University of Virginia Library

Search this document 


collapse section 
Historicism and Critical Editing by G. Thomas Tanselle
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 


Page 1

Historicism and Critical Editing
G. Thomas Tanselle

All scholarly editors must decide to what extent the texts they present in their editions can be permitted to depart from the documentary texts that have come down to them; no more basic theme runs through the history of scholarly editing than the perennial debates over the role of editorial judgment (necessarily subjective, to a greater or lesser degree) in the production of responsible texts. Even those who acknowledge the value, under certain circumstances, of critical texts (that is, texts incorporating the results of critical judgments as to whether alterations are required) sometimes wish to restrict as much as possible the operation of individual judgment. They may say, for instance, that emendations should be limited to the correction of what are thought to be printers' errors in a given text and should not be drawn from the variant readings of other texts, which represent different stages in the history of the work. Critical editing by definition moves one away from documentary texts, because it admits the possibility of emending those texts. This process need not be unhistorical, for the scholarly goal of emendation is to recreate texts that once existed, even if in some details they existed only in their authors' minds. But the fact remains that critical texts (if emendations have actually been made in them) do depart from the particular texts that have survived from the past; and any recreation of something that does not exist is conjectural and inevitably reflects, to some degree, later attitudes. These issues—which, taken together, might be called the question of historicism—have been discussed at length by generations of editors, and they will always be discussed. It comes as no surprise, therefore, to recognize that they have been prominent in editorial debate in the early 1980s; but they have, I think, been approached in these years from some neglected directions, offering new twists to old dilemmas.

This trend in editorial theorizing is not unrelated to what has been happening in scholarly literary criticism. Although recognition of the interdependence of textual scholarship and literary criticism has not advanced as far as one could wish, there is no doubt that recent writers


Page 2
have increasingly explored the connection. Editors have always known implicitly that any actions they took as editors reflected particular assumptions about the nature of literature and of verbal communication; but over the years they have not been inclined to confront this fact very explicitly. An age of criticism that has emphasized theory, however, has naturally provided a setting in which editorial discussion becomes more self-conscious regarding the theory of literature underlying it. That editors must be critics and that critics must understand textual history are truisms just beginning to be understood beyond a small circle of scholar-critics. The elements of a new historicism emerging in literary study have recently been usefully surveyed by Herbert Lindenberger,[1] who contrasts the "suspiciousness and self-conscious playfulness" of the new history with the "detachment and self-effacement" of the old, pre-New Criticism, variety. One of the reasons for this shift of tone, he suggests, is directly connected with textual scholarship: the theoretical questioning of the organic unity of individual works has been supported by some of the evidence produced by editors, evidence showing (as he says the Cornell edition of The Prelude of 1798-99 shows) that a literary work "consists essentially of layers of text—often, in fact, unfinished layers—none of which necessarily commands more authority than the others." This line of thinking leads to suspicion of "authorial authority" (p. 17)[2] and in turn to rejection of "objectivity" and "permanence" as attributes of historical scholarship (p. 22). Another link between critical and textual work is the recent critical interest in reading (that is, in readers' "responses"), which encourages a concern with the texts available to readers in the past and the reception accorded them (p. 20). Although Lindenberger's account does not emphasize textual matters, it does illustrate some of the ways in which developments in critical and in editorial theory and practice have begun to feed each other.

The new historicism in textual matters that I wish to examine has a somewhat different emphasis, however. After all, the "new history" Lindenberger describes is new in part because it recognizes the importance of historical context in literary analysis and comes after a period in which historical considerations were slighted. But textual study has always been, and is in conception, historical. The recent concern with


Page 3
historicism among editorial theorists does not result from a rediscovery of the value of historical research (which was never lost) but from new approaches to the nature of literature that dictate new limitations on an editor's freedom to be eclectic. In the pages that follow, I propose to examine the literature of editorial theory of the late 1970s and the first half of the 1980s from this point of view. This report is conceived as a continuation of my previous surveys of recent editorial discussion.[3] That its focus is on historicism reflects the way the field has developed, for the most significant discussions of the last five or six years can profitably be examined in terms of their stance on this issue. I begin with some discussions (largely by historians) that cannot be considered to have advanced editorial thinking but that are representative of an unsophisticated attitude toward historicism still often encountered. I shall then turn to the two extreme positions that define the recent debates: the view that literature is social and collaborative in nature and therefore that the historical forms in which a work was presented to the public are of primary significance; and, at the other end of the spectrum, the view that literary works are the products of discrete private acts of creation and therefore that their essential forms do not include alterations by


Page 4
others nor even later revisions by the authors themselves. Finally I shall look at some efforts to assert the validity of multiple texts, recognizing that different historical interests may require different approaches to editing.


What I am calling a debate about historicism has chiefly engaged the editors of writings by literary figures, not the editors of statesmen's papers and other historical documents. The reason is that the debate is primarily concerned with critical editions, in which the principles underlying an editor's emendations determine how far the critical text departs from the documentary text[4] that served as its basis. Historians[5] have not generally dealt with these issues because the material they have typically edited consists of letters, journals, and other similar manuscripts, which are more likely to call for literal transcription than critical emendation. For them, the issue of historicism in editing is apt to be whether eclectic texts (products of critical editing) can ever be preferred to diplomatic transcriptions of single documentary texts. Even though many historical editors have practiced critical editing in the sense that they have normalized or regularized certain features of their texts, and have not simply produced diplomatic transcriptions, many of them have not been able to see the value of the further step that literary editors have often taken when dealing with multiple texts of a single work, the step of emending one text with variants from another. Not having progressed beyond this elementary stage in the process of thinking about editing, they have not been in a position to enter into the more sophisticated discussions of historicism in critical editing. It is an unfortunate fact that what historians have published on the subject of editing has not contributed to the development of editorial theory.[6]


Page 5

Just prior to the period under review here, in September of 1978, a Conference on Literary and Historical Editing was held at the University of Kansas, under the joint sponsorship of the National Endowment for the Humanities and the National Historical Publications and Records Commission. That such a conference took place was encouraging, for it was the first organized effort to open the lines of communication between literary and historical editors.[7] But the title of the conference, like the title of the 1981 volume collecting some of the conference papers, Literary & Historical Editing (edited by George L. Vogt and John Bush Jones), was misguided in suggesting that the nature of editing shifts at disciplinary boundaries (assuming they can be located). It is perhaps permissible, if not very felicitous, to speak—in convenient short-hand—of "literary editors" and "historical editors," when referring to the editors who deal with the writings of literary[8] and historical figures; but it is surely illogical to speak of "literary editing" and "historical editing," as if differences in editing arise more from the subjects involved than from the kinds of materials. Letters pose similar problems, if they are from the same period and country, whether they are written by statesmen or by novelists;[9] and works published in a series of editions pose a different set of problems, regardless of whether the author is a politician or a poet. I endeavored to make this point in my contribution: although it was entitled "Literary Editing" (pp. 35-56) because that was the assigned topic, it explained that the real distinction is between writings of the kind normally intended for publication and those of the kind normally not so intended (critical editions often being most appropriate for the former, diplomatic transcriptions for the latter). Blurring this distinction in the title and organization of the conference was effectively to discourage participants from recognizing the extent of their shared concerns.

The paper that represented "historical editing"—at the conference and in the volume—set forth a viewpoint, largely through implication,


Page 6
that has not yet died out from its own illogic and therefore must be glanced at here. George C. Rogers, Jr., entitled his paper "The Sacred Text: An Improbable Dream" (pp. 23-33) and seemed by the title to be implying that too much attention can be paid to texts, an interpretation borne out by his reaction to an essay of mine: "The text, the text, it is always the text!" (p. 33). One gathers that he believes textual details to be less important for historical than for literary scholars, for he seems to think that the former deal with ideas (and are presumably above such minutiae), whereas the latter deal with language. This astonishing position (which, I hasten to add, Rogers is not alone in holding) is apparently what underlies his statement that the "work to preserve the words of the founding fathers . . . provides us with an understanding of our republic," whereas "the work to find out what Shakespeare himself had to say" provides us "with an understanding of our language" (p. 27). On this basis, presumably, he can report with approval that the texts in the Papers of Henry Laurens incorporate silent alterations to increase "readability"—dashes, for example, are deleted "unless it is obvious that they should be retained as they would be in modern writing," and commas are added "only when the editors are sure that the addition will clarify the meaning of a passage" (p. 29). More important to him than offering a record of such alterations is the provision of historical annotation, for a textual apparatus "tends to confine thinking to the text at hand" (mere language, that is), but annotation "tends to release thinking in a thousand new directions" (p. 31). The connection between nuances of language (including punctuation) and nuances of thought is not made, and thus there is no recognition of the fact (which follows from it) that textual details are of equal importance to all who wish to read with the fullest understanding, regardless of the nature of the writings to be read. Much of Rogers's paper—after an introductory section explaining incorrectly what literary editors do[10] —sets forth the practices of the


Page 7
Laurens edition, as an example of what historical editors do. But nothing in that account explains how the "work to preserve the words of the founding fathers" (or other historical figures) is furthered by concealing certain details of the manuscripts (and depriving readers of the opportunity to arrive at their own evaluations of the significance of those details);[11] nor is any coherent rationale offered for preferring annotation, however stimulating, to information about the words and punctuation of the text itself, the text presumably being the reason for the existence of the edition.[12]

Rogers's Kansas paper makes no contribution to editorial thinking nor—unfortunately, given the occasion for which it was prepared—to the promotion of mutual understanding among editors in different fields.[13] Resuscitating it here is no doubt unkind; but its essential position, however ineptly set out in this instance, continues to be argued. Two years later, at the Williamsburg conference of the Association for Documentary Editing (which had been founded a few weeks after the Kansas conference, with the same goals), Robert J. Taylor still found alterations for the sake of readability to take precedence over the presentation


Page 8
of literal transcriptions.[14] He claims that the inclusion in the edited text of "inconsequential" authorial deletions, "incomprehensible" authorial punctuation, and "superfluous" authorial dashes "could well annoy a modern reader" and that "reader annoyance itself could block the reader from sensing a writer's mood" (p. 5). It does not seem to have occurred to him that any serious reader will be more annoyed by an editor's officiousness in withholding documentary evidence and will find incredible the idea that the "burden of proof" should be on those who introduce no alterations rather than on those who do.[15] Curiously, Taylor presents an excellent statement explaining why anyone who is bothered by unfamiliar or inconsistent spelling and punctuation reveals thereby "an unhistorical attitude"—for he does not see that this point demolishes his own argument. He proceeds to say that "slavish copying" can sometimes "get in the way of the meaning of the words and the spirit of the document" and that the "main objective" is "the illumination of history" (p. 6). We thus come back to the same basic misunderstanding that was present in Rogers's paper, but Taylor is more explicit: "the aesthetic interest is central in the study of literary documents," whereas the "overriding concern" in the study of historical documents is "their contribution to the understanding of history" (p. 6); therefore "the principle of the sanctity of the text" is "not necessary for many, perhaps most, of the documents that an historical editor works with" (p. 7). Aesthetics has nothing to do with the issue, of course; what is being missed here is the simple fact that a careful reading of any piece of writing involves attention to details of wording and punctuation, whether or not the writer is generally considered to be an effective user of the language.[16]


Page 9

Those inclined to agree with the views expressed by Rogers and Taylor often make a further point, which in fact renders their brand of historicism rather paradoxical. They are likely to disapprove of eclectic texts on the grounds—as Taylor puts it—that such texts have "no historical validity," not having "a real existence" and not representing "what was" (p. 7). Wayne Cutler has concisely stated this position by saying that "the historical editor speaks only for one document at a time"; "conflation," he says, "breaks down the time factor that is so important in linking written witnesses to particular past events."[17] These editors therefore put themselves in the peculiar position of saying that one loses the evidentiary value of individual documents by any conflation of the texts of two or more of them but that certain kinds of editorial alteration within the texts of single documents are permissible, and indeed can even assist readers in seeing the historical significance of those documents. An additional irony is that the editors who produce eclectic texts generally provide records of variants and emendations (thus recognizing the importance of documentary evidence), whereas those who favor individual documentary texts often (especially in the field of history) furnish no detailed records of their normalizations (thus suggesting a less rigorous concern with such evidence). But the issue should not be how important documentary evidence is: obviously it is fundamental, whether or not one decides to take the next step and make critical use of that evidence. Some editors who do not wish to take that step, however, are not willing to think about its potential usefulness. The result is the sad spectacle of scholars asserting—sometimes with a touch of pride—their own closed-mindedness. Cutler unfortunately serves as an example:

To what uses literary critics may put bastard documents is for them to say, but the saying of the same will not likely change the historical discipline's rules of evidence and citation. I am far from being convinced that a common definition of terms would inform our dissimilar approaches to editing, for it may well be the case that on the subject of methodology we have little of consequence to exchange. (p. 9)
It would be a great misfortune if editorial discussion were to stall for long at this level.

Yet attempts to deal with the supposed differences between "historical"


Page 10
and "literary" editing persist in getting off the track. An egregious instance is Claire Badaracco's proposal for a paper for the first conference (1981) of the Society for Textual Scholarship.[18] One can readily concur in her dissatisfaction with the use of the words "historical" and "literary" to designate two kinds of editing, but it is hard to see what is gained by her substitution of "documentary" and "textual," based on "principles emerging from one's philosophical stance in relation to the problem of VALUE" (p. 43). When she explains that "textual" editors[19] value "the author's intention," whereas for "documentary" editors "it is not the text but the document itself which is of the greatest value" (p. 42), she is merely perpetuating a misguided approach, adding to it some new confusions. Her piece would not be worth mentioning except that it elicited from Fredson Bowers a splendid reply, which in memorable fashion cuts through to the heart of the whole question and says what needs to be said. Naturally the reporting of evidence is central. Bowers concisely makes the essential criticism of historical editors' common practice of omitting any record of the authorial deletions and revisions present in the texts of the documents being transcribed: "All one can ask is, Is this documentary?" (p. 65).[20] These editors, he notes, have repeatedly "turned a blind eye to the superior possibilities for the transmission of information that have come to characterize the new school of editing making its way in the humanities" (p. 49). The emphasis is on the "possibilities for the transmission of information," not on the nature of the edited text, since "for the purposes of historical interpreters it may be moot whether an eclectic conflated text made up from multiple authorities is better suited than a transcript of a single document, provided in both cases an apparatus records the variants" (p. 66). Critical editors, Bowers rightly insists, place just as much value on documentary evidence as diplomatic editors do, but the kind of edition they generally construct, containing both a critical text and an apparatus, meets the varying interests of different audiences and releases the editor from being "the victim of the requirements of only one segment of an audience"


Page 11
(p. 73). Bowers's essay—as this brief summary of a few key points suggests—deserves a wide readership among those who have had suspicions about the scholarly seriousness of critical editors in the field of literature and who have not been able to see that all who deal with texts confront the same problems. The observations he makes are in fact self-evident, as he several times suggests; but past debate does not give one grounds for hope that they will be soon understood, in spite of his effective statement to ease the way.

