GENETIC JOYCE STUDIES - Issue 13 (Spring 2013)



Tackling the “errears and erroriboose”: Another Look at the Rose/O’Hanlon Finnegans Wake

Terence Killeen,
James Joyce Centre, Dublin



Many of the issues that arise in connection with the Danis Rose/John O’Hanlon edition of Finnegans Wake have been comprehensively dealt with by Wim van Mierlo in the Spring 2012 number of Genetic Joyce Studies.[1] The purpose of this essay is to supplement van Mierlo’s piece with some further examples of the editors’ procedures, which it is hoped may help to illuminate the edition’s underlying principles and some of its problems. Some additional information about the strategic marketing and publicity aspects of the edition’s appearance is also offered.

   As van Mierlo notes, the reception of this edition has been – and remains – surprisingly muted. Hence my desire to revisit it here, following on from his fine contribution. One would have thought that a new edition of Finnegans Wake – in fact the first real edition – would be seen as the second most important thing that could happen in Finnegans Wake studies, second only to the discovery of some long-lost manuscript. It is of course the case that scholarly reactions necessarily take time, and this “due process” is not helped by the fact that the edition lacks an apparatus that would explain the decisions taken. This lacuna has also hampered commentary in the newspapers and in the more literary journals.[2] So far, the only scholarly commentary of which I am aware is van Mierlo’s piece in this forum, and his earlier review in the James Joyce Literary Supplement of which his GJS piece is an expanded version.[3]

   If the response so far has been notably subdued, this is not the fault of the editors or publishers. Contrary to the impression that may have been given, a very considerable effort was made to obtain as high a profile as possible for this work. Its launch in the plush surroundings of Dublin Castle in March 2010 was “a great affair”, to coin a phrase, with the late Minister for Finance, Brian Lenihan, doing the honours and considerable domestic publicity. While it might not be quite true to say that the elite of Erin hung upon Rose’s lips, it was certainly, by any standards, an “event”. It contrasted strikingly with the launch in 1997 of Rose’s “Reader’s Edition” of Ulysses, which was much more low key. A cheerleading article by the journalist Bruce Arnold, a consistent supporter of Rose, appeared earlier in the Irish Independent, and was, I think, the first public announcement of the edition.[4] After the launch in Dublin, another highly glitzy event was held in New York: the Houhnyhm website is adorned with pictures of Rose with luminaries such as Sir Christopher Ricks and David Greetham at this occasion, which also, I believe, featured readings by Professor Denis Donoghue and the poet Nick Laird among others. A planned launch in London does not seem to have transpired, but the book was again promoted by Rose on a visit to Sydney, Australia, and another event was held in Boston. All this is evidence of the editors’ and publisher’s wish for at least a succès d’estime; given the book’s price, more could hardly be expected. The idea seemed to be that while very few people would or could read the edition, the degree of prestige and publicity surrounding it, and its obvious and admirable quality as a sheer book, an artefact, would suffice to establish its credentials in the world at large. By and large, that goal was achieved.[5]

   The difficulty for both the supporters and the critics of this project is the void at its centre: the absence of the normal apparatus that would accompany any scholarly, critical edition.[6] Even Bruce Arnold, who finally got around to penning a review of the work on the publication of the cheaper Penguin edition in April last year, did concede that “it is true that it would be of great help to scholars if we could now at last have alongside this clear-reading text Rose’s much desiderated hypertext. . .”[7] Although it was never explicitly stated, it seemed clear that the reason for the absence of this apparatus was the fear of legal action from the James Joyce Estate. This fear, given the Estate’s track record and the history of its relationship with Rose, was a very real one; the editors and publishers were taking a risk in bringing out this publication at all, and the inclusion of the apparatus, with its necessary use of quotations from the drafts and notebooks, might well have inflamed the situation further. In fact, the Estate has not so far taken any action over this edition and its ability to do so has been greatly lessened by the ending of copyright in Joyce’s works in the EU at the start of 2012.

  This change in the legal status of Joyce’s works makes it now very surprising – not to say inexplicable – that Rose has still not released the apparatus – the hypertext, as he calls it – behind this edition. Without pretending to legal expertise, the consensus seems to be that all of Joyce’s published works, in most EU countries, are now out of copyright. It is also generally considered that the appearance of most of the drafts and notebooks in the James Joyce Archive during the 1970s constituted publication of those materials.[8] This should mean that Rose would now be quite safe in publishing the vast bulk of his apparatus – the only parts that might be problematic in some countries would be the very few materials that are not included in the Archive volumes, and this, one imagines, could be easily circumvented. It is therefore a continuing mystery as to why this publication has not occurred, a mystery on which this essay, unfortunately, cannot throw any light.

  We remain therefore in the uncomfortable position of having to try to assess this edition without the benefit of the scholarly apparatus one expects in any serious attempt at editing. There are just two potential sources of information about the edition’s principles and procedures that might be of assistance: one is the booklet that accompanied the publication (its contents were incorporated unaltered into the Penguin version), containing contributions from the people I named in the first endnote, and the other is a short item called “Rationale” that appeared on the Houhnyhm website some little time after the edition first appeared.

