University of Virginia Library

Mind and Textual Matter

by
RICHARD BUCCI

"The things which the textual critic has to talk about
are not things which present themselves clearly and sharply
to the mind.… Mistakes are therefore made which could
not be made if the matter under discussion were any corpo-
real object, having qualities perceptible to the senses." This remark, made
nearly ninety years ago by A. E. Housman in his well-known address
"The Application of Thought to Textual Criticism," suggests the crux of
many recent editorial discussions, in which some of editing's most basic
humanistic assumptions have been challenged with arguments influenced
by the movement sometimes called "postmodern" literary theory.[1] Many
of the challengers are themselves editors, and were motivated at least in
part by a sense that textual criticism was both technically overdeveloped
as a field and falsely estranged from literary criticism. An expressed inter-
est in drawing textual and literary criticism nearer to one another (as if
they were not already interpenetrated dimensions of the same discipline)
was thus a prominent feature of many of the discussions. A second inter-
est, also of an integrating character, was in surmounting the perceived
national or linguistic isolation of Anglo-American editorial scholarship
through an engagement with editorial traditions of other countries, espe-
cially Germany and France. Movements to open intellectual horizons in
this age of overly determined specialization are to be welcomed, and this
one has had its benefits, as readers of Scholarly Editing: A Guide to Research
can attest.[2] Many of the two dozen scholars whom David Greetham as-
sembled for this unusual project exhibited a felt sense of responsibility in
their contributions, which taken together provide Anglophone students
with a useful history of textual criticism across several periods of time
and many languages.

Other expressions of dissatisfaction with a perceived narrowness in
the editorial status quo have been less progressive, however, including

many of the attempts to align textual criticism with contemporary trends
in literary theory. It would not be unfair to say that interest in widen-
ing the horizons of editing has not always and everywhere been sincere,
some of it seeming to mask an anxiety that textual criticism, with its tra-
ditional historical and biographical focus, and rigorous bibliographical
preoccupations, was simply missing out on the wave of postmodernism
that had swept through the English departments of American universities
in the 1980s. While the swell of enthusiasm that marked the appearance
of this wave has largely receded (along with an improbably abstruse and
jargon-filled form of discourse), some of its dehumanizing ideas are still
very much with us. The anti-objectivist, metaphysical conceits of the de-
constructionist approach to texts, for example, found their way across
the imaginary line separating literary from textual criticism, and were
received with favor by some editors of English-language Renaissance and
modern works. The evidence can be seen in the shift of editorial focus
away from the author, even to the point at which the author is completely
abnegated as an object of editorial interest. In recent editorial literature
it is not impossible to find bizarre statements in which the human origins
of texts are seemingly denied, and authors are redefined as abstract con-
cepts conjured, depending on the argument, in the Romantic or Modern
period. Objective reality is consequently denied also to aspects of the past
having to do with authors' lives, and especially to the thoughts of authors
in relation to particular works. These thoughts were an object of editorial
discovery long before the moment in 1939 when R. B. McKerrow gave
prominence to the idea of authors' intentions, and have for just as long a
time served as a focal point for the critical reconstruction of works which
imperfectly survive in documentary witnesses.[3] Editors engaged in such
reconstructions have accordingly bound themselves to approach variant
readings or suspect passages with the author's preferences, rather than the
preferences of others (including their own), in mind.

Where authors themselves are liable to being reduced to posthumous
concepts, author-centered critical texts not surprisingly have been denied
scholarly legitimacy, often with the charge that these texts are "Platonic"
ideals. As Stephen Orgel, one of the most influential opponents of the
traditional bibliographical and editorial focus on the author, remarked in
1996, "the basic assumption of most editorial practice is that behind the
obscure and imperfect text is a clear and perfect one, and it is the editor's
job not to be true to the text's obscurity and imperfection, but instead

to produce some notional platonic ideal."[4] Professor Orgel's views have
some points of coincidence with those of Jerome McGann, who, while not
as regardless as Orgel of the author's importance, is another of the many
voices suggesting that traditional scholarly editorial practices have as their
design the recovery of a perfect text—a pursuit which once posited is
derided as futile. This circular criticism has been heard often enough in
recent years that it has begun to influence the self-consciousness of some
editors whose editorial approach remains author-centric.[5] While mo-
mentarily pervasive due partly to residual postmodern trends in literary
theory, the image of the authorially intended critical text as an ideal text
in the Platonic sense is the outcome of a basic misunderstanding. A clue
to the source of the confusion can be found in the manner in which the
Platonic label has been attached to traditional editorial activity—some
philosophical language is used in the criticisms, but no philosophical argu-
ment is seriously developed. What in Platonic terms is meant by the label
itself is not easy to see: the critics never discuss whether they believe that
author-centered editors are pursuing ideal forms (the apparently separate
Platonic world of objective essences, discussed in many dialogues) corre-
sponding to the sensational forms of the literary works they are editing,
or a universal aesthetic form (αὑτὀ ὁ ἕστι καλόν), such as that presented in
the Symposium, which the author had also presumably pursued. Where the
first pursuit is postulated and criticized, the position of the critics would
seem to admit exclusively of "sensational" (or material) forms for literary
works—an interesting point needing, however, a good deal more elabo-
ration than is ever attempted. As for the second pursuit, it would seem,
since Plato ascribed priority to universal forms, that the critics were main-
taining that traditional editors believed that these forms could be recol-
lected, already perfect, though here too necessary elaboration is wanting.
Most obviously missing from the arguments, however, is some treatment
of Plato's view of art, wherein universals were denied to artistic creations
(on the grounds that most art was mere imitation of natural things), and
their authors refused admission to the ideal republic. This preliminary
examination of the criticisms would lead to a conclusion that Platonic
forms are not being referred to, and the name and thought of Plato have
little significance to the critics beyond a vaguely understood connection to
idealism in general. For while the critics do not agree among themselves
on the editorial alternatives, they do seem mainly united in a view that
traditional author-focused critical editing involves a belief in the existence
of a pure, ideal text, and a willingness to strive after it. The source of the
confusion is the same as that suggested by Housman, when he ascribed a
certain class of editorial blunders and misconceptions to the fact that the
objects of editorial attention are not of a "corporeal" nature.

I

The critical editorial approach to Anglophone literary works which
arose alongside the "New Bibliography" had detractors before the ar-
rival of the postmodern wave, the adherents of which, furthermore,
were not even the first to describe it as a Platonic pursuit. As far back
as 1959, Shakespeare scholar Leo Kirschbaum criticized the editorial vi-
sion of Fredson Bowers for, among other things, its supposed "Platonic
idea of an Elizabethan dramatic text—an absolute, unchangeable, holy
thing.…"[6] Much of the editorial discussion that shaped Bowers's ap-
proach was centered on works of the English Renaissance, and so past
and present opponents of critical editing have sometimes attempted to

support their criticisms by referring to the declining fortunes of certain
hypotheses of tradition advanced by McKerrow and his colleagues and
successors, especially those involving the origins of Shakespeare's plays.
Paul Werstine, an active latter-day critic of the concept of authorial in-
tention, has in several essays sought to discredit W. W. Greg's editorial
recommendations by undermining his investigations into the problems
of Shakespearean transmission. Dissecting Greg's discussions about the
manuscript origins of a few of the early printed texts, Werstine attempts
to show in "Post-Theory Problems in Shakespeare Editing" that they con-
stitute, in Werstine's oft-repeated Lyotardian phrase, a "grand narrative,"
whose supposed purpose is "the conservation of the Romantic view that
the greatest art we have was produced by the individual genius working
in splendid isolation from the rest of his culture."[7] For the tolerant reader
there are some limited lessons to be learned in Werstine's overall skepti-
cism of the discoveries of the New Bibliography, and even in his criti-
cisms of Greg's individual hypotheses about whether certain printed texts
were based on actors' memorial reconstructions or Shakespeare's "foul
papers." What good there is, however, must be carefully strained from
an environment of distorted representations of Greg's motives and even
aspersions on his character (as well as sprinklings of some rapidly aging
postmodern jargon). For readers unfamiliar with the original discussions,
the danger is high that Werstine's misrepresentations will be absorbed
along with his sometimes insightful analyses and reassessments of the evi-
dence. Particularly damaging in this respect is Werstine's unprincipled
amalgam of some speculations of Greg's contemporary J. Dover Wilson
with Greg's well-known contribution to editorial practice. In a later es-
say, Werstine goes so far as to say that "Greg's theory of copy-text rests
to a large extent on his confidence in Wilson's work on Hamlet."[8] In fact,
Wilson's work on Hamlet is not mentioned in "The Rationale of Copy-
Text," Greg's influential statement of editorial principles; nor even do the
plays of Shakespeare figure largely among the examples he uses in that
essay to demonstrate his main insight—that a revised edition of a work,
containing the author's last preferred wording, may not be as faithful to
his or her preferred formal details ("accidentals") as the earlier edition,
if the later typesetting was based on it. Readers of Greg's The Editorial
Problem in Shakespeare, furthermore, will be surprised to find Werstine,
working from a few fleeting and circumspect remarks he has found there,
fashioning a wholesale endorsement of Dover Wilson's views on Hamlet [9]
The masking of Greg's reservations about Wilson's views may mislead
new scholars, who will have to seek elsewhere, at any rate, for an appre-
ciation of the many important and enduring advances in the investigation
of play origins owed to Greg's scholarship.

