Praised as an epochal scholarly event and denounced as a scandal, the critical and synoptic edition of James JoyceТs Ulysses first published in 1984, together with the corrected text that was published separately in 1986, has received extraordinary publicity for a work of its kind.1 Its editing procedures have lifted the general public, students, literary critics, and scholarsЧthe vast majority of whom are not themselves editorsЧto a heightened awareness of textual editing. With readers now beginning to realize that editions should be scrutinized and assessed as carefully as interpretations have always been, users of the 1986 reading textЧwhich in this new printing remains available worldwideЧneed to be aware of how Hans Walter Gabler, supported by an international team of collaborators and advisors, arrived at its text and of how this edition resembles and also differs from others that might be produced. This is crucial now that the copyright protection for the first-edition text of Ulysses has expired in most of the world and will end soon in the United States, with the result that many editions are becoming available.
When dealing with a scholarly edition, readers should know something about the theoretical assumptions behind it and about the procedures that were adopted to produce it. On the face of it, accomplishing the goal of offering a text of a work that is more accurate than any that have appeared before might seem fairly simple: find out what the author wanted, clear away the errors, and you have it. But authors are rarely so cooperatively tidy: they change their minds; they destroy or discard documents once they have moved beyond them; they make changes in person, by phone, or via e-mail. Then other people get involved: a typist types, or a printer sets, something different from what the author wrote; a publisherТs editor changes the text, with or without the authorТs consent or sometimes with the authorТs active encouragement. Moreover, determining the order and relative importance of the surviving documents can be complicated. Is one edition earlier or later than another? Was the author involved at all in a particular editionТs production? Because of gaps in the available evidence and of inconsistencies or other complications in the surviving evidence, an editor needs a theoretical approach to the task and a set of procedures that follow from the assumptions.
The critical and synoptic edition of Ulysses needs to be understood in terms of the assumptions and methods of most Anglo-American editing today, because it both follows them and departs from, even challenges, them in important ways. In the method that has come to dominate Anglo-American editing, an editor studies all the relevant surviving documents for the work in question and selects one version as the copytext. The documents include any notes, drafts, manuscripts, typescripts, and proofs that are extant, plus printed versions in which the author was involved. The copytext, usually the first edition or, if available, the authorТs manuscript, is the basic text that the editor will follow for such matters as spelling, punctuation, etc., in places where the evidence is inconclusive, and for all the words except when differences between documents indicate authorТs revisions and so call for the editor to alter the copytextТs words on the basis of one of the other documents. In the terminology of editing and textual criticism, the words are called Сsubstantives,Т spelling and punctuation are matters of Сaccidentals,Т inconclusive readings are СindifferentТ ones, and the editorТs alterations of the copytext are called Сemendations.Т
The resulting text, eclectically blending authorial corrections and revisions with the system of accidentals from the copytext, was eventually epitomized as fulfilling the authorТs final intentions. This method of copytext editing producing an eclectic text offers the editor a way of dealing with gaps in the historical record and with seemingly equal choices among variant readings (when in doubt, follow the copytext). It strives as well to rescue the authorТs text from its ravagement through time at the hands of the scribes, typists, publisherТs editors, and printers who were allowed to alter, and presumably corrupt, it. But it also tends to suppress the historical determinants that originally affected the work and its production in the name of the authorТs final intentions because the eclectic edited text is an idealized construct that appears to transcend time by recreating the СpurityТ of the authorТs isolated conception. The editor is also able to disappear behind the author, since the edition will likely be presented as the authorТs (the editor fulfilled the authorТs intentions) rather than as the editorТs (the editor started with some basic premises and made many decisions and choices in order to produce the edited text).
