The first publication of Ulysses was conscious of its mistakes, as is evident in Sylvia Beach's caveat: "The publisher asks the reader's indulgence for typographical errors unavoidable in the exceptional circumstances." These "exceptional circumstances"--among them Joyce's deteriorating eyesight, selling his manuscript piecemeal before its publication to make ends meet, and the disastrous contaminations introduced by non-English-speaking typists in Dijon and led by a 'correcting' foreman (Dalton 108; Driver 33)--applied not only to the publication of Beach's edition, but to the creation and development of Ulysses itself. Later editions would include a "definitive" edition printed by Odyssey Press in 1932, which standardised many of Joyce's deliberately unique accidentals; a pirated and error-ridden edition in the United States that accidentally became copytext for the first legal American edition in 1934; and a "corrected and reset" edition published in 1960 (1961 in America) that introduced many errors. The general frustration caused by these less-than-perfect editions culminated in the "Joyce Wars" of the late 1980s, sparked by the "corrected" 1984/86 Garland edition.
1 Even though Beach's cautionary statement was only included in Shakespeare & Co. editions, it should be included in every text of Ulysses. The entire history of the publication of Ulysses is the history of transmitted textual error.

Yet, even if all non-authorial contamination were removed, Ulysses would remain a complicated text. Joyce spent seven years writing and revising Ulysses and the extant pre-publication documents do not make clear which authorial textual variants Joyce preferred. When he consistently duplicated a variant, it is clear he must have preferred it; when he was inconsistent (as was common), or when one of the eighteen episodes exists in more than one form in collateral documents, it is not so clear. This, I would argue, makes it a protean text, one capable of taking many different forms but having no central apparatus or preferred structure. Just as Stephen's thoughts constantly shifted and changed on his walk along sand-shifting Sandymount Strand, so too is Joyce's text multifaceted, "a maze of dark cunning nets" (84U 3:154; 61U 41). In order to understand the complex elements that make Ulysses an unavoidably protean text and what problems this poses for editors who seek a 'corrected' text, it is helpful to examine the complications encountered by the Gabler edition, the most recent attempt at correcting Ulysses.

In 1977, Hans Walter Gabler, with the backing of the Joyce estate, became general editor of a new edition of Ulysses to be published by Garland. Well aware of the difficulty involved in correcting Ulysses, Garland and the Joyce estate hired not just Gabler but a host of others, hoping to ensure that this edition would not fall prey to the same fate as previous editions. The Joyce estate further stipulated that the editorial team--composed chiefly of Gabler and University of Tübingen graduate students Wolfhard Steppe and Claus Melchoir, and aided by Harald Beck, Walter Hettche, John O'Hanlan, Danis Rose, Charity Scott Stokes, and Kinga Thomas--must meet the approval of an "academic advisory committee." This committee consisted of two leading Joyce scholars, Richard Ellmann and Clive Hart, and eminent textual critic Philip Gaskell and was "assisted by" A. Walton Litz and Michael Groden.

The necessity for such a team of experts was clear from the first editorial decision: what would this edition use as copytext? As mentioned above, any previously printed edition was obviously problematic. Yet, as Sylvia Beach had learned, Joyce's manuscripts are problematic in their own right, and the team soon realised that using the manuscript as copytext was simply to exchange one set of problems for another. Joyce's creative process was such that he rarely deleted material after a certain point; he only added more layers (Johnson li). Of the eighteen episodes that comprise Ulysses, only three ("Cyclops," "Circe," and "Eumaeus") ever achieved fair copy status; the rest are only available in draft form (Gabler 1984 "Afterword" 1,880). This means that many episodes were revised in numerous collateral documents, as when Joyce made revisions to his own drafts that were not consistent with revisions made to typescripts, which included errors he did not always correct. Finally, if matters are not yet thorny enough, Joyce's poor eyesight meant he was subject to incorrectly copying his drafts when he acted as his own scribe, making it very difficult at times to discern whether variants created in this process were purposeful or unintentional. As a result, the manuscript is a palimpsest with significant lacunae and is littered with inconsistent revisions.

With these challenges in mind, the Gabler team settled on a compromise between two common editorial approaches, using the earliest extant manuscript (the Greg-Bowers method) and using the first printed edition that was corrected by the author (the Thorpe-Gaskell approach): they would construct a text using all pre-publication material and the 1922 first printed edition (Shakespeare & Co.) and create a text based on a hypothetical lost working draft of Ulysses (Gabler 1984 "Afterword" 1,895). This pre-publication material includes the Rosenbach Manuscript (a holograph copy Joyce composed between 1917-21), extant typescripts, serialisations from The Little Review and Egoist, and proofs for the first edition (Ibid 1,896). Gabler deemed this genetic text the "continuous manuscript text" and claimed this text to be his copytext (Ibid 1,896).