Readers of Stephen Orgel's The Authentic Shakespeare, a collection of
essays warmly received by Professor Werstine, will find themselves con-
fronted with an even less accurate picture of the editorial views and pur-
poses of Greg and the tradition he represents (though Orgel criticizes
Greg's ideas without using his name). At the center of this distorted rep-
resentation, contained mainly in two brief opening essays, "What Is a
Text?" and "What Is an Editor?", is the familiar assertion that critical ed-
iting in the main involves the pursuit of a Platonic ideal text.[10] The asser-
tion is meant to serve as a foil for Orgel's often convincing demonstrations
of the instability of texts, which would fail to achieve the insightful effect
Orgel seeks for them if they were cast against a more balanced assessment
of Greg's views. Orgel also discounts—or rather, attempts to abnegate—
the traditional editorial connotation of the word authority, meaning writ-
ing which came from the author. His objections to this connotation are
representative of a range of similarly strong reactions recorded by other
scholars seeking to diminish the importance of the author in the literary
process. In their discussions the critics typically ignore the restricted but
indispensable traditional editorial meaning, and Orgel is no exception.
To those familiar and comfortable with that meaning, the objections to
it seem, again, to be the result of a misunderstanding—whether willful
or not—rather than of a genuine disagreement. Orgel wants the word to
mean social power or control, to the exclusion of all other meanings. He

accordingly presumes that only those with legitimate or financial power
can have authority, and concludes that authors are without authority
since in his view they are often paid servants mechanically satisfying the
demands of their powerful employers. In the English Renaissance, Orgel
explains, the "authority of the published text was … that of the pub-
lisher: he owned it; the author's rights in the work ended with his sale of
the manuscript. The publisher was fully entitled to alter the manuscript if
he saw fit—the manuscript was his."[11] Here it is easy to see that Orgel's
use of the term authority is simply incongruous with how the term has
been used in scholarly editing. That a misconception is operating is made
plainer by Orgel's attempt to clarify his views with an analogy about
painting. "Once the painting has been sold," Orgel explains, referring
to painting in general, "we do not believe that the artist has any further
rights in it." The owner, according to Orgel, has the authority to mutilate
or even destroy it, as was the fate meted out by Clementine Churchill
to Graham Sutherland's portrait of her husband Winston: "to believ-
ers in the autonomy of the artist, the painting was a Sutherland; but to
Churchill, the painting was a version of himself, and he didn't like it. The
authority in this case belonged to the subject/patron.…"[12] Here a basic
misconception (the confusion of the two meanings of the same word) and
a subordinate misconception (the reference to those who understand the
editorial meaning of authority as believers in something called "the au-
tonomy of the artist") are compounded together with the false equation of
two incongruous artistic genres. A painting is a particular material object;
the destruction of the object is the same as the destruction of the work
of art. A literary work, however, is not a material object, but sequences
of words; the destruction of a particular material object that records the
words does not necessarily destroy the work itself—unless the object is the
only record and no one has committed the text to memory. It must further
be pointed out that for those who understand the editorial meaning of
authority, as a measure of that which came from the author, the authority
of the painting Mrs. Churchill destroyed was unique to that object and
could not be transmitted to another object—even by Sutherland, who in
attempting to do so would only create another Sutherland painting, with
an authority all its own. Whatever aesthetic judgments might be made
of them, hand-painted "copies" of paintings are not copies from an edi-
torial point of view, but unique objects possessing the authority of their
creators—and none of the authority of the "copied" work. The same does
not hold true of texts, where the authority present in the text of one docu-
ment may be transferred in the making of another. Since a literary work
consists of the ordered thoughts of the author (or authors)—conveyed to
the mind of the reader or audience by means of a particular sequence of
inscribed symbols or vocalized sounds—it may be truly represented by
any number of "copies," whether symbolic or spoken, as long as they are
linguistically accurate.[13] If, to respect Orgel's set of parameters, an author
creates a work according to a publisher's general design, it would still
have to be conceded that the design is fulfilled according to that author's
unique artistic inclinations. Literary creations, no less than fashions of
dress or laws, outwardly define a society, but also inwardly reflect its de-
mands. In literature the demands, or influences, may be directly related
to the work at hand, such as a publisher's command or a friend's advice;
or they may be less apparently related to the work but no less formative of
the author's intentions, such as the author's memories of the past, or reac-
tions to contemporary public or private events. That influences play upon
and shape the mind and taste of the author is, however, of keen interest
to the author-centered editor. Furthermore, the author at last must create
the work, and what happens to it after he or she relinquishes it—whether
it is deliberately or accidentally altered by others, for example, or even
altered by the author, under the pressure of an undeniable demand by
some powerful person—does not erase the historical fact of the text as it
was before the alterations were made. How much of the unaltered text
may be recovered is a matter which depends on the totality of the avail-
able evidence, but the recovery of this text has been the purpose of textual
criticism since its earliest known existence.

Other examples of Orgel's use of the word authority might be sum-
moned as further proof of his belief that publishers, producers, patrons,
owners—but not authors—have power and influence in society as a
whole and in its artistic dimensions.[14] Most editors would not argue with
the essentials of this belief (while perhaps feeling that it is stated too cat-
egorically), but neither would they concede that the editorial connotation
of the term authority, as a measure of textual proximity to that which the

author created or intended to create, is much affected by discussions of
other meanings of the word. The exclusive definition proposed by Orgel
might at first seem to be justified by his view of the author—as little
more than an employee carrying out the instructions of the powerful.
Yet even an author (or authors) who creates a play according to someone
else's instructions about plot and characters produces an "authoritative"
manuscript in the traditional meaning of the term. An editor interested
in recreating the author's text would, in addition to repairing it of scribal
or compositorial errors, seek also, evidence permitting, to separate it from
later non-authorial alterations, such as modifications made by others for
performance. For an author-centered critical text, neither the errors nor
the acting changes have authority; an editor pursuing a collaborative so-
cial text would extend the conception of authority to comprehend the
acting changes.[15] Unlike Orgel's conception, however, the conception of
authority modified to guide the social textual editor bears a distinct rela-
tion to the original author-centered one. Authority in the sense of social
power is a self-evidently important concept affecting artistic creations
as much as any other aspect of organized human life. To claim that this
concept negates the editorial meaning of the word, however, is as illogical
as would be a claim that the concept of powerlessness negates the editorial
designation of codex descriptus. Authoritative and derived texts are useful
editorial concepts that cannot be disposed of by value judgments relegat-
ing the author to the margins of the creative process.

Orgel's conclusion about traditional editing, that it is a striving for a
Platonic ideal, emerges from the following set of assertions:

It can be acknowledged, in respect to Orgel's remark about the assump-
tion of the transparency of the printing process, that the discovery and
development of methods of bibliographical analysis in the first half of
the twentieth century tended to foster an overconfident mood concern-
ing possible results. Given early successes some confidence was to be
expected; overconfidence showed itself when bibliography was set upon
problems (especially in Shakespeare) that could not be decisively resolved
within the limits of the existing evidence. One temptation was to see me-
chanical causes in textual variation that all literary sense would suggest
was the result of purposeful alteration, authorial or otherwise. To react as
Orgel does, however, to the early ambitions of analytical bibliography by
positing the simple identification of a work of literature with a "material
embodiment" suggests a kind of idealism in itself, in which documents
necessarily appear, but authorless and without tradition, with the contin-
gent gaps in the historical record transformed into an essential mystery.
Orgel's blanket statement (at the beginning of "What Is a Text?"), that "at
the heart of our texts lies a hard core of uncertainty," seems, however, not
so much an idealist credo about cognitive limits as a casually exaggerated
conclusion about the expectations and results of analytical bibliography.[17]
Whether or not A. W. Pollard, McKerrow, or Greg sometimes expected
too much is a matter of small importance relative to their accomplish-
ments, prominent among which were the establishment of methods of
bibliographic analysis, and the permanent connection of these to the edit-
ing of Anglophone literature. Orgel's awareness of these achievements is
demonstrated in some of his other essays (notably "Acting Scripts, Per-
forming Texts")—even as he attempts to press his main point about the
unimportance of authors.

Orgel's mention of "accidentals" is of course a reference to W. W.
Greg's essay "The Rationale of Copy-Text," in which the authority of
the accidentals (spelling, punctuation, word division, and emphasis) of a
work, was, for certain editorial situations, treated separately from that of
its "substantives" (wording). Anyone familiar with Greg's argument will
at once recognize that it has not so much been misinterpreted as per-
fectly inverted—an unfortunate effect, perhaps, of the disregard of the
editorial meaning of the word authority. Greg conceived of the division
out of his concern that authorial spelling and punctuation be preserved
as far as possible. The concern was connected to the editorial problem

engendered when a revised edition of a work was typeset from an earlier
edition. As Greg observed, the later edition is derivative where it does not
incorporate the author's changes; since accidental variants so often resist
reasonable selection, those of the earlier edition should be preferred in
instances of doubt, as more likely to be authentic. Though the term has
sometimes seemed to invite misinterpretation, accidentals was coined by
Greg in an attempt to indicate greater instability in the transmission pro-
cess, not intrinsic insignificance. Those involved in transmitting the text
seemed either to treat spelling and punctuation with less care than they
did the words, or felt freer to alter them deliberately. Revising authors,
furthermore, were rarely able or willing to halt or reverse the deterio-
rating trend. Finally, since a scholarly editor is a better judge of verbal
variants than accidental variants, Greg recommended that questionable
accidental variants from the revised text be regarded as without author-
ity. A critical edition of the revised text, he furthermore concluded, would
better rely upon the accidentals of the earlier edition in cases of doubt.
The idea, again, was to preserve the most authoritative level of spelling
and punctuation, by "generally" adopting the accidentals of the earliest
surviving document in a series. As for the variant wording, that, said
Greg, "belongs to the general theory of textual criticism and lies alto-
gether beyond the narrow principle of the copy-text."[18]