It is easy to disappear behind the towering figure of James Joyce but difficult to adopt a more visible editorial stance that reveals the editor, as well as Joyce, at work. Yet for Hans Walter Gabler as an editor, JoyceТs methods of writing Ulysses and the surviving evidence regarding that work called for a visible stance. An astonishing array of materialsЧespecially prepublication documentsЧhas survived; they open up the whole process of JoyceТs composition of the work for the purposes of editing, but at the same time they leave tantalizing and important gaps. Joyce wrote Ulysses episode by episode, and the process is almost entirely one of growth and expansion. After compiling notes and rough drafts, Joyce brought each of the eighteen episodes to a temporary finish in a final working draft that he gave to a typist. For eight full episodes and part of a ninth, Joyce apparently made a fair copy of this draft, making some changes as he went along; for these pages the working draft has not survived. (The surviving manuscript, partly the final working draft and partly JoyceТs fair copy, is called the Rosenbach Manuscript after the museum that owns it.) Each episode was transcribed by a series of typists and printers, and some sections were set in proof as many as eight or nine times. Joyce often added to the text as he read and corrected the latest transcription, but as he corrected each transcription he seems not to have looked back to the original manuscript. In addition, as he revised and corrected the proofs in 1921 for the book publication, he was often working on two or three episodes at the same time, reading proofs for early episodes, for example, at the same time as he was drafting the later episodes of СIthacaТ and СPenelopeТ. The printers had to reset much type again and again because of the huge number of JoyceТs corrections, revisions, and additions, and they worked under very short deadlines as they approached the publication date that Joyce wantedЧFebruary 2, 1922, his fortieth birthday.
Gabler decided at the beginning of his work that traditional copytext methods would not work well for the textual situation that Ulysses presents. At least three factors led to this decision: the manuscript, which does provide a beginning-to-end version in JoyceТs hand, is too far removed from the extensively augmented text that Ulysses eventually became; the typescripts and proofs are steps along the way in the process of expansion; and the first edition is too filled with errors.2 Gabler looked to German genetic editing, which is oriented more towards authorial revision than towards transmissional corruption, and also to Fredson BowersТs work within the copytext-editing tradition on constructed, or what Gabler calls СvirtualТ, documents as copytext. Bowers demonstrated that a copytext can be a lost or virtual document when he edited Stephen CraneТs stories and Henry FieldingТs Tom Jones. In the case of Crane two surviving versions of a story that each descend directly from a lost original were used to recreate the lost original document, and the recreated document served as copytext. For Fielding the accidentals of one document (the first edition of Tom Jones) were merged with the substantives of another (the fourth edition), and this constructed hypothetical document became the copytext. In the implications of these examples Gabler saw a way of meeting the challenge of the complex textual situation presented by Ulysses. He reasoned that JoyceТs activity on the prepublication documents from the final working draft through to the final page proofsЧhis manuscript inscription plus all the additions to the typescript and proofsЧcan add up to a manuscript of the whole book, even though a virtual one. In one of the editionТs major innovations he reconstructs this virtual manuscript, calling it the Сcontinuous manuscriptТ of Ulysses, and uses it as the editionТs copytext.
In assembling the continuous manuscript the editor makes an important distinction between a СdocumentТ and a СtextТ. If an author writes out a story and then returns to its pages twice to revise it, there is only one document but three distinguishable texts (the original story, the original with one set of revisions, the original with both sets). Most of the extant documents for Ulysses contain more than one text. Typical cases are manuscripts with JoyceТs handwritten text plus his subsequent revisions, typescripts with the typed version plus one or two rounds of correction and revision in JoyceТs hand, and proofs which contain the printed version plus JoyceТs corrections and revisions. When documents are missingЧas is the final working draft where it served as copy for the typist; three chapters in typescript where the extant fair copy served as copy; and small sections of the proofsЧthe surviving documents can serve as evidence for the text contained on the missing ones that immediately preceded them. For example, the first set of proofs will show as printed text the material that Joyce presumably entered in his handwritten additions on a lost typescript page. Thus, an editor can reason that while the documents may not all survive, all the text is recoverable either through direct evidence or through recreation by extrapolation from the extant documents. The Сcontinuous manuscriptТ (the conflation of all JoyceТs handwriting on the manuscript and in the corrections and revisions on the typescripts and proofs) remains a virtual document, but the continuous manuscript text can thus be created as something now real. This is what Gabler has done. What he calls the Сsynoptic textТ presents, on the left-hand pages of Ulysses: A Critical and Synoptic Edition, the construction of the continuous manuscript text. The synopsis is accomplished through an elaborately coded system that indicates all JoyceТs revisions, additions, and deletions, including the stage at which each change occurred. Also, the editorТs decisions, including his transcriptions of the manuscript and his choices as to which of JoyceТ variant words (when there is more than one possibility at a particular place) are retained in the continuous manuscript text, are on display for all readers to observe and to assess.