Even David Hall, who is particularly interested in the history of books and reading, found it possible in his 1983 Wiggins Lecture to refer sarcastically to "the work of analytical bibliographers and their holy of holies, the text."[21] The depth of his misunderstanding is revealed by his further saying, "The very concept of a perfect text is an invention of the twentieth century, and cannot be imposed upon the past" (p. 335).[22] Students of the history of reading and of the role of books in society are rightly interested in the texts available to readers at particular times in the past; but so are students of the history of literature, and no critical editor of a literary text would pretend that a newly constructed critical text (as opposed to its apparatus) would be appropriate for analyzing earlier readers' reactions.[23] Whether historians are in fact as concerned with past texts as they ought to be is a question one cannot avoid raising, if Rogers's exasperated exclamation "always the text!" and Hall's slighting reference to "holy of holies" are at all representative of a common feeling. The truth of the matter is that, because analytical bibliography developed primarily among literary scholars, many historians have not yet come to understand the lessons it has taught about the role of physical evidence in uncovering textual problems (lessons relevant to the study of manuscripts as well as of printed books) and therefore have not recognized that the task of identifying "the text" read at a given time is often more complicated than the simple location of a single copy. Indeed, the growing numbers of historians interested in what is often referred to


Page 12
(following the lead of the French in this field) as histoire du livre, dealing with books in their broadest social contexts, have surprisingly often failed to see how important analytical bibliography and textual study are for their endeavors.[24]

Despite extensive discussion of these matters in recent years, encouraged in part by the activities of two organizations devoted to fostering interdisciplinary communication among scholarly editors, the split between literary and historical editors regarding the responsible handling of historical evidence has not grown significantly smaller. That so much energy has been invested in debating such elementary—such essentially undebatable—points is regrettable; there are, after all, real issues waiting to be further explored. No one doubts the importance of making transcriptions or reproductions of the texts of certain individual documents (both manuscript and printed); and it seems scarcely credible that anyone would question the desirability, in connection with such transcriptions or reproductions, of reporting as much as possible of the textual evidence those documents contain.[25] Similarly undebatable, one would think, is the idea that in certain instances a further usefulness might result from the production of a text embodying alterations made at the editor's discretion[26] (with the alterations recorded).[27] Both literary and historical editors acknowledge this point, but some do so only in a limited way. Many editors who disapprove of eclectic texts nevertheless produce critical texts, for they make certain kinds of alterations, aimed at bringing a text to what in their judgment is a higher standard (whether of readability, mechanical correctness, correspondence to the author's intention, or something else). Any editor who normalizes or modernizes


Page 13
a documentary text is obviously engaging in critical editing, for the resulting text departs from all the historical witnesses through the operation of the editor's critical judgment. Some of these editors balk at the idea of drawing any of their alterations from another text of the same piece of writing, labeling such a practice "eclectic" and charging that it destroys the integrity of individual documents. But that integrity has already been violated by the editor's own intrusions; "eclecticism" only alludes to a particular source of such violation. No one would argue that editors have any obligation to produce eclectic texts when they find such texts inappropriate; but surely editors who understand the usefulness of even one kind of departure from absolute fidelity to a documentary text can also conceive of the usefulness, under some circumstances, of such eclectic texts. Whatever the field, scholars must engage in interpretation of the raw materials of history, and eclectic texts are one product of such interpretation.[28] Literary scholars may have more occasions for producing them than scholars in other fields; but it seems inconceivable that any scholar can fail to comprehend the rationale for and function of such texts. Yet that is precisely what much of the argument has been about. We are not talking here about which materials are most appropriate for eclectic treatment or what the principles for emendation ought to be but simply whether eclectic texts can ever be justified as historical scholarship. Clearly they can be: critical (including eclectic) texts have a place in the scholar's repertory as surely as diplomatic texts do. And in either case the scholar has an obligation to report the details of the documentary evidence. On this level, there is nothing to debate.


The more interesting and potentially fruitful discussions of textual theory begin at the point where these unsophisticated complaints of historical


Page 14
editors leave off. They start from an acceptance of the value of eclectic texts and consider the problem of historicism from a higher plane. The issue is not whether eclectic texts as a genre are defensible but how best to produce them: when to be eclectic, how much departure from a documentary form of a work is allowable and desirable, whether editors should introduce emendations of their own in addition to readings drawn from other texts, and what principles or standards should underlie alterations of either kind. Discussions of these matters are usually concerned with works of the kind normally intended for publication (usually, in fact, with works actually published, after the fashion of their times), for critical editing is more likely to be of service in connection with such works than with private writings. How one conceives of the nature of such works and what concept of authorship it entails are therefore crucial questions, the answers to which determine the goal one is aiming toward in making alterations to a documentary text of a work.

One family of answers to these questions proceeds outward from the author to the author's social context, tending to make authorship more a social than a private activity and sometimes expanding the concept of text. In some of his recent work, D. F. McKenzie has been moving in this direction, calling for a "sociology of the text." For instance, the paper he presented at a 1977 Wolfenbüttel symposium (the proceedings of which were published in 1981)[29] deals with "Typography and Meaning" and argues, as one illustration of the connection, that Congreve's altered treatment of scene divisions and stage directions in his collected Works of 1710 was inextricably tied up with their typographic presentation, which resulted from "a new and intimate form of teamwork" (p. 110) between Congreve and his publisher (and friend) Jacob Tonson. McKenzie is saying, in other words, that in this instance features of a printed book that are often regarded as nontextual cannot in fact be separated from the words and punctuation in considering the author's textual intentions. Congreve is a particularly interesting case, for the contrast between the original quarto editions of his plays and the Works of 1710 marks the transition between the seventeenth-century "disjunction of playwrights and printers" (p. 82), which resulted in printed forms "insensitive to the problems of mediating a theatrical experience" (p. 83), and the eighteenth-century effort to give "typography a voice in the hand-held theatre of the book," which made the printed form more than a


Page 15
makeshift report of something that had its real existence elsewhere. Despite the special circumstances of the Congreve example, McKenzie intends it to be emblematic of a broader point, not limited to Congreve, the eighteenth century, or printed drama: that the book itself is "an expressive means," in which "the substance of the text" cannot be divorced from "the physical form of its presentation," for the book conveys "an aggregation of meanings both verbal and typographic" (p. 82). As he concisely puts it, "A book's total form is itself a significant historical statement" (p. 99). In support of this generalization he offers a wide-ranging survey, filled with characteristically acute observations.[30]

No one, I think, would dispute the view that every detail of a printed book carries historical meaning, though few critics have adequately integrated the evidence offered by format, paper, type design, and page layout into their readings of works, and McKenzie is quite right to stress the seriousness of this failing.[31] But we can all agree that readers' responses are affected by typography and book design without feeling that such features of books are necessarily inseparable from the works conveyed by the books. The issue turns on whether one is willing to admit the legitimacy of being interested in the artistic intentions of authors as private individuals rather than as social beings accommodating their intentions to various pressures emerging from the publishing process. On this fundamental question McKenzie wavers, and his imprecision weakens the foundation of an essay that is admirable in so many ways. A distinction has to be made between the books that were available to be read and reacted to at particular times in the past and the forms of works as intended by their authors. If one wishes to reconstruct how earlier readers reacted or to analyze their written reactions, there is no doubt that the physical features of the books they read are relevant and can in that


Page 16
sense be considered part of the "texts." But if one is more concerned with assessing a writer's mind and ideas than in examining how they were perceived by the contemporary reading public, some or all of the elements of book design may be nontextual, if the writer was not interested in them or did not put any of them to special use. Even in these cases, a scholar should not ignore such features of original printings, for they are part of the historical setting;[32] but the physical arrangement of the scholar's critical text need not be affected. Situations will vary: there are writers, as Congreve apparently was, who are so intimately involved with the design of the printed presentation of their works that the design (or some parts of it) must be regarded as textual; for other writers, the design does not reflect authorial intention. Critical editors must judge which is true in each case, just as they must make judgments about words and punctuation.

McKenzie's essay suffers from not being built on this distinction.[33] His insistence on the necessity of eclectic texts, for example, is confusing in the context of a concern with the "sociology" of the text. In discussing Congreve's revisions for the 1710 Works, he says that an editor "must seek to serve the play at its fullest and best[34] by restoring a reading when he believes its suppression to reflect a narrow moral, rather than a literary, judgement on Congreve's part": "Conflation is inevitable" (p. 109). Although his defense of eclectic editing is well stated, it is likely to leave readers puzzled. Congreve's self-censorship, he believes, resulted sufficiently from "external pressures" (p. 107)—legal constraints on coarse language and Tonson's attitude toward it—to justify an editorial decision to restore the canceled language; but such a decision emphasizes the author's own wishes over the product that emerged from the social process of publishing. Of course McKenzie is right to assert that responsible textual decisions must be based on an understanding of the "complex of attitudes—personal, social and trade—" (p. 109) that lie behind variant readings; presumably he is referring to the same broad range of considerations later when he says that variants should be "interpreted in the context of book trade history" (p. 117). Editors must naturally be as informed as possible about all aspects of the historical context; but


Page 17
being so informed does not determine what view of authorship they should take as a basis for textual decisions. For in many—no doubt in most—instances one cannot accommodate both the private wishes of the author and the collaborative product of the publishing process simultaneously.[35] If McKenzie can speak of interpreting variants in the light of book-trade history and still opt in the case of Congreve (who worked closely with his publisher) for restoring readings that the author deleted under the publisher's influence, then he is not claiming that one must necessarily give preference to the historical product of the book trade (the "historical statement" conveyed by a "book's total form").[36] He is saying only that editors' knowledge should include book-trade and typographical matters, and he does not go on to confront the fact that more than one responsible approach to critical editing can be followed by editors who have this knowledge.

What he recommends for Congreve is the same, in the end, as what many of the editors who are considered followers of Greg would have done in this situation. All McKenzie is really saying is that editors have frequently had insufficient historical knowledge to recognize the textual role that typography can play. It may be fair to call this lack a "failure of historical imagination," but it is an overstatement to assert that "most recent work in textual bibliography" (p. 105) is guilty of it, for editors routinely consider which (if any) typographical features of the printings of an author's work must be defined as textual; and a decision not to classify them as textual, in an edition focusing on authorial intention, does not necessarily signify that the editor has failed to take the whole book, or the whole historical context, into account. McKenzie's misunderstanding of Greg's rationale[37] has led him to think that "current


Page 18
theories of textual criticism" are "quite inadequate to cope" with the connections between typography and meaning; actually there is no theoretical problem on this score at all, and the value of McKenzie's piece is rather to have provided us with some illuminating examples of these connections and to have reminded us forcefully how essential a knowledge of typographic and book-trade history is to editors (not to mention other things they must know). He makes no case in this piece for the necessity of what he grandiosely calls "a new and comprehensive sociology of the text" (p. 118). But in his impressive Presidential address to the Bibliographical Society five years later, he more calmly and more stimulatingly defines the "sociology of texts," not in relation to a supposed crisis that it can rescue us from but as the "substance of bibliography" (p. 365), thus expanding bibliography to include the study of orality, literacy, and "the recording function of memory" (p. 333).[38] His encompassing view enables him to offer one of the most eloquent testimonies we have to the indispensability of textual eclecticism in reaching some understanding of the past, showing that it in no way violates "historicity" (p. 334).[39] Even those bibliographers who feel uncomfortable with the idea of becoming anthropologists (as well as sociologists) will come away from McKenzie's humane manifesto with an enriched understanding of their role as textual historian-critics.[40]


Page 19

Although McKenzie's position emphasizes the social settings in which authors work, it does not deny the primacy of unconstrained authorial intention as a guide for critical editing. Others who focus on the social side of authorship go farther and believe that the collaborative nature of the publishing process makes artificial any attempt to isolate an author's uninfluenced intentions. Jerome J. McGann has become the most prominent advocate of this point of view, particularly as a result of A Critique of Modern Textual Criticism (1983).[41] He attacks the approach developed by Bowers (and derived from Greg) for editing works of the last two centuries because it emphasizes authorial intention; he believes that it has "tended to suffocate textual studies" by limiting them to a narrow "psychological and biographical context" (pp. 119-120). In its place he calls for an editorial theory that would recognize literary works to be "fundamentally social rather than personal" productions (p. 8). Locating authority in authorial intentions, he says, causes works


Page 20
to be seen in "the most personal and individual way," and "the identity of the author with respect to the work is critically simplified through this process of individualization":
The result is that the dynamic social relations which always exist in literary production—the dialectic between the historically located individual author and the historically developing institutions of literary production—tends to become obscured in criticism. Authors lose their lives as they gain such critical identities, and their works suffer a similar fate by being divorced from the social relationships which gave them their lives (including their "textual" lives) in the first place, and which sustain them through their future life in society. (p. 81)
This passage sets forth, as effectively as any in the book, the view that "words do not by themselves constitute a system of communication" and that "literary works are not produced without arrangements of some sort." Thus, for him, the very existence of works (and not merely their publication) depends on collaborative effort: "the authority for the value of literary productions does not rest in the author's hands alone" (pp. 47-48).

