting prints are said to reflect different, consecutively numbered states. Feeding the art market’s demand for such artifacts, catalogers obsessively document states and inventory their variants. Even a casual survey of available descriptive catalogs for a graphic artist like Rembrandt, some of whose etchings exist in upwards of ten states, reveals the absence of clear formulae for recording image variants. Instead, catalogers of Rembrandt’s works tend to adopt ad hoc methods and styles of collation. The lack of any real set of conventions in such documentation suggests at the very least that the variant has a different (albeit no less important) status in the visual arts than the verbal. If we confine our analysis of Rembrandt catalogs to those published in the nineteenth and twentieth centuries, what becomes clear is that the relationship between collation and collocation fundamentally shifts over time. In the nineteenth century – prior to the rise of photomechanical reproduction – the descriptive catalog was generally devoid of images, and consequently collocation played little or no role in the presentation and analysis of variants. Instead, description and collation combined to serve a substitutive function, standing in for the missing visual objects. A catalogue raisonn´e from this era can thus be considered an ekphrastic work, in a broad sense: a verbal rendering of a visual artifact. As an example, consider the following entry from Charles Henry Middleton’s nineteenth-century catalog of Rembrandt’s etchings, which describes a print of the legendary Dr. Faustus: He is represented standing in his laboratory on the left side of the print. He wears a white cap and academical gown. His writing-desk lies before him upon a table, on which his closed right hand is resting, while the other is placed upon the left arm of the chair, from which he has just risen, attracted by the sudden appearance of a luminous magic circle in the centre of a casement to the right. He is apparently watching the movement of a shadowy hand which points to a reflection of this circle in a mirror held by another hand below. Lower down, on the right, a pile of books lies upon the table, and below, in the right corner, is the upper half of a globe. In the left background are a shelf of books, an hour glass, a skull, etc., while many sheets of paper fastened together hang by the upper part of the casement.9

242

kari kraus

To make sense of Middleton’s distinctive rhetorical strategy, it is helpful to consider it within the context of the various categories of visual description developed by art historian Erwin Panofsky in the first half of the twentieth century.10 Composed of three levels, Panofsky’s hierarchy consists of a literal interpretation of pictorial content (the pre-iconographic level) followed by increasingly more symbolic interpretations (the iconographic and iconological levels), the last of which ambitiously seeks to incorporate knowledge about the civilizations that produced the images. In the quoted passage by Middleton, visual meaning is captured almost exclusively at the preiconographic level: a primary level of factual analysis that is characterized by generic information about the people, places, and things represented in the image.11 Middleton’s enumeration of the desk, chair, books, hourglass, and skull in the Dr. Faustus print makes this a quintessentially pre-iconographic description, likely to strike a reader who has grown up in a cultural milieu saturated by images (and for whom, therefore, Middleton’s approach is alien) as peculiar in its singular focus on object identification rather than formal interpretation. Compare it to the following annotation of the same print, published almost 100 years later, which contains no pre-iconographic information whatsoever: The print is sometimes identified with “the practicing alchemist” in Clement de Jonghe’s inventory of 1679; in Rover’s inventory of 1730, and in the Huls sale catalogue of 1735, it is described as Dr. Faustus. Wolfgang Wegner . . . points to similarities with the engraved title-page of a book on magical practices of 1651.12

This description, published in Christopher White and Karel Boon’s catalog of Rembrandt’s etchings in the second half of the twentieth century, is concerned exclusively with conventional subject matter: it attempts to classify the predominant figure in the print according to a standard type (“the practicing alchemist”) and to assign it a literary identity (“Dr. Faustus”). White and Boon’s analysis is decidedly iconographic,13 never trespassing into the pre-iconographic zone that distinguishes Middleton’s description. The reason for the difference is as simple as it is consequential: the first description functions as a substitute for the image, the second as a supplement to it (White and Boon include relatively high-quality black and white photographs of the print in Volume II). Taken together, these examples demonstrate our shifting reliance on language to communicate visual information. With the emergence of cheap photomechanical reproduction methods, and the increased use of images in printed materials, descriptive text gradually came to fulfill a variety of auxiliary functions with respect to images beyond that of representation.

Picture criticism

243

11.1 Rembrandt, Faust (c. 1652). Ink on paper/Dry-point, etching, and engraving, 21 × 16 cm. Public domain image.

Turning our attention from description to collation, we find that although both catalogs devote considerable space to documenting variants, Middleton’s annotations are more copious than White and Boon’s, a difference once again attributable to the significant role played by photographic evidence in the latter work. Because White and Boon provide impressions of multiple

244

kari kraus

states whenever possible, collocation inevitably modulates collation but does not replace it since the two modes of comparison are not interchangeable. Despite these contrasts, what is striking about both catalogs is the challenge of determining the appropriate unit of analysis for recording variants and drawing the boundary between significant and insignificant difference. Reflecting the assumption that no detail – however minute – is inconsequential, both catalogs err on the side of inclusivity of information. Here, for example, is Middleton’s collation of three known states of the Dr. Faustus print: States : 1st. An upright book is seen to the right, partly cut through by the plate-edge. It is shaded by regular diagonals from right to left, not very close, nor running into each other . . . 2nd [state]. Additional very fine and close lines and interlines shade the book in the same direction as the lines in the 1st State. These close lines frequently touch. Patches of similar lines strengthen the shadow on the shoulder, compensating to some extent for the removal of the burr . . . 3rd [state]. The book is still further shaded at the top and a little way down the side by harsh diagonals right to left, and down the middle of the book is a series of regular horizontals . . . Fine horizontal interlines deepen the shading behind the shoulder to the left; equally fine downward work shades the book.14

It is not iconographic or pre-iconographic variance that is captured here, but rather perceptual variance of a more elemental order, one that is excluded from Panofsky’s typology. Lines and interlines, burr and diagonals: these are the formal properties – all related directly to intaglio processes – that guide Middleton’s collation. Such properties are abundant and technically subtle, occasionally bordering on infinitesimal. Elsewhere, Middleton expands his analysis beyond changes of state, sporadically reporting, for example, on where impressions diverge from one another in terms of paper texture, color, or inking method. The end result is that the catalog’s object of collation becomes a moving target. The struggle to define the depth of analysis is particularly evident in Middleton, for whom attention to descriptive minutiae is the marker both of his own connoisseurship and of the inexhaustibility of meaning in the details of the image itself. At the same time, he is deeply concerned to defend the project against imputations of both pedantry and pure subjectivity: The amateur whose well-directed taste leads him to look for excellence alone may think the time misused that is devoted to the task of recording minute and, in his sense, utterly unimportant peculiarities or variations . . . The student who would really command a knowledge of his subject must be content to descend to particulars, must train his eye to remark distinctions and differences which, if sometimes minute, are sometimes important; must learn what to admire,

Picture criticism

245

and be able to fix the standard by which he decides. To acquire this power unaided is no easy task, and he who attempts to impart it should make it his first endeavour to be exact.15

Here Middleton anticipates charges of pedantry and deflects them by arguing that attention to minutiae is warranted by the fact that even the humblest visual mark can serve as a conduit of meaning. It is a rookie mistake, he suggests, to assume that a metal shaving or an attenuated line has no semantic significance. At the same time, Middleton mystifies the nature of that meaning by insisting that it can only be detected and interpreted by the connoisseur. In discerning intrinsic meaning in the most rudimentary and microscopic of perceptual marks, Middleton would seem to imply that all variation is significant – that there is no such thing as noise in the visual system that may be safely disregarded for purposes of collation. This fundamentalist stance helps explain why it is sins of omission that Middleton ultimately fears, rather than sins of commission. It also makes clear that the underlying model of similarity we adopt when comparing images has direct implications for collation. If the model does not presuppose an overall state of similitude against which difference operates; if it lacks a framework for decomposing images into constituent parts; or if it fails to include provisions for weighting variants or treating some as equivalent members of the same pictorial type or element, despite a few outward differences in detail, such as thickness or depth of line – then the task of collation and the project of picture criticism will be rendered that much more difficult. Moreover, any alternative model to thinking about comparison, transformation, and reconstruction will need to avoid falling prey to the fetishization of difference.

Pictorial transmission: Bartlett’s game of visual telephone Up to this point, allography has been privileged as a semiotic system because it imposes formal criteria that allow us to filter significant properties of texts from insignificant ones. This ability to filter is what regulates many of the canonical practices of textual scholarship, including reproduction, collation, and conjecture. In trying to assimilate pictorial information into this predominantly textual model, however, we risk misrepresenting pictures as a poor man’s words. Might we reach some preliminary conclusions about the significant characteristics of images without taking recourse in loaded terms such as “syntax,” “grammar,” or “language”? Rather than artificially segmenting pictures into discrete, digital units, are there other ways we can investigate them that will effectively advance a science of picture criticism that is compatible with their distinct ontology? In this section, we will look at promising directions for future research coming out of the science lab.

246

kari kraus

Writing about the effects of serial copying on the representation of botanicals in the fifteenth century, William Ivins attempts to characterize the nature of the changes that this class of information cumulatively underwent over time. Transmitted via woodcuts, the illustrations “constitute a remarkably sad example of what happens to visual information as it passes from copyist to copyist”: These herbals, beginning with the Pseudo-Apuleius of about 1480 and coming down through the Grete Herbal of 1526, are extremely interesting from still another point of view. When arranged in families and in a time order they clearly show the operation of what I suppose is one of the basic human characteristics. So long as the illustrators did not return to the original plants as sources of information about their shapes, but confined themselves to such knowledge of the forms as they could extract from pictures made by earlier men – to what may be called hearsay and not first-hand evidence – it was inevitable that they should rationalize their own pictorial accounts and overlook or disregard what appeared to them to be mere irrationalities in the pictorial accounts given by their predecessors. This rationalization most frequently took the form of an endeavor for symmetry, which produced regular shapes that not only lost all verisimilitude of lines and edges but introduced a balanced arrangement of parts and forms, which, however satisfying to mental habits, resulted in a very complete misrepresentation of the actual facts.16

The trend toward symmetry, rationalization, and conventionalization that Ivins bemoans in this passage was documented by the British psychologist Frederic Bartlett two decades earlier, in 1932. Soon after founding a program in experimental psychology at the University of Cambridge, Bartlett published a highly influential book on cognition, described in the preface as a “laboratory consideration of memory,” which takes up “the important group of mental processes . . . usually included under the term ‘remembering’.”17 Through a series of experiments designed to chart the fate of cultural material transmitted via memory, Bartlett demonstrated, with vivid empirical evidence, something we now take as axiomatic: the human organism’s propensity to dramatically transform the raw input of perception and the stored contents of the mind. His findings largely corroborate Ivins’s views, albeit with some important revisions and qualifications. While Bartlett’s intention was to study perception and recall separately, his hypothesis, confirmed in the course of his experiments, was that the two were profoundly inter-related. In The Method of Perception, test subjects were briefly exposed to visual patterns and designs that ranged from the very simple to the more structurally complex. They were then asked to immediately reproduce or describe what they had seen.18 Similarly, in the Method of Description – designed to test the accuracy of recall over longer periods of time – subjects were shown a series of drawings of faces (all men in

Picture criticism

247

uniform) and subsequently asked to describe them after various intervals (after a lapse of thirty minutes, then again two weeks later, then again a week later, and so forth).19 Other studies had volunteers read a folktale or descriptive passage, and afterwards – in a process reminiscent of the children’s game known as “Telephone” or “Chinese Whispers” – attempt to retell it, with each retelling forming the basis for the next volunteer’s version, thereby creating “chains of reproductions.”20 Most relevant within the context of this essay, Bartlett adapted the exercise to include graphical content.21 While the results of all the experiments cross-reference and reinforce one another, it is the game of visual telephone that speaks most directly to Ivins’s concerns. In what follows we’ll focus on key findings that cut across all the studies, and then have a closer look at the Method of Serial Reproduction, which incorporates pictorial material. A tacit question that underlies Bartlett’s research is this: should those properties that remain relatively intact or resistant to change be considered more determinative of a cultural object’s core identity than other properties? For example, if a picture is repeatedly copied by hand, should we infer that attributes remaining stable across iterations are more culturally or aesthetically significant than those which are omitted or altered? Bartlett’s experiments repeatedly show that while the various parts of an image or text change at different rates, we should be wary of establishing a connection between content fixity and relevance. Indeed, the opposite appears to be true: transformation rather than stability serves as a better indicator of importance. If there is a general law to extract, it is that significant properties persist, but in a modified form: a conflict or battle in a story might undergo embellishment; a schematic feature of a face might become more representational; a prominent line or pictorial detail might be transferred to a new location. “The transformation of material,” writes Bartlett, “which is constantly illustrated in recall, occurs most frequently in connexion with the details which individual interest tends to make salient” (my emphasis).22 What this line of inquiry (by Bartlett, Ivins, and others) yields is a set of principles informing the transmission of images: r Repeated copying increases the prominence of certain details, with a concomitant downgrading, dilution, or even omission of others. r The trend in repeated copying is toward stereotyped and conventional reproductions whereby an individual rationalizes the novel, unfamiliar, or exotic element in the direction of the known and familiar. This phenomenon accords with the textual principle of lectio difficilior or “the more difficult reading” – the rule that when adjudicating variants, the editor should favor the more difficult reading on grounds that scribes would tend to normalize complex authorial passages. In a similar vein,

248

kari kraus

art historian Jonathan Alexander has noted that a medieval illuminator might draw on a visual or mental repertoire of stock scenes, character types, and poses to illustrate his text. A falling horse, a sleeping figure, or a knight in arms, for example, would be generically drawn, irrespective of context or circumstance. The same reclining figure might represent a sleeping shepherd in one instance or the biblical Adam in another. Any distinguishing traits would be lost as a result of conventionalization.23 r In serial reproduction, conventionalization and simplification of detail frequently alternate with elaboration and amplification of detail. r One picture may have a priming effect on another, causing a copyist to blend and confuse details in a pictorial series when asked to reconstruct it from memory. A related phenomenon is the visual conflation of two or more sources in a single image – a practice with obvious textual correlates. By way of illustration, in his discussion of the method of serial reproduction (or visual telephone), Bartlett offers a sequence of images produced by volunteer subjects in the lab (Figure 11.2). The first drawing in the sequence is the Egyptian hieroglyph for “owl,” from which the Latin letter “M” is thought to derive. The most conspicuous change is the change in referent as the owl gradually metamorphoses into a cat. The process by which this occurs is one of oscillation between simplification and elaboration: over several cycles of reproduction, the drawing first loses all pretense of representation, becoming increasingly nonrepresentational, before detail is gradually added back in to re-shape the amorphous form into a cat, at which point conventionalization takes over. The tendency to multiply and transfer parts is also evident. The migration and reduplication of the bird’s stylized wing, for example, is illustrative: by drawing number 4, it has been cloned and repositioned on the figure, and by drawing number 9, it denotes a tail rather than a wing, whose location and direction change at least two more times before the series concludes. Bartlett underscores the importance of this type of alteration by noting that multiplication of detail “is present in every single series” that he has collected.24

Caricature and pictorial transmission: D’Arcy Thompson’s method of Cartesian deformation One of Bartlett’s arguments is that test subjects tend to treat pictures as “unitary patterns,”25 not as composite objects made up of distinct parts, despite some evidence to the contrary, such as their privileging certain details over others, as described above. In a similar vein, the twentieth-century biologist D’Arcy Thompson theorized that a holistic rather than atomistic

Picture criticism

249

11.2 Frederic Bartlett, experiment in serial reproduction (1932). Copyright Cambridge University Press.

250

kari kraus

approach to image manipulation could help us understand the evolution of living organisms. Apropos of his dual background in the humanities and sciences, Thompson drew inspiration from the Renaissance artist Albrecht produced the Durer and his geometric alterations of human faces.26 Durer ¨ ¨ caricatures of human heads shown in Figure 11.3, for example, by changing the dimensions of the grid sections on which they are inscribed. These change operations are topological in nature: they involve warping the image along the vertical axis to expand or contract, elongate or squash the features. Thompson postulated that these artistic processes could teach us about biological processes; that we could learn something about evolutionary descent with variation by analyzing the transition from one species to another in terms of topological deformations. His work reminds us that just as biologists in the nineteenth century have borrowed stemmatic methodologies from textual scholarship, so too have they occasionally found inspiration in the visual arts. As astounding as it may seem, the takeaway here seems to be that mechanisms of biological change allegedly resemble caricature. Thompson’s objective, then, was to put Durer’s system to use in the service ¨ of science. Here (see Figure 11.4), he shows us the transformation of one genus of fish into another closely related one.27 The distortion is achieved by modifying the angle of the grid lines and stretching the image to fit its new frame. While the resulting differences between the two drawings certainly warrant attention, the similarities are just as striking. For the topologist, those properties that are preserved under various transformations are as mathematically interesting as those that are not. Such invariance is a function of the rubber sheet techniques that govern the science of topology: the attributes that are most durable are those that would remain constant even if drawn on an elastic sheet that can be stretched in various ways. For example, any simple geometric shape – a circle, a triangle, a square, and so forth – is topologically equivalent to any other, since in all cases they consist of a closed boundary in two dimensions. Topology thus promotes certain kinds of shape properties (such as order, connectivity, and enclosure) and demotes others (such as length, area, verticality, horizontality, and angularity).28 These topological values in turn establish basic criteria by which similitude between two or more pictures may potentially be evaluated. The cognitive primacy of topological content invites us to consider the extent to which this category of information is retained across multiple cycles of pictorial reproduction. With respect to Thompson’s theories, it also begs the question of whether or not image distortion can be profitably understood in terms of caricaturization. Do the topological exaggerations apparent in Thompson’s grid experiments correlate in any way with Bartlett’s

Picture criticism

251

¨ 11.3 Albrecht Durer, selections from Vier Bucher von menschlicher Proportion (1528). ¨ Public domain image courtesy of the US National Library of Medicine.

252

kari kraus

11.4 D’Arcy Wentworth Thompson. A cartesian grid transformation of Argyropelecus olfersi into Sternoptyx diaphana (two related species of fish). Copyright Cambridge University Press.

observation that dominant details tend to “suppress” or “overmaster” other details as pictures are continually redrawn from memory, either by the same hand or by different hands? Clearly the precise function that visual amplification serves in serial reproduction is an area ripe for further investigation, bearing in mind that it is also the sine qua non of caricature.

Pictorial asymmetries: Michael Leyton’s method of recovering history from shape Sharing with Thompson an interest in topology as an explanatory device for cultural and biological change, the cognitive scientist Michael Leyton has proposed a set of rules for reconstructing the history of a two- or threedimensional object based on its current spatial configuration.29 Leyton’s ingenious method involves exploiting the asymmetries (or “distinguishabilities”) of a given shape by topologically manipulating them until they have been erased.30 The resulting symmetrization represents – according to Leyton – the restoration of a prior state. The type of asymmetry of interest to Leyton is the so-called curvature extremum, the point of greatest bend in a curve. Following Leyton’s theories, we might say that the extent of distortion of the right-hand fish indicates the presence of a force, procedure, or event allegedly responsible for creating this distortion: actions which result in physical changes at the topological level such as flattening or extrusion. In principle, therefore, we can imagine reversing these same actions, stretching and pulling the object in order to convert each asymmetry to a symmetry, thereby moving backwards in time to an earlier state of the image. Once all the asymmetries have been removed, we are left with a “symmetrical” fish; the distorted shape carries the record of the causal history that produced it.