It is accurate to say that the continuous manuscript text was assembled or created or recreated or constructed, but it was not created by copytext editing. Once it was assembled, the continuous manuscript text became the copytext for Gabler's edition, but until it came into existence there was no copytext at all. (Much of the discussion of the edition has been confused on this matter, assuming incorrectly that the Rosenbach Manuscript or the typescript is the copytext for the continuous manuscript text.) Gabler constructed the continuous manuscript text, in his words, Сas Joyce wroteТ Ulysses. This means building the text up, stage by stage, from the working draft towards the goal of the first edition text as it would have appeared had no mistakes been made. The straightforward reconstruction becomes complicated when documents are missing or when something Joyce wrote on one document was not typed or printed and so not transmitted through the production process.
Like almost all editors, Gabler dismissed as a working principle the concept of Сpassive authorization,Т the idea that because Joyce left an error standing or did not restore a reading as he read proof he must have wanted the resulting reading in the book. But to accompany such a rejection an editor needs procedures to help determine when to accept a reading that the author did not restore and when not. Two procedures are especially important. First, any text handwritten by Joyce was presumed to be authoritative and hence admissible into the text unless it could be proved to be faulty. Conversely, any transmitted (typed or printed) text was considered to be potentially faulty unless it proved to possess authority. Second is GablerТs Сrule of the invariant context,Т which means that a word or passage from an earlier document could be admitted into the text only if the context around it (words, sentence, paragraphЧthe scope of the context varies from example to example) underwent no change, thus remaining invariant. These procedures have led to an important, and controversial, aspect of the edition. Several words and passages appear in the Rosenbach Manuscript but presumably not in the final working draft that was used by the typist and is now lost; these words and passages thus were never typed or printed. When Gabler judged them to be JoyceТs revisions as he made his fair copy of the working draft and when their context remained invariant, he admitted them into the continuous manuscript text on the grounds that they represent the fullest development of the text. Some examples will be given later on.
One further note about the continuous manuscript text: it was not constructed in order to fulfill what is known as Сauthorial intention.Т Gabler's phrase, СUlysses as Joyce wrote it,Т refers to JoyceТs activity as he created Ulysses both in the extant documents and by inference from those documents to the lost ones. The editor studied what Joyce did, not what the editor thought Joyce meant or intended. This makes his edition one oriented towards the text (the authorТs text in this case more than, say, the published text) but not towards intention. The framework of genetic editing supplies editors with a set of premises and methods in which an edited text is built from the ground up with each stage considered as a version, a distinguishable self-contained text that does not need to be justified in terms of the author and the authorТs intentions. The variants between one version and the next are seen not as errors to be corrected but as revisions in a changing text. On the whole, the variants in a many-layered manuscriptЧsuch as the extreme example of the continuous manuscript of UlyssesЧthat will go together to form each identifiable version will be self-evident from the process of the writingТs development. But enough instances of alternatives usually remain where the editor must exercise critical judgment. The grounds for this judgment can be procedural ones, such as the priority given to JoyceТs own inscription or the rule of the invariant context that determined whether a reading was marked as valid or deleted in the continuous manuscript text, or they can be decisions that the editor had to make on the basis of his understanding of the kinds of revision Joyce was likely to make at the pertinent stage of his work on Ulysses.