Having established a copytext, Gabler and his team then set out to emend it in order to produce a reading text clear of the diacritical marks and multiple variants of the genetic text. Emendations were considered individually so that the editors could prudently and judiciously examine every varying moment in its proper context (Ibid 1,899-1,900). Generally speaking, however, "the final state of the text's development is considered reached when it is last fully and correctly written out in the author's hand" (Ibid 1,901). Gabler confidently proclaimed his edition to be exempt of all non-authorial contamination, representing "Ulysses as Joyce wrote it" (Gabler 1986 "Afterword" 649).

When Garland finally published Gabler's edition in 1984, the gigantic project had given birth to a hefty three-volume, 1,919-page edition which featured the genetic text on verso pages and the reading text on the rectos. Initially, this edition was received with great applause
2 and indeed many long-standing errors were corrected for the first time.3 Yet not everyone was pleased with the edition. In April 1985, not yet a year since its publication, the enthusiasm surrounding the new edition slowed significantly when John Kidd presented "Errors of Execution in the 1984 Ulysses" to the Society for Textual Scholarship. By the summer of 1985, many renowned Joyce scholars--among them former academic advisors to the Gabler team Richard Ellmann and Clive Hart--convened at the Princess Grace Irish Library in Monaco and each voiced different objections to the 1984 edition.4 The summer of 1985 also saw the publication of Jerome McGann's "Ulysses: A Postmodern Text," which neither denounced nor applauded the reading text nor the methods used by Gabler and his team.5

Ultimately, it was Kidd's grievances that attracted most attention. In 1988, Kidd published his scathing dithyramb "The Scandal of 'Ulysses'"
6 in which he cited fifteen specific moments in the reading text of Gabler's edition which he felt unfaithfully represented Joyce. Kidd's most critical indictment was more general: he vehemently objected to the editors' reliance on reproductions of manuscripts and archival material rather than the original documents. Specifically, Kidd's objection was, to use Gabler's own words, that "[a]ll variants...were checked against the original documents or document reproductions in facsimile (for the Rosenbach Manuscript) and photo reprint (The James Joyce Archive for scribal transcripts, typescripts and proofs)" (Gabler, 1984 "Afterword" 1,906). Why no one from the Gabler team checked the variants of their edition against the original Rosenbach Manuscript or original typescripts and proofs can only be interpreted as a major oversight.7 In later attacks, Kidd pointed out that the Gabler edition does not remain faithful to the bibliographic code that Joyce had been sensitive to in the Shakespeare & Co. editions, such as ending the daylight portion of the novel on page 365 (the number of days in a year) and whittling Molly's letter to Bloom on her fifteenth birthday (which fell on the fifteenth of the month) down to fifteen sentences. Nowhere in his Foreword or Afterword does Gabler explain or acknowledge a conscious decision to remain faithful to Joyce's linguistic code while ignoring the bibliographic element. 8

Exchanges between Kidd and Gabler and their respective supporters continued in The New York Review of Books and the Times Literary Supplement for months,
9 and Kidd's fiery attack ignited the passions of Joyce scholars and aficionados for years afterward. Still, Kidd was not the first to express concern with Gabler's edition: as mentioned earlier, Hart and Ellmann first publicly expressed their reservations at the 1985 Monaco conference. Each had concerns that were hallmarks of a protean text. When Gabler showed Ellmann a 'lost passage' he'd recovered from an early draft in Rosenbach, Ellmann was glad to see that it confirmed his hypothesis that love was the "word known to all men"; yet Ellmann also felt "grave doubts" about including the passage in the reading text because it "is most uncharacteristic of Joyce" and was likely purposefully removed from the text by Joyce in later drafts (Ellmann "Crux" 28-34). Hart expressed similar concerns that Gabler's reading text was "an incoherent amalgam of so many different Joyces" (Hart 64).

Ellmann, Hart, and others involved with the Garland edition privately harbored serious misgivings during the production of the new edition which remained private until Charles Rossman published "The New 'Ulysses': The Hidden Controversy" late in 1988.
10 Reconstructing material preserved in Ellmann's papers,11 Rossman found that Gabler's uncompromising insistence on strictly adhering to his editorial method was often at odds with Gaskell and Hart, who frequently opposed the method (Rossman 55). A calamitous meeting on June 4, 1983 revealed that the differences between Gabler and the advisory board could not be reconciled. The Joyce estate, which wanted to produce the best possible text of Ulysses and renew its copyright, was forced to shift its position on editorial decision-making to try to circumvent the impasse: now Gabler, not the advisory board, had final say in editorial decisions and the imprimatur of the advisors was no longer necessary (Rossman 56). The meeting and its consequences precipitated the resignation of both Gaskell and Hart in June; it was months before Ellmann convinced the two men to return (Rossman 56-7).