Picture criticism

253

In 2000, cognitive scientists Douglas Vickers and Adrian Kym Preiss published the results of an experiment that attempted to link the theories of Michael Leyton with the methodologies of Frederic Bartlett: An experiment, modeled on Bartlett’s (1932) method of serial reproduction, was carried out, in which 44 observers, tested in five groups of 4 to 13, were asked to reproduce briefly presented, irregular heptagons, drawn randomly from an original pool of 168 figures. Observers were then presented with each other’s (randomly allocated) reproductions and asked to reproduce them. This process was repeated until each observer had made 20 reproductions.31

In agreement with Leyton’s theorems, the authors observe that test subjects cumulatively symmetrized the figures that they copied. It is as if the act of perception involved recovering some primordial platonic state, and it is this state – not the empirically sensible one presented to the copyists – that is captured in the drawings. Such platonic forms represent, according to Leyton, “undifferentiated eternity from the past into the future.”32 Put differently, they are atemporal, since an “undifferentiated” continuum effectively negates the fourth dimension. The results, moreover, confirm Ivins’s assertions – previously quoted – that pictures which undergo manual transmission are regularized. At the same time, the findings aren’t entirely consistent with Bartlett’s. Some extrema in Bartlett’s experiments seem to be particularly resilient to change, while other extrema diminish, intensify, migrate, or fuse with remaining asymmetries. More experimental studies are needed to clarify and refine these preliminary observations. The strengthening of some pictorial asymmetries at the expense of others would better explain the “caricature” effect (Bartlett’s “over-potency of detail”) previously discussed. It would also imply the inverse of Leyton’s theory of perception, for it points to a perceptual process that involves moving forward rather than backward in time. We are potentially confronted with two distinct regimes operating simultaneously in the act of copying, one recovering history in the form of symmetries, the other deepening the effects of time in the form of asymmetries. Conclusion In The Archimedes Codex – the story of the earliest surviving manuscript of the great mathematician of antiquity – Reviel Netz, Professor of Classics at Stanford University, struggles to convey his sense of awe at the discovery of the battered, time-worn palimpsest, which afforded him the unique opportunity to apply the comparative method to Archimedes’ diagrams.33 Although the palimpsest is the oldest surviving witness to the acclaimed

254

kari kraus

scientist’s work, an interval of more than 1,000 years most likely separates it from any authorial copy. “If our evidence is so far from the original,” Netz asks, “what chance do we have of ever getting there? How can we realistically hope to know what ancient diagrams looked like?”34 For Netz, the conjectural methods of textual scholarship provide the means to telescope time and space, bringing us “to the shores of Syracuse – face to face with Archimedes.”35 Indeed, his confidence in the approach leads him to conclude that his reconstructed diagram of Sphere and Cylinder I is “identical to the one drawn . . . by Archimedes himself.”36 For Netz, this possession of the past – this communion with the dead – is heightened by the fact that what is ostensibly recovered is not a linguistic symbol, but rather a drawing: “There is something particularly ‘tangible’ about diagrams,” he notes, “Words are conceptual, but drawings are physical – they are bodily. This is how he, Archimedes, traced his figure, turning a stick in his hand.”37 Netz’s account of his work with the medieval Archimedes codex provides insight into why scholars continue to be animated by the prospect of picture criticism, despite the fact that it has never achieved the status of a recognized discipline. Those who venture into this area do so in the absence of standard ports of call – extended case studies, dedicated research journals, established methodologies and curricula – that would allow them to position their own efforts within a larger scholarly and historical framework. The allure of picture criticism, however quixotic it may appear, is the same as that of history more generally: to go beyond the vanishing point of the perceptible to retrieve a representable past. While textual scholarship can to some extent serve as a model, it is deficient by virtue of being grounded in an ontology of the text (one of the ironies of Netz’s work, in fact, is that he is forced to disavow the very pictorial properties of his diagrams he elsewhere celebrates in order to make them conform to the stemmatic method). The wider vistas needed to advance a science of picture criticism will need to be supplied, at least in part, by art history, cognitive science, and mathematics. The clinical findings of Frederick Bartlett, for example, suggest that serial copying of pictures introduces systematic and predictable changes, some of which have textual counterparts (for example, the duplication of detail in pictures corresponds to the phenomenon of dittography in texts, while the relocation of detail is similar to the transposition of letters). The topological principle of transformation, by contrast, is peculiar to pictures, and since it furnishes some of the concepts crucial for understanding Michael Leyton’s theories of shape reconstruction, it warrants further study. Finally, picture collation remains an elusive goal, but recent research funded by the Office of Digital Humanities at the National Endowment for the Humanities to develop software for automatically analyzing and collating images is a promising and welcome development – perhaps a sign of things to come.38

Picture criticism

255

Still, the question persists: why a chapter on picture criticism in this volume – rather than, say, the Cambridge Companion to Art History? To answer, consider for a moment Borges’s Ireneo Funes, a fictional character possessed of a memory so detailed and an eye so acute that he can perceive vineyards and grapes in a goblet of wine. If the methods of textual criticism cannot provide an exact template for picture criticism, they can nonetheless serve, like Borges’s Funes the Memorius, as a way of imagining conjectural ways of reading images as well as texts. Just as the disposition of letters and words on a page can tell us something about their prior configurations, so too can the shapes and irregularities of a picture harbor information about the past. Perhaps even more importantly, textual scholarship teaches patterns of thought that help us reckon with “deep time,” time measured in intervals of tens, hundreds, or even thousands of years. It is these patterns and reservoirs of thought – and the value system that accompanies them – that makes textual scholarship a matter of such obvious import to picture criticism. Entangled in the same skein of knowledge, each illuminates and enriches the other.

NOTES 1 Morris Eaves, “Graphicality: Multimedia Fables for ‘Textual’ Critics,” in Elizabeth Bergmann Loizeaux and Neil Fraistat, ed. Reimagining Textuality (University of Wisconsin Press, 2002). 2 A tool that ultimately proved to be too blunt an instrument for its intended purpose. See William Ivins, Prints and Visual Communication, 1st edn. (Harvard University Press, 1953), 13–15, 33, 40–2. 3 Kurt Weitzmann, Illustrations in Roll and Codex: A Study of the Origin and Method of Text Illustration, Studies in Manuscript Illumination, No. 2 (Princeton University Press, 1947), 182–92. 4 Goodman, Languages of Art, 2nd edn. (Indianapolis: Hackett Publishing Co., Inc., 1976). 5 W. J. T. Mitchell, Picture Theory: Essays on Verbal and Visual Representation (Chicago University Press, 1994) 11; Jerome McGann, The Textual Condition (Princeton University Press, 1991), 57. 6 Random Cloud [i.e., Randall McLeod], “Tranceformations in the Text of Orlando Furioso,” in Dave Oliphant and Robin Bradford, eds., New Directions in Textual Studies, (Harry Ransom Research Center and University of Texas Press, 1990), 72. 7 John Unsworth, “Scholarly Primitives: What Methods Do Humanities Researchers Have in Common, and How Might Our Tools Reflect This?” Symposium on Humanities Computing: Formal Methods, Experimental Practice, King’s College, London, May 13, 2000 http://people.lis.illinois.edu/∼unsworth/ Kings.5–00/primitives.html 8 King Lear, ed. Kenneth Muir, Arden Shakespeare 2nd edn. (London: Routledge, 1994; Methuen, 1972), 107.

256

kari kraus

9 Charles Henry Middleton, A Descriptive Catalogue of the Etched Work of Rembrandt Van Rhyn (London: John Murray, 1878), 275. 10 Erwin Panofsky, Studies in Iconology (1939; Oxford: Westview Press, 1972). 11 Panofsky, Studies, 5. 12 Christopher White and Karel Boon, Rembrandt’s Etchings: An Illustrated Critical Catalogue, Vol. i (Amsterdam: Van Gendt and Co., 1969), 124. 13 Panofsky, Studies, 6–7. 14 Middleton, Descriptive Catalogue, 276. 15 Ibid., x–xi. 16 Ivins, Prints, 40–2. 17 Frederic Bartlett, Remembering: A Study in Experimental and Social Psychology (Cambridge University Press, 1932), v–vi. 18 Ibid., 16–33. 19 Ibid., 47–62. 20 Ibid., 118–76. The phrase “chains of reproductions” appears on p. 119. 21 Ibid., 177–85. 22 Ibid., 62. 23 For discussion and examples, see Jonathan Alexander, Medieval Illuminators and Their Methods of Work (Yale University Press, 1992) 85–9, 112, 114–15, and passim. 24 Bartlett, Remembering, 181–2. 25 Ibid., 25. 26 Thompson, On Growth and Form (Cambridge University Press, 1961), 290–2. 27 Ibid., 299. 28 See Patrick Maynard, Drawing Distinctions: The Varieties of Graphic Expression (Cornell University Press, 2005), 55–6. 29 Leyton, The Structure of Paintings (Vienna: Springer, 2006) and Symmetry, Causality, Mind (Cambridge, MA: MIT Press, 1992). 30 For discussion, see Leyton, Symmetry, 1–37 and Structure, 1–42. 31 Douglas Vickers and Adrian Preiss, “Transformational Analyses of Visual Perception,” Proceedings of the Twenty-Second Annual Conference of the Cognitive Science Society, August 13–15, 2000, University of Pennsylvania, ed. Lila Gleitman and Aravind Joshi (London: Psychology Press, 2000), 1063. 32 Leyton, Symmetry, 35. 33 Reviel Netz and William Noel, The Archimedes Codex (Cambridge, MA: Da Capo Press, 2007). 34 Ibid., 94. 35 Ibid., 100. 36 Ibid., 99. 37 Ibid., 100. 38 Randall Cream, et al., The Sapheos Project January 27, 2009, http://sapheos.org/ technicalwhitepaper.html

12 MATTHEW G. KIRSCHENBAUM AND DOUG RESIDE

Tracking the changes Textual scholarship and the challenge of the born digital

Word processors of the Gods William Gibson, in his 2003 novel Pattern Recognition, allows Ngemi, collector and connoisseur of antiquarian computers, to have his joke: “I am negotiating to buy Stephen King’s Wang,” he announces deadpan. “The provenance,” Ngemi continues, “is immaculate, the price high, but, I believe, reasonable. A huge thing, one of the early dedicated word processors.”1 King, in fact, acquired his Wang word processor in 1983, the same year Time Magazine dubbed personal computers the “machine of the year.” The Wang would have included a Z80 processor and 64K of RAM. Clearly the technology was tantalizing to him. The following year, in the introduction to his short story collection Skeleton Crew, King wrote: In particular I was fascinated with the INSERT and DELETE buttons, which make cross-outs and carets almost obsolete . . . . I thought, “Wouldn’t it be funny if this guy wrote a sentence, and then, when he pushed DELETE, the subject of the sentence was deleted from the world?” Anyway, I started . . . not exactly making up a story so much as seeing pictures in my head. I was watching this guy . . . delete pictures hanging on the wall, and chairs in the living room, and New York City, and the concept of war. Then I thought of having him insert things and having those things just pop into the world.2

“Word Processor of the Gods,” the story that resulted, is about a writer who is given a homebrew word processor as a gift from his favorite nephew shortly before the latter is killed in a car accident. Our protagonist, Richard, discovers that the machine has precisely the awesome (or ominous) powers described above: writing about a person or thing can bring it to life in the real world, and deleting the passage likewise causes the person or object to be summarily “erased” from existence. In this manner, before the juryrigged device burns itself out from overload, Richard divests himself of an indolent wife and spoiled teenage son, replacing them with the departed nephew and the nephew’s mother, who turns out to have been his High School sweetheart. Hardly divine stuff as King himself admits, but “Word 257

258

matthew g. kirschenbaum and doug reside

Processor of the Gods,” which first appeared in Playboy in 1983, is one of the first fictional responses to the sudden and rapid domestication of the electronic word. Tellingly, it is also a fiction about a writer who discovers in the technology the means to edit his own life. A year earlier, in 1982, James Fallows, a journalist, had produced a long essay for the Atlantic entitled “Living with a Computer” (in his case a Processor Technology SOL-20).3 For many of the magazine’s readers it would have been their first in-depth exposure to a personal computer, and Fallows takes nothing for granted as he details its operations. “When I sit down to write a letter or start the first draft of an article, I simply type on the keyboard and the words appear on the screen.” Like King, he is also struck by the dramatic way in which text can be purged: “[E]ach maimed and misconceived passage can be made to vanish instantly, by the word or by the paragraph, leaving a pristine green field on which to make the next attempt.” (Later in the piece we learn that to accomplish this “instant” feat he must first place special marker characters at the beginning and end of the passage in question, then execute a series of chorded keystrokes – a procedure unthinkably time-consuming by current standards.) Both King and Fallows’s accounts recall a moment when authors of all stripes were on the lookout for opportunities to work some description of their new tool (or toy) into their prose. Tom Clancy, in the opening pages of his breakthrough techno-thriller The Hunt for Red October (1984), narrates a precocious toddler dumping thousands of words of text from the hero’s Wordstar system. In 1986, Arthur C. Clarke offered up “The Steam Powered Word Processor,” which details the exploits of the Reverend Charles Cabbage (obviously a stand-in for Babbage) and his Word Loom, an ambitious contraption modeled on a pipe organ for automating the composition of sermons. In Foucault’s Pendulum (1988), Umberto Eco begins the plot with files found on the word processor of one of his protagonists. The machine is kabbalistically named Abulafia (“Abu”) and Eco treats us to several exuberant pages of what is really his own celebration of the new technology: “[I]f you’ve written a novel with a Confederate hero named Rhett Butler and a fickle girl named Scarlett and then change your mind, all you have to do is punch a key and Abu will global replace the Rhett Butlers to Prince Andreis, the Scarletts to Natashas, Atlanta to Moscow and lo! you’ve written war and peace.” And later: There, indiscreet reader: you will never know it, but that half-line hanging in space was actually the beginning of a long sentence that I wrote but then wished I hadn’t, wished I hadn’t even thought let alone written it, wished that it had never happened. So I pressed a key, and a milky film spread over the fatal and inopportune lines, and I pressed DELETE and, whoosh, all gone.4

Tracking the changes

259

There is a real-life counterpart to this text in the archives at the Harry Ransom Center, where playwright Terrence McNally has deposited an electronic file that is a stream of consciousness rendering of his first encounter with Word Perfect (in 1988).5 He comments, among other things, on his discomfiture in watching the text slide off of the screen. Another similar electronic document exists within Salman Rushdie’s born digital papers at Emory University. Authors were moved not only to write with their personal computers, but to write about writing with them. References to word processing and other new technologies also began seeping into interviews with contemporary writers. The intrusion of consumer electronics brand names and technical specs into such venues can be jarring, yet they also spur unexpected insight. Joan Didion, in a 1996 interview for Salon.com, was asked about the particulars of her hardware, which turned out to be an IBM ThinkpadTM . She then added: Before I started working on a computer, writing a piece would be like making something up every day, taking the material and never quite knowing where you were going to go next with the material. With a computer it was less like painting and more like sculpture, where you start with a block of something and then start shaping it.6

Some writers like Didion embraced the new technology, finding in word processing the kind of spontaneity and fluidity that Walter Ong would characterize as “secondary orality.” Others were clearly skeptical and shunned computers as engines of abstraction, the mere act of pecking at a keyboard unleashing unfathomably intricate acts of mediation – voltage differentials that are converted to signals for an electron gun to paint phosphors with, the radiant text forever out of human reach behind the cool, uniform surface of the glass. Just as the above examples show how computers challenged and changed a writer’s relationship to his or her own words, so too do they now challenge the textual scholar’s relationship to writers and their written texts. Textual scholars will now need to contend with variants and versions created and destroyed as whimsically as Eco narrated. They must learn not only about word processors but also about email and iPhones and Web 2.0. Authors routinely maintain personal Web sites, including blogs and Facebook pages, often as part of their publisher’s marketing and outreach. Random House now offers all its authors a turnkey “toolkit” for constructing their online presence: Authors can then select different types of content to add to their pages, such as profile or biography information, links to favourite sites, audio and video clips, book reviews, bibliographies, photo galleries, blogs and newsletters. The web pages will be hosted on a community-based website called AuthorsPlace

260

matthew g. kirschenbaum and doug reside

and once authors have created their web pages they can choose whether to interact with other authors on the site, or whether to use their pages as a standalone website.7

What scholar of American letters would not want to read Walt Whitman’s MySpace page, if such a thing existed? Yet efforts to collect, preserve, and curate this kind of born digital literary material are only in their infancy, with near universal consensus among archivists and information scientists that the practical problems involved are growing almost daily with every permutation of new hardware technology and online Web service. The fact is that nearly all literature today is born digital in the sense that it is composed on the screen, saved on a hard drive (or other electronic storage media), and accessed as part of a computer operating system. True, some writers will still employ longhand or even mechanical typewriters as a step in their composition process, but sooner or later the text will be keyed into a computer, almost always to be further revised. Editors edit electronically, inserting suggestions and emendations and emailing the file back to the author. Publishers use electronic typesetting and layout tools, and only at the very end of this process is the electronic text of the manuscript (by now the object of numerous transmissions and transformations) fabricated as a printed book. In the particular realm of literary and textual scholarship, this means that a writer working today will not and cannot be studied in the future in the same way as writers of the past, because the basic material evidence of their authorial activity – manuscripts and drafts, working notes, correspondence, journals – is, like all textual production, increasingly migrating to the electronic realm. The New York Times Book Review has already fretted over the fate of literary history and biography with more and more authors’ correspondence (correspondence with editors, with readers, with each other) shifting to email. A 2005 story details the plight of Deborah Treisman, the New Yorker’s fiction editor: “Unfortunately, since I haven’t discovered any convenient way to electronically archive e-mail correspondence, I don’t usually save it, and it gets erased from our server after a few months,” Treisman said. “If there’s a particularly entertaining or illuminating back-and-forth with a writer over the editing process, though, I do sometimes print and file the e-mails.” The fiction department files eventually go to the New York Public Library, she said, “so conceivably someone could, in the distant future, dig all of this up.”8

In the same piece, the novelist Zadie Smith is quoted as saying: “I don’t have a single early draft of any novel or story. I just ‘saved’ over the originals until I reached the final version.” This is not a new problem. Almost two decades ago communications scholar Daniel Chandler had already noted

Tracking the changes

261

that “writing done with a word processor obscures its own evolution.”9 The only foreseeable remedy at the time was impractical: methodically printing and archiving a hard copy of each successive draft of a work. In fact, however, as word processors matured they began incorporating sophisticated version tracking and date/time-stamping features that (at least in theory) offer unprecedented access to an author’s composition process, allowing editors and scholars to watch a passage take shape keystroke by keystroke. This is the central paradox of the current technological moment: fragile as electronic media are, so too do they create enormous and potentially unprecedented opportunities for scholarship. With the advent of “track changes” and similar features, every electronic document can become a genetic text embedding the history of its own making. The storage space required to keep such records is now trivial, with terabyte-scale drives already on the market. Email, for its part, also demonstrates the balance between problem and possibility: traditional paper-based archives rarely include copies of letters from an author, because most people don’t retain duplicates of their own correspondence. Email, however is different: the contents of an author’s “Sent” folder typically sit right alongside of their Inbox, creating an invaluable resource. The lesson here is that for every challenge raised by born digital material, there are equal and astonishing opportunities. Opportunities and also, we hasten to add, responsibilities. In his novel Lisey’s Story (2006), early computer adopter Stephen King continues his fascination with writers and the writing process. He describes the widow of a deceased author contemplating the “Incunks” (derived from incunabula, it is her improvised and unflattering term for critics and scholars and fans desperate for some new scrap of her late husband’s work, however minor) picking through the contents of his study. Three desktop computers are the center of attention: Each was newer and lighter than the last, but even the newest was a big desktop model and all of them still worked. They were password-protected, too, and she didn’t know what the passwords were. She’d never asked, and has no idea what kind of electro-litter might be sleeping on the computers’ hard drives. Grocery lists? Poems? Erotica?10

Incunks though we may be, it falls to us in this chapter to explore how the task of textual scholarship changes when its primary objects are the born digital detritus of an author’s hard drive, the “electro-litter” scattered across a computer operating system. Are we even still talking about textual scholarship here, or are we now in the realm of computer and information science? What kind of skills will textual scholars need to function in this new environment, and what kind of tools will they need to have at their disposal? What are the ethical issues involved in plumbing the depths of a

262

matthew g. kirschenbaum and doug reside

storage medium that routinely commits all manner of data, often without the user’s awareness, to magnetic memory? Should a scholar be allowed to see an author’s high score on Tetris or their choice of desktop wallpaper? What about the music available on an MP3 playlist? What about choices for fonts and layout? Such details may seem trivial, but in fact scholars often want to know what an author was listening to or what images were important to them during the writing process. And what about pornography or financial data or other sensitive material that turns up on the machine? Is such content germane to the scholarly record? Under what circumstances? The growing effectiveness of forensic information recovery raises the stakes still further: is erased or overwritten data appropriate to include within the scope of textual scholarship if it can be retrieved in a readable form? (According to one account, the late David Foster Wallace was sometimes in the habit of deleting unwanted text from his hard drive lest he later be tempted to change his mind and put it back into a manuscript.11 ) And finally, what can the textual scholar do to ensure that the born digital material of today remains legible and accessible to the scholar of the future? Raising RENT The issues and questions raised here are of course far broader than textual scholarship or even the future of cultural heritage more generally. Every area of human endeavor is coping with the rapid and widespread transition to born digital content. Government archivists must deal with data inflows on the order of many dozens of terabytes each year, encompassing everything from White House email to the most mundane bureaucratic documents and records (a terabyte is 1,000 gigabytes; the content of the Library of Congress is typically estimated at around twenty terabytes of text total). Law enforcement has had to come to terms with collecting and analyzing born digital evidence since computers are now an integral part of crime scenes and criminal investigations. In the corporate world, the issue is especially acute since many firms, as a matter of policy, delete email and other electronic files at set intervals to avoid windows for liability. How would one study the early history of the Silicon Valley Dot Com boom in such an environment?12 What about massive and irreplaceable medical and scientific data sets? Numerous commentators have spoken openly of a “digital dark ages” and the like, with vast amounts of information simply vanished from the human record because of technological obsolescence and the absence of adequate migration paths.13 The scenarios are often devilishly alarming: how will we ensure, for example, that future generations know the location of toxic waste dumping sites when the records pertaining to their existence are stored in electronic databases running on already obsolescent systems?