Only after the continuous manuscript text was assembled did copytext editing come into play, as the continuous manuscript text was then emended, like any other copytext, as a result of the editorТs comparison of it to the other prepublication documents and to the few postpublication documents in which Joyce was involved (primarily errata lists that he helped to prepare and corrections for the 1937 reprint of the 1936 Bodley Head edition). Since most of the collation was done to construct the patterns of writing and revision in the continuous manuscript text in the first place, the copytext editing was largely confined to eliminating errors of transmission and to emending accidentals. Again, it was not done primarily to fulfill final authorial intentions.3 The copytext editing of the continuous manuscript text is indicated in the footnotes to the synoptic textЧthe presentation of the editorТs assembly of the continuous manuscript textЧin Ulysses: A Critical and Synoptic Edition.
The critical edition of Ulysses set as its arbitrary goal the creation of a parallel text to the historical first edition, one that ideally represents the first edition without errors. (Of course, nothing is ideal, and the 1984 edition inadvertently introduced a few errors of its own.) Such a goal was a pragmatic, and not a logically necessary, one; the assembled continuous manuscript text could have stood as the editionТs text. As it is, GablerТs edition offers as the parallel text to the first edition text the assembled and then copytext-edited continuous manuscript text, as displayed on the left-hand pages of Ulysses: A Critical and Synoptic Edition with its system of diacritical codes showing the editorТs assembly, and footnotes revealing his emendation, of the continuous manuscript text. A further extrapolation (again the result of a pragmatic decision on the editorТs pan), offered on the right-hand pages and in this printing, is the editionТs reading text, which comprises the synoptic text without any of its words or punctuation in full or angle brackets (those deleted or changed by Joyce), its diacritical codes, or its footnotes. Episode and line numbers in this printing correspond to those on the right-hand pages of Ulysses: A Critical and Synoptic Edition.
A passage from the СLestrygoniansТ episode (8:654-67; pp. 138-39 in this printing) provides a good, and much-discussed, example of how the continuous manuscript text was assembled (the synoptic text is in volume 1, p. 356, ll. 10-24 of Ulysses: A Critical and Synoptic Edition).
(B)[Squatted] Perched(B) on high stools by the bar, hats shoved back, at the tableso calling for more breado no charge, swilling,
The final working draft for СLestrygoniansТ is lost, so the earliest extant document is the fair copy on the Rosenbach Manuscript. The original text of this passage reads there, СSquatted on high stools by the bar, hats shoved back, at the tables calling for more bread no charge, swilling, chewing gobfuls of sloppy food, their eyes bulging, wiping wetted moustaches. A man with a napkin tucked round him spooned gurgling soup down his gullet. A man spitting back on his plate: gristle: no teeth to chew it. Chump chop he has. Sad booserТs eyes.Т Subsequent revisions and additions changed and augmented the text, with letters B, C, and D indicating, respectively, JoyceТs revisions to the lost final working draft as indicated by the typed text on the extant typescript, the first round of revisions to the typescript, and the second round of typescript revisions. (Letters in parentheses indicate reconstructed text on documents that have not survived.) The numbers indicate the revisions on each subsequent setting in proof. Full brackets show JoyceТs deletions or changes, as in the revision of the manuscriptТs СspoonedТ to СshovelledТ in the second round of typescript revisions (l.15). Carets indicate additions within a single stage, such as JoyceТs addition of СinfantТsТ between СaТ and СnapkinТ on the manuscript (ll. 14-15) or of СSomething galoptious.Т as an addition-to-an-addition on the first set of proofs (l.23). When combined with angle brackets, carets show a revision, as when Joyce revised СchewingТ to СwolfingТ on the manuscript itself (ll. 11-12). The synoptic presentation of the continuous manuscript text is thus an assemblage of inclusion: JoyceТs deleted and superseded readings, as well as those that remain in Ulysses, are all part of it.