The controversy and debate caused by the Gabler edition reminds us of two very important lessons, which can be easily understood in the following example, concerning the debate surrounding the removal of a comma in "Eumaeus." First, the sentence as it appeared in every edition before 1984, with the contested comma (following "tiger") intact:

Nine tenths of them all could be caged or trained, nothing beyond the art of man barring the bees. Whale with harpoon hairpin, alligator tickle the small of his back and he sees the joke, chalk a circle for a rooster, tiger, my eagle eye. (61U 662)

In Gabler's edition, the final comma was deleted, thus concluding the sentence, "tiger my eagle eye" (86U 16:1795-6). Kidd argued that the last four words must be interpreted as an allusion to Dante's jest to Stephen in A Portrait of The Artist, when she quips "eagles will come and pull out his eyes" (P 6) if he does not apologise. Removing the comma, Kidd asserts, simply because it does not appear in a single "earlier draft--one as messy as any other," complicates this allusion and is "an ill-considered conflation of drafts" (Kidd 36). John O'Hanlan, member of the Gabler editorial team, defends the decision to remove the comma because it has nothing to do with Dante's comment or Portrait; it is simply a moment in which Bloom ruminates about mankind's power over animals. The sentence is a list of folk wisdoms regarding animals, the last four words of which indicate that staring down a tiger will cause him self-consciously to turn away. Adding a comma after "tiger" is to break up the phrase, make it anomalous to the three previous, and hamper its meaning (O'Hanlan 80).

The lessons from such an exchange are clear: in the first place, editorial theory is inextricably linked to hermeneutic theory. Editorial theory concerns an interpretive effort that provides the practice of choosing variants, but when more than one variant is authorial, the decision falls upon the editor, and the interpretive power of hermeneutics (hence Garland's academic advisory committee). Neither argument posed by Kidd or O'Hanlan is 'correct' nor 'incorrect,' they are simply divergent readings of the protean text; they are hermeneutic interpretations that prefer one face of the text to another, yet neither face is more valid. This leads to the second point: each variant is only as important as the other variants it stands to replace. When an authorial variant replaces an obvious non-authorial one, the issue hardly breaks waves. But because no variant can be examined without interpreting it in relation to another, all conflicting authorial variants--such as this one--are chosen based on which Joyce an editor chooses to present: early, medial, late, or a conflation.

Thus it is easy to see why any edition of Ulysses, even one that seeks to establish a text free from all non-authorial variants as Gabler's does, is necessarily problematic. As Jerome McGann noted, "editing, including critical editing, is more an act of translation than of reproduction" which "necessarily involves fundamental departures from 'authorial intention,' however that term is interpreted" (McGann Condition 53). One form of praise for Gabler's edition is that it allows us to view this axiom in an extreme example
12: despite his attempt to remove all non-authorial influence--perhaps indeed because of it--Gabler asserts his own influence on the text. The "ideal" but hypothetical text that Gabler seeks to 'reconstruct' is of Gabler's own creation as much as it is Joyce's; after all, as other critics have pointed out, the continuous manuscript text is likely a creation, not a reconstruction, of a form that Joyce himself likely never saw.13

Future editors, taking note of McGann's principles, may seek to edit one Joyce instead of conflating several. These are perhaps the most useful possible editions of Ulysses that have yet to appear, as I will mention in a moment. Editors of such texts must realise, however, that seeking Joyce's authorial intentions is further complicated when one considers that because the readership of Ulysses has always encountered errors in the text, these 'errors' are in fact a meaningful component of Ulysses. Ira Nadel argues that Joyce's "[l]anguage, through dislocation, renews itself and 'error' becomes correction, a new method to establish meaning" making "Joyce's errors...the very 'portals of discovery' Stephen refers to in his discussion of mistakes in Shakespeare"
14 (Nadel 133). Vicki Mahaffey later makes a similar argument and declares that some of Joyce's 'errors' are sometimes intentional, that because "he situates many of his characters' thoughts and utterances on the line between wanting to 'bring forth' a word or idea and wanting to suppress it" we must recognise that "[e]rrancy (modelled on odyssey) provides the very structure of Ulysses" (Mahaffey 183, 186).15 Furthermore, it is widely known that Joyce's Ulysses, like Dante's Inferno, exaggerates many personalities known by the author, in a manner either flattering or deprecating, depending on one's personal relationship with Joyce. Should Ulysses be corrected to portray accurately the real-life individuals it represents? No one since the Middle Ages has attempted to change such views in Dante's verse, much to the dismay of many of hell's shades, I am sure.