Tracking the changes

263

How can a five billion-dollar ship, such as a navy aircraft carrier which might be expected to remain in service for decades, be maintained in the future when all of its blueprints and documentation for the vessel are electronic? The famous Mormon genealogical archives, many millions of feet of microfilm and microfiche housed in the church’s Granite Mountain Record Vault (carved 600 feet into the side of a mountain outside Salt Lake City), are reportedly cataloged and indexed with an Oracle database, something no one is sure they can preserve adequately.14 Specialists are fond of pointing out that electronic data simply cannot be expected to survive in usable form without the benefit of dedicated efforts to ensure that it does. Unlike the diary or photograph found in an attic – brittle or faded perhaps, but still perfectly legible after all these years – electronic information’s threshold for happenstance and serendipity is virtually (so to speak) nil. Abby Smith puts it this way: When all data are recorded as 0’s and 1’s, there is, essentially, no object that exists outside of the act of retrieval. The demand for access creates the “object,” that is, the act of retrieval precipitates the temporary reassembling of 0’s and 1’s into a meaningful sequence that can be decoded by software and hardware. A digital art-exhibition catalog, digital comic books, or digital pornography all present themselves as the same, all are literally indistinguishable one from another during the storage, unlike, say, a book on a shelf.15

For all this, it would be a mistake to think electronic data is hopelessly ephemeral. In point of fact, digital data, in particular magnetic disk storage, is one of the most persistent and indelible forms of inscription we have yet devised. As Kirschenbaum has discussed elsewhere, this is a function of two interrelated qualities: first, there is the durability of the magnetic recording process itself, where physical storage media can undergo all manner of extreme trauma, including fire, flood, and material breakup while still retaining a recoverable magnetic trace; and second, electronic data’s remarkable capacity for proliferation.16 A modern operating system may retain dozens of different copies of what the user perceives as a single, homogenous “file”; internet traffic, including Web sites and email, is cached in numerous locales, often with backup and other redundancies at each site. The conclusion to be drawn here is that electronic data are not necessarily any more fragile or given to vanishing into the ether than any other form of inscription; but they are different from prior forms, chiefly because of the fundamentally symbolic nature of data’s encoding – their dependence, as Smith observes, on appropriate software logics to make sense of the binary values, and the dependency of this software, in turn, on still more software (such as operating systems, which themselves, ultimately, depend upon specific hardware configurations). The technical challenges involved in digital preservation are

264

matthew g. kirschenbaum and doug reside

therefore formidable, but the social challenges are greater still as we all seek to adjust our habits and assumptions to accommodate these new forms of storage and recording. The social challenges are where textual scholars can play a leadership role in literary studies and the preservation of born digital literary material, helping to ensure that documents and records survive in useful as well as usable form, just as textual scholars have long fulfilled advocacy roles in traditional library settings, opposing policies that tamper with bindings, dust jackets, and other forms of material evidence. In 2007, Doug Reside (one of this chapter’s authors) was invited to inspect and help preserve the digital “papers” of composer and playwright Jonathan Larson at the Library of Congress. Larson is best known as the lyricist and composer of RENT, the hit musical that closed on Broadway in September 2008 after a critically acclaimed twelve-year run.17 But Larson had died suddenly of an aortic aneurism resulting from Marfan’s syndrome the day before the show’s off-Broadway opening in 1996. Despite his tragically abbreviated career (he was only 36), his output was extensive. In addition to RENT, he worked on at least ten other full-length musicals and a significant body of incidental music and stand-alone songs. His papers, including over 150 31/2-inch diskettes, were given to the Library of Congress in 2003. In the remainder of this section we will take a closer look at this particular example as a way of providing an extended practical demonstration of born digital textual scholarship. Initially the Library had considered treating Larson’s computer diskettes much like the other media in its audio collections – that is, catalog them according to the label on the object but without detailed listings of the files stored on them. This is not an unusual approach, and, given the strained budgets and workloads of most archives, along with the time and specialized technical expertise required for more detailed cataloging, it may be all that is possible in many cases. However Reside, who is a scholar of musical theater as well as a digital practitioner, suggested an alternative course of action, which was approved by the Library’s administration and the Larson estate. It was agreed that he would create disk images (that is, exact, bit-for-bit copies) of the disks in the collection using the “data definition” (“dd”) imaging utility that is included with most distributions of the Linux operating system. After creating these “images,” he could open virtual versions of the disks on his laptop without needing to work with the actual disks (which would have posed obvious risks). The data images themselves would be stored on a USB flash drive kept in the music library, ensuring that the digital data, like the physical artifacts, would remain in house at all times. It quickly became apparent that about 80 percent of the disks could not be read by external drives readily available on the market today. They were 800K or 400K diskettes rather than the now standard 1.4 Mb versions.

Tracking the changes

265

However Reside had access to a Macintosh Powerbook G3 that had both a floppy drive capable of reading the older disks as well as an Iomega Zip drive – precisely the right hardware constellation for the task. Using this machine, along with an external USB Zip drive, he created images of the floppies on a Zip disk on the old machine, transferred the images to the external Zip drive, and thence to the USB flash drive. Such chains of transfers from medium to medium and platform to platform are common in digital preservation projects, but the fragility of the links that compose them is palpable. In this case, the author’s tragic death meant, at least, that the materials were acquired by the Library fairly quickly and could be preserved using computer hardware very similar to that used to create it. Nonetheless, the process still depended on three hard-to-replace drives and one aging laptop, all available only through serendipity. Had any of these failed, the time and financial expense of replacing them would have significantly slowed the recovery work. Ten years from now reconstructing such a chain of devices and hardware connectors might prove impossible. Even as of this writing, three of the disks have resisted all attempts to recover their data. This is not uncommon in old, magnetic media, and pinpointing the exact cause of the problem, not to mention the employment of more exotic techniques would likely be prohibitive in all but the most important cases. Fortunately, there is a wealth of material on the disks that is still readable. These files may be divided into two categories: those that are, or could relatively easily be, printed and stored in the paper collection (e.g. drafts of librettos, phone bills, letters), and materials that would lose significant meaning if converted out of the electronic medium. The latter category, in this case, consists mostly of electronic music files. Larson recorded a good deal of his music using a Roland U-20 keyboard and an early version of Mark of the Unicorn’s Performer (a music editing computer program now called Digital Performer). These files he would then take to music arranger, Steve Skinner, who would, as Skinner puts it, “add a technically sophisticated thing to his intuitive approach.”18 Unlike an MP3 file, which records a particular performance of a sound, these music files are a set of digital instructions that describe, very precisely, how a piece of music should be played. Computers with different sound cards and speakers will interpret these instructions in different ways, and the sound that they produce will vary. Simply recording these files to a more traditional medium (a compact disc or tape, for instance), therefore, would preserve only a single performance rather than the data that generated it. The disks also contain metadata that contextualizes the drafts in the paper collection. Each file is identified by a date and time of creation, making it possible to tell from metadata on the files when, exactly, Larson created a

266

matthew g. kirschenbaum and doug reside

new file and started working and thereby enabling scholars to examine the artist’s creative process with greater precision than is usually permitted by imprecisely dated paper drafts. For example, from metadata recorded by Larson’s computer, we learn that the meeting between Roger and Mimi, eventually the subject of the song “Light My Candle,” was written in more or less the form it appears in the Broadway version over the course of about a month. In a draft outline of the musical, last modified on January 23, 1992, Larson describes the scene as “The Loft” and writes that it will include a song by Roger called “My Story.” This song does not exist as an independent file, but may refer to the lyrics in a January 13 draft of the complete script as it then existed in which Roger and Mimi meet and describe their own forms of artistic expression (Roger tells Mimi “I write plays you never heard of” and Mimi responds by confessing, “I embroider sunsets onto pillow cases”). Another synopsis, last modified on February 6, 1992, adds a song called “Light A Candle” alongside “My Story,” but here the character of “Mimi” is inexplicably named “Amy.” The first full draft of the lyrics titled, as in the Broadway version, “Light My Candle” appears in a March 5 file. This version closely resembles the text as we know it today, but with a few, comparatively awkward exchanges that Larson would later revise. For example, the March 5 version includes the following dialogue: ROGER THAT WAS MY LAST MATCH MIMI WAIT. OUR EYES’LL ADJUST. LUCKY THE MOON’S OUT ROGER IS IT THE MOONLIGHT? – OR THE FLORESCENT LIGHT FROM THE GAP?

By March 16, 1992 Larson had completed a version of the lyrics almost identical to the version performed in the show: ROGER THAT WAS MY LAST MATCH MIMI OUR EYES’LL ADJUST. THANK GOD FOR THE MOON ROGER MAYBE IT’S NOT THE MOON AT ALL I HEAR SPIKE LEE’S SHOOTING DOWN THE STREET

The paper archive also reveals this change, but without the timestamp on the digital file one would be unable to identify the order of the changes as precisely, and might, for instance, overlook the fact that the song was

Tracking the changes

267

written around the time when Spike Lee was regularly in the news for his work on Malcolm X (partially filmed in Manhattan).19 This is a bit of trivia, of course. What is perhaps more interesting is what the timestamps reveal about the creative process. By collating files created and modified over the course of six years, one can watch as RENT develops from the sort of overwrought and muddled draft so familiar to those of us who have attempted to write musicals into a milestone of the genre, clearly worthy of the Pulitzer Prize it was ultimately awarded. Some of the revisions seem obvious in retrospect: replacing the somewhat awkward lines above about the supposed moonlight being light from Gap, for instance. In other cases, the reason for the change is not immediately clear. In the final digital draft of the script, dated January 15, the songs “Santa Fe” and “I’ll Cover You” immediately follow “The Tango Maureen,” whereas in the previous, January 5, draft, they were sung just before it.20 Perhaps the change was made because Angel introduces “I’ll Cover You” by saying, “Alone at last!” (in the earlier position, the couple have only shared the stage with anyone else for the length of Angel’s “Today for you” dance number), or perhaps it was felt Angel and Collins’s healthy and loving relationship better contrasted with the troubled love triangle of Mark, Maureen, and Joanne when the bad preceded the good. The choice is an interesting one to study for those trying to master the difficult problems of structure and pacing in musical theatre storytelling. Even when the sources are analog, the textual history of a work can demystify the creative process and serve as a guide for those attempting to make a similar sort of artistic journey, however, when the drafts are digital, the signposts can be marked more precisely. Yet there remains ambiguity. Larson would often use a date as a filename for a draft. Often, these filenames do not match the timestamps on the files. In cases where the timestamp date is later than the filename one might assume that Larson simply saved the file several days later without creating a new file, however, there are cases in which the filename suggests a later date than the timestamp. For instance, although the final draft of RENT on the disks was last modified on January 15, the filename is “***1/16/96.” Is this because Larson’s computer’s clock was set incorrectly, or did Larson postdate the file? Although automatically generated metadata (like timestamps) are useful, they must still be interrogated by the textual critic and considered alongside other contextual data. File organization and folder structures can also reveal useful contextual information. For instance, a January 1990 draft of RENT is on a disk titled “Current Work” that also contains files with timestamps between 1989 and 1991 for Larson’s other musicals 1/2 MT House, Superbia, Boho Days, and Manhattan Tower, as well as lyrics for commercial jingles. Although the subject-based organizational scheme the Library of Congress has used

268

matthew g. kirschenbaum and doug reside

for the paper files is certainly logical and useful for finding and retrieving artifacts, the structure of the files on the disks suggests a non-intuitive relationship between, for example, a 1989 commercial jingle for Coke and the 1990 version of RENT, that would be obscured by standard classification schemes. Of course a librarian or archivist could copy the timestamps, directory structures, and other metadata onto the paper, but to do so the curator would need to access the data on the disks at least once, which, as my experience demonstrates, is not always an easy matter. Other contextual information cannot be easily represented on paper at all. Anyone who has composed a document on a word processor and then printed it for proofreading knows that when one looks at text on a monitor the experience is different than looking at the same text in hard copy. The formatting, the amount of text visible at one time, the colors, and the visual context all change (sometimes dramatically) in the shift from pixels to paper. It might be instructive, for instance, to see the early drafts of RENT as Larson saw them when composing. How much of “La Vie Boheme” was visible at once without scrolling? How long could a line of lyric on Larson’s monitor be before it wrapped to the next and distorted the formatting? Did any of these environmental variables affect Larson’s creative process? These questions could only be answered by studying the electronic files in an environment that replicates the one used by Larson at the time of creation. In some cases, original disks can reveal the same sort of information to the scholar as original paper drafts: a sense of the exact affordances and limitations of the medium on which the artist worked and, occasionally, access to deleted or rewritten drafts invisible in copies but which can be recovered from a lower palimpsestic layer with great skill and the right tools. Even an electronic copy as faithful as a bit-level disk image cannot fully replicate all of this information, but it comes closer, at least than a paper printout. For certain questions, even electronic copies of Larson’s disks will be insufficient. Archivists and scholars have long recognized the value of studying the very paper on which earlier authors composed their work – no one would suggest, for instance, that a photocopy of Jane Austen’s manuscript drafts is equivalent to the original pages on which she wrote – but the dissonance between copies of an electronic file is less often considered. Of course, simply preserving the disks on which Larson worked is also insufficient. If all the information on every disk in the Larson collection were preserved forever without a means of accessing and interpreting them, Larson’s work would be equally lost. Like most computer users, Larson used commercial software (e.g. MOTU Performer and Microsoft Word 5.1) to create his work, and he stored it using the default, proprietary formats associated with these programs. As a result, the future scholar would need access, not only to Larson’s files, but also copies of the software used to

Tracking the changes

269

generate them (or at least some way of interpreting files saved in its format). These programs are usually designed to run on a specific operating system (e.g. Macintosh System 7), which must also be preserved. Unfortunately, commercial software is often distributed with non-transferable licenses. That is, at least according to the license agreements, it is not permissible for the original purchaser of the software to give that program to anyone else. While a case might be made that fair use exceptions apply in such instances, many public and academic libraries, often advised by risk-adverse legal counsel, may not wish to assume the liability of a potential lawsuit. Even if a library does decide to preserve all the relevant software, however, the data remains in danger. The programmers of operating systems make important assumptions about the underlying hardware on which it is running. An Intel-based MacBook laptop built in 2010 cannot boot Macintosh System 7 in the same way that Larson’s Macintosh Quadra 605 did. One can, of course, use emulators (computer programs designed to run software written for another type of computer), but these emulators are themselves digital artifacts that must be updated and maintained if they are to remain useful, and they do not provide all of the contextual information one might find from opening a file on the original hardware. Again, the digital archivist has returned to the familiar, if somewhat changed, world of physical preservation. In the Larson collection at least, there are also paper drafts that contain information that is not found in the digital materials. Like many writers, Larson seems sometimes to have found the paper and pencil a more congenial environment than the word processor for early drafts (in the box 11 of the RENT collection, for instance, there is an early handwritten draft of “Light My Candle” which seems to predate the March 5 version on the disks). Perhaps because the disks in the collection appear to have been mostly used for backups or for transferring data from one computer to another, the files they contain tend to represent significant drafts that Larson felt were worth preserving. There are, therefore, printed copies of the script for which there is no apparent electronic analogue, perhaps never copied from his hard drive (which is not held by the library), or replaced when Larson saved his revisions without creating a new file. There is, for instance, a printed draft of “Light My Candle” in the collection that probably follows the handwritten copy but predates the March 5 version on the disks, in which Rogers states, matter-of-factly: IT’S NOT THE MOONLIGHTTHERE’S A GAP STORE FOUR BLOCKS AWAY.

After finishing a draft, Larson would often print it and make initial revisions on paper before returning to the electronic version. Before arriving at the

270

matthew g. kirschenbaum and doug reside

specific reference to Spike Lee in the final version, Larson crossed out the mention of the “florescent light from the Gap,” in a printed copy of the lyrics and replaced it, in pencil, with: MAYBE IT’S NOT THE MOON IT COULD BE A FILMSHOOT

This hybrid process of digital and analogue writing is, of course, very common even today for all sorts of writers, and libraries and archives must prepare for the challenges it creates. What is less common today, when laser printers are ubiquitous in offices and high quality inkjet printers are the standard in homes, is the way in which Larson was forced to consider the printer he had available to print each draft. Before 1993, Larson seems to have had a relatively inexpensive ImageWriter dot matrix printer that he used for day-to-day printing of drafts, but which produced copies that, even in the 1990s, appeared amateurish and, perhaps more importantly, could be difficult to read when photocopied. Therefore, when quality mattered (for example, when submitting a script to a producer or a granting agency), Larson would find a laser printer. However, because the resolution and shape of the characters produced by the ImageWriter differed from the shape and size of those produced by a laser printer, it was necessary to reformat the script for the laser in order to account for the differences in spacing and pagination. Thus, there are folders on Larson’s disks with titles like “RENT Laser” that contain versions of the script formatted for higher quality printing. These reformatted scripts exemplify the awkward hybrid of centuries-old print traditions with the advantages and quirks of new media, and the challenges this transitional moment presents to textual scholars and archivists. Meeting the challenges Of course, as any student of RENT is aware, loss is a part of life, and it is certainly a part of cultural preservation. For all its complications, though, the Larson archive is, in many ways, one of the easiest test cases for the digital archivist and scholar concerned with the “born digital.” In particular, the advent of so-called cloud computing (the increasing popularity of network services for email, blogging, photo-sharing, and social networking), as well as the growing user base for third-party back-up services like DropBox and Carbonite, suggests that content from the first twenty-five years or so of personal computing may in fact represent an anomalous window of opportunity, one wherein the archivist and scholar actually has reasonable prospects for direct hands-on access to the original hardware and storage media.