 Regrettably, the booklet is practically useless as a source of enlightenment as to the edition’s rationale. This is no reflection on the distinguished contributors: only one, Gabler, is a Joyce textual scholar (Greetham is indeed a textual scholar, but not of Joyce’s texts) and Gabler, like the rest, and indeed like all of us, laboured under the disadvantage of not having seen the apparatus either.[9] The contribution by Greetham, lacking both a Joycean textual background and any apparent knowledge of the edition’s apparatus, is particularly pointless, while that of Deane is admirably enthusiastic but would lay no claim to scholarship in this instance. This leaves us with the two contributions by the editors, and these are also quite unsatisfactory as textual introductions to the work. Only towards the end of the second piece, the “Afterword”, are some hints given as to the principles which may underlie the work.[10] The editors declare that the correction of what are normally called “transmissional errors”, which they have undertaken, is actually “the less important of the editorial tasks” which faced them.

    They add: “The greater task lay in the restoration through emendation of the syntactical coherence of the individual sentences as they underwent periodic amplification under the writer’s revising hand. What is important is that the root sentence, considered as a logical linguistic structure expressed through syntax, retains its essential structure irrespective of its often complex expansion. In practice, yet not invariably, damage to this coherence was corrected by Joyce or one of his helpers. Otherwise it is visible in collation as a simple error. In other instances the loss or part-absence of the syntactical structure was not noticed and, as the sentence was further amplified, the damage intensified, often to the extent that its original and essential coherence is irrecoverable short of a full genetic analysis.”[11] We will return to this important statement.

  The “Rationale” that appears on the Houyhnhnm website is more helpful, as far as it goes. It does not in fact go beyond the first page of Finnegans Wake. The first of the two examples given concerns the opening sentence of the book (FW 003.1-3). Two of the three changes that Rose makes in this opening sentence can be debated (in the case of one, the change of “commodius” to “commodious”, the evidence in favour of the emendation seems to me quite straightforward: the word “commodious” is first inscribed in the second set of transition proofs (BL Add.47475-92; JJA 44:253); it becomes “commodius” on the first set of galley proofs (BL Add.47476a-1; JJA 49:005) and it is the kind of slip that could easily be overlooked) but they are certainly the outcome of arguable decisions. One might come to a different conclusion from Rose on the basis of the evidence, and in one case there is evidence that he appears to have ignored or overlooked, but a case has certainly been made.

  In relation to the two other changes that Rose makes in this opening sentence, namely the considerable extra indentation – about ten ems, I would estimate – given to the opening line and the replacement of the “and” in “Howth Castle and Environs” by an ampersand, the much decried issue of “passive authorisation” does arise. Van Mierlo does seem to dismiss this criterion a bit too lightly; in this at least he concurs with Rose, who is similarly dismissive of the concept in his “Rationale” (he would have to be, to justify many of the changes he has made). It is true that “passive authorisation” is almost without heuristic value in textual scholarship: as van Mierlo indicates, it is impossible to ground any positive assertion as to an author’s “intentions” on the basis of an absence, of a non-event. Nonetheless, a reader might well wonder how, if Rose is right about these two changes, Joyce could have failed to notice their absence from the book in its final state, given that this is the opening sentence and opening paragraph of the entire work.

   What Rose does establish with certainty is that up to the first galley proofs of the book (drawn in March 1937, BL Add.47476a-1; JJA 49:005) Joyce on every occasion either wrote or let stand (including in the transition publication) the words “& Environs” and from a very early point strongly indented the opening line. (This included the first printing of the work in transition.) The change to “and Environs” and the change of the opening line’s indentation to a normal-sized “set left” certainly occurred very late in the day and do not appear to have been made by Joyce. The main problem with Rose’s retention of the ampersand, however, is that Joyce frequently uses it throughout the early drafts of this first chapter, and indeed subsequently. In the published text these are all changed to “and”, and in these cases Rose does not seek to retain them.[12] There is every indication that in these instances the ampersand is merely a form of shorthand. It is hard to see what is different about this one case. For instance, in the frequent plays on the title “Arthur Guinness Son and Company Limited”, Joyce never uses the ampersand, and yet this is a case (a formal business title) where an ampersand would be fully justified.[13] It is particularly disappointing that Rose does not address this issue at all in the “Rationale”. As so often, we are left to speculate as to his reasoning – a singularly unsatisfactory situation. In the case of the extra indentation, a strong case for its inclusion has been made, despite the issue of “passive authorisation”.

   The second example Rose gives from this first page concerns the words “devlinsfirst loved livvy” (FW.003. 23/24). He changes this to “devlins first loved livvy.” Once again, Rose shows that Joyce initially wrote or let stand at several levels (including the transition publication) “Devlins first”, and that “Devlinsfirst” emerged only when the printer produced the first galley proofs for the book in 1937 (BL Add.4747.6a-1; JJA 49.005).