In spite of these peculiar difficulties, there are possible editions of Ulysses that are now awaiting an editor, which would undoubtedly be useful to scholars. A synchronic critical edition of Ulysses that seeks not to shape the text but to represent it as it appeared at just one point in history
16 would indeed be very useful. Scholars seeking to read Early Joyce may read an edition based on the earliest drafts and manuscripts; readers of Medial Joyce may appeal to an edition that represents later manuscripts or one that corrects the 1922 first edition; Later Joyce could be found in editions revised by the author (though it may be argued that satisfactory editions of this final category presently exist).

Following Gabler's attempt at representing all pre-publication variants, it would seem that a similar edition--a genetic or variorum edition--representing all post-publication variants would be meritorious.
17 It is true that post-publication variants of Ulysses necessarily include many more non-authorial errors than those found in pre-publication materials because of the participation of editors, publishers, printers, typists and, in one case, an illustrator.18 Some may consider this fact to be an argument against such a text, but there are several reasons why it should be interpreted as such an edition's raison d'être. Obviously, if some of Joyce's errors are intentional (as Mahaffey postulates), it is impossible to distinguish these from non-intentional errors (Joyce did, after all, proofread the mistake-riddled 1922 Shakespeare & Co. and the 1937 Odyssey Press editions). But perhaps a better argument lies in the fact that Ulysses is one of the most studied texts of the twentieth century; all of its variants--non-authorial and otherwise--are part of the text we read and study. Considering a work of literature to be the product of the author operating alone is not only naïve, it is reductive. The historical context in which an author writes is essential for understanding the work of that author.19 Joyce's historical context included, as we have seen, many obstacles that prevented him from producing a text that was free of non-intentional error, but to try to escape this fact is to try to visit Joyce and his text in a time that never existed.

Some critics, McGann among them, claim that the "'perfect' form" of the protean text would be a correction of the 1922 first edition (McGann "Postmodern" 301). Ultimately, however, this would differ from the 1926 second edition (which incorporated into the text the eight-page list of "Errata" attached to subsequent impressions of the first edition) in only a few moments where the editor has sought to represent the Joyce of his choice. Editions that continue to conflate the Joyces of Ulysses, such as Danis Rose's 1997 "Reader's Edition" (Lilliput Press), will simply be open to the same criticisms as Gabler's reading text.
20 Less problematic would be an edition that represents two previous editions on facing pages (most useful would likely be the 1922 Shakespeare & Co. edition and the 1937 Odyssey Press edition--the first and last editions published in Joyce's lifetime). Such an edition would be quite large (perhaps spanning two or three volumes as Gabler's edition does), but undoubtedly would be a useful tool for scholars.

One final and thus far little explored editorial possibility would be a hypertext edition
21 of Ulysses which represents all textual variants. A hypertext edition--unavailable to the 1984 editorial team--would have the unique ability to represent variants without the distracting and confusing diacritical marks that mar Gabler's edition. Yet even this solution is not without its drawbacks. Creating a hypertext edition of Ulysses means intensive labour, similar to the production of the Garland edition, only once again to produce an edition that prefers some variants to others. Current technology (late 2003) only allows one inter-textual variant to be represented in the primary reading text, while all others are represented only in links to them, which of course is up to each reader to explore. A hypertext allows each reader to become her own editor, but this kind of reading is a very tedious process. Like a variorum, a hypertext edition seems more practical to scholars of Ulysses than to general readers, and perhaps is most useful on this level.

In the Odyssey, Odysseus could capture Proteus only by surprising him and holding him until he became exhausted of his form changing (Fitzgerald 66); for a reader to capture Joyce's protean text, exhausting all changes of form is just as necessary. Donald Phillip Verene hinted toward such a resolution to Ulysses's protean nature when he suggested readers study both the 1922 first edition and the 1984 reading text in order to gain a more comprehensive perspective of the text (Verene 217). Editors must realise that any edition printed in the post-Gabler era must stand in relation to all those that have come before it.

The first editors of Ulysses who set out to correct it were very much like Stephen at the conclusion of Portrait: hearing a soft voice spoken in their hearts, they were eager to discover the mode whereby Joyce's text could truly express itself (P 206-7), and in doing so produce a definitive Ulysses that would quiet all textual debate. Yet, like Stephen walking along the Sandymount Strand years later, they realised they have rushed out into the world only to find that they cannot escape the past. The complications posed by the protean text of Ulysses are a necessary element of the text. They are obstacles that were troublesome in the past and will always be so for editors who seek to produce an edition intended to replace all those previous. Future editors must embrace the protean text and welcome the fact that each edition, past and present, is at least as significant in its relationship to every other edition as it is to Joyce. Only by accepting that each edition of Ulysses has its own historicity and by studying the editorial oeuvre can we learn to truly find "Ulysses as Joyce wrote it."

APPENDIX (A Genealogy of Editions of Ulysses)


© George Micajah Phillips, 2004
volume 4, issue 2, 2003-4