McGann's book serves a useful purpose in asserting the importance of seeing one's own scholarly endeavors against the background of the historical evolution of the field and in focusing renewed attention on the social context of literary production. The treatment of the latter is disappointing, however; the book does not achieve its aim of developing "a fully elaborated argument for a socialized concept of authorship and textual authority" (p. 8). Such a book would be valuable, for—despite increased interest in the historical study of the profession of authorship—scarcely any careful and thoughtful analysis has been made of the implications, for textual criticism, of the social structure of authorship. McGann does provide several interesting examples, often drawn from the writings of English Romantic poets, to illustrate some of the ways in which works become collaborative enterprises, and he offers from time to time variations on his fundamental observation that "an author's work possesses autonomy only when it remains an unheard melody" (p. 51); but he expends much of his energy on a criticism of the position of Bowers and those who, in one way or another, have followed his lead. Even if this criticism were well-founded,[42] it is hard to see how the discrediting


Page 21
of one approach amounts to a "fully elaborated argument" for another. The underlying assumption, of course, is that there are only two alternatives and that only one of them is valid. A far more productive way to proceed would be to recognize that a variety of approaches is justifiable and then to concentrate, with positive arguments, on the merits of the one under discussion.

McGann obviously believes that the structure of modern publishing does not admit of any approach to authorship that stresses individual artistic creation and that to take such an approach would be a falsification of history. But surely it is a legitimate (and natural) historical pursuit to be interested in the minds of particular persons from the past, particularly if their writings (or other accomplishments) have any reason to command our attention. However much those writings as published and read were a collaborative effort, we are not being unhistorical in wanting to know just what the initiating mind contributed to that effort. The initiator, by virtue of being the initiator, is forever set apart from those who follow, however necessary they may have been for the completion of the act of communication (and, indeed, however beneficial we judge their ministrations to have been); and if we are concerned with more than one work "by" the same person because we feel that they may illuminate one another, the creating mind is the link between them. The attempt to establish what an author thought and wrote when not making concessions to pressure from others is an essential activity for understanding history.[43] But those who engage in it are not thereby


Page 22
denying the value of examining as well the forms of works that came off the presses and went into readers' hands. No one can reasonably claim (and I am not aware that anyone has tried to claim) that the texts of works as they in fact appeared in successive printings and editions over the years are not important for historical study or that publication is not a social process. The opposite in each case is a truism, and McGann in these respects is stating the obvious (which is not a bad thing to do). But it is equally a truism to say that intellectual (and thus literary) history is concerned with the private as well as the public, with the minds and ideas of individuals as well as with the transmutations of those ideas in their passage through the world. The two approaches are complementary and both are necessary, though one may be more appropriate than the other for certain purposes.

These observations, I believe, provide the proper context for reading McGann's book; with them in mind, one immediately recognizes the fallaciousness or narrowness of many of his statements. It is ironic, given his emphasis on breadth and on the need to free textual study from constricting ideas, that his own position is narrow-minded in limiting the acceptable approaches to a single one. Studying literature as a social product is only superficially more inclusive than studying it as the product of a single creator, since one must explore all the same areas of concern in either case; but one is clearly taking a restricted view if one is not open to the values of both approaches. There is nothing wrong, of course, with being an advocate of one position, so long as one is not blinded to the contribution of the other. Insofar as McGann is committed to denying the usefulness of an author-centered conception of editing, his argument is doomed to failure; and his advocacy of a society-oriented one is weakened by his lack of balance.

It also suffers from a lack of clarity. The role of authorial intention is central to the whole discussion, and yet at the end one does not quite know what McGann is suggesting about its place in editing. In an appendix called "A Possible Objection," he reports that a reader of his manuscript objected to his generalization that a first edition "can be expected to contain what author and publishing institution together worked to put before the public" (p. 125), by citing instances in which D. H. Lawrence was "an unwilling partner in a downright repressive process." One might well wonder how this is a "possible objection" to


Page 23
McGann's general line, for if a work achieves its existence through the interaction of author and publisher, then an expurgated work must be accepted as readily as one we might regard as improved, for both can be the result of the particular chain of historical events that constitute the publishing process.[44] But McGann surprisingly replies that the collaboration of author and publisher does not always turn out well and that "the editor must examine carefully the early publishing history in order to arrive at a reasonable decision" regarding the choice of a "textual version" (p. 127).[45] He adds that authorial intention is "only one of many factors to be taken into account, and while in some cases it may and will determine the final decision, in many others it cannot and must not be forced to perform that function" (p. 128). Similarly, he had earlier stated that "Shelley's manuscripts frequently assert a strong demand to be adopted as copy-text, whereas Byron's rarely do": despite the fact that Shelley "published in a fashion that was normal for his period," it would be "a disservice to Shelley's work . . . if a critical edition today neglected to consider, in the matter of copy-text, the sincerity and integrity of Shelley's manuscripts" (pp. 108-109). If McGann's point is that we should accept the results of the publishing process only when we feel that the work has benefited from it, we are then being asked to engage in a very different kind of editing, in which our choices are dictated by our personal preferences; the results may be admirable artistically but are not designed to increase readers' understanding of the past.[46] If he is saying that we are to follow a published text only when (or at those points where) the author can be thought to have sincerely approved of and desired it, then the focus is on authorial intention after all, and the procedure is not really different from the one that editors in the Bowers tradition have been using. They have habitually made judgments to distinguish which revisions made or suggested by others (whether publishers or other acquaintances) were fully accepted by the


Page 24
author in the spirit of active and welcome collaboration, and they have rejected only those revisions that the author appears to have accepted grudgingly or been forced into accepting. These editors have normally recognized the social side of authorship by acknowledging that an author's intentions can sometimes include the results of collaboration.

Of course these matters require subjective judgments, based on historical evidence, but so do most decisions in critical editing. In the Bowers tradition, however, those judgments are made in a clear conceptual framework, with authorial intention at its center; but McGann offers no equally clear alternative, for he seems to waver in regard to what is at the center. It is not very helpful to be told that authorial intention is one of "many factors" to be taken into account and that sometimes it will be the "determining one": when the author's intentions and the publisher's actions are in conflict, what does one do? Naturally, one can go either way, but these are two different approaches to editing. McGann believes that "to see 'author's intentions' as the basis for a 'rationale of copy-text' is to confuse the issues involved" (p. 128); one should rather say that confusion is promoted by maintaining that an undefined mixture of two distinct approaches constitutes a useful rationale.[47] Clear thinking is better served by recognizing at the outset that individual desires and social pressures, though they may in specific instances be in harmony, are conceptually irreconcilable and that an editor's guiding principle in textual decisions must favor one or the other. It is understandable that many editors have regarded authorial intention as the more sensible choice for a scholarly critical edition, since one can argue that to represent the historical results of the publishing process a diplomatic edition or photographic facsimile would be preferable.


Page 25
But I do not mean to suggest that there are no possibilities in between: one could decide, for example, to accept all the publisher's alterations but to correct typographical errors, thus producing a critical text representing the publisher's intention; or one could decide to remove from the text of a single printing the textual features imposed on it against the author's wishes, thus producing a critical text reflecting the author's intentions at that particular time (regardless of any later authorial revisions). Such alternatives, however, have clear aims, giving priority either to the author's intentions or to outside influences. McGann's account contains no hint of this rational structure of possibilities within which critical judgment can operate purposefully.[48]

The issue here is not subjectivity, for all critical editors—in the Bowers line or any other—make judgments (which are inevitably subjective) at every turn. It is difficult to understand why McGann thinks that "Giving up the rule of final intentions" will "introduce a subjective factor into the critical process" or that his proposal involves "the re-emergence of a 'subjectivity'" (p. 107), as if subjectivity had not always played a central role.[49] Because one of the factors influencing subjective decisions in a given instance is, he believes, the history of the previous editing of the work and what needs the present audience for the work has, it is not surprising that he welcomes modernizing as a legitimate scholarly activity.[50] He accuses those who reject modernizing of not understanding that all editions, even unmodernized ones, are—like all other literary efforts—time-bound, reflecting the concerns and attitudes of the age in which they were produced. All thoughtful people, including many


Page 26
scholarly editors, are perfectly well aware of this fact, but it is irrelevant. If one engages in historical scholarship, one is attempting, through an informed imaginative effort, to escape into the thinking of another time, even though one knows that the escape is never complete and that it will have to be reattempted by others in the future. Scholarly editors of critical editions do not really imagine that they are packaging up textual history for all time or preventing the further alteration of the text (see p. 93). But these realizations do not invalidate the effort of historical reconstruction. Modernized texts, like some kinds of critical essays, are attempts at elucidation, which may be more or less helpful to the readers of a given time; but those readers (even the "nonspecialist" ones)[51] who are interested in a work as testimony from the past will need to have the best results that historical scholarship has achieved in the recovery or imaginative reconstruction of particular texts and versions of that work.[52] McGann considers his discussion of modernized editions to be "the final phase of the argument" (p. 94), and it does epitomize his dual conviction that literary production is "not an autonomous and self-reflexive activity" (p. 100) and that critical editions, like all other works, are "always produced under the pressure of contemporary demands" (p. 96). But it also perpetuates the blurring of essential distinctions that permeates the book. One must deeply regret that a book offering the hope of a systematic


Page 27
exposition of a social view of authorship and its editorial implications must finally be judged to have left the issues more confused than clarified.[53]


A convenient pairing was brought about the following year by the publication of Hershel Parker's Flawed Texts and Verbal Icons: Literary Authority in American Fiction (1984),[54] for that book locates the other pole of the debate over authorial intention and its place in the historical study of texts. Parker takes the position, with a vengeance, that authorial intention is central: he believes that the important intention is what authors manifest during the creation of works and that even authors' own later revisions often have no more right to become part of those works than the alterations initiated by others. His and McGann's books, therefore, serve to define the limits of the area under discussion. What they share is a sense of urgency about the need for a renewed historical orientation in literary studies and a conviction that textual criticism is central to literary criticism. From there on, of course, they are in contrast, and in more ways than one might anticipate. For McGann's book, despite its incoherence, calls attention to a fundamental theoretical issue, whereas Parker's book, though far better written and organized, is principally of interest for its detailed case histories, not because its basic thesis is of theoretical significance.

Parker begins with the assumption that "All valid meaning is authorial meaning" (p. ix);[55] he is concerned not with arguing this general point but only with proceeding from it to a more precise definition of authorial meaning. Drawing especially on John Dewey's Art as Experience (1934) and Albert Rothenberg's The Emerging Goddess (1979), he describes the creative process as one that "begins, continues (as clinical


Page 28
observation records, with varying admixtures and sequences of excitement, arousal, boredom, anxiety, and determination), then ends—ends with stubborn finality" (p. 34). Because this process is "by nature determinate, revising authors very often betray or otherwise blur their original achievements in ways they seldom intend and seldom become aware of" (p. ix). He therefore concludes that Greg's rationale of copy-text is not an appropriate guide to the establishment of texts reflecting their authors' intentions because it assumes that authors' later revisions are, as a general rule, to be accepted in preference to the earlier readings.[56] According to Parker, "all empirical evidence should tell us that Greg is wrong, that in any mature human being, writers included, a state of indefinitely sustained arousal toward one object is unnatural" (p. 35). Thus authors' later revisions—as illustrated by a number of examples, particularly Henry James, Mark Twain, Stephen Crane, and Norman Mailer—may introduce inconsistencies (of detail, characterization, tone, or theme) and produce "maimed" texts. Critical editors, he believes, should not assume that "every author retains full authority over anything he has written for as long as he lives" (p. ix); they should instead reject later revisions that result in "unreadable texts" (p. x).

This argument, as set forth by Parker, is not a helpful or clarifying one. The key to the problem is the fact that even if the creative process is granted to be "determinate," in the sense that it comes to an end, one is not provided by Parker with any historically oriented guideline for defining that ending point. It does not make sense, as he recognizes, to fix any quantitative limit: obviously one cannot say that any revision made within twenty-four hours, or three days, or six weeks, of the writing of the last sentence of a work emerges from the heat of the creative process, and that later revisions do not.[57] Even during what seems a


Page 29
single period of creative intensity, writers going back to an earlier paragraph may not be fully in tune with its context and may make revisions that seem to weaken or confuse the passage. Furthermore, as Parker admits, it is possible for some revisions to appear successful even though they were made at a time when one would assume the creative process to have ended. "No hard and fast limits can be drawn," he says, and he offers a few of the obvious reasons: "some writers have better memories than others" or refresh their memories with more rereadings of their works, and "a great deal always depends on how prolonged and overwhelming the creative process had been and how thoroughly the author is compelled to put it behind him and go on with a new work" (p. 40). The creative process is as various as are the temperaments of creators. Therefore none of Parker's generalizations about late revisions can be unqualified (he says that writers "very often"—but not always—"blur their original achievements"), for some late revisions are not inappropriate. If satisfactory revisions do sometimes occur after the creative process has ended, then clearly some factor other than the heat of sustained creativity is responsible for their success. Alternatively, one could define the creative process so as to include all satisfactory revisions; thus one could say that the existence of such revisions shows that the creative process had not yet ended at the time they were made. Either way, how has the introduction of the idea of a "determinate" creative process been of assistance? The focus of attention has been placed on distinguishing revisions that are satisfactory, or not inappropriate (or however one wishes to describe them), from those that are the opposite; and, it turns out, the concept of the creative process is not the analytical element that enables these distinctions to be made.

What does underlie them is the editor's aesthetic judgment. Deciding whether a revision is successful (or satisfactory, or simply makes sense rather than nonsense) depends on individual judgment. Some revisions that produce clear inconsistencies are indubitably blemishes (when one can be sure that no authorial purpose is served by them), but many of the revisions that Parker discusses can be (and have been) the subject of disagreement as to their worth or effectiveness. The fact that he is asking editors to make literary value judgments is not in itself a problem, for


Page 30
all critical editing requires editorial judgments, which involve literary sensitivity.[58] Nor can one accuse him of simply telling editors to choose whatever variants happen to please them, for he explicitly couches his proposal in historical terms, with its goal the establishment of the author's—not the editor's—intended wording and punctuation. Yet on a more sophisticated level there is a sense in which this accusation is justified. One must remember that Parker is concerned only with revisions that authors made of their own free will; naturally those they were forced into making do not reflect their uninfluenced intention. The essential issue, then, is whether revisions that seem to the editor to introduce inconsistency or incoherence can be considered to represent the author's intention. Normally one is safe in assuming that authors do not intend to make their works incoherent or to make revisions that damage their works. But the kind of intention Parker is rightly concerned with is the intention in the act of writing, the intention to place a certain word or mark of punctuation next, not the intention to produce a particular effect. By definition, then, authors who make revisions of their own free will are producing intended texts (which may of course contain slips), even if those revisions, in the opinion of some or all critics, were executed with insufficient thought and are damaging in their effect. Everyone agrees that authors can make mistakes. But their revisions, however mistaken, are historical facts, and we cannot deny that unsatisfying versions of works are what authors have sometimes left to us as their finally intended texts. Parker believes that in such cases we should be concerned with the earlier intentions that produced better results. In effect, therefore, he is saying that—even as scholarly editors, who inform ourselves with historical facts—we should reject authorial revisions that do not strike us as successful.

