Anda di halaman 1dari 73

0

REATIVE COMPOSITION IN WORDSWORTH,


TENNYSON, AND DICKINSON

SUPy

University of Virginia Press


Charlottesville and London

Umveraty of Virgin^^ University of VitginlJ


'M'bytl..Reo,oradV

Printed in the Urti.ed States of America on ac.d-fr.e paper


Pirst

published 2009

135798642

Library of congress Ca..lo^ng-in-Public.tiDat,


Text as process: creative

siUon'in Wordsworth, Tennyson, and Dickinson /


p. cm.

includes bibliographical refaens ^^^^^


1. Wordsworth,
^ ^Criticism, Textual. 3. Dickinson, Emily,
on,1809-18y^-^""
,
Tennyson. B.'
'809-^
1830-1886-Criticism. Textual. I. Title.
PR5893.B87 2009
821'.7dc22
2008033600

, j,wines from the inside back cover of Trinity


Frontispiece: A page of Tennyson
ermission of the Master and Fellows
Notebook 15 (0.15.15). (Reproduced with
of Trinity College. Cambridge)

"I hardly know anymore who and where I am."


"None of us knows that, as soon as we stop fooling ourselves."
Martin Heidegger, "Conversation on a Country Path"

o.

List of Illustrations

viii

Acknowledgments
Introduction

ix
1

L Contextualizing Process: Three Perspectives on Genetic Criticism


2. Theorizing Process: Origins, Agency, Intention

38

3. Reclaiming Process: Toward a Compositional Method


4. Wordsworth's Process

57

75

5. Tennyson's Process

117

6. Dickinson's Process

168

7. A Philosophy of Composition: The Coming-into-Being


of the Literary Work 215
Notes

239

Bibliography
Index

269

287

PJlLtc^c

0>lyU>

1. Wordsworth: early Prelude draft

94

2. Wordsworth: early Prelude draft

95

3. Wordsworth: draft for the end of The Excursion

113

4. Wordsworth: draft for the end of The Excursion

114

5. Wordsworth: ink scribble on Ruined Cottage draft

127

6. Tennyson: first-draft sequence for "Morte d'Arthur"

129-31

7. A spatial map of Tennyson's The Holy Grail notebook


8. A compositional map for Tennyson's Idylls of the King
9. Collected editions of Tennyson's

142-43
148

10. Tennyson: draft sequence for "Enid's Song"


11. Dickinson: "Fortitude incarnate"

137

161-63

185

12. Dickinson: "As Frost is best conceived"


13. Dickinson: "Those fairfictitious People"
14. Dickinson: "Crisis is sweet and yet the Heart"

189
192-93
200-201

15. Dickinson: "Your thoughts dont have words every day"


16. Dickinson: "A little madness in the Spring"

209

17. Wordsworth: making the pen work in Ruined Cottage draft


18. Dickinson: "The Sea said 'Come' to the Brook"

viii

205

234

222

I want to begin by thanking the various institutions that have supported


this project in substantial ways. Manuscript research at libraries in the
United Kingdom and the United States was funded by an AHRC Innova
tion Award in 2003-4 for which I am extremely grateful. Lancaster Uni
versity also kindly granted me a sabbatical term at the very start and end
of the project. While undertaking manuscript work, I was based at the
Wbrdsworth Trust for three months and wish to thank Jeff Cowton, as cu
rator, and the late Robert Woof, as director, for their professional assistance
but also for the warmth of their welcome into the community. I also thank
Pamela Woof, Ann Lambert, and others at the Trust for their encourage
ment and support. Work on Tennyson's manuscripts took place at Trinity
College, Cambridge, and the Houghton Library, Harvard. The librarians at
Trinity were extremely helpful, and Clare Hall, Cambridge, allowed me to
stay as a visiting scholar and be part of a supportive postgraduate commu
nity there. I would like to thank the staff of the Houghton Library for their
helpfulness, particularly Betty Falsey, Tom Ford, and Susan Halpert. My
thanks also to the curator of manuscripts, Leslie Morris. The Department
of English and American Literature at Harvard University also kindly per
mitted affiliation, and the British Literature Colloquium group there al
lowed me to benefit from their seminar. Finally, my thanks to Amherst
College Library (Special Collections) and to Daria D'Arienzo, curator, as
well as to Tevis Kimball at the Jones Library, Amherst. During my research
abroad, various people made me welcome, in particular James Engell, Leah
Price, and the graduate students at Harvard. I also want to thank Isobel
Armstrong for her company and conversation along the Charles River,

ix

ACKNOWLEDGMENTS

ACKNOWLEDGMENTS

which was a gift I did not expect, and Christopher Ricks for two Hvely and

All images from Emily Dickinson's manuscripts are reproduced by


permission of The Houghton Library, Harvard University. The Presi
dent and Fellows of Harvard College.

constructive conversations.
I am grateful to GilUan Beer and Susan Manning for their encourage
ment at the very earliest stages of the project, when it was most needed,
as well as for constant support throughout my academic career. Others
who are always intellectually generous and deserving of my thanks and
who have contributed here include James Butler, Stephen Gill, and Stephen
Parrish. I am also indebted to James McLaverty and Paul Eggert for their
kindness and willingness to respond to core chapters of the book and to
Peter Shillingsburg for involving me in his annual Symposium on Textual
Studies, which has been extremely stimulating. At Lancaster I have been
fortunate in the support of colleaguesparticularly Simon Bainbridge,
Arthur Bradley, Jo Carruthers, Cathy Clay, Keith Hanley, Tony Pinkney,
and Michael Sandersas well as that of the Reading Group, Tracy Mansell,
Caroline Rose, and Brighid Webster. My thanks to Sheila Fyfe for check
ing my French and German and to Charlotte Avery, Ben Quash, Bart Van
Es, and Wei-Wei Yeo-Lee for being constant friends, critics, and readers.
Finally, I wish to thank my brother, my sister Gill, and my parents for their
unwavering faith in me, and lastbut never leastJohn Hilliard.

Texts of poems are reprinted and transcribed by permission of the pub


lishers and Trustees of Amherst College from The Poems of Emily Dickin
son: Variorum Edition, Ralph W. Franklin, ed., Cambridge, Mass.: Har
vard University Press, Belknap Press, Copyright 1998 by the President
and Fellows of Harvard College. Copyright 1951,1955,1979,1983 by the
President and Fellows of Harvard College.

Some of the material in the book has been previously published elsewhere.
The section on "Denial of Origins' in chapter 2 appears in an extended
form in Textual Cultures (FaU 2007): 100-117. Chapter 3 appeared in an ear
lier form (here significantly revised) in TEXT 17 (2005): 55-91; a much fuller
discussion of unintentional meaning than the one given here in chapter 3
occurs in an article for The Emily Dickinson Journal 14 (2005): 24-61. Mate
rial on Wordsworthian microanalysis in chapter 4 previously appeared in
Studies in Romanticism 44 (Fall 2005): 399-421. Thanks to Wayne Storey at
TEXT and Textual Cultures and David Wagenknecht at Studies in Romanti
cism for allowing reproduction of previously published material.
All images from Wordsworth's manuscripts are reproduced by permis
sion of the Wordsworth Trust, Dove Cottage, Grasmere.
All images from Tennyson's manuscripts held at Trinity College, Cam
bridge, are reproduced with the kind permission of the Master and Fellows
of Trinity College, Cambridge.
All images from Tennyson's manuscripts held at The Houghton Li
brary, Harvard, are reproduced by permission of The Houghton Library,
Harvard University The Presidents and Fellows of Harvard College.
The Tennyson transcription from MS Ashley 2104 pages 24-25 is made
with the kind permission of The British Library. The British Library. All
Rights Reserved.

xi

i
o

This book is about the Hterary text before it becomes a completed work of
art. It takes as its focus the prepublication work, the draft materials, that
constitute text as process. For too long, textual process has been treated as
the province of scholarly editors or drawn on in a largely secondary way by
critics delving below the surface of the published text. The time has come
for a reevaluation of this alternative aspect of the literary work. This study
presents textual process as something not only of interest to editors and
textual theorists but as material worthy of philosophical definition and'a
full critical response.
The book has two primary objectives. First, it seeks to understand the
nature of the text in a state of process and its status relative to the completed
work. Second, it aims to develop a critical method and a hermeneutics for
interpreting the text in this state. Put simply: we need to understand what
text as process is, and then think about what to do with it.
In his essay "Genetic Texts," the German textual critic Hans Walter
Gabler declares, "Fundamentally the issue is whether, critically, the pro
cess of the writing is, or is not, integral to the product of the writing."' Such
an issue is of central importance to this book. Although it is focused on the
draft materials of a work, this study is ultimately concerned with enlarging
our definition of "text" to understand text as process alongside the com
pleted text. It does not seek to define textual process as something entirely
apart from the text in a completed state but as part of the continuum of
the text, for which the concept of a single, stable state is in part an illusion.
Necessarily, however, understanding process demands a recognition of its
difference from the final product as well as its vital connection to it. It must
be allowed to be both a part and apart.

TEXT AS PROCESS

INTRODUCTION

This book presents a methodology for the study of draft materials. In


terms of future use it aims to provide a way of responding to textual pro
cess that can be widely used, disputed, and enlarged, forming the basis of a
new subdiscipline. However, since it is the first study of this kind, and since
such a methodology does not yet exist, it also provides a more fundamental
definition of what scholars understand textual process to be. It examines
the status of this material in philosophical terms, as well as explaining
why a methodology for interpreting it has not existed until now (in AngloAmerican studies at least), and it presents a new approach centered on re
claiming intention as a complex shifting state within process. The book
seeks to articulate a self-reflective critical practice that thinks about what

critical divide, structurally and conceptually; his last three books have all
been concerned with the concept of "versions" of a text and the value of
"multiplicity." At the heart of his theory is a concept of "textual instability"
defined in opposition to a model of "textual stability" involving a single,
authoritative text: "Textual instability, in a similarly simple view, is just
the opposite: the absence (or lack) of a single correct or best or most au
thoritative text."' Stillinger takes this further in relation to Coleridge with
a definition of "textual pluralism" in which "each version of a work em
bodies a separate authorial intention that is not necessarily the same as
the authorial intention in any other version of the same work.""* His book
thus argues for independent multiple versions of Coleridge's poems as the
most appropriate way of presenting (and responding to) works that are the
product of obsessive revision and rewriting.
Stillinger provides a useful and important model for the kind of work
I will be doing here. He convincingly unites literary and textual criticism,
and his work is immensely valuable in establishing such approaches as ac
ceptable. On the whole, though, for Stillinger, the key issue is his concern
with the valuingor democratizingof different draft versions of a text
rather than with allowing the debate to shift to critical use of editorial ef
fort. Of course Stillinger is deeply interested in composition, in process
rather than product, and he also has some interesting discussion of issues
of readership and of multiple intention. In relation to Coleridge, for exam
ple, he wonders "whether he had any intentions at all in the conventional
sense." However, Stillinger's primary interest is less in textual process it
self than in the appropriate presentation of compositional material.
A second Anglo-American study that connects text-critical and
literary-critical thinking is Hershel Parker's Flawed Texts and Verbal Icons.
Parker's "textual-aesthetic approach" seeks to study the text as process
through the use of biographical and textual evidence and an "aesthetic"
judgment about what constitutes the best version of a text. As its title sug
gests, the book argues against New Critical principles and the situating of
meaning intrinsically within a literary text.
Parker almost always focuses on the text immediately before publica
tion or in its development around the act of publication, with authorial
intentionality "built into the words of a literary work during the process
of composition."' This "original intention" establishes an ideal authorial
version of the text that is then in danger of being lost by the author's own
subsequent actions: "All valid meaning is authorial meaning, but in stan
dard literary texts authorial meaning may be mixed in with non-sense,
skewed meanings, and wholly adventitious meanings which result from
tampering with the text, by the author or someone else." He continually

is being undertaken as it undertakes it.


The work is timely. The approaches it suggests could only really have
come about at the end of the twentieth and the start of the twenty-first cen
turies. Only from the 1970s onward has Anglo-American editing advanced
to a point that allows for the editing of literary works not merely in terms of
a final single-state text but in multiple versions and with full reproduction
of draft materials. The kind of critical work this book advocates can only
begin to be widely practiced now, when the draft materials of many major
authors are more accessible than ever before (in facsimile and scholarly edi
tions as well as high-quality web archives).^ It remains the case, however,
that although access to such material has significantly increased, there has
been, as yet, no attempt to respond to it in a systematic way within the
Anglo-American scholarly community. Such books as have been written
in this area tend to be viewed as offshoots from textual criticism and tend
to focus on the writings of a specific author rather than offering larger ex
plorations of the nature and interpretation of the materials.
There are certain practitioners and theorists nevertheless who need to
be acknowledged as influential for this work, primarily those textual crit
ics with a heightened awareness of their relation to literary criticism. These
include Jack Stillinger's Coleridge and Textual Instability and Multiple Au
thorship and the Myth of Solitary Genius and Hershel Parker's Flawed Texts
and Verbal Icons. More editorially centered studies that have also proved
valuable include Philip Gaskell's From Writer to Reader: Studies in Edito
rial Method (with its useful range of editorial/authorial examples); Donald
Reiman's The Study of Modern Manuscripts; John Bryant's The Fluid Text;
D. C. Greetham's excellent Theories of the Text; and Philip Cohen's edited
collection Devils and Angels: Textual Editing and Literary Theory.
Jack Stillinger and Hershel Parker require further consideration here
as the two Anglo-American scholars whose work seems most to anticipate
this study. Stillinger's work consciously straddles the text-critical/literary-

TEXT AS PROCESS

INTRODUCTION

returns to the question of "when intentionahty is built into a passage or a


work as a whole," which is defined by him as being "at the moment of com
position, if at all."' In effect, then, Parker still seeks to establish a hierarchi
cally definitive state for a text (even if this is a state within process), which
creates the major problem of how such a state can be established. (When
is "original intention" fulfilled? What marks the "moment of completion"
within composition after which a text can be "damaged" by the author?)
Parker s own method in Flawed Texts (which is never explicitly defined)
can be seen most clearly in his fifth chapter, on Mark Twain, in which he
explores biographical information about Twain himself (the author's false
presentation of the text's composition), biographical information about the
text under consideration, accounts of different textual versions and textual
history, the sequence of composition, the order of manuscripts, and critical
discussion of compositional and publication decisions.
However, Parker's aesthetic bias leads him to treat the author in a re
markably judgmental way. He responds to the creative process, and to au
thors themselves, in value-laden terms: "If writers fail to achieve their full
intentions during composition, they are even more likely to damage parts
of what they had achieved when they belatedly alter a text."'" It seems to me
that such an approach fails to understand the complex nature of intention,
even as it raises intelligent questions about it. So, on the one hand, Parker
makes helpful statements such as: "[AJuthorial intentionality is built into
the words of a literary work during the process of composition, not be
fore and not afterwards," and "Criticism really is dependent upon textual
knowledge ... evidence in drafts or the earliest complete version of a work
may be of interest equal to that of the first (or any later) printed edition.""
On the other hand, however, he can be extremely judgmental of both au
thors and critics. As Hans Walter Gabler points out: "Parker appears thus
at bottom not to have shed his own allegiance to the modes of interpreta
tion and evaluation of New Criticism. He recognizes, it is true, writing
as process. Yet the acts of revision which the process involves tend to fall
victim as 'flawed texts,' to his evaluative grasp because he has so strongly
privileged the original acts."'^
A more recent work that values textual process in similar ways to this
study is John Bryant's The Fluid Text (2002). Bryant argues that all texts
are "fluid," existing as different versions that flow into each other. He gives
a s h i s g o a l a d e s i r e t o "challenge o u r t e n d e n c y t o d e f i n e a m a t e r i a l t e x t . . .
as a fixed thing, and to suggest new ways of reading, interpreting and
teaching."" In his fifth chapter Bryant defines eight different characteris
tics of the "version" and relates this to three modes of textual production
(creation, publication, and adaptation). His approach shares with my own

a concern to respond to process as process and to recognize the value of


shifting intentions in understanding this state of a text. However, his work
is .primarily focused on the implications of such a position for the editing
and teaching of texts (rather than for critical analysis). He also situates
himself largely within the Anglo-American debate: French and German
practices are defined together as genetic criticism and treated with a strong
German bias. My own interest in the French critique genetique marks a
significant distinction between this study and that of Bryant.
Finally, the highly influential work of Jerome McGann also clearly in
forms this book at a larger level, above all in the belief that text-critical and
literary-critical practice can, and should, be vitally connected through a
"symbiosis of editorial and interpretive work."'^ McGann's success in cri
tiquing the dominant intentional mode of textual criticism and in shifting the editing of texts away from an authorially centered perspective and
toward a social one is, of course, well known. Where my work differs from
McGann's, however, is in seeking to allow greater space for contextual and
compositional knowledge (often "authorial" in nature). This book is con
cerned with trying to determine the nature of meaning and interpretation
for text as process in a way clearly analogous to McGann's textual concerns,
but it focuses on textual meaning at the earlier stages of the writing process,
when the role of the author as individual is inevitably more pronounced.
The work is therefore centered on a dialogue between linguistic/semantic
and intentional or unintentional meaning on the page rather than a dia
logue between linguistic/semantic and social meaning. In part, this is an
inevitable consequence of focusing on the stages of a text's emergence and
development (as I am doing) rather than on the later stages of production
and transmission (as McGann tends to do). The fact that McGann draws
on manuscript materials primarily to explore the context of textual pro
duction can overshadow other contextual ways of reading such material,
in particular the context of the poem's making.In sum, where McGann's
work might be crudely reduced to the statement that "transmission is a
part of meaning," my own could be summarized as "the making is a part
of meaning." Ultimately, however, these two positions are not incompatible
but complementary.

At least as significant as the Anglo-American context, for this study, is


the work of French textual "geneticists." The critique gitietique presents a
considerable body of scholarship and creative thinking about the materials
of composition that lie closer to my own study than any work by AngloAhierican critics. Studies such as Almuth Gresillon's Elements de Critique
Genetique or Pierre-Marc de Biasi's La Genetique des Textes address many
of the issues I consider here, in detail, with an impressive knowledge of a
5

TEXT AS PROCESS

number of major French writers of the nineteenth and twentieth centu


ries. The advantages of French retention of national authors in major col
lections in the Bibliotheque Nationale, as well as the support for focused
research on such collections, are clear. Chapter 1 of this book will engage
fully with German and French advances, and the method of chapter 3 is
strongly informed by the critique genetique. At the same time, however,
this study is not intended to be a mere conduit for French thinking. Rather,
it attempts to articulate a distinctive Anglo-American model. This is in
part a recognition that a genetic, or compositional, criticism is strongly
defined by national traditions of editing and the nature of major writers
within a particular cultural heritage. Although there is much to learn from
other nations, the field must develop its own identity from within.'
Finally, if the study of textual process is to emerge with any kind of
credible status, then it also needs to engage with the theoretical (postHeideggerian, phenomenological) denial of individual creative origins and
of creative agency that has been highly influential within literary studies
through its indirect manifestation in late twentieth-century French think
ing. Where the area of study under consideration is that of creative compo
sition, it seems to me essential to understand it both in terms of will, inten
tion, hnearity, teleology"inauthentic" structures of being (in Heidegger's
terms)flnti in its relation to potential "authenticity." It is my belief that,
even though the aesthetic of author as creator and an expressive model of
literary creativity has been replaced by a fragmentary conception of the
work as multiple texts in multiple dialogues, a full understanding of com
positional material nonetheless demands a response to that material that
creates space for both conceptualizations. Put simply, I would argue that
the core intentional structures of the creative mind (the "author"), even if
a kind of delusion, are a necessary delusion for creative process and one
worthy of study This argument is made in chapter 2 and again, more fully,
in the final chapter.
What does this book do that no previous studies have done? Three
claims can, I think, be made for it. First, it attempts to provide a universal
methodology with larger application than for the three poets used as case
studies here; it aims to provide a firm basis for an Anglo-American "ge
netic" criticism underpinned by a philosophical account of the nature of
process. The book does not try to argue for a rigid methodological struc
ture but for a flexible framework that can be adapted and enlarged in the
light of authorial quirks or generic differences (for example, the writing of
a novel or play as opposed to poetrypoetic composition over time being
the generic focus here). This is a potentially controversial ambition. One
purpose of the framework for the study of process is that it allows for com-

INTRODUCTION

parison across and between writers. There may be those who would ques
tion whether such a framework is necessary and would want to assert that
all writers are unique. However, if the study of process is to develop beyond
the province of editorial specialists working on individual authors, then
enabling such comparison seems to me to be essential. It is only when we
begin to consider what constitutes a first-draft state, say, rather than what
constitutes a Tennysonian, or Joycean, or Wordsworthian first draft, that
textual process itself becomes the central focus of investigation.
Second, this study compares European and Anglo-American editorial
principles and draws on French genetic criticism as an underpinning for
Anglo-American studies in a way that has not previously been undertaken.
It seeks to make an Anglo-American readership far more aware of the use
fulness of the critique genetique but ultimately draws on European prac
tices to develop an approach emerging from its own tradition.
Finally, a third claim for the book's originality lies in its self-conscious
hermeneutical practice when analyzing draft materials. In relation to the
three nineteenth-century case studies," textual analysis moves backward
^^d forward, across draft text and published text in a wide range of ways,
bringing together different kinds of meaning to offer a new form of in
terpretation. Such hermeneutic practice is rarely undertaken on the textcritical side of studies of process, where critical concern is usually with
the relative status of texts rather than with their interpretation. On the
literary-critical side, while there are some excellent individual readings of
literary texts that draw on underlying draft materials, (particularly for au
thors for whom there is a strongly established interest in the drafts), this
is rarely undertaken in a consciously methodological way or as part of a
larger field of study.
I stated earlier that the core concern of this book is to understand what
text as process is and then to decide what should be done with it. This
concern underpins the overall structure and organization of the study.
The first three chapters introduce, contextualize, and develop the com
positional method presented in chapter 3, beginning with a comparative
overview of Anglo-American, German, and French editing principles and
attitudes toward genesis. The next three chapters apply that method in
different ways to the poetic process of Wordsworth, Tennyson, and Emily
Dickinson. The final chapter concludes with a full philosophical account
of the nature of textual process, of the making of meaning and the kind of
interpretation it requires. The book thus seeks to illustrate the usefulness
of its own theoretical conclusions by direct application of such ideas to
draft materials and further reflection on them.
Ultimately, this work intends to move "genetic criticism" forward in

6
7

TEXT AS PROCESS

Anglo-American scholarship by bringing it to the attention of a much


larger audience. It is to be hoped that such analysis will begin to be viewed
less as a rarefied practice and more as an intrinsic part of literary studies.
In the end, the book aims to redefine and enlarge literary-critical under
standing of the object of its investigation. Such recognition should lead to
a further reintegration of text-critical and literary-critical responses to the
literary work of art as well as an enlargement of interpretative practices to
include the special demands made by the materiality of draft materials and
the nature of text in a state of process.
This book addresses itself to multiple audiences. It will be working
within familiar territory for textual scholars and theorists, who should be
its most immediately receptive audience. It is also to be hoped, however,
that the book will reach out to a larger readership by means of the indi
vidual authors used as case studies here and engage those working in the
field of nineteenth-century literature who see before them a rich domain
of surviving manuscript drafts but are not always sure how to journey
through it.

A.oc.e^^

THREE PERSPECTIVES ON GENETIC CRITICISM

(M]any a literary critic has investigated the part ownership and


mechanical condition of his second-hand automobile, or the pedigree
and training of his dog, more thoroughly than he has looked into the
qualifications of the text on which his critical theories rest.

Fredson Bowers, Textual Criticism and Literary Criticism

This chapter sets out to explain the importance of recent advances in tex
tual criticism and theories of editing for the emergence of a compositional
method and a study of text as process that is dependent on the way in
which such materials are presented. Without the shift in attitude of the last
twenty-five years, away from the privileging of final authorial intention
and in favor of multiple textual versions, this book could not exist. It is
important, therefore, to understand the history and nature of accessibility
to draft material alongside the development of any method for interpreting
that material.
This chapter gives three accounts of the development of textual criti
cism and its relation to "genetic criticism" in terms of Anglo-American,
German, and French practices to make clear the text-critical foundations
that underlie this study and the timeliness of it as a way of thinking about
process that could emerge only once editing principles allowed for the
valuing of such material. It also seeks to illuminate, by comparison of three
traditions, the development of editorial principles out of a particular his
tory and culture in a way that will allow, in the ensuing chapters, for a look
back at the Anglo-American tradition to redefine the way ahead.

.
^
>

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

ANGLO-AMERICAN TEXTUAL CRITICISM:

sions in terms of accidentals because successive editions become increas


ingly divergent from the original (or lost original). Where there is more
than one substantive text, then one must acknowledge that an alternative
choice was possible. Hershel Parker gives a good clarification of the differ
ence between McKerrow and Greg and the significance of Greg's distinc
tion: "Greg had taken a crucial half step beyond Ronald B. McKerrow, who
had realized that the first edition should be copy-text but had assumed that
once you identify an author's hand in a later text you should adopt all of the
verbal variants in that text except the ones which are manifestly erroneous.
Greg by contrast allowed for compositorial error and casual alterations:
the editor would choose to adopt only those variants which seemed clearly
to be authorial."

"BEST TEXT" TO MULTIPLE TEXTS


Key positions in the development of Anglo-American editorial principles
and textual criticism, before Jerome McGann's work, have been articulated
elsewhere in considerable detail.' My aim here, therefore, is to give only a
brief overview of early positions before moving to a consideration of more
recent developments in the field that bear directly on textual process.
Modern textual criticism in English studies finds its origins in bibli
cal scholarship, from which two key stages for textual criticism originally
emerged: "recension," the relationship between manuscripts in terms of
a "stemma" used to reconstruct a text to a state as close as possible to the
lost original: and "emendation," applying understanding of the problems
involved in the processes of transmission to try and ehminate corruptions
from the text.^ Drawing on this structure in the nineteenth century, Karl
Lachmann developed a genealogical method for the editing of classical texts
with the aim of determining the most authoritative text through a line of
textual descent. The purpose of such a method was purification: "to 'clear
the text' of its corruptions and, thereby, to produce (or approximate)
by subtraction, as it werethe lost original document, the 'authoritative

editions in which those variants appeared."'*


The resulting rigidity of McKerrow's approach was softened by W. W.
Greg in his famous article "The Rationale of Copy-Text" (1950-51). Greg
made an argument designed to give editors more room for some degree
of individual judgment within a controlled framework, opposing what he
described as "the tyranny of the copy-text," which encouraged editors to
place too much weight on a single authority. As is well known, Greg distin
guished between two critical acts: the necessity of adherence to copy-text
for the accidentals (spelling, punctuation, and so on) and the text-critical
decision to choose between substantive (content-based) readings. Accord
ing to Greg, the copy-text should be the earliest in a series of textual ver-

In the 1960s and 1970s Fredson Bowers took up Greg's principles and
adapted them further in relation to nineteenth-century texts. However,
whereas Greg (like Lachmann and McKerrow) was dealing with lost origi
nals and derived texts, many of the nineteenth-century manuscripts Bow
ers was working on had survived, so that at this point, the debate over
what the copy-text should be began to shift more firmly to grounds of
intentionality and questions of authority: issues that arise from the pres
ence of manuscript material. In "Some Principles for Scholarly Editions of
Nineteenth-Century American Authors," Bowers defines Greg's distinc
tion between accidentals and substantives as being between two kinds of
authority, with the authority of the word (content) divided from that of
the forms (spelling, punctuation, and so on). He differed from Greg in as
serting that the final manuscript version of a text should generally be the
copy-text because this was closest to final authorial intentions: "An au
thoritative edition is one set directly from manuscript, or a later edition
that contains corrections or revisions that proceeded from the author."'
Bowers also codified his principles into rules of practice and articulated a
kind of scholarly apparatus that was then adopted by the Center for Edi
tions of American Authors (CEAA, founded in 1963). The Center defined
a series of principles and standards for critical editing in strongly intentionist terms in the Statement of Editorial Principles and Procedures (1967,
rev. 1972). This document gives a clear account of the Center's main aim;
"[T]o make the works of important American writers available, and avail
able in texts which reflect the authors' intentions as fully as surviving evi
dence permits."' Speaking of the editor's role, it states: "The editor aims to
establish, as far as surviving evidence permits, a text which presents, both
in its accidentals and its substantives, the author's intention."' The kind
of edition that emerges from such principles is described by G. Thomas
Tanselle: "What is now referred to as a 'CEAA edition,' then, is the specific

10

11

text.'"'
In the early twentieth century, R. B. McKerrow adapted the Lachmann
method to the editing of Elizabethan texts, first using the term "copy-text"
in 1904 to make clear the importance of deciding on the most authori
tative manuscript (the "best" text) by careful study. He established quite
strict principles for deciding on a single authoritative text, reacting against
the unreliable judgments otherwise involved in creating eclectic texts:
"McKerrow's attitude doubtless sprang from his overreacting against the
abuses of some nineteenth-century editors, who felt free to choose among
variant readings without adequate study of the nature and origin of the

TEXT AS PROCESS

combination of two elementsa text edited according to Greg s theory,


combined with an apparatus providing the essential evidence for examin
ing the editor's decisions."'" From the mid-twentieth century onward, and
at the heart of Bowers's principles and those of the CEAA, lies the concept
of final intentions.
The question of what exactly constitutes final intentions, in editorial
terms, has been debated widely, and with considerable intelligence, by crit
ics such as Tanselle, Morse Peckham, E. D. Hirsch, Philip Gaskell, and
James Thorpe, among others." Gaskell opposed Bowers's principle of favor
ing the final manuscript version over the first published text, questioning
how one can know that final intentions are embodied in the manuscript:
"At first glance it might seem that the manuscript will be the obvious choice
for copy-text, for it is what the author actually wrote. But does it represent
the text as the author wanted it to be read?"'^ In From Writer to Reader,
Gaskell further distinguished between authorial intention and expectation
and placed an emphasis on readers' needs and the audience. The text pre
sented by the editor should be "as authoritative as the evidence allows ...
in the form best suited to his intended audience."" James Thorpe has also
argued in favor of including the process of publication as part of the move
toward final intention (and thus as a factor to be taken into account when
choosing copy-text): "In many cases the author expected that his inten
tions would be completed by the agency of editor or printer in the matter of
accidentals.... It is clear that a reversion to the authorial manuscript would,
in such cases, actually thwart the author's intentions."'" These two critics
thus directly anticipate Jerome McGann in viewing the social aspects of
the text's production as an element constituting authority, but they are still
bound to a theory that has a model of intention at its heart.
In the context of the Anglo-American tradition, McGann's social
theory for textual criticism represents a major shift from the principles
established before his participation in the debate. He defines the limita
tions of the author-centric intentionist approach and counters this with an
alternative model in which authority is located not with the writer of the
text alone but in the social forces and communal activity that bring the
text into being. McGann very clearly works through the positions of his
predecessors in A Critique of Modern Textual Criticism to pull together all
previous approaches and take a significant step forward into territory that
demands not only a new approach to editing but a different response to
the texts being edited, and a widening definition of what textual criticism
is. McGann thus represents a major challenge to the previously dominant
view of the author's final intentions as the ultimate model of authority. For

12

CONTEXTUALIZING PROCESS

him, this concept of artistic autonomy is fundamentally flawed, since au


thority is "a social nexus, not a personal possession."" McGann argues that
the text-critical debate should not be so strongly focused on the location
of final authorial intention but that authority can equally well be located
in the act of publication. Toward the very end of the Critique, McGann
reminds us of "Greg's shrewd comment: 'Authority is never absolute, but
only relative.'"' The editor must choose his text according to particular
needs and not attempt to create a single ideal text.
Peter Shillingsburg somewhat critically assesses the nature of Mc
Gann's intervention: "Because the net defined by 'authorial' authority
eliminated readings essential to the text he wanted to produce, McGann
redefined authority, creating a net that would hold alterations introduced
by the production crews. . . . This expedient served his needs and those
of other like-minded editors so well that his definition of authority has
quickly gained the status of an alternative to the Greg/Bowers editorial
school."" Shillingsburg suggests that McGann replaced one model of au
thority with another more suited to the kind of texts he was working with
(particularly in the case of Byron) and to his own ideological sympathies.
However, the full significance of McGann's challenge is not simply that it
replaces one position with another but that it allows the possibility of more
than one way of approaching a text and of more than one authoritative
version of a text existing, according to the nature of that approach. What
this also implies is an end to a view of editing itself as a practice in which
the editor attempts to adopt a position of objectivity or uhimate judgment.
Potentially, it allows the editor more space for individual expression and
' for a closer integration of literary-critical thought with text-critical acts.
Editorial openness to "versions" of a text clearly affects the value accorded
to compositional material. The tendency to view a work as more than a
single entity leads the way to the next step of considering the value of the
parts that make up the whole as of an interest equivalent to the whole itself
(the identity of which is now also brought into question).
In Scholarly Editing in the ComputerAge, Shillingsburg provides a clear
overview of the major approaches to scholarly editing in the mid-to-late
twentieth century. Shillingsburg's own position (like that of Gaskell) turns
away not only from an author-centered model for editing and textual criti
cism but also from the need to assert any single approach as the only cor
rect one: "[N]o single approach is the right approach. Critics and schol
ars need texts for different purposes."" He divides text-critical responses
intO' four "orientations": historical/documentary, sociological, aesthetic,
and authorial. In each case, authority is located in a different place: in the

13

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

document itself, collaboratively between the author and publisher, in the


editor and the author, and in the author. The first two approaches will lo
cate authority in a specific text; the aesthetic and authorial approaches are
more likely to produce eclectic texts. Shillingsburg argues that the selec
tion made "depends almost entirely on where the editor locates or finds
the textual authority by which he proceeds to purify the text," although
he also points out that these approaches are not exclusive." One key point
that Shillingsburg makes here is that, ideally, an edition need not be lim
ited by the approach that has been taken: "Having chosen the most ap
propriate orientation for the editing of a particular text, it is possible to
prepare an apparatus that will make the edition useful to persons wishing
that another orientation had been employed."^" This is an important point,
since it suggests that material in a really well-presented edition is not nec
essarily trapped within the editorial principles that underlie it. Ultimately,
Shillingsburg articulates a movement in editing toward a more subjective
sense of the editor's role, one that allows for increasing interaction between
textual and literary criticism. At the heart of his book is the assertion that
"[t]he two keys releasing editors from the tyranny of their own theories are
recognition of the multiple nature of literary texts and of the fundamen
tally critical nature of editorial theories."^' As he puts it in a later work,
"editing is a critical enterprise that not only involves criticism but is in fact