   Rose has a particular problem here because it is evident that subsequently, on the page proofs, Joyce altered “Devlinsfirst” to “devlinsfirst” and yet did nothing about the joining up of the two words. Here “passive authorisation” returns with a vengeance: we are obliged to believe that Joyce made the change from “Devlinsfirst” to “devlinsfirst” yet did nothing about the joined-up word. Rose at this point does address this issue: he argues that the joined up word is spread over two lines on the page proof – “Dev-/linsfirst” – making it more likely that Joyce might have missed the joining up. He might also have argued, but does not, that one of these changes might be called a “creative” one – the change of “Devlinsfirst” to “devlinsfirst”, while the other could be seen as editorial – the splitting of the joined-up word into two again – in other words, a correction. One was the kind of change that Joyce was perhaps more likely to want to make; the other involved a different mode, mere correcting, in which he might be less interested. In any case, Rose here, unlike in the first instance, has at least addressed this issue, and to that extent his account of this second example is more satisfactory.

   Whatever one may think of the “Rationale” that Rose, subsequent to his edition’s publication, has offered on the Houyhnhnm website, it remains the case that it concerns only the first page of the book, and a mere four changes in total. Given that Rose/O’Hanlon have stated that their edition contains some 9,000 changes[14], that leaves about 8,996 unaccounted for, or, in terms of pages, 627 pages of the standard text. This situation could hardly be called satisfactory. For the rest, we are left to infer, as best we can, what Rose’s reasoning may have been; and as we have seen, even where he has rendered some account of his procedures, there are gaps. It is possible, however, by following the abundant paper trail mainly contained in the James Joyce Archive, to trace some of these processes, and indeed van Mierlo has already done this, in some four instances, one a very brilliant piece of detective work.



In what follows, I want to give some further examples of Rose’s editorial procedures and finally to draw a few general conclusions, some of which may already be implicit in the foregoing. Before embarking on these examples, it is important to refer to the only other systematic attempt at emending Finnegans Wake, Robbert-Jan Henkes’s and Erik Bindervoet’s list of “transmissional departures” given at the end of their Dutch translation of the work and selectively reproduced in the Oxford World’s Classics Finnegans Wake published last year.[15] In their introductory “Note on the Text” to the Oxford edition, Henkes/Bindervoet make it clear (p. xlviii) that this is not intended as a fully emended edition: their listing is “random” and “circumscribed by our ad hoc practical queries” as translators. So in comparing what they have done with the Rose/O’Hanlon edition we are not comparing like with like. Nevertheless, their listing does provide an interesting and valid point of comparison.

            It swiftly emerges that there are vast differences between the two outcomes, indicating strongly how many different paths can be taken to the editing of Finnegans Wake and confirming how unwise it would be to take the Rose/O’Hanlon edition simply on trust. It may well be that the only real solution is a full variorum edition listing all the recorded states of the text. For instance, the Henkes/Bindervoet list does not incorporate any of the changes made by Rose/O’Hanlon in the opening paragraph. On the first page they list two emendations – “venisoon” to “venigsoon” (FW 003.10) and “green ever and evermore since” (FW 003.23) – which Rose/O’Hanlon do not incorporate. They concur with Rose/O’Hanlon in changing “devlinsfirst” to “devlins first”. Again, they emend FW 004.11, to read “false jiccup, what rosycrucians contested of simily emilies!”, a change which Rose/O’Hanlon do not make. These are by no means the only points where Rose/O’Hanlon and Henkes/Bindervoet part company – and quite often, the reasons for the divergence are far from clear. Thus, on FW 033.29, Henkes/Bindervoet would alter “knew and loved” to “knew and, knowing loved”. The manuscript evidence is quite unambiguous: on the first set of transition proofs (BL Add 47475-15v; JJA 45:106) Joyce clearly marks this insertion. This change was not carried over to the second set of proofs and never appeared subsequently. From what one understands of the principles underlying the Rose/O’Hanlon edition (correction of “transmissional errors” being one of its stated goals) this would seem an obvious emendation for them to make. Yet they do not make it. In the ensuing set of examples, I will indicate where Henkes/Bindervoet concur with Rose/O’Hanlon and where they diverge. For subsequent reference, it will be convenient to number the examples as they occur.


1. FW 074.16/17 reads “Words weigh no no more to him than raindrips to Rethfernhim”. Rose/O’Hanlon’s text (FW2010 059.26) reads “Words weigh no more to him than raindrips to Rethfernhim.” Thus Rose/O’Hanlon have removed the second “no”. Examination of the drafts show that Joyce first wrote “Words say no more to him” (BL Add.4742-155v; JJA45:198). In the first fair copy (BL Add.4742-193; JJA 45:219) this becomes “Words weigh no no more to him” and remains in that form thereafter. The first “no” is at the end of one line and the second “no” at the start of the next. Rose/O’Hanlon must here have reasoned that Joyce, in writing out the fair copy, failed to notice that he had already written “no” at the end of a line and thus repeated it at the start of the next. Here Rose/O’Hanlon are renewing a distinction first introduced by Hans Walter Gabler between Joyce as creative artist and Joyce as scribe: in this instance, Joyce is acting as scribe, copying out his own work for the use of a typist. Henkes/Bindervoet do not incorporate this change.