This position is not necessarily untenable, but we are not likely to be persuaded of its cogency by an argument enveloped in an irrelevant discussion of the creative process (interesting concept though it is). The more direct way to justify such an approach is to say that, if one is producing a critical text (in which the choice among variants involves the editor's judgment in any case), one should present the work in the most successful of the various forms dictated by the author's shifting intentions. (If the goal is to remain historically oriented, the editor must be


Page 31
limited by what the author wanted at some point, whether at the last stage of revision or not.) One way this decision has previously been approached[59] is in terms of whether a set of revisions can be judged to have turned a work into a different work (or an independent version): if grounds do exist for believing that particular revisions (regardless of their extent) affect a work so profoundly that the result ought to be thought of as a separate work, then one is free, if one chooses, to edit the earlier work rather than the later. This approach gives editors considerable scope for exercising their literary judgment—first, in making the subjective decision that the revisions produce a new work and, second, in valuing the earlier work more highly. But there is a great difference between this line of thought and Parker's, even though both rely on editors' aesthetic judgments. This one places those judgments squarely within a historical framework (admittedly, and necessarily, a subjective historical construct): one can reject later revisions if they appear to represent a new departure and turn a work in a new direction, for the result is being postulated as a historically separate entity requiring independent editing. If, however, one cannot make such a case, one must then consider the revisions as attempts to perfect or complete the work along the general lines already manifested in it, and one cannot reject some of them, simply because they seem misguided, if one's historical aim is the author's last intentions for each work (or aesthetically independent version).[60] Parker's approach, on the other hand, reverses the priorities and places the editor's aesthetic judgments above the historical succession of authorial revisions: those revisions considered by the editor to damage the work are not, in Parker's plan, allowed a place in a critical text meant to show the author's intentions. This way of proceeding involves shifting "final intention" in one of two ways: either the term means artistically final intention (however early or late such intention revealed itself) rather than the chronologically final intention of each independent version, or else it refers to the intended results, thus excluding revisions that do not, as they turned out, promote the intention of improving the work. In either case editors are asked to determine when authors' free-willed revisions were not in the best interests of their works and to construct their critical texts accordingly. It is not uncommon for critical editors to have to decide what authors "really" wanted, as opposed to what they actually did, in situations where their freedom of choice is suspect; but to protect authors from the literary consequences


Page 32
of their unconstrained textual decisions, as Parker recommends, is in fact to remove the author from the center of attention.

The role of subjectivity in Parker's plan is thus fundamentally different from its place in the approach that requires—to justify the absence of late revisions from a critical text—a demonstration that they actually produced a new work (or an independent version, whichever one wishes to call it). Parker's rejection of that approach is characteristic of his reasoning. In discussing the 1907 text of James's The American, he asks whether a work can be "new" when "a good many lines at a stretch, occasionally, are wholly unrevised, while some other revisions respect the structure of a long paragraph while altering it stylistically sentence by sentence" (p. 107)—as if the presence of unrevised passages, or passages with unrevised structure, are incompatible with the existence of a different work. He later ridicules critics who seem to think that an author "confers meaning on a completed text with the wave of a hand or by ripping a book apart and reordering a hunk of it" (p. 218). When "hunks are re-used, unaltered or only slightly altered," he asserts, "what goes unrevised to a greater or lesser extent goes unrethought, unrestructured, carrying its original intentionality in a new context where that intentionality is more or less at war with the different intentionality in the altered or newly written passages" (pp. 228-229). These comments suggest a quantitative test, implying that a new work must be largely constructed of new sentences: "If unaltered or scarcely altered hunks of the original text remain in the later text, that later text is not truly a 'separate' one" (p. 229). Of course unaltered passages do sometimes clash with their new contexts, but at other times they do not. A writer may make only a few crucial changes in a work, leaving the rest unrevised, and yet the import of the whole may be changed; the unaltered passages will have taken on altered significance, and the fact that they are unaltered does not necessarily mean that they are "unrethought." Parker's firm belief that "textual meaning is not something living in a text" as an autonomous entity (p. 219) should be seen to argue against, not for, his view: a passage, not being autonomous, will not necessarily carry its original meaning into a new context. If it seems to fit successfully into the new context, are we to say it has been rethought and that the result is a new work? And if it does not, that it is unrethought and the result not a new work? Parker is really saying that a new work can come only from successful revision, making "new work" an evaluative term referring to a successful work. His argument seems to run as follows: writers infuse meaning into works in the process of composition; moving passages around is not a part of the composition process unless the shifted passages are thoroughly rethought; we can tell that they have been


Page 33
thoroughly rethought if they function well in their new contexts; if they do not function well, we cannot consider the new versions to be new works, regardless of the nature of the revisions that were made, but only botched versions of the original work. Ironically, if he could accept the idea that revisions can produce unsuccessful new works, he would have a stronger argument for offering—what he favors—a critical text containing unrevised readings.

Parker's main thesis is built on a patent inconsistency. He repeatedly insists that the only meaning we should be interested in is what the author put into the text: we must think of the author as "a human being who worked meaning into the text line by line and page by page," for "authorial power is the only literary power there is" (p, 219). Nevertheless we are also told that authors frequently do not know when to stop revising their work and that what some of their revisions produce is nonsense rather than valid meaning: "the creative process, like any other process, has bounds, beyond which no author, however fine a craftsman, is apt to intervene with impunity" (p. 51). There is apparently a higher authority than the author, after all. Because Parker believes that unrevised passages in new contexts, not having been built up word by word anew, exemplify the doomed effort to prolong the creative process and are thus unlikely to succeed, he is scornful of authors and critics who claim—as in the instance of Tender Is the Night—that a rearrangement of sections has resulted in dramatic improvement. "It is," he says, "as if the right order were latent in what Fitzgerald wrote, awaiting only a magic touch to restore it to a platonic ideal which had had no reality during the years of composition" (p. 219). But Parker, too, has his platonic ideal, for he states that "the working assumption of any student of the creative process is that the direction of any developing aesthetic object is toward unity" (p. 217). The creative process has an end, and the work is then at its most unified, though authors do not always realize that they have reached this point and may try to make further improvements. Parker professes to put the author back into literary criticism: he deplores the tendency to see a work as "a verbal icon, a unique, perfect, and essentially authorless entity" (p. ix). But he also attacks the idea "that an author had the right to do whatever he wanted to do to his text" (p. 60),[61] and he believes that allegiance to the unity of the created work (whether or not we call it an "icon") finally takes precedence over a concern with what the author in fact did to the work.


Page 34

Parker surely senses the trap he has set for himself, for at one point he says that we often "revere" the creator and the icon at the wrong times, the creator "after he has gone from creator of his own work to merchandiser or promoter of it" and the icon "not in the form it had when the artist was most in control of it" (p. 49)—a statement that epitomizes all the inconsistencies of the argument, even while it almost acknowledges them. The theoretical scaffolding collapses; but we are left with a series of incisive discussions of individual works, showing how a knowledge of their compositional and publishing histories (including what can be inferred about them through close reading) brings one in touch with basic critical questions about each work. These analyses, along with similar ones that Parker has published separately,[62] stand on their own, and it is regrettable that he tried to erect a theory around them. He thinks of himself as promoting a new movement and wishes he had found a name for it, recognizing that his earlier term, "the New Scholarship," was not right (p. 241).[63] In fact his achievement is to have provided some telling examples of the connections between good textual scholarship and good criticism; no label is needed to dignify that accomplishment, and his awkward attempt at a unifying theory of the creative process only detracts from it.

Once we see that Parker's approach actually moves away from the author, we should not be surprised to find that a reader-response critic takes the same general line.[64] Steven Mailloux is a rarity among such critics in feeling the need to discuss "Textual Scholarship and 'Author's Final Intention'"—as he entitles one chapter (pp. 93-125) of his book Interpretive Conventions: The Reader in the Study of American Fiction (1982).[65] He sees himself as offering a "third alternative" for the editor


Page 35
between the "originating view" of the author as an "isolated figure freely expressing his uniquely individual and privileged intention" and the "collaborative view of the writer as merely one among many text-producers": the "convention-based" view, in which authorship is a "convention-governed role" (and thus "socially constituted") but one that "individuals can take on" (p. 107). Although this approach combines some aspects of the other two, he considers one "advantage" of it to be that "it preserves scholarly editing's traditional emphasis on the 'author's intention'" (p. 108). There is no question that he takes the author's side when external pressure is involved; changes made under such pressure are invalid, he says, and editors are "fully justified" in rejecting them (p. 118). From there on he gets himself into the same predicament as Parker. Some revisions freely made by the author result in an "illogical or inconsistent restructuring of reader response" (p. 117), which "provides an aesthetic reason for giving the original version priority" (p. 120). In other words, "editors can take control of the intended text out of the author's hands if justified by an examination of historical evidence and the intended structure of the reader's response" (p. 121).[66] This approach "preserves" the "traditional emphasis" on authorial intention in a strange way, by shifting textual authority from author to editor (who is also a reader). Mailloux, in the end, makes editorial decisions the same way that Parker does, but he does not begin from the same premise, since he believes that "literary authorship is socially constituted" (p. 107).


Page 36
Here we have a social view of authorship that nevertheless sanctions editorial interventions on the supposed grounds of authorial intention. And thus have we come full circle.


These proposals to treat literature either as a social product or as the outcome of a private creative process are marred by various flaws of argument, but the weakness they share is their assumption that only one approach is valid. Editorial theorists who recognize (as most of them do) that a critical text is based on critical judgment should also see that no one critical text can be the best one from everyone's point of view or for all purposes. If there is a legitimate interest both in a writer's process of creation and in the vicissitudes that writings undergo in the process of initial and subsequent publication—as there obviously is for any historically minded person, since both are historical processes—then various approaches to textual criticism, emphasizing one or the other of these interests, must be acknowledged to be acceptable, depending on their own internal logic. Thoughtful editors have long recognized this point, and even if they have become advocates of a particular approach they have not denied the usefulness of multiple editions reflecting different approaches.[67] Yet the lure of the single text is so strong[68] that it has made many, perhaps most, editorial debates less fruitful than they might have been. Among recent writers on editorial matters, two—Hans Walter Gabler and Peter L. Shillingsburg—have particularly addressed themselves


Page 37
to the question of how to accommodate multiple texts of a work. For both of them a theory of literature supports the argument for the necessity of multiple texts; but they both also give serious thought to the practical matters of apparatus, because the chief problem for this capacious view is the challenge of adequately displaying variant readings.[69]

Gabler's position, far less well thought out than Shillingsburg's, serves to focus some of the issues. In a paper for the 1981 Society for Textual Scholarship conference,[70] Gabler argues that we take too limited a view when we concentrate on the "synchrony" of a particular version of a work and that we ought to be alert to the "diachrony" of the evolutionary stages through which a work develops. We must distinguish, he says (and few would contradict him), between transmissional variants (which are deviations from what the author intended) and authorial variants (some of which are too often regarded as further deviations from an ideal text). The former should be corrected through emendation, but all of the latter, he believes, have a place in the literary work, which "may be said to comprise all its authorial textual states." What Gabler calls "a natural condition of the literary work" is "the manifest existence of discrete authorial versions of a text." In his pretentious language, the "total text" of a work "presents itself as a diachronous structure correlating the discrete synchronous structures discernible." The variant is not an "extraneous irritant" but an "integral textual element of pivotal significance in the textual totality of the work" (p. 309). One sees what he is getting at here, despite the expression and despite two serious conceptual flaws: first, he continually refers to transmissional errors in such a way as to suggest that it is no problem to separate them from authorial revisions, when of course making that distinction is in many instances a central editorial activity;[71] second, he repeatedly speaks of "discrete textual states" (even claiming that "there


Page 38
must always have been discrete textual states, in temporal succession, of a literary composition") without facing the fact that revision frequently does not proceed in readily separable "discrete" stages and that—when it does—such stages do not necessarily coincide with those represented by surviving documents.

Having set forth this view of the nature of literary works, Gabler proceeds to outline a method for presenting the evidence, for he believes that the traditional apparatus is at odds with that view. Since "revisional variation" is "meaningful only in its contextual relations," we need to "devise modes of apparatus presentation which leave the contextuality intact": "Lemmatised fragmentation is categorically not suitable for the purpose" (p. 311). Instead, what "would seem an absolute necessity" is "some manner of 'integral apparatus' for the visualisation of revisional variance in invariant contexts, which in this case should display the work's entire shape, or sequence of variant shapes, in apparatus form before the critic's eye." Whatever form is chosen, it must not lose sight of "essential tenets, such as those of the situatedness in context of the revisional variant and its integrity to the work's total text" (p. 313). The point that emerges from this verbiage is that authorial variants, being part of the literary work, are more appropriately reported in the running text than in an appended list. He illustrates some of the possibilities, such as texts in parallel columns and the "synoptic" text he has adopted for Ulysses. However much Gabler wishes to make the presentation of variants a matter of theory, it remains a practical issue. All editors, whatever their theory of literature, recognize that variant readings, to be understood, must be placed in context; and when editors decide that the most suitable form of report under the circumstances is an appended list, they are not suggesting that the variants are somehow less significant than those that other editors choose to print (marked with various symbols) in the text. Readers who have used texts with integrated apparatuses (such, indeed, as the Ulysses) know that they are not necessarily easier to follow than appended lists.[72] Gabler implicitly recognizes the difficulty


Page 39
of reading such texts when he provides for "an orientation text for the apparatus," or a "reading text" (p. 316), or even "the edition text" (p. 318).[73] But why, one is bound to ask, should there be a separate "reading" text if all the variants are an essential part of the work? Why should "the object of scholarly and critical analysis and study" (which is "the totality of the Work in Progress") be seen as "opposed" to "a general public's reading matter"? If the "work of literature possesses in its material medium itself, in its text or texts, a diachronic as well as a synchronic dimension" (p. 325), does it make sense for the "general public's reading matter" to be something less than the whole work? Is Gabler saying that scholars and critics need the real work, but ordinary readers can make do with what amounts to an abridged version, offering less than the full aesthetic pleasure that the work in its entirety provides? Gabler's position has obviously not been carefully thought through, but it is worth noticing here because it calls attention to questions that must be faced by anyone who hopes to defend the idea of multiple texts.