GENETIC CRITICISM

a form of literary criticism."^^


This leads us to the most recent phase of textual studies, in which tex
tual criticism and literary criticism begin to be brought together and the
Anglo-American position moves ever closer to Continental practices. Both
McGann's and Shillingsburg's positions can be seen to be underpinned by
the work of the German critic Hans Zeller, and, as we shall see, ideas con
cerning textual versions and the study of draft material have been at the
heart of German editing for some time. A more recent German critic, Hans
Walter Gabler, also makes the case for much fuller critical use of draft ma
terial in "Text as Process and the Problem of Intentionality" when he states
that "the published form of a work need not categorically be an editor's
main, and overriding, point of orientation."" As Gabler emphasizes, inten
tion has imposed a strongly chronological and hierarchical structure on
the critical edition and encouraged a focus on the establishment of a single
text at the expense of the textual matter that precedes it. German scholar
ship is far more advanced in its ambitions for this material: it is willing to
edit manuscripts as manuscripts {Handschriftenedition) without needing
them to be "justified" by a final reading text, in a method that "emphasises
the presentation of textual matter over the critical establishment of text,
or texts."^"

14

Anglo-American Attitudes
Having established that editing practices clearly anticipate the emergence
of a "compositional" or "genetic" criticism in Anglo-American studies, we
need also to consider why such a criticism has been slow to emerge. The
term "genetic" was first used in its literary sense by Wimsatt and Beardsley
in their seminal essay "The Intentional Fallacy."^ In that (New Critical)
work, it is equated directly with a narrow, limited interest in biography
and the personal history of the poet: "[W]e submit that this is the true
and objective way of criticism, as contrasted to what the very uncertainty
of exegesis might tempt a second kind of critic to undertake . . . the way
of biographical or genetic inquiry."^ In the next essay of The Verbal Icon
("The Affective Fallacy"), the question of a genetic criticism is returned
to again: "The Intentional Fallacy is a confusion between the poem and
its origins, a special case of what is known to philosophers as the Genetic
Fallacy. It begins by trying to derive the standard of criticism from the
psychological causes of the poem and ends in biography and relativism."^'
Here, "genetic" is taken to refer to the author's reasons for writing rather
than the materials of the process.
Both critics go on to discuss the idea of "genetic criticism" at greater
length elsewhere and to give fuller definitions of it. In Aesthetics, Beardsley
states: "I call a reason Genetic if it refers to something existing before the
work itself, to the manner in which it was produced, or its connection with
antecedent objects and psychological states."^ This offers a potentially
broader definition. However, Beardsley then proceeds to condemn such
an approach, primarily by narrowing the focus of meaning to evaluative
issues: genetic reasons are used to assess the fulfilling or not of intention,
or to judge the work as successful, skillful, original, or sincere. On this
basis, he argues, "Genetic Reasons... cannot be good, that is, relevant and
sound, reasons for critical evaluations."^' Thus, he adheres to the argu
ment as first given in "The Intentional Fallacy" that while such material
may be used primarily for evaluation of the author's "success," this is not a
valid critical act.
Wimsatt's later essay "Genesis: A Fallacy Revisited" is equally condem
natory of intentionist critics "wishing to throb in unison with the mind of
the artist."'" As in "The Intentional Fallacy," Wimsatt sets up a contrast
between, on the one hand, response to an artwork as private, personal, and
a reflection of the personality of the creator, and, on the other, response to
it as objective and public, with meaning embodied in the text itself Like
Beardsley, he condemns the use of genetic material for any evaluation of

15

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

the work on the basis of the author's abiUty to express personal events, or
to fulfill his original intention, and asserts that authorial intention can, in
any case, never be fully known. Even where external knowledge of the au
thor is used in conjunction with textual interpretation, Wimsatt states that
such knowledge is really embodied in the text: "[I]nterpretations appar
ently based upon an author's 'intention' ofl:en in fact refer to an intention as
it is found in, or inferred from, the work itself."'' In sum (and as one would
expect from a founding father of New Critical methods), Wimsatt consid
ers that if a work of art requires the support of external evidence to com
municate its meaning, then it is a poor work of art and that the author's
intentions at the time of planning or writing may well be very different
from what is actually produced (and therefore are largely irrelevant since it
is only what is produced that matters). He affirms that "[t]he intention out
side the poem is always subject to the corroboration of the poem itself"'^
All of these points are convincing as they stand, but only in relation
to a view that the primary aim of genetic criticism is "the search for the
author's generative intention."" Wimsatt fails to allow for the possibility
of replacing a search for "what the author meant" with analysis of actual
genetic material. Both critics are unable to recognize that a very different
kind of interpretive act might take place in relation to the material itself,
one that would be interested in valuing the process revealed by it rather

enologists, structuralists, reader-response critics, deconstructionists, fem


inist critics and theorists, post-structuralist Freudians, and the rest, the
influence of Wimsatt and Beardsley ... remained manifest in a pervasive
distrust of biographical, textual, and bibliographical evidence."' However,
as I discussed in the introduction to this work, Parker's own method for
responding to the New Critical position is not totally satisfactory either.
German "Versions"
In The Term and Concept of Literary Criticism," Rene Wellek gives a his
torical overview of the use of the term "criticism" in literary studies and
compares its all-embracing use in England with that of other countries:
France, Italy, and Germany. In relation to German criticism he informs
us that, although the term "Kritik" traveled from France to Germany in
the early eighteenth century, "something happened to dislodge the term
and concept and to narrow it more and more till it came to mean only
day-by-day reviewing, arbitrary literary opinion."" As a consequence: "In
'Germany the term 'Literaturwissenschaft' took the place of 'criticism' as
used in the West. It succeeded while similar combinations such as 'science
de la litterature' or 'science of literature' failed in the West."" This issue of
terminology is telling in relation to the mode and methods of scholarship
that such terms describe and can be extended to textual criticism, as Hans
Waher Gabler makes clear when he compares the Anglo-American "criti
cal edition" to the German historisch-kritisch Ausgabe:

than the designing cause that created it.


The earliest use of the term "genetic criticism" is, then, very specific
and wholeheartedly negative, and may have had considerable long-term
impact on the study (or lack of study) of this material in Anglo-American
scholarship. Remarkably few Anglo-American critics have shown any real
interest in reconsidering the use of the term or in turning their attention to
draft material. The best-known exception to this is Hershel Parker, whose
approach, as already mentioned, is both text-critical and literary-critical.
Parker's basic argument, that material concerned with creative process has
been overlooked as a result of New Critical practice, is a fair one: "It is
customary and correct to attribute to the triumph of the New Criticism
the severing of American academic criticism and theory from almost all
(not just stupid and maudlin) biographical scholarship, and from almost
all (not just incompetent and pedantic) bibliographical and textual schol
arship."'" He points out that "'The Intentional Fallacy' has of course cast
a longer pall than its companion essay, 'The Affective Fallacy,' which the
reader-response critics had to repudiate."' Parker reminds us, too, that
the New Critical assumptions of the first essay significantly outlasted New
Criticism itself, with obviously negative consequences for the study of early
textual material: "As New Critics were joined (and succeeded) by phenom-

Other elements of terminology also reveal the assumptions underlying ed


iting and the influence of theory. So, for example, the use of the term "wit
ness documents (Zeuge) in German editions to denote manuscript mate
rials has its basis in the semiotic grounding of German editorial principles
by which the text is released from authorial possession into its existence
as a sign, or system of signs: "Manuscripts and prints are only the carriers
of the received record, the 'witnesses'; they provide the basis for establish
ing texts.""" In Anglo-American scholarship the same material was, until

16

17

What Anglo-American editing has upheld is the adjective critical, reap


plying it to competing authorial variants and their treatment. The Ger
man type of scholarly edition, by contrast, hasin the face of modern
textsstrengthened the adjective element historical (historisch). Con
comitantly, the element critical (kritisch) is understood to apply not so
,,much to the establishment of the text as to the analysisthe critiqueof
the text's genesis and history."

CONTEXTUALIZING PROCESS
TEXT AS PROCESS

recently, commonly called "prepublication material, reflecting the under


lying intentionist model (draft material is of interest primarily in leading
up to and helping to explain the first definitive text). In French genetic
criticism, where textual process is given the highest independent status by
geneticists, the term used for it is the "avant-texte," which points to a ma
jor division between private and public textual material but also implicitly
gives such material value (without this, the text would not exist).
A brief historical overview of German editing is necessary."' As in
Anglo-American scholarship, German editing principles developed out of
a lengthy philological tradition with origins (shared by Anglo-American
editing) in Karl Lachmann's adaptation of editorial principles for medieval
manuscripts to the German writer G. E. Lessing in the early nineteenth
century. German editorial theory and practice began to diverge from ite
Anglo-American counterpart in the 1940s with Friedrich Beif?ners edi
tion of HolderUn, which first allowed for the concept of textual variance
in a model of "organic growth toward unity and superior aesthetic integ
rity.""^ A sense of Beif^ner's own position is given through the translation
of Gunter Martens's 1971 essay from Texte und Varianten, when he says of
BeiCner; "He emphasized the 'methodological principle of following, as if
at the poet's side, the development and growth of lines and stanzas and of
capturing the work of art at its birth and thereby penetrating to the fullest
depths of its meaning.'""'
Hans Zeller outlines a slightly different line of development for German
editing. He compares the dominance of Shakespearean editing in AngloAmerican scholarship to that of the editing of Goethe in German schol
arship where the final revised edition, overseen by Goethe himself, was
initially adopted uncritically as the "standard text. Zeller tells us that this
edition was revised in the 1950s, on principles similar to those of Greg, by
Ernst Grumach: "[E]ach individual variant is investigated to see whether
it originates from the author. If this cannot be made to seem probable
the variant is not adopted in the text.""" However, this traditional model,
which bears comparison with Anglo-American editing, was replaced by
new principles for editing Goethe as defined by Siegfried Scheibe. Such
ideas, involving a theory of versions, were taken further by Zeller himself
The new principles were developed in the 1960s under the influence of
structuralism."' The work of the Prague structuralists led to a significant
shift of interest away from the author to the text itself as a system of signi
fying practices. Zeller and Martens's edited collection Texte und Varianten
(1971) is a seminal text that establishes the core principles of modern Ger
man editorial theory." These principles view the text as a multiple con
struct coming into being over time (diachronic), with each version of a

18

text being valued for its particular historicity (synchronic). Early material
is thus to be reproduced as an intrinsic part of an edition and treated as
necessary documentation of the development of the whole. The reader of
the edition is seen as performing an active role, engaging with and using
the edition to release the text for interpretation. Concerns over authorial
intention were removed from the German editorial debate at this point,
and a clear distinction between Anglo-American and German approaches
began to emerge, so that while "the Anglo-American endeavor has tended
to edit the author, the central German concern over the past decades
has increasingly become to edit the text.""' For some time, German edi
tors have been concerned with such questions as the distinction between
"readings" (Lesarten) and "variants" {Varianten) of the text and with what
constitutes the documentary body rather than with questions concerning
which authorial version to privilege. Zeller sums this up: "What distin
guishes more recent German editions both from most earlier editions and
from recent English ones is fundamentally a different understanding of
the notions of version (Fassung) and of authorial intention and authority
{Autorisation\ and ultimately a different theory of the literary work and
its mutations.""
It will be helpful to look more closely at the basic principles that in
form the German editorial theory of "versions" in the work of two critics:
Siegfried Scheibe and Hans Zeller. Siegfried Scheibe's article "Zu einigen
Grundprinzipien einer historisch-kritischen Ausgabe" (published in Texte
und Varianten) is an influential piece that establishes the new concept of
editing in terms of versions."' His position is then further defined by Zeller
in his seminal 1975 article "A New Approach," published in Studies in Bib
liography (and thus made available to an Anglo-American readership). In
his essay, Scheibe sets out to define core issues in German editing such as
the question of authorization {Autorisation), the status of the text, and the
nature of textual fault {Textfehler). In a later essay, "On the Editorial Prob
lem of the Text" (1989), Scheibe makes an argument in favor of authorized
versions on the basis of a model of staged authorial intention, allowing for
a temporal dimension. He states of different versions that they each con
tain "the work in the form that the author considers right, good, and rep
resentative of his or her intention in this new phase of labor. But the first
textual version, the first writing down of the text, also showed the work
in a form that the author believed to be right at that given moment."'" He
concludes: "Each version shows the work as it represents a specific point in
time and a concrete phase of the author's personal, artistic, and ideological
development.""
Scheibe's essay also raises the question of what constitutes "the text"

19

TEXT AS PROCESS

and of how much equahty to grant to different textual "versions."^ He reininds us of a distinction between the author's view of equality and that of
the editor: From the author's point of view the versions are decidedly not
equally valid. . . . But for the editor, all versions are in principle equal."
he German position repeatedly emphasizes the difference between the
author's and the editor's perspective on the text rather than attempting
to conflate them. For the most part the editorial perspective, in terms of
reception, is privileged over the authorial one, in terms of production. This
emphasis is reversed in French genetic criticism, as will be seen."
Finally, Scheibe reminds us of an unequal factor in German editions
that is easily overlooked:
Logically, therefore, all textual versions should be reproduced in a
historical-critical edition, insofar as they survive. At the same time,
such editions do not give all textual versions equal prominence, since
as a rule they make a distinction between 'Edited Text'... and so called
apparatus. Usually only the versions chosen as 'Edited Text' are printed
in their entirety; all others are more or less reduced to their variant pas
sages with respect to the 'Edited Text'. But these variants themselves are
Text exactly like the 'Text' of the fully reproduced version.
This is an important point for Anglo-American readers. In spite of the
strong arguments m favor of multiple versions and the principle of textual
appearance of the text in many editions is not equal at
all Scheibe goes on to state that in principle any text can be chosen as the
edited text, so long as its place within the chronology is made clear, but
certain versions are far more likely to be privileged.
Turning to Hans Zeller's work, one of the most surprising facts about
It is how early it seems to be in pointing out the limits of the Greg-Bowers
principles and how clearly in advance of McGann (who does acknowledge
Zeller s 1975 article in Critique). Zeller also seems to be unusual among Ger
man editors m that he orients himself in relation to the Anglo-American
position rather than in relation to French critique genetique, which pro
vides a stronger focus for most of the other commentators.
In 'A New Approach to the Critical Constitution of Literary Texts,"
Zeller argues against the creation of an eclectic text from a number of au
thorized versions on the basis of final intention. He prefers to recognize
difference between those authorized versions and to designate each one as
a distinct version" rather than combining them to create a single "final"
text. Zeller's position is underpinned by semiotics, with the work as a whole
being viewed as "a complex of elements which form a system of signs."^ He

20

CONTEXTUALIZING PROCESS

goes on: "Seen in this way a version is a specific system of linguistic signs,
functioning within and without, and authorial revisions transform it into
another system
Since a text, as text, does not in fact consist of elements
but of the relationships between them, variation at one point has an effect
on invariant sections of the text." In other words, the concept of text here
is of a whole body of interrelated structures within a system rather than
a discrete final object. The aim of the German historical-critical edition
is thus to reproduce a complete textual history over time, with a corre
spondingly fluid sense of the text as a developing object, rather than to
attempt to create a single "definitive" state of a text. Further definitions
then followin terms of authorization (those versions of a text approved
by the author) and textual fault (the suspension of authorization on certain
grounds). In general, the German approach is to limit editorial interven
tion and decision-making wherever possible.
Zeller's earlier (1971) article, now translated and published in Con
temporary German Editorial Theory (CGET) as "Record and Interpreta
tion," is equally valuable and seeks to outline a method for producing the
historical-critical edition. As in all his work, he shows a good knowledge
of Anglo-American principles and defines his own position partly in op
position to them: "Instead of the author's intention, which can only am
biguously be inferred or suspected based on the written records, we choose
as a model of text constitution the intention that the author attested to in
those records."' Such a comment reminds us of the strong influence that
German reception and reader-response theory has had on these editorial
principles. At its heart, this is a methodology based on giving the reader
access to as much material as possible, and with a corresponding lack of
critical commentary and analysis on the part of the editor: "Not from an
interpretation but only from the record and its documentation may new
interpretations be gained."" Again, this is made explicit toward the end of
Zeller's earlier essay: "The text is delivered to the user not for permanent
ownership, but rather as a task in which to participate."'
Finally, Gunter Martens's 1989 essay "What Is a Text?" in CGET pulls
together many earlier principles and clearly defines two German positions:
an bid-fashioned one that recognizes the need to reproduce variants but es
sentially sees them as outside the text; and a theory of "versions" that views
the text as a dynamic whole in which variants are different versions of the
text and revision is incorporated into the idea of the text itself He states
of this position: "In part, it also assumes that the various stages of a work's
genesis and revision belong to one and the same text, simply representing,
that is, various versions of the one text.... Every transmitted version of a
text is, in principle, equally valid."^ Martens's essay then goes on to con-

21

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

sider the emergence of this model from semiotics, reaching the position
that "[p]recisely because the text is a fixed sign in the traditional sense and
at the same time breaks or negates the sign, and unfolds its productivity in
this tension, it cannot be reduced to one pole of this dialectical movement.
It is both static and dynamic, is itself a unity of opposites."^ This last sen
tence, then, provides the essential idea underpinning the concept of ver
sions, meaning that the edited text must allow for the "doubleness" of the
text's existence. The editor "must structure a scholarly edition in ways that
articulate both aspects of a literary text: The static and the dynamic, the
fixedness and the breaking up of this fixedness.""
This brief consideration of principles of German textual editing shows
that it develops in the latter half of the twentieth century in a very dif
ferent direction from that of Anglo-American scholarship and editing.
Editorial principles encompass the materials of process and incorporate
them within the very concept of the "text" to be edited. German editorial
theory thus displays an extremely positive attitude toward compositional
material; "The sui generis edition of working drafts and manuscripts (the
Handschriftenedition) has become a central concern, indeed sometimes an
autonomous objective, of German scholarly editing."^ At the same time,
however, this does seem to raise the question of why, if editorial principles
bring such attention to manuscript material, a literary-critical method and
school has not also developed to make full use of such material in German

editing, the aim at all times is to provide a clean and clear metastructure
containing distinct elements with minimized editorial commentary. Ac
tivities that in Anglo-American editing would be defined as part of the
critical editor's role (and thus as "editorial," falling somewhere between
editing and critical activity) German editors would define as "criticism"
and separate from the presentation of the text or manuscript. The confu
sion this creates can be felt in the very use of the term "genetic criticism,"
which, in the German tradition, refers primarily to textual criticism and
i's concerned with how best to present the genesis of a text, or with the dis
tinction between reception and production texts and their relative status.
It does not refer to any kind of literary-critical (interpretative, analytical)
use of that genetic materialas an Anglo-American scholar might assume

literary studies.
Traditionally, in an Anglo-American scholarly editionsuch as the
volumes of the Cornell Series for Wordsworththe contents of a single
volume (following CEAA guidelines) would consist principally of; a schol
arly introduction; explanation of editorial procedure; the reading text;
headnotes and transcriptions for each manuscript contributing to the text,
in chronological order; and appendices (possibly including shorter read
ing texts of passages not included in the final published text). In such a
structural model it is possible that a reader might simply work from the
reading text without ever really engaging with the lengthy apparatus at the
back of the book. Variant material is represented, often quite fully and as
it appears on the manuscript page, but usually in a different place from the
primary reading text, or texts. More significant than the structural organi
zation, however, is the editor's conception of his or her role. In the AngloAmerican tradition the borders between "objective" editing activity and
"subjective" critical commentary are relatively fluid, so an introduction to
the text will almost certainly include some critical thinking, and this is also
likely to be the case for interpretation of the order of text and manuscripts
as articulated in the headnotes to the manuscript apparatus. In German

that it would.
Two English-language editions that attempt to connect AngloAmerican and German editing are worth brief consideration: Hans Walter
Gabler's synoptic and critical edition of Ulysses and James Mays's edition
of Coleridge's Poetical Works for the Bollingen Series. Gabler's edition
combines the German model of genetic editing for the synoptic text (which
appears on the left-hand pages of the work) with Anglo-American critical
editing for the reading text (on the right-hand pages); "Tying the edition
into the two editorial traditions simultaneously was necessary to meet the
complexities of the very task of editing Ulysses."^^ Gabler's German roots
can be seen at a structural level in the way that the edition goes straight
into the text in volumes 1 and 2 with all editorial commentary placed as
apparatus in the back of the third volume. More important, however, it is
representative of modern German editing in its attempt to represent pro
cess visually on the page, with parallel versions of texts and time dimensio;is brought into play. The synoptic text, which involves a "procedure of
telescoping several transmitted records in a single presentation," is com
mon within more recent German editions.^ Gabler makes this text the
focus of the edition. Controversially, his aim is not to produce a text that
resembles the first published version but one that draws on and privileges
Joyce's autograph manuscripts (before typing, copying, and printing). Ga
bler claims that this text is the "document text of highest overall authority"
above the first edition of 1922 and states that "the edition endeavours to
recover the ideal state of development as it was achieved through the traceable'processes of composition and revision at the time of the book s publica
tion." He distinguishes the published text from the text he is representing
here: "The first edition comes closest to what Joyce aimed for as the public
text of Ulysses. Yet it does not present the text of the work as he wrote it."'
One of the strengths of Gabler's edition is the appropriateness of a synoptic

22

23

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

approach for the material, a point made by Philip Gaskell in From Writer to
Reader: "Such a genetic presentation is particularly appropriate for Ulysses
because from the final draft stages onwards Joyce developed the book very
largely by adding to the text."^" It is relatively easy to follow Gabler's text
once one has adjusted to the symbols indicating different kinds of revision
on the left-hand pagealthough any serious comparative work on those
changes cannot be carried out unless the reader also refers at the same time
to the manuscripts (also available in a separate facsimile edition).
James Mays's Coleridge edition clearly follows the German modelhe
presents the Poetical Works in three volumes with two parts each devoted
to Poems {Reading Text) and Poems {Variorum Text). In his introduction.
Mays engages with recent editorial debate and with German and French
editing, stating, "I have adopted some of the features of this GermanFrench tradition in editing, I hope without prejudice."'' For the Variorum
Text Mays gives a brief headnote for each poem and then uses symbols on
the transcription page to denote kinds of revision, sources, and different
textual versions. At the bottom of the page are also further variants, the
division between the two positions being explained in the introduction as
"one of convenience, since both kinds of variation affect meaning."''^
In both Gabler's and Mays's editions a key issue is the relationship be
tween the synoptic text and a single-state reading text. Is the coexistence of
these two presentations of material a major conceptual contradiction? Tan
selle considers it to be so, asking of Gabler's edition. Why, one is bound to
ask, should there be a separate 'reading' text if all the variants are an essen
tial part of the work?"'' However, D. C. Greetham, responding to Tanselle,
argues that Gabler's synoptic text successfully "moves in both horizontal
and vertical axes at once" (i.e., it is both synchronic and diachronic in itself)
and that also having the reading text (as a purely diachronic text) allows
the edition to offer "both of Jakobson's disorders (similarity and contigu
ity) and both of Saus;sure's axes (vertical and horizontal) simultaneously."'"
He continues, "[T]his dual reading forces the reader's eye and concentra
tion to recognize the two dimensions of textual warp and woof.""
In the case of Mays's edition, however, matters are complicated by the
editor's own account of the relationship between synoptic and reading
texts (which are not on opposing pages but in two different volumes). In
his introduction. Mays clearly aligns his synoptic text with the German
structuralist model (although with some reservations) and states: "I be
lieve that the essential advantages of the European manner of proceeding
have been retained. These enable a reader to hold in mind a sense of the
way the poems move, as they often do, simultaneously in several planes:
that is, the way the poems move laterally, as a series of independent ver

sions, and vertically, as one version overlays and succeeds another."' This
synoptic mode of presentation is partly chosen because of the nature of
Coleridge's compositional habits: "The method is specially suited to texts
which underwent continuous revision, which was not always directed to a

24

25

single end.""
In Gabler's Ulysses the synoptic text is privileged over the reading text
in what McGann calls "the most dramatic representation of the work's
postmodern textuality. In an earlier edition the 'clear text' on the rectos
would have been the editor's ultimate object of interest."' However, in
the Coleridge edition, when Mays turns from the variorum volume to the
reading text, he asserts, "The Variorum sequence is the foundation of the
p'resent edition, but the Reading Text is what it supports," and again, that
"it is equally part of the edition, indeed it is the edition in the strict sense
(the Variorum sequence merely contains the material for it)."" Such state
ments seem to give priority to the reading text over the synoptic one. Thus,
the German structuralist principles of the variorum (which not only pre
sents texts synoptically but also goes so far as to collate textual misprints
and nonauthorial changes) are now countered by a strongly authorial ap
proach. For the reading text. Mays also chooses a principle of "variable
copytext," again to reflect Coleridge's inconsistent attitude toward his own
material." However, the principle followed is either "to give the version of
the poem which reflects Coleridge's concern, up to the time he lost inter
est" or to select a version that depends on recognizing "Coleridge's mean
ing before it was modified by second thoughts or other circumstances."'
Thus, selection of reading texts is decided on highly intentionist grounds,
and grounds that are subjectively determined. Mays willingly acknowl
edges this, and he also justifies it on the grounds that "[t]he problem is
more acute in theory than in practice. The text for each poem in all but a
few cases selects itself."^ Nevertheless, the edition does not so much serve
to unite Continental and Anglo-American editing practice as to leave the
two volumes, at best, as two entirely separate texts and, at worst, in conflict
with each other. This is not to condemn Mays, who clearly has considered
all of these issues with care and who has good reasons for thinking that this
editing approach is particularly suited to this poet. However, it does show
the limitations of the synoptic approach (if it always has to be supported by
an alternative text that works against it) and the ongoing conflict between
what sounds good in theory and what is actually of use to readers.
On the whole, then, at a practical level, the German model allows for
good localized textual comparison (line by line); it creates a sense of text
as process and introduces a time dimension into editing; and it tries to
be democratic in relation to different versions of a text. However, it is not

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

allow it).
A second limitation in the German model, already touched on, is raised
by Paul Eggert in his review article of CGET: "A great deal of thought in
Germany has gone into enunciating the principles behind the display of
historical versions of a work. Relatively little seems to have been devoted to
the empirical question of which text ought to become more widely avail
able to the reading public through the editor's efforts."' Determining what

will serve as the "edited text" is left open to the German editor, even though
this gives considerable prominence to whichever text is chosen within the
edition. There is no real debate about the status or reason for choosing a
particular edited text (as there is, of course, within Anglo-American schol
arship). Hans Zeller also makes a related point that the Anglo-American
f, tradition is "capable of differentiating between authorial and nonauthorial
variants more reliably," and he repeatedly acknowledges the usefulness of
analytical bibliography in Anglo-American editing (even if it has been put
to misplaced ends).'
We can conclude this section of the argument by comparing German
and French responses to manuscript material, which many German com
mentators discuss. The first point to bear in mind has already been made
that for the Germans, "genetic criticism" does not denote critical analysis
of texts but instead refers to textual criticism for genetic material. Gabler
ihakes this point by implication when considering the question in relation
to the French critique genitique: "What needs stressing, particularly for
anyone accustomed to taking the text-critical view of manuscripts, is that
its points of departure as well as its aims are critical. Critique genetique
does not innately exploit manuscripts as documents of the written for pur
poses of deriving scholarly editions from them. It discourses the analysis
of manuscripts as writing." From a German perspective, the main advafttage for French scholars working in this way is that they are not bound
by a long tradition of scholarly editing, as German, British, and Americah'scholars are. In an article that goes some way toward trying to define
the need for interpretation of genetic material in German criticism, Klaus
Hurlebusch states:"From the beginning, the path followed by critiquegenetique'differed from that of the German editorial genetics. Unlike the latter,
the French approach did not have to ask itself the hermeneutical question:
what sense does it make to turn one's attention toward textual materials
preserved in manuscripts that originally had been significant only to the
author?"' Hurlebusch seems to look wistfully across to the freedom of
French thinking: "One can hardly imagine the like within the Germanist
-discipline. The independence of such investigations is here way-laid and
blocked off by a reception-oriented hermeneutics that appears tolerable at
most as a preliminary to critical editing."'" German editing theory thus
makes a strong case for integrating manuscript material and the repre
sentation of text as process into the edition, but it does not appear to go
further and consider the critical status, or use, of such material. It is time,
then, to turn our attention to the French critique ginetique, which does
make this its central concern.

26

27

helpful when one wants to consider a long text across different versions or
if one requires a whole sense of the text at each stage. The mode of editing
necessarily condenses different layers of revision, and does so intelligently
and well, as far as such things go. But it could be argued that the actual
physical mass of a text, even of a revised text that contains much invariant
material, is important, if impractical to reproduce.
It seems to me that one fundamental reason for the lack of German
literary-critical engagement with compositional material is the physical
presentation of the variant texts." The linear, synoptic version developed
by the German tradition is difficult to work with because it destroys the
spatial whole of each version of a text on the page and removes the compo
sitional context. Perhaps the ideal, in this editorial model, is to work with
the combination of a synoptic text alongside full facsimile material (as in
the case of Gabler's Ulysses and The James Joyce Archive).
What are the limitations of German editing theory in the wider sense?
Gunter Martens makes clear the extent to which German theory has moved
ahead of the current enthusiastic optimism for textual versions in AngloAmerican scholarship, to the point of a strongly articulated resistance to
such a theory by certain German editors. This leads Martens himself to
raise some of the objections to it: It is impossible to avoid the questions of
whether the immense expenditure of many editions is justified; whether
the expectations for the representation of textual development are fulfillable; and whether the large investment of time and effort disregards the es
sential interests of scholarship and the primary needs of readers."" Again,
in his later essay "What Is a Text?" (1989), after clearly articulating the need
for editing principles that allow the text to exist both as process and as
product. Martens is still compelled to conclude: It will not be possible to
implement fully this theoretically based model of an edition of the textgenetic and text-arresting relationships in every case. Compromises will
have to be made; foreshortenings of the comprehensive conception will be
unavoidable."' So one limitation of the theory of versions is that the prin
ciples of editing are not generally able to be fully implemented within the
book edition (although of course electronic archives and hypertext would

TEXT AS PROCESS

French "Critique Genetique"


As might be expected from a still-developing field, French genetic criticism
frequently raises questions about its own status, value, purpose, and scope,
all of which are directly relevant to this study. Questions include:
. What is writing? How does one write? How does one analyse writ
ten language when the document stacks paradigm over syntax? What
is literary writing? In terms of theory, what is the status of these
"pre-texts"?
Is this a new incarnation of philology? A heuristic study of manu
scripts? A critical procedure? An autonomous discipline?
Is genetic criticism a theory of criticism or just helpful advice, some
thing like: keep in mind that manuscripts can also contribute to the
understanding of literature? Can an original methodology be extrap
olated from an individual case and applied to another beyond stan
dard scholarly procedure?"
Unlike German criticism in this area, French criticism is released from the
issues involved in a totalizing theory of versions, because it largely (though
problematically) conceives of the avant-texte in autonomous terms as apart
from the texte and viewed as a distinct ontological object.'^ French genetic
criticism does not emerge out of French scholarly practice in editing but is
placed in opposition to an outmoded traditional model: Geneticists react
to the French editorial tradition, which opts systematically for the final
version revised by the author (few exceptions confirm this rule).""
The critique genetique emerged in France in the 1970s. Avant-texte, the
term for prepublication material, was first used by Jean Bellemin-Noel in
1974, and the term critique genetique was coined by Louis Hay in 1979.
The history of critique genetique is bound up with the acquisitions of the
Bibliotheque Nationale from the late 1960s onward.'" The acquisition of
the manuscripts of Heinrich Heine led Louis Hay to establish a research
team, funded by the Centre National de la Recherche Scientifique (CNRS),
to work on the material. In 1974 a number of groups working on different
authors' manuscripts combined to create the Centre d'Analyse des Manuscrits Modernes (CAM), and then, in 1982, this center became a major re
search institution housed close to the Bibliotheque Nationale and called
the Institut de Textes et Manuscrits Modernes (ITEM).''
The scientific context for funding is interesting in that it has shaped
the nature of the work to quite a considerable degree; as French commen
tators point out, "genetic studies were from the start a team-effort. This
28

CONTEXTUALIZING PROCESS

Style of research, wary of personal intuitions and expressions, is in hne


with the dominant scientific research structure of CNRS, which favors the
laboratory model and well-defined methodological processes focusing on
a well-defined corpus."' Louis Hay, in the course of outlining broad inter
national and intellectual sources for genetic criticism, also emphasizes the
importance of its being located outside the academic world of the univer
sity and as part of a more direct line of connection with writers themselves:
"Et surtout: c'est dans la reflexion des ecrivains bien plus que dans celle des
universitaires que la critique genetique a pu trouver des moddes" [And
above all: it is in the reflection of writers far more than in that of academics
that genetic criticism has been able to find its models]."
Theoretically, French genetic criticism locates its origins in a line of
developing self-consciousness about poetic process that moves from Ger
man Romanticism (Goethe, Schlegel) through Madame de Stael and Poe.
Precursors to the field exist in the form of manuscript studies of individual
authors, but it emerged as a distinct area of study only when the focus
turned away from the treatment of such material authorially, as a source of
biographical or psychological information about an author, and toward a
much more text-based approach in the 1970s.
In the seminal article "Does 'Text' Exist?" Louis Hay outlines the paral
lel development of theories of the text (Barthes, Derrida, Kristeva) along
side that of genetic criticism, appearing to define the latter in opposition
to a structuralist or post-structuralist approach: "Both the method and the
object of study of genetic criticism are quite distinct [from the theory of
the text]. Its method is the result of extensive empirical work dedicated
to authors' manuscripts. .. . Genetic criticism retains from its origins an
inductive approach, which builds up general models from a series of con
crete observations."' At its narrowest, then, French genetic criticism is
defined as an empirical, grounded, scientific approach to the materials of
process. There are, however, obvious difficulties with this. Laurent Jenny,
a skeptical commentator, points out: "Genetic criticism is searching for a
phenomenon that is in effect unobservable, unobjectifiable: the origin of a
literary work. Its object of inquiry is essentially unstable."" Moreover, in
spite of Hay's apparent definition alongside, but against, a theory of the text
in "Does 'Text' Exist?" such theoretical ideas have significantly influenced
the field and its development. Pierre-Marc de Biasi describes how: "Struc
turalism allowed us to specify the decisive character of certain processes at
work in writing.... Research advances in semiology culminated in a more
dynamic analysis of these structural processes, making the text the center
of a theatricized process of meaning."'"" In his article. Hay himself turns
away from the rigid separations of structuralism, determining instead to

29

TEXT AS PROCESS

look "not at the opposition between pre-text and text, but at the relation
ship between writing and the written work." This leads him to his famous
description of the avant-texte as a "third dimension" of tfext: "the writing is
not simply consummated in the written work. Perhaps we should consider
the text as a necessary possibility, as one manifestation of a process which
is always virtually present in the background, a kind of third dimension of
the written work."""
In spite of repeated attempts at self-definition, critique genetique as a
movement always "remains paradoxical,"'"^ in part because of a conflict
between its material, empiricist focus and its more intangible aspect. In the
introduction to the collected essays of Genetic Criticism, the editors com
ment on such paradoxes in the field: "It grows out of a structuralist and
post-structuralist notion of 'text' as an infinite play of signs, but it accepts
a teleological model of textuality and constantly confronts the question
of authorship. ... it examines tangible documents such as writers' notes,
drafts and proof corrections, but its real object is . . . the movement of
writing that must be inferred from them."'"' Such methodological contra
dictions also lead to the treatment of the material itself as having a double
identity. In "Does 'Text' Exist?" Hay defines the field as locating itself both
in "the material given" (document) and as an "intellectual construction"
(avant-texte). This "dual status" means that genetic criticism is often con
cerned with methods of access to the material at the same time as working
with it more critically.'"" Biasi also registers a "double objective" for the
geneticist, in terms of rendering material readable as well as reconstruct
ing genesis; he describes the avant-texte as "no longer a set of manuscripts
but an elucidation of the logical systems that organize it, and it does not
exist anywhere outside the critical discourse that produces it."'"' Even the
central term for the materials of study is thus problematically doubled.
In a roundtable debate of 1995, transcribed as "Archive et Brouillon" in
Pourquoi La Critique Genetique? Michel Contat poses the initial question
of whether it is possible to substitute the concept of avant-texte for that of
archive. In the course of Contat's discussion with Jacques Derrida, the lat
ter seeks clarification of what is meant by avant-texte:
JD: Si j'ai bien compris, vous appelez avant-texte un etat d'ecriture qui
precede I'etablissement legal de publication.
[JD: If I have understood correctly, what you call the avant-texte is a
state of writing which precedes the legal act of publication.]'"

