Until his death in 1958, Harry Thrift was known around Dublin as the man in the bicycle race in Ulysses, where he was runner-up in the quarter-mile flat handicap in the Trinity Races of June 16, 1904. Joyce doesn’t mention that the young man also placed second in the first heat of the 120-yard handicap, but out of the money in the final. Ulysses is a big book, but not big enough for all the newsworthy events of Bloomsday, 1904. Harry’s transit in the track of the sun was brief. By 1917, when Joyce was composing the “Wandering Rocks” episode in Zurich, Thrift had been a Fellow of Trinity College for ten years, having won tenure in the 1907 Madden Prize examination. Harry devoted himself more to the advancement of rugby and cricket than to the sciences. Eventually he was bursar, then auditor of the college, serving fifty years on the faculty. One of the great collegiate sports boosters in Ireland this century, the teen-age cyclist of 1904 grew into the old boy of quad and green. He died on February 2, 1958, the thirty-sixth anniversary of the publication of Ulysses.

Those who remember Thrift are themselves rounding the bend, most recently (in 1987) Roger McHugh, Joycean and director of the School of Irish Studies. Harry Thrift’s good works, sturdy frame, and jolly demeanor may fade to a misremembered blur, because he is deposed in Ulysses: The Corrected Text (1986), the only form of the novel being printed in the world today. The versions with H. Thrift cycling in the last segment of the “Wandering Rocks”—the only one now in the classroom—have all been replaced with versions putting an H. Shrift near the head of the pack:

Thither of the wall the quartermile flat handicappers, M.C. Green, H. Shrift, T.M. Patey, C. Scaife, J.B. Jeffs, G.N. Morphy, F. Stevenson, C. Adderly and W.C. Huggard, started in pursuit. [1986 Ulysses, p. 209]

Production of all the old Thrift versions was halted in order to replace them with the Shrift version, which is claimed to be part of “Ulysses as Joyce wrote it” (1986, p. 649 and elsewhere). Did it occur to anyone to check whether Thrift was a real person before changing him to Shrift? Apparently not.

The disappearance of Harry Thrift is only one of literally thousands of unfortunate features of Ulysses: The Corrected Text (1986), prepared by Hans Walter Gabler, a professor at the University of Munich, with two graduate students, Wolfhard Steppe and Claus Melchior. In 1986 the Random House hardcover and Vintage paperback editions of Ulysses, last “corrected and reset” in 1961, were replaced by a very different text claiming to correct 5,000 errors. Outside the US, the Bodley Head Ulysses and the Penguin paperback editions of 1960 and 1968 were also replaced with The Corrected Text. Thus a uniform text became the only trade version in print worldwide, although this version differs radically from what Joyce had himself brought forth in 1922 and approved again in 1926, 1932, 1935, and 1936.

The transparent scheme to replace Ulysses outright with another version was apparently partly motivated by the hopes of the Joyce estate—which is represented in business matters by the Society of Authors in London—for a new copyright to run seventy-five years from 1984. This could be accomplished only by creating an entirely new work, which in an unintended sense has been done. The new edition’s supporters tried to persuade the public and professoriate that all previous editions were unusable. The assault on the old text centered on a handful of typos that were rehashed again and again, as if the slip “beard” for “bread” (found only in 1961) proved all editions since 1922 unworthy of attention. Because the first edition was printed in Dijon, some whined about battered Joycean coinages, inciting us to curse the French printer. But in fact the 1922 text was not defective in the way that was suggested. In Michael Groden’s “review” in The James Joyce Quarterly, one paragraph lists ten Joycean coinages restored “for the first time” by Gabler in 1984. Yet for six of these, the majority, the 1922 and 1984 spellings are identical. Is no one awake at the wheel?

During Joyce’s life hundreds of typographical errors in the first edition were corrected, but a similar number of new variants slipped in unnoticed in later editions, meaning that the first and last lifetime texts were equally distant from what Joyce intended. Suddenly, in the 1980s we were told that there were “thousands” of errors all along, at least seven on every page. That claim is a sales pitch for the new edition, and has no basis in the texts themselves. Ulysses: The Corrected Text is not a purified text (new blunders like “Shrift” aside), but a different version from what Joyce conceived, authorized, and saw into print.

On the first page alone the following changes of wording were made in 1986: by to on, up:out, country:land, low:slow. On what basis was this done? The final typescript of the opening of Ulysses, which Joyce is known to have revised, is lost. The editors returned to the words of an earlier manuscript, overruling the form that Joyce elaborated on and ultimately passed several times in proof for the 1922 edition. With the loss of the revised, final typescript, we will never know if the four words rejected in 1986 were or were not Joyce’s own final revisions of the typescript. In an edition so different on the first page from what Joyce himself saw into print, we get not a “corrected text” but another version altogether. At best such a new text (were it accurate) could stand beside the version published during Joyce’s lifetime only as an alternative—not as a replacement. But commercial and not scholarly considerations are behind the disappearance of the version known to the author and his audience of sixty years.


The editors of this radically new Ulysses, working in Munich, first prepared Ulysses: A Critical and Synoptic Edition (1984) for Garland Publishing in three volumes totaling 1,920 pages. The 1984 edition has facing pages intended to show “synoptically,” i.e., by using a complex code of symbols on the left-hand pages, the growth of Ulysses from manuscripts through final proof revisions. On the facing right-hand pages are the same passages in an edited form or “clear text” without notes or symbols. This right-hand reading text without apparatus became in 1986 The Corrected Text.

Both the 1984 Synoptic Edition (84U) and the 1986 Corrected Text (86U) are based on facsimiles of Joyce’s manuscripts in American collections, all of which were transcribed in Munich. The Rosenbach Manuscript, named after the Philadelphia foundation where it is housed, is a collection of almost seven hundred pages spanning the eighteen episodes of Ulysses in Joyce’s own hand. For some chapters, however, the Rosenbach Manuscript is a copy that Joyce made for sale to a patron: it is not the final version, now lost, that Joyce gave his typist. The Munich group worked from a 1975 facsimile of the Rosenbach Manuscript, which was edited by Clive Driver with an introduction by Harry Levin.