Peter Shillingsburg handles these issues with far greater sensitivity. His book, Scholarly Editing in the Computer Age: Lectures in Theory and Practice (1984),[74] is so thoughtful and refreshing that one hesitates to mention its problems, for they (as opposed to the flaws in so many more limited efforts) do not stand in the way of the salutary effect it can, and should, have. Like Gabler, Shillingsburg believes that "literary works of art, unlike some other forms of art cannot safely be treated as single end-products" and that variants resulting from authorial revision form "an important part of a reader's experience of the work" (p. 31);[75] thus he, too, considers the handling of apparatus crucial to the reader's understanding. In contrast to Gabler, however, Shillingsburg bases his conclusions on a thorough reexamination of all the concepts that underlie textual and editorial work, and he does not limit his recommendations to a single approach. His rethinking of the whole process of editing gives weight to his recognition that the various approaches to


Page 40
editing, and the various approaches to literature from which they arise, are irreconcilable and must all be accepted.[76] He deplores the "single-mindedness of each school in thinking theirs is the higher course" and urges that editors "acknowledge, not the correctness, but the legitimacy of opposing viewpoints." Editors ought to "explore the felt needs that the different approaches seek to fulfill" and then to "create editions that at least acknowledge the potential of other approaches" (p. 73).

Since a single critical text can represent only one approach, it falls to the apparatus to show that potential. Shillingsburg does not object to appended lists, but he strongly believes that variants should be identified more fully than they normally are in lists, associating each one not only with a "source document" but also with a "source agent (author, editor, compositor)" and a "source time" (p. 32). He argues that lists in which all variants, authorial and nonauthorial alike, are indistinguishably mixed do not encourage readers who are interested in a different approach to explore the evidence: "trivialized tables confirm critics in the habit of not using the apparatuses of critical editions because they cannot imagine what to use them for" (p. 75).[77] Shillingsburg's view here is analogous to Gabler's, both in its emphasis on the separation of authorial and nonauthorial variants and in its feeling that lists can sometimes make information seem less important than it is. Everyone naturally is in favor of making apparatuses as useful and convenient as possible, but whether the segregation of authorial from nonauthorial variants will always help readers is far from obvious: because the distinction between the two is by no means clear-cut and thus depends on critical judgment, readers will have to examine both categories in any case to determine whether their own judgment agrees with the editor's. And even when, as usual, only the documentary source of variants is identified in the apparatus, readers normally know—from the explanation of textual policy for the edition—the agent (or class of agent)[78] to whom the editor has attributed each type of variant in each document. I would not deny that in some instances dividing a long list of variants into categories could be helpful, but the gain does not seem as dramatic as Shillingsburg suggests;


Page 41
and it is hard to believe that readers who do not understand the purpose of apparatus in one form will readily see it when the form is shifted. The impulse to use apparatus is not instilled in readers by the inviting qualities of the apparatus but by those readers' habits of mind; readers who are interested in history will want to examine the evidence, and whatever form it is presented in will not deter them (though of course they are glad to have it in as convenient a form for their particular purpose as possible).

One can accept Shillingsburg's advice on apparatus or not, as one chooses: the issue is one of convenience, not substance. His significant point is simply his emphatic reaffirmation of the essential function of apparatuses as supplements to reading texts. Justifying a reading text without interpolated variants poses less theoretical difficulty for Shillingsburg's position than for Gabler's. Although Shillingsburg does not deal directly with this matter, the rationale would be, I think, that in order to accommodate all approaches one has to recognize that for some of them alternative readings cannot be considered part of the text. Apparatuses, then, must be looked at in two ways, as Shillingsburg's varying manner of speaking about them suggests: they may be seen as recording "utterances that were or remain a part of the work"[79] or else as presenting "significant information about the work" (p. 75). Some readers will see them one way, and some the other. What Shillingsburg is saying could be reduced to the standard point that editors normally profess: apparatuses are crucial because they enable readers to take different approaches and make different judgments from the editors'. But this point has not always been made in a spirit of tolerance. Shillingsburg's contribution lies not in his advice on constructing apparatuses but in his restatement of an old truth in an uncommon context, one that stresses genuine openness to alternative approaches.

He also makes a contribution in his sane review of the fundamental theoretical issues that all editors must take a position on. Even when they cannot fully agree with him, editors and other readers will benefit from working through his intelligent analyses of such matters as intention, ontology, and what he regards as the four basic approaches to editing. His discussion of intention, for example, is helpful in its emphasis on a writer's "intention to do" (that is, to write "a specific sequence of words and punctuation") rather than an "intention to mean" (pp. 27-29); but in the process Shillingsburg overstates the recoverability of the intended sequence of words and punctuation (he says it is "almost completely


Page 42
recoverable")[80] and underestimates the degree to which establishing an "intention to do" involves postulating an intended meaning.[81] In his


Page 43
chapter on "Ontology," Shillingsburg makes thoughtful distinctions among the concepts of work, version, draft, and text and shows how the entities envisaged in each case have no material existence,[82] though they may be stored in a physical medium (each instance of such storage being a document). His discussion falters, however, in not accounting for the possibility that some literary works are also works of visual art (which therefore must be acknowledged to have a material existence)[83] and in not sufficiently confronting the fact that versions and drafts need not coincide with the texts of surviving documents. He states that a version "is represented more or less well or completely by a single text as found in a manuscript, a proof, a book, or some other written or printed form" (p. 36) and that drafts "are represented more or less well by the manuscripts containing them" (p. 38). The phrase "more or less" alludes specifically to the fact that texts can incorporate slips or errors not intended by the author to be part of a version or its drafts;[84] it does not seem also


Page 44
to refer to the fact that versions and drafts (in the abstract and, I think, appropriate sense employed by Shillingsburg) cannot be tied to the texts of physical documents.[85] Another chapter[86] enumerates the basic approaches to editing (or "orientations") as the "historical," the "aesthetic," the "authorial," and the "sociological." Much of what is said about them, as one would expect, is sensible; but the categories, as set out, are not entirely satisfactory. The "historical" approach includes texts with editorial emendations as well as diplomatic editions and facsimiles; how, then, does this approach differ from the "authorial" and the "sociological," both of which are critical and historical?[87] The "aesthetic" seems similarly to be blurred in conception, for it encompasses the activity of "Commercial editors, literary agents and other merchandizers of literary works"[88] as well as that of scholarly editors who limit


Page 45
themselves to readings "already existing in historical documents" (p. 16). But does not a distinction need to be made between those who improve a work according to their own aesthetic judgments and those who use aesthetic judgments in the historical task of reconstructing authors' or publishers' intentions?[89] Such shortcomings in Shillingsburg's discussion do not prevent its offering some enlightenment and do not lessen the persuasiveness of its tolerance for differing approaches.[90]

It may seem that nothing new has been said by these various writers, and in many ways that is true. The basic issues that confront textual critics and scholarly editors are unchanging, and the attitudes that may be taken toward those issues, though occasionally appearing in altered guises, remain the same. There will be no end to debates over these issues, because they are genuinely debatable; and the process of debate is the way in which each generation of editors thinks through the questions for itself. Some recent editors have claimed that the field is at present in a state of crisis. But the fact that different people hold different opinions about basic issues is not a sign of crisis; it points to the perennial situation in any challenging and lively field. The repeated advocacy of particular viewpoints is not wasteful, except when the advocates do not mend the internal flaws of their predecessors' arguments. Advocates of differing positions need not give up their positions if sound arguments can be made for them; but we do have reason to be discouraged when the arguments continue to suffer from the same defects. There is a prima facie case for the legitimacy of more than one approach to the editorial treatment of historical evidence: editors can produce diplomatic or facsimile editions of individual documentary texts; or they can through emendation create new texts that attempt to be historically faithful either to authors' or to publishers' intentions at particular times. Acceptance of this multiplicity does not, and should not, end debate, for one still has to decide which approach is to be followed in a given situation, and many factors bearing on that decision can usefully be discussed. But the debate thereby moves to a different, and


Page 46
more productive, level. Advocacy of one position then occurs in the context of valid alternative positions; and criticism leveled at a given argument springs not from a belief that only one approach is correct but from the detection of logical flaws in the argument. Enlightened editors who understand this point have always existed; but there is also room, in the field of editing as in any other, for the further spread of enlightenment. Recent writings on editorial theory, like those in times past, provide the basis both for exasperation and for hope. If the acceptance of multiple approaches, as well as the insistence on rigorous argument in support of each, can become more widespread, the quality of the debate will improve, even as the points debated remain (as they must) the same. The recognition that all approaches to the past are partial and complementary helps one to appreciate the full complexity of the issues editors struggle with. Viewed in this way, scholarly editing appears, with more justice than ever, as one of the most demanding and rewarding forms of the critical study of history.



"Toward a New History in Literary Study," Profession 84 (1984), pp. 16-23.


Lindenberger lacks precision in making this point. His complete sentence is, "Given the suspicion we have developed in recent years toward authorial authority, even an author's authorized text need have no more authority than we choose to give it." However, it is not necessary to question the authority of authors over their texts in order to be wary of "authorized texts," which for many reasons may not fully reflect their authors' intentions. The long tradition of critical editing has involved questioning authorized texts but not normally doubting the authority of authors.


The previous two brought the story from the time of W. W. Greg's "The Rationale of Copy-Text" (SB, 3 [1950-51], 19-36) to the late 1970s: "Greg's Theory of Copy-Text and the Editing of American Literature," Studies in Bibliography, 28 (1975), 167-229 (reprinted in Selected Studies in Bibliography [1979], pp. 245-307); and "Recent Editorial Discussion and the Central Questions of Editing," SB, 34 (1981), 23-65. As in those pieces, I am again concentrating on general theoretical discussions and normally exclude essays on textual problems in particular authors' works and reviews of individual editions. During the period under review several useful checklists appeared that do include such material: T. H. Howard-Hill, British Bibliography and Textual Criticism: A Bibliography (2 vols., 1979, continuing the series that also contains Shakespearian Bibliography and Textual Criticism: A Bibliography [1971]); Ross W. Beales, Jr., "Documentary Editing: A Bibliography," Newsletter of the Association for Documentary Editing, 2, no. 4 (December 1980), 10-16; Laurel N. Braswell, Western Manuscripts from Classical Antiquity to the Renaissance: A Handbook (1981); and David Madden and Richard Powers, Writers' Revisions: An Annotated Bibliography of Articles and Books about Writers' Revisions and Their Comments on the Creative Process (1981). There were also several anthologies of essays on editorial matters (other than those mentioned later in this essay), such as N. John Hall's collection on editing the Victorians (volume 9 of Browning Institute Studies, 1981) and recent volumes in the Toronto Editorial Conference series: e.g., William Blissett (ed.), Editing Illustrated Books (1980); A. H. de Quehen (ed.), Editing Poetry from Spenser to Dryden (1981); Trevor H. Levere (ed.), Editing Texts in the History of Science and Medicine (1982). Also during this period Donald H. Reiman published two perceptive accounts of editorial history: "The Four Ages of Editing and the English Romantics," Text 1 (1981), 231-255; and "Romantic Bards and Historical Editors," Studies in Romanticism, 21 (1982), 477-496. And two general introductory treatments of textual and editorial scholarship came out in Modern Language Association publications: my essay on "Textual Scholarship" in Introduction to Scholarship in Modern Languages and Literatures, ed. Joseph Gibaldi (1981), pp. 29-52; and William Proctor Williams and Craig S. Abbott's chapter on "Textual Criticism" in An Introduction to Bibliographical and Textual Studies (1985), pp. 52-90.


A "documentary text" can of course be the text of either a manuscript or a printed item: in either case the text is documentary in that it is what has been preserved in the historical record.


Editors of literary works are historians, too; but for convenience I use "historians" to refer to editors of the documentary texts largely used by students of history.


I do not mean to suggest that all historical editors think alike: I am aware that there are many who do understand what the literary editors have been discussing in recent years. But the most vocal historical editors, and those who publish their views, have not generally come from this group. An exception is John Y. Simon, who showed his open-mindedness in his ADE Presidential address, "Editors and Critics," Newsletter of the Association for Documentary Editing, 3, no. 4 (December 1981), 1-4. Jon Kukla, in his review of the Kansas conference volume (see below), recognized the irony that "as the so-called literary editors push for more thoughtful handling of textual evidence, they are forcing all scholars to confront the fundamentals of historical method" (Newsletter, 4, no. 1 [February 1982], 5-7).


Other lines, within the field of literary studies, need to be opened as well: see Conor Fahy, "The View from Another Planet: Textual Bibliography and the Editing of Sixteenth-Century Italian Texts," Italian Studies, 34 (1979), 71-92. On the need for interdisciplinary discussion, see also my comments in Text, 1 (1981), 1-9.


This shorthand use of "literary" encompasses works in other genres as well. Much of the recent scholarly editing of philosophical works, for example, is in the tradition of "literary" editing. See Fredson Bowers, "Editing a Philosopher: The Works of William James," Analytical & Enumerative Bibliography, 4 (1980), 3-36.


Recent discussions of letters include J. A. Dainard (ed.), Editing Correspondence (1979); Robert Stephen Becker, "Challenges in Editing Modern Literary Correspondence," Text, 1 (1981), 257-270; A. R. Braunmuller, "Editing Elizabethan Letters," Text, 1 (1981), 185-199; Norman Fruman, "Some Principles of Epistolary Interpretation," Centrum, n.s., 1 (1981), 75-94; and Ernest W. Sullivan II, "The Problem of Text in Familiar Letters," PBSA, 75 (1981), 115-126.