2. FW 058.24/25 reads in part “three tommix, soldiers free, of the Coldstream. Guards were walking…” FW2010 047.14/15 removes the full stop between “Coldstream” and “Guards”. Through all the drafts and right up to the final page proofs, Joyce wrote “Coldstream guards”. Then, in making corrections to the first edition, Paul Léon indicated that “guards” should be printed with a capital “G”: “Guards”:




Unfortunately, Léon’s marking is ambiguous;[16] a printer could be forgiven for thinking that Léon was also indicating a full stop after “Coldstream”. It is possible, though this is speculative, that Léon assumed that because Joyce requested the capitalisation of “Guards” it followed that a full stop would precede it. Rose/O’Hanlon’s reasoning here seems to be at least partly, perhaps largely, inferential: they are working both from the inherent implausibility of a full stop between “Coldstream” and “Guards” and also from the more solid textual fact that Joyce himself never wrote such a full stop, nor can Léon’s mark be clearly interpreted as an instruction to a printer. Henkes/Bindervoet do include this emendation; in fact Henkes has written separately about it[17]


3. FW 100.13 has “Parteen-a-lax Limestone. Road”. FW2010 079.33 reads “Parteen-a-lax, Limestone Road,”. This phrase originally appears as an addition written horizontally in the left-hand margin of the galley proofs (BL Add.47476a-202; JJA 49:425). The addition in part reads: “Parteen-a-lax, Limestone Road and cried Abies Magnifica!, not, noble fir?” Because this addition was squeezed into the margin, the dot on the “i” of “fir” could easily be mistaken for a full stop after “Limestone”, the word just above it on the previous line. This is what seems to have happened, and subsequent to this point the full stop remained on all printings of the text until Rose/O’Hanlon, rightly, removed it. And, indeed, one would like to ask those who oppose the very idea of an edition of Finnegans Wake on what basis they would justify retaining this full stop other than a nostalgia for the text we have come to know and love over such a long span. Henkes/Bindervoet do list this emendation.

            Having said that, there are certain other anomalies in this one line, as edited by Rose/O’Hanlon, that are worth mentioning. They do not insert a comma after “Magnifica!”, yet, unless my reading of the line is faulty, the comma is there. On the other hand, they do add a comma after “Road”, where no comma is indicated in the text. One can only presume they do this in the interests of clarity and correct syntax, a matter to which we will return.


4. FW 088.20 reads “And with tumblerous legs, redipnominated Helmingham” etc. FW2010 070.26/27 has at this point: “And with a stopper head, bottle shoulders, a barrel bauck and tumblerous legs” etc. The words Rose/O’Hanlon insert are first inscribed on one of the typescripts (BL Add.47472-273; JJA 46.115) done prior to the transition printing. On the final set of transition proofs (BL Add.47475-126; JJA 46. 209) they take up a full line. On the galley proofs for the book itself (BL Add.47476a-154; JJA 49.117), set from the revised transition proofs, the whole line has dropped out: there has clearly been an eyeskip on the part of the compositor. The line remained absent from Finnegans Wake between transition and the Rose edition. The four characteristics mentioned here are meant to balance the four given at FW 088.17/18 as further attributes of the accused in the trial of (probably) the Shem figure. In these four his resemblance to a bottle or barrel is being stressed, via “stopper”, “bottle”, “barrel” and “tumbler”. Once again, it is hard to see on what basis the continued exclusion of this line could be justified. And once again, Henkes/Bindervoet include it.


5. FW 029.1/2 reads “haunt of the hungred bordels, as it is told me. Shop Illicit, flourishing like a lordmajor or a buaboabaybohm, litting flop” etc. FW2010. 023.04/05 reads “haunt of the hungred bordels, as it is told me, Shop Illicit, flourishing” etc. Thus Rose/O’Hanlon have replaced the full stop after “me” with a comma. The fair copy (BL Add.47472-39; JJA 44.136) reads clearly “as it is told me, Shop Illicit, flourishing” etc. (in other words, the punctuation mark is clearly a comma). On the next level, namely the typescript made from the fair copy (Private collection, JJA 44.171), a full stop has replaced the comma after “me” and there is a sizeable gap between “me.” and “Shop”. The text remained thus throughout all its subsequent levels (including publication in transition) until Rose/O’Hanlon, who have restored the original reading. Rose/O’Hanlon’s decision here has to ignore or bypass a great deal of “passive authorisation”, in the sense that Joyce must frequently have seen the text in what became its Faber form and done nothing about it. On the other hand, it is impossible to avoid the feeling that “me. Shop Illicit” etc is so clearly wrong, it goes so contrary to the way the text runs (it in fact destroys a perfectly coherent, syntactical sentence in a way that is almost counter-intuitive), bearing in mind, moreover, that Joyce’s initial inscription supports the reading “me, Shop Illicit” that I for one would not hesitate to endorse Rose’s decision. Further implications of this editorial procedure will be discussed in my conclusion. Henkes/Bindervoet also list this change.