Additional drafts, typescripts, and proofs of Ulysses are reproduced in sixteen volumes of The Joyce Archive (1978–1979) published by Garland. The Ulysses documents were microfilmed in ten American cities. The “archive” consists of high-contrast facsimiles, which allow faint markings to show up, although at the cost of distinctions between Joyce’s ink and the printer’s pencil. The frequent remark that facsimiles of the Rosenbach Manuscript and The Joyce Archive made possible a “Corrected Text” relies on the assumption that the editors compared all their work in Munich against the originals in America. They did not, with disastrous consequences.

Joyce is reputed to have said that were Dublin destroyed it could be re-created brick by brick from his novel. An exaggeration on three counts: some streets and shops aren’t mentioned; the author occasionally made mistakes; and in many instances Joyce maliciously contoured the cityscape to get even with enemies. Dubliners couldn’t be published in Dublin because Joyce named names, putting publishers at risk. The same happened to A Portrait of the Artist. The historical facts of the first two books are fewer than in any single episode of Ulysses, itself an encyclopedia, street directory, dialect dictionary, census, pub guide, ordnance survey, and vade mecum bound up in blue and white wrappers.

Any scholarly edition worth its price would try to account for differences between historical Dublin and Joycean Dublin. No edited text can escape choices among authorial variants, and most editions do emend slips of the pen or momentary confusion. Where Joyce changes the spelling of a name three times, we must know. Where he slips up, we want to know. If he falsifies the date of the Phoenix Park murders to 1881 to avoid memorializing a massacre near his birth in 1882, the edition should take note.

Ulysses: The Corrected Text is unsupported by such research. Notes on historical errors and justification for dozens of editorial changes of names, places, and dates are never given. If there is an editorial policy, it is neither stated nor evident from the whimsical “correction” here and blunder there. The implications go far beyond the ignorant manufacture of Shrift from Thrift. The failure to distinguish Joyce’s slips from his revisions leads to bizarre assumptions about the lost drafts, the role of the typists, and Joyce’s last-minute changes.

So who’s who in Ulysses? Consider the words “Flowers of idleness” (61U 71), a phrase that occurs early in Leopold Bloom’s interior monologue in “The Lotus Eaters,” the fifth episode, as Bloom notices the heat of the morning, while standing in front of the Belfast and Oriental Tea Company. His thoughts turn to “Ceylon brands” of tea and to the lethargic heat of the Far East:


Too hot to quarrel. Influence of the climate. Lethargy. Flowers of idleness. [p. 71]

The standard commentaries on Ulysses gloss “Flowers of idleness” as an allusion to Byron’s Hours of Idleness. This appears to be the first literary allusion of the “Lotus Eaters” episode, and Hours (1807) was Byron’s first published book: Joyce is priming his hidden machinery. Continuing his walk, Bloom will soon head for the Turkish baths, a blatant Homeric motif of Lotuses, Flowers (rhymes with Hours), and Idleness.

Doubtless there are more Byronic allusions bobbing near the surface; it only remains for a scholar to scoop them up. Since Byron is at the head of the chapter, Joyce’s symmetry fetish would make the end of “Lotus” a good fishing pond. On the last page of the “Lotus” proofs Joyce added:

Cricket weather. Sit around under sunshades. Over after over. Out. They can’t play it here. Duck for six wickets. Still Captain Buller broke a window in the Kildare street club with a slog to square leg. [61U 86]

Hours of idleness under a blue sky, the endless round of play. Why the broken window? Is the broken window real or Joycean? (“I hear the ruin of all space, shattered glass and toppling masonry” [61U 24].) Who is Buller? Is he the army captain in the Finnegans Wake roster of legendary players, “barrackybuller” (FW 584)? Dublin’s only Captain Buller is in Thom’s Official Directory—the equivalent of today’s phone book, which Joyce plundered shamelessly when his memory faltered—on page 1819:

Buller, Captain, Byron lodge, Sutton.

The Byron family, unrelated to the poet, had a tavern that is mentioned in A Portrait of the Artist as a Young Man. Later than 1904 someone must have assumed the poet lent his name to Byron Lodge when a neighboring property was renamed Shakespeare Lodge. Joyce twigged to the filiation of a Dubliner named Byron and a poet named Byron, and Captain Buller is his cunning way of linking them. He expected the professors would seek out Buller in Thom’s. We have proved him right.

If The Corrected Text remains the only version being printed, taught, and studied, few will make the leap from “Flowers of idleness” to Byron to Buller to Byron Lodge, because the 100,000 copies most recently printed all read “Captain Culler.” There was no Culler in Dublin 1904, captain, sergeant, or civilian. Whence this Shriftian blunder? In a word, facsimiles. The recent text was compiled from published facsimiles, photocopies, and microfilms. The “research” was done in Munich, thousands of miles from Joyce’s scattered drafts, final fair copies, typescripts, and proofs, and the transcriptions were not checked against originals. The proof sheet on which Joyce wrote Captain Buller in ink is unambiguous in the original at Harvard. Printer’s pencil markings nearby, although no distraction in the original, are so dark and unruly in the facsimile that a careless transcriber might for a moment think that there was a capital C on the page. (See Joyce Archive, Vol. 17, p. 93.) Neither Joyce nor the printer wrote a C. The printer put an “e” in a circle, attempting to clarify “Kildare” for his typesetter, who set Buller accurately but came up with “Kildere” (soon enough set right). No one associated with the production of Ulysses—not author nor shop foreman nor compositor—wrote, read, or rendered a C in Buller. The first edition has Buller; Thom’s has Buller; Joyce wrote Buller. Just as the 1922 text had Thrift, the Evening Telegraph had Thrift, and Frank Budgen wrote Thrift from Joyce’s dictation.