Michel Contat responds by defining the limits and textual status of the
avant-texte in these terms: "Avant son impression, avant la decision de

30

CONTEXTUALIZING PROCESS

1 auteur de publier I'avant-texte etant en plus une construction critique"


[Before its printing, before the decision of the author to publish. ... the
avant-texte being, in addition, a critical construction].'"' Derrida goes on:
JD: Mais est-ce que le materiau brut, avant que vous fassiez un travail
sur un heritage de brouillons, par exemple, sera un avant-texte?
[JD: But the raw material before any work was done on a study of the
history of rough drafts, for example, would that be an avant-texte^]
MC: En principe, non. On parlerait de dossier preparatoire, de dos
sier documentaire. L'avant-texte serait deji le resultat d'une activite
critique.
[MC: In principle, no. One would call these a preparatory dossier, or a
documentary dossier. The avant-texte would already be the result of
critical activity.]""
The debate makes two key points very clear. The first is that the French
genetic model is not only aware of a distinction between avant-texte and
text but that it also distinguishes carefully between diflferent states within
the body of compositional material, defined as the manuscript, rough draft
materials, and the avant-texte. In so doing, it looks back to distinctions
originally made by Jean Bellemin-Noel in 1977 between the cultural and
material object (manuscript), the gathering together and ordering of ma
terials (dossier), and the presentation of such materials in a reconstructed
order (avant-texte).Second, it emerges that, by 1?95, there is a degree of un
certainty about such distinctions, with the possibility that all texts should
be considered as one text. Unsurprisingly, Derrida rejects distinctions
between one kind of text and another, asserting that no text is ever fully
stable and that "I'archive est toujours un texte" [The archive is always/still
a text].'"'
At its most extreme, the Derridean position raises fears that a possible
Shift of emphasis onto process rather than product, the notion of "writing"
as something equal to the completed work of art, will be achieved at the
cost of the destruction of both. This anxiety is raised by Almuth Gresillon,
who points to ^'an internal contradiction that genetic criticism helps to ex
pose. If one is interested in the manuscripts of works, it is because there is
a link between the pre-text and the text and that the study of one will lead
to an increased knowledge of the other. But at the same time, the impor
tance given to pre-texts undermines the sacrosanct auctoritas of the text,
because it is reduced to the status of just another state among others."""
Laurent Jenny also notes that there is a paradox underlying the private/
public relationship between the two kinds of text: "A pre-text derives its

31

TEXT AS PROCESS

value from the consecration of the text that it precedes. But paradoxically,
the establishment of the pre-text tends to dissolve the textual entity that
was precisely the one that gave it this value."'" This is well put, but .it is
rebutted by Biasi, who states of genetic criticism that it in no way seeks
to strip the text of its poiesis. Its aim is to widen the concept of writing
by opening up access to writing's temporal dimension: a dimension that
would allow us to support the structural study of the text with a poiesis of
the avant-texte."^^^
Such near-contradiction in terms of the relative valuing of textual pro
cess and final text lies at the heart of any attempt to work critically on draft
materials and is present in this study as well. On the one hand, I want to
validate process as an object of analysis in its own right and to consider
textual material in a state of process in a way that allows for its difference
from the published or completed text. On the other, I am arguing for criti
cal integration and movement across and between avant-texte and text,
seeking an enlargement of the definition of literary studies to include this
material. The critique genetique reveals that such delicate negotiations of
position are part of the nature of working with compositional material.
The double focus of French genetic criticism makes it highly distinctive
and important for this study. Unlike both Anglo-American and German
traditions, where a genetic approach is for the most part only concerned
with the production of editions, the critique genetique was from the outset
as much critical as text-critical in focus. A sense of how relatively advanced
the French are in their literary-critical use of genetic material can be felt in
Gresillon's account, in EUments de Critique Genetique, of different theo
retical approaches brought to bear on draft materials in France (narratological, thematic, psychoanalytic, socio-historical)."' Such a range of criti
cal responses seems to place the critique genetique significantly ahead of
Anglo-American writing, which is still at the stage of justifying such mate
rial as being of significant critical value and has not yet achieved the level
of a coherent, recognized field of study. Work in this area is far more likely
to be "assimilated to a form of textual criticism or automatically assumed
to be a branch of it.""" However, the range of approaches in France reflects
an ongoing tension between theories of the text and critique genetique and
the question of the extent to which the avant-texte is, or is not, to be con
sidered as just another kind of text. Gresillon expresses concern that such
critical responses simply respond to compositional material in the same
way as the published text: "Traiter l'avant-texte comme du texte, n'est ce
pas forclore par avance tout espoir de th6oriser I'approche gen^ique pour
elle-meme?" [Isn't treating the avant-texte as a text to foreclose in advance
all hope of theorizing a genetic approach for its own sake?]."' This seems

32

CONTEXTUALIZING PROCESS

to me an important concern, one that is, again, always present in critical


response to the text in a state of process.
A brief look at the French genetic edition may be helpful in considering
the extent to which theoretical ideas emerge in the presentation of texts.
French genetic editing asserts the need to make active choices, to select
and to present a dossier as opposed to the all-inclusiveness of the German
documentary method. This difference is immediately felt in the size of the
editions. Unlike the substantial multivolume German structures, French
editions seem to try and pack as much material as possible into a relatively
small space.
My first example is Paul Valery: Cahiers, 1894-1914, edited by Nicole
Celeyrette-Pietri and Judith Robinson-Valery The Cahiers were written by
Valery, alone and in private, every morning as part of his daily routine
from 1894 onward. They are neither a journal nor a rough draft for a pub
lished work but represent the poet's lifelong dialogue with himself They
^re, then, the ultimate genetic text of a genetic text that is "consciemment
et volontairement ouverte et comme intrins^quement inachev^e" [con
sciously and voluntarily open and as though intrinsically incomplete]."
This edition has been preceded by a full facsimile edition of all the 260
notebooks in twenty-nine volumes and a selected edition of 3,000 pages.
As a result, it is free to be partial, choosing to cover only the first twenty
years of the notebooks up to the First World War. The principal justifica
tion for the edition is its presentation of the text in a typographically ac
curate form. The editors state that their main objective is "de presenter au
lecteur un texte facile et agreable a lire" [to present to the reader a text that
is easy and pleasant to read]."' In this, they certainly succeed. The page
itself is faithful to the manuscript in terms of layout, with minimized code
and reproduction of images and of manuscript originals throughout. The
editors have, however, made judgments as to which images are significant
(and thus reproduced) or not, and in relation to deletion (major deletions
are marked, but minor ones are silently removed). What the edition pre
sents, then, is a kind of "genetic reading text" of notes, ideas, and doodles
in a spirit of openness appropriate to the writer. The reader is invited to
"les utiliser tout simplement comme tramplin pour pousser plus loin sa
propre reflexion" [use them simply as a springboard to push his or her own
reflections even further]."'
Another influential example of a French genetic edition is provided by
Pierre-Marc de Biasi's 1988 edition of Flaubert's Garnets de Travail [work
ing notebooks] in a single volume of 999 pages. The edition contains a sub
stantial genetic framework of some complexity. In the front section Biasi
presents the study explicitly as a genetic edition, gives a detailed history

33

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

of the manuscript collection and of editions of Flaubert, and then turns


to his construction of the Garnets de Travail out of two previous group
ings of Flaubert's manuscripts [Garnets des Notes de Voyage and Garnets
des Notes de Lecture). Once recategorized, the notebooks break down into
two groups: thirteen small pocket notebooks {calepin d'enquete [notebook
of enquiry]) that Flaubert took with him when he went out and four large
notebooks (grand carnet d'idees [large notebook of ideas]) in which much
of the major research and planning of the novels occurred. In a scientificlooking table Biasi represents Flaubert's use of these two kinds of notebook
over time, and then looks in detail at the characteristics of each kind.
In the text of the edition itself, Biasi, like the editors of Valery's Gahiers,
emphasizes his desire to make the material available to all readers. The
notebooks are arranged chronologically in groupings relating to a major
work, with a framework for each one giving a biographical and genetic
context. The edition is surprisingly authorial in this respect, with Biasi giv
ing, in part, as his rationale the fact that "partir de 1851, la vie de Flaubert
coincide assez precisement avec la genese de ses oeuvres" [starting from
1851, the life of Flaubert largely coincided with the genesis of his works]."'
In spite of this impressive superstructure, however, the text itself is dis
appointing. In his theoretical work La Genetique des Textes, Biasi defines
different kinds of transcription as either "transcription linearisie" [linear
transcription] or"transcription diplomatique" [diplomatictranscription].'^"
The former uses code to represent deletion and addition and reads as a
continuous text (not respecting line endings on the manuscript page). The
.latter reproduces the manuscript text as accurately as possible with mini
mal code. Clearly, in the Flaubert edition, Biasi opts for the first mode, even
though in Genitique he states that for Flaubert, "I'id^al est la transcription
diplomatique accompagn^e du fac-simile du manuscrit" [the ideal is a dip
lomatic transcription accompanied by a facsimile of the manuscript].'^' As
Biasi makes clear in his introduction to the edition, if he had attempted a
diplomatic transcription, the edition would have run to four or five vol
umes instead of one (and certainly the fact that the book is a single volume

the French editions have a tendency to go too far in the other direction,
compressing the transcription material and giving only a limited sense of
the materiality of the manuscript. This is understandable in the light of
theoretical distinctions between manuscript, dossier, and avant-texte, but
it can make the final presentation frustrating for the user.
Certain key concepts and anxieties emerge over and over again in writ
ings about the critique genetique by its own practitioners. These include:
debate over the relative status and relations between avant-texte and texte;
ambivalence about allowing room for psychological states and processes
and about using a teleological model of development (both clearly autho
rial); concern over the doubled definition of,the materials of composition
as both document and avant-texte; and anxiety over the theoretical under
pinnings of the field and the mode of interpretation for it.
Some critics, however, have made use of the potential tension between
authoriality and textuality in the critique genetique to develop nonau
thorial teleological structures in ways that are of considerable value to
this study. As a second-generation genetic critic, Daniel Ferrer is unusual
within the French tradition in expressing interest in overcoming the "cu
rious relation of exclusion and reciprocal comprehension [that] remains
between production and product, genesis and work."'" In "Clementis's
rCap: Retroaction and Persistence in the Genetic Process," he outlines a
model of "multiple teleology" (i.e., two-directional) with an anticipatory
and retrospective vision for the avant-texte in which "every act of notetaking occurs with the expectation, however vague, that the note will
somehow be used; therefore every part of the avant-texte in some way reyerts back to the projective logic."'^' Ferrer argues that genetic criticism
can run "downstream" but also "in a retrograde motion" by means of this
projective logic: "each fragment of draft projects itself onto the horizon of
its completion."'^" Again, unusually, Ferrer explicitly asserts the necessity
of a teleological perspective: "[I]t is in vain that genetic criticism regularly
exhorts itself to renounce a teleological vision of genesis. Teleology is not a
critical artifactit is inherent in the genetic mechanisms. However we will

is a major attraction).
These two examples reveal editions much less bound to documentary
accuracy and absolute inclusivity than the German model, much freer in
spirit and with a strong emphasis on the usability and attractiveness of the
volumes produced. The primary objective is making materials accessible
to a wide audience that is expected to want to read and enjoy the text as
process (not merely use it for scholarly purposes). A critical comparison
of the two Continental models might suggest that, if German editions are
unwieldy in their full facsimile pages with facing transcriptions and notes.

see that its teleology is multiple."'^'


In "Post-Genetic Joyce," Ferrer and Michael Groden revisit Groden's
Ulysses in Progress (1977) in the light of historical developments away from
an authorially centered teleological response to compositional materials
and toward a more materialist one. Groden articulates this shift: [T]he
sense of a teleological movement from early stages to finished product can
be replaced at least provisionally by one of a textual field that extends back
wards and forwards between avant-texte and text. Seen in this way the
process need not be interpreted as heading towards 'one great goal . . .

34

35

TEXT AS PROCESS

CONTEXTUALIZING PROCESS

and the pubUshed text can be reconceived as a provisional central point, a


'caesura' in the line of writing."'^ This kind of flexible use of a teleological
macrostructure for process draws close to the model of integrated working
across and between avant-texte and text that I am interested in and will
outline in chapter 3 of this work. Finally, another critic making (slightly
less) flexible use of teleology is Pierre-Marc de Biasi, who argues that "the
production of a text can be understood teleologically, but the process of its
development reflects a logic intersected by many possible becomings."'^'
His ideas, which significantly influence this study, will be explored in more
detail in chapter 2.
What are the criticisms and limitations of the critique ginetique'? As
we have seen, many of these are being self-consciously raised by those in
volved within the field, but from an outsider's perspective, there are a few
more points to be made. Clearly, the major feature of this critical move
ment is that it has a very distinct identity and seems to stand aloneboth
in its emergence from a research center outside the academy and in its
apparent (lack of) relation to traditional editing. In many ways, this posi
tion is both the strength and weakness of the critique genetique: while it
is autonomous, it is also potentially marginalized as a movement. Within
French genetic criticism there is a lack of engagement with Anglo-American
criticism (perhaps not surprising) but also with German criticismwhere
cross-referencing is largely historical rather than theoretical. This is par
ticularly striking by comparison with the German position, where many
commentators were keenly aware of French developments.
Finally, critique genetique does tend to be strongly methodological in
focus, that is, to be about ways of organizing and categorizing material
rather than developing a critical analysis of its own emerging from a full
understanding of the nature of the material. Although I see nothing wrong
with a range of critical approaches to composition, there is a need for a
genetic criticism to clearly define the hermeneutics of reading the avanttexte, and I am not sure that this has yet been undertaken. Gresillon re
peatedly raises this issue in Elements, and in her final chapter ("Towards a
Genetic Theory"), she goes so far as to state:

My attempt in this study to offer a philosophical account of the nature


of creative process (in part to validate textual process as a field of study)
and to explore the hermeneutic aspects of composition may therefore have
something to contribute here.
Chapter 1 has shown that the emergence of a "genetic" or "composi
tional" criticism is strongly reliant on changing theoretical and editorial
conceptions of the text in the late twentieth and early twenty-first centu
ries. However, the chapter has also revealed that the relationship of genetic
criticism to "Theory" is far from straightforward. French genetic criticism
aligns itself with such approaches when it attempts to explore its subject
matter as another form of Barthesian"writing": representative of the "birth
of the text." At the same time, as we have seen, the critique genetique does
retain something of a positivist attitude toward its subject matter, seeking
to treat it "scientifically," to assert an essential difference between avanttexte and text, and to organize and categorize responses to the material
in ways both teleological and implicitly authorial. Critique genitique thus
distances the material it studies from a conception of it as the thoughts on
paper of a creative individual, but at a larger level it still implicitly respects
all kinds of authorial acts and decision-making.

Si la genetique est bien une position critique et non une mode ou une
science n^o-positiviste, comme on a pu le pr^tendre, elle doit se donner les moyens appropri^s pour ^lucider, evaluer et interpreter la genese
des oeuvres. [If genetic criticism is truly critical and not a neopositivist method or science, as it has been made out to be, it must adopt the
appropriate means to elucidate, evaluate, and interpret the genesis of
works].'^

36

37

THEORIZING PROCESS

/lOCeyOyO

ORIGINS, AGENCY, INTENTION

Follow the accident, fear the fixed planthat is the rule.


John Fowles, Wormholes
An Anglo-American compositional criticism would be in danger of reject
ing its own past if it simply accepted a German theory of versions, or the
French critique genetique as its model. It must develop a method of its own
for responding to the materials of process, a method that builds on, and
emerges from, its own text-critical history.
For French geneticists the charge of any return to "philology" is insult
ing because in that historical and national context, the field has defined
itself in opposition to such a concept of the text and of editingand has
advanced very rapidly as a result. The movement is, however, constantly
anxious about, and vulnerable to, such a charge: If genetic criticism is to
become something other than an extension of old, dependable philology, it
must answer the questions it raises."' For an Anglo-American school such
a charge is less problematicgenetic criticism in this context could emerge
out of "old, dependable philology" as a positive, more intellectually and
theoretically enlightened approach. That is, the study of textual process
can draw on recent advances in text-critical and editorial thinking as well
as locating itself in relation to a long tradition. A revisiting of the concept
of intentionwhich has lain at the heart of the Anglo-American tradition
for so longis one possible way of developing a distinctive model. What
is meant by this, however, is a more sophisticated model of intention than
that of "final intentions," one that conceives of intentionality as a complex
of mental states or acts fundamental to process and embodied in the mate
rials of composition. In "Post-Genetic Joyce, Daniel Ferrer gestures in the

38

direction I am taking when he states, "Intention is revealed as a fluctuat


ing, time-bound transaction between a series of writing events and a series
of external constraints."^ He also goes on to make a further leap from a def
inition of genetic studies primarily in terms of its materiality to consider
that "genetic criticism (as opposed to strict textual criticism) is concerned
not only with textual material but also (or primarily) with actions: speech
acts, or rather acts of writing."' Again, such a position is very close to what
I will be trying to articulate here.
This second chapter sets out to confront some of the theoretical prob
lems facing the Anglo-American literary critic who wishes to work with
textual process. In particular, it addresses the question of how such work
can position itself in relation to the "death of the author" and the appar
ent rejection of authorial intention as an element of interpretation. If we
discard the authority of the author altogether, then how do we deal with
process? If we respond to process simply in authorial terms, then how do
we avoid a simplistic and limited understanding of the nature of language
and meaning?
There are two main concerns for this chapter, as it attempts to pave the
way for the method that follows in chapter 3. The first is to look at the treat
ment of creative process by late twentieth-century theorists and to explore
some of the reasons this aspect of the text has been overlooked. A second
concern is to revisit the concept of authorial intention in order to under
stand what is meant by such a term, the misconceptions associated with it,
and its importance in relation to the study of draft: materials. Clarification
is needed not only in relation to the literary debate over intention but also
in understanding intention in terms of actions and philosophical intention
ality. Only once we have a clear sense of the role of both intentional and
unintentional kinds of meaning can we begin to articulate anykind of com
positional or "genetic" method for the study of text as process. Chapter 3
will begin to articulate such a method, outlining a schematic structure by
which to explore different phases of process. This is by no means presented
as the only, or the uhimate, way of interpreting creative materials but is of
fered in the belief that some kind of universal framework (however limited)
is required."
THE DENIAL OF ORIGINS
We can begin by addressing the way in which process has been overlooked
as an area of literary study, in large part because of a theoretical position
that might be termed "the denial of origins." As is well known, Jacques
Derrida's position in relation to authoriality and authorial intention can

39

TEXT AS PROCESS

best be understood by means of his opposition to a metaphysics of pres


ence" and "logocentrism" that privileges speech over writing by assuming
that it is closer to the origin of meaning in consciousness. Using Saussure,
Derrida shows that speech and writing are both defined by their existence
as part of a system of language based on arbitrary signifiers. There is no
transcendental signifier. Instead, meaning is created by the relationship
of signifier to other signifiers, which also means that it has a limitless
context.
Such an account of language and meaning inevitably brings with it a
denial of authorial origins: "The concept of origin or nature is nothing
but the myth of addition, of supplementarity annulled by being purely
additive."' A "metaphysics of presence" seeks to locate creative origins
in the author s mind, but Derrida is committed to a doubled presence in
which things are determined as much as by what is not there (absence) as
by what is. The individuality of the sign is determined by its difference
from all other signs, which it is not.
In an account of literature that views text as emerging in this way, little
emphasis is given to individual human origins for the work. Instead, there
is only the "trace" as a kind of remainder of origins provided by meaning
in a limitless context with no embodied existence. Derrida states that a
meditation upon the trace should undoubtedly teach us that there is no
origin, that is to say simple origin; that the questions of origin carry with
them a metaphysics of presence." The Australian textual critic Paul Eg
gert, in a paper entitled "The Work Unravelled," argues that a Derridean
concept of "writing" collapses the distinction between subject (material;
documentary texts) and intended object (transcendent; the work) in a way
that is problematic for any study of documentary materials: "[F]or editorial
work to be defensible, subject and object must be separable; only then can
explanations about the construction of the objectat least partly in terms
of the personal agency of the subjectbecome possible."' Eggert seems to
be articulating what I also want to argue here; that Derridean "writing," by
treating all textual material as part of one text, denies any distinctiveness
to text as process and prohibits the full exploration of it in both its actual
and ideal conceptions.
The most directly influential declaration of this denial of creative
agency, for literary studies, comes in Roland Barthes famous essay The
Death of the Author" when he states that "writing is the destruction of
every voice, of every point of origin."' Barthes goes on to declare of the
modern "scriptor": "For him, on the contrary, the hand, cut off from any
voice, borne by a pure gesture of inscription (and not of expression), traces
a field without originor which, at least, has no other origin than lan-

40

THEORIZING PROCESS

guage itself, language which ceaselessly calls into question all origins."'"
The author as God, sending out a message to be accurately deciphered and
understood, is replaced by a decentralized model of language as an open,
ceaseless system. Words are no longer the utterances of a creator but offer
a shared act of participation in a preexistent totality. As with Derrida, the
search for a final signified is replaced by a focus on the instability of the
signifier. From such a perspective there is no absolute meaning, and there
can be no meaning-origins. Instead, Barthes conceives of text as a perfor
mative event, arguing that "writing can no longer designate an operation
of recording, notation, representation, 'depiction' .. . rather, it designates
exactly what linguists... call a performative."" Later, in The Pleasure of the
Text, he asserts, "On the stage of the text, no footlights: there is not, behind
the text, someone active (the writer) and out front someone passive (the
reader); there is not a subject and an object. The text supersedes grammati
cal attitudes."'^ The writer is emphatically "lost in the midst of the text (not
behind it, like a deus ex machina)."^^
The easy adoption of Barthes' famous phrase and the brevity of his es
say have meant that his account has had considerable influence, although
in a contradictory way for genetic criticism. The Barthesian "theater of
"production" appears to have implicitly influenced genetic criticism in
France but to have undermined it elsewhere. On the one hand, such a po
sition seems to centralize process as an essential part of existence (as open
event). On the other, it implicitly opposes the concept of a teleology or
history leading up to the present moment and thus works against any sys
tematic study of materials of process. For Barthes, a productive conception
of text places reader and writer on the same level: "[I]t puts the (writing
or reading) subject into the text."'" Such a thesis seems to elevate process,
but it does so largely in terms of reading and at the expense of writing.
Moreover, because it treats all kinds of engagement with the material in
the same way, it silently omits "productivity" in, or for, a state of textual
process. Perhaps it was because French genetic criticism already existed
before any articulation of "writing" as the space of production that it was
able (partly) to assimilate such ideas. Elsewhere, however, where there had
been no systematic attempt to establish any kind of genetic criticism, such
an account tended to stifle any exploration of process.
' Barthes strongly privileges the present moment in a way that allows the
elision of the duration or sequence essential to understanding composi
tional process. He states:
The Author, when believed in, is always conceived of as the past of his
own book: book and author stand automatically on a single line divided

41

THEORIZING PROCESS
TEXT AS PROCESS

into a before and an after. The Author is thought to nourish the book,
which is to say that he exists before it, thinks, suffers, hves for it, is in
the same relation of antecedence to his work as a father to his child. In
complete contrast, the modern scriptor is born simultaneously with the
text... there is no other time than that of the enunciation and every text
is eternally written here and now.''
Such a position elides process. It involves a denial of will and agency
(agency is now located in language rather than in the user of it) and a de
nial of duration for meaning-production. It does not allow for the concep
tion of writing as a highly repetitive process in which the now is built on,
and exists, only as a result of repeated past "nows." This position is char
acteristic of existential phenomenology, and of a particular conception o
time (as well as language).'
Is it possible to counter such positions and create a space that wdl allow
for the study of textual process in ways that respect philosophical advances
while also retaining space for authorial activity and a teleology of devel
opment? We might begin the attempt by agreeing with Barthes rejection
of a certain kind of criticism that results from an overpnvilepng of the
author. Equally, we can agree that Barthes' account must, rightly, lead to a
refinement of the extent to which creative consciousness is allowed, or un
derstood, to control process (the location of agency as not entirely interior)
and a redefinition of composition in terms of the nature of individual entry
into the language totality. Process must, in part, be about the locating of
the self within, and through, an all-encompassing verbal mode of being.
However, the problem with a fully de-centered account for the study of
process clearly lies in two areas; the denial of duration (which elides the
materiality of texts) and the denial of agency. It seems to me that a study
of process must necessarily retain some sense of human creative agency.
Equally, a temporal model that extends beyond the "now" of the
moment is necessary to fully allow for the nature of creative processes. The
privileging of the "now" asserts the spontaneous act over the necessary re
turn to the "now" to reflect on and regenerate from it. Duration and return,
as essential elements in the motive toward production of the literary work,
demand the presence of creative agency, or at the very least a partial belief
in agency on the part of the creator.
,, r.
One way of responding to the "denial of origins" articulated by Barthes
is to incorporate a more complex account of it from his own later writings.
Ultimately, this can lead us toward the possibility that philosophical and
linguistic accounts of language and meaning need not entirely rule out the

In "From Work to Text" Barthes allows for the (admittedly negative)


possibility of an alternative position to text in which the work is defined as
something closed, contained, institutional, and "in a process of filiation"
identified with a need to "try to find the 'sources', the 'influences' of a
work."" Text by contrast is that which has no center, is open-ended, par
ticipates in discourse. The work is a tangible, fixed object; text is an active
process in itself, a field or space.
Barthes tries to argue that the distinction is not historical (books of an
earlier literary period and aesthetic are works, while the right kind of mod
ern book is a text), but this is only partly convincing in the light of his han
dling of examples in Writing Degree Zero. It is clear, however, that text and
work are not simplistically opposed. As he states elsewhere, "the text is, in
the work, what secures the guarantee of the written object."' In phenome
nological terms, text corresponds to an attempt to conceive of language as
an object in its own right, while work stands for a response to the object as
it relates to human need and use of it. These positions in turn determine
the hermeneutic model applied to each of them.
Responding to Barthes, in "What Is an Author?" Michel Foucault takes
up the problem of the disappearance of the authorial subject into "writ
ing." His distinction between "work" and "writing" follows Barthes' own
distinction between work and text but, rather than assuming the auto
matic privileging of the latter, Foucault is more open in allowing for the
problems created by the denial of authoriality. In relation to the work, Fou
cault makes this interesting statement; "A theory of the work does not ex
ist, and the empirical task of those who naively undertake the editing of
works often suffers in the absence of such a theory."" Such a suggestion is
clearly relevant to this study, which is, in part, about defining a "theory of
the work" to allow for the study of process as a response to textual mate
rial that requires some sense of the authorial. It reopens the question left
hanging in Barthes' essay as to whether material of any age can be both
work and text. If both are allowed, then there is the possibility of placing
emphasis on one kind of nexus or another for the study of textual process
(that of the social discourse of intertext or the more personal discourse of
compositional states and acts).
In his later study The Pleasure of the Text, Barthes' distinction between
"pleasure" and "bliss" partly mirrors the earlier distinction between work
and text. In the poetics of reading offered here by Barthes, reader and
writer are both positioned around the play that is the text. Two kinds of
relationship are envisaged; a false and a true contract. In the first, the writ
ing does not actively "desire" the reader, and what emerges as a result is
"prattle" or a "frigid" text.^ The second relationship is subdivided. The

need for individual participation with the language totality.

42

43

TEXT AS PROCESS

text of pleasure concerns a sequence of utterances, a rapid consumption


by the reader of another's discourse, while that of bliss is about the expe
rience of utterance itself, a slow, full reading, and one in which language
is the reader's as much as it is the author's. The text of bliss is outside,
beyond articulation, uncomfortable: "Text of pleasure: the text that con
tents, fills, grants euphoria; the text that comes from culture and does not
break with it, is linked to a comfortable practice of reading. Text of bliss:
the text that imposes a state of loss, the text that discomforts."^' In The
Pleasure of the Text Barthes explicitly allows for the partial collapse of the
distinction between two kinds of reading (between work and text?) within
the reader who "keeps the two texts in his field" to become "a subject split
twice over" and "a 'living contradiction.'"^^ The ultimate pleasure exists, or
is created by, the tension between such contradictions. The reader's aware
ness of them is, in the end, the delicate essence of existence: "to be with the
one I love and think of something else: this is how I have my best ideas.""
Barthes' presentation of readerly contradiction leads him to acknowledge
that "'pleasure' . . . sometimes extends to bliss, sometimes is opposed to
it" and to consider whether the difference between them is "only a dif
ference of degree."^" Thinking back to the earlier work/text distinction, it
is clear that this is a crucial question for any kind of compositional criti
cism, because if the distinction between them is not absolute, then there is
the strong possibility that we can move between a response to "work" and
"text" for the same object.
It is clear that a study of textual process, of origins, must define it
self partly as work in opposition to text. Indeed, the distinction between
avant-texte and text in French genetic criticism implicitly corresponds
to the work/text distinction in a way that creates some of the contradic
tions informing the discipline. Foucault's definition of writing as not being
about the acts of writing or the meanings of an individual negatively con
firms that it is just these rejected elements (of a work) that need to be in
cluded in a full interpretation of compositional process. At the same time,
a "theory of the work," if this is what is needed to understand process, does
not have to conform to Barthes' original picture of an outdated, closedoff, limited approach. A contemporary understanding of a work must be
colored by knowledge of the existence of text against which it is defined,
so that a return to a simplistic conception of authorial origins is not to
be expected. Rather we can follow Foucault's suggestion to "return to this
question, not in order to reestablish the theme of an originating subject
but to grasp the subject's points of insertion, modes of functioning, and
system of dependencies."^^ Foucault wishes to do this in such a way as to
retain discourse as the center and focus of study: "[I]t is a matter of depriv-

44

THEORIZING PROCESS

ing the subject. .. of its role as originator and of analyzing the subject as
a variable and complex function of discourse.''^ Such a position is unsur
prising given Foucault s definition of the "author-function" in terms of a
self-limiting discourse of control.^' However, a study that seeks to value
and validate the process of creative composition may need to allow more
room to the subject as agent and motive force.
Returning to Roland Barthes, it emerges that, in relation to process it
self, Barthes repeatedly plays off sequence (and implicitly, duration) against
act/experience, so the pleasure of the text "supposes a whole indirect pro
duction" where bliss is "a pure production."^' His use of the term "produc
tion" is potentially misleading, however. Barthes states explicitly: "The text
is a productivity This does not mean that it is the product of a labour
but the very theatre of a production where the producer and reader of the
text meet."2' Barthes' use of the theatrical metaphor can, I think, work to
illustrate the difficulty his theory presents for process, as well as a deliber
ate blindness built into the very heart of that theory A theater production
at the time that it is experienced (on, say, the first night) is a shared act
between players and audience. However, even as we experience it in this
way, we are aware that one reason it is called a "production" is because it
has been produced, over time, by means of repeated rehearsals, individual
preparation, and so on. Thus, the "now" of the first night exists only by
virtue of an extended sequence of acts existing as duration behind and
beneath it. Barthes and other theorists of the text insistently emphasize the
first meaning of immediate active "production" over the second meaning
of temporal extension (which may in a sense still be "present" in the "now"
of the first night but also exists as prior to it and detachable from it). Thus,
when Barthes writes of "process" in "Theory of the Text" he refers to the
active engagement with, and in, something and not to the process of that
process by which such engagement is achievable. Barthes' position allows
only for process as active experience for both writer (writing) and reader
(reading). Yet the writer is also a reader; his or her own reader and rereader,
over time. Might the writer not be able at the same time to know that lan
guage exists outside and beyond his control (in an eternal "now"), and yet
experience it as if it were the expression of a past thoughtand to know
that there is no possibility of authorial intention being the only meaning
in a text, and yet to write as if there were? In other words, the writer does
not, cannot, only experience text as "Text" but has to move between a con
ception of writing as work and text, responding to it as a grounded object
being produced as well as an ideal one.
In The Pleasure of the Text, Barthes asserts the necessity of readerly
pleasure being created by contradiction. Describing his own response to

45

TEXT AS PROCESS

THEORIZING PROCESS

tragedy, he states: "I take pleasure in hearing myself tell a story whose end
I know: I know and I don't know, I act toward myself as though I did not
know."' This last position, "I act toward myself as though I did not know,"
is particularly useful in relation to both the reader's and the writer's rela
tionship with language. Barthes writes: "Many readings are perverse, im
plying a split, a cleavage. Just as the child knows its mother has no penis
and simultaneously believes she has one... so the reader can keep saying: I
know these are only words, but all the sameP^ We need to consider whether
this contradiction is capable of application not only to the reading process
but also to the writing process, and to the writer as his or her own reader. In
fact, the contradiction might then be capable of application in three ways:
to the reader of the text (reader-text nexus); to the writer as reader and
writer of a text (the writer-reader nexus); and to the reader of the writ
er's draft materials (reader-process nexus). The distinction between what
is known and what is felt is crucial because it allows for the possibility in
language that what is true (metaphysically) may not be felt to be true by
those experiencing it, and, further, that the misapprehension itself may be
essential to the production of the work of art (or ability to enter creatively
into the language totality). Something similar seems to be advocated by
Jerome McGann when he asserts: "We read in the same spirit that the au
thor writ, or at least we try to; and thenthere and thenwe also read not
in that spiritwe read in different spirits. We are right to do these contra
dictory things because the texts themselves are, ab initioand as the poet
(Byron) said'antithetically mixed' themselves."'^
This kind of tension, of both reader and writer capable of being held in
simultaneous, yet conflicting states, might also be allowed for the readersubject in relation to compositional material. The state of "I know and I
don't know" could equally apply to the reader engaging with the authorial
process of "death through writing" or, the "coming-into-being" of a text.
In relation to compositional process, it is capable of creating space for a
necessary authoriality without that authoriality becoming the dominant
reason and motive for interpretation.

directly associated with an outmoded conception of the author as controlhng presence and of writing as an expressive act by a unique individual,
in favor of a giving up of the self to a preexisting language totality. Inten
tion is discredited primarily because of its association with a search for an
absolute fixed meaning behind the text rather than an understanding of
meaning as a shared experience.