Tracking the changes

271

In the meantime, however, opportunities for born digital textual scholarship are expanding rapidly. Notable authors represented by at least some born digital content already housed within an archival collection include Russell Banks, Samuel Beckett, Lee Blessing, John Crowley, Robert De Niro, Michael Joyce, Thomas Kinsella, Bernard Kops, Norman Mailer, Terrence McNally, Tim O’Brien, Salman Rushdie, Ronald Sukenick, Leon Uris, Alice Walker, and Arnold Wesker. All of these (and more) are at either the Harry Ransom Center at the University of Texas at Austin or Emory University, two institutions that were part of an exploratory study directed by Kirschenbaum and funded by the National Endowment for the Humanities.21 Other libraries and archives are also beginning to acquire born digital content, almost always – as was the case with the Larson materials – as sub-sets of larger collections of “papers.” In most cases, the storage media consists of diskettes; in a few, including such notable instances as Mailer and Rushdie, entire computers are present. (Mailer himself did not actually use a computer, but his personal assistant did.) There is important activity in England as well as the United States, with the Bodleian, for example, pioneering best practices in digital papers from political figures, a domain where there is obvious cross-over application to literary scholarship.22 The British Library, meanwhile, has concluded a pathbreaking project on Digital Lives, “focusing on personal digital archives and their relationship with research repositories.”23 But many challenges exist for which there is simply not yet enough accumulated wisdom and experience to formulate best practices. One of the clear conclusions of Kirschenbaum’s NEH study is that meeting the challenges (and realizing the opportunities) of the born digital will require close cooperation and communication among a triad of stake-holders: authors and other content creators; archivists and information specialists; and finally scholars, who also represent the interests of the wider public. In the current climate, where the expertise for handling born digital materials is still idiosyncratic and widely distributed, appropriate roles for all of these are still very much in the process of being sorted out. Reside’s involvement in the process of transferring data from original media to a contemporary format is probably exceptional (in general such work is and will probably be performed by archivists rather than individual researchers), but it seems reasonable that the work of interpreting bits into human readable information may always be the responsibility of the scholar rather than archival staff. It is easy, on the one hand, to pity the hapless scholar without any specialized training who arrives at an archive only to find that he or she is expected to grapple with an arcane array of antiquarian file formats in order to perform even the most basic research activity, and yet scholars working with original primary source materials are regularly expected to

272

matthew g. kirschenbaum and doug reside

bring some knowledge about the material processes and the context that generated them. We do not, in general, expect an archivist to translate an ancient Chinese manuscript for a visiting scholar or explain to a library patron that one has to read a Hebrew scroll from right to left. For the casual reader interested only in easily accessible content there will be annotated editions, but an expert understanding of media has always been an expected part of a textual scholar’s training. Textual scholars working with content composed after 1980 should therefore be expected to acquire the expertise relevant to the materials in the corpus they are studying. Most will need to know how to search for human readable information in a binary file (at the moment most easily done using a hex editor). In working with the Larson files, for instance, Reside first realized that a file (without an identifiable extension) was in MOTU Performer format when he opened the file in such an editor and saw the name of Roy Groth (an early developer of the Performer software). Familiarity with a low level operating system (a Linux shell, for instance) is also invaluable as such systems remove many of the abstractions between the user and the data that graphic operating systems (Windows or OS X, for instance) establish. Such abstractions arguably make many common tasks more efficient, but the efficiency often comes at the cost of the control over data so essential for digital textual scholarship. Most of all, familiarity with the history of operating systems, common file formats, and media hardware is as important to the contemporary textual critic as was an understanding of paper production, watermarks, and bookbinding techniques in earlier eras. To expect that archivists alone will have this knowledge and be able to communicate it to every researcher is not only unrealistic, it is also an abdication of scholarly responsibility that borders on professional negligence – it is difficult to imagine a scholar working with born digital sources making a convincing textual argument without the skills described above. Still, there is a crucial education and advocacy role for archivists to play, both among those scholars who are still learning (or relearning) their trade and do not yet appreciate the significance of born digital material or the opportunities it affords – and among authors, who may be confused or intimidated in the face of a decision like whether or not to hand over an entire laptop to a repository. Dealers and auction houses, meanwhile, used to trafficking in an author’s literary papers and estates, are only just beginning to think about how to appraise the monetary value of born digital material.24 The “challenge” of the born digital is thus at least as much social as it is technological. New textual forms require new work habits, new training, new tools, new practices, and new instincts. Details regarding file formats and media types are no more or less exotic from the purview of textual

Tracking the changes

273

studies than the minute particulars of letterpress, paper making, and bookbinding. We look forward to textual scholarship’s embrace of this moment when the material foundations of the field are so completely in flux, even as practitioners equip themselves with the knowledge and training necessary to carry their discipline forward into the new century. NOTES 1 William Gibson, Pattern Recognition (New York: Putnam, 2004), 216–17. 2 Stephen King, Skeleton Crew (New York: Signet, 1983), 20–1. 3 James Fallows, “Living with a Computer,” Atlantic (1982): www.theatlantic. com/magazine/archive/1982/07/living-with-a-computer/6063/ 4 Umberto Eco, Foucault’s Pendulum (London: Picador, 1988), 24, 26. 5 Terrence McNally Papers, Harry Ransom Humanities Research Center, Disk No.: 22a, File name: NEWLIGHT. 6 www.salon.com/oct96/didion2961028.html 7 http://toc.oreilly.com/2008/10/how-should-authors-promote-the.html 8 www.nytimes.com/2005/09/04/books/review/04DONADIO.html 9 Daniel Chandler, “The Phenomenology of Writing by Hand,” www.aber.ac.uk/ media/Documents/short/phenom.html 10 Stephen King, Lisey’s Story (New York: Simon and Schuster, 2006), 5–6. 11 D. T. Max, “The Unfinished,” The New Yorker, March 9, 2009. Also available online: www.newyorker.com/reporting/2009/03/09/090309fa fact max 12 See David Kirsch’s DotCom Archive: http://dotcomarchive.org/ 13 For example, www.physorg.com/news144343006.html 14 See http://blog.longnow.org/2007/04/09/the-granite-vaults-of-geneology/ 15 Abby Smith, “Preservation in the Future Tense,” CLIR Issues, 3 (May/June 1998), 1, 6. 16 Mechanisms: New Media and the Forensic Imagination (MIT Press, 2008). 17 RENT is a registered trademark of Skeeziks, LLC. This chapter contains excerpts from various versions of “RENT” by Jonathan Larson. The copyrights in which are owned by Skeeziks, LLC and the Estate of Jonathan Larson and are protected by United States copyright law. These excerpts may not be reproduced, distributed, transmitted, displayed, published or broadcast without prior written consent of the appropriate entity. 18 Allan S. Larson, Nanette Larson, and Julie Larson McCollum, Rent (New York: Rob Weisbach Books, 1997), 15. 19 www.imdb.com/title/tt0104797/locations 20 In the Broadway libretto, “Santa Fe” and “I’ll Cover You” appear even later in the act. 21 Available here: www.neh.gov/ODH/ResourceLibrary/WhitePaperRepository/ tabid/111/Default.aspx 22 Paradigm, Personal Archives Accessible in Digital Format: http://www.paradigm. ac.uk/ 23 See www.bl.uk/digital-lives/about.html 24 See www.newscientist.com/article/dn16589-how-do-you-value-a-historicalemail.html?DCMP=OTC-rss&nsref=online-news

JEROME MCGANN

Coda Why digital textual scholarship matters; or, philology in a new key Why does textual scholarship itself matter? Most students of literature and culture who worked in the twentieth century would have thought that a highly specialized question, and many still do. One hundred years ago the question would hardly have been posed at all. Until the early decades of the twentieth century what we call literary and cultural studies was called philology, and all its critical and interpretive procedures were clearly understood to be grounded in textual scholarship. What is philology? Literally, it means the love of the word, or, more specifically, the love of articulated thought. In the context of Greek oral culture, it also means speech. Since philology as a discipline emerged through print culture, philology’s “word” is primarily textual. The term (and the discipline) grew to eminence in the nineteenth century and defined what we now call “Literary and Cultural Studies.” The great German philologist August Boeckh famously defined the word as “Die Erkenntnis des Erkannten” – The Knowledge of What is Known.1 It is a brilliant formulation. We study and pass on the human record that others before us have studied and passed on to us. Distinguishing two critical procedures – the Higher and the Lower criticism – philology studies the documentary record on the assumption – in the words of the late, great textual scholar D. F. McKenzie – that the documents carry the evidence of “the history of their own making.”2 To the philologian, all possible meanings are a function of their historical emergence as material artifacts. The investigation of those artifacts is the foundation of literary and cultural studies. The Lower Criticism devotes itself to the study of the documents per se; the Higher Criticism considers the documents as they are the product of human endeavor under specific historical circumstances. That comprehensive historical approach to literary and cultural studies was gradually displaced in the twentieth century, and the very term “philology” fell into disuse. In the horizon of Modernism scholars turned to hermeneutics of many kinds and thence – after World War II – to the 274

Coda: Why digital textual scholarship matters

275

meta-interpretive interests that played themselves out, in equally diverse ways, under the general banner of Theory. In the perspective of modern and postmodern hermeneutics and theory, then, “Why does textual scholarship matter?” is perceived as a technical or subdisciplinary (or even pre-disciplinary) question, as Ren´e Wellek’s influential mid-century handbook Theory of Literature (1949) makes very clear.3 But the emergence of digital media in the late twentieth century is forcing a shift of perspective back to the view of traditional philology, where textual scholarship was understood as the foundation of every aspect of literary and cultural studies. Here I shall give a brief explanation about why this shift in perspective is happening and how we should respond. A simple observation throws the issue into sharp relief: In the next 50 years the entirety of our inherited archive of cultural works will have to be re-edited within a network of digital storage, access, and dissemination. This system, which is already under development, is transnational and transcultural.4

Published in 2001, that comment underscores the massive changes that digital technology is bringing to our cultural archives – to the libraries and museums that are the operational source and end of everything scholars do. And because scholarship underpins all of our educational work, intra- as well as extra-mural, these changes in the fundamental documentary record have a correspondingly significant effect on our work as educators. With relatively few exceptions, this editorial work – the migration of the paper-based archives to digital forms – is being undertaken either by librarians or by the agents of commercial entities like Google, ChadwyckHealey, and Gale. The commercial examples are the most troubling from a scholar’s and educator’s point of view. For example, while The English Poetry Database (Chadwyck-Healey) provides a vast depository of books of verse, all the prose materials in those books were removed as an initial – a disastrous – editorial decision. For its part, Gale’s ECCO database (Eighteenth-Century Collections Online) was generated by what is called “dirty OCR” – i.e., Optical Character Recognition software was run on a set of (old) microfilms of the original documents. Google Books is the same. The resulting digital transcriptions are often lamentable. When materials are migrated through entities like university presses – here JSTOR is the outstanding example – the results are much more satisfying because the presses have always operated in close collaboration with scholars. The cost of these commercially produced materials is another issue of concern for the scholar and educator. Unlike JSTOR and Project Muse, Chadwyck-Healey, Gale, Kluwer, and other online publishing agents work

276

jerome mcgann

to generate as much profit from their capital as they can. But so far as a scholar or educator is concerned, knowledge and information should not be viewed as fungible assets. The free-culture policy of the Library of Congress reflects an enlightened social consciousness that scholars and educators share. And while Google Books has begun its work in a relatively benign way, it is a monopolistic enterprise and as such fraught with danger from a scholar’s and educator’s point of view. Research libraries – University of California, University of Michigan, and University of Virginia are notable examples – are also committing themselves to the migration of the paper archive to digital form. In these cases the transcription of the primary documents is generally executed with care. Accuracy is a great concern here because reliable preservation is the primary goal of the work. But from a scholar’s and educator’s point of view, textual accuracy (along with basic documentary metadata about the process of transcription) is only a minimal sine qua non for textual materials. Librarians do not readily move beyond that specialized focus, nor should scholars or educators expect them to. Filling out and exposing the richness of our cultural inheritance is precisely the office of the scholar. These examples remind us about two very broad obligations that scholars have by virtue of their vocation as educators. We are called to surveille and monitor this process of digitization. Much of it is now being carried out by agents who act, by will or by mistake, quite against the interests of scholars and educators – and in that respect, against the general good of society. So we must insist on participating. At the same time – as Umberto Eco, Robert Darnton, and Roger Chartier, among others, keep reminding us – we have to protect our paper-based inheritance.5 Digitizing the archive is not about replacing it. It’s about making it usable in new ways, and making sure we don’t forget the old. A major task lying before us – its technical difficulties are great – is to design a knowledge and information network that integrates, as seamlessly as possible, our paperbased inheritance with the emerging archive of born digital materials. One doesn’t have to be a textual scholar or New Philologist to register alarm at the general situation I’ve just described. Nonetheless, it should be obvious that we won’t bring reliable judgment to the complex problems involved here without a good understanding of manuscript, book, and digital technologies – and traditional audio and visual technologies as well. Because in this essay I want to discuss textual criticism specifically, I shall set aside the issues – deeply significant though they are – that relate to audio and visual materials. So let’s begin by thinking again about these two technologies, bibliographical and digital. It is a commonplace now, even among scholars, to distinguish between “static” paper-based documents and “dynamic” digital

Coda: Why digital textual scholarship matters

277

environments. If we negotiate these media at the level of their immediate self-representations, the distinction is acceptable. If, on the other hand, we want a more critical understanding of these two great modes of expression and communication, we have to expose the social agencies that power and structure them. When we do that we discover how the distinction is seriously misleading. Today such a critical analysis is important, perhaps even imperative, for the reason already given: the entirety of our paper-based inheritance is being digitally remediated. A “Digital Condition” is overtaking and in certain respects completely replacing our “Textual Condition.” How will this unfold? No one knows because no one yet understands the social and institutional implications of these technological changes. But while our Digital Condition is now only beginning to emerge, we have inhabited a Textual Condition for millennia, the Age of Print being the culminant phase of its astonishing emergence. That history has much to teach us now. Our Textual Condition has forged an extensive record of its own making – which is to say, a process of critical reflection in manuscript and print that is itself part of the character, shape, and meaning of the history. In the West we conveniently organize the history of textuality from Socrates’ famous attack on writing in The Phaedrus. I reference that quasilegendary text because, as we shall see, it continues to have relevance for us who see digital technology invading our long-established world of texttechnology. Socrates objects to writing because it introduces an inhuman element into the field of human communication. Anticipating Kate Hayles’s question about “How We Became Posthuman,”6 Socrates answers: we become posthuman by writing, i.e., by estranging our thinking from our own persons, by coding it in scripts. For Socrates – in this respect he anticipates the views of many contemporary digital enthusiasts – written forms are dead forms: static, incapable of responding the way human beings do when they converse. Like “the painter’s works,” written texts are not only silent, mute to our questions; they are self-identical, “they go on telling you the same thing over and over forever” (Phaedrus 67). Today – certainly since Montaigne – we see The Textual Condition very differently. Montaigne’s famous revisions to his essays demonstrate something that Socrates didn’t imagine: that no text goes on telling the same thing over and over forever, that no text, written or otherwise, is self-identical. Covered with inscriptions from the works of antiquity, the walls and ceiling of Montaigne’s library were an emblem – they are still there to be seen – of his passage to India, an inexhaustible resource for his thinking and writing. And so for our own. Socrates’ model for human communication is radically individuated, the one-on-one exchange between a master and a pupil. The object of a Socratic

278

jerome mcgann

exchange is the birth of self-knowledge in the ephebe, with Socrates playing midwife to that event. Montaigne’s model is different. His object remains self-knowledge but the path to its achievement is set outside the self – back to a past he can study as a textual inheritance, and out to a future in which his own writing is borne along. In Montaigne’s pursuit of endless self-rediscovery, the ancient texts say different things to him over and over forever. He does not talk with Socrates, he reads Plato and then writes down his thoughts as the first move in a process of remediation that has not ceased to this day. This is the general context which Kate Hayles recently engaged when she began urging scholars and critics to stop seeing “literature as immaterial verbal constructions” and become “alert to the importance of materiality in cultural productions.”7 That idea and all its corollaries have been the foundation of philology and literary studies for hundreds of years. But as the postmodern Hayles has come to realize – she takes her own life-experience as her central example – the material focus of the philologist fell from attention in the twentieth century. It fell before the coming of the interpretive interests of the New Criticism and, later, nearly all forms of Literary Theory. Shaped as her work has been in that theater of cultural interests, Hayles’s voice – the story she tells us – is important. It signals the return of the philological repressed to the abstract regions of Theory. Let me pause with Hayles, and in particular with her celebrated book Writing Machines. The book is a highly personal account of Hayles’s dawning realization that “The physical attributes constituting any artifact are potentially infinite.”8 Hayles demonstrates that idea with close readings of three literary works – Lexia to Perplexia, A Humument, and House of Leaves – the first a born digital artifact, the other two paper-based. The examples (as well as their sequence of presentation) come to show how a literary scholar steeped in contemporary theory was led by her interest in electronic media to see paper-based textualities in a new way. Electronic text had its own specificities, and a deep understanding of them would bring into view by contrast the specificities of print, which could again be seen for what it was, a medium and not a transparent interface.9 It would take years and many more experiences with electronic texts to understand that electronic literature operated in fundamentally different ways than print and required new critical frameworks to assess its reading and writing practices.10

Those formulations are interesting. The electronic medium is the lever of enlightenment throughout Hayles’s book, the means by which she began to rethink traditional paper-based textualities: “Digital media have given us an

Coda: Why digital textual scholarship matters

279

opportunity . . . to see print [media] with new eyes.”11 The insight into print media has shifted Hayles’s general theoretical views. Now she is able to see that Literature was never only words, never merely immaterial verbal constructions. Literary texts, like us, have bodies, an actuality necessitating that their materialities and meanings are deeply interwoven with each other.12

Armed with that vision, Hayles moves “to think deeply about the interplay between verbal and nonverbal components in electronic literature and art” “as we work toward critical practices and theories appropriate for electronic literature.”13 Looking beyond the “words,” Hayles draws our attention to their perceptual condition – specifically, to certain details of their bibliographical interfaces. As we know, early printers used manuscript models in designing their type fonts and page layouts, and print technology has had a similar influence on first generation born digital works, both in their interface design and even in their underlying logical structures. The influence is apparent in foundational hypermedia projects like The Perseus Project, The Rossetti Archive, and The Canterbury Tales Project – all clearly designed by persons who have worked in research depositories and documentary archives.14 Like the online OED, such works have informational and critical capabilities that overgo in various ways the paper-based materials they remediate. For one thing, IT allows one to supply – in a given plane space like the computer screen (with its obvious equivalence to the printable page) – many more explicit links or connections than are possible in a book space. This is what makes “hyper”textual space appear so dynamic, particularly when set beside the wall of words a reader sees on the page, for instance, of an Austen novel or a typical book of nonfiction. In a digital textspace, readers get described as users, “actively” (so to speak) choosing to follow links to other materials (often located elsewhere), which may themselves be multiply-linked further or looped back in complex ways. In face of this textual body-electric, a book is commonly viewed as a static object and its reader as an agent passively receiving those fixed, self-identical printed characters, which – as Socrates argued – appear to go on saying the same thing over and over for ever. Socrates’ thought about print technology is more true for his fifth-century bc oral culture than it is for Montaigne’s sixteenth-century world, or for ours today. And the difference is important. We can see this if we reflect on two dangers, recognized very early, that threaten digital textspace: the interface may become a confusion of options for the user; and the user may get lost in navigating the looped and linked network of materials. These problems arise because so much information can be immediately and explicitly

280

jerome mcgann

accessed – the adverbs are crucial – through any given unit of a digital environment. The temptation is very strong to instantiate, immediately and explicitly, that network of possibilities. Exploring how to design human/digital interfaces that will meet the many different demands we require is a large, complex, and active disciplinary field. In this work, however, interface is largely investigated not as an aesthetic environment but as a means for delivering information. It is thus not surprising that we lack an electronic literature to vie with aesthetic work produced in print media, or to match the considerable success achieved by advertising and game designers. Only a small band of angels is working to design effective interfaces for humanities materials, whether remediated print materials or born digital literary and scholarly works; and those angels, like most angels I suppose, typically don’t get paid for their efforts. Five hundred years of experience in print and graphic design have produced a variety of models for optimizing bookspace to different works and purposes. That history explains why book technology for aesthetic networks is at this moment so much more flexible and sophisticated than digital technology. This fact about the two media can be seen even if we look at a book only in relation to its most immediate phase space – i.e., as a physical object we can hold in our hands. “Literature was never only words,” Hayles rightly observes, for she wants us to see the relevance of the literary work’s printed and graphical form to its aesthetic character. But as we shall see, that form is only one of the dimensions of a literary work and by no means the richest dimension. So Hayles’s formulation and view, which is widespread, has to be drastically revised and extended. It is true and important to realize that “Literary texts, like us, have bodies,” and that those bodies are different in kind, as e-texts and print-texts are different. It is also true, however, and not widely understood, that these different bodies are all multidimensional. In any human frame of reference, embodiment entails more forms and conditions of being than can be perceived or measured in an immediate and explicit physical environment. Every text, as Roland Barthes insisted, has an indefinite potential for meaning because it lives in a “readerly” environment, which is one part of its con-text. But the same is true for every textual document. Documents are n-dimensional because they are socially determined – constituted (and repeatedly reconstituted) via what the Marxists called – bless them and thank Somebody – their “materials, means, and modes of production.” Whether we work with bibliographical or with digital technology, scholars want an optimal access to and understanding of our documentary materials. And precisely in this respect we can see, at our historical moment, the relevance of philology and its textual critical foundation. We have to go

Coda: Why digital textual scholarship matters

281

back to the future. We need to recover what Trotsky called “the privilege of historical backwardness.” In the 1980s editorial theory and method made a significant move into what became known as a “social theory of text.” For Anglo-American scholarship, two books defined this turn: A Critique of Modern Textual Criticism (1983) and Bibliography and the Sociology of Texts (1986).15 The move entailed rethinking a crucial distinction we traditionally make: the distinction between a text and its context. For the past twenty-five years many scholars have been exploring the fault lines of that distinction. Recent work in antebellum American literature – for instance, studies by J. Gerald Kennedy, Jonathan Elmer, Terence Whalen, Eliza Richards, and Meredith McGill – have been especially interesting for me, as I shall explain in a moment. At the outset of her study of “the Poe Circle” Eliza Richards gives an admirable summary of a social text approach to literary study: “the poetics of creation are inseparable from the poetics of reception.”16 The complete genetic information about any cultural work is coded in the double helix of its DNA, that is, in the codependent relation of its production history and its reception history. While much more could and should be said about the structure of that codependent relation, the essential point to realize is that each strand of this double helix is produced by the collaboration of multiple agents. The terms “the poet” and “the reader” are high-level generalized descriptors of a dialectical process of various persons and institutions. In that frame of reference we can lay out a complete matrix for a sociohistorical interpretive method, in which the social text sits at the center of a multidimensional field of interpretive action.17 In different ways, all of the antebellum scholars who were my point of departure are either explicitly or implicitly calling for a “model of literary production that . . . is intersubjective and interactive.”18 What would such a model look like? The question can be sharply defined if we pose it with respect to the essential philological form that any “model of literary production” must be able to take. That basic form is the scholarly edition. The quest for Richards’s model has been pursued with greatest rigor in the tight little island of textual theory and editorial method. D. F. McKenzie became The Hero of Our Own Time not because he discovered the sociology of the text – we’ve known about that for a long time. He became The Hero because he knew that the idea of the social text had to be realized as a scholarly edition. Such an edition would be addressing and answering some key – basically philological – questions. Could one develop a model for editing books and material objects rather than just the linguistic phenomena we call texts? To pose that question, as McKenzie did, was to lay open the true dimensions