The Corrected Text correctly spells all Dublin names in the first three pages of Ulysses, but soon meddles with the early morning exchange between Buck Mulligan and Stephen Dedalus. The new version claims to correct Conolly (one “n”) Norman to Connolly Norman. As all editions stood before now:

—That fellow I was with in the Ship last night, said Buck Mulligan, says you have g.p.i. He’s up in Dottyville with Conolly Norman. General paralysis of the insane. [61U 6]

Thom’s Directory, page 1381, clarifies Dottyville:

Richmond District Lunatic Asylum, North Brunswick-Street./ For the Country and City of Dublin, Counties of Wicklow and Louth./ Resident Medical Superintendent, Conolly Norman, F.R.C.P.I.; ex-F.R.C.S.I.

Telling us that Conolly Norman is an ex-Fellow of the Royal College of Surgeons in Ireland is one of the quainter touches in Thom’s. Conolly Norman is named correctly on Thom’s pages 826, 840, 865 (with gynecologist Professor Kidd), 867, 879, 1381, and 1438.

A bad repair job inevitably unmasks itself on the first long trip out of town. That Conolly to Connolly is the first name change in The Corrected Text shows you won’t get far without a fanbelt. Norman and Thrift and Buller are ill-served by these corruptions, and their names were letter perfect in all previous editions.


It is a commonplace that Ulysses has suffered hilarious typos—I have already mentioned “beard” for “bread” (61U 147). The odd thing about all the essays decrying the pre-Corrected text is that most of the rehashed examples are trivial, far more trivial than Shrift and Culler. From its context one can tell a beard from a bread, a bear, or a bard. But the Buller-Byron allusion cannot be recovered from the name Captain Culler. Harry Thrift is unrecognizable as Shrift. Conolly Norman gets off lightly in The Corrected Text. He might have been rechristened Donolly Gorman.

As a purported transcription of “manuscripts” (as the advertisements and jacket blurbs boast) the Synoptic Edition is worthless because it relies so heavily on facsimiles. The Joyce Archive reproductions are intentionally high-contrast, darkening to legibility anything detected by the film. Pencil markings take on the appearance of flowing ink and colored pencil looks like black lead. High-contrast development of the film wipes out distinctions that are visible to the naked eye.

To be precise, there are two distinct types of lead pencil. The spidery thin lead is definitely the printers’; the thick lead will be harder to apportion among Joyce, his helpers, and the Dijon printers, Hirschwald and Darantiere. A documentary edition would distinguish markings in the thin lead and the thick, and specify colors of pencil and ink. Not one of thousands of markings was thus labeled in the Synoptic Edition of 1984. Even if the originals had been transcribed, no one could confirm or reject hypotheses about the various hands, authorial or not, without this data.

All studies of Joyce’s drafts, typescripts, and proofs must now begin anew, as if Ulysses had never been edited. Only then will we know who underlined a word for italics, diddled an accent, inserted a comma, deleted a preposition, or remade “you’re a doner” into “you’re a goner” (61U 104). Why argue, as some have done for the past three years, whether “goner” has the true Dublin ring if we have no idea whether the change is in ink, in pencil, in blue pencil or red? Many fantasized “corruptions” will turn out to be Joyce’s own revisions.

Failure to consult originals not only means that the guesswork leading to the five-thousand-times renovated Corrected Text is unsound, but that students of Joyce’s drafts will be misled by the falsified record of the Synoptic Edition. Throughout the 1984 volumes, erasures and illegible deletions are recorded as if detected. Yet facsimiles cannot be used to spot erasures, or decipher blurred and faint writing. The 1984 apparatus alleges erasures that simply are not in the originals. In the third episode, “Proteus,” Stephen recalls the “fading prophecies of Joachim Abbas” in Marsh’s library (61U 39). Five lines below is the variation, “Abbas father, furious dean, what offence laid fire to their brains?” The 1984 edition uses a symbol that means Joyce wrote another word, then made a deletion beneath the first Abbas. Did Joyce toy with Joachim’s name? He is also known as Joachim Flora, Floris, Flore, or Fiore. One could speculate on lost revisions except that there is no erasure beneath, beside, behind, or before Abbas in the original.

So it stands with the entire record of erasures for the 690-leaf Rosenbach Manuscript. The 1984 apparatus records erasures that never existed and misses even more that do. If one goes to Philadelphia and holds up to a lamp each leaf of the second and shortest episode of the book, “Nestor,” one can find in minutes ten erasures not detectable in the facsimile, and not recorded in 1984. Twenty pages from the start of the book and the transcriptions are in a mess.

The Synoptic Edition, then, is a study not of Joyce’s manuscripts but of inadequate facsimiles. At spots where the manuscripts have worn away the 1984 edition pretends to decipher Joyce’s hand. One example should suffice. When Joseph Prescott wrote his 1944 Harvard dissertation, he examined proofs of Ulysses in the Houghton Library. His Exploring James Joyce of 1964 has a footnote about this garbled mouthful: “I munched hum un thu Unchster Bunk un Munch-day” (61U 170). The 1984 edition claims this is what Joyce wrote on the first proofs (84U 358). As Prescott noted, the first “un” does not exist in Joyce’s hand—the sheet’s folded corner has been torn away for decades. The 1978 facsimile that the Munich team worked from has a blur (Joyce Archive, Vol. 18, p. 123). It seems they peeped ahead to what the printer set in the next proof and doubled back to substitute this “un” for the blur. Voila! Ulysses As Joyce Wrote It.