WHAT'S WRONG WITH AUTHORIAL INTENTION?


As "the denial of origins" debate illustrates, early twentieth-century at
tempts to validate the discipline of literary studies by means of objectivity
and a search for normative meaning are replaced in the mid-to-late twen
tieth century by an approach, underpinned by phenomenology, that views
both the reading and writing of literature in terms of an "event." Such a
position inevitably brings with it a rejection of authorial intention, which is

46

In the Anglo-American literary-critical debate, the problem with in


tention was first raised in Wimsatt and Beardsley's famous essay "The In
tentional Fallacy," which I discussed in chapter 1. This essay argued that
' the "fallacy" lay in a critical approach that assumed a vital relationship
between authorial meaning and the meaning of a text. Instead, the essay
stated that "the design or intention of the author is neither available nor
desirable as a standard for judging the success of a work of literary art."" In
a later restatement of this position, Wimsatt made clear that such a fallacy
applied to interpretation as well as evaluation.'" A secondary argument
made by Wimsatt and Beardsley was that of the extent to which, and the
means by which, such intention could be revealed: "[T]he closest one could
ever get to the artist's intending or meaning mind, outside his work, would
be still short of his effective intention or operative mind as it appears in the
work itself and can be read from the work."' This is an interesting com
ment since it makes a distinction between intention in the mind and inten
tion in acts on the page, here defined as "effective intention." Ultimately,
though, Wimsatt simply wants to assert the primacy of intrinsic meaning
in the text itself over that of the writer.
We can move on, beyond literary-critical rejection of authorial inten
tion on New Critical grounds, by turning to more recent philosophical ar
guments. The rejection of authorial intention on the grounds of its interpre
tative shortcomings is argued very persuasively by Hans-Georg Gadamer
in Truth and Method. In this work he looks back to nineteenth-century
interpretative practices, particularly those of Friedrich Schleiermacher, to
argue against a hermeneutic reconstructive model based on creative ori
ginsthat is, a model in which "[h]ermeneutics endeavours to rediscover
the nodal point in the artist's mind that will render the significance of his
work fully intelligible, just as in the case of other texts it tries to reproduce
Ae writer's original process of production."' Such an approach is emphat
ically rejected by Gadamer: Ultimately, this view of hermeneutics is . . .
nonsensical."" He views such reconstruction as "handing on a dead mean
ing" and his own arguments in favor of the work of art as active experience
and "event" emerge in opposition to such nineteenth-century principles."
Rather than a contrast existing between "art" and "life," Gadamer insists
that art is life: "[T]he work of art is not some alien universe into which we

47

TEXT AS PROCESS

are magically transported for a time. Rather we learn to understand our^


selves in and through it.'"' Gadamer insists that meanmg is not unbodied
in the work itself, or in the mind of the author, but in each individual s
engagement with the work from the perspective of their own developing
understanding: "[W]hat is formed by the poet, represented ^7 ^e ac^
and recognized by the spectator is to such an extent what is meant
fwSh. Significance of representation Ues-tha.^

or the actor's prowess as such are not foregrounded from it. The philo
sophical concept of the work of art as an event denies value o agency
means by which a work comes into being. As Gadamer Puts^^^ Tote
diation means that the medium as such is superseded. Thus,
herme
neutics a parallel position to that found m existential phenomenol gy
in the debate over the "death of the author" is found to exist, leaning is
to be located in the activity of engaging with a text and making he expe^
rience of it part of our selves and cannot be seen to exist in an attemp to
understand the creative mind that generated that text. The mtentiona i y
of the writer's mind is rendered nonessential for a full response to the liter
ary work, which is now allowed to stand as an open space ready to unclose
'turrpoXlmerges from Heidegger's account of being in Bdn,
and Time and is therefore (unsurprisingly) in agreement with the account
of language that Heidegger gives and which, again, has implications for
weight and value allowed to creative intentions. As he does for other ph
nomena, so for language too, Heidegger seeks to respond to it no in terms
of its human use value but in respecting it as a thing apart. Thus, he as .
"'In what way does language occur as language?' We answer: Language
speaks:"' Traditionally, language is viewed as a
ceeding from man and representing inner feelings and thoughts m exter
nalized form. From this viewpoint, "[w]hat is spoken in the poem is what
the poet enunciates out of himself What is thus spoken out, speaks by
enunciating its content. The language of the poem is a manifold enun
ating Language proves incontestably to be expression."" Heidegger sums
up the traditional position: "No one would dare to declare incorrect, let
alone reject as useless, the identification of language as audible utterance
of inner emotions, as human activity, as a representation by
concept.- In phenomenological terms, however,
ception of language distorts by viewing it merely as man s tool Authentic
existence requires that the individual no longer views himself as ^gent but
now has a role in terms of "responding" to the preexistent language total
ity The result is to overturn an expressive conception of man s '"elationsh p
to language. Thus, "Man acts as though he were the shaper and master of

48

THEORIZING PROCESS

language, while in fact language remains the master of man. When this
relation of dominance gets inverted, man hits upon strange maneuvers.
Language becomes the means of expression. As expression, language can
decay into a mere medium for the printed word."*'
Both Heidegger and Derrida are sympathetic to the ways in which writ
ers want to feel that they do possess the language with which they work.
Derrida makes clear the way in which the nature of speech leads us toward
a personal identification with it: "It produces a signifier which seems not
to fall into the world, outside the ideality of the signified, but to remain
shelteredeven at the moment that it attains the audiophonic system of
the otherwithin the pure interiority of auto-affection."" Derrida con
stantly articulates the doubleness of an authorial understanding of lan
guage and meaning (discourse) as opposed to the meaning totality (lan
guage as system)indeed, one purpose of deconstruction is to bring to
light the inevitable contradictions between these two conceptions. Thus,
on the one hand, he asserts that "the person writing is inscribed in a de
termined textual system," but on the other, he is interested in the ways in
which authors deliberately resist a full recognition of that fact."' Nicholas
Royle states that "the logic of the supplement dictates . . . that the writer
is always susceptible to being taken by surprise. A writer can never have
complete command or mastery over what s/he writes."*' The phenomeno
logical denial of origins to authorial consciousness is therefore not strictly
a denial of intention, but it must act to radically redefine the nature of in
tention, which can no longer simply be about the externalization of inner
thought.
Gadamer's account of meaning, and Heidegger's and Derrida's of lan
guage, confirm two ideas with significance for the study of creative pro
cess: first, that the intention of the mind that creates cannot be assumed to
provide an understanding of the thing created; and second, that the con
cepts of intention, will, externalization, and expression in relation to lan
guage and meaning form part of a limited understanding of the world. The
creation of an "authentic" work of art occurs by giving up the will, not by
asserting it."' Such a position, if it does not altogether discredit intention,
certainly has the effect of dramatically marginalizing it.
This still leaves us with a paradox, however: how can even those who
write against intention do so without having the intention of doing so? To
deny intention altogether to creative engagement with language seems to
be at odds with our own experience of such activity. This in itself raises the
possibility of a response already touched on. One explanation of the para
dox must be that whether or not language really exists as a tool over which
we have mastery, our experience of using it is as though this were the case

49

TEXT AS PROCESS

THEORIZING PROCESS

at the time of use. Even if it is philosophically true that "language speaks,"


this is not how we experience it, and how we experience it is a crucial part
of understanding creativity. So, although language as a thing existing be
yond and above individual use denies agency, an individual sense of agency

need to be a necessary element of our understandingas they are not for


the completed work of art. This is not to ignore altogether the phenome
nological redefinition of language and meaning as essentially external and
preexistent but to suggest that we may need to respond to them in terms
of our experience of them (internal, expressive) as well as in terms of their
separate existence (external, prior to consciousness).

may be necessary in order to create.


In conclusion, we could say that that there is everything wrong with
authorial intention when:

PROCESS, INTENTION, AND A CONTEXT FOR MEANING


1. It is assumed to be the primary purpose of interpretation.
2. It is understood to represent a normative absolute meaning held by
the text.
3. It is assumed to be a single, static, identifiable thing.
4. It is thought to be capable of complete reconstruction.
5. It is based on an understanding of language as purely expressive.

It is easy to agree that intention should not be used as the fundamental ba


sis for interpretation of a work of art and that the truth of that work must
exist apart from an understanding of its maker. However, the question re
mains whether this is true in the same way, and to the same extent, for the
interpretation of the coming-into-being of the work of art (i.e., for the draft:
materials of textual process). In interpreting process, our interest is, at least
in part, in the way in which meaning comes about. We are no longer an
alyzing language merely in a context of absolute detachment from con
sciousness but in the (sometimes clumsy and awkward) participation of
being with language. When the focus of study is not the "final" text but the
draft materials, then authorial intentional acts, embodied on the page, may

It is still not clear exactly what is meant by "authorial intention," and it is


necessary to pursue this further. The term "intention" encompasses two
kinds of meaningor two dimensionsthat are often conflated and con
fused.'" First, there is a common understanding of intention in terms of
everyday action: behaving in a consciously determined way that has con
sequences in the world. This kind of ordinary intention is understood in
terms of a mental cause resulting in a physical effect. Its importance is
made clear from the reliance of the law upon it. Legal intention draws on
such a meaning of intention to determine the degree to which individual
consciousness can, or should, be held responsible for its actions (and pun
ished accordingly).
Philosophical accounts of action also use intention in this sense but
make a further subdivision between self-conscious deliberate intentions
and intention as a part of the action. The distinction is first made by
G. E. M. Anscombe in her early work on intention in which she distin
guishes between the intention with which a man does something and what
he actually does: "In general we are interested, not just in a man's intention
of doing what he does, but in his intention in doing it."" This is a distinc
tion between a preplanned and internally anticipated event and the more
immediate putting of that aim or purpose into practice through action. I
will return to it later in this discussion.
A second meaning of intention concerns the more specialized use of it
in terms of phenomenological "Intentionality," particularly in the work of
Edmund Husserl. The term is used to describe a certain state of being that
is concerned with the way consciousness is directed toward an object. The
state itself (belief, desire, fear) is the primary focus of interest rather than
thd object of that state (which may or may not even be allowed to exist at all).
Intentionality in this sense denotes a property of being that distinguishes
the way in which humans respond to the world. It could also be understood
in terms of a state requiring satisfactionso that intention can be satisfied
or fulfilled (or not) for the intentional consciousness. It differs from phil
osophical accounts of intentional action in that it is concerned primarily

50

51

However, it seems perfectly acceptable to draw on authorial intention


when:
1. It is understood by the creator as a motivating force for the cominginto-being of a literary work (without necessarily being fully under
stood by consciousness).
2. It informs our understanding and interpretation of the manuscript
page as one dimension of it, along with other nonintentional mean
ing contexts.
3. It is understood to be dynamic and contradictory (forming part of a
constantly changing stream of meaning).
4. It is understood to be capable of reconstruction primarily in terms of
acts on the page.
5. It is based on an understanding of language as both expressive and
experiential (performative).

TEXT AS PROCESS

THEORIZING PROCESS

with an inner state relating to consciousness and not with causal links be
tween mind and act. It follows from such a position, considered in relation
to language, that language is understood to be referential, based on acts of
consciousness, with an inner state giving meaning to a linguistic act."
Opposition to an intentional account of being in terms of both causal
events and Husserlian intentionality has already been touched on in dis
cussions of literary meaning and language. The later phenomenological
rejection of intention as the core constituent of consciousness is based on a
holistic conception of Being that denies a divide between mind and world.
Instead, it is asserted that much of our activity in the world, including men
tal activity, does not involve conscious intentions at all. We might act in a
way that appears causal, but in fact we are simply participating in a larger
whole that does not demand deliberate actions and is therefore not pre
ceded by internal cause. In this account, meaning is grounded in a shared
subject matter that preexists the creator. Textual meaning is understood to
change those who participate in it and to exist not in consciousness but in
engagement with the world.
Located somewhere between the two first understandings of intention
comes speech act theory and John Searle's work in Speech Acts and Inten
tionality. We need to look at such ideas more closely since they do provide
a useful way of thinking about intentional acts on the manuscript page.
Before Searle, J. L. Austin's influential 1955 lecture series, published as How
to Do Things with Words (1962), made clear the extent to which language
is performative by distinguishing between words as utterance (locutions)
and the way in which utterances also perform illocutionary (nonverbal)
acts. Since "to perform an illocutionary act is necessarily to perform a locutionary act," it follows that meaning is understood to encompass not just
the verbal content of an utterance but also the context of its performance.'^
Searle takes such ideas further in Speech Acts when he defines the illocu
tionary act in terms of a preexisting background of shared conventions so
that "speaking a language is performing acts according to rules."'^ In Speech
Acts Searle acknowledges, "It might be objected to this approach that such
a study deals only with the point of intersection of a theory of language
and a theory of action."' gy treating the use of language by consciousness
as a performative event (action through words), speech act theory resituates intention as embodied meaning in the world (i.e., an expressed utter
ance can be understood as the fulfillment of an intentional act).
Searle does not explicitly situate his account in relation to phenomeno
logical intentionality, although it is clear that much of his work is indebted
to Husserlparticularly in the account of intentionality as "directedness"

toward an object capable of being satisfied.' Instead, in Searle's own expla


nation, intentionality can best be understood by means of direct compari
son with speech acts. The entire speech act involves a locutionary utterance
in the context of shared conventions and rules that are illocutionary. The
locution bears with it the force of an illocution. Speech acts are capable of
being understood as "intentional performances" so that the performance
of a speech act is the expression of an intentional state."
Searle's account of intentionality works to bring together two principal
elements: a state of prior mental directedness (intention of doing) result
ing in ordinary actions (intention in doing). Searle states, "Intentionality is
directedness; intending to do something is just one kind of Intentionality
among others."" Intention "in doing" can occur with or without a prior
directedness. This has the important consequence that all actions can be
defined as partaking of intentionality in one way, although they may not
all'involve self-conscious directedness. Searle's distinction between "prior
intention" and "intention in action" emerges from this.'' Prior intention
consciously causes intention in action, which in turn causes and is bound
up with a bodily movement that will result in the satisfaction of the inten
tion. Searle states: "We say of a prior intention that the agent acts on his
intention, or that he carries out his intention, or that he tries to carry it out;
but in general we can't say such things of intentions in action, because the
intention in action just is the Intentional content of the action; the action
and the intention are inseparable."" Searle considers carefully the means
by which intention is bestowed upon action; it operates through the need
to satisfy the intention "by intentionally conferring the conditions of sat
isfaction of the expressed psychological state upon the external physical
entity."' His account is useful for underpinning an intentionalist posi
tion that wants to present the caseas he doesthat "there are no actions
without intentions" and that there is no meaning without intention.^ This
can be achieved if all language is a kind of act, and all action is intentional
(though not necessarily involving conscious intention). I do not want to go
so far, however. This study deals with intentionality as part of a causal or
teleological account of human actions that offers one way of responding to
language and meaning but not the only way.
Such ideas can be translated back into textual terms with the aid of
Peter Shillingsburg's distinction between "intention to mean" and "in
tention to do" (a distinction I draw on at a fundamental level throughout
this book). In the chapter titled "Intention" in Scholarly Editing, Shillings
burg emphasizes the need to give a more complex account of intention,
in Ways that strongly anticipate my work here, when he states: "Theorists

52

53

TEXT AS PROCESS

have tended to think of authorial intention as having a single goal. They


have tended to de-emphasize both the development of intention through
stages, towards completion, on the one hand, and the change or contradic
tion of intentions on the other."" This leads him to develop a distinction
between two different concepts of intention. The first of these, intention
to mean," is defined in terms of fluctuating and changing authorial inten
tions that are experienced internally and therefore not recoverable (and
contain unconscious elements in any case)."The second, "intention to do,
is defined as "an intention to record on paper, or in some other medium, a
specific sequence of words" that is "almost completely recoverable."' The
latter form is held in acts on the page undertaken by the author and is
therefore far more capable of interpretative reconstruction. In sum, Shil
lingsburg states: "The intention to do is, with the three exceptions noted,
[scribal errors, "Freudian slips," shorthand elisions] conclusively recover
able from the signs written; the intention to mean is inconclusively recov
erable through critical interpretation."'
For the purposes of this book, the question to be asked is simple. To
what extent is an intentional context a necessary context for the interpreta
tion of meaning? I want to suggest that we can agree with Heidegger. Ga
damer, and Derrida that it is not a necessary context for the interpretation
of a completed work of art but to acknowledge that it is a necessary context
for the interpretation of the coming-into-being of a work of art. An un
derstanding of textual process has to recognize that its meaning is in part
concerned with the making of meaning. So, for the "completed work of art
presented to the public, we can accept an ontological definition of mean
ing in terms of the open experience of the text, requiring no knowledge
of origins, but when we direct ourselves toward textual process, then the
experience of the object is also the experience of the making of the object
(however limited or contradictory that experience might be). As such, it is
necessary to respond to process both as partaking of the openness of lan
guage and as a sequence of intentional acts on the page that bring the object
into being. We do so not because we seek to understand the mind of the
creator but because we seek to understand the kind of meaning that exists
when language is used as a tool for creative making, as well as understand
ing process as it participates in the openness of the completed work.
Finally, it is important to note that the relation of the reader to textual
process is of a different order than that of the reader to the completed text,
and this is also true in relation to creative intentions. When we read the fi
nal work, we are free to respond to the textual object in terms of whatever
meaning we find in it. When we read the materials of process, however, it is
clear that such materials were (probably) not written with the expectation

54

THEORIZING PROCESS

that they would be read. This positions the reader very differently in rela.tion to intention as an element of interpretation. Rather than intention be
ing clearly directed toward an audience to lead to a certain reading of a text
(in Searle's terms, "the Intentional state expressed"), it is directed, at least
in part, at the production of meaning, in other words, at itself ("the inten
tion ... with which the utterance is made"). Thus, the nature of intention
involved in the coming-into-being of a text is potentially of a different kind
than intention capable of being satisfied. As such, it is no longer marginal
to our understanding of the materials but becomes an essential part of it.
The writer's intention within process is directed toward understanding his
or her own intentions (which may, of course, never be fully achievable). We
need to allow that the making of something through language appears to the
writer to operate on the basis of speech acts (as the satisfaction of individual
intention through acts on the page), while the meaning-content of language
remains part of an open directednessexperienced in an uncovering that
has no limits and can never be fully satisfied.
I want to argue that the making of meaning is in large part about recog
nizing, understanding, and redirecting meaning in an interplay of intended
(planned) and unintended (spontaneous/unwilled) meaning. A pure phenomenologist might argue that such an account merely retrospectively im
bues a "nonintentional" object (draft text) with human intentionality. This
is because an account of language as a system detachable from the living
beings who participate in it (and essentially detachable by its very nature)
leads to an account of intention as something that is retrospectively read
into, or onto, the text. Geoffrey Bennington explains: "Insofar as you are
inclined to attribute intentions to me ... you construct them retroactively
on the basis of the text read, and the text read functions 'mechanically', in
dependently of the intentions you attribute in fact, after the fact, to its sup
posed author."' Such a concept of intention defines it as reader-produced
and suggests that attempted reconstruction of it has severe hermeneutical
limitations. I want to argue, however, that some space must be allowed
for such reconstruction, and further, that there is a distinction between
readerly reconstruction of the intentions of another and a writerly recon
struction of his or her own intentions in the act of creating, both of which
are of interest to the study of process. Reader and writer are both retroac
tively placing intention onto text, but the difference is that for the writer,
this is not just a question of interpretation or self-interpretation but part of
an active process and event: a sequence of acts will follow from that reappropriation. For the writer at the time of writing, unlike for the reader, the
draft text is still open to change so that the reading of intentions into it also
results in acts on the page and changes to the language. I accept that there

55

TEXT AS PROCESS

are limits to the retroactive reconstruction of such intentions on the part


of the reader (and that we can still only access a writerly experience indi
rectly). Nonetheless, intention as a complex and changing sequence of acts
must emerge as having a considerable role to play in our understanding of
text as process.

3.
TOWARD A COMPOSITIONAL METHOD

Books of rules can prove delusive guides.


Philip Gaskell, From Writer to Reader
PREINTENTIONAL AND NONINTENTIONAL MEANING
It emerges from chapter 2 that there must be three contexts for meaning
that bear on the understanding of literary composition and the text in a
state of process. These can be defined as follows:
L The Compositional Context (transparent, preintentional)'
Writer's own history, cultural knowledge, philosophical knowl
edge, literary knowledge, and so forth
History of the world at the time of writing
Objects around the writer
People around the writer
People engaged in the production of materials
People engaged in the publication of materials
2. The Intentional Context (deliberate, intentional, unconscious,
unintentional)
Context of authorial meaning as understood by the writer through
his own past acts
Developing narrative of composition
Meaning on the page understood sequentially/teleologically as acts
of consciousness occurring over time
Causal relation between mind and word on the page, and between
thought and action

56

57

TEXT AS PROCESS
. Unintentional (accidental) meanings occurring as a secondary
consequence
Unconscious meanings and acts on the page
3. The Language Totality (unwilled, nonintentional)
. Ontological whole that allows the literary work to be brought forth
from it
. Totality with which the writer already unknowingly participates
(into which he or she enters and that preexists him or her)
. Sense in the writer of giving him or herself up to the totality
(unwilled openness)
. Emergence of the work of art as an open work of truth accessible
to all
I want to argue that creative process emerges as a result of the writer mov
ing across and between these three contexts. To understand the process
fully therefore, we need to understand it not from an exclusive approach,
dwelling only within one of them, but in terms of the interrelations of cre
ative intention with different kinds of undeliberate acts. The first two con
texts might be understood as preparatory conditions for a creaUve giving
up of the self in the third but, equally, the second context provides a nor
mative structure within which the writer feels himself in control, and upon
which the other contexts bear.^ Thus, there is a dynamic between what the
writer wills and intends and what is unwilled (and allowed to be unwilled),
which drives forward the creative process. This dynamic could also be
viewed as the writer's relationship to the world (on one side through t e
compositional background) and to language and meaning (on the other).
Only the interaction of one with another-the event field held within a
certain frameworkcan bring the creative work into being. It is because
of this movement between a core intentional structure and the constant
engagement of this structure with preintentional and nonintentional con
texts that process must be viewed as a constantly shifting state, resulting
in changing acts on the page, acts that are always potentially subject to
further change.
.
The method, thenwhich is really a suggested framework for interpre
tative activity-is explained here in terms of a core sequence of intentional
states and acts, necessarily offset by nonintentional contexts on either side.
Creative intentionality is enclosed hy a compositional context within whic
intentional acts occur, but it also "contains" the potential for an escape into
the limitless context of the unwilled totality to which it is ultimately sub
servient. This allows us to respond to the materials of process both referentiallytreating the manuscript materials as an intentional object bearing

58

RECLAIMING PROCESS
meaning in the form of intentional acts on the pageand ontologically,
since those same materials also exist as part of the open truth of the work
of art and of the individual in response to it. The framework suggested
here as a way of thinking about compositional material as it develops over
time is intended to have universal application insofar as all writers must
work with, and through, such intentional states if they are to create. It is
emphatically not intended to be prescriptive or to be used as a rigid, regu
latory structure.
We can begin by clarifying different ways in which the intentional core
of'apparently controlled self-expression is constantly engaging with, or al
lowing space for, other kinds of meaning of which it is not fully aware.
Where creative intention occurs as a (normatively) teleological structure
and sequence, preintentional and nonintentional meanings are of a differ
ent kind altogether. First, then, we can address the preintentional composi
tional context. This is necessarily present as part of the extended existence
of the individual who creates, but that individual may be unaware of such
a C9ntext and so will engage with it to achieve everyday objectives without
any conscious intention taking place. Such a concept corresponds to a phe
nomenological account of how we interact with our daily environment. In
^eing and Time Heidegger asserts that we understand the things around
us to exist in terms of a "totality" of interrelated objects, all of which refer
to some human task: "Equipment... always is in terms of its belonging to
other equipment: ink-stand, pen, ink, paper, blotting pad, table, lamp, furnilfure, windows, doors, room."^ When we use a single object within that
totality, we are not at all concerned with its separate nature but only with
it as something "ready-to-hand" and forming part of a work on which we
are focused. The totality accounts for physical objects but also for more
abstract entities such as language and meaningwhich we can then rec
ognize to be communicable by means of shared social and cultural con
ditioningand for skilled acts we already know how to do and so can do
"without thinking." Acts can thus occur that look like the product of will
but may have been undertaken without conscious deliberation. Unplanned
or spontaneous acts may also occur in response to unexpected events im
posed on us by the equipment totalityand these might also affect com
position as a result of interruption or other external factors. Finally, nonauthorial intentions of various kinds will also come into play by means of
the compositional context. These might be unnoticed at a level of domestic
support, reassurance, or even practical assistance. As long as that support
is contributing positively to the development of the work it will largely be
taken for granted, but if it should start to affect it negatively, then the back
ground ceases to function as an invisible support-structure.'*

59

TEXT AS PROCESS

So, to take the example of John Keats writing the sonnet "On Sitting
Down to Read King Lear Once Again," it is immediately apparent from the
title that such a poem could not have been written without Keats having
previously read King Lear. Indeed, in a letter to his friend Benjamin Bailey
on January 23,1818, the poet presents the writing of the poem in terms of
an intertextual cause-effect relationship: "I sat down to read King Lear yes
terday, and felt the greatness of the thing up to the writing of a Sonnet pre
paratory theretoin my next you shall have it." A context of charged im
mediacy and response is further enhanced by the material location of the
fair copy of the poem "opposite the first page of King Lear in his facsim
ile reprint of Shakespeare's First Folio." However, this prior contextto
which the poem self-consciously directs usis only the most obvious selfdetermined one. We might also consider that Keats couldn't write a sonnet
without knowing what a sonnet was, or without being aware of his ability
to articulate his feelings in poetry, that he couldn't do any of these things
without being able to read and write in the first place and, beyond that,
without having some conception of the value of language and meaning as
fundamental elements of human communication. All of this preexisting
background has to be in place for the creative act to occur, so that even the
most apparently "spontaneous" act of composition occurs within a frame
work that both writer and reader tend to take for granted and not even to
notice. An everyday ability to participate in creative process demands this
compositional background.
When Keats encloses a copy of the sonnet in a letter to his brothers the
next day, he gives a clearer account of its origins: "Nothing is finer for the
purposes of great productions, than a very gradual ripening of the intel
lectual powersAs an instance of thisobserveI sat down yesterday to
read King Lear once again the thing appeared to demand the prologue of a
Sonnet, I wrote it & began to read."' The earlier letter had seemed to sug
gest a relationship between background and creative act in which reading
led to writingbut in fact this is not what the letter actually does describe.
Rather, the relationship is between a past reading and an anticipated future
one, but it is the intermediate state of anticipation (between one reading
and another) that stimulates the new creative act. In this case, the com
positional context is not merely "background" preparation. Instead, Keats
seems to consciously manipulate the "background" so that it becomes ex
plicitly bound up with the process and the content of the writing (which,
in the case of this poem, is itself about that process).
On one side of the embodied intentionality of the materials of process
there is a preexistent compositional context. On the other, there is an un
willed giving up of the self to the language and meaning totalities. Here,

60

RECLAIMING PROCESS

the writer no lonpr conceives of him or herself as a creative agent "producing an externalized representation of inner consciousness. Instead, such
distinctions fall away as the individual participates in a preexistent holism
It can only be in the ongoing moment of active composition that the writer
4as the potential to open him- or herself up to language and no longer to
feel as if m control of process. Such moments constitute the experience
of inspiration," since identity is lost and the self is felt to be possessed in
some way outside it. Thus, the "spontaneous" acts as a kind of gateway out
of-the preintentional and intentional states of being and into something
other Such a nonintentional experience can only occur in an individual
capable of operating intentionally, but also (crucially) of willingly giving up
that mtention. Once experienced, however, nonwilled Being asserts itself
as prior to an intentional directed sense of self, and of a higher order.
CREATIVE INTENTIONALITY
AND UNINTENDED MEANING
Having understood the ways in which preintentional and nonintentional
contexts are essential to creative process considered as a whole, I want fi
nally to return to the definition of intention in terms of a teleological core
of self-conscious and unselfconscious intentional actions resulting in a
kind of history of process on the page. If we respond to the materials of
process in terms of applied speech act theory, then we can break process
down into constituent states and acts, the core elements of which are em
bodied on the manuscript page. This allows for a kind of microanalysis of
those acts and of unconscious and unintentional acts that come into exis
tence as a by-product of intention within the materials of composition At
the same time, creative intentionality should be understood as a complex
ot states and acts continually subject to each other and capable of both co
existence and conflict. Creative intention is always open intention for this
reason: intentional elements are continually subject to other intentional
elements.' This distinguishes it strongly from the concept of authorial in
tention within a completed work (which exists only as one possible reading
ot the meanmg of the text among many others). Intention itself is an ongo
ing event within creative process, constantly being changed and redirected
by the unintentional contexts with which it engages. In effect, it operates at
the level of a motivating framework (itself supported by a prior unnoticed
tramework) within which the author feels able to "let go."
In an earlier paper I outlined an intentional model emerging from
Michael Hancher's paper "Three Kinds of Intention." "> I am indebted to his
work for the way in which it defines different kinds of intention, although

61

TEXT AS PROCESS

it holds a more strongly intentionist position than my own. Another excel


lent discussion of intention as a shifting complex of states is given by Peter
Shillingsburg in Scholarly Editing, and his account also anticipates mine
We can begin by acknowledging the likelihood of some kind of holis
tic aim or programmatic intenh'onparticularly for any work of length.
Such intentions might be internal (involving pre-textual composition), or
they might be formulated in notes or letters as some kind of plan. Pro
grammatic intention is only ever going to provide the broad framework
for a work, but it also probably represents the writer's wider ambitions and
could be viewed in terms of a "challenge" the writer sets for him or her
self (and may fail to live up to). This "long-term" intention remains open
for the entirety of the creative process, is subject to redefinition, and may
never be fulfilled. Also present before writing (as part of this program
matic intention) is an anticipatory communicative aspect, an awareness of
anticipated audience."
Second, it is possible to identify stages of partial fulfillment of localized
writing, which we might term contingent intention}' I would view contin
gent intention as combining a series of discrete intentional acts (intentionas-process) with a sense of those acts as part of a sequence or section of
work (either of a single work across drafts or of a section within a larger
work). Intention is contingent in that, although a short-term intention may
have been satisfied, its fulfillment and its value within the whole work re
main dependent on other parts of the process and the wider context of the
developing work and cannot be known until later. This is an ongoing state
in which issues of priority, such as which version to privilege, are not yet
active: a "holding" state. Such a concept is particularly important for a long
work, representing stopping points that are nonetheless only provisional
and known to be so by the mind that creates them. It must also be borne
in mind that contingent intention is occurring in conjunction with acts of
"spontaneous" (immediate) composition.''
If we follow John Searle in allowing that "every action has an inten
tion in action as one of its components,"'^ then we can also designate a
kind of micro-intentionality, or intention-as-process within continent
intentionand this corresponds more to Anscombe's "intention in doing
rather than "intention 0/doing." Alongside acts of spontaneous composi
tion there must exist descending levels of intentional activityfrom the
writer's intentions for a particular day. or hour, down to his intentions m
relation to a line, or phrase, or word. This intention-as-process involves
both the physical act of lifting a pen and applying it to paper and the men
tal intentions involved at a word-by-word level. Micro-intentionality thus