He begins by saying that the literary editor "usually works with materials quite different from those with which the historical editor works, for the latter is interested in manuscripts" (p. 23). Literary editors do frequently deal with works that have been printed; but, if manuscripts survive, these editors not only are "interested" in them but often make them the basis for a new edition. Not realizing this fact, Rogers proceeds to assert, "Literary editors now agree that the first printed version has the best claim to be the copy-text." There is no hint here of the extensive debate, then and now, over this point, and no suggestion that individual situations might affect the choice of copy-text. Literary editors, he claims, make "no attempt to present a facsimile version" (in fact, of course, they often do). instead preferring critical editions, in which "the editors take liberties with the copy-text in order to obtain what the editors consider a text closer to the authorial intention" (p. 24)— a statement that would have been fair enough if the process of emendation had not been misunderstood and trivialized as "taking liberties." In the collation of texts that provides some of the evidence for emendation, we are told, substantives "receive more attention than accidentals" (the naivete of expressing this view without further elaboration clearly goes unrecognized). "In this scheme of editing," Rogers adds, "there tend to be no silent changes" —as if the mechanical reporting of emendations is a part of a theory of editing. There is no point continuing this recital: nearly every statement he makes about the work of literary editors is incorrect or seriously askew. Similar misunderstandings crop up elsewhere in the conference volume: in one of the two introductions, George L. Vogt says, "Most historical editors probably still balk at the literary editors' idea of a completely 'recoverable' text, with all the back-of-the-book baggage that that implies" (pp. 4-5); but no one would claim that every feature of a manuscript text can be recovered from any edition, and the most detailed transcription imaginable need not have any appended apparatus, either at the back of a volume or at the foot of a page.


In a review of the conference volume in the Bulletin of the Bibliographical Society of Australia and New Zealand (7 [1983], 188-190), Harold Love agrees that the faithful reproduction of manuscript texts is desirable but adds that "the overriding considerations must always be the kind of use such documents are to be put to by historians"; it is hard to believe, however, that any serious use of the texts of documents is well served by the elimination of evidence that could play a role in their interpretation. Gordon S. Wood has concisely said, "For historians, convenience of use apparently overrides their concern for literal accuracy" (p. 875 in "Historians and Documentary Editing," Journal of American History, 67 [1980-81], 871-877).


I am not in any sense denigrating annotation, which is one of the kinds of commentary that an edition should set in motion (whether within the covers of the edition or elsewhere). Textual and other kinds of historical annotation are not, in any case, easily separable entities, since textual questions cannot be handled sensitively apart from the historical context.


Rogers, as a historian, is surely going out of his way to be hostile when he claims not to understand that the value of research is not determined by how dramatic the results are: he asks (irrelevantly, but with an implied answer that is incorrect), "Have the editors of nineteenth-century texts come up with any earth-shaking discoveries?" (p 25).


"Editorial Practices—An Historian's View," Newsletter of the Association for Documentary Editing, 3, no. 1 (February 1981), 4-8. The counterargument was provided by Don L. Cook, in "The Short Happy Thesis of G. Thomas Tanselle," pp. 1-4 (the reference being to my piece on "The Editing of Historical Documents," SB, 31 [1978], 1-56—reprinted in Selected Studies in Bibliography [1979], pp. 451-506).


It is no defense of editorial interference to take the line (implied by Taylor on p. 5) that such alterations do not affect the meaning. As I have previously pointed out (SB, 34 [1981], 58), why then go to the trouble of making the changes at all?


Literary historians will be surprised to hear that "No one feels the need to study the poems of mediocre poets or to run through their letters, unless they made better friends than they did poems" (p. 6). A number of historians have made statements similar to Rogers's and Taylor's. Fredrika J. Teute, in "Views in Review: A Historiographical Perspective on Historical Editing," American Archivist, 43 (1980), 43-56, says that "the historical editor treats the document as a fact. While perhaps slighting the nuances which literary editors appreciate, he does not produce bowdlerized versions claiming to represent the author's true, though unexpressed, intent" (p. 49). And Nathan Reingold, in "Reflections of an Unrepentant Editor," American Archivist, 46 (1983), 14-21, states that "the literary editors have a sense of the sacredness of the words they process. The historical editors, even those imbued with origins and essences, have a belief in the importance of the purport of the words" (p. 18). In the first number of the Newsletter of the Association for Documentary Editing (March 1979), William B. Willcox had referred to the "high priests of ipsissima verba, who bemoan our textual impurities" (p. 5).


"The 'Authentic' Witness: The Editor Speaks for the Document," Newsletter of the Association for Documentary Editing, 4, no. 1 (February 1982), 8-9. Cutler's piece was written in reply to David J. Nordloh's 1979 conference paper, "The 'Perfect' Text: The Editor Speaks for the Author," printed in the Newsletter, 2, no. 2 (May 1980), 1-3.


"The Editor and the Question of Value: Proposal," Text, 1 (1981), 41-43. Although the paper she actually delivered at the conference was different from the one foreshadowed by this proposal, the proposal is printed in Text because Fredson Bowers based his commentary for the conference ("The Editor and the Question of Value: Another View," pp. 45-73) on it.


Bowers properly objects to the solecism "textual editor" and points out that a "documentary editor" must also be concerned with texts (e.g., p. 50).


Or, as he puts the point more explicitly, alterations occurring within the text of a document "are as much an integral part of the document as the final form of its inscription" (p. 49). By neglecting such alterations, and variants between separate drafts, historians have not learned as much as they might have from "the writer's veerings of thought" and "changes of mind" (p. 48).


"On Native Ground: From the History of Printing to the History of the Book," Proceedings of the American Antiquarian Society, 93 (1983), 313-336 (and reprinted as a pamphlet, 1984). I have made some further comments on Hall's views in "The Bibliography and Textual Study of American Books," Proceedings of the American Antiquarian Society, 95 (1985), 291-329. Hall once described Charles Francis Adams's 1894 "literal reproduction of the texts" of some antinomian documents and then observed, "Scholarship no longer rests upon such antiquarian exactness" (The Antinomian Controversy, 1636-1638: A Documentary History [1968], p. 21).


He states that American printers and readers of the past "were quite indifferent" to the question of authorial intention. But that observation, however true it may be, has no bearing on our legitimate interest in discovering what a particular author meant to say at a given time in the past.


Though of course present reactions are only the latest stage in the history of the reception of various texts of a work.


I have tried to set forth this point in my Hanes Lecture, The History of Books as a Field Study (1981)—also printed in the Times Literary Supplement, 5 June 1981, pp. 647-649.


As much as possible, that is, within the limits imposed by typography and photography. Some evidence, of course, cannot be reproduced.


Stephen E. Wiberly, Jr., in discussing the interesting question of the editing of maps does recognize critical editions; indeed, he asserts—without qualification—that "edited maps are superior to facsimiles" (p. 509). But although he refers to the debates over the editing of verbal historical documents, he seems not to have grasped some of the issues involved in them. It does not seem promising, for example, to distinguish maps from personal papers on the grounds that "a map purports to tell us something about a reality outside the mind of its creators" and to conclude that it is therefore "logical to correct the contents of old maps" but "not logical to change the contents of personal papers" (p. 502). What he advocates is to "take an old map, accept some of its data without reservation, verify and correct other of its data, and then include both types of data in the edited map" (pp. 501-502). See "Editing Maps: A Method for Historical Cartography," Journal of Interdisciplinary History, 10 (1980), 499-510.


Whether variant readings and emendations are noted within the body of an edited text or in appended lists is a mechanical matter—not unimportant, but certainly not of the same order as the considerations just mentioned.


Bowers says of documentary texts: "Students of history may read these texts to gain a firsthand acquaintance with the undigested material, but at higher reaches these documents provide professional scholars with the necessary data from which formal eclectic interpretation can be made in written histories and biographies" (p. 46). The essential role of subjective judgment in the historical enterprise of reconstructing past texts out of the documentary texts that happen to have survived is affirmed (with somewhat different emphases) by two recent essays on the textual study of early manuscripts: my attempt to show that recension is no less conjectural than "conjectural emendation," in "Classical, Biblical, and Medieval Textual Criticism and Modern Editing," SB, 36 (1983), 21-68; and Lee Patterson's demonstration that the standard dichotomy between external and internal evidence does not coincide with the distinction between objective facts and interpretations, in "The Logic of Textual Criticism and the Way of Genius," in Textual Criticism and Literary Interpretation, ed. Jerome J. McGann (1985), pp. 55-91. (For a more traditional summary of some of the flaws in the stemmatic approach, see Paul Oskar Kristeller, "The Lachmann Method: Merits and Limitations," Text, 1 [1981], 11-20.)


Buch und Buchhandel in Europa im achtzehnten Jahrhundert, ed. Giles Barber and Bernhard Fabian (1981), pp. 81-125. McKenzie's paper, as he states, "extends the argument" of the third of his unpublished Sandars Lectures for 1975-76 ("The London Book Trade in the Later Seventeenth Century").


James McLaverty, who acknowledges the influence of McKenzie, has also found the early eighteenth century to be the time "when authors first became conscious of the anomalous nature of literature and of the importance of printing" (p. 95). He examines the Dunciad Variorum as Pope's "exploitation of the medium" of the scholarly edition, as "an imitation not of spoken discourse but of written discourse" (p. 96). Like McKenzie, he argues that the visual presentation of a text "can carry special associations, and that a richer understanding of the relation of author, book trade, and public may lead to better interpretation of literary works" (p. 105). See "The Mode of Existence of Literary Works of Art: The Case of the Dunciad Variorum," SB, 37 (1984), 82-105. Another writer who is concerned with the effect of typographic layout, especially lineation, on the "reader experiencing the play in the 'theater of the mind'" (p. 69) is Paul Bertram, in White Spaces in Shakespeare (1981).


Analytical bibliographers, of course, have used physical evidence in an attempt to uncover facts about the printing history of individual editions (and this activity can be regarded as one kind of "reading"—see Ross Atkinson, "An Application of Semiotics to the Definition of Bibliography," SB, 33 [1980], 54-73). But the concern here is with the way in which the physical characteristics of a book play a role in the reader's reaction to the piece of writing contained in it.


Scholars should, after all, be aware of the shifting conventions in the use of particular formats and page layouts.


It is implicit in his reference to decisions that authors and booksellers "take, or impose on one another" (p. 103), but the point is not developed.


The word "best" is troubling here and recalls his earlier statement, "Every variant must of course be scrutinized for what it adds or loses in vitality of character and acuity of language" (p. 108). One has to wonder whether this way of stating the matter sufficiently distinguishes editors' own preferences from their judgments regarding what authors preferred.


McKenzie does not acknowledge this point when he refers to "an historic and contextual accuracy in the recovery of every possible element of meaning as intended by the author and perceived by an intelligent and sensitive reader of his time" (p. 92).


Though it may seem that he is at other times, as when he speaks of "the most important one [concern] of all—what, exactly, an author in his own age did say to his readers and how he and his printers directed them to respond" (p. 123). What the author did say and what the author wanted to say are not always the same.


McKenzie devotes the first section (pp. 83-92) of his paper, rather irrelevantly, to a critique of Greg and three commentators on him (Morse Peckham, Hans Zeller, and Tom Davis). The ostensible relevance is that Greg dealt with the drama of a period in which typographical care was not bestowed on printed plays; and the argument is that recent editorial theory, greatly influenced by Greg, is therefore not equipped to handle situations in which text and presentation are integrated. McKenzie gets off the track in emphasizing the distinction between substantives and accidentals and treating it as a distinction between meaningful and formal elements, which (at least in its use by Greg's followers) "has been utterly divisive, shattering any concept of the integrity of the book as an organic form, a material statement in which all its elements participate" (p. 84). Although he is right to find the criticisms of Peckham, Zeller, and Davis deficient, his own analysis fails to recognize that the distinction between substantives and accidentals is actually based not on meaning but on a generalization about human behavior in the past, that it comes into play only when one has no other means of reaching a textual decision, and that accidentals are in fact treated as extremely important elements in texts. There is nothing in Greg's approach, even as extended and followed by others, that prevents one from regarding typography as a textual matter or from preserving the integrity of particular texts (late or early), if one believes the typography or integrity to have resulted from authorial intention; Greg's rationale is no more divisive than any other plan for critical editing, because it recommends eclecticism only when a unity of elements does not exist and only to bring about greater unity. It perfectly well accommodates McKenzie's demands for a textual theory, and there is thus some irony in his charge that the tradition deriving from Greg suffers from "intellectual timidity" (p. 92).


"The Sociology of a Text: Orality, Literacy and Print in Early New Zealand," Library, 6th ser., 6 (1984), 333-365. Another sensitive treatment of oral texts—of "the pleasures and advantages of preliteracy"—is John Miles Foley's "Editing Oral Epic Texts: Theory and Practice," Text, 1 (1981), 75-94.


Indeed, it moves outside books and for him becomes a pattern in life. There is, he believes, "a principle of textual criticism operative in the real world which implies the concept of an ideal text that the versions have failed fully to express": "The physical versions and their fortuitous forms are not the only testimonies to intent: implicit in the accidents of history is an ideal text which history has begun to discover, a reconciliation of readings which is also a meeting of minds. The concept of an ideal text as a cultural and political imperative is not imposed on history but derives from it and from an understanding of the dynamics of bibliography as a study of the meanings 'books' make" (p. 364).


Another writer who has stressed the importance of original typography is Randall McLeod, but his point is different from McKenzie's. His concern is not primarily with the way typographic layout carries historical meaning or with the possibility that it is an integral part of a literary work; rather, he has shown in a series of articles how typographic evidence can be essential in making textual decisions (as when a particular spelling was caused by the necessity of inserting a type between two kerned types that would not fit together) and how scholarly editions that fail to report such evidence deprive readers of information they need for evaluating editorial judgments. Unmodernized (or "old-spelling") scholarly editions, in other words, do not go far enough toward providing historical evidence, for they separate spelling and punctuation from typography. Facsimile editions are desirable, in this view, not specifically because the original typography is part of the text or is what emerged from the original process of publication but because they make available to readers the evidence present in the original typography (that is, all of it that is reproducible). Whereas McKenzie is concerned with determining what details constitute a literary work, McLeod is talking about the presentation of evidence. If we are to have critical editions at all, we must resign ourselves to the fact that their apparatuses will not contain every piece of evidence used by their editors; but we have a right to expect that editors will recognize, take into account, and comment on the kind of evidence McLeod discusses, and his articles are valuable in calling attention to uses of typographic evidence often neglected. His criticism of the practices of some editors is well taken, but it does not in fact invalidate the concept of critical editions in modern typography. See his "Spellbound: Typography and the Concept of Old-Spelling Editions," Renaissance and Reformation, n.s., 3 (1979), 50-65; "Editing Shak-speare," Sub-Stance, no. 33/34 (1982), 26-55; and "Gon. No more, the text is foolish.", in The Division of the Kingdoms (see note 68 below), pp. 153-193.