This particular charge of "Platonism," at least, falls on one of its at-
tempted justifications—an accusation of carelessness about accidentals.
While utterly misplaced in a discussion of Greg's editorial practices, the
accusation, not surprisingly, becomes undisputed fact when its direction is
reversed. As he intimates in The Authentic Shakespeare (p. 17), and demon-
strates in the Pelican Shakespeare, which he co-edited, Orgel is a "mod-
ernizing" editor, preferring to alter Elizabethan spelling and punctuation
according to a necessarily personal notion of what modern readers will
find "accessible."[19] The modernized Shakespeare, a fixture in university

classrooms for generations, was called into question in the first half of the
last century by Greg and his colleagues, who began to pay deserved atten-
tion to period spelling and punctuation. Following them, Fredson Bowers
made old-spelling the practical norm for scholarly editions, and in his edi-
tions always adopted what he believed were the author's accidentals (or
the nearest recoverable semblance thereof).[20] This ought not be forgot-
ten, though Orgel's mistaken statement about accidentals may have been
partly triggered by Bowers's perverse habit of declaring, in contravention
of his practical attitude, that accidentals lacked basic significance. Never-
theless, when editors with social concerns favor modernization—and in
this Orgel joins Jerome McGann—they would seem not only to forswear
all logical criticism of the reasoned eclecticism of author-focused editing
(not to mention criticism of the seriousness with which that orientation
has treated accidentals), but also to create the very dichotomy between
literary and textual criticism that they profess to oppose.[21] The authori-
ally intended critical text is eclectic insofar as it is a reconstruction of a
particular moment in the history of a work of literature to which no one
document bears adequate witness. The editor therefore needs the free-
dom to bring together variants from as many sources as necessary (includ-
ing his or her own mind), but the process of selecting and arranging them
takes place within a strict historical framework. The process is focused on
recovering the author's preferences; one could criticize it for sacrificing
the preferences of society (as represented by scribes, actors, publishers,
proofreaders, printers, etc.), but in a well-constructed critical edition these
are not far to seek, since rejected variants are recorded in the apparatus.
It should not need saying, but by removing the historical limits, mod-
ernizing editors deny readers not only the authorially intended text, but
also the social text. Social textual theory, as articulated by McGann and
endorsed in one or another form by Orgel and many others of like mind,
holds that literary and dramatic works are collaborative products, created
as much by commissioning patrons, acting companies, and publishing
companies as by authors; it should follow from this that the text presented
in a scholarly edition ought to reflect the concept of collaboration. The
implications of the approach remain unexplored, however, since practical
demonstrations in the form of social textual editions are slow in appear-
ing. The expressed preference for modernization suggests, however, that
some social textual theorists would rather not fully extend theory into
practice. Reticence to provide editorial demonstrations of the social text
may arise from a realization that, in discounting the importance of the
author's preferences with respect to accidentals (and substantives), and el-
evating those of society in their place, the new focus leads practically to a
text with the linguistic attributes of the same period as those of the author-
centered text. For Renaissance works, these attributes include early and
irregular forms of spelling and punctuation, unfamiliar to modern eyes at
first glance. Experience has shown, however, that with less patience than
might at first be expected, modern readers can gain adequate facility in
reading typographically faithful renderings of English-language Renais-
sance texts. Nevertheless, some scholars halt after that first glance and de-
cide that the forms of words (and certain words themselves) are obstacles
to understanding and should be removed and replaced with forms the
scholar finds more familiar. This line of reasoning proceeds from some
glaring prejudices: that modern readers will not understand older forms
of the written language, that these forms constitute a discrete rather than
integral part of the language of the original, and that they are inessential
to the meaning and can be exchanged for modern equivalents which will
make the meaning clearer. Whether or not modernization has advantages
in certain circumstances, its blanket application at the university level
obviously impedes literary study. It pulls textual and literary criticism
apart, and then casts the former discipline aside, not recognizing that its
most basic function is textual interpretation; as for the latter discipline,
it is left to beg original meanings of linguistic anachronisms, texts whose
origins have been partly effaced according to elusive notions of what will
be familiar to modern readers. In The Editorial Problem in Shakespeare Greg
discussed the essential importance of accidentals, and of the need to pre-
serve their original forms—even if these reflect only the intrusions of the
author's contemporaries. Of the spelling he had this to say:
Of the need to preserve original punctuation, Greg simply observed, "Just
as the language of an Elizabethan author is better represented by his
own spelling than by ours, so the flow of his thought is often more easily
indicated by the loosely rhetorical punctuation of his own day than by
our more logical system." It is this line of reasoning that has been twice
missed: first when Greg is falsely criticized for believing that accidentals
may be altered without altering meaning, and then when the misinformed
critics circle back and favor or practice modernization.

II

Textual criticism, like all forms of scholarly investigation, is an activity
in which standing hypotheses are tested and revised or replaced, often by
the authors of the original hypotheses. Where the amount of evidence is
limited and fixed, difficult problems may over time be judged insoluble,
but only after an extended period of examination. Greg's scholarship is
in a sense a model of this process, as it is in part a record of corrective
second thoughts prompted by his own or others' further research. The
history of The Editorial Problem in Shakespeare offers exemplary testimony
to Greg's ever-growing views on the text of Shakespeare, most obviously
in the preface to the second edition (written in 1950), in which is of-
fered a catalogue of corrections of discussions appearing in the first edi-
tion, along with references to further and better scholarship on particular
points. The work had its origins in the series of Clark Lectures he deliv-

ered at Cambridge in 1939, but before the lectures could be published,
R. B. McKerrow's Prolegomena for the Oxford Shakespeare appeared. This
important book obviously demanded Greg's attention, especially since
he found himself in some disagreement with McKerrow's conservative
editorial principles. In preparing the lectures for publication in 1942, he
expressed a "more elaborate and less rigid" approach in a preface, and
in his own "Prolegomena" ("Prolegomena.—On Editing Shakespeare")
gave a detailed introduction which contested some of McKerrow's recom-
mendations, but also clarified many of his own. An appended essay, "Note
on Accidental Characteristics of the Text," is a snapshot of Greg's think-
ing on this subject at a point a few years in advance of "The Rationale of
Copy-Text," and can deepen understanding of its better-known successor.
Both the "Prolegomena" and the "Note" are helpful in demonstrating
why the problem of "accidentals" became for Greg a key to unlocking
the theoretical doors behind which the progress of critical editing had
stalled. In his "Note" Greg confesses "strong 'conservative' instincts,"
and an editorial disinclination "to violate" the "integrity" of particular
editions or manuscripts, but explains that these attitudes are reflections of
the "bibliographer's outlook … only a stepping stone, though an essential
one, towards the truly critical position."[23]

Through the development and application of methods of the physi-
cal analysis of documents, this "stepping stone"—which, of course, was
the New Bibliography—allowed for the possibility that judgment might
again play a serious role in the editing of English literature. Judgment
had been robbed of its privileges by foregoing editorial abuses, in which
choices among variants were made haphazardly or according to the aes-
thetic preferences of the editor. The conservatism of Greg and McKerrow
was a natural enough reaction to the undisciplined editorial past, though
they both strained under self-imposed restrictions. The strain was most
felt in encounters with works surviving in multiple authorities, whose in-
terrelationships could sometimes be revealed through a combination of
bibliographical analysis and critical insight. On occasion the presence
of authorial revision was suggested with more or less certainty, and so
emerged the editorial problem of multiple recoverable moments in the
life of literary works. The problem was new, for the most part. For most
surviving ancient and medieval works the possibility of identifying more

than one authorial version is excluded. Noteworthy exceptions are Ci-
cero's Academica, which survives in partial original (Priora) and revised
(Posteriora) versions, and the "Prologue" of the work by Chaucer tradi-
tionally known as "The Legend of Good Women," for which it is ac-
cepted that MS Gg contains the revised version of the earlier text (F)
preserved in several other manuscripts (see also note 33). When A. W.
Pollard and the others began developing methods to analyze printed
books, they could not help but draw lessons from the editorial experi-
ence of earlier literature, while taking note of differences. The differ-
ences seemed first of all to have to do with the amount and quality of the
textual evidence: while authors' manuscripts of works from the English
Renaissance are rare, the surviving early printed editions were, by and
large, made during or shortly after the authors' times. The apparently
better quality of the evidence suggested that methodical investigation of
mysteries of composition and tradition would be well rewarded; where
early expectations were not satisfied, it was not for want of trying.[24] Where
clarity was obtained, however, the nature of the new editorial problem
took shape. In this respect, historical importance attaches to McKerrow's
judgment that certain of Nashe's works survive in an earlier, uncorrected
edition, and a later edition which, while mainly derivative, also reflects
Nashe's corrections and, as with The Vnfortunate Traveller, revisions. This
may have been the first time that a scholarly editor, fully cognizant of the
survival of multiple moments in the history of the work he was editing,
was able to make a conscious choice to focus his critical text on one or
the other.

McKerrow decided on the later versions, and his decision to use them
as his copy-texts was most famously recalled and criticized by Greg forty-

odd years later, in "The Rationale of Copy-Text." This essay (appearing
in the third volume of Studies in Bibliography) opened a new vista in edi-
torial thought, one in which greater responsibility would be referred to
the judgment of the editor—the "truly critical position" of Greg's earlier
"Note on Accidental Characteristics of the Text." "The Rationale" was
one of the many important products of Greg's maturity, appearing in his
seventy-fifth year, more than fifty of which had been given over to the
study of Renaissance literature. In his youth he had been a pioneer of
bibliographical analysis, and in the extremity of age he tirelessly advo-
cated for an editorial approach fitted to its advantages and limitations.[25]
To the point when McKerrow and Greg began to consider the problem
typified by The Vnfortunate Traveller and present in certain other Renais-
sance works, the widely accepted goal of editing—the movement of the
received text closer to that which the author wrote—was of necessity
minimally defined. Though multiple levels of authorial intention are not
generally discernable in Renaissance works, Greg's recommendation in
"The Rationale of Copy-Text" for the construction of a critical text of
a revised work was ideally suited to the few, like The Vnfortunate Traveller,
which evidently survive in sequential series of documents. While many
Renaissance plays have two or more independent, often widely varying
witnesses, their relationships cannot often be established, with the result
that Greg's special copy-text recommendation can rarely be applied as
prescribed.[26] It was, however, in the editing of works of later periods,
which are generally witnessed by a more ample documentary record, that
the implications of Greg's discussion could unfold. The editorial experi-
ence of American literature proved that Greg was in the right frame of
mind, in emphasizing both textual genealogy and editorial judgment, in
the approach to works for which ample evidence had survived. As the
level of evidence rises, however, so does the complexity of the problem
of authorial intention; for much of the literature of the nineteenth and
twentieth centuries, the boundaries of discussion established in "The Ra-
tionale of Copy-Text" are too narrow.