Another pitfall of the 1984 edition’s reliance on facsimiles is exposed on the 1922 Ulysses typescript now at Buffalo, in which Joyce added, in black ink, Crême de la crême. The phrase is in Bloom’s mind, so the circumflexes instead of the usual accents graves (è) may be his slip, not Joyce’s. In any case, someone has blue-penciled in the “correct” modern accents. How does the Synoptic Edition distinguish Joyce’s ink from the blue pencil? It doesn’t (84U 370). Joyce’s circumflexes are unreported because they are obscured by the blue pencil, which appears black in photocopies (Joyce Archive, Vol. 12, p. 316).

Aside from the claim that it’s based on original manuscripts, the most misleading boast of The Corrected Text is that it reproduces letter by letter and comma by comma Joyce’s own spellings and punctuation. In more than two thousand places, or three times a page, the new version is demonstrably not what Joyce wrote in any manuscript. Rather than confess to extensive alteration of Joyce’s spelling, punctuation, italics, capitals, compounds, and foreign language phrases, the editions insist the words are Joyce’s own:

Emendation is essentially confined to the removal of unquestionable errors…. This results in an unstandardised and unmodernised text. Inconsistencies of usage and orthography, unconventional spellings, obsolescent word forms and of course the idiosyncracies of Joyce’s punctuation are largely left standing. [84U 1898]

As admirable as this policy might be, it is not the one followed in 1984 and in 1986. Setting aside cases where Joyce may have revised now-lost drafts, and looking only at uncontested final manuscript readings, the newly edited texts overrule what Joyce actually wrote two thousand times. (In another thousand places we encounter dubious decisions made in the 1984 Synoptic Edition, where authoritative documents conflict, and the first edition of 1922 seems to embody final revisions.)

Although the refusal to print precisely what Joyce wrote usually affects only two or three readings a page, the alterations commonly swell to five, six, or seven per page. In chapters typed directly from the Rosenbach Manuscript, expanded on typescript, and passed in several proofs, The Corrected Text can depart from the final version in Joyce’s hand as many as seven times per page—the frequency, we were told, with which the first edition was corrupt.

For a useful sampling of the misguided policies of the editors, take the “corrected” page 541 of the 1986 edition, whose text begins on page 662 of the 1961 edition. (Yes, the new edition compresses the novel into one hundred fewer pages by using smaller type and more lines per page. An “unreadable book” made doubly so.) Bloom and Stephen are slouching toward Eccles Street after midnight. Stephen is stunned from a brawl and Bloom is in a stupor, although a loquacious one. The original documents consist of a draft, the Rosenbach Manuscript made from it, the typescript from the Rosenbach Manuscript, and the printer’s proofs. This typist, who entered at the sixteenth episode, was assailed for adding six hundred commas and generally restyling the text toward standard (non-Joycean) English. Since the Rosenbach Manuscript is heavily revised but carefully written out, the early draft would normally be drawn on only for emendations of “obvious errors,” none of which I could detect in the final manuscript for this page. Of interest are final elements in Joyce’s own hand that have been rejected in The Corrected Text.

On this page, 541 or 662, notable for a Homeric allusion to the treachery of Sirens in a couplet sung by Stephen, Joyce’s final hand is overruled eight times in the 1984 and 1986 editions. The Rosenbach Manuscript itself is rejected six times; one late addition by Joyce in proof is not followed; and his insertion of an ellipsis—to suggest a gap in the dogtired exchange of Bloom and Stephen—is ignored. Beyond these eight “corrections” are three more cases in which typist or printer first made a slip that Joyce touched and varied. I dissent from but won’t quibble here with the 1984 edition’s return to earlier versions than the one finally reached by a mixture of corruption, correction, and last-minute inspiration.

It is a claim of the 1984 and 1986 editions that fearsome “transmissional departures” (jargon for “nonauthorial changes”) are not accepted, regardless of how many times Joyce saw them, varied them, or added to them. (This is more posture than postulate: just as the typist tried to rein in Joyce so have more recent editorial hands.) On “corrected” page 541 only one among eight editorial changes, a comma to close a subordinate clause, even begins to suggest an “obvious error” (whatever that means). Allow the comma, and there remain seven editorial alterations of what Joyce last wrote. (Sometimes in overruling Joyce’s manuscript the 1984 text arrives at the same reading as the 1961 text, which it dismissed as corrupt.)

The seven “corrections” of Joyce’s final manuscript and his proof additions are as follows:

  1. The song title Youth here has end is altered to capitalize End. Yet lower case was standard British style in song and book titles.
  2. Joyce’s underlined name for the German composer Johannes Jeep is altered to roman type, though in Jeep’s lifetime at the turn of the seventeenth century most books would have italicized it. A Baroque Joyce might have spurned that rule and used roman. Modernist Joyce made a typographical play by italicizing the songster’s name. Not once in seventeen prior typesettings did a printer think italics an error. Nor are they.
  3. Still bucking Joycean syntax by page 541, the edition has a comma after Monday in this typical Bloomism, an entire paragraph in itself. In Joyce’s hand:

Still, supposing he had his father’s gift as he more than suspected, it opened up new vistas in his mind such as Lady Fingall’s Irish industries, concert on the preceding Monday and aristocracy in general.

The typist, alarmed by this hefty boa, added commas after “gift,” “mind,” and “Monday.” The 1984 synopsis omits the first two, without even a footnote, but plugs in the typist’s third, after “Monday.”

  1. The typist also inserts a colon that Joyce neither wrote nor deleted. It is adopted in The Corrected Text, although elsewhere such colons are not provided by Joyce and are not supplied in the 1984 synoptic edition. Stephen will sing his couplet:

about the clear sea and the voices of sirens, sweet murderers of men, which boggled Bloom a bit:

Von der Sirenen Listigkeit
Tun die Poeten dichten.
[From sirens’ cunning
Fashion poets verses.]
[86U 541]

The harmless colon is not Joyce’s but the typist’s. Fair enough, if one feels Joyce saw and accepted the change. (The Corrected Text claims not to be what Joyce passed but what he wrote without interference by the typist.) Once one has accepted the colon before the song found in all editions, would it not make sense to do the same two pages later? All editions before 1984 read:

Stephen singing more boldly, but not loudly, the end of the ballad:

Und alle Schiffe brücken
[And every ship ?broke up
?crossed over].