62

RECLAIMING PROCESS

combines self-determined actions with a preintentional "know-how,"


drawing on the compositional context.
When we consider these micro-intentional acts in the immediate
context of contingent stages, a distinction seems to emerge between a
willed intention to do or make, without knowing exactly what, and a sub
sequent intention to communicate a certain meaning. This is a distinction
between intention as motive force (to write something) and as meaning
force (to communicate a particular idea). Alongside these, the actual act of
creative composition occurs without any conscious intention. Contingent
intention and intention in action depend on a relationship between giving
oneself up to meaning and attempting to reassert control of it. Contingent
intention therefore describes an interaction between immediate writing
in an unwilled state, in which there is no need of conscious intentions,
and a subsequent intentional return to that writing to achieve a sense that
meaning-intention is satisfied through it.
One other consequence of the contingent nature of creative composi
tion, particularly for a long work written over time, is the possibility of a
failure of motivation occurring as a consequence of the need to pause and
recommence labor on a work. There are numerous accounts of poetic nonf ompletion for the Romantic periodwhen the acceptability of presenting
an incomplete work to the world emerges for the first timeand clearly,
the study of incompletion and "creative failure" (or the perception of it) is
a potentially rich area of investigation for compositional criticism. Asked
why he had never completed "The Recluse," Wordsworth is recorded to
have remarked to an American visitor, George Ticknor: "Why did not
Gray finish the long poem he began on a similar subject? Because he found
he had undertaken something beyond his powers to accomplish. And that
is my case.""
If the intentional core is structured on provisional and partial satisfac
tion for the writer (experiencing process as agent), then we also have to
consider whether that satisfaction is finally achieved in terms of a total
state of final intention. One question that the study of the literary work
from' the direction of the text as process allows to be asked is, how "final"
is the published text? I would question whether such a concept really exists
for the writer, particularly in works of poetic length. In a brief discussion
of Alexander Pope's "An Essay on Man," Richard Wendorf raises this issue:
"If, as Valery suggested, 'a poem is never finished, it is only abandoned,'
then authorial manuscriptseven more than their printed counterparts
should provide us with the materials for a poetics of abandonment. Such a
poetics would be based of course, on a principle of imperfection."' There
is, certainly, the first presentation of a work to the public, which means

63

TEXT AS PROCESS
that an endpoint of some kind is achieved and that a sense of fixedness
and finality attaches itself to the work by virtue of its material form and
reception. Without question, this brings into play a whole train of specific
activities and anxieties for the author that are embodied in the composi
tional material in different ways. However, the finality represented by the
moment of publication is, in effect, imposed externally. The writer might
have gone on changing things, but now, time constraints, the fixing of type
on the page, and other physical and practical needs determine the text in
one form, the final' form of the first published text. For the writer, a sense
of finality may be far more tenuous than for the publisher, printer, critic,
and reader. Compositional material contains the potential and possibil
ity for many different kinds of poem, not just the one the world knows.
Of course, various decisions led the poet to create this text and not that
one, and those decisions were unlikely to have been entirely arbitrary, but
the text is something more than the final productas the very survival of
compositional material illustrates. For the writer, I would suggest that the
idea of final intention is really only one possible stopping point in the con
tinual process of contingent intention through which the material evolves.
Potentially, such a process is endless, and for this reason, the writer may
well go on changing the text after publication and right up to the end of
his, or her, life.
Finally, we should briefly consider the concepts of unfulfilled intention
and revised intention, which may or may not be allowed to exist at all (and
which have the danger of assuming an overly linear/teleological structure
for creation). The first describes a state similar to that of programmatic
intention, existing at a distance from the period of core creative activity
but occurring at a different moment in time within the compositional pro
cess. At some later pointpossibly after the publication and reception of
a workthe writer is forced to acknowledge that his original ambitions
cannot be met because of the way the material itself has emerged. This may
well result in future action, through revision or rewriting, in a further at
tempt to fulfill the original holistic aim. Arguably, unfulfilled intention is
therefore no different from a further state of contingent intention, unless it
is accepted that the return to a text after publication is of a fundamentally
different order from the return to a text before publication.
Revised intention is also temporally situated after an act of publication
or fixed completion of a text. This is similar to unfulfilled intention but im
plies that the writer, rather than still trying to meet his original objectives,
returns to the work with changed objectives.'^ Such a change may occur as
the result of a considerable time delay between first finishing the work and
returning to it, so that the writer has lost sight of, or forgotten, his original

64

reclaiming PROCESS
mtentions. Alternatively, the changed context of his life and other works
may have made the writer dissatisfied with those original intentions. Re
vised intention will result in material that is effectively defined as a separate work from the original and that differs from it intellectually as well as
textually Richardson s Pamela might be cited as an example. According
dip Gaskell, The plot of Pamela, which had served as the vehicle for
Richardson s first headlong inspiration, proved difficult to reconcile with
is later urge to give the novel a stronger moral purpose."" The novel thus
exists m a first edition that presents Pamela "at her most natural but with
crudities of tone and structure," in revised editions in which her "manners
were refined, and in a rewritten version finally published in 1801 "in which
Richardson used his mature technique to tackle the contradictions in the
plot and to enhance its morality."^

When compositional process is considered as a complex structure of


interrelated states and acts, an important connection between time and in
tention also emerges. It might be argued that developing any kind of time
structure for intention simply takes us back to an old authorial model of
linear progress toward a "final ideal" text with the danger of falling into
the teleological trap, in which a draft can only be read as a function of the
printed text. However, what I am describing here is a structure within
which different kinds of time perspective on a work compete or coexist.^^
n so doing, I am partly influenced by the handling of teleology by French
genetic critics as a necessary but flexible element in genetic studies, as
discussed at the end of chapter 1. At one level, of course, the text comes
into being chronologically, but diflferent kinds of intention relate to dif
ferent points withm, or perspectives on, that chronological time span.^'
ontingent intention and intention-as-process involve a sense of fluid time
m which intention is being acted out directly, or being rapidly overwritn. The Other intentional states are either anticipatory of this core state
(programmatic) or look back retrospectively on it across the divide created by the act of publication/completion (unfulfilled intention, revised
intention). Intention itself changes according to whether it relates to a
xed point m time from which process is considered or to a fluid, changing
process in action.

I would also suggest that in composition over time (as, for example
wi h the writing of a long poem), a drive toward intention in any particular
block of work IS often counterbalanced by an almost deliberate "resistance
to m ention at a creative level through the piling up of indeterminate ma
terial and the creation of multiple possible creative paths (one of which
may e fixed by the act of publication, but which is not the only possible
aping of the text). Michael Hancher makes an interesting observation

65

TEXT AS PROCESS

that is relevant to such issues: "There are casesmost obviously, certain


long worksin which the author never does face and reconcile his con
flicting tentative intentions^ for different parts of his text. Intention im
plies in itself a forward dynamic, a sense of purpose, an objective to be
attempted or attained. But the creative artist may not want to be thinking
in this way about the whole text at the point of writing one part of the text.
It is highly possible, then, that programmatic intention might be at odds
with contingent intention, or that there will be conflict within contingent
intentional material. At certain points, the writer may want simply to pro
duce a mass of material with no particular shape or order that he can then
draw on later. At this point, he may not want to be writing within an ac
tively shaping mass of material but to write with a deliberate lack of shape
(within the wholethere will still probably be clear intentions for what he
is writing locally). There may then need to be a denial of any sense of wider
intention in the short term for programmatic intention to be achievable in
the long term.^'
In conclusion, it is necessary to consider unintended meaning, which
occurs within, and as a by-product of, the intentional complex. It exists
in the form of embodied acts on the page but contains meaning not nec
essarily attributable to the writer's intentional acts. In Intention, G. E. M.
Anscombe states of the example of "offending someone" that "one can do
this unintentionally, but there would be no such thing if it were never the
description of an intentional action."^ Anscombe lists various kinds of
"happening" that may be intentional or unintentional, such as intruding,
offending, or kicking. In these scenarios, the possibility of unintended
meaning comes into being in all its ambiguity. Such ideas are also explored
by Jack Meiland in The Nature of Intention where he asks of intentional
actions, "Can the agent try to do X and yet unintentionally perform the Y
in question?"^' Again, like Anscombe, Meiland suggests that unintentional
action occurs only within a framework of intended action. He tells a nar
rative in which one individual arranges to meet another at a certain time
in a certain place. The individual goes there an hour early (as he thinks)
to prepare but finds the person there anyway because it is in fact an hour
later than he had realized. Meiland concludes: "Because he did what he
did unintentionally, what he did is not the carrying out of his intention"
(although it appears to be so to the other person).'" John Searle also briefly
addresses the question of unintentional action with relation to the Oedipal
narrative:
Oedipus intended to marry Jocasta but when he married Jocasta he was
marrying his mother. "Marrying his mother was not part of the Inten

66

RECLAIMING PROCESS

tional content of the intention in action, but it happened anyhow. The


action was intentional under the description "marrying Jocasta," it was
not intentional under the description "marrying his mother"
[T]he
total action had elements which were parts of the conditions of satisfac
tion of the intention in action and other elements which were not.''
This account makes clear once more that the unintentional is brought into
being as a kind of "by-product" of intention and that the outcome of an
action can appear to be that of satisfied intention without actually being
so. Searle goes on to offer a useful definition of the unintentional: "[W]e
count an action as unintentional under those aspects which, though not
intended, are, so to speak, within the field of possibility of intentional ac
tions of the agent as seen from our point of view."'^ Unintended action su
perficially appears to be in opposition to intended action, but it is, in fact,
dependent on it.
Jack Meiland is relevant here in that he makes clear the importance of
understanding intention in a temporal way:"These cases also show that we
should not speak of the agent's doing or not doing X intentionally, but only
of his doing or not doing X intentionally at a certain timeP^ Again, this
is a very important point for understanding intention in relation to crea
tive process as a sequence of acts. Meiland continues: "[A]n agent may be
performing an action intentionally at one moment and not intentionally at
another moment while performing the action continuously."'"' This shows
us that unintended meaning is strongly temporal and temporary within
the creative process. It may exist only in a single state in an isolated way
on a single manuscript page, or it may be converted by the author into
intended meaning and then carried forward within the text for all future
drafts. It is, however, likely to be either removed or incorporated at the next
stage of process.
If the "unintentional" exists within compositional material as actions
held in their physical marks on the page, then two kinds of unintended
meaning result from this: "accidental" and "unconscious." The first refers
to marks, stains, or other physical and material attributes of the manu
script materials that do not appear to bear any communicative intention.
The second refers to unintended meanings, such as misspellings, miscopying, unreadable words, repetitions, or occurrences of one word entered in
place of another. In a reading of Keats's manuscript of "To Autumn," Helen
Vendler shows that "some misspellings are suggestive. The sun is naturring (for maturing): orr (for or) has been proleptically contaminated by
the upcoming/wrrow; red becomes (by contamination from reap) readP^
Such "errors" can prove extremely fruitful if the "wrong" word turns out

67

TEXT AS PROCESS

RECLAIMING PROCESS

to be more striking and unexpected than the one intended to be written. A


psychoanalytic approach to process would of course focus on such acts as
ways in which the unconscious breaks through to find expression.'
A third kind of unintended meaning is more complicated and might
be described as "consciously intended unintentionality." For example, a
poet may intend to minimize intention in a work by introducing random
elements, such as cutting lines up and jumbling them about. In this case,
the writer manipulates the unintentional possibilities of meaning. Equally,
the deliberate introduction of others' intentions might occur here if the
writer consciously chooses to let another's intentions bear on the materi
als and so resigns his meaning in favor of someone else s.'^ Finally, and
more conventionally, we should remember that traditional poetry always
develops as a kind of compromise between meaning-content and sound or
shape. Elements of formmeter, line length, rhyme patternare all ways
in which semantic meaning is forced to confront more-or-less arbitrary
boundaries and to allow itself to be redirected and reformulated by an ex
ternal structure that is "unintended" in a sense (although the poet has, of
course, chosen to work within such a structure).
John Searle's final definition of unintentional action as "within the field
of possibility of intentional actions of the agent" points to the ambiguity
and uncertainty implicit in the unintentional.^ This is relatively clear in a
case like that of Oedipus (where we can assume that there is no likelihood
of him wanting to marry his mother so that the action is verifiably unin
tended) but far less clear when we consider it within the creative process,
particularly in material terms. We cannot be sure whether unintended
meaning occurs indirectly as an unconsidered consequence of, say, writing
on a page of a certain size that may limit or constrict or shape that work,
or whether such a shaping was actively envisaged and chosen by the writer
from the start. One core characteristic of unintended meaning, therefore,
is that it is highly ambiguous as to whether or not it is unintended. Certain
writersEmily Dickinson for oneseem to make use of such ambigu
ity." However, even where unintended meaning may pass unnoticed by
the writerexisting only as part of the functionality of the compositional
contextit always has the potential to be reinterpreted by later readers (or

and some kind of "failed" communication. Such a model privileges fulfill


ment of intention through successful communication (in which intention
is satisfied), with the result that "[n]onplenitude will be treated as though
it were an extrinsic accident, even if it in fact occurs frequently, even if it
takes place everywhere."^" Derrida argues that the relationship between
intention and "plenitude" (satisfied intention?) is in fact a divided one; if
plenitude is achieved, then intention "dies." Nonplenitude is therefore not
a mere by-product of the drive toward fulfillment but an essential part of
the nature of intention: "If nonplenitude (the non-telos) is therefore not an
empirical accident of the telos, or even a simple negativity, one cannot take
it into account as one might a contingent accident
[P]lenitude is at once
what orients and endangers the intentional movement.""" This is a classic
Derridean reading, in which the fulfillment of intention, which appears to
be its essence, is shown to be defined equally well by its nonfulfillment, or
at least dependent on the possibility of it. For Derrida, then, unintended
meaning precedes the intentional rather than existing as a by-product

by the writer as a later reader).


It may be helpful to conclude this discussion of unintended meaning by
a brief recontextualizing of the concept in relation to Derrida, as a reminder
of the dangers of an overly intentionist approach. In a critique of speech
act theory, Derrida opposes the way in which traditional speech acts set up
a hierarchized opposition between serious, meaningful communication

68

of it.
How does such an account relate to my definition of "unintended"
meaning? Derrida seems to be discussing the satisfaction or nonsatisfaction of intention rather than the accidental nature of the "unintended" (for
which the satisfaction of intention was not directly the goal). However, the
way in which he asserts that the potential failure of intention to be fulfilled
is bound up with its successful fulfillment could also apply to unintended
meaning, which exists as an alternative articulation of one of the many
possibilities held in language. A Derridean account of unintended mean
ing would presumably deny that it is a by-product of intention and see it
simply as the alternative coming into being of latent meaning.
It is important to note that I am not arguing for the superiority of in
tended over unintended meaning in any way. Rather, I seek to show that
it is often through the cross-interpretation of acts on the page that can
be reconstructed as "intended" and those that are "unintended" that we
can fully respond to the materials of process. In both cases, I would also
want to allow that we cannot absolutely know the intentional act (though
we can probably reconstruct intentional sequences) and cannot always be
sure that the unintended is unintended. In a sense, then, I position myself
somewhere between Searle and Derrida in relation to acts on the manu
script page. I find the distinction between intended and unintended com
positional acts useful and worth retaining, but I do not want to assert that
only intentional acts are of value for our response to the manuscript object,
not at all.

69

TEXT AS PROCESS

RECLAIMING PROCESS

A TYPOLOGY FOR THE STUDY

implicitly or explicitly articulated in the scholarly framework. If the edition


works successfully, however, it should provide the scholarly reader with
enough access to the material for him orher to be able to check and confirm
the editor's conclusions for himself and to then use this material to explore
the text further, or to release the material from a teleological perspective
(which may be necessary in the first instance). If the reader is working with
a facsimile text that has no scholarly framework, then some of these more
"text-critical" tasks may need to be undertaken before analysis can occur.
If the reader is working with an electronic presentation of the text, in the
form of multiple versions, then the question of whether it is necessary to
establish a teleological perspective on the manuscripts, or whether a radial
structure is able to be immediately used, can be explored.
Once an initial sense of the structural relationships between composi
tional materials has been established, the reader can then begin to explore
more interpretative, literary-critical ways of responding to that material.
"Text-critical" activities represent a valuable preparatory element for work
on compositional material. While not explicitly articulated within the
framework I will give, such activities are implicitly present within it.
At the other "pole" is the use of materials of process to clarify and pur
sue cruxes within the published text that are revealed, explained, or con
tradicted by knowledge of the shape, structure, and development of the
work in a state of process. This is a vital act, and one in which literary criti
cism and text as process are bound together. It is probably the way in which
draft material is most frequently used by literary critics.
Between these two poles, and forming the core ground for a composi
tional or genetic criticism, are activities relating to specific study of the act
of composition and of process considered either independently of "prod
uct" or the final meaning of the work. Such activities might focus on a par
ticular phase of composition, on the materiality of process, or on the way
in which meaning emerges for a particular writer or form.
In chapter 2 of La Genetique des Textes, Pierre-Marc de Biasi formulates
some of the stages and phases involved in the study of genetic material. He
defines four stages of reahzation within the process;

OF COMPOSITIONAL MATERIAL
I want to conclude this chapter by offering a typology that will clarify dis
tinct phases within the compositional process and thus provide a practical
framework for critical response to such material. Thinking about the use
of such a framework returns us to the question of what a compositional
method is for and takes us back to the relationship between text-critical and
literary-critical activity. Historically, a distinction has been made between
literary criticism (an act of interpretation that has as its aim the produc
tion of meaning within a work) and textual criticism (an activity concerned
with the presentation of a work and with details of form). However, as
Peter Shillingsburg asserts; "The central concern of both textual critics and
literary critics is meaning. The central focus or locus of that concern is the
text. The problematic nature of meaning agitates literary critics and theo
rists; the problematic nature of texts agitates textual critics and theorists.
Both should agitate us all."''^ The relationship between textual and literary
criticism in the study of draft materials can be seen as a continuum with, at
one extreme, controlled "textual" tasks (such as decisions about the spell
ing and presentation of words on the page) and, at the other, the subjective
interpretation of a literary text in terms of its content and context. Between
these two positions, a range of activities occur that involve varying degrees
of critical intelligence and judgment, and in which the two areas frequently
overlap. The outline I am about to provide is intended to suggest that the
study of textual process could potentially range across the whole contin
uum (linking textual process to textual product) or involve engagement at
a particular point or stage within process. A compositional method also
has obvious application to editing practices (allowing the possibility of an
edition structured to explore the materials by means of the framework or
one phase within it), but it really aims at enlarging interpretative practice
through a new form of critical analysis in which textual and literary-critical
activity can both be employed for the full understanding of process.
Activities within editing that are defined by a teleological underpin
ning and nearer to a "text-critical pole" might include; clarifying the order
of the manuscripts; developing a model for stages of composition relating
manuscripts to each other; mapping this model onto biographical informa
tion; understanding a particular manuscript notebook or a particular se
quence within that notebook; or identifying characteristics of a particular
text in terms of its composition by a particular author. For the literary critic
working with an Anglo-American scholarly edition, many of these tasks
are likely to have already been performed by the editor and to have been

70

phase pr^redactionnelle [precompositional phase]


phase redactionelle [compositional phase]
phase pre^ditoriale [prepublication phase]
phase ^ditoriale [publication phase]^'
Only the first three of these phases involve the avant-texte. The prepub
lication phase ends at the point of "bon a tirer" [passed for press/ready to

71

TEXT AS PROCESS
RECLAIMING PROCESS

print], at which point the avant-texte becomes a text. In "What Is a Literary


Draft? which argues for the value of studying the rough draft, Biasi sets
out to provide "a general table of the stages, phases, and operational funcions that enable the classification of different types of manuscripts accord
ing to their location and status in the process of a work s production." The
table, which can be found in Yale French Studies 89 (1996), provides a useful
summary of Biasi's work and a model for my own typology It is defined
y Biasi as a chronotypology" and thus works horizontally, to describe
different aspects of each phase, and vertically, to denote the relationship of
each phase to the next in a teleological line. Biasi insists that a causal struc
ture is essential in classifying genetic material, but he also asserts that
1 operation consistant kfaire comme si chaque brouillon successif chaque
operation du scripteur representait une etape vers le but final qui est le
texte. Cette representation heuristique est necessaire, mais elle n'est evidemment pas adequate pour decrire la realite des conflits, des desirs, des
hesitations
qui caracterisent la genese.
[the operation consists in acting as if each successive draft, each act of
the writer, represented a stage toward the final goal, which is the text.
This heuristic representation is necessary, but it is obviously inadequate
to describe the reality of the conflicts, desires, and hesitations ... that
characterise genesis]
In oAer words, a larger teleological framework needs to be in place to cat
egorize and to negotiate the materials, but the fact that it is present does
not commit one to a rigid, goal-centered response to that material
Biasi also makes it clear that he is aware of the generic Hmitations of
his table in What Is a Literary Draft?" and in his detailed account of the
four phases (which I just outlined) in La Genetique des Textes, where seneric emphasis on the novel comes through to color his account of each
stage. This marks a key distinction between his typology and my own
(which has Its own generic bias in favor of poetry). In comparison to
Biasis tpology, then, my outline focuses more on the potential activi
ties of the critic than the characteristics of the textual material. It does
not aim to be complete, only to give suggestions of possible activities that
might be undertaken."^
A. The Compositional Context
Compositional environment (habits, place, space, people)
Equipment (pen, pencil, size and shape of materials, nature of
entry)

72

. Breakdown of equipment (on page, in instruments, in larger


context)

Events/circumstances bearing on composition


B. The Intentional Core Structure
Programmatic Intention

.
.

Pre-textual composition (mental and oral)


Motivation for composition (internal, external)
Relationship of text to other texts (of the author, of other authors)
Potential failure, change, or redirection of programmatic inten
tion

Planning, notes, outlines, lists, synopses


False starts, undeveloped projects
Contingent Intention
Moment of first draft
Phases ofwork on a text (distinctions between first draft, draft fair
copy, printer's copy)
Compositional propulsion, means of stimulating or restarting
composition

. Points of contingent completion (fair copy-texts of part or whole


work)
Structural organization and reorganization

to other text (of the author, of other authors)


Multiple versions, rewritings
Intention-as-action (Micro-intentionality)
PhysKal aspect and appearance of text on the page
Physical aspect and appearance of manuscript notebook
Localized structural organization
. Revision: immediate, interniittem, long-term, shaping, propulsive
hftect of changed context on meaning of words
Replacement of one word with another (reflecting changed
intention)

Creative judgment (changing intentions, how defined, how di


rected)
Use and role of amanuenses
Abandoned works
Act of Publication/Point of Completion
Concept of completion
Prepublication preparation of text
Distinction between revision and editing
Revision of proofs
Private prepublication printed texts
Increased effect of nonauthorial interventions

73

TEXT AS PROCESS
Unfulfilled Intention/Revised Intention
. Response to critical (or domestic) reception
. Revisions to first edition and later editions
Enlargement of a work (once or more than once, over time)

Restructuring and rewriting


Revision or preparation of texts by author and others
Significance of last published lifetime edition
Significance of posthumous publications

/loce^^

Unintended Meaning
Degree to which the unintentional is allowed to enter into the cre
ative process
Local level within the text
. Local level within the text by another (result of dictation, copying)
. Effect of external factors, human and nonhuman (interruption,
mood, weather)
. Effect of material factors (shape, size, form of paper)
. Effect of printing and publishing methods and conventions
Unconscious intentions on the page (slips of the pen, misreadings,
miswritings)
Nonauthorial Intention (This also comes under A)
Indirect effect of others on authorial intention, prepublication
(opinions of friends, family, editor)
. Direct effect of others on the text domestically (copying, punctua

How it staggered me to see the fine things in thin ore! interlined,


corrected! as if their words were mortal, alterable, displaceable, at
pleasure! as if they might have been otherwise and just as good! as if
inspiration were made up of parts, and these fluctuating, successive,
indifferent! I will never go into the workshop of any great artist again.
Charles Lamb, "Fragments of Criticism:
The Disenchantments of an Original MS"

tion, spelling)
Direct effect of others in preparation for publication (compositors,

The study now moves from an editorial and theoretical account of com

typesetters, publishers)
Direct effect of material aspects of publication (number of pages,

chapters take as their focus three major nineteenth-century poets and a

layout, shape and size of text)


Indirect effect of others on future authorial intention, postpublication (reception by critics, journals, wider circle of friends)
C. The Language Totality
Interpretation of meaning content
. Consideration of process as a "becoming"
. Whole/part relations, hermeneutic circle
. Consideration of the "spontaneous" (immediacy of writing)
. External and authorial intertextuality

positional process into critical analysis of draft materials. The next three
major edition (or series) for each through which to explore editorial and
critical treatment of text as process. My aim is not to work through each
stage of the typology given in chapter 3 for each writer but to discuss those
elements of the typology that seem most applicable to the particular writer
concerned. This also serves to make the point that I do not see the method
as offering absolute, rigid principles. Rather, it provides a framework that
can be drawn on with varying emphasis according to the characteristics of
the materials under consideration.
As we saw in discussions of the material by practitioners of the critique
genetique in chapter 1, text as process is freer, more fluid, more experimen

In the three chapters that follow, the framework I have just given will be

tal, more circular, and more repetitive than a "final" text, and working

applied and tested in various ways to the poetic draft material and com

with such material is of a very different order. This is true at the obvi

positional practices of William Wordsworth, Alfred Tennyson, and Emily

ous physical level of reading manuscript pages but also in terms of the
shape and content of material on those pages, in the relationship between

Dickinson.

different revisions of the same passage, the relationship between separate

74

75

TEXT AS PROCESS

WORDSWORTH'S PROCESS

passages of work toward the same section of a poem in a notebook, and


so on. At the same time, it is worth bearing in mind, always, the double
paradox that textual process exists only as a consequence of the existence
(and status) of the completed text, and also that editorial construction of
compositional material (from the order of manuscripts, through the order
of drafts, down to the words on the transcription page) constantly relies
on and uses the completed text as a guide through which to achieve an
accessible form.' This chapter centers on Wordsworth s creative practice
for his two long poems The Prelude and The Excursion, revisiting different
elements of the typology in relation to Wordsworthian composition.
The question of the extent to which elements of compositional material
can be dissociated from the final form of the text (published or unpub
lished), and from the extended compositional process, is raised directly in

The first is "a composante genetique mais centrees sur le texte et enrichies
par des transcriptions" [of a genetic composition, but centered on the text
and enriched by transcriptions].'' By contrast, a genetic edition does not
aim to present the publication of a textual work or to establish a text but is
centered on a presentation of genesis itself. The structural organization of
a Cornell volume clearly privileges a reading text (or texts), with an intro
duction that provides a chronological account of how that text came into
being and transcriptions revealing the full process. At the same time, how
ever, the underlying principle running throughout the Cornell Series for
Wordsworth involves the privileging of earlier versions of texts over later
versions revised by the poet. Thus, the series does have a strong interest in
textual genesis but tends to be directed toward establishing a stable text or
texts and is therefore not purely "genetic."
Jack Stillinger is the most vocal critic of the editorial principles un
derpinning the series. In "Textual Primitivism and the Editing of Words
worth," Stillinger criticizes the Cornell Wordsworth on the grounds that
"its understandable eagerness to discover, promulgate, and extol early ver
sions to take the place of later ones, is in the process of doing away with
the later Wordsworth once and for all." Stillinger then runs through four
problems that he identifies in the project: the difficulty of defining an "ear
liest complete state," the annotation, the exclusion of Wordsworth's final
edited texts, and the influence of the series on future study. In relation to
the last point, Stillinger expresses concern over the "inadvertent standard
izing of these early texts."' His conclusion, closely related to the German
theory of versions, is: "A healthier reaction would be to stop this nonsense
about 'the worst of Wordsworth' and grant the legitimacy and interest, in
trinsic or in connections with other texts, of all the versions of The Prelude
and the rest of the poems in the canon. Recent textual theory... favors this
more catholic view, and it has the additional support of common sense:
Wordsworth did, after all, write the 1805 version and the 1850." In fact,
ten of the twenty Cornell volumes present more than one version of a text
as parallel reading texts, sometimes involving different manuscript ver
sions, admittedly, but also sometimes including both an early and a signifi
cantly later version (e.g.. The Salisbury Plain Poems, The Borderers). Two
volumesr/ie Ruined Cottage and The Salisbury Plain Poemsprovide

the editing of Wordsworth's poetry.


THE CORNELL SERIES AND CONTINGENT COMPLETION
Over the last thirty years, the Cornell Series for Wordsworth has had a sig
nificant effect on understanding of the poet's canon and awareness of his
writing process. The original proposal for the series was made by Stephen
Parrish and John Finch to Cornell University Press in 1966 (although, ow
ing to various circumstances, the first volume of the series did not appear
until 1975). The first three volumes proposedSalisbury Plain Poems, The
Prelude, 1798-99, and Home at Grasmereas well as The Ruined Gottage,
were published with a clear aim in mind: to "rescue these lost poems and
to display for the first time a full and accurate record of the growth of
Wordsworth's other poems, from his original drafts down to the final life
time, or first posthumous printings."^ These editions presented authorially
unpublished material that had been overlooked or marginalized by earlier
editions and yet was felt to be both substantial and significant for an un
derstanding of Wordsworth's poetry.
In the light of chapter 1, we might want to ask why a Cornell edition is
not defined as a "genetic edition." Pierre-Marc de Biasi's distinction be
tween two kinds of genetic edition (horizontal/vertical) describes the latter
as "la publication chronologique des documents se rapportant a la serie
integrale ... des transformations successives qui permettent de comprendre sa genese" [the chronological publication of documents relating to an
integrated series of successive transformations that allow understanding
of genesis].^ This sounds very much like a Cornell edition for Wordsworth.
However, elsewhere Biasi makes an important and useful distinction be
tween a critical edition, genetically composed, and a fully genetic edition.

76

four reading texts.