The superscript circles in the synopsis point to the footnotes (not reproduced here), where the editor has recorded his editorial emendations to the continuous manuscript text. For example, at l. 14, he emended the manuscriptТs СaТ to СanТ preceding СinfantТs napkinТ on the basis of his conjecture of JoyceТs activity on the lost final working draft, the text on the surviving typescript providing the evidence. The edited text differs from all earlier editions of Ulysses in one place: the word Сgums,Т with the subsequent colon (l. 17 of the synopsis and l. 660 of the reading text), is restored to the text for the first time here.
The presence or absence of СgumsТ might seem like a minor matter, but it is indicative of all the decisions involved in editing Ulysses. The editor admitted the word into the continuous manuscript text, and it became part of the edited text, on the basis of its appearance in the serialized version of СLestrygoniansТ in the Little Review; he argues that its appearance there is evidence that Joyce added the word onto a lost typescript page. The wordТs appearance here is consistent with GablerТs procedures. In a review of the edition, Jerome J. McGann made the important observation that СgumsТ is correct here but that an edition that follows other principles would be equally correct without the word. This word can stand for the many that appear in GablerТs edition, often for the first time in printed versions of Ulysses, because of his editorial principles and the consistent application of the procedures that follow from those principles.
Several examples can indicate how the editor arrived at particular readings and also how other editions might read differently. First, on the opening page of this edition, Buck Mulligan calls СoutТ to Stephen (l. 6) and blesses the СlandТ (l. 10), whereas in earlier editions he called СupТ and blessed the Сcountry.Т In both cases, the editor follows the Rosenbach Manuscript (which here was the typistТs copy) and reasons from a bibliographic analysis of the transmission text that the typed СupТ and СcountryТ were unauthorized departures from JoyceТs text. In the first case, he additionally surmises that the typist was looking ahead to СCome up, Kinch!Т in the following line. Likewise, in this edition the telegram that Stephen Dedalus recalls in СProteusТ reads, СNother dying come home father.Т (3.199), whereas earlier editions show the first word as СMother,Т more correct but failing to image the curiosity of the telegramТs orthographic error. The editor follows JoyceТs inscription of СNotherТ on the Rosenbach Manuscript (again the typistТs copy), which Joyce insists on once more in his revisions to the first set of proofs, and rejects the reconstructed typed text on the lost typescript and the СcorrectionТ to СMotherТ entered in a hand other than JoyceТs on the fifth and final set of proofs. The best known passage in this edition that is not part of any previous printed edition of Ulysses is the so-called СloveТ passage in СScylla and Charybdis.Т In the middle of his discussion of Shakespeare, Stephen asks, СЧWill he not see reborn in her, with the memory of his own youth added, another image?Т and then thinks, СDo you know what you are talking about? Love, yes. Word known to all men. Amor vero aliquid alicui bonum vult untie et ea quae concupiscimus . . .Т (9.427-31). The passage is in the Rosenbach Manuscript; the final working draft used by the typist is lost. Gabler reasons that the working draft did not differ from the surviving fair copy at this point and that the typist skipped from one ellipsis at the end of an underlined passage indicating italics in the line before StephenТs question (the line ends СLТart dТкtre grandp….Т)to a similar nearby ellipsis after another underlined passage (StephenТs Latin thought ending with СconcupiscimusТ),thus omitting StephenТs question and subsequent thought. In each case, and in the case of СgumsТ as well, the editorТs justification for his choices was textual and bibliographical, not critical; none of these examples presented a problematic or ambiguous textual situation. It is important to note, though, that an edition prepared under other assumptions (for example, one privileging the transmitted text over the written one) might in each case choose the reading that this edition rejects.
These few details are part of the large system that makes up any editing project. The full system includes not only the editorial assumptions and procedures that are visible in all the particular readings but also responses to broader questions about the nature of literary works and their texts, the relationship of the author to the work, the role of the editor, and the nature of authority in an edition. In being a text-based, rather than an author-based, edition; in its use of genetic editing theories and methods; and in its synoptic presentation, this edition of Ulysses offers an alternative to dominant Anglo-American methods of editing that questions and challenges the accepted paradigms. As Gabler has acknowledged, the edition can be discomforting.