Only two pages after inserting the typist’s colon, The Corrected Text offers one of its five thousand improvements by removing the same mark provided by the same typist in the same circumstance. Being neither what Joyce wrote nor what he passed, pages 541 to 543 of the 1986 edition generate new inconsistencies. (The mysterious verb brücken sails by without a textual note, silent as Death’s ship.)

  1. In the wee wicked hours, the Dubliners have cheek-to-cheek fur on their tongues and fog in their skulls. The narration, too, deteriorates, as in the paragraph broken off (in editions 1922–1983) with three dots indicating an ellipsis:

Stephen went on about the highly interesting old…

[61U 662; cf 86U 541]

Joyce added the ellipsis on second proof; in 1984 it was removed. These narrative suspensions first appear in Joyce’s early draft of the episode, and he multiplied them in the Rosenbach Manuscript. Joyce gradually reshaped the “Eumaeus” episode, adding ellipses throughout. Another ellipsis that Joyce added in proof is disallowed in The Corrected Text:

Still, to cultivate the acquaintance of someone of no uncommon calibre who could provide food for reflection would amply repay any small…

[61U 646; cf 86U 528]

Joyce never says what small compensations boon companionship may demand. On the third proof he thought three dots there would do just fine. The first edition and all others have thirty ellipses in “Eumaeus,” every one coming from Joyce. Twenty-seven are removed in 1984 without even an explanatory paragraph. Typical of the confusion, the 1984 text notes (p. 1749) assert that the typist added “frequent” ellipses. The typist added none—all are in Joyce’s hand. The 1984 accusation against the typist has already entered Joycean lore, and is credulously retailed in Derek Attridge’s Peculiar Language (1988, p. 176).

  1. The hallmark of the “Eumaeus” episode is cliché, euphemism, redundancy, and mashed metaphor ambered in a gooey sentimentanty. Kindhearted Bloom’s thoughts, and maybe his actual words to Stephen, take this form:

But such a good poor brute he was sorry he hadn’t a lump of sugar but, as he wisely reflected, you could scarcely be prepared for every emergency that might crop up. He was just a big nervous foolish noodly kind of a horse, without a second care in the world. [1986 version]

Bloom then taggles out some nonsense about the plight of animals. Again the episode snarls itself into gibberish:

Nine tenths of them all could be caged or trained, nothing beyond the art of man barring the bees. [So far so good.] Whale with a harpoon hairpin, alligator tickle the small of his back and he sees the joke, chalk a circle for a rooster, tiger my eagle eye. [1986 version]

Those last four words make little sense on the surface and no explicator has taken them up. On the second page of A Portrait of the Artist young Stephen fears “the eagles will come and pull out his eyes,” and the Anglo version of mon oeil is loitering here, too. The “tiger my eagle eye” had a slightly different, albeit murky, form in Joyce’s final draft, the typescript, and the editions of 1922 to 1983: “tiger, my eagle eye.” An earlier draft—one as messy as any other—lacks the comma, so in 1984 the final written and published form is overthrown. To remove Joyce’s comma and to create the confounding “tiger my eagle eye” is hardly an obvious correction. Rather, it is an ill-considered conflation of drafts.

  1. The next fiddling with Joyce’s final draft and published version on page 541 raises fundamental questions about editorial theory. The blathery sentence-paragraph in its 1922–1983 form:

He looked sideways in a friendly fashion at the sideface of Stephen, image of his mother, which was not quite the same as the usual blackguard type they unquestionably had an indubitable hankering after as he was perhaps not that way built.

The claimed corrections of 1984 and 1986 are here given italics:

He looked sideways in a friendly fashion at the sideface of Stephen, image of his mother, which was not quite the same as the usual handsome blackguard type they unquestionably had an insatiable hankering after as he was perhaps not that way built.

First the addition of “handsome.” It is in neither the final Rosenbach Manuscript which Joyce gave his typist nor any published in his lifetime. Exhumed from the early draft, it plumps out the passage with no allowance for context. Bloom knew May Dedalus, and sees a resemblance in her son. Stephen is attractive, but not a hunk. Despite the redundancies and muddles peculiar to the episode, to say Stephen’s beauty “was not…the usual handsome” plasters over one of the rare glimpses of the physical Stephen we get. Ulysses is more than pure style: a portrait of Stephen is emerging and Joyce deleted a word to sharpen up that portrait. The Corrected Text is marbled with the fat of such pseudo-restorations from shoulder to shank.

Like the gratuitous “handsome,” the substituted “insatiable” for “indubitable” is born of inattention to context. Joyce’s typescript addition makes clear that Bloom thinks of the “blackguard type” whom women “unquestionably had an indubitable hankering after.” In this episode of windless repetition, is language anywhere as effectively becalmed as with “unquestionably…indubitable”? (See the reproduction of the corrected typescript on this page.)

The palpable genius of Joyce’s typescript revision is absent from the 1984 synopsis. The breakdown is traceable to the odd theory of the Synoptic Edition that all the fragments in Joyce’s hand from early drafts through last proof insertions can be assembled like mosaic chips. In service of this ill-conceived theory is a “synoptic” apparatus that omits entirely words or punctuation not in Joyce’s hand. Thus the typist’s error “indubitable”—which inspired Joyce to stitch in “unquestionably”—is banned from the display of revisions in the Synoptic Edition. We are told instead that Joyce added “unquestionably” two words to the left of “insatiable hankering after.” Not so. The “insatiable” was nowhere on the page when Joyce thought up “unquestionably.” The synopsis and footnotes have no record of the typed “indubitable” which catalyzed the revision. The fact that the typist mistakenly changed “insatiable” to “indubitable” is irrelevant once Joyce adopted “indubitable” as part of his own revision. A true-Bloomism is skewered on the 1984/86 trident of bad theory, apparatus breakdown, and literary insensitivity.