He had earlier expressed some of the same ideas in essays, as in "The Text, the Poem, and the Problem of Historical Method," New Literary History, 12 (1980-81), 269-288, where he says that "every work of art is the product of an interaction between an artist, on the one hand, and a variety of social determinants on the other" (p. 275). In two other papers presented at textual conferences in 1981 and 1982 (but not published until after his book) he explores the "social nexus" of literary works and the relations between "historicist textual criticism" and literary criticism: "Shall These Bones Live?", Text, 1 (1981 [published in 1984]), 21-40; and "The Monks and the Giants: Textual and Bibliographical Studies and the Interpretation of Literary Works," in Textual Criticism and Literary Interpretation, ed. McGann (1985), pp. 180-199. The latter piece presents a useful (if rather melodramatic) statement of the role of textual criticism in literary study, depicting textual criticism as a broad field of historical scholarship, in which the production of editions is only one of many agenda (a view I agree with, though McGann believes I do not).


It is not, however. But the fact that his criticism of this school of editing is often superficial or incorrect is a secondary point; the more fundamental matter to be examined is his general attack on an author-based approach to editing (whether or not he has the details of the Bowers line right). I shall simply note here three of the deficiencies in his treatment of the Bowers position. The most serious flaw is his failure to acknowledge its flexibility, for it does not demand that a manuscript be used as copy-text in preference to a first edition, or a first in preference to a later edition, when the circumstances suggest otherwise, and it is capable of handling the examples he cites. A prime instance of this flexibility is Bowers's own practice in his William James edition and his own theoretical statement in "Greg's 'Rationale of Copy-Text' Revisited," SB, 31 (1978), 90-161—an essay cited once by McGann but not really used in his analysis. It is thus incorrect to state flatly, without qualification, that "the Bowers position is that the author's manuscript is a higher authority" than the first edition (p. 55). McGann also stresses the word "final" in "final intention" in an extremely literal way that I think the editors following Greg have not normally done; in practice the term has excluded the intentions reflected in unfinished drafts but it has not been taken to rule out the possibility of different intentions manifested in different completed versions. And several times he speaks of "the rule of final intentions" governing "the choice of copy-text" (as on p. 55), without recognizing that an early copy-text chosen for its accidentals may be very far indeed from the author's "final intentions" with respect to substantives.


McGann's unwillingness to accept this point underlies his belief that the Bowers system, based on a classical model (derived, via Greg, from Lachmann), is inappropriate for dealing with modern materials: because the textual critic of modern works "actually possesses the 'lost originals' which the classical critic is forced to hypothesize, his concept of an ideal text reveals itself to be—paradoxically—a pure abstraction, whereas the classical critic's ideal text remains, if 'lost,' historically actual" (p. 57). The force of this observation as a criticism is not clear, for a "pure abstraction" can be a valid goal; and textual critics of classical works (or works of any other period) have the option of attempting to construct texts reflecting their authors' intentions, as well as attempting to recreate lost documents. The concept of "final intention" is not merely a device to help guide editors of modern materials through "the mass of documentary evidence" that often confronts them (p. 56); it leads to a goal worth working toward, regardless of the simplicity or complexity of the surviving documents, a goal that is historical whether or not the text being sought ever existed in written form. (Cf. McKenzie's point in note 39 above.)


Even to summarize McGann's position, as McGann himself does, in terms of an author and publisher working together to put something before the public does not, one would think, catch the spirit of McGann's argument, if what is implied here is that they necessarily worked together harmoniously and cooperatively. To think of the production of literature as a social or collaborative process is, one supposes, only to say that a number of people are involved and that any of them may influence the published text—not that they are never in conflict with one another.


He says that instances of the kind under discussion "go to the issue of textual versions rather than to the rationale of copy-text," but he is in fact talking about the choice of copy-text, and his summarizing paragraph speaks of "the decision on copy-text." At another point he distinguishes, without explanation, between "copy-text" and "version of a text we choose to work from" (p. 56).


They are thus not the product of historical scholarship, in the usual sense that it attempts to reconstruct past events. Of course any activity in the present is historical in the sense that it in some degree reflects its own time (as I shall note further below).


Another example of the attempt to mix these approaches is Donald Pizer's "Self-Censorship and Textual Editing," in Textual Criticism and Literary Interpretation, ed. McGann (1985), pp. 144-161. Pizer's four "'tests' for accepting the belief that self-censorship has occurred and that restoration of an earlier state of the text is required" are: evaluating evidence regarding the composition and publication history of a work, evaluating evidence bearing on the author's motives in making revisions, deciding through critical analysis whether one version is "better" than others, and considering whether "the first published version of a major work" is "a historical artifact that should continue to occupy the role of general reading text even if it has been subject to self-censorship" (pp. 150-151). The latter two are not in fact relevant to "accepting the belief that self-censorship has occurred"; they do represent two possible approaches to making textual decisions, but neither can be combined with the first two to provide a coherent rationale for choosing a single text. Pizer recognizes the value of publishing editions of alternative texts, but he still thinks ultimately in terms of a single text that he would like to see "generally read" (p. 155). (In itself, his account of "the interaction between self and world, which is inseparable from the expressive process" [p. 154], and of the "historical resonance" [p. 156] of long-established texts, is well stated.)


The "set of interconnected guidelines" he refers to at one point (p. 107) is not the same thing: it consists of considerations to be taken into account in choosing a copy-text, considerations that include "the character of the audience of the edition," "the early printing history" of the work, and "the current state of textual criticism" (p. 106).


He seems to think that he has opened up a wider range of options and thereby increased the area in which subjectivity can operate, but editors have always faced the same options, for there have always been valid alternatives to critical texts reflecting authorial intention (which in themselves present many alternatives).


The opening sentence of his chapter on modernizing betrays some confusion in terminology, if nothing else. He says that a "sharp distinction" is usually made "between the scholarly or critical edition on the one hand, and modernized or noncritical editions on the other" (p. 95). Generally the term "critical edition" is used to mean an edition containing a critical text, which is a text that does not agree precisely with a single documentary source because it incorporates alterations reflecting the editor's critical judgment (exercised to make the text more nearly conform with some desired goal). Thus an edition can be scholarly without being critical, and all modernized editions contain critical texts. Of course one can define these terms anew if one wishes, but it seems inappropriate to associate the word "noncritical," which already seems to refer to the absence of critical judgment, as in a facsimile, with "modernized," which involves a great deal of judgment and refers to editions that are at the opposite extreme from facsimiles.


McGann continually refers in this discussion to "nonspecialist editions," by which he means "nonspecialist texts," for he is referring to the adjustment of the text, not just the apparatus, for different classes of readers.


Arguments for modernizing are always doomed to failure, just as the practice of modernizing is bound to be inconsistent. Yet defenses of it continue to appear. Stanley Wells, in Modernizing Shakespeare's Spelling (1979), believes that modernizing "removes unnecessary barriers to understanding, making it possible for the reader to concentrate on the text itself, undistracted by obsolete and archaic accidentals of presentation" (p. vi). This view of "the text itself" links Wells with generations of compositors and publishers' editors that preceded him in feeling free to alter spelling and punctuation but less free to alter words (the behavior Greg described in his "Rationale"). But after going through Wells's discussion of the problems with which modernizing is fraught, readers are likely to be convinced that the real barrier between them and the text is being erected by the modernizer. Wells ends by explaining that modernizing editors have to give "hard thought" to the meaning of each word, and their choice of a form in each case "communicates the results of such thought." Modernization, he concludes, "may thus be seen not, as some would have it, as a work of popularization, even of vulgarization, but as a means of exploring Shakespeare's text that can make a real contribution to scholarship" (p. 34). It is difficult to see how the case for modernization is strengthened by saying that it forces editors to think hard about every word, for responsible editors of unmodernized critical editions must do the same thing. Editors who modernize, Wells says, "may be surprised to find how much that is of interest in Shakespeare's language has gone unnoticed." But modernizing seems too high a price to pay for a promise that editors have done the work we expect them to do anyway. (Other unconvincing arguments for modernizing have recently come from R. M. Flores in "The Need for a Scholarly, Modernized Edition of Cervantes' Work," Cervantes, 2 [1982], 69-87.)


In "Shall These Bones Live?" (see note 41 above), McGann has shown how powerfully he can write about the role of history in literary understanding. I believe that what I have suggested here regarding the validity of two fundamentally opposed emphases in textual criticism is analogous to his acceptance of the necessity for two approaches to literature ("two classes of men [scholars and critics] are always upon the earth of humane letters, and whoever seeks to reconcile them seeks to destroy the existence of their shared world" [p. 25])—and indeed seems to me more in the spirit of his comments than his own Critique is. But he also notes that a productive symbiosis, rather than a sterile co-existence, requires constant prodding from each side. His Critique—and, I hope, my remarks here—can be seen as part of this constructive process.


Some parts of the book had been published earlier, as in (among other places) "The Determinacy of the Creative Process and the 'Authority' of the Author's Textual Decisions," College Literature, 10 (1983), 99-125.


Thus "some perfectly real aesthetic frissons" are "spurious" (p. 11).


His criticism of Greg, the Center for Editions of American Authors, and Bowers is often unfair, so much so that it is self-defeating: for example, he speaks of editors who were moved "to stake their reputations on the validity of the rationale of copy-text set forth by W. W. Greg" (p. 59), as if reputations were the dominant concern; he says that "no editorial formula, even one as appealing as Greg's, can substitute for the expertise which comes only from years of conscientious (and preferably loving) biographical and critical study of the author whom one presumes to edit" (p. 67), as if editors who follow Greg substitute a formula for knowledge; he complains about certain critical judgments by CEAA editors "with well-nigh limitless federal funds" (p. 82), as if the magnitude of funds (which he knows he is exaggerating) were relevant to this issue; and he refers to the "hubris of the Newest Bibliography which at its worst barricaded editing within a mad-scientist laboratory more isolated from the author than any critic or theorist had yet been" (p. 240).


Of course one cannot limit the discussion exclusively to whole works, because the process of original composition may be spread out over a long period, or may be interrupted, with the result that the earlier and the later parts of a work may be the products of separate bursts of creative activity and may conflict with each other in various ways. (John McClelland, in "Critical Editing in the Modern Languages," Text, 1 [1981], 201-216, takes a position somewhat akin to Parker's in advocating as copy-text a text representing the point "where the work has acquired its definitive structure, either as the result of the accumulation of variants or by a dramatic revision" [p. 206], for "we are probably being more faithful to the text itself by presenting it in the state at which it was receiving the author's closest attention," since "extensive revision requires as much concentration as does composition" [p. 207]. McClelland, however, is more optimistic than Parker about the possibility of coherent later revision and thus would approve later copy-texts more readily than Parker. In a dubious argument, McClelland also claims that the variants would be "easier to handle" if the copy-text were "the text that will actually be read" [p. 204].)


Thus it is no objection to his line of reasoning to say that in rejecting late revisions the editor may not see what the author was getting at. Undoubtedly it does happen that editors find revisions thoughtless or damaging because they fail to follow an author's changed, and perhaps unconventional, mode of expression. But all critical editors take the risk of misunderstanding the authors they hope to serve well. (And they are not to be blamed—as Parker suggests on p. 12 and elsewhere—for trying to make sense of the texts they encounter: literature can be hard to make sense of.)


See G. T. Tanselle, "The Editorial Problem of Final Authorial Intention," SB, 29 (1976), 167-211; reprinted in Selected Studies, pp. 309-353.


An interest in such last intentions does not of course mean that there is no legitimate historical interest in earlier stages of a work.


He also attacks—a different matter—"the superstition that the author is infallible" (p. 51). Whether very many people have ever taken the position that the author could make no mistakes in revision is a real question, but is beside the point here. (Parker does understand that a "maimed text" can still retain power: see p. 8.)


Such as "Evidences for 'Late Insertions' in Melville's Works," Studies in the Novel, 7 (1975), 407-424; "Aesthetic Implications of Authorial Excisions: Examples from Nathaniel Hawthorne, Mark Twain, and Stephen Crane," in Editing Nineteenth-Century Fiction, ed. Jane Millgate (1978), pp. 99-119; and three articles by him and Brian Higgins: "Sober Second Thoughts: Fitzgerald's 'Final Version' of Tender is the Night," Proof, 4 (1975), 129-152; "Maggie's 'Last Night': Authorial Design and Editorial Patching," Studies in the Novel, 10 (1978), 64-75; "The Flawed Grandeur of Melville's Pierre," in New Perspectives on Melville, ed. Faith Pullin (1978), pp. 162-196.


See "The 'New Scholarship': Textual Evidence and Its Implications for Criticism, Literary Theory, and Aesthetics," Studies in American Fiction, 9 (1981), 181-197; and Brian Higgins and Parker, "The Chaotic Legacy of the New Criticism and the Fair Augury of the New Scholarship," in Ruined Eden of the Present, ed. G. R. Thompson and Virgil L. Lokke (1981), pp. 27-45.


Parker is not receptive to reader-response criticism: see, for example, pp. 219-220.


He admits that this discussion "will temporarily submerge the issue of reader response" (p. 94), but he believes that reader-response theory can finally contribute to the concept of authorial intention. Not surprisingly, he characterizes the kind of intention editors are primarily concerned with as "operative intentions," encompassing not only "the actions that the author, as he writes the text, understands himself to be performing in the text" (or "active intentions") but also "the immediate effects he understands these actions will achieve in his projected reader" (p. 99). (He unnecessarily elaborates his set of definitions by introducing "inferred intention" to "characterize the critic's description" of what the author is aiming to achieve, even though he recognizes that "operative intentions" are also necessarily inferred. "I am trying," he says, "to emphasize the inferential process an editor uses to posit a specific authorial intention." Surely this process need not be called an "intention" or brought into the definition of authorial intention at all. Mailloux's confusion is shown by his illustration of how the term is useful: it allows him, he says, "to talk about the operative intention an author claimed he had and a perhaps differing inferred intention an editor arrives at after evaluating all the relevant evidence." But the first of these is simply an author's "stated intention," which editors are always wary of, and the latter is the "operative intention" that the editor infers on the basis of all evidence, including the author's claims.) The upshot of his discussion hardly seems a revelation, because editors have regularly been able to understand, without being schooled by reader-response theorists, that authors use language conventions and therefore have expected their readers to react in certain ways to certain locutions. Explicitly defining the concept of authorial intention so as to include the "immediate effects" to be produced in the reader during the reading process is, perhaps, unobjectionable, but it is also unnecessary; it represents no shift in what has always been meant by the concept of the author's intention to convey a meaning through writing down particular words and punctuation. (Why a concept of authorial intention should imply a "specific rationale for critical interpretation"—a point made several times, as on pp. 108, 110, 112, 113—is puzzling.)