Without question, the interest in early versions of Wordsworth's texts,
evinced by Jonathan Wordsworth, Stephen Gill, and the Cornell editors,
has brought to light much early poetry that previously existed only in man
uscript and to which previous readers of Wordsworth had no access. But
how far should this go? The important distinction made in Continental

77

TEXT AS PROCESS

genetic editing between text as process and text as product is partly under
mined by the creation of reading texts from draft material in the Corne
Series. This becomes more problematic if texts such as The Rumed Cottage
and The Prelude, 1798-99 are then removed from a compositional context
to be reproduced elsewhere as discrete pieces. Donald Reiman, in his initial
report on the early volumes of the Cornell Series for the Center for Schol
arly Editions, expresses such concerns. Of the inclusion of the Two-Par
Prelude" within the third edition of the Norton Anthology, he comments;
"To have these early texts available for the scholar and student is valuable;
to have the two-part Prelude of 1789-1790 [sic] the only version of the Pre
lude available to students encountering WW for the first time seems to me
less unambiguously so."'
rj
On the one hand, it might be argued that separate publica ion of draft
material achieves one of the ambitions of genetic criticism by releasing
manuscript material from its connection with the "final text, thus al ow
ing it to be considered and valued in its own right. On the other, the danger
is that it does so at the expense of losing its identity as manuscript material.
So the "good news" is that many students of English literature are now
studying draft material of Wordsworth's poetry; the "bad news" is that
they may not know that they are doing so, or, even if they do, that they wil
simW respond to it as if it were a "final" text rather than developing spe
cific skills for the study of compositional material.
This is the kind of position that I think the French genetic critic Laurent
Jenny is gesturing toward when he states; "To present a pre-text for reading
is obviously to inaugurate it as a text. ... In this sense, a pre-text cannot
be read and still remain a pre-text."'" The question it raises in relation to
manuscript material is whether a critic's or editor's ambition for the mate
rial should be to create maximum reader awareness and accessibility tor a
completed block of manuscript work (in which case its reproduction m the
Norton Anthology is to be seen as a good thing), or whether the ambition
should be to allow for full exploration of that material in a context where
its status as a particular kind of text (representative of process, not product)
is most valued. My own concern with valuing textual process would lead
me toward the latter, rather than the former, position.
When we consider the editing of Wordsworth in relation to the typol
ogy given in chapter 3, it is clear that the concept of acte of contingent
completion, occurring within the compositional process, is hig y signi
cant for the editorial position adopted by the Cornell Series. Continent
completion describes stopping places within the compositional develop
ment of a major work and might be represented most clearly by a fair copy
manuscript that brings together disparate notebooks and unites a number

78

WORDSWORTH'S PROCESS

of books for the final poem. With The Prelude, firm points of contingent
completion are represented by final fair copy in the form of MSS U and V
for The Prelude, 1798-99; MSS A and B for The Thirteen-Book Prelude; and
MSS D and E for The Fourteen-Book Prelude}^ Such acts provide the con
tinuous (usually fair copy) textual material for Cornell as well as the justifi
cation for the establishment of a "fixed" textual version as the reading text.
In The Music of Humanity, Jonathan Wordsworth makes the importance
of such acts of completion explicit: "One's aim in presenting a text must
clearly be to reproduce the poem exactly as it stood at a single point in
time. The obvious point to choose is the moment of completion, before any
revision has taken place."'^ The difficulty with this position, however, is
the question of what constitutes a poem in a state of completion within the
compositional process.'' Ultimately, the point to bear in mind is that for a
writer such as Wordsworth there is no such thing as absolute completion;
even a completed state is always contingent. Ray Carney, in an excellent
review of the Cornell Series, makes this point; "But of course one of the
things these manuscripts show is the extent to which the concept of a final
text is itself a critical fiction in a case like Wordsworth's."''*
TWO LONG POEMS AND PROGRAMMATIC INTENTION
I want now to return to the concept of programmatic intention to consider
it in relation to Wordsworth. Programmatic intention concerns the writer's
anticipated plans and expectations for his work before any active, recorded
composition. It has a partly abstract identity, corresponding to the writer's
highest ambitions and standing as a measure by which he or she assesses
the success, or otherwise, of the final achievement. It may also exist in a
'.more tangible state in the form of notes, plans, scenarios, and so forth.
In Wordsworth's case, the nature of programmatic intention for either
The Prelude or The Excursion is complicated by the fact that major individ
ual works are composed within a structure of programmatic intention for
an even larger whole ("The Recluse") and are measured against the ambi
tions in play here. Wordsworth's decision not to publish The Prelude, made
during the process of completing the poem in 1804-5, is related to the dis
placed importance given to the wider structure; "[I]t seems a frightful deal
to say about one's self, and of course will never be published, (during my
lifetime I mean), till another work has been written and published, of suf
ficient importance to justify me in giving my own history to the world."''
The 1814 title page presentation of The Excursion as "A Portion of The Rec
luse" stands as a public statement of this position, demanding that read
ers and critics be aware of the greater ambition.' So the presence of the

79

TEXT AS PROCESS

projected "Recluse" creates a superstructure and an abstract spatial model


within which composition on parts of that model (two long poems) takes
This massive structure of programmatic intention, publicly declared,
both helps and hinders Wordsworth as a writing poet. In part, it works to
release individual works from the extreme pressure of poetic ambition be
cause it locates the satisfaction of programmatic intention elsewhere. The
writing of The Prelude clearly reflects this. Wordsworth is able to work on
it because he knows he really ought to be doing something else. In a letter
of 1805 he says: "I began the work because I was unprepared to treat any
more arduous subject and diffident of my own powers. Here at least I hoped
that to a certain degree I should be sure of succeeding, as I had nothmg to
do but describe what I had felt and thought, therefore could not easily be
bewildered."'^ Beth Darlington notes of the decision to expand The Prelude
beyond five books that "the enlargement of The Prelude enabled him to
compose poetry associated with 'The Recluse' but to avoid direct confron
tation with his sense of inadequacy to proceed with it."' Creativity in one
area is partly stimulated by evasion of another.
Wordsworth's emotions on completing the 1805 Prelude also strongly
indicate a sense of unsatisfied programmatic intention:
I have the pleasure to say that 1 finished my Poem about a fortnight ago.
I had looked forward to the day as a most happy one; and I was indeed
grateful to God for giving me life to complete the work, such as it is; but
it was not a happy day for me I was dejected on many accounts; when
I looked back upon the performance it seemed to have a dead weight
about it, the reality so far short of the expectation; it was the first long
labour that I had finished, and the doubt whether I should ever live to
write the Recluse and the sense which I had of this Poem being so far
below what I seem'd capable of executing, depressed me much."
The poet's internal value judgment means that programmatic intention
is unfulfilled, and this seems to be a common experience in relation to
this kind of long-term, ideal, preconceived aim. However, in Wordsworth's
case, programmatic intention is not merely the poet's own but is also partly
external and embodied in another man s mind.
The origins of "The Recluse" project are to be found in discussion be
tween Coleridge and Wordsworth in 1798 but remain unknowable in any
detail because they were unrecorded at that time. From this point onward,
Coleridge and Wordsworth's personal relationship is bound up with the
development of "The Recluse," and there is a strong (and unusual) sense

80

WORDSWORTH'S PROCESS

of shared programmatic intention that creates difficulties in terms of dis


tinguishing between Coleridgean intention on Wordsworth's behalf and
either Wordsworth's understanding of this, or his own intentions. Stephen
Gill describes the dynamic that is created between the two men in relation
to Coleridge's letter of1799 to Wordsworth (in which he expresses the hope
that early Prelude composition is for "the tail-piece of 'The Recluse'").^
Gill states: "This letter and Wordsworth's reaction set a pattern for future
years. After a lapse of time Wordsworth surfaces with something demon
strably achieved. Coleridge, who in the same time has completed nothing,
seems to welcome it, but actually disparages the work by wishing it were
something else. In immediate response Wordsworth is checked, but then
he reasserts more vigorously the validity of what he has been doing."^' As
well as Coleridge's approval, the presence or absence of Coleridge in Words
worth's life is consistently significant for periods of active composition and
of inactivity on the long poems.
Coleridge's involvement in Wordsworth's programmatic intention for
"The Recluse" gives him enormous critical power because his opinion has
the remarkable status of a kind of external, subjective judgment for the
poet. He is not the creator of the work, but he has a creative status in re
lation to it. As a result, when the failure to fulfill programmatic intention
invoiced externally, this acts as a powerful block on any further attempts,
by Wordsworth, to continue with the work. For this reason, Coleridge's
criticism of The Excursion must play a significant part in Wordsworth's
inability to attempt to fulfill the original intention and to complete "The
Recluse."^^
A programmatic element is present in both French and German theo
retical accounts of process when they attempt to make a universal differ
entiation between two major kinds of writing (and, by implication, writer).
The critique genetique distinguishes between a "programmatic" form of
writing and a "process-centered" one in which the programmatic writer
relies on planning and creative anticipation to produce the work, while
the process-centered writer is able to jump straight into drafting and re
drafting. The distinction has consequences for the nature and shape of the
genetic corpus, so that a programmatic writer will tend to work within
a larger structure of anticipated writing and conceptualization, while a
process-based writer produces successive versions of a work building on
itself.
In German hands, a corresponding distinction is offered but is articu
lated with a slightly different emphasis. Scheibe, in "On the Editorial Prob
lem of the Text," distinguishes between a work that is mentally composed
before it is written down (resulting in a relatively clean text) and a work for

81

TEXT AS PROCESS

which active creative thought occurs on paper through the act of writing
(resuhing in a heavily revised and rewritten text). The Dutch critic Anne
Marie Kets-Vree defines these two models as that of the "Mozart versus the
Beethoven procedure," with the former involving genesis in the mind and
the latter genesis on paper.^'
Scheibe's distinction has been taken forward by another German critic,
Klaus Hurlebusch, who has made the fullest attempt thus far to define
the two types in detail in "Understanding the Author's Compositional
Method." His aim is to bring together the two aspects of writing that ex
ist in a dialectical tension between "the mainly reproductive, work-genetic
writing process ('poiesis') and the mainly constructive, psycho-genetic writ
ing Cpraxis')."^^ In relation to its reproductive aspect, the writing process is
viewed as a communicative act, anticipating readership from the start and
being externalized and intersubjective. This mode of writing, Hurlebusch
argues, values directedness toward a goal (text as product) and is likely to
be teleological. The second mode (constructive) is focused on productivity
within a more solitary, internal model of writing. From this perspective,
texts are transitional stages in an unending process. Creative goals are not
met by the completion of a text but are linked to personal self-development,
and as a result, a writer for whom this aspect is dominant will be unable to
separate him or herself from the work, which maintains a private signifi
cance even when published.^^ The heart of the distinction, then, is between
an internalized or externalized self-conception in which the first type of
writer will hold the writing within until it is perfect, while the second type
is inclined to externalize and rework it in an external form.
Direct criticism of the attempt to articulate a universal distinction be
tween "two kinds of writer" has been made by Almuth Gresillon and Dan
iel Ferrer. Gresillon suggests that such a distinction cuts across all periods
and literary movements and ignores the relationship between writing and
its historical context.^ Ferrer proposes that in fact the "process" writer is
simply a subcategory of the "programmatic" type, so that "process writing
appears to be a particular case, or minimal form, of programmatic writ
ing."^' Pierre Marc de Biasi, who advocates the "two kinds" model, freely
admits that there are many writers who fall somewhere between the two,
or partake of both.^ This suggests that the distinction is really one of de
gree or tendency, offering a spectrum of activities rather than an opposi
tion, and, to be fair to Hurlebusch, his model does try to articulate this. In
the terms of this book, the distinction could be easily understood as that
between Searle's prior intention (programmatic) and intention-in-action
(process-centered)for which it will be recalled that prior intention is
not essential for intention-in-action to occur. This would seem to support
82

WORDSWORTH'S PROCESS

F.errer s point that the "two kinds" are really concerned with a preexistent
stage that may or may not be present.
If we take Wordsworth as an example and attempt to align his practice
to these models, it is clear that he might initially appear as a "program
matic writer, with a heightened awareness of the macrostructure (thus
falling into the first category), but he is also very clearly a writer who views
creative process as undetachable from the self and unending, and so could
be defined as "psychogenetic." Moreover, his actual practice is often pro
cess-centered, building on itself through messy redrafting, although this
is clearly not how he likes to think of himself. Thus, the complexity of
the two kinds" model and the self-perception linked to either category
ultimately draws attention to the compositional contradictions between
'speech and writing that so inform Wordsworth's poetry (in ways that are
^explored later in this chapter).
In terms of the compositional contexts outlined in chapter 2,1 would
suggest that both of these "kinds" are participating in the doubled or framed
intentional/nonintentional structure of composition, but to different de
grees. The writer who relies on plans is trying to maintain control of pro
cess (or the illusion of control through inner speech) as long as possible
b^t must eventually enter into open engagement with language. The writer
wjio simply writes is able to enter more freely into the unwilled structure
and depends less on the willed anticipation-and-return structure of intenhon around it. This relationship is also in effect an interpretativeor selfinterpretative one that demands that the writer move between two very
different relations to the material produced: one in which he or she is the
writer writing (experience of the making) and one in which he or she is the
writer reading the written draft (experience of the thing made).
A literally constructive sense is felt in Wordsworth's compositional
method in relation to his long poems. So, for example, in his introduction
to- the Cornell edition of The Prelude, 1798-99, Stephen Parrish describes a
slow process of assembly" for the first part of this early work.^' He states:
Fpr Wordsworth it was mainly a matter of fitting pieces of verse together,
like parts of a puzzle."'" Such a model is, I think, fundamental for Words
worth and informs on a larger scale his long poetic composition from the
very earhest work for The Prelude onward. Rather than viewing a long work
as a continuous fluid whole, he tends to view it in terms of building blocks
that can be moved around, repositioned, or taken out. Hurlebusch states of
the constructive writer: "Larger texts result from a joining of text segments.
This is due to the self-reflexivity of the writing process. The author's gaze
oil the written has a decisive genetic significance for the writing ... texts
emerge from acts of writing."'' As a writer, Wordsworth rehes heavily on

83

TEXT AS PROCESS

what is already written, so that the text, in Hurlebusch's words, "fertilizes


itself."32
For Wordsworth, first written composition is a difficult part of the po
etic process. In his biography of the poet, Stephen Gill remarks; Through
out his life Wordsworth complained of physical symptoms whenever
struggling with embryonic ideas or revising work whose particular form
no longer satisfied him."'' In the Letters Wordsworth repeatedly describes
his difficulty with the act of composition in physical terms. Writing to
Coleridge from Goslar in 1798, he famously states; "As I have had no books
I have bL obliged to write in self-defence. I should have written five times
as much as I have done but that I am prevented by an uneasiness at my
stomach and side, with a dull pain about my heart. I have used the word
pain, but uneasiness and heat are words which more accurately express my
feeling. At all events it renders writing unpleasant."'" Physical symptoms
are used as an explanation for underachievement, and the poet here par y
seems to recognize that he translates into terms of illness symptoms not so
much about being physically unwell as a mental state of disturbance and
excitement that could be viewed as a natural part of the imaginative and
creative process.
. .
Wordsworth's difficulties with first composition may, again, be par y
explained by the definition of him as a "constructive" writer in the sense
that he requires composition to build on itself. When we think about first
written composition within the whole compositional process, it is clear
that what makes it such a unique stage is the fact that it has to generate
its own material. Paul Eggert, in the article "Textual Product or Textual
Process," makes this clear; "Revision is accorded the sanie importance
as composition despite the likelihood that the author will often have been
more intently engaged with the text when first writing it. Put another way,
composition starts from scratch whereas revision is done to an existing
text; at the time of revision the author is subject to a crucial textual influence-the existence of a manuscript-that was lacking when he or she first

^''^mrrswSisapoetwho atboth intellectual andphysicallevels makes


considerable use of a preexisting context for composition. It is not surpris
ing, in light of this, that the stage of composition he finds most physica y
disabling should be that for which there is no preexistent material,
is
notable too that the letters expressing these feelings most strongly con
cern the early writing of the early Prelude. When Wordsworth comes to
work on his second long poem. The Excursion, jh^blemys partly side
stepped by the presence of book 1 in the form of The Rumed Cottage. This
provides him with a body of material in a fair copy state about which he

84

WORDSWORTH'S PROCESS

feels confident. The poet can thus begin a new major work by revising and
making changes to this piece, preparing it for its position as the first book
of many, and in so doing leading himself on into fresh composition from
an already established base.
COMPOSITIONAL INTERTEXTUALITY
A second aspect of Wordsworthian composition, to be considered across
both The Prelude and The Excursion as developing long poems, concerns
compositional intertextuality. In a consideration of what he calls "intertex
tual genetics," Gene W. Ruoff makes clear the ways in which, for a poet like
Wordsworth, intertextuality is wide-ranging; "an adequate genetic criti
cism would have to understand a text as a confluence of diverse waves of
influence. Any text has strong connections to its immediately preceding
versions and to the prior body of a writer's work. It also has connections
to a surrounding literary climate, which in the case of Wordsworth and
Coleridge is embodied especially strongly but not totally in their mutual
influence upon one another."' Such ideas are allowed for in French genetic
criticism through the distinction between "endogenetics" and "exogenetics." The former describes the development of draft materials out of other
writing, the latter, development of draft materials by means of external
sources and stimuli (but only other texts, not empirical objects).'^
My interest here is not so much in the author's relationship with exter
nal influences as it is with intertextuality as it exists within his own com
positional material ("endogenesis"). Paul Eggert, again, makes some inter
esting observations on this subject; "Textuality will be better understood,
I believe, when it is opened outnot only when different textual states of a
developing work (let us say, a novel) are compared to one another but also
when elements of them are compared to other things the author was work
ing on while the novel was in progress.... These relationships within and
between the author's writings form, as it were, an authorial intertextual
ity."' A clear example of this in Wordsworth's case exists in relation to the
epitaphic books of The Excursion. Wordsworth's writing here is strongly
anticipated by his earlier poem "The Brothers" but is also colored by prose
writing in the form of the Essays on Epitaphs written for The Friend, as well
as his translations of the epitaphs of Chiabrera, which were all taking place
when he was first drafting the poetry. The author later makes the connec
tion explicit by publishing the first Essay in the Notes to The Excursion.
However, I would suggest that there are (at least) two further kinds of
authorial intertextuality existing within the compositional process. The
first concerns intertextual activity involving the re-employment of mate-

85

TEXT AS PROCESS

rial that is written for one context, in another. This is a core compositional
activity for Wordsworth and occurs across a wide range of texts. The sec
ond concerns the juxtaposition of different materials within a single manu
script notebook. It is worth noting that the Cornell Series for Wordsworth
structures each volume around a single or multiple textual product so that
a range of manuscripts are represented according to whether a particular
text is entered into them. What the Cornell Series therefore does not do
is provide the alternative compositional perspective that also exists, but
is easily overlooked, in the form of a single discrete manuscript that con
tains material toward a number of different texts within its covers. I want,
therefore, to adopt the alternative perspective in order to explore material
intertextuality in a single notebook.
The earliest Ptelude notebook is that of MS JJ (DC MS 19), written in
Goslar in 1798-99. DC MS 19 bears within it a mixture of poetry and prose,
of William's writing and Dorothy's, and of the mundane, the personal, and
the poetic. Toward the front of the notebook, entered while in Germany,
there is a prose account of a visit to Mr. Klopstock; Dorothy s account of
leaving Hamburg and going to Goslar; German verb tables and grammar
work as the young people attempt to learn the language. The Prelude draft
runs forward from the back of the notebook, starting with Was it for this
on the inside back page (89'), until it meets two pages of German verbs
and then the "Essay on Morals," both inverted. Some years later, Dorothy
reused the notebook for journal entries from Sunday, 14 February 1802, to
Sunday, 2 May 1802, and this fills the middle portion of the notebook.
Looking across the contents of MS JJ tells us quite a lot about the na
ture of composition for this poet. Above all, it stresses that the notebook is
sharedeither William or Dorothy makes use of it at different times and
for widely different purposes. The nature of these entries in physical terms
(sometimes in pencil, at the back, in a shared book) also suggests a quite
unaffected attitude toward the poetic draft material. Poetry is not accorded
a particularly heightened status within the manuscript.
One kind of intertextuality within the notebook might be termed "spa
tial." In MS JJ, Prelude draft (working inward from the back) is followed
by German verbs and the "Essay on Morals, and this raises an interesting
question about the effect of materials on each other. It is uncertain whether
the German verbs and the inverted "Essay" were already in the notebook
before the poetry was entered." However, it is at least possible that Words
worth started writing the Prelude material in what was a limited space.
This suggests that the physical context of the notebook may be significant
for the psychologically difficult stage of first composition. Starting at the
back seems to make a far less confident statement than starting at the front

86

WORDSWORTH'S PROCESS

of a notebook. Starting in a space that does not stretch forward indefinitely


may be easier than starting in an empty book.
The spatial entry and development of the early Prelude draft itself in
this notebook is also significant in relation to the difficulty of first com
position for Wordsworth, and I want to think about this in a little more
detail here. Both Stephen Parrish and Stephen Gill describe MS JJ Pre
lude material as developing "in a zig-zag fashion," but it is unclear what
exactly this means."" The pattern disclosed is one in which Wordsworth
seems to anticipate how much space he wants to fill in advance, working
largely in blocks of three or four sides. So, he makes the first continuous
entry on Z', working backward onto the opposite page, T. He then turns
over a page and works forwards from X' to Y', at which point he runs out
of space for reworking and finishes a rewriting of Y' on the bottom of W
^nd the top half of W. The next block is entered across V* and Win pen
cil and then recopied on W and X'' at which point Wordsworth turns the
notebook sideways for no apparent reason. The third block is a clear run
of writing over three sides from U' to V, again with the notebook turned
sideways. The fourth block again assumes a three-sided run, starting on S*
and working forward for three sides, breaking off onV with recopying of
the boat-stealing scene of P on R'. A final entry occurs with a continuous
three-sided run from P^ to
again, with the notebook turned sideways.
Although this is the earliest surviving Prelude material, it is not all firstdraft material but represents a mixture of copied text and draft. This serves
partly to explain the way the poet works, since he may know, from work
already written, roughly how long he expects a certain section to be. What
emerges, though, is not really a structure that zigzags about but a carefully
constructed method of building backward into the notebook from work
already written. Each new accretion of the text is both constrained and re^ assured by the presence of the previously entered text that exists behind it
(working inward from the back of the notebook). It thus provides a kind of
stepped structure of support. The need for this, for Wordsworth, is clearly
suggested by his extreme anxiety over first-draft composition and, at this
early stage of his career, presumably an anxiety over writing a piece of any
considerable length. He thus chooses to undertake a piece of cumulative
extended writing within a strongly proscribed space.
The characteristic of working backward, illustrated here in a physical
way, is also employed more generally in his draft materials at a level of
poetic content, according to Duncan Wu. Commenting on the structural
anomaly of Wordsworth having written the ending to the whole ThirteenBook Prelude at the Five-Book stage, he states that "the technique of be
ginning with the conclusion and working backwards is common to other

87

TEXT AS PROCESS

WORDSWORTH'S PROCESS

poems, including The Ruined Cottage and 'The Idiot Boy'; the knowledge
that, with the ending safely committed to paper, he could retrace his steps
at leisure, evidently gave Wordsworth a feeling of security.""'
Another kind of authorial intertextuality involving intellectual and
physical juxtaposition exists within DC MS 19 in the close cohabitation of
Wordsworth's fragmentary "Essay on Morals" and the early Prelude work.
The "Essay" is an explicit expression of Wordsworth's dissatisfaction with
Godwinian thought and with the form and mode of its communication.
Its central, anti-Godwinian concern is that actions depend on habit, not
reason, and he argues that it should therefore be the role of literature (and
philosophy) to change the formation of habits for the better rather than to
impose on them an unnatural rationalism. The moral value of the forma
tion of individual habits clearly compares to the kind of learning presented
in the early Prelude; its scenes of guilt and imagined punishment are con
tained immediately before the "Essay" fragment, working in from the back
of the notebook. In terms of the mode of communication as well, the sense
that The Prelude not only presents an account of a certain kind of upbring
ing but also seeks to communicate it actively, must be felt as a poetic al
ternative to Godwinian or Paleyan schemes, rejected here as insufficient
because "[t]hey contain no picture of human life; they describe nothing.""^
Poetry can convey moral truth in a more accessible way than philosophy,
"purifying thus/The elements of feeling and of thought.""' In such ways,
then, the argument of the prose fragment clearly bears a strong relation
ship with the poetic "pictures of human life" with which it cohabits in the
notebook. The "Essay" bemoans the fact that "I know no book or system
of moral philosophy written with sufficient power to melt into our affec
tions [?s], to incorporate itself with the blood 8f vital juices of our minds &
thence to have any influence worth our notice in forming those habits of
which I am speaking.""" The poetic draft material describes how:

One other characteristic of this early Prelude material in MS JJ is worth


consideration, and that is the turning of the notebook sideways (noted by
Parrish in the Cornell edition but not commented on), in conjunction with
Writing in pencil. The decision to do this seems to occur as a result of writ
ing on pages
and W, where Wordsworth writes in pencil over the verso/
recto page divisions. In further pencil drafts of this material, on pages A/B
and the inside back cover, the page is always turned sideways, and this then
seems to prompt the copying of the passage in ink, also sideways, on X' as
well as later separate ink entries.

The scenes which were a witness of that joy


Remained, in their substantial lineaments
Depicted on the brain,
By the impressive agency of fear,
By pleasure and repeated happiness,
So frequently repeated."'

Why write in pencil? There might be a number of reasons for this at a


general level both practical and psychological. If writing outside, pencil is
.more practical, particularly in an early nineteenth-century context." None
of Wordsworth's actual writing implements (apart from his inkstand) have
survived, so we cannot be sure exactly what he was writing with, or even
whether he was using a quill pen or a steel-nibbed pen, though the former
is'iargely assumed to be the case. Either way, these pens did not have any
kind of reservoir and so if used outside would have required carrying ink
as well. Pencil is both easier and more convenient. But there is also the
sense that in writing in pencil one does so with the knowledge that the
words can be erased if desired. This knowledge alone may have a liberating^effectwhether or not the act of erasing is ever put into practice. If we
compare the pages on which Wordsworth writes vertically in pencil with a
horizontal (sideways) entry, we can see that on the vertical page, lines often
slip down the right-hand margin because of lack of space, while, when he
writes horizontally, his hand is larger and looser, apparently enjoying the
luxury of length."''
I have considered this notebook so far primarily in order to work across
the" disparate contents of a single manuscript (thus revealing an alternative
perspective to that created by the Cornell editions). However, I also want
to suggest that the real power of the manuscript and, I hope, of the kind of
analysis I am arguing for, lies in the fact that we do not have to view it only
in terms of the developing meaning of the whole but also in a discrete way
that values the uniqueness of each page. A fine example occurs on one of
the turned-sideways pencil draft pages, B', where the nature of the material
form, the fading, half-readable pencil words, directly enacts the fragmen
tary meaning it contains:

While the prose piece argues for the need to act upon habit to good effect,
the poetic draft describes and exemplifies the desired process. Whichever
text was entered first, each bears upon the other.

When shape was [?not ?no] figure to be seen


Low [?breathing]
and steps
and sounds
Of undistinguishable motion, steps"'

88

89

TEXT AS PROCESS

WORDSWORTH'S PROCESS

I want now to turn to a more detailed analysis of compositional activity


on the manuscript page. Instead of responding to draft materials of The
Prelude in terms of its contingent stages, or of comparison between them,
I want here to explore the "micro-Prelude," responding at a level of com
position in action.'" This will allow for the breakdown of textual process
into its smallest constituent parts so that we can consider it as a series of
widening circles, moving outward from a close focus on a single word in
a line, to lines on the page, and then to blocks of developing work within
the manuscript. It also allows for a refinement of our understanding of the
nature of composition as a process that is both mental and physical by ap
proaching it in terms of a sequence of intentional acts. Finally, from such a
perspective, activities such as revising and redrafting can be more closely
considered, and distinct textual stages can also be defined. First-draft ma
terial is particularly appropriate for this kind of work.
It is necessary to return to John Searle's account of intention as dis-

cussed in chapter 2. In Intentionality, Searle looks closely at the relation


ship between intention as a mental state and the action that results from
it and argues for an "intention in action" that forms a part of every act,
^however small: "[T]he intention in action just is the Intentional content of
jHe action; the action and the intention are inseparable."" Intention exists
Both as a state in the mind and as the embodiment of that state as an event.
|t combines an internal condition and its external fulfillment. Drawing on
{lis earlier work in relation to speech act theory, Searle states: "There is a
double level of Intentionality in the performance of the speech act. There
'is first of all the Intentional state expressed, but then secondly there is the
intention in the ordinary . . . sense of the word with which the utterance
is made."'^ In the same way, in an act of oral or written composition at
a localized level, it could be argued that a double level of intentionality
exists, embodying compositional "speech acts" on the manuscript page.
As well as this kind of intention, there is also the presence or absence of
conscious intention on the part of the person who actsin Searle's terms,
V "prior intention."'' According to Searle, it is not possible to perform any
action without intention occurring, but it is possible to perform an action
without conscious intention: "[TJhere can be actions without correspond
ing prior intentions.... But there can't be any actions, not even uninten
tional actions, without intentions in action."'"
This study does not go quite so far as Searle in arguing that all action is
intentional. Rather, as explained in previous chapters, it envisages a move
ment between willed, conscious intention and a giving up of the self to
language (in the state of creative composition). Nevertheless, my account
still assumes a return to conscious intention and a reclaiming of perceived
ownership of language by the writer through acts of revision (which might
also lead to a further giving up of the self, and so on). In relation to this in
tentional return, Searle is very useful. He sums up the small-scale activity
involved in these terms: "[T]he whole action is an intention in action plus
a,"bodily movement which is caused by the intention in action and which
is the rest of the conditions of satisfaction of that intention in action.""
The process can be understood as a sequence, the last three parts of which
constitute the intentional act. Every return to a text (at the detailed level of
"intention in action") will be undertaken with a very clear, localized focus
on intended meaning. In terms of literary composition, then, we could say
that Searle's "prior intention" effectively corresponds to intended acts on
the page at a local level (word by word, line by line) in the core composi
tional processes of revision, correction, and rewriting.
The process can be more clearly understood in relation to a particular
example from the earliest continuous Prelude material in MS JJ:

90

91

When we respond to the page in isolation, it has a kind of beauty, I think,


not only because of the evident fragility of the meaning it contains but
also because of the unexpected and unintended enactment of its semantic
meaning in its visual form. On B' this is felt in the "stepped" appearance of
"and steps / and sounds"; in the almost faded form of "Low breathing"; and
in the "undistinguishability" of the whole as it slowly rubs away over time.
A similar effect occurs at the bottom of the preceding page, on A^ where
redrafting ends with the lines:
That looked upon me how my bosom
beat
With expectation"'
Here, the physical layout of words on the page enhances the meaning of
the words, as does the fact that the line is left suspended. The question of
whether Wordsworth intended such a meaning on the manuscript page is
not the issuealthough it seems at least likely that the powerful effect of
reading the line "How my bosom/beat/with expectation" was not lost on
him and that the page may have been left in such a way so it could act as a
stimulus for further reworking. The point I want to make is simply that a
certain kind of meaning does exist here, which is only held on this page as
a unique object.
THE MICRO-"PRELUDE": FIRST-DRAFT ACTIVITY

TEXT AS PROCESS

WORDSWORTH'S PROCESS

ridge cliff
alone
While on the perilous edge I hung"

part of the creative process because it is through small, subtle changes to


the base material that the poet gradually reclaims and refines the nature of
communication and understands it for himself

In reconstructing composition here, we could assume that Wordsworth


begins by rereading the original line and finds himself dissatisfied with the
word "edge," so he crosses it out and replaces it with one, or two, alterna
tives. In its smallest component parts, the process thus runs: prior intention
(I intend to cross the word out); intention-in-action (I am about to cross
the word out); bodily movement (pick up pen/place pen on paper); action
(physically mark a line through the word)." In this particular example the
sequence is either immediately followed by, or perhaps coterminous with,
a second sequence: prior intention (I intend to replace an unsuitable word
with another); intention-in-action (I am replacing an unsuitable word with
another); bodily movement (place pen on paper); action (physically enter

If we read a compositional text just for its relationship to the final draft,
then repeated words, deletions, and aborted passages or lines may have a
localized significance but are not considered in terms of what they them
selves actually signify. If, however, we read a compositional text in such
a way as to want to look closely at those lines because of the underlying
process and actions they reveal, then we find ourselves undertaking a dif
ferent activity and becoming part of a different kind of reading process.
The final completed version of the text is still of immense importance here
in helping to retrace authorial activity, but it is now primarily being used,
and valued, for that purpose. Reading the text in a state of process becomes
a kind of puzzle in which words on the page signify a sequence of actions,
of rapidly changing small-scale acts that can be reconstructed. This kind
of reading allows for a full intellectual engagement with draft material as
representative of creative process.

the word "cliff^' on the page).'


In literary-critical terms, the critic would interpret this intentional se
quence by analyzing the resulting changes of meaning. So, it would appear
that the word "edge" was too specific and narrow for Wordsworth's desired
communication. He therefore replaced it with the word "clifT to create
a more topographical sense of context and scale. Another intentional se
quence then followed, resulting in the addition of an optional replacement
word, "ridge," which further enlarged the context and sense of place.
It would be tedious to undertake this level of microanalysis of inten
tional acts at any great length, but it is, I think, helpful to be aware of it and
to see that the creative process for written composition is capable of being
broken down to this extent. It is also worth noting that in another artistic
mediumsuch as painting or sculpturethis kind of detailed tracing of
intentions is likely to be impossible. In the written form, however, a con
nected narrative of intentional acts is recorded on the page and can be re
constructed with reasonable accuracy. Of course, we still cannot be sure of
the amount of time occurring between any of the changes (apart from the
evidence provided by changes of ink or hand), and we cannot access the allimportant initial point of composition, which in this case produced the line
"While on the perilous edge I hung alone." We can only ever reconstruct
from beyond the point of first composition, reminding us how important,
and unique, that stage of the creative process is.
For the writer, the process of replacing one word with another, and
then, later, of physically revisiting his or her own small-scale changes of
intention, helps him or her to see a developing meaningwhich moves on
from first composition and could even, potentially, not include any of the
original lines in the final version. The words on the page become an active

92

I want to pull back a little further now, to look not at a single line but at
a single page of manuscript composition. The next two examples are both
taken from MS JJ, the manuscript containing the earliest continuous work
for book 1 of The Prelude. They concern two very well-known passages
the "raven's nest" and "boat-stealing" episodes (see figs. I and 2). The ear
liest version of the first piece is found on page X*, where a version written
above the line across the page is immediately revised below it (see fig. 1).
The first version runs:

{W
{[?]ith what strange utterance did
wind
the loud dry
Blow through my ears, lyhat colours
vfhat motion did
The CO

thedond

the lou
the colours of the sky
{not

Wh

The sky was {then


no sky
Of earth & whith what motion move the cloud,
As on the perilous brink cliff"

93

ri
'

*^2

"Jit'

"

"'k-'-Tir

t"2i-

Fz^. 1. Draft from DC MS 19 {Prelude MS JJ), X". (Reproduced by permission of the


Wordsworth Trust, Dove Cottage, Cumbria)

Fig. 2. Draft from DC MS 19 (Pre/w^^e MS JJ), P. (Reproduced by permission of the


Wordsworth Trust. Dove Cottage, Cumbria)

TEXT AS PROCESS

The passage is a continuation of material on Y' (running forward from the


back of the notebook), where the lines are written in regular blank verse.
This local context for X', as well as the nature of the work on the page, is
such that it looks very much like an example of first written composition
with no preceding oral work (an unusual state for Wordsworth).
To ascertain the order of writing and to reconstruct the compositional
process, the usual editing procedure is to look ahead to the next version
where authorial decisions and localized changes of intention are embodied
in the new text.' In this example it is easy to do so, since the next version
is written out immediately below:
As on the perilous brink cliff
ridge cliff
alone
While on the perilous edge I hung
With what strange utterance did the loud
dry wind
Blow through my ears the sky seemd not
a sky
Of earth, and with what motion moved
the clouds^

What emerges on
is a model of half-line composition for this first-draft
material, which provides the poet with a number of different possible
combinations. The initial entries on the page above the line are probably
the two top half-lines "With what strange utterance did/Blow through
my ears," with the word "wind" either also entered then, or at the time
the words "the loud dry" were entered below the line. Initially, then, the
lines could have read "With what strange utterance did blow through my
ears/The loud dry wind" as well as "With what strange utterance did the
loud dry wind / Blow through my ears.""
The next revision on the top half of the page perhaps occurs after the
decision to place "loud dry wind" between the first two half-lines, and so
now concerns the second half of the second line, which needs to be filled.
Wordsworth plays with two options ("the colours of the sky" and "what
motion did the cloud"). When he finally rewrites the passage, the first halfline option ("the colours of the sky") disappears. This may suggest that the
line immediately below these workings, beginning "the sky was then no
sky," is a revision of "the colours of the sky," which then leads to the entry
of the two lines below, one using some of the crossed-out revision ("whith

96

WORDSWORTH'S PROCESS

what motion move the cloud"), one of them a new half Hne, "As on the
perilous brink cliff."
In the third stage of composition, below the line across the page, Words
worth (either immediately, or on returning to the passage) reaUzed that
this Hne would work well as the start of this section and so began again,
placing it first. Changes here are detailed"brink" is rejected, perhaps becaiise it suggests the position of feet rather than hands. "Edges" (as we have
already seen) is crossed out as being too unspecific. "Ridge" or "cliff" are
retained as options, for both of which assonance with "perilous" immedi
ately before them in the line creates a poetic effect.
An editor of a critical edition reading this will be asking, "How is this
different from the kind of activity I undertake?" The answer is that it is not,
in essence, different. In part, this book aims to create more space for such
activities to be articulated, providing points of connection between textcritical and literary-critical activity The distinction occurs only in terms
of the purpose for such analysis. Where the editor undertakes such tasks
primarily to determine the sequence of draft material within a notebook or
across manuscripts, and is unlikely to articulate it for the reader, the com
positional" or "genetic" critic can undertake such tasks to understand the
rj'ature and process of composition, or to enrich understanding and analy
sis of the text by knowledge of its development.
A second example will perhaps clarify ways in which the detailed study
of composition can provide an alternative perspective on the completed
text. The famous boat-stealing scene provides an interesting example of
revision being used in different ways within the compositional process. On
the manuscript page of MS J J, V, this passage reads:
Of sparkling light
When from behind that rocky steep till then
The bound of the horizon just between
e}
The summit & the stars a hugh} high cliff
As if with voluntary power instinct
the oarsand

Upreared its head I truck again

struck
again

And growing still stature the huge


With measured motion like a living
thing

Strode after [?me]


Rose up between me & the stars & still

97

TEXT AS PROCESS

WORDSWORTH'S PROCESS

With measured motion like a living


thing
Strode after me"

When one reads this manuscript page for the first time, the amount of rep
etition within it is striking. The example I want to look at is on the sixth
line, where "struck again" is repeated at its end. The nature of this repeti
tion, and the author's compositional intention here, is not absolutely clear.
It may well be that Wordsworth intends the second "struck again" to be a
deliberate repetition of the first,,for poetic effect, as it is used in the final
version. Alternatively, though, it may be an example of simple recopying
over a rejected revision (probably "struck the oars again") that had made
the original words unclear on the page, so that the author needed to reen
ter them. In this case the powerful repetition of "I struck & struck again,"
as incorporated in the next revised version (on R^), could come about as a
result of recopying for clarification on the page, which, half-accidentally,
reveals it as a creative option to the poet. Further repetitions of "With mea
sured motion," "strode after me," and "huge clifF' also occur on P.
The repetitions on the page of DC MS 19 are fairly typical of composi
tional text, where redrafting obviously creates similar versions of lines or
repeated words on the page. What is less typical is that two repetitions in
close proximity should then be used for dramatic effect in the completed
text, in which the passage acquires a large part of its power from the use
of repetition that unites the boy's movement on the lake with the shifting
perspectives created by it and the consequent internal effects on himWhen from behind that rocky steep, till then
The bound of the horizon, a huge Cliff,
As if with voluntary power instinct,
Upreared its head; I struck, and struck again,
And, growing still in stature, the huge cliff
Rose up'
Is it simply coincidence that so much repetition should occur on the man
uscript page for this passage? Possibly But it is also possible that authorial
activity here is actually concerned with "unintentional" composition. That
is, the misreading of a revisionary act on the manuscript page can reveal a
creative development to the author that he had not consciously intended.
It is interesting, however, that we as critics can only really access this free,
random aspect of creativity through a highly structured and close analysis
of intentional acts. The emergence of the "unintentional" depends on its
identification through the structures of intention.