Without expecting to approach it too nearly, McKerrow and Greg
proposed as their editorial goal a text reflecting the moment when the
author submitted a finally revised and corrected manuscript for publi-
cation (or performance)—the "finally intended" text.[27] When editors of
American literature adopted Greg's recommendations they also adopted
this goal, which for them was more nearly realizable—the literature of
the last two or three centuries being represented not only by published
forms, but often by prepublication documents, from authors' drafts to
corrected proof sheets, and sometimes illuminating collateral material
such as authors' notes and publishers' correspondence. The history of
some works is therefore attested to well enough for editors to recognize a
creative process and be able to reconstruct several stages in its life. The
American editorial experience, with Greg's essay providing the theoreti-
cal point of departure, was shaped in many practical engagements with
the copious evidence, and accompanied by many more discussions and
controversies. One of its great fortunes has been to have as its chronicler
G. Thomas Tanselle, who was an important participant in both the prac-
tical engagements (as an editor of the Melville edition), and the discus-
sions and controversies, to which he made many of the most learned and

satisfying contributions. His recent collection Textual Criticism since Greg
(2005) is an indispensable guide to these discussions, which are carefully
recounted, dissected, and clarified in six thick survey-essays written at
intervals during thirty years.[28] The volume contains nearly six hundred
footnotes with references to perhaps two thousand editions of authors and
books and articles on editing. The titles of three successive essays are sug-
gestive of the problems that distinguish the editing of modern literature:
"Historicism and Critical Editing" (covering discussions from the period
1979–1985), "Textual Criticism and Literary Sociology" (1985–1990),
and "Textual Instability and Editorial Idealism" (1990–1995). The titles,
that is, call attention to the problems of authors' changing intentions and
the nonauthorial alteration of texts. Authors of earlier literature changed
their intentions too, of course, and their contemporaries and near-
contemporaries also influenced and altered their works, but this kind of
history only becomes generally visible—and therefore of central editorial
concern—for scholars of recent literature. Many of the most important
pages of Tanselle's surveys beginning with the "Historicism" essay are
therefore devoted to elaborating the conception of authorial intention to
address this history, and discussing the emerging editorial alternative of
social textual theory.[29] Together, the three essays constitute an extended,
many-sided exploration of what becomes visible to the editor of modern
literature: the complexities of the artistic process and the dynamic inter-
action of author and society. Tanselle's handling of several decades' worth
of often rancorous editorial disputes has been exceedingly judicious and
impartial, always seeking to clear away thoughtless or counterproductive
verbiage so that the most useful insights might be better appreciated. The
advantage to any discipline of having its scattered discussions brought
together in the careful analyses of an eminent scholar need hardly be em-
phasized; that much of the rancor which the author defused was directed
at his own views—or at what detractors mistook for them—only in-
creases the overall achievement. Tanselle's many original essays on edito-
rial and bibliographical subjects had become essential reading for editors
of American literature, and so those who began to question the "Greg-
Bowers tradition" naturally saw the author as its inheritor and principal
exponent. This identification was not inappropriate in itself, though the
picture of the tradition painted by many of its critics might have been less
distorted had more practicing editors engaged themselves in the discus-
sions. The Greg-Bowers editorial model has been repeatedly criticized by
a few editorial polemicists for its supposed interest in establishing a single,
exclusively valid constructed text, reflecting only the author's last inten-
tions, while denying the social, collaborative dimension of the creative
process. To sustain this criticism, early, abstract, and simplified formula-
tions of editorial theory were used as points of departure, while practical
developments in the editions themselves were little accounted for.

The effect of these misconceptions was heightened by an initial lack
of clarity on the part of some author-focused editors about the need to
deepen the definition of their editorial goal; later on, as the best editions
began to present critical texts reflecting complex histories of composition
and tradition, much of the insight within their covers went unnoticed
and unpublicized. Apart from those of Bowers and Tanselle, that is, few
contributions to the scholarly journals were made by practicing author-
focused editors, while the journals published little in the way of mean-
ingful assessments of the editions. These factors probably contributed to
the emergence of social textual theory, since it was conceived partly as
an alternative to a perceived narrowness in the authorial focus. In fair-
ness to those who were moved by the arguments of Jerome McGann,
social textual theory's chief progenitor, it should be noted that author-
focused critical texts were sometimes viewed as necessary outcomes of
an objective process, a fallacy that has caused confusion in more than
one generation of scholarly editors. Tanselle began the process of clear-
ing away some of the confusion and rectifying the theoretical disjuncture
between the goal McKerrow and Greg had conceptualized and the edi-
torial realities of more recent literature in his 1976 essay, "The Editorial
Problem of Final Authorial Intention."[30] In characteristically systematic
fashion, the author drew upon practical lessons, most obviously from his
own engagement with Melville's works, as well as pertinent philosophical
and editorial discussions, to sharpen the concept of artistic intention and
demonstrate its connection to scholarly editing. From the practical side
it had already become apparent that many modern works survived in a

series of variant documents, some of which seemed to suggest a progres-
sion, whether of a steady or irregular character, toward a single final state,
while others revealed complex forms of development, in which multiple
"final" states could be identified as possible objects for attempted critical
reconstructions. Not surprisingly, the evidence frequently revealed the
involvement, welcome and not, of persons other than the author in the
creative process. Tanselle discussed the expurgation of Melville's Typee
and Stephen Crane's Maggie, calling attention to a large class of modern
works whose themes and realistic language were more or less censored by
or at the behest of publishers or their agents. A whole range of uncoerced,
genuinely collaborative interactions between authors and others was also
being discovered at this time.

Of the discussions Tanselle examined in this early article, perhaps
the most provocative and helpful in its way was "A New Approach to
the Critical Constitution of Literary Texts" by Hans Zeller.[31] Zeller pro-
ceeded from an understanding that modern works present editors with
rich and visible histories; he endorsed the concept of authorial intention
as the proper focus for a critical edition, but rejected what he under-
stood as the automatic preference of the Anglo-American approach for
the author's last intentions. He furthermore disavowed editorial eclec-
ticism, believing that each variant document, even if distinguished by
but a single authoritative alteration, corresponded to a particular set of
authorial intentions, and was therefore a discrete version of the work in
question. In supplying a brief informative historical sketch of German
editorial practices, especially in the determining field of Goethe criticism,
Zeller found cause to examine the edition of Goethe's works prepared by
the Deutsche Akademie der Wissenschaften zu Berlin (DDR), from 1949.
The coincidence of this date with the appearance of "The Rationale of
Copy-Text" enhances Zeller's comparison of Greg's recommendations
with the practices of the edition's first editor, the learned German Jewish
philologist Ernst Grumach (1902–67).[32] Zeller found much in common
in the approaches of the two contemporaries, though followers of Greg
might find more correspondence between Grumach's views and those

of R. B. McKerrow, at least in the matter of copy-text choice, since, in
Zeller's description, Grumach used revised editions as copy-texts for the
two works he edited (the epics Reinecke Fuchs and Hermann und Dorothea).
Zeller's main reason for pairing Greg and Grumach, however, seems to
have arisen from the willingness of both editors to draw readings from
different authoritative documents to serve their goal of an authorially in-
tended critical text. Zeller's theory of insular documentary-based versions
prevented him from endorsing their approaches, which he equated with
"contamination"—the term used in the editing of ancient texts to indicate
the classic problem of a single document carrying variants from alternate
traditions and so resistant to stemmatic classification.

Despite his rejection of the logic of Greg's rationale, Zeller did state
the editorial problems of modern literature with admirable clarity, and
was especially helpful in stressing that modern works could survive in
multiple versions. A weakness for exaggeration sometimes misdirected
attention when an issue of importance was being discussed: it is unhelp-
ful, for example, to consider the "editorial principle of final authorial
intention" in light of Zeller's suggestion that its "logical conclusion" would
be the "annihilation" of works such as The Aeneid, whose authors willed
that they be destroyed.[33] The belief that "final intention" must mean the
author's last expressed wishes for his or her work has proved to be a
persistent one. Certainly when J. P. Postgate spoke in 1911 of the "ul-
timate intention" of Marlowe with respect to the variants of the widely
divergent 1604 and 1616 versions of Doctor Faustus, something like last
(known) wishes was meant (see note 27); instructions to destroy works,
however, are rarely of use to the editor of a critical text. Zeller's view
that each variant document (and even each layer of alteration on each
document), constitutes a discrete version, was in part a reaction to his
mistaken belief that the Greg-Bowers approach held either a "morpho-
logical" or "teleological" view of literary works.[34] These notions—which
seem suggestive of a work of art perfecting itself (or maturing), in a ba-
sically involuntary natural or philosophical movement—are of course
foreign to modern scholarly editing. Yet few would deny, as Zeller seems

to do categorically, the existence of an artistic process, in which artists
develop works, with one or another degree of deliberation, toward points
of completion. It is a process that most would agree can be revived, in
pursuit of a revised end. Whether all the development within this process
is constructive, or should be described as progress, are not strictly edito-
rial questions, though they will likely demand the attention of the editor
of a modern work, at least when it comes to deciding which point of the
process should be the object of the attempted reconstruction. For Zeller,
however, authors' interest in perfecting their works, or their response to
others' suggestions or demands, are, in the distinction he draws, matters
for literary rather than textual analysis. The editor's job as he sees it is
limited to recounting the history of the work in its versions and selecting
the version to be presented as the critical text, which is defined as the
text of a particular document, cleared by the editor of simple errors. He
criticizes attempted reconstructions that purport to go beyond particular
documents, to reflect the author's uncompromised intentions. His reasons
are clearly stated: referring to Fredson Bowers's relatively straightforward
restorations in The Blithedale Romance of manuscript deletions of a few ref-
erences to liquor and sex, Zeller writes, "I find the idea perplexing, that
the editor should feel obliged not only to make inferences about the final
intentions of an author, but also about the causes behind these intentions.
I cannot regard the psyche of an author and its analysis as a substantial
foundation for editing."[35] This remark is more consistent with Zeller's
practical recommendations than is his earlier avowal of the goal of an
authorially intended text. The creative human mind does not always re-
spect documentary boundaries. The original documentary record, never
having been complete, is liable to have eroded by the time a work comes
in for scholarly attention. Editors of modern works commonly seek to
reconstruct texts of lost documents—fair-copy manuscripts, typescripts,
proof sheets—from descendents and ancestors. Only by drawing readings
from both an early printed edition and a draft manuscript, for example,
can an editor reconstruct the text of the intervening lost authoritative
document. In other words, for some historical purposes, only an eclectic
text will answer. There are limits of course.