We have seen how seven changes of Joyce’s hand cluster on one page of what purports to remove seven “corruptions” per page. One could easily show that other pages have been similarly mangled.


Next we must lay to rest the claim that Ulysses: A Critical and Synoptic Edition is a record of Joyce’s own revisions. If we look at Joyce’s documents behind the one page of The Corrected Text just examined, it is obvious that the 1984 apparatus is incomplete. The newly edited text inserts a word (“handsome”—from the working draft) that Joyce had understandably omitted from his final copy. The “tiger, my eagle eye” also lost the punctuation of the Rosenbach Manuscript because of a trust in the same rough draft. But other telling variants from the working draft pass unmentioned. We can’t bemoan the insensitivity that led to replacing “indubitable blackguard” with “insatiable” (example 7) without noting that the working draft read “incurable.” Neither Joyce’s finally intended “indubitable” nor his earliest “incurable” are mentioned in the synopsis.

When the 1984 edition added capitalization to Joyce’s Youth here has end (example 1 above), it did not inform us that the working draft reads Youth has here and (sic) end. If we assume that Joyce slipped in writing and for an, we get a different sequence from the Rosenbach Manuscript:

Youth has here an end [draft—not recorded in 1984]

Youth here has end [Rosenbach Manuscript]

Youth here has End [1984/1986 emendation; also in 1961]

Joyce is translating Sweelinck’s Mein junges Leben hat eine Ende. Not a coincidence, then, that End is capitalized by the editors, though Joyce’s usage requires no capital. There are two long books and scores of articles on Joyce’s musical allusions, yet none cite the draft title. The grandly named Critical and Synoptic Edition does not provide the variants to set others scurrying for Joyce’s source of Sweelinck (the spelling of which is another hint). The same seekers would track the Johannes Jeep couplet and the utterly unidentified, barely translatable “Und alle Schiffe brücken.” Maybe all three airs are in an anthology Joyce consulted.

The single most important linguistic clue about these lyrics is in the working draft’s archaic (but once standard) dieresis on Jeep’s Poëten. That Joyce omitted it from his final draft is regrettable, since the baroque accent has a scent of Kammermusik absent in plain modern Poeten. No one interested in Joyce’s allusions and linguistic talents can proceed without notice of Poëten. The variant is unrecorded in the Synoptic Edition or anywhere else in print. For a study of Joyce’s Ulysses drafts, the self-styled genetic apparatus in the Synoptic Edition (like the editors’ transcription of the Rosenbach Manuscript) is, to repeat myself, worthless.

For foreign languages and accenting in Joyce’s drafts, see my “Gaelic in the New Ulysses,” Irish Literary Supplement (Fall, 1985). The Irish Gaelic is botched in The Corrected Text, despite unattributed corrections of 1984 errors taken from my article and worked into the 1986 text. A majority of the thirty changes made between 1984 and 1986 (Irish, Latin, Spanish, French, and German phrases, and crucial punctuation) are lifted from my post-1984 articles and interviews. With these borrowings, Ulysses became a corrected text in a sense not explicit in the book’s subtitle.

Close scrutiny of individual pages of the 1986 edition reveals seven, six, or five editorial changes per page of Joyce’s final manuscripts. A more general perspective turns up entire classes of emendation not discussed in the Synoptic Edition or The Corrected Text. Changes from what Joyce actually wrote fall into at least thirteen classes:

  1. changes in Joyce’s correct, contemporary spelling
  2. compounding into one word what Joyce wrote in two words
  3. creation or removal of italics

  4. addition of unneeded punctuation

  5. changes of dates, money, and other numbers

  6. changes of personal names

  7. changes of place names

  8. rejection of specific typographical features ordered by Joyce (and intentional suppression of Joyce’s instructions to the printer, making a patchwork of Joyce Archive transcriptions)

  9. changes of Joyce’s idiosyncratic abbreviations

  10. changes of capitalization

  11. changes in literary allusions

  12. changes in Joyce’s spelling or accenting of French, German, Greek, Irish Gaelic, Italian, Latin, Scots English, and Spanish, while other errors in each of these languages are neglected and unidentified (oddities in Swedish and Hebrew stand unnoted and uncorrected)

  13. illusory improvements when a character misspeaks or makes a Freudian slip, where Joyce wrote and clearly intended the slip.

The haphazard execution of these unneeded changes leads to scores of new inconsistencies in The Corrected Text. That Joyce himself may have been irregular and unruly in spelling and punctuation is no license for an edition to outdo him and introduce new muddles. Yet that is precisely how the new text differs from the old. Where we once had Joyce and his printers tangled in their own lapses and second thoughts, a third force asserts itself to add more knots of hazard and mischance to the whole.

The printing history of a work can be as important in editing as manuscript study, particularly when friends of the author help to correct the text. Such alterations were passed on by Sylvia Beach in 1927, Stuart Gilbert in 1932, and again for the 1935 text illustrated by Matisse, and Paul Léon in 1937. (Their contributions are unmentioned in the 1984 edition.) Every different state of the text in an author’s life and soon after must be first detected, then examined, analyzed, and recorded. Text experts say, “Locate, Collate, Relate.” The compilation of a lifetime printing history will also uncover posthumous editions of interest to collectors and historians of the book.