282

jerome mcgann

of what he was after: a model for editing texts in their contexts. So the initial question is a palimpsest concealing other salient questions. Could one develop a model for exposing and comparing relationships between phenomena that are radically discontinuous: different authors and their authorized texts, say, as well as the relations between various agents – individual as well as institutional – in an eventual field disposing more than just textual or bibliographical things? Could the model expose and examine relationships between phenomena – various works and their various agents – located in fields that are discontinuous in social time and space? Finally, could such machines be designed and actually built, the way the critical editions we inherit were designed and built? The (alphanumeric) critical edition – developed and modified over many centuries, and well before the coming of print technology – is one of the most remarkable machines created by the ingenuity of Man. Its importance cannot be overstated, for it is a complete model of the enginery that has come to sustain nearly the whole of the cultural memory of the West. Its design flexibility is amazing both with respect to the numbers and kinds of materials it can represent, and to the network of social relations it both models and engages. But that is to put the matter too abstractly. We can see what these machines involve if we reflect briefly on a representative example – for instance, J. C. Mays’s recent edition of Coleridge’s poetical works.19 This great edition deserves, and has received, the closest kind of study. Let me here focus on certain key formal properties – features it shares with scholarly editions in general. The object of such works is to supply readers with a more or less comprehensive view of the current state of what we know about the works to be edited. Mays’s critical and synoptic edition of Coleridge aspires to give a comprehensive view of the production and reception histories of all Coleridge’s poetical works. This means that the edition is obligated to examine and represent every salient documentary witness of every poetical work and, in addition, to represent the social and historical relations in which these witnesses stand to each other. The key critical means for displaying that information is the text/apparatus device. Linked to that elementary machinery are various commentaries, annotations, indexes, bibliographies, maps, chronologies, and appendices that help to elucidate the materials in brief or extensive ways. Most important – and rarely noticed as such – is the use of various symbols and abbreviations in these works. So in Mays’s edition we have notes like the following (to the title of poem no. 156, “This Lime-Tree Bower My Prison”: “C’s predilection for TP’s bower – or ‘Jasmine Harbour’ – is corroborated by other sources, e.g., Poole i 202, H Works xvii 119, Cottle Rem 150–1. TP’s walnut-tree and elms (lines 52, 55) likewise had a real existence.” A

Coda: Why digital textual scholarship matters

283

list of abbreviations in the edition fills out these truncated references, and the procedure allows the editor to save a good deal of paper-space. But the references, abbreviated or not, are far more significant than merely being space-saving devices. They are in fact vehicles for traversing space and time – the edition’s hyperlinks to the received socio-historical network of materials with which the author’s work is meshed and implicated. They are the elementary signs that Coleridge’s works are social texts and do not make explicit all of their meanings or all of their conditions of meaning. Joseph Cottle, Thomas Poole, and William Hazlitt – the lives of these men are closely bound up with the life and the works of Coleridge. More to the specifically textual point of the matter, we know about their relationships through specific manuscript or printed sources, some direct sources, most indirect. The abbreviations in the quoted passage reference those sources and, in so doing, expose yet another remarkable truth about scholarly editions: they assume the existence of reference depositories of a highly sophisticated kind. “Literature was never only words”: although Hayles says this to draw attention to a book’s typographical and design features, a work like Mays’s edition plunges far more deeply into the precise significance of bibliographical materiality. The edition is a model, a theoretical instantiation, of the vast and distributed textual network in which we have come to embody our knowledge. So The Critical Edition – the idea of it that Mays learned about through his reading – supplies him with a machinery for delivering a summary representation of the entire production and reception histories of Coleridge’s poetical work. It then links that representation – along with all its imbedded parts and sub-links – to the social and institutional networks that made those histories possible. The edition thus involves a comprehensive metainterpretation of Coleridge’s poetical work – at once a reflection of what it has come to mean in its historical emergence, and a forecast of how it might be taken to mean in the future. In observing this of Mays’s edition I am not saying that the work is (for example) empirically comprehensive in its summary of the poetical materials or the meanings that have been or could be built from those materials. It is not nor could it be. Mays scants some materials and overlooks others, and of course there are materials – production and reception materials both – that undoubtedly exist but have not yet come to light. Nonetheless, the edition is both theoretically and methodologically complete. What is not there could be there, nothing about Coleridge’s poetry has been alienated as such. We are now called to design and build digital equivalents of such a machinery. These scholarly devices will have to comprehend works like The Rossetti Archive, which is a born digital representation of non-digital work, as well as works like Lexia to Perplexia, which is without explicit connections to

284

jerome mcgann

non-digital materials. Such machineries have at this point only the most primitive existence. A library that houses manuscripts, books, and digital works is precisely such a machine, storing these materials and providing access to them. But the machinery of our emerging digital environments is at this point primitive to a degree since we have yet to demonstrate how these different kinds of materials can be integrated for study. Nor does anyone yet have a good idea about how online scholarly works will be sustained beyond a twenty year horizon. And while that may be an entrepreneur’s horizon, it is not a scholar’s. We don’t have the knowledge of what we need to know. Philology and textual criticism can, I believe, help us acquire that knowledge. The first thing to learn is that the general conceptual character of the machinery we require can be theorized. We are looking to realize the system of philology as a digital emergence. Second, because the history of the emergence of traditional scholarly editions represents the material history of pre-digital philology, we can be confident about realizing the theory of a new philology by building models of what we now think the theory must involve. The Perseus Project, The Rossetti Archive, and The Canterbury Tales Project are early attempts at building such models. In the case of The Rossetti Archive – about which I can speak from an inner standing-point – building the model exposed the theoretical limitations of the model. That exposure led next to the undertaking of NINES (The Networked Infrastructure of Nineteenth-century Electronic Scholarship), which attempts a more comprehensive realization of the theory of the new philology.20 As a philological endeavor, this machinery has to meet the following functions/requirements. The first two reflect what traditional philology calls the Lower Criticism, the second two, the Higher Criticism. 1. The depository of artifacts must be comprehensive. D. G. Rossetti’s original works, The Rossetti Archive, and Lexia to Perplexia all share the common space of an emergent cultural history. Although I’ve set aside all discussion of audio and visual materials (including film and television), I mention them here in passing to underscore the requirement that the archive be comprehensive. 2. Its different parts must be organized in a network of internal links and external connections that can be represented as conventions. Fundamentally this is a process of identifying and classifying artifacts and their component parts. Any given artifact will have many identifiable parts and can be variously classified. Such and such an object is (for example) at once a poem, a printed page, a sonnet (of a certain kind), a proof sheet (corrected or uncorrected, authorial or non-authorial), a section of a larger

Coda: Why digital textual scholarship matters

285

poetical object, a translation, and so on and so on. Its parts are similarly multiple. Daniel Pitti, the Associate Director of the Institute for Advanced Technology in the Humanities at the University of Virginia, approaches the teaching of XML text encoding by handing out a recipe printed on a single page and asking each member of the class to take five minutes making a list of the object’s formal features. The ensuing class discussion quickly reveals the range of possibilities. For materially different types of artifacts – printed texts, maps, photographs – divisioning and classifying become yet more complicated when the ultimate purpose is to arrange them in a system that permits coherent analysis and study. The problem is greatly amplified when the manipulable physical properties scale to radically different measures, as is the case with a depository that includes paper-based objects and born digital objects, 3. The total system must rest in a single perspective that reflects the conception of the system generally agreed upon by its users. In terms of traditional philology, this rule explains the codependent relation that holds between the Lower and the Higher Criticism. That relationship reflects the social character of the system generally considered. In a nonhistorical system – Aristotle’s for example – the parts and classifications are conceived a priori. That is the theory of the system. In philology, however, Aristotle’s categories and topoi are understood as socially inflected and historically emergent. Designing a system in the horizon of philology, then, requires building critical devices that require the system to be modifiable through use. If the system is not “open” in that way it is not, in the philological perspective, theoretically complete. 4. From that general conceptual vantage, the system must have the flexibility to license, and ultimately store, an indefinite number of particular views of its artifacts and their relations, including different views of the system as a whole. That set of functions reflects the fact that a philological system is fundamentally a system of social software. It may be modeled and then instantiated in the paper-based form that we have inherited and still use, or in the online network that continues to emerge today, and acquire more precise definition. Realizing this important homology between bibliographical and digital networks is crucial as we try to design the latter to the needs of scholars and educators. Thus, the initial design of the Collex software that powers the NINES initiative was conceived to allow users to

286

jerome mcgann

search, browse, annotate, and tag electronic objects and to repurpose them in illustrated, interlinked essays or exhibits. By saving information about user activity (including the construction of annotated collections and exhibits) as “remixable” metadata, the Collex system writes current practice into the scholarly record and permits knowledge discovery based not only on predefined characteristics or “facets” of digital objects, but also on the contexts in which they are placed by a community of scholars.21

Developing a practical interface for realizing that general goal will only emerge from building, testing, and – necessarily – modifying design models. The Collex design process has focused from the outset on access and usability of its key functions. Redesign will doubtless continue as new search and repurposing functions – some planned, some not yet conceived – are implemented. Let me close with some reflections about the quality of online information and the knowledge we expect to gain from it. Scrupulous minds regularly, and rightly, urge caution in using online materials which, as we all know, can be quite unreliable, or worse. Enthusiastic educators add that the problematic state of online information presents a great opportunity to help train students in methods of critical assessment. Scholars will take it for granted that if we make something available for study – online or in print makes no difference here – the material must be accurately disseminated. A stable URL for an online work ensures a certain level of authority for materials that are so easily copied and represented; but of course the data at that URL can also be easily modified and remodified without notice. It can also disappear. A scholar necessarily sees that situation as a problem. So if s/he creates a website, reliability and sustainability are (or should be) primary concerns. In fact, however, most scholars who create online materials – the “content” is often excellent – give little serious thought to either issue. For instance, most – nearly all – websites created in html will not outlive their creators, and the duration of the materials may well be much shorter even than that. Furthermore, suppose the materials are corrected or modified. Has provision been made to preserve those changes or even to keep an accurate record of them? Most scholars who use online resources don’t think about these matters for a very good reason: they are accustomed to working in a paper-based environment, which has worked out – built into its machinery – forms and procedures for dealing with such issues. These forms and procedures are not infallible but they are reliable. As Don McKenzie knew, our paperbased documentary record bears within itself the history of his own making. How will the digital record do that? We don’t yet know. What we do

Coda: Why digital textual scholarship matters

287

know – listen to the philologists – is that most of the record will be lost. So the problem is not to save this record: the problem is what we will choose to save, and why, and how will we do it? One final remark along these lines. If you look at most online scholarly editions – I leave aside the question of their structural stability – you will access a certain text or set of texts the scholar has chosen. The edition will perhaps be meshed with commentaries and notes and linked to other materials remotely located. Assuming all this is done knowledgeably and with care, how useful – in a scholarly sense – is the result? The answer comes if we compare such a work with an edition one might publish in print. Not without reason does the profession sanction the latter and look skeptically on the former. The printed book is immediately gathered into a complex network of trusted information and institutional relations that prepare it for further critical and scholarly engagements. The online work is more widely accessible, it’s true, but only in a relatively abstract sense. It is what the digerati call a “siloed” object – linked everywhere . . . and nowhere. It lacks the professional infrastructure that the scholarly book possesses by virtue of the mature social network in which it is located. And there are further problems with the way online editing is regularly imagined. With rare exceptions, when scholars create such works they transcribe a single text. Indeed, the common “digital view” of scholarly editing is that all one needs is a single, carefully produced digital transcription – complete with facsimiles of course – of a good copy of a particular work. Put that online, make it stable, and voila, ` a scholarly edition. But think again about the Mays edition of Coleridge, which is a critical interpretation of all the salient witnesses for each of Coleridge’s poems. More than that, the edition assumes that every single copy of every published (and unpublished) Coleridge work is potentially relevant to the critical survey. That assumption is what makes scholars assemble libraries of early materials and pressure their institutional depositories to augment their collections. One can certainly design and, to an extent, build an online version of a work like the Mays edition – The Rossetti Archive and The Canterbury Tales Project are good examples. More, digital resources have brought exciting new critical opportunities to scholarly works of that kind. Replete or innovative as they may be, however, they lack the institutional infrastructure that would integrate them, equally and simultaneously, to their bibliographical inheritance and their emerging digital network. Consequently, necessarily, these projects starve as scholarship. That situation is rapidly changing – for better and for worse, I’m quite sure. I sketch this history, and end with these problems, to help guide those changes along lines that will meet our scholarly and educational needs – the

288

jerome mcgann

traditional needs of what we might still want to conceive as our philologia perennis. NOTES ¨ und Methodologie der gesamten philologischen 1 August Boeckh, Enzyklopadie Wissenschaften, ed. Ernst Bratuschek (Leipzig: B. G. Teubner, 1877), 11. 2 D. F. McKenzie, “What’s Past is Prologue,” in Peter D. McDonald and Michael Suarez, SJ, ed., Making Meaning: “Printers of the Mind” and Other Essays (University of Massachusetts Press, 2002), 254. 3 Ren´e Wellek, Theory of Literature (New York: Harcourt Brace, 1949). 4 See Jerome McGann, “Textonics. Literary and Cultural Studies in a Quantum World,” printed in The Culture of Collected Editions, ed. Andrew Nash (Basingstoke: Palgrave Macmillan, 2003), 249. 5 Roger Chartier (trans. Teresa Lavender Fagan), “Languages, Books, and Reading from the Printed Word to the Digital Text,” Critical Inquiry, 31.1 (Autumn 2004), 133–52; Umberto Eco, “Afterword,” in Geoffrey Nunberg, ed., The Future of the Book (University of California Press), 295–301; Robert Darnton, “The Library in the New Age,” New York Review of Books 55.10 (June 12, 2008), 72–80. 6 N. Katherine Hayles, How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics (University of Chicago Press, 1999). 7 N. Katherine Hayles, Writing Machines (MIT Press, 2002), 19. 8 Ibid., 33. 9 Ibid., 43. 10 Ibid., 37. 11 Ibid., 33. 12 Ibid., 107. 13 Ibid., 101, 33. 14 Each is accessible online respectively at: www.perseus.tufts.edu/; www. rossettiarchive.org/; and www.canterburytalesproject.org/. 15 Jerome J. McGann, A Critique of Modern Textual Criticism (University of Chicago Press, 1983) and D. F. McKenzie, Bibliography and the Sociology of Texts (London: British Library, 1986). 16 Eliza Richards, Gender and the Poetics of Reception in Poe’s Circle (Cambridge University Press, 2004), 1. 17 See Jerome J. McGann, “Literary History and Editorial Method: Poe and Antebellum America,” New Literary History 40.4 (Autumn 2009), 825–42. 18 Richards, Gender, 5. 19 J. C. C. Mays, ed., The Collected Works of Samuel Taylor Coleridge: Poetical Works, Vols. i–iii, Parts 1, 2 (Princeton University Press, 2001). 20 Accessible online at www.nines.org 21 Bethany Nowviskie, “Collex: Facets, Folksonomy, and Fashioning the Remixable Web,” www.digitalhumanities.org/dh2007/abstracts/xhtml.xq?id=152

GUIDE TO FURTHER READING

The following list of further reading has been drawn from titles provided by the contributors of this volume and is intended to supplement the works mentioned in their endnotes. While not designed to be a comprehensive bibliography of textual scholarship, these readings should prove rewarding to those who wish to deepen their exploration of the key issues discussed in this volume.

General orientations and foundational texts Bajetta, Carlo M. “The Authority of Editing: Thoughts on the Function(s) of Textual Criticism,” Textus 19 (2006), 305–22. Burnard, Lou, Katherine O’Brien O’Keeffe, and John Unsworth, eds. Electronic Textual Editing. New York: The Modern Language Association of America, 2006. Cerquiglini, Bernard. In Praise of the Variant: A Critical History of Philology, trans. B. Wing. Baltimore: The Johns Hopkins University Press, 1999. Chernaik, Warren L., Caroline Davis, and Marilyn Deegan, eds. The Politics of the Electronic Text. University of London, 1993. DeFrancis, John. Visible Speech: The Diverse Oneness of Writing Systems. Honolulu: University of Hawaii Press, 1989. Eliot, Simon, and Jonathan Rose, eds. A Companion to the History of the Book, Oxford: Blackwell, 2007. Finneran, Richard J., ed. The Literary Text in the Digital Age. Ann Arbor: University of Michigan Press, 1996. Greetham, David C., ed. Scholarly Editing: A Guide to Research. New York: The Modern Language Association of America, 1995. Textual Scholarship: An Introduction. New York: Garland, 1992, 1994. Havelock, Eric A. The Muse Learns to Write: Reflections on Orality and Literacy from Antiquity to the Present. New Haven: Yale University Press, 1986. Johns, Adrian. The Nature of the Book: Print and Knowledge in the Making. Chicago: University of Chicago Press, 1998. Loizeaux, Elizabeth, and Neil Fraistat, eds. Reimagining Textuality: Textual Studies in the Late Age of Print. Madison: University of Wisconsin Press, 2002. Martin, Henri-Jean. The History and Power of Writing, trans. L. G. Cochrane. Chicago: University of Chicago Press, 1994.