98

Jonathan Wordsworth, in a discussion of revision in the long poem,


states; The history of the long poem serves to emphasize that composi
tion is in its nature backward-looking. Every time a poet seeks for the right
word, he revises an earlier expression, on paper or within the mind. Every
time he seeks to clarify an idea, he revises an earlier conscious thought,
an earlier pre-verbal hunch." This account (while strongly authorial)
seems to support the kind of recqnstructed model of intentional acts that
I have undertaken here, as does his comment that "revision, in the sense
of changing authorial intention, seems to be inevitable in the making of a
lotog poem."''
Close study of compositional material for its own sake begins to raise
more complex questions about the nature of revision and redrafting. Is the
kind of revision taking place in this early draft material, a[t the most active
stage of the compositional process, of the same nature as that which takes
place later in the process, or even at a much later date, when Wordsworth
returns to an earlier text?*
The difference may be merely one of degree, so that in early composi
tion a large number of words within any section of a text are changed and
a larger number of possible options exist within a line and between lines.
In later revision, the number of changes is considerably reduced, as is the
likelihood of redrafting at any length. This would seem to suggest that the
nature of revision is the same, only lessened in intensity, at a later time.
One of the characteristics of early draft material is a willingness to allow
change and experimentation. Different alternatives exist on the page, and
there is a full range of possibility both for the localized composition and
within the wider structure. As the text develops, particularly for a long
poetic structure, the degree of possibility is closed down, because each in
dividual section is affected by what occurs around it. This suggests that a
context of phases or stages of composition is an important element of the
compositional act, and one that affects the nature of the revision under
taken. For first-draft material, the wider context is likely to be undefined
and not yet physically embodied, and revision is therefore still able to lead
to considerable changes in meaning, mood, or tone, if the poet requires it.
The example given earlier ("struck and struck again") may illustrate how
Creatively revision can function within this early stage. For later mate
rial, however, revision becomes increasingly circumscribed, because local
meaning is affected by the wider contextin terms of mood and style as
well as contentand the poet will not want to make a local change that
might destabilize the whole. By the time of a full fair copy-text of a ma
jor work, or revisions to a published edition, the existence of the stable,
,fixed material around it must affect the nature and extent of the localized

99

TEXT AS PROCESS

WORDSWORTH'S PROCESS

revision. So, although the physical act on the page remains the samethe
crossing out of words or lines and replacement by othersthe potential
ramifications of the revision must be controlled, and this will affect the
nature and range of revisionary acts.

I want now to look at the ways in which the poet's own sense of his role
and his representation of it within the "final" text can seem at odds with
the compositional acts and strategies that have produced it. I will move
between the draft materials and the published work, between writerly selfidentity and projected poetic identity. In Wordsworth Writing, Andrew
Bennett considers two related issues: the "paradox of the poet for whom
words are immaterial" and the linked conception that Wordsworth is "a
poet who doesn't write poetry." My concern here is also with this para
doxical, doubled position in relation to Wordsworthian composition, but
I will be exploring it in, and through, the materials of process. My aim is
to consider Wordsworth's "compositional contradictions" at both a repre
sented level in the completed work as well as in the underlying drafts. Two
key areas of contradiction for Wordsworth are the linked issues of orality
and spontaneity.
Orality allows the poet to retain a sense of the words as "his" in a way
that twentieth-century philosophy has shown to be an illusion. The French
phenomenologist Maurice Merleau-Ponty asserts: "For the speaking sub
ject, to express is to become aware of; he does not express just for others
but also to know himself what he intends."^ In other words, "the listener
receives thought from speech itself."^' For Merleau-Ponty, "spontaneity,"
if it can be said to exist at all, does so by means of the self-generation of

ing "above all other external powers a dominion over thoughts."^" Much
critical work has of course been undertaken in this area (by Paul de Man,
J. Hillis Miller, James Chandler, Frances Ferguson, Mary Jacobus, and
most recently Andrew Bennett) in relation to the self-division of language
as spoken and written discourse and Wordsworth's doubled consciousness
of it.
Oral composition stands as a kind of bridging state for Wordsworth
between what he sees as pure interior poetry and the debasement of the
written word. The spontaneous ideal is bound up with orality so that an
"overflow" can pour forth directly from the body of the poet with emo
tional expression and verbal expression still connected. In a discussion of
lyric voice in The Prelude, Mary Jacobus describes Wordsworth's reliance
on poetic voice as a "surrendering to an auditory myth of self-presence."^'
She points out that Wordsworth does not go so far as to envisage himself
as the ancient poet composing to the harp but that he draws instead on
the idea of listening to himself, "the myth of the 'inner ear,'" so that com
position is understood as the external recording of what has already been
experienced as sound.^ She concludes: "No wonder Aeolianism pervades
The Prelude; it is Wordsworth's defence against that inability to hear one
self think (or speak) involved in writing itself."'''' Orality creates an illusion
of spontaneityas a direct outpouring of self-generated creativity. Even
if lines are actually composed one by one and then written down, or even
if there is considerable oral and mental revision before writing, the poet
can feel as though the process depends on internal rather than an external
(textual) stimulus and as though creativity is somehow occurring outside
the words in which it is uttered. The act of dictation, or of dictated revi
sion, is another way of creating this illusion in relation to the written word,
and Wordsworth employs all of these methods in his own creative
processes.^

language.
In a Derridean sense, Wordsworth is clearly "phonocentric," viewing
written words as secondary to speech and valuing an "absolute proximity
of voice and being."'^ When Derrida describes a Rousseauvian position
as tending to "confine writing to a secondary and instrumental function:
translator of a full speech that was fully present," he could equally well be
describing Wordsworth's position (which is hardly surprising in view of
the two writers' shared historical and cultural context).^' At the same time,
though, Wordsworth's very anxiety with an internal, self-generated model
of composition and with an assertion of mental and oral over written
creativity suggests that he recognizes the "awful" power of words, hold

There is, clearly, an element ofself-deception here. How"oral" is Words


worth's poetry? Certainly he composes outdoors often, and by uttering
words aloud, but does he memorize large sections or does he work by
speaking a line or a few lines aloud and then writing them down? Detailed
evidence does not seem to be available as to how long or how much Words
worth composed orally before writing. Andrew Bennett skeptically sug
gests that "the image of Wordsworth engaged in walking and composing
has as much to do with what he told people, particularly later in life, as it
does with his actual compositional practice of at least the early years."^' In
any case, in whatever way Wordsworth made the transition from voice into
word, this activity is very different from being a truly "oral" poet, who, in

100

101

COMPOSITIONAL CONTRADICTIONS (SPEECH, WRITING,


AND THE PRODUCTION OF THE LONG POEM)

TEXT AS PROCESS

an oral culture, would never seek to create a written version of the text and
for whom each telling would be a new and unique experience.
Occasionally, and in an indirect way, Wordsworth seems to acknowl
edge that the "spontaneous" is a false compositional construct. In letters to
younger poets, he repeatedly asserts the need for the poet to "labour" and
take time over composition. In a letter to William Rowan Hamilton in 1816,
he states; "Again and again I must repeat, that the composition of verse is
infinitely more of an art than Men are prepared to believe
Milton talks
of 'pouring easy his unpremeditated verse.' It would be harsh, untrue and
odious to say there is anything like cant in this; but it is not true to the
letter, and tends to mislead. I could point out to you 500 passages in Mil
ton upon which labour has been bestowed, and twice 500 more to which
additional labour would have been serviceable."" There is a contradic
tion, then, between Wordsworth's own representation of the poet as orally
"spontaneous" and the production of that (self)representation which, as he
is well aware, depends heavily on a written, revisionary process. Words
worth, too, "tends to mislead."
A model of "spontaneous" oral composition can be represented in a more
convincing way for a poem of no great length than for a long work. There
is the possibility of direct correlation between poetic composition and the
representation of poetic composition. This is not possible in the same way
for a long poem-where both the poet and the reader know that such a poem
cannot have been created in a single spontaneous act. The Prelude (1805) is
written over a seven-year period; The Excursion is probably written over an
eight-year period, with the poet, inevitably, laying work down and picking
it up again for both poems during this time. Yet in The Excursion, as I have
previously discussed elsewhere, the Poet character is not represented as an
actively creative figure, and not at all as a writer.' In The Prelude, which
takes as its subject the means by which its subject becomes a poet, Words
worth does refer to the lengthy poetic process, but he does so in ways that
either suggest orality or avoid representation of the written.
In The Prelude, the emphasis on poetry as a spoken rather than written
discourse can be dramatically illustrated by referring to Lane Cooper's A
Concordance to the Poems of William Wordsworth. For the word "speak"
there are thirty examples in The Prelude, as well as eight uses of the word
"speaking" and nine of "speech." By contrast, for the words "write" and
"wrote," there are no uses whatsoever in either The Prelude or The Excur
sion, and the word "writer" is not included in the Concordance at all.^
There can be no doubt that the poet is to be seen to speak freely rather
than write his message.

102

WORDSWORTH'S PROCESS

As is well known, opening sections of books within The Prelude often


refer to the development of the poem itself, and the sense of its unfolding
length, in terms of an oral discourse. Each time the poem pauses to discuss
its own progress, usually in the first section of a book, it evades any men
tion of the act of writing:
Of these, said I, shall be my Song, of these
If future years mature me for the task
Will I record the praises, making Verse
Deal boldly with substantial things
[T]hus haply shall I teach.
Inspire, through unadulterated ears
Pour rapture, tenderness and hope'
Wordsworth often uses words that describe the act of writing but without
dwelling on the physical aspect of it in any way. One word commonly used
in this manner, which occurs frequently in both long poems, is the verb
"to tell"the poet writes of what is "told" and "untold."" This emphasizes
what is being done (a tale being narrated, an account being given) rather
than how it is being done. As such, it is probably understood to describe an
oral rather than a written act, but it could describe either. A particularly
clear example of the deliberate use of it to avoid direct representation of
the act of writing occurs with the very last word of The Excursion where
the Poet-character concludes with the line "My future Labours may not
leave untold."' The line refers to the written act of the recording Poetcharacterand implicitly also to Wordsworth as the potential author of
future parts of "The Recluse"but the poem refuses to directly acknowl
edge either as a writer. Other words, such as "record" and "relate," are sim
ilarly employed as verbs that are concerned with communication but place
emphasis on the purpose of an action rather than on its means.
On the rare occasion where the act of writing is represented within
The Prelude, it is with negative connotations. So in book 4, on his return
to Hawkshead after Cambridge, the poet compares the brook, controlled
within the garden and flowing without effort and at the loss of its true
character, to himself:
And now, reviewing soberly that hour
I marvel that a fancy did not flash
Upon me, and a strong desire, straitway

103

TEXT AS PROCESS

WORDSWORTH'S PROCESS

At signs of such an emblem that shew'd forth


So aptly my late course of even days
And all their smooth enthralment, to pen down
A satire on myself.

I want to return to Wordsworth's draft material to locate within it exam


ples of first written composition. Such material represents the moment of
transition and translation from thought, or voice, to written word and thus

the vital link between oral and written poetic composition about which
Wordsworth is so uneasy.
First-draft material is often the least likely manuscript material to sur
vive, for two possible reasons. The first is simply that first draft is of least
value to the poet during the active creative process: he or she rapidly moves
away from it and no longer needs it. At a practical level, first draft is there
fore far more likely to get lost or mislaid than, say, the final fair copy-text
of a poem. Second, a poet might well choose to destroy first-draft material,
from an anxiety to protect the source of creativity or from a sense that
the material is much improved at a later stage. Moreover, there is always a
question of uncertainty over first-draft status: we can always identify the
earliest surviving draft of any passage, but we cannot always be sure that
it is actually the first written composition. The precarious nature of first
draft in terms of survival is likely to be reflected in its physical characteris
tics on the page. In general, first-draft material involves rough handwriting
(the poet is writing fast and getting ideas down), and it may contain a con
siderable amount of crossing out and revision, though this is not always
the case (if the writer undertakes considerable mental composition first,
for example). We also need to remember, then, that any discussion of "first
draft" is always about "first written draft" and that it may well have been
preceded by considerable mental or oral "rewriting."
The Prelude manuscript I want to draw on here is Prelude MS WW (DC
MS 43), a pocket notebook containing early draft material for books 3-8
and book 12 of The Prelude, along with Dorothy's notes for the Scotch Tour
of1803. The Cornell edition of the Thirteen-Book Prelude gives such a clear,
enlarged photographic reproduction of the manuscript that it is a surprise
when faced with the material object to see how fragile it is. The notebook
embodies the vulnerability of its material in various ways. First, it exists in
loose leaves, having been taken apart because the siblings were sharing it:
"The dismantling of the notebook may have resulted from a tug between
Dorothy's need for her notes . . . and Wordsworth's wish to use its blank
pages for notes while out walking."" Second, its entries are all written in
pencil, making it vulnerable to loss of clarity, and the words are extremely
faint on the page. Third, it seems likely, as suggested earlier, that it was car
ried around outside: "The roughness of Wordsworth's drafts strongly im
plies, as Jonathan Wordsworth suggests, that many of them were written
outdoors."'^ The pages are also extremely small (10.5 X 8.45 cm). Because
of its loose-leaf nature, a clear order of entry within the notebook cannot
be established, although blocks of work are clearly entered together, and
sometimes the paper is partially joined.

104

105

Writing occurs at a level of doubled mockery: the adult poet judges his
youthful self and envisages his own self-judgment. The written act is imag
ined and indirect. Elsewhere in the poem, a sense of the "written" is more
commonly associated with the indirect act of reading words than with the
direct act of producing them. In book 5, Wordsworth describes the power
of books on the mind but bemoans the limitations of words as the only
medium of external communication:
Oh! why hath not the mind
Some element to stamp her image on
In nature somewhat nearer to her own?^
As a result, books are viewed as frail objects ("Poor earthly casket of im
mortal Verse!") in which (written) words wait to be released by the liv
ing voice. The act of writing is secondary to feeling^"Why call upon a
few weak words to say/What is already written in the hearts/Of all that
breathe!"and to the true communications of natural men:
Theirs is the language of the heavens, the power.
The thought, the image, and the silent joy;
Words are but under-agents in their souls'
The Prelude is a poem that values its own creation, but not the medium
through which it finds expression. The final poem's refusal to explicitly ac
knowledge the creative power of writing and the poet as writer means that
the relationship between completed textual product and the process that
brings it into being mirrors the model of divided self-representation at the
heart of the poem. The process of making voice into word produces words
that present poetry as voice in what Bennett calls the "inevitable paradox
of a writer writing about his poetry as speech."'
"THE PRELUDE" AND FIRST (WRITTEN) DRAFT

TEXT AS PROCESS

{ten
Of holy chare {a

a shepperd
[?pries]

[?e]
[?When there co]
[?When]
A father of his people
When [?he]
good
Is [?Said]
and as the [?best] King''
This looks like an example of material that has not previously been articu
lated anywhere else. Instead of having lines whole and entire, the sound of
them clear in the head before he begins to write, Wordsworth finds himself
grappling here with each word. Where oral composition might be con
cerned with the overall sound and fluidity of a short passage (perhaps the
unfolding of a contained narrative, or scene), this passage of written com
position concerns a more prosaic moment of introduction and the estab
lishment of character. This suggests that mental or oral first composition
may not work with meaning in the same way as written first composi
tion can. The practical use of memory to retain text in the mind means
that oral composition is likely to be concerned with short blocks of work,
memorized images, or core lines. Although revision and redrafting can, of
course, also take place in the mind, written composition is far better for
allowing the playing out of simultaneous options, as can be clearly seen in
this case. The first written act thus allows for a different kind of imagina
tive activity from the oral one. It is less fluid, and perhaps less impressive
as an "overflow," but it also operates in a different way.
Finally, in Prelude MS WW, the famous "Imagination" passage from
Wordsworth's crossing of the Alps is interesting because in its original con
text it seems to be an example of first written composition commenting on the
act of first written composition. Wordsworth presents the effects of "Imag
ination" as activelyintervening on spoken or written composition ("verse"):
A little while [?Imagination] crosd
me here
{n

Like a{[?] unfatherd vapour, & my


verse
{aused

Halts in mid course. I p{a

108

[?was]
[?in] cloud

WORDSWORTH'S PROCESS

[8 draft lines here]


as in

I paused was lost awhile [ ? ]


a cloud
lost awhile

A cloud which ere my verse


came to [?an end]
Some azure mist & [?yet tis]'
Originally, the passage articulates a sense of the writer as "lost" in lan
guage and possessed by it. When the passage finally appears in 1805 after
the "Crossing of the Alps," however, "verse" (which could be oral or writ
ten) has become Song, and the effect of the imagination is described in
much less creatively interventionist terms:
Imagination! lifting up itself
Before the eye and progress of my Song
Like an unfather'd vapour; here that Power,
In all the might of its endowments, came
Athwart me; I was lost as in a cloud,
Halted without a struggle to break through"
Wordsworth does still attempt to describe the nature of poetic process ex
plicitly in the "final" text, but the very nature of that text as a stable product
works against what he is communicating. By comparison, the uncertain,
fragmentary repetitions of the draft version enact what they describe on
the manuscript page as the verse "Halts in mid course" and the lines repeat
themselves: "in cloud/I paused was lost awhile/a cloud/lost awhile/A
cloud."

SPEECH, WRITING, AND COMPOSITIONAL


MOMENTUM IN "THE EXCURSION"
The elision of poet as writer also occurs in The Excursion. As I have previ
ously discussed elsewhere, the Poet-character is not represented as an ac
tively creative figure, and not at all as a writer, so that a discrepancy be
tween past and present, between the apparent "now" of the represented
utterance and the "then" of writing, is strongly felt.'"" We can explore such
issues further across the final poem and the manuscript materials of the
poem by looking at the handling of openings and endings of individual
books that are directly connected to the oral representation of the Poet.
One structure that contributes to compositional momentum for a long

109

TEXT AS PROCESS

poem is the sequence of repeated endings and new beginnings for each
book division within it. Such a pattern provides a kind of structural rep
resentation of the spontaneous act as sequence, a constant sense of fresh
beginnings within and across a larger framework. Beginnings and endings
are often very firmly marked on the page by both William and Dorothy in
Wordsworth's manuscripts. An early example of this occurs with DC MS
16. It contains a decoratively written heading for "Adventures on Salisbury
Viain I Part First" on 28' in Dorothy's hand and again for "Adventures on
Salisbury Plain/Part second" on 40'. Also in this notebook, The Ruined
Cottage receives the same treatment with "End of the first Part" on 49" and
"The End" written very clearly on 56''. Of course, at one level these head
ings exist simply for clarification, but they also emphasize the way in which
a sense of regular "beginnings" and "endings" provides one of the compo
sitional rhythms of the manuscript notebook.
In The Excursion, endings andbeginnings are directlyrelated to speech.
We can see this clearly in an example from the end of book 6 and start of
book 7 where the break occurs as a shift from direct speech to the Poet's
reflection on what he has heard:
'But each departed from the native Vale,
In beauty flourishing, and moral worth.'
While thus from theme to theme the Historian passed,
The words he uttered, and the scene that lay
Before our eyes, awakened in my mind
Vivid remembrance""
The same structure is dramatically reversed for the end of book 8 and start
of book 9:
as One
Who from truth's central point serenely views
The compass of his argument.began
Mildly, and with a clear and steady tone.
'To every Form of Being is assigned,'
Thus calmly spake the venerable Sage,
'An active principle''"^
The inversion at this point means that the opening of book 9 coincides
with direct speech for the only time within the poem, placing considerable
emphasis on the Wanderer's important final speech.

110

WORDSWORTH'S PROCESS

For every book of The Excursion, the ending consists either of the end
of a speechthe half-spoken apostrophe of the Poet ostensibly "thinking
aloud"or the Poet's framed response to the spoken utterance. Such mate
rial forms a "written" framework that must be taking place at a time distant
from the dramatic action of the poem, yet the poem refuses to acknowledge
the distancing act directly, so that even these frames are presented as if they
1 occur at a time contemporaneous with the actual utterance. In part, this
krves perhaps to make a central Wordsworthian point about the nature
of poetry: that you can be a poet "silently" without writing words down;
that true poetry is about feeling and shared response. In part, this creates
a work in which the poet must be recording, memorizing, and recalling all
that is narrated, but the main part of the narrative is always presented as if
it is speech or response to speech occurring at the time that we hear it.
When one looks across the poem as a whole, two basic models for book
endings in The Excursion emerge. The dominant model is of an ending
that concludes abruptly with the last line of a speech or story, and this
holds true for books 3,5,6, and 7. In these cases, the ending of the previous
book determines the nature of the beginning that follows it. The start of
book 7, already discussed, provides an excellent example of the interlockipg structure of speech and response, in which the Poet's account allows
the narratives to flow on around the opening of the next book.
A second kind of ending concerns a doubled structure, with a con
clusion first of speech and then of the book. For book 4 in DC MS 73,
workings on 6^-7' and 7" all relate to the conclusion of the Wanderer's dis
course, which is presented as an ending within the ending: "Here closed
the Sage."'"^ The contents of the book are framed by the presence of the
Poet, who comments on what he has heard and returns the reader to a parHcular time and place. In the final version of book 4, the close of speech
is followed by the close of day, and the ending of the book. This kind of
ending also occurs for books 1 and 2. Again, the book ending is preceded
by a clear statement of the end of a speech-"He ceased"; "So ends my do
lorous Tale.
These books finally have a triple closure (speech, day, book).
Such endings also affect the openings of the ensuing books (2, 3, and 5),
which do not have to define themselves directly in terms of linkage to what
precedes them and tend to concern themselves with an aspect of scenery,
nature, or the journey.
Significantly, the conclusion of the entire poem takes this sense of more
than one ending even further. At the end of book 9 we are given, first, the
conclusion of the Pastor s speech"This Vesper service closed"'"'then
the Solitary's brief words and departure, and finally, the conclusion to the
poem itself as given by the Poet-character:

111

TEXT AS PROCESS

From this communion with uninjured Minds,


What renovation had been brought; and what
Degree of healing to a wounded spirit,
Dejected, and habitually disposed
To seek, in degradation of the Kind,
Excuse and solace for her own defects;
How far those erring notions were reformed;
And whether aught, of tendency as good
And pure, from further intercourse ensued;
This(if delightful hopes, as heretofore,
Inspire the serious song, and gentle Hearts
Cherish, and lofty Minds approve the past)
My future Labours may not leave untold.'"
The sense of repeated endings, or of an absolute ending repeatedly de
ferred, is felt in the published text at a syntactical level in the irresolution
of the final sentence with its "What... & what... And whether."
Draft toward the ending occurs in only one surviving notebook (DC
MS 73), where it is entered twice. The first version occurs on 46'; the second
is a fair copy of this on IS'. The conclusion is strongly visual on the page
in both cases; as a result of the word "End" written below the first version
and "Finis-" below the second (see figs. 3 and 4). The poem, then, emphat
ically affirms that it is to end in an unresolved way.'"^ On the manuscript
page, such irresolution is further reinforced by changes made to that last
line:
And gentle hearts & lofty minds approve,
{may
{[?may]

{leave

My future Labours shali not {[?le]


untold
End.
untold

My future Labours may not leave tmtold.


Finis
Thus, the final line bears the meaning "My future Labours may tell" but is
unable to voice that meaning directly and instead brings with it the strong
possibility of what might not be told.
All these levels of uncertainty also point toward the sense of an exter
nal reference for the poem's ending. Susan Wolfson notes of the Poet: "The
muted interrogatives of his concluding lines, moreover, include himself

112

Fig. 3. Draft for the end of The Excursion. DC MS 73, 46'. (Reproduced by permis
sion of the Wordsworth Trust, Dove Cottage, Cumbria)

WORDSWORTH'S PROCESS

few^Sc^'i-.

.4/^M

prffeia#, ,iS' r'. .'


't . ,

l'
'^i^i't i

jr; ,:-rf;',

\.. :...

4. ir

'% 4^: ^^; "'"M3:> iv"!;

i''-j.. -;. <.51. . s^'- ' ' "Sf"^


'
' ' V'.ss-::i;i;:tPig. 4. Draft for the ead of The Excursion. DC MS 73,18'. (Reproduced by permis
sion of the Wordsworth Trust, Dove Cottage, Cumbria)

as well as the Solitary: 'What renovation,' 'What/Degree,' 'How far,' and


'whether aught' consign the Solitary's reformation to uncertainty, and the
conditionals if and may suggest the Author's doubt about the reception and
renewal of his own efforts.""" It certainly seems as if at this point, there is a
sense of Wordsworth's own impinging anxieties about his wider ambitions
outside the dramatic framework of the poem. From this perspective, the
doubts in the Poet-character's conclusion about his ability to "tell" any
thing further also read like a self-fulfilling prophecy for Wordsworth writ
ing "The Recluse." The nature of the ending works against the very idea of
there being any such confident assertion, either at a dramatic level in the
poem or for the poet himself.
It might be assumed that the kind of fluid, dialogic response structures
of the endings and beginnings of each book of The Excursion reflect a com
positional structure driving the writing of the poem, but for the later books
at least, this does not seem to be the case. Instead, we find many of these
endings and beginnings clustered together. A single manuscript notebook,
DC MS 74, contains a total of seven beginnings and endings in the follow
ing order: the end of book 7 and start of book 8; the end of book 8 and the
start of book 9; the end of book 6 and the start of book 7; and the end of
book 5. For each of the endings and beginnings within this manuscript, the
version contained there is either the only, or the earliest, surviving version.
This seems to signal that the bringing together of the end and start of the
later books is a conscious, and late, act within the compositional process
the fact that the manuscript contains such a number of these points of
connection (without containing the whole of each book concerned) sup
ports such an idea. The nature of the beginnings and endings in DC MS
74 suggests that the act of linkage is not integral to the creative process, in
this work at least. Beginnings and endings are largely a means of stitch
ing larger blocks of material together once that material is written rather
than providing the impetus to write material in the first place. At the same
time, such a structure is, in a sense, highly appropriate for this particu
lar poem. The self-conscious stopping and starting, the continuous "new
beginnings," reflect the uncertainties and changes of position embodied in
the character of the Solitary whose possible redemption lies at the heart of
the narrative.
Alongside this account of a late linking-together of books, however, we
do need also to be aware of an alternative model of self-generating begin
nings that exists for earlier books of the poem, by drawing on material
previously written for another context. The openings of books 2, 4, 5, and
9 all import material from elsewhere. A number of interesting examples of
The Excursion's use of "prebeginnings" as part of long poetic composition

115

TEXT AS PROCESS

can be found in DC MS 70. This is evident with the very first entry within
that notebook, for book 4 on 16' beginning "Happy is he." This passage is
originally found in Home at Grasmere MS R where these lines are not in
corporated into MS B. Wordsworth thus begins book 4 of The Excursion.
which was to be the philosophical heart of the poem, with a dedaratwn o
the need for human understanding of all forms of life that was
^
written for an entirely different context. As such, it can be compared to the
opening of book 2, in which the description of the wandering Minstrel is
lifled from DC MS 48, and the opening to book 9, in which the Wanderer s
speech concerning "an active principle""" is taken from unused early Pre
lude material in DC MSS 15 and 16. In each case, the piece of earlier writ
ing seems to provide a base from which fresh composition (at a key point
within the later poem) can develop."'
Is there a sense, then, in which Wordsworth's "compositional contra
dictions" lead him to want to envisage long poem composition as some
kind of "spontaneously" self-generating structure? Certainly it does look
as though the concept of "The Recluse" works through The Excursion hj
providing an intertextual "charge." Different elements of text, written for
different parts of "The Recluse" and existing as a large central mass of unsituated material, can be drawn on in a process of textual self-generation.
The characteristic Wordsworthian processof drawing on the past to cre
ate in the presentis here played out in literal, material terms.

yioce^^

He spoke of the diseased craving to have all the trifles of a man of


genius preserved, and of the positive crime of publishing what a poet
had himself deliberately suppressed. If all the contents of a poet's
waste-basket were taken out and printed, and issued in a volume, one
result would be that the things which he had disowned would be read
by many to whom the great things he had written would be unknown.
William Knight, "A Reminiscence of Tennyson"
As with the chapter on Wordsworth, this chapter is concerned with apply
ing theoretical ideas discussed earlier to the materials of process and thus
with using Tennyson, in part, as a second "case study." However, it also
hopes to offer a full critical exploration of the Tennysonian process. The
first half of this chapter will therefore consider the different ways in which
Tennyson "self-translates" as an essential part of his compositional practice,
exploring his process in terms of a going out from the self and a return to it
mediated by different forms of self-translation, as seen in Idylls of the King.
The second half moves on to a consideration of the complex hermeneuti
cal structure created by Tennyson's long-term composition of the Idylls. It
is still necessary to begin, as we did with Wordsworth, by considering the
state of contemporary access to Tennyson's manuscript materials.
Policies of compositional protection, initiated by Tennyson and Hallam
Tennyson, have partly retarded Tennyson scholarship. Tennyson deeply
resented the treatment of the poet as a literary celebrity, complaining in
his own lifetime about the intrusions created by unwanted visitors and
autograph hunters. In anticipation of Wimsatt and Beardsley, he seems
to have viewed the study of draft materials as an extension of misplaced

116

117

TEXT AS PROCESS

development of a "compositional hermeneutics" that might explore further


the extent to which a specialist hermeneutic is needed for text as process,
the extent to which this can or should be reconstructive rather than deconstructive, and the question of how far interpretative activity should seek to
separate process from the final work of art. In this work, my primary aim
has been to "come into [the circle] in the right way,"^ but the ceaseless,

o4.e>:>

turning act of understanding still goes on.

ABBREVIATIONS
Bd-T
CGET
PLT

Martin Heidegger. Being and Time


Hans Walter Gabler, George Bornstein, and Gillian Borland Pierce,
eds. Contemporary German Editorial Theory
Martin Heidegger, Poetry, Language, Thought
INTRODUCTION

1. Gabler, "Genetic Texts," 62.


2. Such accessibility raises a number of questions concerning the paradox
that increased electronic access can lead to decreased access to the original man
uscript and about the value of the original physical object over accurate virtual
representation.
3. Stillinger, Coleridge and Textual Instability, vi.
4. Ibid., 119.
5. Ibid., 25.
6. Hershel Parker, Flawed Texts, x.
7. Ibid., 23.
8. Ibid., 3, ix.
9. Ibid., 182.
10. Ibid., 3-4.
11. Ibid., 23, 81.
12. Gabler, "Unsought Encounters," 158.
13. Bryant, Fluid Text, 2.
14. McGann, "Pound's Cantos," 38.

238

239

NOTES TO PAGES 5-13

NOTES TO PAGES 13-19

15. John Bryant comments on two aspects ofMcGann's approach that impede
ana ysis of text as process: "[H]e de-emphasizes the role of authorial intention in
the shaping of a social text and seems to remove the crucial element of an individ
ual writers creativity from the shaping of bibliographic codes" (52).
. w o recent collections o fessays give a g o o d overview o fthe historical
thTfh

translating key papers in

and Boria^nd

bornstein.

textes (2004), ed. Deppman, Ferrer, and Groden. Special editions of journals are
Yal T hZ
^
overview of critique genetique; see in particular
Yale French Studies 89 (1996) and Romanic Review 86.3 (May 1995).
1. CONTEXTUALIZING PROCESS
L See particularly Tanselle, "Greg's Theory of Copy-Text" and MrC^nn

Critique of Modern Textual Criticism.

t.
of "recensio" and "examinatio" at work see Maas, Textml Crttiasm. The practice is also described by Gaskell. New Introduction to
CritiST'
3.
4.
5.
6.