Given his view of literary history as cut-and-dried matter, Zeller un-
surprisingly found Bowers's treatment of Stephen Crane's Maggie espe-
cially untenable. Maggie is uncharacteristic of Bowers's work in that it

comes dangerously close to breaching the historical limits that give criti-
cal editing its essential meaning. From a single round of revisions Crane
made in the text of the privately printed 1893 version of his story for
publication by D. Appleton in 1896, Bowers sought to preserve only what
he regarded as "stylistic changes as well as literary improvements," while
rejecting bowdlerizing changes the publisher had demanded.[36] For his
critical reconstruction of The Blithedale Romance, Bowers believed he could
reasonably restore three references to liquor and sex that were deleted
in manuscript, on the grounds that, unlike the bulk of the alterations,
these three did not reflect Hawthorne's own inclinations. They seem to
have been made out of deference to polite taste (possibly as represented
in the person of Sophia Peabody, Hawthorne's wife). A simple case like
The Blithedale Romance, in which a few revisions seem clearly out of step
with the rest, allow us to conceive of an author separately serving his or
her own intentions and someone else's while engaged in a single round of
revisions of a work. Bowers argued that two sets of intentions were also
served in Crane's revision of Maggie, but the editor was unable to identify
the two types of changes convincingly, as he had done for the Hawthorne
work. The basis of his critical reconstruction was questioned, politely by
Tanselle in "The Editorial Problem of Final Authorial Intention," and
more forcefully elsewhere.[37] His acceptance of the bowdlerized version
of the ending of chapter 17 attracted particular unfavorable notice. One
would have to assume that the evidence for alterations serving the pub-
lisher's rather than Crane's intentions was in the toning down or removal
of highly realistic language or subject matter—curse words, primarily,
and, in chapter 17, Maggie's encounters with prospective customers. By
restoring only some of these deletions in the critical text, Bowers seemed
to negate the controls he had established for his reconstruction.

Bowers called his critical text the "'ideal' text of Maggie as a literary
fact, not a limited 'ideal' text either of the 1893 or of the 1896 edition."[38]
This statement is not worded very clearly, nor was its meaning clarified
elsewhere in Bowery Tales, the volume containing the story, but it seems
to indicate a belief that the critical text was an ideal reconstruction of
the work as a whole, as opposed to a rendering of one or the other ver-
sion. If this surmise about Bowers's meaning is accurate, then he was

either basically confused on this occasion, or momentarily at odds with
the general direction of the editorial tradition he was otherwise involved
in upholding. Within the author-focused model, a critical text can be
said to represent the work as a whole only when the evidence indicates a
creative process defined by a single culminating point or when there is not
enough evidence to do anything else. Otherwise, a critical text can only
represent a particular moment in the history of the work. Bowers's Maggie
was more truly an attempted reconstruction of a presumed, intermediary
third version of the story, which, in the editor's view, existed once in the
author's mind, but only there. It might, however, be recovered in the
comparison of the text of the first edition with that of the revised one. The
authorially intended text is often partly defined by the rejection of altera-
tions made by or at the behest of publishers or their agents—especially
those with the suspected purpose of censoring coarse language or sensitive
subjects. Yet a scholarly editor must construct a critical text according to a
viable historical hypothesis, whereby it is argued that the author intended
for the work to take the chosen form at one time or another—though the
factor of time cannot be defined too narrowly, since a particular version
may have taken shape over an extended period. Bowers's claim for the
critical text of Maggie indicates that he came to believe that Crane had
corrected and improved the text according to his own tastes, quite apart
from simultaneously fulfilling the bowdlerizing demands of his publisher
(possibly with the assistance of Ripley Hitchcock, Appleton's editor). Had
he neatly separated the bowdlerizations and the improvements, and ad-
opted only the latter, his text might have attracted less criticism. Even
better would have been a clear, convincing demonstration, if such were
possible, that Crane had conceived of such a text. As it was, the edition
of Maggie must be counted among those few early occasions when Bowers
did seem to misrepresent the critically reconstructed text in metaphysical
terms, as the definitive synthesis of a literary work's constituent histori-
cal moments. Opponents of the "Greg-Bowers" editorial approach have
tended to fix their arguments to these early missteps, allowing themselves
thereby to miss a world of subsequent experience, in which editorial un-
derstanding developed and matured.

III

Scholarly attempts to go beyond the texts of existing documents to-
ward closer re-creations of what authors wanted have always provoked
doubt and criticism. Those in the field of Anglo-American literature who
have recently imagined the "hegemony of the Greg-Bowers-Tanselle re-
gime" (to use a typical description), seem unaware that the controversy
over editorial freedom and eclectic practices winds backwards in a cycle

measured in centuries, even to the days of Poliziano and probably be-
yond, wherein, furthermore, the favorable position has been less than
dominant.[39] Even the leveling of the charge of "Platonism" against edito-
rial eclecticism is not a new pastime, as the experience of Bowers's first
critical texts nearly half a century ago demonstrates. A few new features
of the opposition to editorial eclecticism have been observed, including
the belittlement of the author and the corresponding aggrandizement of
non-authorial contributions to the text. The striking feature of much of
the criticism, however, is the absence of careful analysis of editions and
critical texts and the presence in its place of sweeping generalities, often
cast in philosophical language, and leaning, to one or another degree, on
jargon. The following crowded passage, taken from David Greetham's
Theories of the Text, epitomizes most of the arguments.

While it may not seem so at first glance, this passage is organized in a
manner helpful to understanding the basic position held by many recent
opponents of critical editing. The usual elements of the false image of the
"Greg-Bowers" editorial approach are present: the Platonic and Roman-
tic idealism, the overriding faith in the "single unitary consciousness of an
originary creator" (i.e., the author), and the denial of the social dimension
of texts. Yet Greetham goes beyond repetition of this often-seen litany,
when he introduces a charge of Hegelianism, which, while surely sugges-
tive to many readers of an intention to obfuscate matters, can be used to

clarify a pervasive misunderstanding about textual criticism. Greetham
describes the "Greg-Bowers" editorial approach as Hegelian in two ways.
For the first he attempts to fit two familiar clichés used to describe the
approach—the "Romantic" notion of the ideal text and its companion,
the solitary genius creator—into the Hegelian dialectic. His second use
of the description is the more telling. Accordingly, the reader imagined
by the Greg-Bowers editor is one who accepts that literary works have
their origins in the mind of the author, their material manifestations be-
ing concrete extension ("superstructure") of the original notion (intension,
"base"). Greetham attempts to identify traditional editorial approaches
with "pure" Hegelianism, assuming that to those readers for whom Hege-
lian dialectics is at all meaningful, the meaning is dependent on the great
inversion accomplished by the Left Hegelians (Strauss, Bauer, Stirner,
Feuerbach), and especially, of course, by Marx and Engels. The dialectic
was not set aside in this inversion (except by Feuerbach), but its origina-
tor's ideal starting point was repositioned as extension—as the historically
determined outcome of being. Priority is granted not to the self-moving
absolute idea, but to a dynamic material world—to concrete natural and
anthropological reality. The assumption in Greetham's passage is that
any viewpoint adopting a "pure" Hegelian dialectic—in which nature
and society are mere developmental shadows of the unfolding idea—will
be regarded as quaint or absurd. But absurdity more obviously attaches
to the suggestions that the "Greg-Bowers" editors are wrong, and that
the thoughts of the author somehow do not precede the physical records
made of them. For those who accept the inverted or materialist dialec-
tic, the real begins with the material, but it does not end there. The
thoughts of the author, while not prior in an absolute sense, obviously
precede the records that were made of them, though the records—e.g.,
paper and ink—are basically material in nature and the thoughts are
not. While not images of their obviously posterior records, the thoughts,
however, are images of prior experience of the natural and social world.
The paper and ink display the markings needed for reconstituting these
thoughts, and the fidelity of the representation is the concern of textual
criticism. While the intellectual premises of traditional critical editing
need not be explained within the framework of the historical materialist
outlook, they are fundamentally historical in nature and consequently
do not exceed the boundaries of philosophical materialism. An editor's
personal endorsement of philosophical idealism, on the other hand, does
not directly bear upon his or her reconstruction of a historical moment
in the life of a literary work, since that activity is necessarily governed
by the pertinent evidence and the soundness of the editor's judgment in
working upon it. The ideal character of the reconstruction, furthermore,
has to do with the nature of thought and language, and is not indicative
of Platonic or Hegelian belief.[41]

In arguments asserting that author-focused editing proceeds from a
belief in an ethereal perfect text, the eclectic critical text is not easily
discernable. It is described falsely, as the embodiment of the editor's aes-
thetic preferences—an ahistorical amalgamation of favorable readings
taken from various sources. It should be not be necessary to repeat at this
point that what is of concern to the author-focused editor are the prefer-
ences of the author, and how they developed over time, and whether and
where these conflicted with the preferences of others involved in copying,
publishing, or performing the work being edited. The counter-arguments
gather some force, perhaps, in the word "eclectic," which retains the
ancient connotation of a selection of pleasing philosophical beliefs. One
part of the connotation, selection, does apply to the editorial tradition, as
G. Thomas Tanselle emphasized in his outstanding essay, "Editing with-
out a Copy-Text."[42] There editors were advised to serve their historical
ends with critical texts wholly constructed by selecting readings from per-
tinent sources, rather than emending a pre-existing copy-text. In reason-
ing beyond the copy-text, a document that has often proved debilitating
of editorial judgment, Tanselle was seeking to sharpen the historical focus
of critical editing. With no particular document automatically determin-
ing questionable readings, the editorial process remains keenly centered

on the history of the work and the point in its development chosen as the
object of the reconstruction. To say that this process is "eclectic" is simply
an acknowledgment that in the effort to account for human intentions and
frailties, and gaps and errors of all kinds, the critical reconstruction nec-
essarily adopts a historical principle of greater compass than that which
might be defined by the texts of particular documents.