289

290

guide to further reading

Morgan, Nigel J., and Rodney M. Thomson, eds. The Cambridge History of the Book in Britain, Vol. ii: 1100–1400. Cambridge: Cambridge University Press, 2008; Vol. iii: 1400–1557, ed. L. Hellinga and J. B. Trapp, 1999. Niles, John D. Homo Narrans: The Poetics and Anthropology of Oral Literature. Philadelphia: University of Pennsylvania Press, 1997. Pfeiffer, Rudolf. History of Classical Scholarship from 1300 to 1850. Oxford: Oxford University Press, 1976. History of Classical Scholarship from the Beginnings to the End of The Hellenistic Age. Oxford: Oxford University Press, 1968. Reiman, Donald H. Romantic Texts and Contexts. Columbia: University of Missouri Press, 1987. Rouse, Mary A., and Richard H. Rouse. Authentic Witnesses: Approaches to Medieval Texts and Manuscripts. Notre Dame, IN: University of Notre Dame Press, 1991. Schreibman, Susan, Ray Siemens, and John Unsworth, eds. A Companion to Digital Humanities. Oxford: Blackwell, 2004. Shillingsburg, Peter L. From Gutenberg to Google: Electronic Representation of Literary Texts. Cambridge: Cambridge University Press, 2006. Scholarly Editing in the Computer Age. 3rd edn. Ann Arbor: University of Michigan Press, 1996. Tanselle, G. Thomas. A Rationale of Textual Criticism. Philadelphia: University of Pennsylvania Press, 1989. Textual Criticism and Scholarly Editing. Charlottesville: University of Virginia Press, 1981. Wells, Stanley, and Gary Taylor with John Jowett and William Montgomery. William Shakespeare: A Textual Companion. Oxford: Oxford University Press, 1987. Williams, William Proctor, and Craig S. Abbott. An Introduction to Bibliographical and Textual Studies. New York: The Modern Language Association of America, 1999. Theoretical context Amodio, Mark C. Writing the Oral Tradition: Oral Poetics and Literate Culture in Medieval England. Notre Dame, IN: University of Notre Dame Press, 2004. Anderson, Benedict. Imagined Communities: Reflections on the Origin and Spread of Nationalism. 1983; London and New York: Verso, 1991. Baron, Dennis. A Better Pencil: Readers, Writers, and the Digital Revolution. Oxford: Oxford University Press, 2009. Barthes, Roland. Image–Music–Text, trans. Stephen Heath. New York: Hill and Wang, 1977. Bartlett, Francis Charles. Remembering: A Study in Experimental and Social Psychology. New York: Macmillan, 1932. Bellemin-No¨el, Jean. Le texte et l’avant-texte. Paris: Larousse, 1972. Benjamin, Walter. The Work of Art in the Age of its Technological Reproducibility and Other Writings on Media, ed. Michael W. Jennings, Brigid Doherty, and Thomas Y. Levin; trans. Edmund Jephcott, Rodney Livingston, Howard Eiland et al. Cambridge, MA: Harvard University Press, 2008. Bornstein, George, and Theresa L. Tinkle, eds. The Iconic Page in Manuscript, Print, and Digital Culture. Ann Arbor: University of Michigan Press, 1998.

guide to further reading

291

Burt, E. S., and Janie Vanpee, eds. Reading the Archive: On Texts and Institutions, Yale French Studies 77 (1990). Chandler, Daniel. The Act of Writing: A Media Theory Approach. Aberystwyth: Prifysgol Cymru, 1995. Chartier, Roger. Forms and Meanings: Texts, Performances and Audiences from Codex to Computer. Philadelphia: University of Pennsylvania Press, 1995. Cohen, Philip G., ed. Texts and Textuality: Textual Instability, Theory and Interpretation. New York: Garland, 1997. Croce, Benedetto. The Aesthetic on the Science of Expression and of the Linguistic in General, trans. C. Lyas. Cambridge: Cambridge University Press, 1992. De Certeau, Michel. The Poetics of Everyday Life, trans. S. Rendall. Berkeley: University of California Press, 1984. Derrida, Jacques. Archive Fever: A Freudian Impression, trans. Eric Prenowitz. Chicago: University of Chicago Press, 1996. Of Grammatology, trans. G. C. Spivak. Baltimore: Johns Hopkins University Press, 1974. Writing and Difference, trans. A. Bass. Chicago: University of Chicago Press, 1978. Dillon, Janette. “Is There a Performance in This Text?”, Shakespeare Quarterly 45 (1994), 74–86. Drucker, Johanna. The Word Made Flesh. New York: Granary Books, 1996. “Picture Problems: X-editing Images 1992–2010.” Digital Humanities Quarterly 3.3 (Summer 2009). http://digitalhumanities.org/dhq/vol/3/3/000052/000052. html. Eaves, Morris. “‘Why Don’t They Leave it Alone?’ Speculations on the Authority of the Audience in Editorial Theory,” in Cultural Artifacts and the Production of Meaning: The Page, the Image, and the Body, ed. Margaret Ezell and Katherine O’Brien O’Keeffe. Ann Arbor: University of Michigan Press, 1994, 85–99. Eco, Umberto. The Open Work, trans. A. Cancogni. Cambridge, MA: Harvard University Press, 1990. Eliot, Simon, Andrew Nash, and Ian Willison, eds. Literary Cultures and the Material Book. The British Library, 2007. Goodman, Nelson. Languages of Art: An Approach to a Theory of Symbols, 2nd edn. Indianapolis: Hackett, 1976. Goody, Jack. The Interface between the Written and the Oral. Cambridge: Cambridge University Press, 1987. Grafton, Anthony. Worlds Made by Words: Scholarship and Community in the Modern West. Cambridge, MA: Harvard University Press, 2009. Greetham, David C. Theories of the Text. Oxford: Oxford University Press, 1999. Grigely, Joseph. Textualterity: Art, Theory, and Textual Criticism. Ann Arbor: University of Michigan Press, 1995. Hall, Gary. Digitize This Book! The Politics of New Media, or Why We Need Open Access Now. Minneapolis: University of Minnesota Press, 2008. Hayles, N. Katherine. Writing Machines. Cambridge, MA: MIT Press, 2002. Honko, Lauri, ed. Textualization of Oral Epics. Berlin: Mouton de Gruyter, 2000. Ivins, William Mills. Prints and Visual Communication. Cambridge, MA: Harvard University Press, 1953. Jed, Stephanie H. Chaste Thinking: The Rape of Lucretia and the Birth of Humanism. Bloomington and Indianapolis: Indiana University Press, 1989.

292

guide to further reading

Josipovici, Gabriel. Touch. New Haven: Yale University Press, 1996. Kirschenbaum, Matthew G. Mechanisms: New Media and the Forensic Imagination. Cambridge, MA: MIT Press, 2008. Kuskin, William. Symbolic Caxton: Literary Culture and Print Capitalism. Notre Dame, IN: University of Notre Dame Press, 2008. Leyton, Michael. Symmetry, Causality, Mind. Cambridge, MA: MIT Press, 1992. The Structure of Paintings. Vienna: Springer, 2006. Lord, Albert Bates. Epic Singers and Oral Tradition. Ithaca: Cornell University Press, 1991. The Singer of Tales. Cambridge, MA: Harvard University Press, 1960; 2nd edn., 2000. The Singer Resumes the Tale, ed. Mary Louise Lord. Ithaca: Cornell University Press, 1995. Manguel, Alberto. A History of Reading. New York: Penguin, 1996. Manoff, Marlene. “The Materiality of Digital Collections: Theoretical and Historical Perspectives,” Libraries and the Academy 6.3 (2006), 311–25. McGann, Jerome J. A Critique of Modern Textual Criticism. University of Chicago Press, 1983. Repr. Charlottesville: University of Virginia Press, 1992. Radiant Textuality: Literature After the World Wide Web. New York: Palgrave, 2001. “Texts in N-Dimensions and Interpretation in a New Key,” Text Technology 12.2 (2003), 1–18. The Textual Condition. Princeton: Princeton University Press, 1991. ed. Textual Criticism and Literary Interpretation. Chicago: University of Chicago Press, 1985. McKenzie, D. F. Bibliography and the Sociology of Texts. The Panizzi Lectures. London: British Library, 1986. Modiano, Raimonda, Leroy Searle, and Peter L. Shillingsburg, eds. Voice, Text, Hypertext: Emerging Practices in Textual Studies. Seattle: University of Washington Press, 2004. Mundal, Else, and Jonas Wellendorf, eds. Oral Art Forms and Their Passage into Writing. Copenhagen: Museum Tusculaneum Press, 2008. Murphy, Andrew, ed. The Renaissance Text. Manchester: Manchester University Press, 2000. Murray, Simone. Mixed Media: Feminist Presses and Publishing Politics. London: Pluto Press, 2004. Nunberg, Geoffrey, ed. The Future of the Book. Berkeley and Los Angeles: University of California Press, 1996. Ong, Walter. Orality and Literacy: The Technologizing of the Word. London and New York: Routledge, 1982. Panofsky, Erwin. Studies in Iconology. New York: Harper & Row, 1962. Reid, Martine. “Editor’s Preface. Legible/Visible,” Yale French Studies 84 (1994), 1–12. Reiman, Donald H. The Study of Modern Manuscripts: Public, Confidential, and Private. Baltimore: The Johns Hopkins University Press, 1993. Said, Edward. Orientalism. New York: Pantheon, 1978. Sampson, Geoffrey. Writing Systems: A Linguistic Introduction. Stanford: Stanford University Press, 1985.

guide to further reading

293

Stock, Brian. Listening for the Text: On the Uses of the Past. Baltimore: The Johns Hopkins University Press, 1990. Street, Brian V. Literacy in Theory and Practice. Cambridge: Cambridge University Press, 1984. Sutherland, Kathryn. Jane Austen’s Textual Lives: From Aeschylus to Bollywood. Oxford: Oxford University Press, 2005. “Revised Relations? Material Text, Immaterial Text, and the Electronic Environment,” TEXT 11 (1998), 17–39. Tedlock, Dennis. The Spoken Word and the Work of Interpretation. Philadelphia: University of Pennsylvania Press, 1983. Thompson, D’Arcy Wentworth. On Growth and Form. An abridged edn. Cambridge: Cambridge University Press, 1961. Vandendorpe, Christian. From Papyrus to Hypertext: Toward the Universal Digital Library, trans. P. Aronoff and H. Scott. Champaign: University of Illinois Press, 2009. Willats, John. Art and Representation: New Principles in the Analysis of Pictures. Princeton: Princeton University Press, 1997. Zumthor, Paul. Oral Poetry: An Introduction, trans. K. Murphy-Judy. Minneapolis: University of Minnesota Press, 1990. Historical context Alexander, Jonathan J. G. Medieval Illuminators and Their Methods of Work. New Haven: Yale University Press, 1994. Arber, Edward, ed. A Transcript of the Registers of the Company of Stationers of London 1554–1640 ad, 5 vols., 1875–94. Repr. Gloucester, MA: P. Smith, 1950. Bataillon, Louis Jacques, Bertrand G. Guyot, and Richard H. Rouse, eds. La pro´ duction du livre universitaire au moyen age: exemplar et pecia. Paris: Editions du Centre National de la Recherche Scientifique, 1991. Battles, Matthew. Library: An Unquiet History. New York: Norton, 2003. Bender, John, and Michael Marrinan, eds. Regimes of Description: In the Archive of the Eighteenth Century. Stanford: Stanford University Press, 2005. Bloch, R. Howard, and Stephen G. Nichols, eds. Medievalism and the Modernist Temper. Baltimore: The Johns Hopkins University Press, 1996. Bradbury, Nancy M. Writing Aloud: Storytelling in Late Medieval England. Champaign: University of Illinois Press, 1998. Camille, Michael. Image on the Edge: The Margins of Medieval Art. Cambridge, MA: Harvard University Press, 1992. Carlson, David R. English Humanist Books: Writers and Patrons, Manuscript and Print, 1475–1525. Toronto: University of Toronto Press, 1993. Casson, Lionel. Libraries in the Ancient World. New Haven: Yale University Press, 2001. Chartier, Roger. Inscription and Erasure: Literature and Culture from the Eleventh to the Eighteenth Century, trans. A. Goldhammer. Philadelphia: University of Pennsylvania Press, 2007. “Languages, Books, and Reading from the Printed Word to the Digital Text,” trans. T. L. Fagan, Critical Inquiry 31.1 (Autumn 2004), 133–52.

294

guide to further reading

The Order of Books: Readers, Authors, and Libraries in Europe between the Fourteenth and Eighteenth Centuries, trans. L. G. Cochrane. Stanford: Stanford University Press, 1994. Clanchy, M. T. From Memory to Written Record: England 1066–1307, 2nd edn. Oxford: Blackwell, 1993. Cohen, Charles Lloyd, and Paul S. Boyer, eds. Religion and the Culture of Print in Modern America. Madison: University of Wisconsin Press, 2008. De Hamel, Christopher. The Book: A History of the Bible. London: Phaidon, 2001. Medieval Craftsmen: Scribes and Illuminators. Toronto: University of Toronto Press, 1992. Doane, Alger Nicolaus, and Carol Braun Pasternack, eds. Vox Intexta: Orality and Textuality in the Middle Ages. Madison: University of Wisconsin Press, 1991. Duckett, Kenneth W. Modern Manuscripts. Nashville: American Association for State and Local History, 1975. Echard, Sian. ˆ Printing the Middle Ages. Philadelphia: University of Pennsylvania Press, 2008. Eggert, Paul. Securing the Past: Conservation in Art, Architecture and Literature. Cambridge: Cambridge University Press, 2009. Eisenstein, Elizabeth L. The Printing Press as an Agent of Change: Communications and Cultural Transformations in Early-modern Europe. Cambridge: Cambridge University Press, 1979. Fussel, Stephan. Gutenberg and the Impact of Printing, trans. D. Martin. Farnham, ¨ Surrey: Ashgate Publishing, Scolar Press, 2003. Gellrich, Jesse M. The Idea of the Book in the Middle Ages: Language Theory, Mythology, and Fiction. Ithaca, NY: Cornell University Press, 1985. Hamburger, Jeffrey F. The Visual and the Visionary: Art and Female Spirituality in Late Medieval Germany. New York: Zone Books, 1998. Hanna, Ralph. London Literature, 1300–1380. Cambridge: Cambridge University Press, 2005. Pursuing History: Middle English Manuscripts and Their Texts. Stanford: Stanford University Press, 1996. Kuskin, William, ed. Caxton’s Trace: Studies in the History of English Printing. Notre Dame, IN: University of Notre Dame Press, 2006. Love, Harold. Scribal Publication in Seventeenth-Century England. Oxford: Clarendon Press, 1993. Machan, Tim William, ed. Medieval Literature: Texts and Interpretation. Medieval & Renaissance Texts & Studies. State University of New York at Binghamton, 1991. Piper, Andrew. Dreaming in Books: The Making of the Bibliographic Imagination in the Romantic Age. Chicago: University of Chicago Press, 2009. Raven, James, ed. The Destruction of Great Book Collections Since Antiquity. London: Palgrave Macmillan, 2004. Zumthor, Paul. Toward a Medieval Poetics, trans. P. Bennett. Minneapolis: University of Minnesota Press, 1992. Editions and discussion of specific editions Bertram, Paul, and Bernice Kliman, eds. Three-Text Hamlet: Parallel Texts of the First and Second Quartos and First Folio. New York: AMS Press, 1991.

guide to further reading

295

Braun-Rau, Alexandra. William Shakespeares King Lear in seinen Fassungen, Beihefte zu editio 20. Tubingen: Niemeyer, 2004. www.textkritik.uni¨ muenchen.de/abraun-rau/lear/index.html. Clayton, Thomas, ed. The Hamlet First Published (Q1, 1603): Origins, Form, Intertextualities. Newark: University of Delaware Press, 1992. Eaves, Morris, Robert Essick, and Joseph Viscomi, eds. The William Blake Archive. www.blakearchive.org. Franklin, Ralph William, ed. The Manuscript Books of Emily Dickinson, 2 vols. Cambridge, MA: The Belknap Press of Harvard University Press, 1981. Gabler, Hans Walter, ed. Ulysses: A Critical and Synoptic Edition. New York: Garland, 1984. Gurr, Andrew. “Maximal and Minimal Texts: Shakespeare v. the Globe,” Shakespeare Survey 52 (1999), 68–87. Holderness, Graham, and Bryan Loughrey, eds. Shakespearean Originals – First Editions. Hemel Hempstead: Harvester Wheatsheaf, 1992. Honigmann, E. A. J. The Stability of Shakespeare’s Text. London: Edward Arnold, 1965. Howe, Susan. A Bibliography of the King’s Book or Eikon Basilike. Dexter, MI: Paradigm Press, 1989. Kahan, Jeffrey, ed. King Lear: New Critical Essays. London and New York: Routledge, 2008. Kane, George, ed. Piers Plowman: The A Version. Will’s Visions of Piers Plowman and Do-Well. An Edition in the Form of Trinity College Cambridge MS R.3.14 Corrected from Other Manuscripts, with Variant Readings. London: University of London, The Athlone Press, 1960. Kane, George, and E. Talbot Donaldson, eds. Piers Plowman: The B Version. Will’s Visions of Piers Plowman, Do-Well-, Do-Better and Do-Best. An Edition in the Form of Trinity College Cambridge MS B.15.17, Corrected and Restored from the Known Evidence, with Variant Readings. London: University of London, The Athlone Press, 1975. Lachmann, Karl, ed. Lucretii De Rerum Natura Libri VI. Berlin, 1850. Lernout, Geert. “Controversial Editions: Hans Walter Gabler’s Ulysses,” Text: An Interdisciplinary Annual of Textual Studies 16 (2006), 229–41. Netz, Reviel, and William Noel. The Archimedes Codex: How a Medieval Prayer Book is Revealing the True Genius of Antiquity’s Greatest Scientist. Philadelphia: Da Capo Press, 2007. Paradigm Project. Workbook on Digital Private Papers. Oxford: Bodleian Library, 2005–7. www.paradigm.ac.uk/workbook. Parry, Milman. The Making of Homeric Verse, ed. A. Parry. Oxford: Clarendon Press, 1971. Patterson, Lee. “The Logic of Textual Criticism and the Way of Genius: The KaneDonaldson Piers Plowman in Historical Perspective,” in Negotiating the Past: The Historical Understanding of Medieval Literature. Madison: University of Wisconsin Press, 1987, 77–113. Reiman, Donald H., and Neil Fraistat, eds. The Complete Poetry of Percy Bysshe Shelley. Baltimore: The Johns Hopkins University Press, 2000–. Reynolds, L. D., and N. G. Wilson. Scribes and Scholars: A Guide to the Transmission of Greek and Latin Literature, 2nd edn., Oxford: Oxford University Press, 1974.

296

guide to further reading

Ross, Charles L., and Dennis Jackson, eds. Editing D. H. Lawrence: New Versions of a Modern Author. Ann Arbor: University of Michigan Press, 1995. Russell, George, and George Kane, eds. Piers Plowman: The C Version. Will’s Visions of Piers Plowman, Do-Well, Do-Better and Do-Best. An Edition in the Form of Huntington Library ms hm 143, Corrected and Restored from the Known Evidence, with Variant Readings. Berkeley: University of California Press, 1997. Taylor, Gary, and Michael Warren, eds. The Division of the Kingdoms. Oxford: Oxford University Press, 1983. Thompson, Ann, and Neil Taylor, eds. Hamlet and Hamlet: The Texts of 1603 and 1623, 2 vols. London: Arden Shakespeare, 2006. Tronch-P´erez, Jesus, ´ ed. A Synoptic Hamlet: A Critical-Synoptic Edition of the Second Quarto and First Folio Texts of Hamlet. Val`encia: Universitat de Val`encia; Zaragoza: SEDERI, 2002. Urkowitz, Steven. Shakespeare’s Revision of King Lear. Princeton: Princeton University Press, 1980. “‘Well-sayd olde mole’: Burying Three Hamlets in Modern Editions,” in Shakespeare Study Today, ed. Georgianna Ziegler. New York: AMS Press, 1986, 37–80. Werstine, Paul. “The Textual Mystery of Hamlet,” Shakespeare Quarterly 39 (1988), 1–26. Wilson, F. P. “Shakespeare and the New Bibliography,” in The Bibliographical Society 1892–1942: Studies in Retrospect, ed. F. C. Francis. London: Bibliographical Society, 1945. Revised as Shakespeare and the New Bibliography, ed. Helen Gardner. Oxford: Clarendon Press, 1970. Discussions of editorial method B´edier, Joseph. “La tradition manuscrite du Lai de l’Ombre: R´eflexions sur l’art d’´editer les anciens textes,” Romania 54 (1928), 321–56. Blecua, A. “Defending Neolachmannianism: On the Palacio Manuscript of La Celestina,” in Variants, ed. P. Robinson and H. T. M. Van Vliet. Turnhout: Brepols, 2002, 113–33. Bornstein, George. Material Modernism. Cambridge: Cambridge University Press, 2001. ed. Representing Modernist Texts: Editing as Interpretation. Ann Arbor: University of Michigan Press, 1991. Bowers, Fredson. Essays in Bibliography, Text and Editing. Charlottesville: University of Virginia Press, 1975. “Multiple Authority: New Problems and Concepts of Copy-Text,” The Library, 5th ser., 27 (1972), 81–115. Bryant, John. The Fluid Text. Ann Arbor: University of Michigan Press, 2002. Busby, Keith, ed. Towards a New Synthesis? Essays on the New Philology. Atlanta, GA: Rodopi, 1993. Cohen, Philip G., ed. Devils and Angels: Textual Editing and Literary Theory. Charlottesville: University of Virginia Press, 1991. Dahlstrom, Mats, Espen S. Ore, and Edward Vanhoutte, eds. Electronic Scholarly ¨ Editing – Some Northern European Approaches, A special issue of Literary and Linguistic Computing 19.1 (2004).

guide to further reading

297

De Biasi, Pierre-Marc. “What is a Literary Draft? Toward a Functional Typology of Genetic Documentation,” Yale French Studies 89: Drafts (1996), 26–58. Dennis, Rodney G., and Elizabeth A. Falsey, eds. The Marks in the Fields: Essays on the Uses of Manuscripts. Cambridge, MA: Houghton Library/Harvard University Press, 1992. Deppman, Jed, Daniel Ferrer, and Michael Groden, eds. Genetic Criticism: Texts and Avant-textes. Philadelphia: University of Pennsylvania Press, 2004. Dow, Elizabeth H. Electronic Records in the Manuscript Repository. Lanham, MD: Scarecrow, 2009. Eggert, Paul. “The Book, the E-text, and the ‘Work-Site’.” Text Editing, Print, and the Digital World, ed. Marilyn Deegan and Kathryn Sutherland. Farnham: Ashgate, 2008, 63–82. Falconer, Graham. “Genetic Criticism,” Comparative Literature 45.1 (1993), 1–21. Gabler, Hans Walter. “Beyond Author-Centricity in Scholarly Editing,” Journal of Early Modern Studies 1.1 (2012), 15–35. “Questioni: The Primacy of the Document in Editing,” Ecdotica 4 (2007), 197– 207. “Theorizing the Digital Scholarly Edition,” Literature Compass 7.2 (2010), 43– 56. Gabler, Hans Walter, George Bornstein, and G. B. Pierce, eds. Contemporary German Editorial Theory. Ann Arbor: The University of Michigan Press, 1995. Greg, Walter W. The Calculus of Variants: An Essay on Textual Criticism. Oxford: Oxford University Press, 1927. “The Rationale of Copy-Text,” Studies in Bibliography 3 (1950–1), 19–36. Gumbrecht, Hans Ulrich. The Powers of Philology: Dynamics of Textual Scholarship, Champaign, University of Illinois Press, 2003. Hay, Louis, ed. Essais de critique g´en´etique. Paris: Flammarion, 1979. Jenny, Laurent. “Genetic Criticism and its Myths,” Yale French Studies 89 (1996), 9–25. Kelemen, Erick, ed. Textual Editing and Criticism: An Introduction. New York: Norton, 2009. Kenney, E. J. The Classical Text: Aspects of Editing in the Age of the Printed Book. Berkeley: University of California Press, 1974. Lernout, Geert. “Genetic Criticism and Philology,” Text: An Interdisciplinary Annual of Textual Studies 14 (2002), 53–6. Loewenstein, Joseph F. “Authentic Reproductions: The Material Origins of the New Bibliography.” Textual Formations and Reformations, Laurie E. Maguire and Thomas L. Berger, eds. Newark: University of Delaware Press, 1998. MacDonald, Claire, and William H. Sherman, eds. On Editing, special issue of Performance Research 7.1 (March 2002). Mac´e, Caroline, Philippe Baret, Andrea Bozzi, and Laura Cignoni, eds. The Evolution of Texts: Confronting Stemmatological and Genetical Methods, Linguistica Computazionale 24–5. Pisa: Istituti Editoriali e Poligrafici Internazionali, 2006. Smith, Martha Nell. “Software of the Highest Order: Revisiting Editing as Interpretation,” Textual Cultures 2.1 (Spring 2007), 1–15. Steding, Soren A. Computer-Based Scholarly Editions: Context–Concept–Creation– Clientele. Berlin: Logos Verlag, 2002.