The 1984 edition provides a brief description of ten editions and concludes, “The present edition is the eleventh edition of Ulysses” (p. 1856). Not by a long shot is this true. It is the eighteenth edition. Seven separate typesettings were overlooked, though most are available in the New York Public Library, the British Library, and the Library of Congress. Some were listed in Alan Cohn’s periodic checklists in The James Joyce Quarterly. A more amateurish result cannot be imagined. (An entirely different list of Previous Editions is at the front of The Corrected Text. This is a longer but equally confused mishmash, capped by the misspelling of and as und near the page’s bottom.)

Does it matter that the printing history in the 1984 edition is so shoddy? Aren’t the editions published in Joyce’s lifetime what really count? If seven typesettings are overlooked, those few that were seen may have been only skimmed. No collations of variants in the 1934, 1935, or 1940 printings were made. Those who have never prepared a critical edition may not shudder at the dangers of a sloppy printing history. In the case at hand, collations of all the printings are more important than for any edited version in recent memory because The Corrected Text relies on editions after 1922 for four hundred emendations of Joyce’s manuscripts.

Both Joyce and the first edition are overruled to borrow from the 1926, 1932, 1936, 1960, or 1961 editions. In each of four hundred instances we have Joyce’s own hand and his approval of what he wrote through several proofs. Variants of later printings were substituted for his manuscripts, although the Synoptic Edition has not one paragraph on the individual merits or demerits of any of the 1926–1961 editions. Joyce’s style is thrown out in favor of spellings, punctuation, accents, and italics of unknown authenticity. Absolutely no justification is attempted for these four hundred changes. Some are so patently un-Joycean that the very fabric of Ulysses, its allusions, linguistic texture, and tone are endangered.

To take a typical example, Joyce alluded to Aelfric’s Homilies by lifting the word twey from his copy of Saintsbury’s History of English Prose Rhythm and wrote out twey—an archaic spelling of “two”—in his own hand four times. Someone attempting to correct the text of the Bodley Head edition in 1937 checked a dictionary, located tway but not twey. An early draft manuscript, the Rosenbach Manuscript, the typescript of the lost final working manuscript, and Joyce’s notesheets for “Oxen of the Sun” in the British Museum all have twey. Neither the early draft nor the notesheets are recorded in the Synoptic Edition, yet the inauthentic tway of 1937 is adopted against Joyce’s own hand. The Bodley text contains many such bad guesses, some incorporated in the 1986 text.

A vowel in Joyce can speak volumes, and tway is not the only slight to world authorship. The 1984 and 1986 editions mangle Joyce’s use of Aelfric, Dante, Shakespeare, Milton, Byron, Burns, Goldoni, and Goethe (and others, I suppose). There may be a record here: Most Authors Abused in a Single Edition.

The Corrected Text draws more emendations from the 1932 Ulysses than any other (but denies it outright: 1984 Afterword, p. 1899). When Joyce quotes Dante, Goethe, or Robert Burns, the “corrected” readings do not follow the manuscript but go against Joyce to follow an edition set up in Hamburg. If there is any merit at all to the Odyssey Press text printed in Hamburg in 1932, it would come from Joyce’s close contact with Stuart Gilbert, who looked over the Hamburg proofs. Gilbert also provided the introduction, text, and further changes for the Limited Editions Club edition illustrated by Matisse in 1935. In the Limited Editions Club newsletter for October 1935 we read that it is “the most complete possible text” with “corrections suggested to Mr. Gilbert by James Joyce himself!” (exclamation in original). The correspondence of Joyce, Gilbert, Paul Léon, Matisse, and the Limited Editions Club is extant. (Mr. Erik Stocker of the St. Louis Public Library traced these letters for me.) More than one blunder in The Corrected Text can be traced to ignorance of this material and failure to collate the Limited Editions Club text with the Hamburg text of 1932. Thus the 1984 and 1986 editions overrule Joyce’s manuscripts 110 times to follow Gilbert’s 1932 editing, explicitly deny the debt to Gilbert, and then neglect his further work in 1935 for the Limited Editions Club.

While on the subject of unexamined archives, I should note some preliminary findings. Extraordinary materials that go unmentioned in the Synoptic Edition have turned up in my search of holdings in American collections. Joyce’s correspondence with Ezra Pound in which he discusses changes in the first episode of Ulysses is accessible. Letters to a typist of the “Circe” episode are extant, and two copyists of that episode are identifiable from their handwriting. The identity of another typist for the drafts that were written in Zurich is now known. A set of Shakespeare and Company proofs with Sylvia Beach’s changes exists. An annotated copy used to check the proofs of the first American edition has been found in another archive; its existence is explicitly referred to in yet another. An edition made while Joyce was alive had trial pages set from one source, then work resumed from scratch with another. The earlier pages still exist. Typescripts with Joyce’s revisions that were exhibited in 1975 at the Joyce Symposium in Paris and prominently cataloged are not lost, as the Synoptic Edition claims (p. 1740, lines 1–3).

Even correspondence about the 1937 Bodley Head text (the twey/tway botch) gullibly embraced in 1984 is available. I turned up a postcard from Joyce to the typist Claud Sykes with textual alterations; its existence was denied in The New York Times, April 29, 1985, page B2. Robert Bertholf, the curator of rare books at SUNY–Buffalo, produced the card. None of these documents is mentioned in the Synoptic Edition. The overlooked documents are listed in the registers of libraries in Austin, Buffalo, Carbondale, New Haven, New York City, Princeton, and Tulsa. What I tracked down to Paris has already migrated to Austin, where Joyce has a champion in the director of the Humanities Research Center, Decherd Turner. Scavengers, good luck! Remember, facsimiles don’t lie. But liars facsimile.