"Classical, Biblical, and Medieval Textual

McGann, Critique of Modern Textual Criticism, 15.


Tanselle, "Greg's Theory of Copy-Text," 176n7.
Greg, Collected Papers, 382.
Parker. Flawed Texts, 59.

7 Bowers. "Some Principles for Scholarly Editions " 223

9. Ibid.. 4.
10 Tanselle, "Greg's Theory of Copy-Text." 194. In this article. Tanselle de
fends the princip es of the CEAA and responds to various criticisms of the center
f "TK r
.
of responses to the CEAA principles is given on pp 591-92
of The Center for Scholarly Editions: An Introductory Statement."
. Tanselle s article The Editorial Problem of Final Authorial Intention"
thiltimr
of footnote references to other participants in the debate at
a te
of mtendonality in terms of literary critical and philosophi
cal terms will be reconsidered in chapter 2 of this work.
12. Gaskell, New Introduction to Bibliography, 339

) GaskelJ attacks CEAA principles (183-95). This leads to a critique of his


own work by Tanselle in "Recent Editorial Discussion." 51-55.
14. Thorpe. Principles of Textual Criticism, 193.
15. McGann, Critique of Modern Textual Criticism, 48.

240

16. Ibid., 104.


17.
18.
19.
20.
21.

Shillingsburg. Scholarly Editing in the Computer Age, 88.


Ibid., 77
Ibid., 17
Ibid., 29.
Ibid., 89.

22. Shillingsburg, "Inquiry into the Social Status of Texts," 74.


23. Gabler, "Text as Process," 108.
24. Ibid., 112.
25. "The Intentional Fallacy" was first published in the Science Review 54
(1946): 468-88. It was revised and republished in The Verbal Icon (1956).
26. Wimsatt and Beardsley, "Intentional Fallacy," 18.
27 Wimsatt and Beardsley, "Affective Fallacy." 21.
28. Beardsley. Aesthetics, 457
29. Ibid. 458.
30. Wimsatt, "Genesis." 194.
31. Ibid., 210.
32. Ibid., 222.
33.
34.
35.
36.

Ibid., 224.
Parker, Flawed Texts, 214.
Ibid, 216-17
Ibid., 216.

37 Wellek, "Term and Concept of Literary Criticism." 30.


38. Ibid., 32.
39. Gabler, introduction to CGET, 3.
40. Martens, "What Is a Text?" CGET, 217.
41. For more detailed information see Gabler's introduction to CGET; Mar
tens's two essays within that collection; and Plachta, "German Literature."
42. Gabler, "Unsought Encounters," 156.
43. Martens, "(De)Constructing Texts by Editing," 125.
44. Zeller, "New Approach," 235.
45. For an excellent account of the relation of structuralism and semiotics to
theories of editing, see the chapter "Structure and Sign in the Text"in Greetham's
Theories of the Text, 276-325.
46. Martens and Zeller. eds.. Texte und Varianten. This work is not currently
available in translation, although three essays from the collection are translated
and republished in CGET.
47
48.
49.
50.
51.

Gabler, introduction to CGET, 2.


Zeller. "New Approach." 231-32.
This essay is not available in translation.
Scheibe. "Editorial Problem," 197
Ibid., 197

241

NOTES TO PAGES 20-24

52. Scheibe distinguisiies between "variants" and "readings." The term


"readings" allows for comparison of unauthorized versions of a text or a com
parison between an unauthorized and authorized version. The term "variant"
denotes two alternative, clearly authorized versions of a text. See also Martens,
who takes these ideas further in "What Is a Text?" 210; 226n3.
53. Scheibe, "Editorial Problem," 199.
54. The distinction between reception- and production-based approaches is
made by Scheibe in "Editorial Problem" (195) and developed at length by Hurle
busch in "Understanding the Author's Compositional Method." I engage with
the latter in some detail in chapter 4.
55. Scheibe, "Editorial Problem," 201. The same point about the problem of
selecting the "edited text" is made by Zeller in "Structure and Genesis": "In prin
ciple, one must edit versions, and must edit every version either in extenso or
by recording its variants in relation to another completely edited version of the
work. (As a rule, therefore, only one single witness document provides the edited
text)" (107).
56. Hans Walter Gabler is another exception to insular tendencies in Ger
man editing since he engages closely with both traditions (Anglo-American and
French).
57. Zeller, "New Approach," 240.
58. Ibid., 240-41.
59. Zeller, "Record and Interpretation," 25.
60. Ibid., 47
61. Ibid., 51.
62. Martens, "What Is a Text?" 211.
63. Ibid., 221.
64. Ibid., 222.
65. Gabler, introduction to CGET, 4.
66. Gabler, "Genetic Texts," 69.
67. For a helpful discussion of the term "synoptic," see Hopker-Herberg, "Re
flections on the Synoptic Mode of Presenting Variants," 79.
68. Gabler, with Steppe and Melchior, eds., Ulysses, 3:1894,1892.
69. Ibid., 3:1891.
70. Gaskell, From Writer to Reader, 235.
71. Mays in Coleridge, Poetical Works, cxxii.
72. Ibid., cxxxiii. Mays states: "Changes in rhythm caused by punctuation
are as significant as verbal changes, and changing the form of a word (underlin
ing it or putting it in capitals or quotation-ma/ ' modifies its meaning. Changes
of tense go in the text, and variant forms, like iKidst' for 'mid', 'ne' for 'nor', 'ee'
for 'eye', in the apparatus" (cxxxiii). In general, anything that affects meaning is
in the text on the page rather than the apparatus at the bottom.
73. Tanselle, "Historicism and Critical Editing," 39.

242

NOTES TO PAGES 24-29

74. Greetham, Theories of the Text, 312,313.


75. Ibid, 313.
76. Mays in Coleridge, Poetical Works, cxxiii.
77. Ibid., Ixxx.
78. McGann, "Ulysses as a Postmodern Text," 297.
79. Mays in Coleridge, Poetical Works, cxiiv, cli.
80. Ibid., cxlv.
81. Ibid., cxlvi.
82. Ibid., cxlix.
83. In "Genetic Editing, Past and Future," the French genetic critic Louis
Hay also notes the limited emergence of a fully critical German genetic criticism:
"This is partly for technical reasons: genesis presented in apparatus typography
does not always lend itself easily to analysis" (120).
84. Martens, "(De)Constructing Texts by Editing," 126.
85. Martens, "What Is a Text?" 225.
86. Eggert, "Shadow across the Text," 320.
87. Zeller, "Structure and Genesis," 97. Zeller notes in passing that German
editing "does not distinguish in so marked a way between accidentals and sub
stantives as does Anglo-American practice" ("New Approach," 235) and suggests
that this may be due to the nature of the German language itself (235n4).
88. Gabler, "Genetic Texts," 68.
89. Hurlebusch, "Understanding the Author's Compositional Method,"
75-76.
90. Ibid.,8L
91. These questions are posed by Gresillon, "Slow," 114; Hay, "History or Gen
esis?" 192; and Compagnon, "Introduction," 395, respectively.
92. At the same time, there are some historical links between French genetic
criticism and German theory. In "Genetic Editing," Hay notes the significance
of the Beif^ner edition and of German editing as one starting point for French
thinking (118-20). H. T. M. van Vliet also points out the author's German origins
in a review of Almuth Gresillon's book (translated into German) and the fact
that Hay was originally a Heine scholar. See Van Vliet, "Gresillon, Literarische
Handschriften," 339.
93. Compagnon, "Introduction," 397.
94. In "Critiques de la Critique Genetique," Hay states that in the early 1970s,
for the first time, the number of modern manuscripts in the national library's
collection exceeded the number of medieval manuscripts (408).
95. For detailed information on the French context for genetic criticism, see
Pierssens, "French Genetic Studies at a Crossroads"; Compagnon, "Introduc
tion"; Lernout, "Genetic Criticism and Philology"; Hay, "Critiques de la Critique
Genetique."
96. Pierssens, "French Genetic Studies at a Crossroads," 620,

243

NOTES TO PAGES 29-39

NOTES TO PAGES 39-45

97. Hay, "Critique de la Critique Genetique," 407.


98. Hay, "Does 'Text' Exist?" 68.
99. Jenny, "Genetic Criticism and Its Myths," 10.
100. Biasi, "Toward a Science of Literature," 41.
101. Hay, "Does 'Text' Exist?" 73,75.
102. Deppman, Ferrer, and Groden, introduction to Genetic Criticism, 2.
103. Ibid., 2.
104. Hay, "Does 'Text' Exist?" 68.
105. Biasi, "Toward a Science of Literature," 42,43.
106. "Archive et Brouillon. Table Ronde," 192.
107. Ibid.
108. Ibid.
109. Ibid., 207.
110. Gresillon, "Slow," 115.
111. Jenny, "Genetic Criticism and Its Myths," 23.
112. Biasi, "What Is a Literary Draft?" 54.
113. Gresillon, Elements de Critique Genitique, 147-75.
114. Deppman, Ferrer, and Groden, introduction to Genetic Criticism, 10.
This is not to say that Anglo-American critics are not working in a range of ways
with genetic material for a range of authors but that such work occurs in isola
tion and without any sense of it being a distinct field of study. How many AngloAmerican literary critics even know what is meant by the term genetic criticism!
115. GrtsiWon, EUments de Critique Genetique, 161.
116. Pietri and Valery, eds., "Preface," in Paul Valery: Cahiers, 1894-1914, 22.
117 Ibid., 26.
118. Ibid., 24.
119. Biasi in Flaubert, Carnets de Travail, 102.
120. Biasi, Genetique des Textes, 65.
121. Ibid., 66.
122. Ferrer, "'Clementis's Cap,'" 224.
123. Ibid., 227.
124. Ibid., 223, 227
125. Ibid., 230.
126. Ferrer and Groden, "Post-Genetic Joyce," 503.
127. Biasi, "Toward a Science of Literature," 38.
128. Gresillon, EUments de Critique Genetique, 204.

4. The final chapter will return to some of the issues raised here to consider
them more deeply by means of a philosophical enquiry into the nature of creative
process.
5. Derrida, Of Grammatology, 167.
6. Ibid., 74.
7. Eggert, "Work Unravelled," 53.
8. It is important to bear in mind, however, that Derrida is not denying
the existence of authorial intention as one meaningful context; he simply (and
rightly) denies it any absolute authority: "[T]he category of intention will not
disappear; it will have its place, but from this place it will no longer be able to
govern the entire scene and the entire system of utterances" ("Signature Event
Context," 326).
9. Barthes, "Death of the Author," 142.
10. Ibid., 146.
11. Ibid., 145. Here, Barthes partly seems to anticipate my application of
speech acts to the text as process. However, his account elides the materiality of
writing and thus the possibility of multiple returns to grounded acts.
12. Barthes, Pleasure of the Text, 16.
13. Ibid., 27
14. Barthes, "Theory of the Text," 38.
15. Barthes, "Death of the Author," 145.
16. A full philosophical account of phenomenological temporality, agency,
language, and meaning in relation to understanding process is given in the final
chapter of this book.
17. Barthes, "From Work to Text," 160.
18. Barthes, "Theory of the Text," 32.
19. Foucault, "What Is an Author?" 207-8.
20. Barthes, Pleasure of the Text, 4,5.
2L Ibid., 14.
22. Ibid, 14, 2L
23. Ibid., 24. For a fuller exploration of such contradictions in relation to
Freudian disavowal, see Bushell, "Textual Process."
24. Barthes, Pleasure of the Text, 19,20.
25. Foucault, "What Is an Author?" 221.
26. Ibid.
in. See also "The Order of Discourse," in which Foucault states, "It would of
course, be absurd to deny the existence of the individual who writes and invents"
(59), but he views this individual as one who "takes upon himself the function
of the author" (59). He thus distinguishes between "the sense of the speaking
individual who pronounced or wrote a text" (58) and "the sense of a principle of
grouping of discourses" (58).
28. Barthes, Pleasure of the Text, 25,26.

2. THEORIZING PROCESS
1. Gresillon, "Slow," 123.
2. Ferrer and Groden, "Post-Genetic Joyce," 509.
3. Ibid.

244

245

NOTES TO PAGES 45-53

29. Barthes, "Theory of the Text," 36.


30. Barthes, Pleasure of the Text, 47.
31. Ibid. See also note 23.
32. McGann, "Case of the Ambassadors," 163,
33. Wimsatt and Beardsley, "Intentional Fallacy," 3.
34. In "Genesis," Wimsatt states: "The design or intention of the author is
neither available nor desirable as a standard forjudging either the meaning or the
value of a work of literary art" (222).
35. Ibid., 221-22.
36. Gadamer, Truth and Method, 166.
37 Ibid., 167
38. Ibid.
39. Ibid., 97
40. Ibid., 117
41. Ibid., 120.
42. Heidegger, PLT, 190.
43. Ibid., 197
44. Ibid., 193.
45. Ibid., 215.
46. Derrida, Of Grammatology, 166.
47 Ibid., 160.
48. Royle, Jacques Derrida, 56-57
49. Such ideas are addressed fully in the final chapter of this book.
50. Useful overviews of intention from a literary, editorial, or philosophical
standpoint, are given by Patterson, "Intention" Tanselle, "Editorial Problem of
Final Authorial Intention," 173-74; Greetham, "Intention in the Text," in Theo
ries of the Text, 156-205; McLaverty, "Concept of Authorial Intention in Textual
Criticism": and Siewart, "Consciousness and Intentionality."
51. Anscombe, Intention, 9. See also Meiland's distinction between purpo
sive and nonpurposive intention in The Nature of Intention, 7-11; Kemp's distinc
tion between ulterior intention and immediate intention in "The Work of Art
and the Artist's Intentions"; and Searle's distinction between prior intention and
intentions in action in Intentionality, 84-98.
52. In this way Husserlian intentionality anticipates speech act theory. How
ever, although Husserl's account of language is referential to consciousness, he
does also allow for an ideal "species" meaning. See Smith, "Towards a History of
Speech Act Theory."
53. Austin, How to Do Things with Words, 114.
54. Searle, Speech Acts, 36-37.
55. Ibid, 17
56. In Intentionality Searle has one reference to Brentano on p. 14 and one to

246

NOTES TO PAGES 53-61

Husserl on p. 65. He claims to be giving a logical rather than an ontological ac


count, but that account clearly corresponds to Husserl's in many respects, while
going much further in terms of its causal conclusions.
57. Searle, Intentionality, 27.
58. Ibid, 3.
59. Hershel Parker adopts Searle's two kinds of intention in Flawed Texts,
allowing prior intention to apply to action "long prior to or momentarily prior
to the act of writing" (23) and intentions in action to "the actual composing pro
cess" (23).
60. Searle, Intentionality, 84.
61. Ibid., 27
62. Ibid., 82.
63. Shillingsburg, Scholarly Editing in the ComputerAge, 33.
64. Ibid., 37,
65. Ibid., 36.
66. Ibid., 37
67 Ibid,, 38,
68. Searle, Intentionality, 27.
69. Bennington, Interrupting Derrida, 10,
3, RECLAIMING PROCESS
1. The compositional context corresponds to Heidegger's "equipment to
tality" in the ready-to-hand environment. Searle's "Background" for language
would form part of such a context.
2. In a Heideggerian sense, the third context must always preexist the other
two, and the being who creates. My account, however, while drawing on Heideg
ger, clearly moves away from him, particularly in asserting a movement into and
out of "authenticity" at the heart of the creative act.
3. Heidegger, Be^T, 97.
4. This would correspond to Heidegger's description of malfunction in the
equipment totality and the point at which the "ready-to-hand" becomes conspic
uously "unreadiness-to-hand" (B&T, 103). It may seem odd to apply such an idea
to human intervention, but when others are viewed insofar as they contribute to
the creative process, then they do seem to function in this way.
5. Keats, Letters, 1:212.
6. Stillinger in Keats. Poems of John Keats, 588.
7. Keats, Letters, 1:214.
8. For a discussion (and rejection) of the idea of intentionless meaning, see
Knapp and Michaels, "Against Theory." They argue that intentionless meaning
is "radically counterintuitive" (727) because meaning and intention are insepa-

247

NOTES TO PAGES 61-65

NOTES TO PAGES 65-71

rable. However, I think their account of intentionless meaning on the beach is


misleading for the reason given here: it does not allow that unintended meaning
can only occur where there is already intention.
9. Open intention corresponds in editorial terms to "textual instability" as
described in Hans Waher Gabler's article "The Text as Process." Gabler states:
"[T]he text of a work under the author's hand is in principle unstable. Instability
is an essential feature of the text in progress" (111).
10. See Bushell, Intention Revisited," for a more detailed engagement with
Hancher's paper and for the emergence of my terminology out of his. Hancher
distinguishes between programmatic intentions, active intention, and final inten
tion. Hershel Parker also engages with Hancher but criticizes him for not allow
ing for a fully active intention within composition: "Hancher can accommodate
the period before the composition, the moment of completion, and the indefinite
period afterwards; but duringthe ongoing creative process itselfhas no place
in his theory" (Flawed Texts, 22).

13. Throughout this section it is to be remembered that the intentional com


plex represents a return to what is already written by the writer and, in a sense, a
reclamation of nonintentional being by acts of will.
14. Searle, Intentionality, 107.
15. Wordsworth, Letters: Later Years, 3:583nl. One factor in relation to incom
plete longer works concerns the poet's ability to move from the kind of method
required for shorter rather than longer composition. On this, see Thomas Gray's
comment that "he [Gray] had been used to write only Lyric poetry in which the
poems being short, he had accustomed himself, & was able to polish every part;
that this having become habit, he could not write otherwise; & that the labour of
this method in a long poem would be intolerable" (Correspondence of Thomas
Gray, appendix Z, 3:1291). Keats's difficuUies over "Hyperion" probably fall into
the same category.
16. Wendorf, "Alexander Pope, An Essay on Man," 47.
17. Keats s Hyperion" and "The Fall of Hyperion" again come to mind.
18. The existence of revised intention is debatable, and bound up with edito
rial debates over the distinction between a "variant" and a "version" of a text.
19. Gaskell, From Writer to Reader, 75.
20. Ibid., 75, 76.

21. Gresillon, "Slow," 117.


22. James McLaverty's work makes good use of the temporal dimension in or
der to retain an intentionist element while releasing a text (or editorialpresentation
of it) from absolute linear organization. See "Issues of Identity and Utterance."
23. Again, see McLaverty, who uses the philosophical example of "Theseus's
Ship" (if all parts of the ship are reconstructed gradually over time, is it still the
same ship at the end?) to reconsider the structural conception of the work in
terms of "overlapping versions and of coexisting rival claimants to be the work"
("Issues of Identity and Utterance," 141).
24. Wordsworth's three principal versions of The Prelude would correspond
to such a model. See also my discussion of contingent intention in relation to
Wordsworth in chapter 4.
25. Programmatic intention anticipates contingent intention but also clearly
overlaps with it and continues to be present "behind" it. Contingent intention is
in a sense also "unfulfilled" at the time of writing and may only be definable as
^contingent" retrospectively.
26. Hancher, "Three Kinds of Intention," 831nl0.
27. See Shillingsburg, Scholarly Editing in the Computer Age, 35-36.
28. Anscombe, Intention, 84.
29. Meiland, Nature of Intention, 70.
30. Ibid., 81.
31. Searle, Intentionality, 101.
32. Ibid., 102.
33. Meiland, Nature of Intention, 87.
34. Ibid., 88.
35. Vendler, "Reading Keats in Manuscript," 40.
36. See Freud's Psychopathology of Everyday Life, in which chapter 6 is enti
tled "Misreadings and Slips of the Pen."
37. Philip Gaskell makes a relevant observation concerning Charles Dick
ens's practice of giving a single manuscript draft over to the compositors:
"[I]ts sheer illegibility did sometimes lead to verbal errors. Indeed it is curious
that Dickens, who must have been aware of some of these mistakes and of the rea
son for them, did not think it worth while to increase the legibility of his manu
scripts by using a little more paper" (From Writer to Reader, 143-44).
38. Searle, Intentionality, 102.
39. See Bushell, "Meaning in Dickinson's Manuscripts."
40. Derrida, afterword to Limited Inc., 128.
41. Ibid., 129.
42. Shillingsburg, "Autonomous Author," 22.
43. For a full account of each of Biasi's four phases (in terms of processes,
function of researcher, and documents) see the second chapter of Genetique des
Textes. His account does influence my own given here.

248

249

11. In The Study of Modern Manuscripts, Reiman bases his core distinction
between "public," "confidential," and "private" materials of process on the "social
intentions of the writer" (40), which corresponds to an anticipation of audience.
Reiman defines a literary work (such as a poem) as "public," even if it is never
published, on the basis of such social intention: "Authors who intend that their
work be published write with that end in view" (43).
12. My thanks to Michael Sanders for suggesting the use of the term contin
gent intention.

NOTES TO PAGES 72-79

44. Biasi, "What Is a Literary Draft?" 32.


45. Ibid., 33.
46. Biasi, Genetique des Textes, 87.
47. My thanks to James McLaverty for suggestions and comments that di
rectly inform the typology at certain points.
4. WORDSWORTH'S PROCESS
The section entitled "The Micro-'Prelude': First Draft Activity" has been previ
ously published as a paper entitled "Wordsworthian Composition: The MicroPrelude" in Studies in Romanticism 44 (Fall 2005): 399-421. In that paper, a
second section explores "first fair copy" activity in DC MS 21 (Prelude MS RV).
There has not been space to retain it here.
1. My thanks to James Butler for reminding me of this important point.
Something similar is suggested by the French critic Laurent Jenny's comment
that "paradoxically, the establishment of the pre-text tends to dissolve the textual
entity that was precisely the one that gave it this value" ("Genetic Criticism and
Its Myths," 23).
2. Parrish, "Worst of Wordsworth," 91.
3. Biasi, Ginetique des Textes, 79.
4. Ibid., 69.
5. Stilhnger, "Textual Primitivism," 14. By "later Wordsworth," Stillinger
means the later editions and revisions of earlier works by the older poet, rather
than later works.
6. Ibid.
7. Ibid., 20.
8. Ibid., 27. A position similar to that of Stillinger is articulated by Zachary
Leader in the chapter "Wordsworth, Revision and Personal Identity" in Revision
and Romantic Authorship, but instead of concluding in favor of textual pluralism.
Leader concludes with a return to respect for final authorial intention.
9. Reiman, Romantic Texts and Contexts, 135. Actually, Reiman is in error
here in criticizing the Cornell editors for this reproduction of the early Prelude
text. The "Two-Part Prelude" was first published in the third edition of the Nor
ton Anthology in 1974 and then in the Norton Critical Edition of The Prelude,
1799,1805,1850 in 1979. In each case, the text was the one prepared by Jonathan
Wordsworth and Stephen Gill, not that of Parrish for the Cornell Series.
10. Jenny, "Genetic Criticism and Its Myths," 15.
11. Other manuscripts are more ambiguous. Manuscript M looks superfi
cially like a point of contingent completion for a five-book Prelude but is actually
copied out after the decision to write a distinct five-book poem has been aban
doned. It represents a contingent state within the longer process but not a discrete
"state" for the poem. Manuscript C is another arguable case, an interim manu

250

NOTES TO PAGES 79-84

script copied in 1819-20 and used for revision in 1831-32. Both MS M and MS C
were, initially at least, copied for someone else rather than forming a core part of
the poem's development (ahhough MS C does come to do this).
12. J. Wordsworth, Music of Humanity, 31.
13. Jack Stilhnger raises this issue in "Textual Primitivism and the Editing
of Wordsworth" when he comments on the "elusiveness of the 'earliest complete
state' of a work" (14). Stephen Gill has also raised concerns about how far one
can go in giving independent status to texts that were not initially intended as
texts. According to Gill, "It is clearly legitimate to rescue from oblivion poems
Wordsworth excised from his canon or did not publish at all" ("Wordsworth's
Poems," 54).
14. Carney, "Making the Most of a Mess," 634.
15. Wordsworth, Letters: Early Years, 470.
16. Wordsworth, The Excursion, 35.
17. Wordsworth, Letters: Early Years, 586-87.
18. Darlington's introduction to Wordsworth, Home at Grasmere, 14.
19. Wordsworth, Letters: Early Years, 594.
20. Coleridge, Letters, 1:538.
21. Gill, Wordsworth: A Life, 171.
22. For more detail on Wordsworth's abandonment of the Recluse project
as recorded in the letters, see Darlington's introduction to Home at Grasmere,
26-32.
23. Scheibe, "Editorial Problem," 195; Kets-Vree, "Dutch Scholarly Editing,"
143. Kets-Vree states, "Manuscripts of the Mozart type contain few or no variants
because the genesis did not take place on paper, but in the author's mind. The
author who works according to the Beethoven method, by contrast, starts with a
rough version and then proceeds to delete and refine" (143).
24. Hurlebusch, "Understanding the Author's Compositional Method,"
85-86. It is worth noting, however, that the French and German distinctions
don't directly map onto each other (e.g., "constructive" seems to correspond to
"programmatic," yet it is the French "process-centered" writing that builds on
itself.)
25. Ibid., 85-98.
26. Gresillon, Sliments, 104-5.
*27. Ferrer, "Clementis's Cap," 225.
28. Biasi, Genetique des Textes, 33.
29. Parrish in Wordsworth, The Prelude, 1798-1799, 26.
30. Ibid., 21.
31. Hurlebusch, "Understanding the Author's Compositional Method," 96.
32. Ibid., 92.
33. Gill, Wordsworth: A Life, 160. See also Andrew Bennett's discussion of
Wordsworth's resistance to writing in his recent book Wordsworth Writing, and a

251

NOTES TO PAGES 91-100


NOTES TO PAGES 84-90
comparable piece by Mark L. Waldo, "Why Coleridge Med to
book came out after this study had gone to press, so I have not been

51. Searle, Intentionality, 84.


Y

incorporate a response to it as I would have liked.


34. Wordsworth, Letters: Early Years, 236.
^
35. Eggert, "Textual Product or Textual Process, 62-63.
36 RuofF, Wordsworth and Coleridge, 16.
fy. ?;:e distinction Is first made by Raymonde D'
tlque et Poitique," in Essah d, CritiqM Gtel.jue, ed. Louis Hay.
developed by Biasi in "What Is a Literary Draft? 42 49.

52.
53.
54.
55.

Ibid., 27.
Ibid., 84.
Ibid., 107.
Ibid., 106.

fl. SSrierdltoSon between "prior intention" and "intention-in;

39 wT^olm.'inWorlwlrth'sProse Worte gives theorderof^^^

edgew^Sff'ratherthanjust"toremove'edge.-"Searlesta^^^
tional component can be as complex as you like"
59. Of course, the text will also contain a meaning eyo

^^e poet's

conscious understanding.

M to

and 23?eb 1799-

!e'of

tdeologlcal model dearly shapes our re^

the
material as lplrobabVo.tlre!ywrte between 6^
1799- (30). Since the "Esssy" is also Inverted and entered a random number of p g
;tiismore,lk=lythatitas.,.r^
40. Parrish, introduction to Prelude, 1798-17^% i, <^ui.
^ 41. Wu in Wordsworth, Five-Book Prelude, 8.
42. Wordsworth, Prose Worfcs, 1:103.
43. Wordsworth, Prelude. 1798-99, 46, Reading Text, lines 137 38.
44. Wordsworth, Prose Wor;:s, 1:103.
4^3-35
45. Wordsworth, Prelude, 1798-99,53, Reading Text, lines 429-31,433 ^
46. In "The Five-Book PreiuJe," Jonathan Wordsworth first suggests that
Prelude material (for MS WW) was written in pencil because it was wri
""47'TS\ooserhorizontalhandislesstrueofcopiedverticalinfcpassages,h
ever Solm^bethatWordsworth initially turns thenotebooksidewaysmorde
to try andrelease himself creativelybut then simply decides itiseasiertowrue
the lines in this way (although his use of it for copying remains inconsistent) _
48. DC MS 19 (MS JJ), B'. All transcriptions essentially follow those of th
Cornell editions.

OTL'i^teS="'*sth.tengageintelllgentlywlth^

as part of a teleological structure.


fs The femhlr

th^^^^^

dry" is crossed through and yet retained is confus

ing. One p^X is that WordLorth crossed ^

Cornell transcription page.

. .r i:poio7-i2
65. Wordsworth, Prelude, 1798-99, Reading Text, 45, hnes
12.
66. J. Wordsworth, "Revision as Making," 36-37.
68 B^h Gresillon and Biasi give detailed accounts of the nature of revi
sion within process. These will be discussed in more detail in relation to mi y

HSSSSSS
Bennett Wordsworth Writing, 3. Clearly, there is a degree of overlap

positional pracuceb.

^ i

m"* mv work is more

pre tLS'Zaterlal and issues of revision include: Brinkley


Hanley. e^
pre-textuai mare
Revision"; Brinkley, The
and

-Tttcident in the Simplon Pass" Wolfson

-The Illusion of Mastery"; and Schell, "Wordsworth's Revisions of the Mcent ot


Q wdnn " For full-lensth studies, see Bennett, Wordsworth Writing; Manning,
ZdingRomantics; Ruoff, Wordsworth and Coleridge; Magnuson, Colertdge an

Merleau-Ponty, "On the Phenomenology of Language," 90.


71. Merleau-Ponty, Phenomenology of Perception, 207.
72. Derrida, Of Grammatology, 12.
73. Ibid., 8.

Wordsworth; and Sheats, The Making of Wordsworth s Poetry.

253
252

NOTES TO PAGES 101-110

74.
75.
76.
77.

Wordsworth, Prose Works, 2:84.


Jacobus, Romanticism, Writing, and Sexual Difference, 169.
Ibid.
Ibid., 173.

78. See also Bennett, "Appendix on Poetic Dictation," in Wordsworth Writ


ing, 175-77.
79. Bennett, "Wordsworth Writing," 6.
80. Wordsworth, Letters: Later Years, 2:454.
81. See Bushell, Re-reading 'The Excursion,' 169-78.
82. There are also no uses of the word "writing" for The Prelude, though
there are two for The Excursion. There are nine uses of the word "written" in The
Prelude, and two in The Excursion. See Cooper, Concordance.
83. Wordsworth, Thirteen-Book Prelude, 1:310, AB-Stage Reading Text, book
12, 231-34, 237-39.
84. There are twenty-one uses of "told" and fifteen of "tell" in The Prelude;
sixteen of "told" and sixteen of "tell" in The Excursion.
85. Wordsworth, The Excursion, 299, book 9, 795.
86. Wordsworth, Thirteen-Book Prelude, 1:152, AB-Stage Reading Text, book
4,49-54.
87. Ibid., 1:163, book 5, 44-46.
88. Ibid., 1:166, book 5,164.
89. Ibid., 1:166-67, book 5, 185-87; 1:311, AB-Stage Reading Text, book 12,
270-72.
90.
91.
92.
93.

Bennett, Wordsworth's Writing, 4.


Reed, introduction to Wordsworth, Thirteen-Book Prelude, 1:19.
Ibid., 1:19.
MS WW (DC MS 43), 18'-.

94. Wordsworth, Thirteen-Book Prelude, 1:173, AB-Stage Reading Text,


482-91.
95. MS WW (DC MS 43), 18\
96. According to Bryant, "[T]he fact of the shifting of the words (not just the
shifted words themselves) has meaning" (Fluid Text, 97).
97. DC MS 70,7'.
98. DC MS 43 (MS WW), 28''. See also Keith Hanley's reading of this draft in
"Crossings Out," 120.
99. Wordsworth, Thirteen-Book Prelude, 1:189-90, AB-Stage Reading Text,
book 6,525-30.
100. See Bushell, Re-reading 'The Excursion,'169-78.
101. Wordsworth, Excursion Reading Text 229, book 6,1307-8;230, book 7,1-4.
102. Wordsworth, Excursion Reading Text 275, book 8, 608-11; 276, book

254

NOTES TO PAGES 111-119

103. DC MS 73. 6^.


104. Wordsworth, Excursion Reading Text 76, book 1, 985; 103, book 2,930.
105. Ibid., 296, book 9,753.
106. Ibid., 296-97, book 9, 783-95.
107. WiUiam Galperin also discusses the multiple endings of the final text in
"'Imperfect While Unshared,'" 212.
108. DC MS 73,46^; DC MS 73,18^
109. Wolfson, Questioning Presence, 123.
110. The line "an active principle" is from Excursion Reading Text 276, book
9, 3.
111. The use of earlier material as a stimulus certainly looks to be true for
book 2; see DC MS 47, 33'. The only surviving version of the book 4 opening
in DC MS 74 (260 is probably copied from elsewhere, however, so the initial
context of integration is not known. For discussion of these passages, see also
the Cornell Series Excursion "Manuscript History: First Stage" (429); "Seventh
Stage" (464).
5. TENNYSON'S PROCESS
1. Knight, "Reminiscence of Tennyson," 267.
2. Quoted in H. Tennyson, Memoir, 1:198. In the light of nineteenth-century
publishing habits this is not quite so destructive as it sounds, since the publishing
house required a work in manuscript to be taken apart and printed in separate
blocks of loose pages. (The high cost of handmade type in the first half of the
century limited the amount of a work that could be printed at a time.) See Dooley,
Author and Printer in Victorian England, 42-44.
3. Again, see Dooley, who writes of "the indifferent attitude many Victorian
authors took toward their manuscripts once a work was printed" (Author and
Printer in Victorian England, 127).
4. Quoted in H. Tennyson, Memoir, 1:198.
5. The only explicit written statement by the poet himself exists as a note in
the In Memoriam manuscript (Trinity Notebook 13) and concerns Spedding's
notes within it: "[T]he private notes are not to be shown by her [Lady Simeon]
to anyonenor is anything to be copied" (not reproduced in The Tennyson
Archive).
6. Christopher Ricks's The Poems of Tennyson in Three Volumes provides the
clearest information about compositional order and dates for the manuscripts in
relation to individual poems rather than particular notebooks. Far less useful is
John Pfordresher's A Variorum Edition of Tennyson's "Idylls of the King."
7. Omissions are usually listed at the front of an individual volume but not
referenced at the points where they occur within it. A notable example is the

255

Anda mungkin juga menyukai