6. FW 114. 16-20 reads “But by writing thithaways end to end and turning, turning and end to end hithaways writing and with lines of litters slittering up and louds of latters slettering down, the old semetomyplace and jupetbackagain from tham Let Rise to Hum Lit. Sleep, where in the waste is the wisdom?” FW2010 091.1/2 has “the old semetomyplace and jupetbackagain from Ham Let Rise to Hum Lit Sleep, where in the waste is the wisdom?” By removing the full stop after “Lit.” and by capitalising “Rise”, (and of course by changing “tham” to “Ham”) Rose/O’Hanlon have ensured that the two phrases “Ham Let Rise” and “Hum Lit Sleep” are read in apposition, and the reader is not led to suppose that the full stop after “Lit.” means that “Sleep, where in the waste is the wisdom?” is a separate sentence, though of course it in theory could be. On BL Add.47475-46; JJA 46.445, transition proofs we find the words already in their standard FW form. Unusually, traces of any earlier inscription are missing. Rose therefore must have proceeded entirely in accordance with an editorial “hunch” here. It certainly now “fits” the text much better, but whether that is adequate grounds for such a change – not only from “Lit.” to “Lit” but also from “tham” to “Ham”, which seems equally without textual foundation – is questionable. This is a case where one would dearly like to see Rose’s textual apparatus, in order to discover how he grounds this emendation. This matter is discussed further in the conclusion. Significantly, Henkes/Bindervoet do not list this variant.


7. The final example concerns FW 061.36 to FW 062.21, which reads as follows:


The seventh city, Urovivla, his citadear of refuge, whither (would we believe the laimen and their counts), beyond the outraved gales of the Atreeatic, changing clues with a baggermalster, the hejirite had fled, silentioussuemeant under night’s altosonority, shipalone, a raven of the wave, (be mercy, Mara! A he whence Rahoulas!) from the ostmen’s dirtby on the old vic, to forget in expiating manslaughter and, reberthing in remarriment out of dead seekness to devine previdence, (if you are looking for the bilder deep your ear on the movietone!) to league his lot, palm and patte, with a papishee. For mine qvinne I thee giftake and bind my hosenband I thee halter. The wastobe land, a lottuse land, a luctuous land, Emeraldilium, the peasant pastured, in which by the fourth commandment with promise his days apostolic were to be long by the abundant mercy of Him Which Thundereth From On High, murmured, would rise against him with all which in them were, franchisables and inhabitands, astea as agora, helotsphilots, do him hurt, poor jink, ghostly following bodily, as were he made a curse for them, the corruptible lay quick, all saints of incorruption of an holy nation, the common or ere-in-garden castaway, in red resurrection to condemn so they might convince him, Humpheres Cheops Exarchas, of their proper sins.


At its core, (BL Add 4747-1b; JJA 45.151) this passage reads as follows (carets are used to mark Joyce’s marginal additions):


The city of refuge whither he had fled to forget & expiate manslaughter, the land in which by the commandment of with promise his days ^apostolic^ were to be long, murmured, wd rise against him ^with all that in it were^, do him hurt ^ghostly & bodily, poor jink^, as were he made a curse for them, the ^corruptible^ lay quick, the saints ^of incorruption^ of an unholy nation, the castaway in resurrection of damnation to convince him of their proper sins.  


In the second fair copy (BL Add 47472-126; JJA 45.177) this passage reads (caret marks ^ indicate the the beginning and ending of Joyce’s additions on the fair copy):


The city of refuge whither he had fled ^from the Eastmen’s land^ to forget in expiating manslaughter, the ^wastobe^ land in which by the fourth commandment with promise his days apostolic were to be long, murmured, would rise against with all which in them were, do him hurt, poor jink, ghostly following bodily, as were he made a curse for them, the corruptible lay quick, all saints of incorruption of an holy nation, the common or back garden castaway in red resurrection of damnation, so they might convince him of their proper sins.


Following extensive revision and augmentation on the first typescript, (BL Add.47472-152; JJA45 191) this passage became in the next fair copy (BL Add.47472-182; JJA45 209):


The seventh city, his citadear of refuge, whither (would we believe the laimen and their counts), beyond the outraved gales of Atreeatic, he had fled shipalone, a r[18] of the wave, from the ostmen’s d[19] on the old vic, to forget in expiati  g[20] manslaughter; the wastobe land in which by the fourth commandment with promise his days apostolic were to be long by the abundant mercy of Him Which Thundereth From On High, ^murmured^, would rise against him with all which in them were, do him hurt, poor jink, ghostly following bodily, as were he made a curse for them, the corruptible lay quick, all saints of incorruption of an holy nation, the common or ere-in-garden castaway, in red resurrection to condemn so they might convince him, first pharaoh Humpheres Cheops Exarchas, of their proper sins.