298

guide to further reading

Sturges, Robert S. “Textual Scholarship: Ideologies of Literary Production,” Exemplaria 3.1 (1991), 109–31. Sutherland, Kathryn. “Being Critical: Paper-based Editing and the Digital Environment,” in Text Editing, Print, and the Digital World, ed. Marilyn Deegan and Kathryn Sutherland. Farnham: Ashgate, 2009, 13–25. Tanselle, G. Thomas. “Editing without a Copy-Text,” Studies in Bibliography 47 (1994), 1–22. Rpt. in Tanselle, Literature and Artifacts. Charlottesville: Bibliographical Society of the University of Virginia, 1998, 236–57. “The Editorial Problem of Final Authorial Intention,” Studies in Bibliography 29 (1976), 167–211. Taylor, Gary. “The Rhetoric of Textual Criticism,” TEXT 4 (1988), 39–57. Timpanaro, Sebastiano. The Genesis of Lachmann’s Method, trans. G. Most. Chicago: University of Chicago Press, 2005. Tisseron, Serge. “All Writing is Drawing: The Spatial Development of the Manuscript,” Yale French Studies 84 (1994), 29–42. Van Hulle, Dirk. “Compositional Variants in Modern Manuscripts,” Digital Technology and Philological Disciplines, Linguistica Computazionale 20–1 (2004): 513–27. MSS: Modern Manuscript Studies, http://modernmanuscripts.blogspot.com/. Textual Awareness: A Genetic Study of Late Manuscripts by Joyce, Proust, and Mann. Ann Arbor: University of Michigan Press, 2004. “Textual Descent with Intentional Modification: Genetic Criticism and Genetic Engineering,” Linguistica Computazionale 24–25 (2006), 215–28. Van Hulle, Dirk, and Wim Van Mierlo (eds.) Reading Notes. Variants: Journal for the European Society for Textual Scholarship 2.3 (2004). Vanhoutte, Edward. “Every Reader his Own Bibliographer – An Absurdity?” in Text Editing, Print and the Digital World, ed. Marilyn Deegan and Kathryn Sutherland. Farnham: Ashgate, 2009, 99–112.

Studies of specific authors and problems Arnheim, Rudolf, et al. Poets at Work: Essays Based on the Modern Poetry Collection at the Lockwood Memorial Library. New York: Harcourt, Brace and Company, 1948. Bustarret, Catherine. “Paper Evidence and the Interpretation of the Creative Process in Modern Literary Manuscripts,” L’Esprit Cr´eateur 41 (Summer 2001), 16–28. Butterick, George. “Modern Literary Manuscripts and Archives: A Field Report,” Credences New Series 1 (1980), 81–103. Cameron, Sharon. Choosing Not Choosing: Dickinson’s Fascicles. Chicago: University of Chicago Press, 1992. Coe, Michael D. Breaking the Maya Code, revised edn. New York: Thames and Hudson, 1999. Darnton, Robert. “Google and the New Digital Future,” New York Review of Books 56.20 (December 17, 2009). “The Library in the New Age,” New York Review of Books 55.10 (June 12, 2008). Eaves, Morris. “Picture Problems: X-Editing Images 1992–2010,” Digital Humanities Quarterly 3.3 (2009). www.digitalhumanities.org/dhq/vol/3/3/000052/ 000052.html.

guide to further reading

299

Foley, John M. The Singer of Tales in Performance. Bloomington: Indiana University Press, 1995. Fraistat, Neil. “Illegitimate Shelley: Radical Piracy and the Textual Edition as Cultural Performance,” PMLA 110 (May 1994), 409–23. Greetham, David C. Textual Transgressions: Essays Toward the Construction of a Biobibliography. New York: Garland, 1998. Hodson, Sara S. “Private Lives: Confidentiality in Manuscripts Collections,” Rare Books and Manuscripts Librarianship 6 (1991), 108–18. Hymes, Dell. “In Vain I Tried to Tell You”: Essays in Native American Ethnopoetics. Philadelphia: University of Pennsylvania Press, 1981. Now I Know Only So Far: Essays in Ethnopoetics. Lincoln: University of Nebraska Press, 2003. Jowett, John. Shakespeare and Text. Oxford: Oxford University Press, 2007. Kirschenbaum, Matthew G., et al. Approaches to Managing and Collecting BornDigital Literary Materials for Scholarly Use. Office of Digital Humanities, National Endowment for the Humanities, 2009. www.neh.gov/ODH/Default. aspx?tabid=111 & id=37. Kirschenbaum, Matthew G., Richard Ovenden, and Gabriela Redwine. Digital Forensics and Born-Digital Content in Cultural Heritage Collections. Council on Library and Information Resources, 2010. Luthin, Herbert W., ed. Surviving Through the Days: Translations of Native California Stories and Songs. Berkeley: University of California Press, 2002. Maas, Paul. Textual Criticism, trans. B. Flower. Oxford: Oxford University Press, 1958. Masten, Jeffrey. Textual Intercourse: Collaboration, Authorship, and Sexualities in Renaissance Drama. Cambridge: Cambridge University Press, 1997. McGann, Jerome J. Black Riders: The Visible Language of Modernism. Princeton: Princeton University Press, 1993. Middleton, Charles Henry. A Descriptive Catalogue of the Etched Work of Rembrandt van Rhyn. London: Murray, 1878. Ross, Seamus, and Ann Gow. Digital Archaeology: Rescuing Neglected and Damaged Data Resources. Glasgow: Humanities Advanced Technology and Information Institute, 1999. www.ukoln.ac.uk/services/elib/papers/supporting/pdf/ p2con.pdf. Smith, Martha Nell. Rowing in Eden: Rereading Emily Dickinson. Austin: University of Texas Press, 1992. Viscomi, Joseph. Blake and the Idea of the Book. Princeton: Princeton University Press, 1993. Weitzmann, Kurt. Illustrations in Roll and Codex: A Study of the Origin and Method of Text Illustration, 2nd edn. Princeton: Princeton University Press, 1970. Werner, Marta. Radical Scatters: Emily Dickinson’s Fragments and Related Texts, 1970–86. Ann Arbor: University of Michigan Press, 1999.

INDEX

Abbot, Craig S. An Introduction to Bibliographical and Textual Studies, 2 accidental(s), 49, 50, 52, 103, 105, 107, 188 Aeschylus, 19, 23 Alcaeus, 21 Alcuin of York, 26 Aldine press, 27 Alexander, Jonathan, 248 Alexandria, 4, 5, 20, 21, 22, 26, 39, 63 Library of, 19, 23, 229 Allemann, Beda, 68, 69 allography, 245 Ambrosian Library, 33 American Folklore Society, 218 Amory, Hugh, 36 Anacreon, 21 analogy, 4, 21, 22, 39, 80 Angelo, Jacopo, 27 anomaly, 4, 22, 39 apparatus, 5, 7, 46, 52, 53, 68, 70, 83, 98, 99, 102, 103, 106, 107, 108–9, 114, 125, 220, 232, 236, 240, 282 Appian Civil Wars, 39 Aquila, 229 Arbeitsgemeinschaft fur ¨ germanistische Edition, 75 archetype, 29, 32, 65, 80–1, 212, 214, 215, 237, 239 Archilochus, 21 Archimedes, 253, 254 Archive of Folk Culture, 221 Archive of the School of Scottish Studies, 221 Aristarchus of Samothrace, 21 Aristophanes, 23 Aristotle, 31, 285 Athenaeus The Learned Banqueters, 23

Augustine In psalmos, 26 Aurispa, Giovanni, 27 Austen, Jane, 4, 43, 52, 268, 279 Mansfield Park, 43 Sense and Sensibility, 43 Bancroft, Richard, 33 Banks, Russell, 271 Barbi, Michele, 74 (ed.) Vita nuova, 74 Barthes, Roland, 71, 280 “Death of the Author,” 230 “From Work to Text,” 54 Bartlett, Frederic, 246–8, 250–3, 254 Beckett, Samuel, 271 Bede, 26 B´edier, Joseph, 6, 8, 10, 22, 32, 70, 76, 130, 232, 234 (ed.) Song of Roland, 121–4, 214 Beißner, Friedrich, 68 (ed.) Holderlin, 67 ¨ Bellemin-No¨el, Jean, 6, 71, 72 Bengel, Johann Albrecht (ed.) Greek New Testament, 232 Prodromus, 232 Benjamin, Walter, 230 Bentley, Richard, 64, 66, 67 (ed.) Paradise Lost, 32, 42 (ed.) Terrence, 32 Beowulf, 120, 211 Bible, 63, 64, 66, 197, 198, 199, 229 Complutensian, 30 Dead Sea Scrolls, 23, 24 distinction between Catholic and Protestant, 199 Greek New Testament, 25 Greek Septuagint, 23, 24, 25, 229 Gutenberg’s printing of, 36, 230

301

302

index

Bible (cont.) Hebrew (and Torah), 23, 24, 25, 198, 228, 229 Jerome’s Vulgate and Origen’s Hexapla, 25 King James version, 30, 199 Latin, 26, 230 Masoretic, 24 Nag Hammadi, 23 The Souldiers Pocket Bible, 197 Bibliographical Society (London) The Library, 46 Bibliographical Society of America, 36 bibliography, 35, 44, 46, 47, 56, 81, 82, 116 analytical, 2, 9, 19, 21, 36, 37, 50, 71, 79, 81, 87 enumerative (descriptive), 2, 23, 34 historical, 2 links to textual criticism and literary criticism, 47 manuscript, 17 New Bibliography, 1, 2, 4, 6, 20, 45, 47, 48, 54, 55, 79, 80, 81, 82 of Shakespeare, 46 socio-cultural context of, 37 Biblioth`eque Nationale, 23, 33, 71 Gallica project, 73 Blackstone, William, 192 Blake, William, 107 Blessing, Lee, 271 Boas, Franz (ed.) “The Grizzly Bear”, 218, 219 Boccaccio, 27, 28, 194, 195 Bodleian Library, 33, 197, 271 Bodley, Thomas, 33 Boeckh, August, 274 Bohmer, Eduard ¨ (ed.) Song of Roland, 121 Bonaparte, Napoleon, 126 Boon, Karel catalogue of Rembrandt etchings, 242, 243 Borges, Jorge Luis, 201, 255 Funes the Memorius, 255 born-digital material, 3, 11, 17, 259, 260, 261, 264, 272, 276, 278, 280, 283 challenges of, 272 classification of, 285 and cloud computing, 270 definition of, 260 first generation works, 279 handling of, 271 significance of, 272

and textual scholarship, 261, 271 transition to, 262 Borromeo, 33 Bowers, Fredson, 2, 4, 35, 37, 45, 47, 50, 51, 56, 57, 58, 106, 107, 233 Principles of Bibliographical Description, 50 Studies in Bibliography, 51 “Textual Criticism,” 50 The Works of Stephen Crane, 52, 53, 56 Bracciolini, Poggio, 28, 229 Brantome (Pierre de Bourdeille) ˆ Livre des dames, 191 Brault, Gerald J. (ed.) Song of Roland, 214 Brepols, 124 British Library, 23, 33, 200 Digital Lives project, 271 British Museum Library, 33 Brod, Max, 69 Bronson, Bertrand, 206, 208 The Traditional Tunes of the Child Ballads, 208 Bront¨e, Charlotte Villette, 53 Bryant, John, 125 The Fluid Text, 108 Buchner, Georg, 69 ¨ Bud´e, Guillaume, 33 Buzzetti, Dino, 3 Byzantium, 27 Cabrera Nunez ˜ de Guzman, Melchor de, 188, 189 Callimachus Pinakes, 23 Canfora, Luciano, 76 Capella, Martianus Nuptiae Philologiae et Mercurii, 24 Catullus, 27 Caxton, William, 34 Center for Editions of American Authors (CEAA), 51, 52, 104 Committee on Scholarly Editions (CSE), 51 Center for Manuscript Genetics (Belgium), 74 Centre d’Analyse des Manuscrits, 71 Cerquiglini, Bernard, 38, 76, 234 In Praise of the Variant, 38, 76, 233 Cervantes, Miguel de, 190, 193 Don Quixote, 190, 200, 201 Chadwyck-Healey, 275 English Poetry Database, 275

index Chamberlain’s Men, 93, 110, 112 Chandler, Daniel, 260 Chapman, R. W., 4, 42, 43 (ed.) Mansfield Park, 43 (ed.) Novels of Jane Austen, 52 (ed.) Sense and Sensibility, 43 Charlemagne, 26, 126 Charles IV, 33 Charles V, 33 Chartier, Roger, 9, 276 Chaucer, Geoffrey, 34 Canterbury Tales Project, 279, 287 Legend of Good Women, 25 Child, Francis James, 206 Chrysoloras, Manuel, 27 Cicero, 28 De Republica, 26 Epistolae ad familiares, 231 Verrine orations, 232 Clancy, Tom Hunt for Red October, 258 Clarendon Press, 43 Clarke, Arthur C, “The Steam Powered Word Processor,” 258 Clayton, Thomas, 113 codex, 22, 32, 39, 40, 194, 199, 200, 254 coinage, 17 Coleridge, Samuel Taylor, 282, 283, 287 collation, 11, 21, 26, 28, 31, 43, 52, 103, 105, 206, 228, 237, 239, 240–5, 254 and collocation, 241 definition of, 240 of images, 237 textual vs. pictorial, 240 Coll`ege de France, 70 Collex interface, 285, 286 collocation, 234, 240, 241, 244 definition of, 240 Columbus, Christopher, 125 Commena, Anna, 26 Committee for Scholarly Editions (CSE), 104 Condell, Henry, 88, 196 ´ Condillac, Etienne Bonnot de, 191 Constantine, 28 Constantinople, 26 Contini, Gianfranco, 74 copy-text, 5, 35, 47, 48, 49, 53, 68, 92, 105, 106, 107, 113 definition of, 49, 104 Cottle, Joseph, 283 Cousin, Victor, 69 Crane, Greg, 1, 2

303

Crane, Stephen, 52, 53, 56 Maggie, 52 Croce, Benedetto, 73 Crowley, John, 271 cuneiform, 17 Curtis, Richard Skinhead Hamlet, 197 d’Anjou, Ren´e, 195 Damasus I, 26 Daniel, Samuel The Queene’s Arcadia, 196 Dante Alighieri, 73, 74, 193, 194, 195 Divina commedia, 74, 194 Vita nuova, 74 Darnton, Robert, 276 De Niro, Robert, 271 De Smedt, Marcel, 74 deep editing, 238 Deleuze, Gilles, 195 Derrida, Jacques, 225 Dickens, Charles, 52 Oliver Twist, 52, 56 Dickinson, Emily, 107 Dickinson Electronic Archive, 200 Diderot, Denis, 192 La Religieuse, 191 M´emoire sur le commerce de la librairie, 192 Didion, Joan, 259 Dillon, Janette, 113 Doane, A. N., 209 (ed.) Wiþ Færstice, 209, 210 Donaldson, E. T., 38 (ed.) Piers Plowman, 37, 233 Donation of Constantine, 28 Doran, Madeleine, 79, 82 Dryden, John, 231 Duggan, Joseph, 214 (ed.) Song of Roland, 124 Duguid, Paul, 38 Durer, Albrecht, 250 ¨ Early English Books Online (EEBO), 46 Early English Text Society (EETS), 8, 34, 126 Eaves, Morris, 236 Eco, Umberto, 259, 276 Foucault’s Pendulum, 258 Ecole Normale Sup´erieure, 70 e´ criture, 6, 72 edition (concept of), 2, 3, 7, 8, 9, 13, 58, 70, 74, 88, 90, 91, 92, 94, 95, 97, 98, 99, 100, 102–3, 104, 108, 109, 124, 125, 126–7, 128, 129, 131, 232

304

index

edition (concept of) (cont.) critical, 5, 6, 7, 8, 46, 49, 51, 52, 53, 64, 66, 67, 68, 70, 74, 104–7, 109, 116, 124, 212, 214–15, 221, 232, 233, 234, 282, 283 diplomatic, 108 eclectic, 22, 48, 52, 53, 57 electronic (digital), 8, 10, 12, 13, 58, 74, 95, 97, 98, 105, 116–17, 125, 221, 234, 279, 287 facsimile, 16, 95 genetic, 72, 73 oral text, 10, 205, 221 print, 8, 10, 234, 287 scholarly, 7, 27, 35, 97–9, 104, 105, 116, 205, 215, 220, 281, 282, 283, 284, 287 synoptic, 282 textual, 8, 12 trade, 97 Eggert, Paul, 6, 7, 8, 58 Ehrmann, Bart, 25 Eichhorn, Johann Gottfried Einleitung in das Alte Testament, 33 Eikhenbaum, Boris, 75 Einstein, Albert, 35 El Cid, 120 Electronic Textual Editing (Modern Language Association), 39 Eliot, T. S., 100 The Waste Land, 56 Elizabeth I, 193 Elmer, Jonathan, 281 emendation, 17, 26, 29, 33, 43, 44, 52, 63, 72, 105, 114, 115, 216, 220, 260 vs. recension, 44 Encyclopaedia Britannica, 35, 40 Enfield, William, 192 English Association, 42 Erasmus, 29, 31, 63, 64, 231 Eratosthenes, 63 Erne, Lukas, 111 Eschenbach, Wolfram von, 66 Euripides, 19, 23 European Society for Textual Scholarship (ESTS), 61, 75 Eustathius, 26 Fallows, James, 258 “Living with a Computer,” 258 Ferrer, Daniel, 71, 72 Fichte, Johann Gottlieb, 190, 191 Fielding, Henry, 50 Fierabras, 126 Fiesoli, Giovanni, 76

