Denis Donoghue examines the new edition of ‘Ulysses’

Denis Donoghue

Joyce’s Ulysses was published on his 40th birthday, 2 February 1922, in a limited edition of 1000 numbered copies. The text was full of misprints, as Joyce irritatedly knew. As late as November, he had been tinkering with the last chapters, getting further detail from Dublin – ‘Is it possible for an ordinary person to climb over the area railings of No 7 Eccles Street, either from the path or the steps, lower himself down from the lowest part of the railings till his feet are within 2 feet or 3 of the ground and drop unhurt?’ he wrote to his Aunt Josephine – and the galleys were demanding attention he couldn’t give them. On 6 November he complained to Harriet Shaw Weaver that ‘working as I do amid piles of notes at a table in a hotel I cannot possibly do this mechanical part with my wretched eye and a half.’ He evidently decided that he couldn’t do much about the printer’s errors in time for the birthday, but he hoped they would be corrected ‘in future editions’.

Joyce wrote Ulysses by hand, and his arrangements for having the manuscripts typed were so loose that errors were inevitable. Some of the typists thought the writing would be improved by more orthodox punctuation. Further errors were made in the printing-press. None of the compositors at Dijon knew English, except for the foreman, Maurice Hirchwald, who knew only enough to decide that he could correct Joyce’s vagaries. Joyce tried to undo some of this damage, but the job was too much for him. He corrected thousands of misprints on the galleys, but missed about two thousand.

Later editions of Ulysses haven’t been much better. Between 1922 and 1933, the book couldn’t be published legally in the United States, so there was no merit in doing heavy work on the text. When Judge John Woolsey lifted the ban on the book – 6 December 1933 – Bennett Cerf set about issuing a new edition. But the text he gave his printers was a copy of Samuel Roth’s facsimile pirated edition, printed in New York: its errors remained to corrupt Cerf’s Random House edition of January 1934.

I needn’t recite the history of the publication of Ulysses in England. The gist of the whole matter is that the editions which most people read, the Penguin ‘reprinted with corrections’ in 1971 and the Random House Vintage Books edition, ‘corrected and reset’ in 1961, are about equally erroneous: seven errors to a page, according to Professor Gabler’s count. Some misprints are common to both editions, but each edition has its own errors, too.

On 6 December 1966 Jack Dalton gave a lecture at Cornell University on the text of Ulysses, quoting several instances of its corruption, and promising to produce, on contract to Random House, a new and satisfactory edition. In the event, he didn’t live to keep his promise, and the work passed to a team of bibliographers, led by Professor Gabler.

The main problem in devising a reliable text of Ulysses is that there is no complete manuscript. For some chapters there are drafts, jottings, notebooks, the British Library Note-sheets, fair copies – the most important things are in the Rosenbach Manuscript at Philadelphia, and other collections at Cornell, Buffalo and other universities. Rosenbach is the crucial holding, its holograph notation, as Gabler says, ‘marks a decisive point of consolidation in the compositional development’. But it doesn’t solve every textual problem. For one thing, Rosenbach is ‘full of erasures indicating revisions during the fair-copying’. For another, Joyce continued to work on chapters even after he had given the manuscript to a typist: he did not let the typescript go without further ado to the printer.

Gabler’s aim has been ‘to uncover and to undo the first edition’s textual corruption’. His main principle has been to distinguish ‘the documents of composition’ – which he regards as authoritative, unless they can be shown to be faulty – from ‘the documents of transmission’ – which he regards as potentially faulty, unless they can be proved to be authoritative. So Joyce’s autographs are separated from the typescripts, the serial versions in the Little Review and the Egoist, the proofs of the first edition, and the first edition itself. But the distinction is hard to maintain, since transmission becomes composition as soon as Joyce tinkers with it. In any case, Gabler has tried to assemble, as his copytext, ‘a continuous manuscript text for Ulysses, extending over a sequence of actual documents’. His principle is a narrative one, as if he were reconstructing a story. Or an archaeological one, deducing a complete structure from related fragments. In the new edition, the left-hand pages record the entire history of each word and accidental, so far as it can be established and indicated by a complicated system of notation: the right-hand pages give the clean text of the book without interruption or comment. Textual explanations and justifications are set out at the end.

I’ll give a few examples, in a minute, of the differences the new edition makes. But it’s worth saying at once that they’re not merely a matter of correcting ‘Steeeeeeeeeephen’ to ‘Steeeeeeeeeeeephen’ in the Telemachus chapter and ‘Pprrpffrrppfff’ to ‘Pprrpffrrppffff’ in the Sirens chapter, as Craig Raine pretended to think a few weeks ago in the Sunday Times. If you are totally indifferent to misprints, you won’t even consider buying or otherwise consulting the new Ulysses. ‘On the whole,’ Raine claimed, about misprints, ‘I couldn’t give a fuppenny tuck.’ Not even if the botched printing were of his own verses? More to the point: suppose it were discovered that the printing of Paradise Lost is botched, with lines and half-lines dropped, wouldn’t English poets, critics, and common readers agree that a new edition should be produced, especially if the hard labour involved were to be done, as it probably would be, by German or American scholars?

The full text of this book review is only available to subscribers of the London Review of Books.

You are not logged in

[*] A two-part essay by William Empson, advancing this interpretation, was published in this journal: Vol. 4, Nos 15 and 16.