Firstly, some textual remarks. The comma between “manslaughter” and “the wastobe land” has now become a semi-colon, indicating a certain loosening and subdividing of the syntax of this sentence.

            Also, it is evident that the omission of “him” after “against” in the first fair copy and the typescript is an error, that “him” was always understood and is indeed supplied in this second fair copy. (It is of course present in the very first draft.) Again, the initial omission of “murmured” in this second fair copy (it is supplied as an addition, as the transcription makes clear) is also an oversight and it was always meant to be here, having occurred in all the earlier versions. These again are effects of “Joyce as scribe”. (This may be the point at which to remark on the particularly happy trouvaille of “wastobe land”, which incorporates both “waste land” and futurity.)


Secondly, complex though this passage is, it is in fact a coherent sentence, with multiple additions: the subject is “The seventh city”, with “ his citadear of refuge” and “the wastobe land” as subjects in apposition. The predicate (main verb) is “murmured” (as in “murmured in discontent”), with “would rise” and “do him hurt” in apposition, and the object (twice; once indirect) is “him”. This basic sentence structure, going back to the first inscription of the passage, has survived up to now despite all the elaborations that have come to surround it, and despite even the omission of one of its basic words on two of the drafts. Its burden is the by this point familiar one of HCE as civic scapegoat, as a castaway who had come to a city of refuge and now finds that city turning against him. So what is being said is not at all obscure.

On the next typescript (BL Add.4742-238; JJA 45.233) there are just a couple of additions (far fewer than on the first typescript). One, to follow the words “expiating manslaughter”, reads “and to league his pagan lot, palm and patte, with a papishee”. The second is the addition of “franchisables and inhabitands,” after “all which in them were”.  

The first of these additions is not incorporated into the transition proofs (BL Add.47475-27; JJA 45.294) so Joyce writes it in again at the same place he had previously indicated. However, while doing so, and in a fine example of the endlessly expanding nature of this work, he adds some more text: “and to league his pagan lot, palm and patte, with a papishee” becomes “and to league his pagan lot, palm and patte, with a papishee: For my qvinne I thee rape gifttake and bind my hosenband I thee haltar.”

It is at this point that the sentence falls apart: it is unable to sustain itself as a sentence under the weight of all the material that has been packed into it. The necessity (a characteristic Finnegans Wake necessity) to tell the entire story of HCE once again in this one sentence (including a Viconian cycle and another self-referential commentary on the book’s own technique) has finally overwhelmed the syntactical structure. The break is signalled on the final transition page proofs (BL Add.47475-116, -201; JJA 45.318, 319) by the placing of a full stop after “halter” (as it has become) and the capitalisation of the first letter of the next word, “The”. Similarly, the colon after “papishee” appears to have become a full stop and we now have three separate sentences, as in the current standard Finnegans Wake text, where we had begun with one.

Moreover, in the first sentence, the subject is left stranded: it has no main verb. “The seventh city, Urovivla, his citadear of refuge” is without a predicate. Similarly the second sentence “For my qvinne” is not really a standalone sentence either, since it is clearly linked to the preceding (non)-sentence as a kind of complement to it. The place where the original syntactic structure survives is in the third sentence, which does indeed have a subject, predicate and an object, indeed several of each: namely, “The wastobe land” etc as subjects, “murmured”, “would rise” and “do him hurt” as main verbs, and “him” (in “against him” and “do him hurt”) as objects.

            Rose/O’Hanlon’s procedure here is to establish (I will not say “restore”) a syntactical unity and to turn these three sentences into one:


The seventh city, Urovivla, his citadear of refuge, whither (would we believe the laimen and their counts), beyond the outbraved gales of Atreeatic, changing clues with a baggermalster, the hejirite had fled, silentiousissuemeant under night’s altosonority, shipalone, a raven of the wave (be mercy, Mara! A he whence Rahoulas!), from the ostmen’s dirtby on the old vic, to forget in expiating manslaughter and, reberthing in remarriment out of dead seekness to devine previdence (if you are looking for the bilder deep your ear on the movietone!), to league his pagan lot, palm and patte, with a papishee (for mine qvinne I thee giftake and bind my hosenband I thee haltar), the wastobe land, a lottuce land, a luctuous land, Emeraldillium, the peasant pastured, in which by the fourth commandment with promise his days apostolic were to be long by the abundant mercy of Him Which Thundereth From On High, murmured, would rise against him with all which in them were, franchisables and inhabitands, astea as agora, helotsphilots, do him hurt, poor jink, ghostly following bodily, as were he made a curse for them, the corruptible lay quick, all saints of incorruption of an holy nation, the common or ere-in-garden castaway, in red resurrection to condemn so they might convince him, first pharaoh, Humpheres Cheops Exarchas, of their proper sins. (FW2010, 49.37-50.15)


Rose/O’Hanlon have made this sentence cohere grammatically (and it now does cohere grammatically) by the device of putting the awkward intrusive phrase “for mine qvinne I thee giftake and bind my hosenband I thee haltar” into parentheses and then running the rest of the passage on. There is no textual warrant that I can find for these parentheses. There is one mark just before the words are written out for insertion on BL Add.47475-27; JJA 45.294 but this mark is clearly meant to indicate the point of insertion and does not signal the start of a parenthesis. In any case there is no closing parenthesis.