With the recent death of Richard Ellmann, Joyce’s biographer and trusted adviser to the Joyce estate, we lost the figure who might have quickly cut short the textual travesty of Ulysses: The Corrected Text. It was Ellmann’s endorsement of the “synoptic” 1984 text that set the stage for the 1986 trade release. Under the title “The Big Word in Ulysses,” in these pages (The New York Review, October 25, 1984)—reprinted as “Finally, the Last Word on Ulysses” (The New York Times Book Review, June 15, 1986)—Ellmann emphasized the addition of a single passage alleged to alter the meaning of Ulysses. So compelling was Ellmann in support of inserting the words “Love, yes. Word known to all men” that his review of 1984 became the Preface to Ulysses: The Corrected Text. The passage was actually uncovered and published by Clive Driver in 1975 in his facsimile of the Rosenbach Manuscript, and is not a 1984 “discovery” as Ellmann first thought. Stephen muses:

Do you know what you are talking about? Love, yes. Word known to all men. Amor vero aliquid alicui bonum vult unde et ea quae concupiscimus… [86U 161]

What he endorsed in 1984 Ellmann was quick to renounce. Only a handful of specialists know that the core of Ellmann’s support for the edition and the most heralded textual addition soon eroded. In The Georgia Review (Summer, 1986), Ellmann turned against “The Big Word” Love:

It is extremely helpful to have Joyce confirm that the word known to all men is love…. But there are reasons for arguing that, however much it may clarify Joyce’s outlook, it should not be included in the final text.

He then confronts the Latin love gumbo added in 1984 without an editorial translation or source. John T. Noonan, in a response to Ellmann’s first review, identified Joyce’s hash of two fragments of Aquinas. Now aware of the contorted Latin that had stumped the Synoptic Edition, but not impeded its insertion, Ellmann decided:

Unfortunately it puts together two passages that are separate in the original…in looking at the passage again [Joyce] could see that the Latin was without further explanation unintelligible…. By striking out the passage he could avoid drawing undue attention to a weakness in Stephen’s argument.

In a paper read at the Princess Grace Library in Monaco in May 1985 and published in that conference’s proceedings, Assessing the 1984 “Ulysses” (edited by C.G. Sandulescu and Clive Hart), Ellmann’s objection was more forceful. The second of the two phrases from Aquinas, he wrote,

seems vaguely confirmatory…or simply incomprehensible. Since conjecture is in order, it seems reasonable to surmise that Joyce recognized what he had written as tortured and self-defeating…. [We are] entitled to follow Joyce’s example and give up love.

So much for the single most important “restoration” in Ulysses: The Corrected Text and the central theme of its already outdated Preface. Ellmann showed uncommon courage in his reversal.

The spurious “love” addition aside, what did Ellmann think of the edition as a whole? Full page ads in the literary weeklies quoted him in banner blurbs:


—Richard Ellmann

This is far and away the most widely known puff for the edition and is still in use in both America and England. Yet Ellmann never spoke such words. In a letter to me dated “22.iv.985” (sic) Ellmann wrote:

Incidentally, the quotations from me in the press were inaccurate. “Stunning” was not one of my words, though of course I was approbatory.

After this clarification, those promotions can blaze:


—Richard Ellmann

Today even the “approbatory” epithet seems too strong as Ellmann’s final word on the text that he watched disintegrate.

It was unfortunate that The New York Review of Books assignment fell to Richard Ellmann in 1984, because he was the principal adviser to the Synoptic Edition, was named on its double title page, and under most circumstances would be expected to decline reviewing a work he had advised for seven years. That his piece became the 1986 Random House Preface and appeared in The New York Times Book Review means that the godfather to the text was the only reviewer in the two most influential American literary journals. The pattern recurred in England when the Times Literary Supplement ran Hugh Kenner’s fulsome accolade, utterly unaware of Kenner’s relentless American campaign in favor of “The Computerized Ulysses” (as he called it in Harper’s, April 1980). For years, most Joyceans (myself included) knew about the long-awaited but darkly shrouded monument-in-progress only through Kenner’s writings.

Without appearing to build a conspiracy theory, I must add that the James Joyce Quarterly assigned its only review of the new Ulysses to Michael Groden, who had long been a collaborator in the edition and was named on the title page of both the 1979 edition prototype and the finished 1984 text.

And so it fell out that the players reviewed the play and found it pleasing. No wonder it has taken a while to uncloak this imposture. (My own study, “An Inquiry into Ulysses: The Corrected Text,” for the Papers of the Bibliographical Society of America, has taken more than three years to prepare. It will list, by page and line number, hundreds of errors not mentioned in the present article.)

In 1986, instead of the promised “Ulysses as Joyce wrote it,” a radical version forged of shoddy scholarship and puffed out with grandiose claims pushed Joyce’s authorized version off the shelves. Flawed it may be, but classic Ulysses is the better text. Because The Corrected Text is the only version being printed today, and is selling about 100,000 copies per annum through Random House and Penguin in England, the steps to withdraw it require some thought.

The publishers are victims as much as Joyce’s readers and certainly are not to blame for this fiasco. Fortunately, they stand to lose only the printing costs of the corrupted copies now warehoused. The stock should either be destroyed or all copies given a new title page and jackets without the word “corrected.”

The choice of a replacement is easy. Of the Modern Library imprint of Random House—priced only a dollar above the Vintage paperback of the 1986 text—there are some copies remaining. Orders of Ulysses can henceforth be filled with Modern Library copies reproducing the text of 1961. I propose that starting immediately the 1986 text be withdrawn and all American orders for Ulysses be filled with the Modern Library text, except for special requests for The Corrected Text by scholars and research libraries.

With the deceptive and inaccurate Corrected Text out of the way, Joyce’s publishers and estate will need to consider editing Ulysses afresh. If a foundation agrees to help, the scholarship can get under way. Until then, the Modern Library edition, the book roughly as Joyce last saw it, is the best we have, and Random House can start shipping copies tomorrow if it wishes.

This Issue

June 30, 1988