A generation earlier, Fredson Bowers, searching for a useful defini-
tion of eclecticism, noted that in the "purest sense" the term described
"any text admitting emendation of error," since the "principle of choice
has been invoked."[43] Aware of contemporary opposition to eclectic texts,
however, Bowers recognized that what was at issue involved more than
the correction of obvious typographical errors in a single document. "Re-
marks on Eclectic Texts," the essay in which he expressed these thoughts,
was therefore given over to discussing exemplary situations taken from
British and American literary works dating from the end of the sixteenth
to the end of the nineteenth centuries. The diverse examples—including
works from Shakespeare, Beaumont-Fletcher-Massinger, Fielding, Haw-
thorne, William and Henry James, and Stephen Crane—allowed Bow-
ers to discuss the significance of eclectic editorial practices as they grew
more complex along a range of single- and multiple-authority traditions.
He recognized that editorial eclecticism had not gained general accep-
tance at the time of his "Remarks" (1973), writing of the "rigid adherents
of faithful reprints of one authority in American literature and among
English scholars in relation to nineteenth-century works."[44] On the other
hand he noted that at the time, opposition to eclecticism did not extend
to the works of Shakespeare. The source of the inconsistency—a failure
to recognize that transmissional carelessness and intentional tampering
could erode the authority of a work no matter what period it was created
in—was well understood (if not always clearly explained or handled) by
Bowers, who had scholarly familiarity with works of many centuries.

As he acknowledged more than once, Bowers took his conception of
"authority" from R. B. McKerrow, principally from Prolegomena for the
Oxford Shakespeare, which is to say that he agreed that authority came from
the author. This definition, as we have seen, runs against recent attempts
to extend the concept, and also contradicts attempted new interpretations
of McKerrow's views. His views on this matter were, however, already
clear enough when he made them explicit. His skepticism concerning the
editorial ways of predecessors and contemporaries, furthermore, did not
prevent him from distinguishing the text the author intended to publish
(or have performed) from the existing, flawed testimony of it—the texts of

documents. He wrote of documents "having" authority, and of the most
authoritative document being that which in the editor's opinion must
have deviated least from the author's autograph manuscript—the line
of reasoning that led him to name and formalize the concept of "copy-
text." He intended to emend this document, however, as he made clear in
Prolegomena and even demonstrated in the two tantalizing appended pages
from his projected edition of Richard the Third (to be based on the Folio
text). The sources for the emendations would include authoritative docu-
ments other than the copy-text and also early derived documents—since,
as the author helpfully emphasized, for literature from remote periods
these contain valuable early corrections (or endorsements) of doubtful
readings.

While sometimes viewed as having identified authority with its mate-
rial (documentary) embodiments, McKerrow was aware of a less tangible
essence of literary works. The conception of final authorial intention as
expressed in Prolegomena—an imagined "fair copy, made by the author
himself, of his plays in the form which he finally intended to give them"
(p. 6)—did not equate an author's intended text with the text of any
document existing or lost, including those inscribed by the author. In at-
tempting to account for the obviously corrupt condition of certain printed
Renaissance dramas, McKerrow explored the problem of control of the
text in its manifold details, as it moves from copy to copy. In this context
he had cause to discuss the relative merit of a fair copy of a draft manu-
script made by a professional scribe, as opposed to the author, recogniz-
ing in the problem both psychological and practical dimensions.

It is doubtful that even today's critics of the early hypotheses of McKerrow
and Greg concerning the origins of the poorly printed plays would take
issue with these seemingly self-evident observations. On the other hand,
in pointing out the obvious practical difficulties of producing an accurate
copy of a literary work, and recognizing that even (or especially) authors
are prone to confusion and error when attempting to copy their own
work, McKerrow called attention to something essential about the nature
of the written language. To speak only of modern alphabetic writing, this
consists of complex systems of signs representing single sounds, which
are combined into groups representing words, and their nonverbal ap-
purtenances, including inflection (emphasis) and pauses (punctuation), all
further sequenced within longer syntactical constructions. In long written
works these constructions are embedded within still larger arrangements
created according to the author's greater expository, aesthetic, rhetori-
cal, and logical purposes (whether or not these purposes are partially
responses to the preferences or demands of others invested with a measure
of influence or control over the text). At the most narrowly material level,
the outcome is a great sequence of thousands or hundreds of thousands
of letterforms, other symbols, and spaces. An author attempting to cre-
ate a precise symbolic representation of his or her thoughts thus faces
substantial practical difficulties, which also await anyone trying to make
a faithful copy of the original.

Yet these marks are intended to convey linguistic meanings, and
while the meanings are not easily recollected without their corresponding
marks, they have priority in relation to them. From an awareness of the
linguistic and material complexities that make all texts difficult to control
and vulnerable to unintended error, can follow the basic conceptual dis-
tinction between documentary texts and the works they represent. This
distinction makes the editorial goal of authorially intended texts under-
standable, along with the need for judgment-based eclectic practice in
order to attempt them. The distinction has been thoughtfully explored by
G. Thomas Tanselle on many occasions, beginning as early as his 1981
essay "The Texts of Documents and the Texts of Works,"[46] and most

memorably in his Rosenbach Lectures of 1987, widely known in their
printed form, A Rationale of Textual Criticism.[47] One of Tanselle's purposes
has been to make plain that the term "verbal work" describes a fairly open
concept, with a formally limitless chronology and content, while a docu-
mentary text is a particular material object, created at a specific point or
points in time. A second purpose has been to clarify the nature of verbal
works using a comparison with works of the plastic arts: unlike paintings
or sculptures, verbal works cannot be directly experienced by beholding
material objects. Our experience of them does depend on documentary
texts, which are material objects, and so may be appreciated for their
physical appearance, and may contain illustrations and decorative fea-
tures intended to be integral parts of the works they represent. The verbal
work per se, however, is not a material object, nor is it communicated by
purely sensational means. Comprehension is basically an intellectual pro-
cess, which begins with a literal interpretation of the sequences of symbols
as transmitted in the documents. These prompt the reader to recall the
words and then re-imagine the thoughts of the author, but the potential
for understanding is regulated from the beginning by the level of accuracy
attained in the sequences of symbols.

While the practical complexity of the written language is the cause of
much error, it also provides textual criticism with one basis for the reason-
able analysis of the text, whereby puzzling readings are either determined
to be errors and possibly corrected, or endorsed as true. Questionable
passages can be tested against expected linguistic patterns, by considering
them in the light of literary factors such as meaning and prosody, or fac-
tors relating to aspects of the physical document, including handwriting
or typography, as well as the dimensions of the page or column. These

factors are further considered within the context of what is known of the
movement of the text from copy to copy or edition to edition. In reflect-
ing on his restoration of the true name of Sextus Cloelius—a confederate
of Cicero's enemy P. Clodius Pulcher—the late D. R. Shackleton Bai-
ley reviewed the conflicting manuscript readings in Cicero's letters and
speeches, in which the name appears mostly in forms either of Cloelius
or Clodius. He noted that "[c]opyists were under no temptation to make
'Cloelius' out of 'Clodius'. The temptation was all the other way, espe-
cially in view of the association of Sextus with P. Clodius Pulcher."[48] In
addition to the historical context—the association of the two like-named
men—Shackleton Bailey attributed the error—Clodius for Cloelius—to
the likelihood of "el" being miscopied as "d", rather than the other way
around, and demonstrated that the archetype of the various witnesses
must have contained forms for Cloelius.

The title of the article in which this discussion appeared, "Mump-
simus—Sumpsimus," makes a generally useful point about editing, in
referring to the supposed riposte of an errant English priest on being
corrected of his nonsensical misreading in Mass of the phrase, "quod in
ore sumpsimus" (in the mouth we have taken): "I will not change my old
mumpsimus for your new sumpsimus" (related by Richard Pace in 1517,
as quoted in the OED). Erroneous readings tend to live on after they
have been conclusively disestablished, since they can usually claim what
is sometimes (and misleadingly) referred to as "documentary authority."
The presence of a reading in an authoritative document is obviously of
basic editorial interest, but the term is a deceptive syllogism, since docu-
ments can contain errors of every imaginable kind. Many of these can be
detected and repaired by further editorial investigation, the outcome of
which might be a defense of a variant reading present in an alternative
document, or an editorial conjecture—a reading which, though not pres-
ent any existing document, might be the author's own expression, recalled
from oblivion by editorial erudition and brain power. Shackleton Bailey
first published the evidence for his recollection of Cloelius in 1960; he
wrote "Mumpsimus—Sumpsimus" in 1973, partly to record the subse-
quent persistence of "Clodius" in editions of Cicero made by those who
should have known better. W. W. Greg's warnings about the "tyranny" or
"mesmeric influence" of the copy-text, well familiar to readers of Studies in
Bibliography, are meant to teach the same lesson as the title of Shackleton

Bailey's article, a coincidence indicating that, as Greg believed, general
editorial principles belong to all historical periods.