Along similar lines, Jerome McGann in his review claimed that the edition Сraises all the central questions that have brought such a fruitful crisis to literary work in the postmodern periodТ and suggested that it should be Сa required object of study for every scholar working in English literature.Т As an object of study, GablerТs workЧhis assumptions and his proceduresЧcan be discussed and debated, but, as Vicki Mahaffey has noted, the controversy that erupted over the edition deflected the kind of questioning that McGann envisioned. Specific details were discussed apart from their relationship to the editorТs basic assumptions and methods as a whole. More important, as Mahaffey argues,
many of the most widely publicized attacks are based on premises about textual editing that the general reading public takes for granted, so that when a critic proves that Gabler has violated these guidelines, his editorial competence is implicitly or explicitly called into question. It takes a reasonably specialized reader to realize that the weakness of such arguments, which seem logically convincing on their own terms, is at the level of the premise, since Gabler does not share many of the premises on which the critique is based.
GablerТs loudest and most persistent critic, John Kidd, has since 1988 steadily and relentlessly attacked the edition. With a great deal of rhetorical flurry and a few oft-repeated examples, Kidd captured a great deal of attention. But all his pages of supposed analysis, and the sixty pages of tables and charts of GablerТs alleged errors and inconsistencies in his СInquiryТ into the edition, managed finally to demonstrate only two errorsЧmistranscriptions of the names СBullerТ at 5.560 and СThriftТ at 10.1259Чand to point to one reading that resulted from the editorТs inconsistency in following his editionТs own stated rules of procedure. The passage in questionЧdiscussed in GablerТs СNote on the TextТЧis at 16.1804-5: Сwas not quite the same as the usual handsome blackguard type they unquestionably had an insatiable hankering afterТ should be Сwas not quite the same as the usual blackguard type they unquestionably had an indubitable hankering after.Т In this instance, the editorТs diminished attention to the rule of the invariant context and his mistaking of an authorial revision based on a transmission error for a mere correction led him astray. The items on KiddТs long lists can be checked individually and will possibly lead to exposure of other errors or debatable readings or decisions, but the tables are constructed so capriciously and idiosyncratically, with so little demonstrated understanding of GablerТs theoretical assumptions and procedures, and with no coherent or consistent indication of KiddТs own working assumptions that they can point to errors or misjudgments only by accident. KiddТs campaign forced a great deal of negative attention on this edition but has ultimately revealed very little at all about it. It is to be hoped that the kind of inquiry that McGann and other critics have called for can now come to the forefront.4
Such an inquiry is possible because, like any responsible editor, Gabler discussed his editorial procedures and laid out his decisions fully in Ulysses: A Critical and Synoptic Edition. He defines a Сcritical editionТ by Сthe complex interdependence of a text established from the ground upТ as opposed to marking up and correcting an existing text Сand its interfacing apparatus.Т Many different kinds of critical editions are possible, including a copytext edition or a different kind of nontraditional edition, but for all of them the text itself constitutes only one part. Equally essential is the apparatus, which acknowledges the hand of the editor. Readers should be extremely suspicious of any edition that presents itself as a reading text without an apparatus spelling out all its editorТs assumptions and decisions.
Anyone wishing to follow the logic and procedures that produced the readings in this edition, in other words to listen to the editor speaking as editor, is strongly urged to use the line numbers here to find the corresponding passage in Ulysses: A Critical and Synoptic Edition, with its synoptic text on the facing left-hand page. Likewise, more detailed explanations of GablerТs assumptions and procedures are available in his Afterword to the three-volume edition and in his articles СOn Textual Criticism and Editing: The Case of JoyceТs UlyssesТand the more extended СWhat Ulysses Requires.Т
Michael Groden
August 1993