            At this point it is relevant to quote again from the booklet accompanying their edition Rose/O’Hanlon’s description of what they saw as their principal editorial task in this edition: “The greater task lay in the restoration through emendation of the syntactical coherence of the individual sentences as they underwent periodic amplification under the writer’s revising hand.”[21]   

            We are here at the nub of the issue. A great deal hangs, for Rose/O’Hanlon, on the word “restore”. It can be inferred to be their view that whenever syntactical coherence is broken, this is because of a “transmissional error”, an error made either by Joyce as scribe (and their distrust of Joyce as scribe is much greater than Gabler’s) or by some other “scribe” involved in the transmissional process. The example just cited shows the complexity of this process. This is clearly a case where these editors saw a need to “restore” “syntactical coherence”. Is the word “restore” justified in this instance? In a sense, yes, because, as we have seen, this passage did originally form one syntactically coherent sentence. However, in order to accept what Rose/O’Hanlon have done we have to ignore what looks very like a full stop after “halter” on BL Add.47475-201; JJA 319 and the definite capitalisation of “The” in “The wastobe land” (BL Add.47475-116; JJA 318).

            Based on the statement in the booklet already quoted, it is apparent that Rose/O’Hanlon reasoned that the beginning and ending of this passage had once formed part of the same sentence, that they were still umbilically linked, and that their hidden unity or coherence should be restored. In this sense van Mierlo is certainly correct to highlight Rose/O’Hanlon’s continuing preference for Joyce’s initial inscriptions over all subsequent ones, despite the many obstacles in the path of retrieving these and of making them fit with the numerous later additions and interpolations that befall them. Clearly, Rose/O’Hanlon believe (and they may well be right, at least in theory) that although Joyce would massively elaborate on an initial syntactical core, he usually did not alter this fundamental basis: it continues to underlie the additions and elaborations with which it is festooned. Hence the need to respect this core, stubborn structure, even when Joyce himself as scribe, not to mention the many other collaborators, fail to do so. Put like that, it sounds straightforward: in practice, as the last example shows, it is anything but.

            In this instance, Henkes/Bindervoet would also make changes, though not the same ones. Rather than inserting parentheses, they list a comma after “halter”, linking that part of the passage with the succeeding sentence. After “papishee” they place a colon, rather than the start of a parenthesis. A colon is, in fact, what Joyce clearly marks on BL Add.4745-27, JJA 45.294, although he equally clearly marks a full stop after “halter”, whereas they insert a comma. (And later on, on BL Add 47475-116/201 Joyce appears to have replaced the colon after “papishee” with a full stop.) In this way, they too put this sentence “back together again”. And it must be said that their method involves a less drastic alteration to the text than does the Rose/O’Hanlon version.



It is to be hoped that the examples cited above at least indicate the range and complexity of the problems facing a potential editor of Finnegans Wake. The basic situation is explained very well by van Mierlo: “Syntactical consistency is a feature of Finnegans Wake, but it is one that is hampered by persistent revision and corruption.”[22] There is a fundamental difference between “revision” and “corruption”: where syntactical incoherence can be convincingly assigned to corruption, then it is indeed valid to “restore” it. Where syntactical incoherence is due to revision, the matter is much more debatable. And of course there is a considerable grey area where it could be either factor that is behind this incoherence, and where the choices are even more problematic. (In fact, van Mierlo’s assertion of his belief that “editors correct texts; they do not correct their authors, even when the author makes a blatant mistake” is perhaps the one point where his position does not quite match the complexity of the object he is discussing, namely Finnegans Wake: the trouble is that Joyce wrote many different things, and sometimes these things are contradictory, so that an editor committed to producing a “reading text” has to choose; it is not so much a question of correcting “mistakes” as of trying to ensure a certain minimal consistency in the text one is editing.)

            Using the criterion of corruption/revision, then, and taking each of the examples in turn, we can say that Example 1 is an instance of corruption, corruption introduced by Joyce the scribe, and it is therefore valid to emend it. There remains a residual possibility that this inscription is intentional, and should therefore be preserved, but all editing with a view to producing a reading text is a matter of balancing probabilities, and corruption is here the stronger probability.

            Example 2 can also be put down to corruption, due either to a slip of the Léon pen or perhaps to confusion by Léon as to Joyce’s intentions. Again, it’s perhaps a marginal call, but such calls are precisely the function of an editor.

            Example 3 is clearly a case of textual corruption, and, as mentioned already, it is hard to see on what basis the retention of the standard FW reading could be justified.