Or, perhaps more straightforwardly: "it is possible to reject some authorial revisions because they deface the text after it was finished" (p. 112).


Certainly those editors responsible for the wording of the CEAA emblem ("An Approved Text," not "The Approved Text") felt this way; but this gesture has not been taken as seriously as it deserves to be. Tom Davis, in a thoughtful review of the Lawrence and Hardy editions ("Textual Criticism: Philosophy and Practice," Library, 6th ser., 6 [1984], 386-397)—which begins with promise and descends to anticlimax—describes two models of "a text and its history": one emphasizes growing deterioration and corruption, the other a positive process of collaboration. Both, he says, are "equally 'true'" (p. 389).


As Randall McLeod puts it, "our tradition is unwilling to allow multiple textual authorities to rest as a simultaneous set of existential entities to be encountered absurdly by the reader" (p. 422 of "The Marriage of Good and Bad Quartos," Shakespeare Quarterly, 33 [1982], 421-431). He goes on to speak of "the infinitive text, which we may define as a polymorphous set of all versions, some part of each of which has a claim to substantive status." The recent controversy over the text of King Lear calls attention to the perennial competition between the attractions of a single text and the messier possibilities of independent versions. Although those who believe that the Lear variants provide evidence of Shakespeare's revision have overdramatized their case, there is no doubt that the possibility of authorial revision must always be in the editor's mind in evaluating variant readings. See Steven Urkowitz, Shakespeare's Revision of "King Lear" (1980); Gary Taylor and Michael Warren (eds.), The Division of the Kingdoms: Shakespeare's Two Versions of "King Lear" (1983); and Ernest A. J. Honigmann, "Shakespeare as a Reviser," in Textual Criticism and Literary Interpretation, ed. McGann (1985), pp. 1-22.


Fredson Bowers has recently summarized his views on apparatus in "Notes on Editorial Apparatus," in Historical & Editorial Studies in Medieval & Early Modern English, for Johan Gerritsen, ed. Mary-Jo Arn and Hanneke Wirtjes, with Hans Jensen (1985), pp. 147-162. (He and Paul Werstine have also had an exchange on historical collations that include readings from posthumous editions: Werstine, "Modern Editions and Historical Collation in Old-Spelling Editions of Shakespeare," Analytical & Enumerative Bibliography, 4 [1980], 95-106; Bowers, "The Historical Collation in an Old-Spelling Shakespeare Edition: Another View," SB, 35 [1982], 234-258.)


"The Synchrony and Diachrony of Texts: Practice and Theory of the Critical Edition of James Joyce's Ulysses," Text, 1 (1981), 305-326. As the title makes clear, this paper is intended to provide the rationale for the plan followed in Gabler's edition of Ulysses (1984); but it also includes illustrations from Faulkner and Milton and is meant to be a theoretical statement of wide applicability.


At one point, for example, he distinguishes between "the authoritative text free of corruption" and "the critically constituted text of final authorial intention established by bibliographically controlled editorial eclecticism" (pp. 310-311; see also p. 318).


A distinction should be made between a "genetic" text and the kind of "synoptic" text that Gabler is recommending. A genetic text aims to show the development of the text or texts present in a single document by providing a running text that indicates cancellations, interlineations, and other alterations. Gabler's synoptic text, on the other hand, aims to bring together in a single running text the authorial readings from all relevant documents. The symbols in the synoptic text, therefore, have to serve two functions: to indicate (as in a genetic text) the status of alterations within documents and also (as the sigla in a list do) to identify the various source documents and show their sequence. Furthermore, the synoptic text contains editorial emendations, for it is concerned only with authorial revisions, not with "corruptions"—which are therefore to be corrected in the synoptic text and recorded "in the type of subsidiary apparatus best suited to the purpose, i.e., an appended lemmatised emendation list" (p. 318).


In his edition of Ulysses he calls his "reading text" a "new, critically established text," consisting of "the emended continuous manuscript text at its ultimate level of compositional development" (p. 1903).


Issued as Occasional Paper No. 3 by the English Department of the Faculty of Military Studies, Royal Military College, University of New South Wales. A revised version will be published by the University of Georgia Press in 1986.


As he states this point elsewhere, "reading a single text of a work of art as if it adequately represented the work or was in fact the work may limit the reader's access to the whole work of art" (p. 71).


He does not, however, always give adequate recognition to the fact that critical editing is a form of literary criticism—as when he says that "textual critics tried to provide texts that would deserve the new scrutiny" of the New Critics (p. 7), as if the providing of a text is a pre-critical activity.


Shillingsburg forsakes his usual even-handedness here by calling lists that include nonauthorial variants "trivialized"; to the person interested more in the text that emerged from the publishing process than one reflecting authorial intention, these variants are not trivial.


Often it is not possible to identify the agent more precisely than "publisher's editor or compositor" or "someone other than the author."


Or, as he puts it elsewhere, editors should shift their "emphasis from 'the right text' to 'the whole work'" (p. 42).


If there were as little doubt about this kind of intention as Shillingsburg suggests, those editors whose aim is to establish authorial wording would not disagree as often as they do. Since manuscripts, as Shillingsburg recognizes, may contain "scribal error, Freudian slip, or shorthand elision" (p. 27), one must use critical judgment, based on an interpretation of intended meaning, to determine where these flaws occur.


Editors who talk about authorial intention do in fact generally mean an intended sequence of words and punctuation, but it is understandable that their discussions often contain references to intended meaning: the one is inextricable from the other. Recognizing this point in no way leads editors away from the active authorial intention they should properly focus on and toward such other levels of intention as "an intention to be brilliant or successful, or to write a novel or poem." But Shillingsburg oversimplifies the issue when he adds to this series: "or to convey an idea or emotion or attitude" (p. 26). Of course he is right to say that "several alternative texts of more or less satisfaction to the author" might convey the same meaning; nevertheless, the intention to convey a particular idea is not irrelevant to the intention to write down a particular sequence of words. Shillingsburg (p. 28)—like Mailloux (pp. 96-97) and Parker (p. 22)—complains that my concept of active intention in my 1976 essay (see note 59 above) does not encompass the author's intention in the process of composition. All I can say is that I meant for it to: when I spoke, for example, of "the intention of the author to have particular words and marks of punctuation constitute his text" (p. 182 [324]), I was assuming that such intention manifests itself continuously throughout the writing of a work (and often shifts, producing revisions—a problem I took up in the later parts of the essay). (I am aware of the irony: this situation illustrates the fact that a writer's statement of intention does not always match what readers find in the work.) Shillingsburg himself, though he emphasizes the composition process, is inclined to think in terms of "stages" (as on p. 24) rather than—more realistically—a continuous process, in which any pronounced points of stability that do exist grow out of numerous subordinate stages. Another writer who defines the intention relevant to editors as what the author "intended to write or what he intended to constitute his text" (p. 127) is James McLaverty, whose article on "The Concept of Authorial Intention in Textual Criticism," Library, 6th ser., 6 (1984), 121-138, is an unusually intelligent, if occasionally mistaken, treatment of the subject. McLaverty believes that every published version of a work reflects a different intention and that an editor must choose one version (not necessarily the "final" one) to edit (p. 130); but he does not understand that the editor can draw emendations from the texts of other versions without necessarily mixing versions, for in many cases the only documentary evidence of corrections in one version is the report of them embedded in the documentary text of the next version (cf. note 85 below). His reluctance to allow such eclecticism springs from a general tendency to wish to restrict the area in which editorial judgment operates. Thus he finds a way to argue against any correction of errors of external fact: an author's intention to be correct, he feels, is outside the scope of the kind of intention (to have a certain sequence of words and punctuation) with which the editor is concerned (p. 129). Critical editors, however, if their approach is to be truly critical, cannot eliminate any part of a text—such as references to external fact—from critical scrutiny, even though there will be no universal agreement as to which factual errors are, and which are not, expressed in words that the author intended to write. (See also my comments on "External Fact as an Editorial Problem," SB, 32 [1979], 1-47; reprinted in Selected Studies, pp. 355-401.) His complaint that no principle has been established for determining which versions of a work are so different as to require separate editions similarly represents a desire for rules that would replace judgments. In another interesting essay on intention, James E. May ("Determining Final Authorial Intention in Revised Satires: The Case of Edward Young," SB, 38 [1985], 276-289) attempts to identify those revisions in satires that "presuppose an audience and state of affairs different from those for which the work was originally intended" (p. 277) and to show that they (unlike the ones "compatible with the work's original conception" [p. 289]) should "be denied final authority," thus "preserving the historicity of the earlier version" (p. 277). Despite the repeated emphasis on "final authority," his work appears to recognize that both the original and the revised versions are of independent value as satiric works.


Even a text, which is "the actual order of words and punctuation as contained in any one physical form," is immaterial because the text "can exist simultaneously in the memory, in more than one copy or in more than one form" (p. 38). Shillingsburg's point is correct, if awkwardly stated—for if a single text can exist in more than one form (as it can), then it should not be defined as being "contained in any one physical form." A text is simply one particular sequence of words and punctuation (which may or may not accurately represent a version or a draft). James McLaverty has also recently explored the ontological status of literary works—with more sophistication than Shillingsburg—in "The Mode of Existence of Literary Works of Art" (see note 30 above), an essay that helpfully summarizes the debate over whether printed texts of literary works are scores (notational systems) or instances of the work. In his essay on intention (see note 81 above), he makes use of the idea that the printed text is a score (p. 127).


I think he is unwise, here and elsewhere, to contrast literary art with pottery, suggesting that a vase does not undergo revision in the process of its creation. ("There is a pleasing simplicity in the notion that texts grow or develop or are shaped toward a final form—rather like a potter shaping a vase on a wheel. But the analogy is misleading. A book does not come in final shape directly from the artist's hands like a vase" [p. 36]). All created works, however, are subject to revision during their creation. The reason we sometimes have versions and drafts of literary works is that the physical forms in which they are recorded are separate from the work itself, whereas the earlier versions of a vase are necessarily destroyed in the process of shaping the final version. If earlier versions are part of the total work, then in many cases the full work cannot survive, and the extent of what does survive is fortuitous. Some versions of literary works never existed on paper but only in the minds of their creators; and often writers who use word processors destroy all trace of earlier versions or drafts as they create new ones. The total work, inclusive of all versions and drafts, is probably never fully recoverable; but in the case of literary works (and other works of which versions are recorded on paper) there is the chance that some of the documents containing versions and drafts have survived. This point has not been adequately dealt with by those who hold that variants are integral parts of works.


As Shillingsburg concisely puts it, "Texts may contain non-authorial parts; versions do not" (p. 39).


A single set of revisions, for instance, could be spread over two documents, with the result that the text of neither document would reflect the entirety of a particular version or draft. When Shillingsburg describes an eclectic text as one that "mixes material from two or more versions" (p. 40), one wonders whether he is rejecting the usual definition of "eclectic" as referring to the mixing of material from two or more documentary texts or whether he is equating versions with the (sometimes defective) texts of individual documents. A further hint of this problem occurs in his inaccurate description of the "historical orientation," when he speaks of "points in creation when the text reaches stasis in a document" (p. 61). The stasis conferred by a physical document does not necessarily represent a point of stasis in the development of a work. (I call the description inaccurate because the value of reproducing the text of a document does not depend on whether a "documentary form is a complete record of the work at some stage in its development"; no justification is needed other than the fact that the document survives and is therefore a piece of historical evidence.) In his "Key Issues" article (see note 90 below), Shillingsburg deals more incisively with this matter, making clear that he does indeed wish to define "eclectic" as referring to the mixing of versions, not merely of documentary texts. But in doing so he is moving beyond what scholarly editing normally encompasses and into "creative" or "aesthetic" editing: "Eclecticism," he writes, "does not attempt to conform to a single ascertainable version ever existing at one time even in the author's intention (or conception)" (pp. 7-8). If the term "eclectic" is to be applied to scholarly editing at all, it must refer to the mixing of documentary texts to produce a postulated version. Shillingsburg does recognize in his article that the integrity of a version—being an abstraction—is not tied to an individual documentary text; but that recognition should also have caused him to see that an editor who mixes readings from different texts in an effort to produce a text conforming to the author's intentions at a given time (and not to produce simply what the editor likes better) is not mixing versions but rather is reconstructing one.


It is rather puzzlingly called "Forms." The opening of the chapter suggests that forms are "details of presentation" as opposed to "substance" (p. 13). But the subject matter treated is certainly not limited to "accidentals."


Shillingsburg admits that "all violations of documentary historical forms (including Bowers' and McGann's) are supported by appeals to one or more competing formal orientation which is seen to take priority over the historical, even if only in some limited way" (p. 15). In that case, if the "orientations" are meant to represent the basic positions that are often mixed in practice, why not define the "historical" approach as simply the diplomatic (or noncritical) reproduction of documentary texts?


Despite the fact that Shillingsburg says in his introduction that he is dropping "from further consideration" what he calls "commercial editing and copy-editing" (p. 4).


Shillingsburg says that "an editorial concern for the 'best' text is always an appeal to an aesthetic orientation" (p. 16) and at another point refers to "readings from several versions selected on the basis of the editor's aesthetic values in his attempt to present the author's 'best' text" (p. 10). But these statements are not helpful because they do not make clear that the best text for each of various historical purposes requires the editor's aesthetic judgment to be put to the service of deciding what someone in the past would have preferred.


Because Shillingsburg's book represents a fuller working out of the ideas he had earlier set forth in "Key Issues in Editorial Theory," Analytical & Enumerative Bibliography, 6 (1982), 3-16, it is not necessary here to discuss that generally less satisfactory essay. Some passages in it, however, are clearer than corresponding ones in the book (see note 85 above).