Hans Zeller's stricture, mentioned in section II, against editorial at-
tempts to analyze authors' minds would attract many latter-day endorse-
ments, but presumably from those who would reject the author focus of
Zeller's editorial approach. The incongruity applies in both directions,
since most author-focused editors assume their discipline has a psycho-
logical dimension. The concept of authorial intention in editing involves
the editor in an investigation of the author's mind, as well as of the minds
of those involved in copying or publishing the author's work. There are,
obviously, many potential limits on what can be known of the author's
habits, artistic inclinations, and stylistic preferences, and a prudent edi-
tor is always conscious of them. It is, however, in the nature of the edi-
torial temperament to test those limits. In the editing of classical texts,
one of the forms that the concept of authorial intention takes is known
as usus scribendi, or the author's usual practice, a criterion for judging
variants which has been worked since at least the days of Aristarchus. As
with the example from Cicero of paleographic analysis, usus scribendi was
developed by editors of works from the remote past, and then adopted
by editors of literature of all periods. Whether called by its Latin name
or not, it describes an important dimension in the evaluation of variant
readings from alternate texts believed to have descended independently
from a lost common ancestor. Given what is known, for example, of Mark
Twain's early style and preferences, the doomed cat, sensing that she was
about to be "et up" (rather than "eat up") was more likely to have been
"a grabbling" (as opposed to "grabbing") for the flower pot from which
"Fitz Smythe's Horse" was attempting to snatch her.[49] The pairs of vari-
ants come from independent contemporary reprintings of a lost original
printing, and it is nearly certain that only one is authoritative. In judging
between them, a second criterion, related to usus scribendi, might also sug-
gest itself to an editor: lectio difficilior potior —the more difficult reading is
preferable. This criterion is likewise associated with the editing of older
texts, but is equally valid for texts of all periods; it arises from the obser-
vation that unusual expressions unfamiliar to copyists (or typesetters) are
exposed to the danger of being miscopied or replaced by more common
expressions, whether inadvertently or by interpolation. The difficult vari-

ant, in other words, is more likely than the common one to be the original.
The two criteria converge in the little example from Mark Twain, since
the "author's usual practice" involved a wide vocabulary which included
more than its share of uncommon words or phrases (lectio difficilior). The
convergence is useful to keep in mind, since in the classical field the op-
posite observation about the two criteria has been made: that they work
in different directions, and lead toward divergent results—usus scribendi, it
has been said, moves an editor toward smoother, linguistically normalized
texts, while lectio difficilior favors textual dissonance and idiosyncracy.[50]
This opposition can also be observed in very recent texts, as when Bow-
ers, who used both criteria (though perhaps without awareness of their
Latin names) when he edited Stephen Crane, came upon "glintered" and
"glinted" in alternate authoritative texts. He chose the more common
variant, and thus usus scribendi over lectio difficilior, since he was unable
to locate in Crane another example of "glintered." The dissatisfaction he
expressed over his choice almost certainly arose from his awareness of
Crane's preference for unusual words and spellings.[51]

Paleographic analysis, lectio difficilior, and usus scribendi have been
widely applied in the editing of works from all periods, though their mer-
its and limitations tend to be most actively discussed by editors of an-
cient works. Like bibliographic analysis (of manuscripts as well as printed
books), methods of linguistic study use the physical evidence of particular
documentary texts, along with the general properties of the written lan-
guage and the means by which it is conveyed, to understand the inten-
tions and actions of those who created the documents—the author as
well as others involved in copying or publishing the work. Intention is
a basic editorial issue, since linguistic complexity makes control of the
text hard to achieve. The obligations of the scholarly editor, whether in

reconstructing the archetype of an ancient work or a point in the history
of a modern one, are to understand these intentions and actions as far
as possible, and divulge the results of the analyses that have been under-
taken. These obligations would extend, furthermore, to the hypothesized
social textual edition as well, since the goal of such an edition no less than
an author-centered one would be to lay bare the details of the collabora-
tive and conflicting relationships of the principals involved in the early
performance or publication of the work.

IV

Since critical editing had its beginnings in the treatment of works
from the remote past, for which authoritative evidence is comparatively
slight, editors who began the process of applying scholarly methods to
more recent works were faced with some problems of adaptation. Means
for handling more evidence from authors' times had to be developed, but
more obviously the critical text had to be redefined with greater preci-
sion, so that unmistakable was its status as a conjectural reconstruction
of a single moment in the life of a work that in its totality may be defined
by several recoverable moments. The historical focus of the critical text,
that is, must narrow and sharpen as the level of original evidence rises;
it must not be conceived of or presented as a supra-historical synthesis
of the work as a whole—as Bowers had mistakenly done when he edited
Maggie. Time was needed for the historical purpose of critical editing to
reassert itself in its new surroundings. After Greg's examples of autho-
rial revision in "The Rationale of Copy-Text," Bowers, fresh from his
experience editing the dramatic works of Thomas Dekker, pushed un-
derstanding forward in "McKerrow's Editorial Principles for Shakespeare
Reconsidered."[52] Here Bowers indicated that a dramatic work could have
several "final" forms, prepared for either literary publication or stage pre-
sentation. He recognized that the critical text would aim at reconstructing
one of these versions, with the variants of the others recorded in the ap-
paratus. Shakespeare critics usually identify performing versions by their
comparative brevity, and few editors—Stanley Wells, Gary Taylor, and
their associates of the Complete Oxford Shakespeare being the notable
exceptions[53] —have been willing to set aside suspected authorial passages

in order to present the cut-down theatrical version. While Bowers (like
McKerrow) was with the editorial majority in his orientation toward re-
constructing the fuller versions, his method called for careful investigation
of the sources of variation. In a few multiple-authority plays of Shake-
speare, certain passages are not present in both or all three sources, giving
rise to questions of who made the cuts, and why: whether they were made,
for example, to serve the (potentially related) interests of dramatic struc-
ture and theatrical performance. The questions sorely test the evidence
in Shakespeare, but their difficulty does not alter the fact that they are of
basic interest to the author-centered editor.

In light of the foregoing consideration it becomes apparent that an em-
phasis on performing versions of plays—such as that exhibited by Wells,
Taylor, and associates—while requiring uncommon editorial choices,
does not carry an editor as far from the tradition of Greg and Bowers
as the tenor of recent editorial disputes would suggest. The distinction
between social, collaborative, or performing texts on the one hand and
authorially intended texts on the other has been drawn so starkly in re-
cent times that the two sides are at best thought of as mutually exclusive
alternatives. Logic and experience, however, suggest a less than absolute
divide, which varies from situation to situation. The possibility of con-
structing an authorially intended critical text is unaffected, for example,
by the matter of whether the work being edited has one author or several:
collaborated works have been the object of this approach since Bowers's
edition of Dekker.[54] Wells and Taylor and their colleagues have been criti-
cized for their belief that the shorter versions of Shakespeare's plays were
made either by the author or by others with his approval, in the pur-
pose of improving their "overall structure and pace.…"[55] The criticism

has come, not surprisingly, from both traditional Shakespeareans and
those caught in the prevailing current of questioning the importance or
even the existence of authors.[56] The evidence for or against the presence
of authorial revision in Shakespeare is scant and much in dispute, but
where authorial revision has been reasonably hypothesized, nothing
prevents an author-focused editor from constructing critical texts of the
revised versions. On this subject, it is worth remembering that Greg for-
mulated his ideas on copy-text in response to the problem of authorial
revision.

As works created in ever more recent periods began to draw editorial
attention, the levels of evidence encountered demanded critical texts and
apparatuses with an increasingly precise historical focus. The widest edi-
torial experience in modern letters has been gained with the works and
papers of American authors; results have been uneven, but in places the
field has given bloom to sound, exciting scholarship. Reference to a few
exemplary bright spots will, it is hoped, serve to indicate the dimensions
of the editorial experience of recent works, for which multiple signifi-
cant moments are recoverable. Here comparatively plentiful evidence of
composition and printing history, while never amounting to a complete
record, throws into relief a range of textual problems that remain mostly
hidden for literature from more remote times. The detection of errors
remains of primary importance in the editing of modern works, but clari-
fication of the history of the intentional alteration of the text becomes an
equally important objective. The authorial orientation of the traditional
approach can and should concentrate editorial attention on the interests
of others involved in the publishing process as well as on the trajectory of
the author's changing intentions. Editors of modern works soon become
familiar with a great range in the sources of alteration, including purely
authorial revision, the author's acceptance of the suggestions of an invited
reader, his or her acquiescence to the textual demands of others whose
approval was needed for publication or performance, and the imposition
of publishing- or printing-house styles in matters of spelling, punctuation,
or even grammar. The many colorfully worded complaints about printers
and proof-readers made by exacting writers like Mark Twain and A. E.

Housman attest to their keen interest in textual control, the satisfactory
attainment of which eluded their most diligent efforts.[57]

The point at which authors seem finally to accept the alterations of
others does not necessarily or even usually equate with the editorial idea
of authorial intention, and scholarly editors who focus on the latter are
justified, evidence permitting, in reconstructing texts as they stood be-
fore being so altered. This was the course followed by Harrison Hayford,
Hershel Parker, and G. Thomas Tanselle, editors of the Northwestern-
Newberry Typee (1968), when they presented a critical text of Melville's
story repaired of the textual contributions and subtractions which had been
forced upon the author by his American publisher, John Wiley of Wiley &
Putnam.[58] Typee presents, for Melville's works, a textual situation of un-
usual complexity, and though the volume was the first in the Northwestern-
Newberry edition, it comes very close to providing a model of historical
editing. The first American edition of Typee (A), hurriedly set from proof
sheets of the English edition (E), suffered not only from careless typeset-
ting but also from small changes apparently made by Wiley, who softened

language or gouged out passages that offended him. Some months later,
seizing upon Melville's modest suggested alterations (and restorations)
for a revised edition, Wiley demanded of him a thorough cleansing of
most wording touching on sexual and religious matters. Melville had little
choice but to acquiesce, and so the American revised edition (AR), while
containing some authorized changes (including "The Story of Toby"), is
chiefly distinguished by euphemizing corruptions and missing passages.