Figulus, Nigidius Commentarii grammatici, 24 Fitzgerald, F. Scott, 50 Flaccus, Verrius, 28 De orthographia, 24 De verborum significatu, 24 Flaubert, Gustave, 71, 234 “Un Coeur simple”, 73 Florence, 27 Foakes, R. A., 84 Foerster, Wendelin, 122 Folger Library, 196 folio (folium), 194 Fontainebleau Library, 33 Foucault, Michel, 102, 120, 200 The Archeology of Knowledge, 120 The Order of Things, 120 Francis I, 33 French National Center for Scientific Research (CNRS), 71 French Revolution, 64 Furnivall, Frederick, 34 Gabler, Hans Walter, 6, 7, 8, 57, 113 (ed.) Ulysses, 57, 69, 75 Gale, 275 Eighteenth-Century Collections Online (ECCO), 275 Garber, Marjorie, 128 Gaskell, Philip, 107 A New Introduction to Bibliography, 46 Gautier, L´eon, 122 (ed.) Song of Roland, 121, 122, 214 Genese research group (Belgium), 74 genetic criticism, 6, 57, 69, 71, 72, 73, 84, 85, 200, 232, 234 Gibson, William Pattern Recognition, 257 Gilgamesh, 226 Giovanni, Norman Thomas di, 201 Goedeke, Karl (ed.) Schiller, 67 Goethe, Johann Wolfgang von, 67, 68, 74 Akademie-Ausgabe edition, 68 Weimar edition, 67, 68 Goodman, Nelson Languages of Art, 237 Goody, Jack, 213 Google, 275 Google Books, 73, 275, 276 graffiti, 17 Grafton, Anthony, 63

index Gramsci, Antonio, 195 Granite Mountain Record Vault, 263 Greetham, David, 3, 4, 5, 125 Textual Scholarship: An Introduction, 2 Greg, W. W., 2, 4, 35, 44–51, 54, 58, 79, 80, 81, 82, 105, 107, 233 Calculus of Variants, 46 Clark Lectures (1939), 49 Editorial Problem in Shakespeare, 47 (ed.) Marlowe’s Doctor Faustus, 1604–1616, 47 “Principles of Emendation in Shakespeare,” 44 “Rationale of Copy-Text,” 47, 50, 56 Shakespeare First Folio: its Bibliographical and Textual History, 47 Gr´esillon, Almuth, 70, 71, 72 Grigely, Joseph, 56 Grimm, Jacob and Wilhelm, 220 Groden, Michael, 72 Groth, Roy, 272 Grumach, Ernst (ed.) Goethe, 68 Guattari, Felix, 195 Gumbrecht, Hans Ulrich, 76 Gutenberg, Johannes, 17, 36, 79, 81, 230 Haraway, Donna, 130 Harkness, Bruce, 51 Harmon, Samuel and Polly “George Collins,” 206–8, 210, 211 Harry Ransom Center, 271 Havelock, Eric, 212 Hawthorne, Nathaniel, 50 The Blithedale Romance, 50 The House of the Seven Gables, 50 The Marble Faun, 50 Hay, Louis, 6, 48, 71 Hayles, N. Katherine, 277, 278, 280, 283 Writing Machines, 278 Hazlitt, William, 283 Heine, Heinrich, 6, 71, 234 Heminges, John, 88, 196 Henry, Mellinger, 206, 207, 208, 210 hermeneutics, 16, 274, 275 Hesiod, 21 Higher Criticism, 16, 274, 284, 285 Hoffmann, August Heinrich, 65 Holderlin, Friedrich, 67, 74 ¨ Frankfurt edition, 69 “Great Stuttgart Edition,” 67

305

Homer, 18, 21, 24, 27, 64, 211, 213, 214 Odyssey, 228 Honoratus, Servius, 24 Hornschuch, Hieronymus Orthotypographia, 189 Houghton Library, 33 House of Leaves, 278 Housman, A. E., 31, 36, 42 Human Genome Project, 128 Humboldt, Wilhelm von, 64 Humphrey, Duke of Gloucester, 33 A Humument, 278 Huntington Library, 33, 200 Hurlebusch, Klaus, 67 Hyginus, Julius, 24 Hymes, Dell, 218 hypertext, 39, 73, 98, 279 Innocence Project, 128 inscriptions, 17 Institut des Textes et Manuscrits Modernes (ITEM), 71, 72, 73, 75 Institute for Advanced Technology in the Humanities, 285 intention (authorial), 5, 9, 19, 20, 37, 38, 43, 45, 48, 49, 50, 52, 53, 54, 56, 57, 66, 68, 71, 76, 93, 94, 97, 100, 106, 107, 108, 111, 113, 189, 201, 228 Ireland, William Henry forgeries of Shakespeare, 192 Isidore of Seville, 26 Etymologiae, 24 Ivins, William, 246, 247, 253 Prints and Visual Communication, 236 James, Henry, 43 The Novels of Henry James (New York edition), 43 Jerome, 26, 198, 199 Vulgate, 25, 29, 229 Johnson, Samuel, 43, 79, 82 conjectural criticism, 42 (ed.) Shakespeare, 87 Jonson, Ben, 49 Joyce, James, 71, 75 Ulysses, 57, 69, 75 Joyce, Michael, 271 JSTOR, 275 Kafka, Franz, 69, 195 Kane, George, 38 (ed.) Piers Plowman, 37, 233 Kant, Immanuel, 188, 190, 191 Metaphysics of Morals, 188

306 Kastan, David, 201 Kelvin, William Thompson, 35 Kennedy, J. Gerald, 281 Kenney, E. J., 63, 64, 65 Kindle (e-reader), 39 King, Stephen, 257–8 Lisey’s Story, 261 Skeleton Crew, 257 “Word Processor of the Gods,” 257 King’s Men, 90, 93, 110, 111 Kinkead-Weekes, Mark (ed.) The Rainbow, 108 Kinsella, Thomas, 271 Kirschenbaum, Matthew, 11, 38, 39, 58, 263, 271 Kittredge, George Lyman, 206 Kliman, Bernice “The Enfolded Hamlet,” 113 Kluwer, 275 Knowles, Richard, 84 Kops, Bernard, 271 Kraus, Kari, 10 Lachmann, Karl, 5, 6, 29, 31–3, 35, 38, 64–6, 70, 74, 76, 122, 232 (ed.) De rerum natura, 31, 65, 232 (ed.) New Testament, 66 (ed.) Nibelunge Not und die Klage, 32 Laclos, Pierre Choderlos de Les liaisons dangereuses, 191 Lacnunga manuscript, 209 Wiþ Færstice, 209–11 Lambeth Library, 33 Lanson, Gustave, 70 Larson, Jonathan, 271, 272 1/2 MT House, 267 Boho Days, 267 Manhattan Tower, 267 RENT, 264–70 Superbia, 267 Laufer, Roger, 36 Laurentian Library, 33 Lawrence, D. H., 101 The Rainbow, 108 Sons and Lovers, 101 Women in Love, 100 Lee, Spike, 270 Malcolm X, 267 lemma, 115, 234, 240 Leo XIII, 31 Lernout, Geert, 4, 5, 35 Lessing, Gotthold, 66 Lewalski, Barbara, 94 Lewis, Sinclair, 50

index Lexia to Perplexia, 278, 283, 284 Leyton, Michael, 252, 253, 254 Library of Congress, 23, 34, 221, 262, 264, 267, 276 Lilly Library, 34 Lindahl, Carl, 221 Ling, Nicholas, 110 lithography, 18, 106, 224 Livy, 28 Lord, Albert, 214, 220 Louvre Library, 33 Lovati, Lovato, 27 Lower Criticism, 16, 274, 284, 285 Lucan, 26 Lucretius, 27, 28, 31, 35, 65, 232 De rerum natura, 5, 65 Lupus of Ferri`eres, 26, 228 Lycurgus, 19 Lydgate, John, 34 Maas, Paul, 29, 32, 38 Textkritik, 74 Mabillon, 29 Machan, Tim, 213, 215 Macrobius, 26 Mailer, Norman, 271 Malone, Edmond, 46, 87, 193 Malory, Thomas, 34 Manilius, 28 Astronomica, 31 Mann, Thomas, 75 Marcellinus, Ammianus, 28 Marlowe, Christopher Doctor Faustus, 49 Mary-Lafon, Jean-Bernard, 9 (trans.) Fierabras, 126 Matociis, Giovanni de, 27 Mays, J. C., 283, 287 (ed.) Collected Works of Samuel Taylor Coleridge, 282, 283, 287 McGann, Jerome, 1, 2, 3, 4, 5, 11, 13, 22, 37, 54, 57, 125, 238 Critique of Modern Textual Criticism, 37, 79, 281 McGill, Meredith, 281 McKenzie, D. F., 1, 4, 5, 23, 37, 57, 188, 274, 281, 286 Bibliography and the Sociology of Texts, 281 McKerrow, R. B., 2, 4, 44, 46, 48, 49, 51 Introduction to Bibliography for Literary Students, 46 Prolegomena for the Oxford Shakespeare, 46

index McLeod, Randall, 9, 37, 238 McNally, Terrence, 259, 271 media, 2, 233, 239, 270, 272, 278 allographic vs. autographic, 237 changes in, 12 consciousness, 56 electronic (digital), 1, 5, 6, 12, 40, 233, 234, 261, 275, 278 hardware, 272 history of, 24 print, 280 relations among oral, script, and print, 211 social, 11 storage, 11, 260, 263, 264, 265, 270, 271 time-based, 14 types of, 17, 224, 277, 280 Metzger, Bruce, 25, 31 Michel, Francisque, 122 (ed.) Song of Roland, 122 Middleton, Charles Henry catalogue of Rembrandt etchings, 241, 242, 243, 244 Milton, John Paradise Lost, 32, 42, 67 Mitchell, W. J. T., 11, 238 Montagnore, Geremia da, 27 Montaigne, Michel de, 277, 278, 279 Essais, 191 Monumenta Germaniae historica, 34, 126 Moxon, Joseph, 189 Muir, Kenneth, 94 Muller, Theodor ¨ (ed.) Song of Roland, 121, 122, 123, 214 Murray, James, 34 Nashe, Thomas, 46 National Union Catalogue, 23 Neihardt, John G., 216 Black Elk Speaks, 216–18 Networked Infrastructure of Nineteenth-century Electronic Scholarship (NINES), 284, 285 Netz, Reviel Archimedes Codex, 253, 254 New Criticism, 71, 83, 106, 214, 278 new historicism, 36, 83 New York Public Library, 33 New York Review of Books, 17 New York Times Book Review, 17, 260 New Yorker, 260 Newton, Issac, 64 Nicholas V, 33 Niles, John, 10

307

¨ ¨ Editionsfilologer, 75 Nordiskt Natverk for Nutt-Kofoth, Rudiger, 66 ¨ O’Brien, Tim, 271 Office of Digital Humanities, 254 Ong, Walter J., 211, 259 Opie, Iona and Peter, 18 Origen, 229 Hexapla, 25 Oxford Classical Texts, 42, 43 Oxford English Dictionary (OED), 34, 279 Oxford English Texts, 43 Palatine Library, 24 palimpsest, 22, 26, 253, 282 Panofsky, Erwin, 242, 244 papyrus, 17, 21, 22, 198, 224, 227 parchment, 22, 27, 194, 228 Paredes, Alonso Victor de, 188, 189 Paris, Gaston, 70, 73, 74 (ed.) Song of Roland, 122 Parrish, Stephen (ed.) Cornell Wordsworth, 56 Parry, Milman, 214, 220 Pascal, Blaise Pens´ees, 191 Pasquali, Giorgio, 74 Pavier, Thomas, 197 Pearl, 34 Peisistratus, 18, 19, 33 Percy, Thomas, 220 Reliques of Ancient English Poetry, 212 Pergamon (also Pergamum), 4, 5, 21, 22, 39, 228 Perseus Project, 279, 284 Petrarch, 27, 28, 193, 195 I Trionfi, 195 Pfeiffer, Rudolf, 63 philology, 5, 12, 28, 35, 64, 65, 66, 69, 70, 75, 76, 122, 123, 205, 212, 214, 274, 275, 278, 280, 284, 285 Picard, Raymond, 71 pictographs, 17 picture criticism, 11, 236–55 comparative, 239 definition of, 236 relation to textual criticism, 237, 238 Pidal, Ramon Men´endez (ed.) Song of Roland, 214 Pierpont Morgan Library, 33 Piers Plowman Athlone edition, 37, 233 Pindar, 21

308

index

Pitti, Daniel, 285 Pizan, Christine de, 195 Planudes, Maximus, 27 Plato, 27, 278 Phaedrus, 277 Platt, Arthur, 42 Plautus Pseudolus, 189 Pliny, 26, 27 Plutarch, 27 Poincar´e, Henri, 238 Politian, 28, 29, 31 Poliziano, Angelo, 28, 231 Pollard, A. W., 4, 44, 46 Shakespeare Folios and Quartos, 46 Poole, Thomas, 283 Pope, Alexander, 94 positivism, 19, 35, 37, 70, 73, 85, 221 postmodernism, 221 post-structuralism, 36, 54, 71, 72, 75, 76, 107 potsherds, 17 Preiss, Adrian, 253 Priscian Institutiones grammaticae, 24 Project Muse, 275 Propertius, 27 Proust, Marcel, 71, 75 Psellus, Michael, 26 Ptolemy Soter, 19 Quintilian, 28 Ars grammatica, 24 Racine, Jean, 123 Ramus, Petrus, 31 recension, 4, 8, 10, 49, 232, 233 vs. emendation, 44 Redgrave, R. G., 46 Reed, Issac (ed.) Shakespeare, 87 Reichl, Karl, 221 Rembrandt, 241, 242 Renan, Ernest, 70 Rerum britannicorum medii aevi scriptores, 34 Rerum gallicarum et francicarum scriptores, 34 Reside, Doug, 11, 38, 264, 265, 271, 272 Review of English Studies, 46 rhapsodes, 18, 19 Richards, Eliza, 281

Robins, William, 129 Rosello, Mireille, 130 Rossetti, Dante Gabriel, 284 Rossetti Archive, 279, 287 Rousseau, Jean-Jacques Dialogues ou Rousseau juge de Jean Jacques, 191 Essay on the Origin of Languages, 224 Nouvelle H´elo¨ıse, 191 Rovere, Giuliano della, 39 Rowe, Nicholas (ed.) Shakespeare, 87 Rudler, Gustave, 70 runes, 17 Rushdie, Salman, 259, 271 Russell, George (ed.) Piers Plowman, 233 Rychner, Jean, 214 Sade, Marquis de Cent vingt journ´ees de Sodome, 191 Said, Edward, 76 Orientalism, 131 Saint-Pierre, Bernadin de Paul et Virginie, 191 Salutati, Coluccio, 27, 28, 229 Sandys, J. E., 63 Sargent, Michael, 10 Sattler, D. E. (ed.) Holderlin, 69 ¨ Scaliger, Joseph Justus, 31 Scheibe, Siegfried (ed.) Goethe, 68 Schiller, Friedrich, 67, 74 scholasticism, 31 Scott, Walter, 220 scroll, 22, 194, 198, 199, 228, 272 Segre, Cesare, 123 (ed.) Song of Roland, 123, 124, 215 Seneca, 27, 29 Sernine, Daniel Chronoreg, 130 Shakespeare, William, 6, 7, 8, 36, 42, 43, 46, 49, 74, 110, 113, 114, 128, 190, 193, 194, 196, 197, 198 1 Henry IV, 91, 93 2 Henry IV, 91, 93 Arden edition, 95, 115, 116, 240 (Old) Cambridge Edition, 87 Cymbeline, 89 First Folio, 89, 195, 196, 197 forgeries of, 192 Hamlet, 8, 99, 109–17, 190

index hand of in The Book of Sir Thomas More, 190 Henry V, 20, 91, 93 King Lear, 6, 79–86, 87, 90, 93, 94, 113, 192, 197, 240 Merry Wives of Windsor, 91, 93 Norton edition, 87, 94 Oxford edition, 86–95, 112, 113 The Passionate Pilgrim, 197 Pericles, 196 print forms of, 45 The Rape of Lucrece, 196 Thomas Pavier editions of, 197 Troilus and Cressida, 89 Venus and Adonis, 197 Sharp, Cecil, 207 Shaw, David, 36 Shelley, Percy Bysshe, 35, 50 Sheridan, Richard, 50 Shillingsburg, Peter, 75 Short Title Catalogues, 23, 34 Eighteenth-Century Short Title Catalogue, 34 English Short Title Catalogue (ESTC), 34 Short Title Catalogue of Books Printed in England, Scotland, and Ireland, and of English Books printed abroad, 1475–1640, 46 Short, Ian (ed.) Song of Roland, 124, 215 Sir Gawain and the Green Knight, 34 Skinner, Steve, 265 Smith, Abby, 263 Smith, Zadie, 260 Snow, C. P., 64 Socrates, 277, 279 Song of Roland, 8, 10, 120–4, 214–16, 233 Sophocles, 19, 23, 43 Spinoza, Baruch, 64 Stallybrass, Peter, 9 Statius, 26 Steevens, George, 87 stemma, 5, 29, 65, 80, 81, 130, 214, 232, 233, 239 stemmatic method, 80, 122, 123, 124, 254 Stengel, Edmund, 122 (ed.) Song of Roland, 121, 122, 214 Stillinger, Jack, 56 Street, Brian V., 213 structuralism, 36, 71, 72, 75 Studies in Bibliography, 47 Stunica, 30

309

Sukenick, Ronald, 271 Sutherland, Kathryn, 4, 5, 35, 61 Symmachus, 229 Tanselle, G. Thomas, 2, 4, 21, 36, 37, 39, 45, 50, 54, 56, 233 Rationale of Textual Criticism, 53 Tarrant, R. J., 38 Tate, Nahum adaptation of King Lear, 197 Taylor, Gary, 115 (ed.) Division of the Kingdoms, 7, 84, 85 (ed.) Oxford Shakespeare, 6, 87, 112 Taylor, Neil (ed.) Hamlet, 115 Tedlock, Dennis, 218 Terence, 32 TEXT (journal), 79 Text Encoding Initiative (TEI), 13 textual criticism, 6, 11, 16, 18, 20, 22, 31, 35, 36, 37, 39, 42, 44, 46, 47, 48, 51, 54, 57, 58, 63, 65, 66, 79–85, 87, 90, 91, 93, 94, 97, 117, 205, 214, 215, 230, 232, 233, 234, 236, 237, 238, 255, 276, 284 textus, 18, 20, 25, 30, 66, 229, 234 Theobald, Lewis, 20 Theodotion, 229 Thompson, Ann (ed.) Hamlet, 115 Thompson, D’arcy, 248, 250, 252 Thorpe, James, 56, 107 Principles of Textual Criticism, 48 Tibullus, 27 Tillotson, Kathleen, 52 (ed.) Oliver Twist, 52, 56 Times Literary Supplement, 17, 42 Timpanaro, Sebastiano, 64, 65, 74 Toelken, Barre (trans.) “The Grizzly Bear,” 218, 219 Tomashevskii, Boris, 75 Treisman, Deborah, 260 Trevisa, John On the Properties of Things, 38 Triclinus, Demetrius, 27 Tronch-P´erez, Jesus, ´ 115 (ed.) A Synoptic Hamlet, 113, 114, 115, 116 Trotsky, Leon, 281 Tzetzes, John, 26 Unsworth, John, 239 Uris, Leon, 271 Urkowitz, Steven, 83

310

index

Valla, Lorenzo, 28, 29, 31, 63, 229 Adnotationes in Novum Testamentum, 29 analysis of Greek New Testament, 29 Van Hulle, Dirk, 74, 75 Vanhoutte, Edward, 74 variant(s), 5, 10, 17, 18, 21, 38, 46, 48, 50, 52, 62, 63, 67, 74, 83, 91, 92, 97, 101, 102, 103, 105, 107, 108, 111, 114, 115, 116, 205, 206, 208, 212, 213, 214, 215, 228, 232, 236, 237–41, 243, 244, 245, 247, 259 Varro De lingua latina, 24 Vatican Library, 33 Verrall, A. W., 42 Vickers, Douglas, 253 Vinaver, Eug`ene, 22 Virgil, 24, 26, 193 Aeneid, 193 Vogelweide, Walther von der, 66 Walker, Alice, 271 Wallace, David Foster, 262 Warren, Michael (ed.) Division of the Kingdoms, 7, 84 Warren, Michelle, 6, 8, 9, 12 Weitzmann, Kurt, 236 Wellek, Ren´e Theory of Literature, 275 Wells, Stanley (ed.) Oxford Shakespeare, 6, 87 Werstine, Paul, 112

Wesker, Arnold, 271 Western Historical Manuscript Collection, 217 Whalen, Terence, 281 White, Christopher catalogue of Rembrandt etchings, 242, 243 White, Hayden Metahistory, 128 Whitman, Walt, 50, 260 “Shakspere-Bacon’s Cipher”, 200 Leaves of Grass, 200 Whitman Archive, 200 Widener Library, 34 Williams, William Proctor An Introduction to Bibliographical and Textual Studies, 2 Wolf, Friedrich August, 33 Prolegomena ad Homerum, 33 Wordsworth, William Cornell edition, 56, 104 Prelude, 25 WorldCat, 23 Yeats, William Butler, 56 Cornell edition, 104 Zenodotus of Ephesus, 19, 21 Zola, Emile, 71 Le Rˆeve, 73 Zumpt, Carl (ed.) Cicero, 232 Zumthor, Paul, 10, 209, 215, 234 Toward a Medieval Poetics, 233