After analyzing this history, the editors decided to construct a critical
text which referred to the moment when Melville submitted his changes
to Wiley. The decision was motivated neither by aesthetic considerations
nor personal preferences (accusers of editorial "Platonism" take note),
but by an interest in carrying out, as far as the evidence permitted, the
author's last known wishes for his work. The decision of course also meant
the denial of Wiley's wishes, and so as is often true of author-centered
critical texts, the outcome of an earlier, coerced agreement of publisher
over author would be reversed. The editors' goal was ambitious given
the limits imposed by the available evidence (rich as this evidence might
seem to an editor of works from the remote past). The complete loss of
Melville's final manuscript[59] was one obvious limitation, and another was
the near complete absence of evidence of the early revisions made by
the author's brother, who in London arranged for first publication, as
well as of those of John Murray, publisher of that first, English, edition,
and Henry Milton, Murray's reader. The editors began with E, the earli-
est extant complete authoritative text, but this text already incorporated
more than one level of non-authorial alteration. As often happens in lit-
erary archaeology, several earlier moments in the history of a work were
found in E, but in a collapsed state. Only the last of these moments was
meaningfully visible; from internal and external evidence the erstwhile
existence of the earlier moments could be recognized, but little could be
done further to distinguish them. The editors restored American spellings
to some "our/or" words,[60] but made little headway in identifying what
had been altered, let alone recovering Melville's original version. In A,

on the other hand, Wiley's depredations were easily identifiable, both
because enough evidence survived and Wiley's interest was understood
(illustrating not only the editorial relevance of psychology, but also the
author-focused editor's appreciation of the intentions of persons other
than the author). Thus were some substantive variants in the first Ameri-
can edition identified as Wiley's work; as for what Melville called the
"Revised … Expurgated?" edition (AR), the editors subjected this text to
much analysis. In addition to the thorough bibliographic examination of
AR, aesthetic, social, and psychological factors were taken into account,
with the result that the editors were able to distinguish, in the matter of
the substantives at least, some or all of Melville's earlier, uninfluenced
revisions from those Wiley had forced upon him.[61]

The role and importance of the critical text has rarely received so
dramatic a demonstration as it does in the scholarly edition of Melville's
Typee. The evidence presented there concretely reveals that the two first
editions imperfectly represented Melville's early intentions (E a good deal
less imperfectly than A), and that the revised edition clearly betrayed his
later ones. These pre-existing documentary texts, records of the author's
interactions with his publishers and first readers, are of permanent and
undeniable value. Yet few would deny equal value to the constructed
critical text, which convincingly approaches the Typee that Melville at last
wanted. The totality of the evidence displayed by the scholarly edition, in-
cluding the critical reconstruction of the finally intended text, heretofore
denied the author and his readers by contingent factors no longer obtain-
ing, allows for the study of the recoverable life of this work through its
development. Careful review of the Northwestern-Newberry Typee should
give pause to those who profess that author-centered editing is ignorant of
either the "instability" of literary works or their social context.

The editors of Typee had little choice but to accept the early alterations
of Melville's brother and his publisher, since evidence enough survived to
know that they were made, but not enough to identify them. When the
details of an author's interactions with contributors are known to an edi-

tor, criticism becomes unavoidable. Unconsidered endorsements of the
"social text" sometimes assume that all nonauthorial contributions, once
identified, are rejected by an editor taking an author-centered approach.
This assumption proceeds, however, from far too abstract a view of the
authorial focus. Welcome verbal (and pictorial) contributions to literary
works made by persons other than the author have usually been endorsed
by traditional critical editions. The works of Mark Twain offer many
examples of collaborative relationships in which the author was a will-
ing partner. He wanted his books to be illustrated, and invited artists to
depict scenes of their own choosing, and his publishers' agents to select
or create captions for them. Intolerant as he was of unsolicited printing-
house alterations of his manuscripts, he sought textual advice from friends
and family members whose literary talents or ear for coarse language
he valued. Charles Henry Webb assembled and edited the sketches for
The Celebrated Jumping Frog (1867); Bret Harte criticized the manuscript of
The Innocents Abroad (1869), and Olivia Langdon, then the author's fiancée,
read the proofs—thereby beginning a lifetime of service to the literature
of her husband-to-be; Adventures of Huckleberry Finn (1885) received read-
ings from both William Dean Howells and Mrs. Clemens. Where the per-
tinent evidence survives, it usually shows an inclination on Mark Twain's
part to accept the suggestions of his readers. The author was aware that
his vivid use of English was ahead of prevailing notions of politeness, and
while he wanted to test those notions, he did not want far to exceed them.
The nature and limits of the counsel he willingly followed in this area is
explored fairly thoroughly in the volumes of the University of California
edition of Mark Twain's works. The author's relationships with his arbi-
ters of taste could be lively, as an outstanding episode in the history of
A Connecticut Yankee in King Arthur's Court (1889) reveals. The details are
fully discussed in the California scholarly edition, edited by Bernard L.
Stein and his colleagues at the Mark Twain Papers.[62] It was a controversy
which, in the editor's words, reproduced "in miniature one of the crucial
issues in the nineteenth-century debate over realism."[63]

Connecticut Yankee received readings from Mrs. Clemens (manuscript),
W. D. Howells (proof), and, in between, from poet and chronicler of
American literature Edmund Clarence Stedman (typescript). Not quite
an invited contributor, Stedman was foisted upon the author by nervous
first-time publisher Fred Hall—though once arrangements were made,
Mark Twain welcomed his involvement. Stedman professed deep admira-
tion for the novel, but a studied commitment to respectability prevented

him from letting pass an opportunity to warn the author that others might
not understand the book as he did. In keeping with his assignment, he
suggested a few minor changes of a polite nature, which Mark Twain duly
accepted. The author probably accepted most of Howells's suggestions
too, though this is a guess, since the proof sheets that Howells read and
marked are lost—with one spectacular exception. A single sheet from
the novel's conclusion (ch. 43, "The Battle of the Sand-Belt") preserves
printed and holographic evidence of the author's struggle, in November
1889, with one of Howells's queries. In the printed text of the sheet, the
Yankee's "head executive," Clarence, enlightens his boss by imagining the
reaction of the commander of the royalist insurgency after a messenger
reads him the Yankee's order to surrender: "'Disembowel me this animal,
and convey his kidneys to the base-born knave, his master; other answer
I have none!'"[64] Howells found this language too strong, a reaction he
registered by underlining, more or less, the first eight words, and writing
a question mark in the margin. Mark Twain responded by inscribing the
following message on the sheet, which he then sent to his publisher:

Hall returned the sheet inscribed with a note of his own: "Mr. Stedman
says that this is stronger as it is, but that it had better be changed, he
suggests the words 'Disembowel' & 'Kidneys' might offend some." The
sheet also bears Stedman's attempt to revise the offending sentence, which
reveals that he objected also to the word "knave": "Cleft Cleave me this
man in twain, & convey his lights (?) to the the base-born hind, his, etc."
Mark Twain, good to his word, returned the sheet again, now inscribed
with his final note:

The first American edition, which went to press a few days after Mark
Twain wrote this last note, followed the instruction but further substituted
"who sent him" for "his master", possibly without authorization.

The scholarly critical text followed the instructions in the last note ex-
actly, though it would have been within its purposes to preserve the earlier
and more vivid expression, given Mark Twain's strongly stated prefer-
ence for it. On the other hand, the caution governing this decision likely
arose from an awareness of the particular circumstances under which this
work was first produced. Compared to the author of Typee before him or
of Maggie after, Mark Twain was in an unusually powerful position. He
could take or leave requests coming from the publishing house, Charles L.
Webster and Company, because he was its senior partner. As for the
recommendations of his wife or best friend, these were freely asked for,
freely given, and not binding. The author was more or less at liberty to
follow his own artistic lights, but knowing that in places these burned
too brightly for the times, was willing to accept moderating advice from
trusted sources in the matter of potent words and phrases. About Adven-
tures of Huckleberry Finn, for example, he told Charles L. Webster, "I want
Howells to have carte blanche in making corrections."[65] Evidence of the
contributions of Howells, Mrs. Clemens, or others to the texture of Mark
Twain's prose, and the author's reactions to them, has rarely survived in
so dramatic a form as the tell-tale proof sheet from Yankee. Yet whatever
form it takes, surviving evidence of the give and take between author and
invited reader has little caused the editors of the University of California
edition to go back on the author's original decisions.[66] Though today

we are keenly interested in Mark Twain's original language, he, it must
be recognized, had a genuine disinclination to inflame public sentiments
uselessly (as he saw it) by a few casually chosen words and phrases. He
admitted to having "dug out" and "throttled" most of his own "darlings"
from A Connecticut Yankee, but felt nonetheless that Howells's help was
needed to complete the job; this feeling can be (and generally has been)
regarded as part of his artistic intentions.[67] Discovery of the social context
in which Mark Twain worked is one of the purposes of the scholarly edi-
tion; the serving of this purpose rather clarifies than obscures the nature
of the author's genius.

The American experience has confirmed much past editorial wis-
dom, but also demonstrated that in the approach to works represented
by abundant original textual evidence, scholarly attention shifts: as sev-
eral recoverable moments in the life of these works are recognized, the
demand arises for critical texts of greater historical focus, accompanied
by revealing historical apparatuses. The reasonableness of the traditional
editorial emphasis on authors was repeatedly reconfirmed, finding new
definition in authors' visible struggles with publishers over realistic scenes
and language. These struggles also gave new meaning to the editorial
concept of intention (and not only authorial intention). The relevance of
this concept was further strengthened by a deeper understanding of the
problem of textual control—a benefit derived from the ability to study
multiple authoritative documents of the same work. These lessons were
not learned everywhere and at once. Textual history that might have
been understood was missed, and sometimes even deliberately simplified.
The importance of revealing this history was overshadowed in places by
the emphasis on establishing finally intended texts for works then avail-
able only in bad editions, if at all. That the shelves of American libraries

were being filled with reliable texts of the nation's literature—often for
the first time—may have led some editors to regard the critical text as
a metaphysical ideal, rather than a conjectural reconstruction of a par-
ticular historical moment. (Certainly advertising departments of some of
the participating university presses fostered such a viewpoint, with copy
announcing sales of the "only authoritative" texts.)

On the other hand, the historical purposes of scholarly editing did find
many worthy demonstrations. In a classic appreciation of those purposes,
Jo Ann Boydston, then editor of the John Dewey edition, spoke of the
essential importance of the critical apparatus: "Every apparatus has a
story to tell that is, like that of Dewey's Psychology, a story of suspense and
discovery, a true textual drama."[68] Her edition of the work she names was
purposefully designed to reveal a complex history of authorial revision.
The result, which appeared in 1967, fulfills the promise of critical edit-
ing: it presents a critical text of the work at the moment in 1891 when the
author submitted final revisions to his publisher, and displays the other
recoverable moments in the accompanying apparatus. Together, critical
text and apparatus recount the motion of Dewey's early thought, as he
began to modify a youthful Hegelian outlook. Far from obliterating dis-
tinctions between the moments of literary history—in the service, as de-
tractors would have it, of an idealistic abstraction—the well-constructed
critical edition brings that history into view. When the editors are skilled,
the history is presented truly and in a manner conducive of understand-
ing; at this point, if the history remains obscure in places, it is owing to
the limits of the preserved evidence, and the elusive relationship between
thought and its records.