            Example 4 is also a case of textual corruption (in this case an eyeskip) and the restoration of the missing line is fully justified.

            Example 5 is again a transmissional error; the only basis for its possible retention would have to be “passive authorisation” and as we have seen, and as van Mierlo stresses, this is a highly suspect criterion.

            Example 6 is problematic, because, to use a phrase that appears in many such essays as this, “the unfacts, did we possess them, are too imprecisely few to warrant our certitude.” (FW 057.16/17). Unusually, the full textual record for the form “Hum Lit. Sleep” is not available nor is the record for the form “tham”. While one appreciates that the passage in Rose’s edition reads a good deal more clearly than it did in the standard text, and while acknowledging an instinctive sense that something is “wrong” with the passage as it stands in the standard text, the alteration introduced by Rose seems on the evidence unwarranted.

            Example 7 is clearly an example of “syntactical consistency” being hampered by “persistent revision”, to use van Mierlo’s terms, rather than by corruption, so different considerations apply. The editors’ project to ensure “the restoration through emendation of the syntactical coherence of the individual sentences as they underwent periodic amplification under the writer’s revising hand” would appear to be perfectly exemplified by this passage. Ultimately, my own feeling is that this “restoration” is not here justified, because the scope and extent of Joyce’s revisions finally go beyond the goal of syntactic coherence, and such an overstepping of the limit cannot be inadvertent. But in the best traditions of having one’s cake and eating it, I would like to append a note to any edition explaining what had caused this situation.

            And such a wish may go some way to clarifying the particular difficulties and delights of getting involved with Finnegans Wake: if one gets into the book at all, one tends to get very far in, so that one is willy-nilly involved in textual complexities which were not at first on one’s agenda at all.

            A final remark: van Mierlo, in citing the statement about syntactic coherence already quoted twice in this essay, is right to draw attention to the uncomfortable echoes it contains of Rose’s practice in his 1997 “Reader’s Edition” of Ulysses, namely the dreaded “copyreading” whereby the editor apparently felt justified in making whatever changes he wished at any point where he felt the text was unsatisfactory – unsatisfactory according to obscure criteria. In fact, though, as van Mierlo also concedes, this is not a factor, as far as can be ascertained, in the Rose/O’Hanlon Finnegans Wake. The examples cited have a textual basis of some kind. The closest one comes to copyreading is perhaps Example 6, where the textual basis seems very scanty and where it looks as if Rose/O’Hanlon is proceeding by instinct. But even in this case, the change is not without a certain textual ground. So the highly controversial Rose Ulysses should not be used as a stick with which to beat this edition.

            Nevertheless, it must be said that these editors have set themselves two tasks: one, the correction of transmissional errors, is a traditional one for an editor; the other, the “restoration of syntactical coherence”, is not, and is bound to be controversial. Controversy, however, is nothing new in Joyce editing, and I do feel that for its very rational and generally well grounded approach to the first of these tasks, this edition is to be welcomed. It would be doubly, trebly welcome if only the accompanying apparatus were available.



[1] Wim van Mierlo, “ For polemypolity’s sake”: Editing Finnegans Wake — A Consideration and Review” Genetic Joyce Studies, Issue 12 (Spring 2012). Danis Rose’s and John O’Hanlon’s edition of Finnegans Wake was published in March 2010 by Houhynhm Press, Dublin in a luxury-type format, accompanied by a booklet with contributions from Rose and O’Hanlon, Seamus Deane, Hans Walter Gabler and David Greetham. The edition was republished in May 2012 by Penguin Books without alteration at a much more affordable price.

[2]  Michael Wood, “Quashed Quotatoes”, London Review of Books, 16 December, 2010; Gordon Bowker, “Restoring James Joyce’s book of the night”, TLS, 4 July, 2012. Considering the difficulties under which they laboured, the reviews, that by Wood in particular, are very perceptive.


[3] Wim van Mierlo, “Editing the Wake”, James Joyce Literary Supplement, Volume 25, Number 2, Fall 2011.


[4] Bruce Arnold, “Finnegans Second Wake”, Irish Independent, 6 March, 2010.


[5] The only sour note in this general harmony was struck by my own review in The Irish Times, (“A Flawed ‘Finnegans’ Wake-up Call”, The Irish Times, March 13, 2010). In that review, I felt it was important, “preparatory to anything else”, to highlight the absence of the standard scholarly apparatus that should accompany an edition such as this. It was also necessary to take issue with the bland assumption that all the 9,000 changes in this edition should be taken on trust. The fable of the “Emperor’s New Clothes” came to mind: one was not supposed to mention the glaring fact that no support was being provided for the assertions made. Nothing in the review indicated that the changes made were necessarily wrong – the point was that they were not accounted for. It is to address the question of the actual changes made and their validity that the present essay – and my contribution to the Prague International James Joyce Foundation symposium of 2010 – have been undertaken.   


[6] The obvious contrast for Joyceans is the meticulous presentation by Hans Walter Gabler, in his critical and synoptic edition of Ulysses of all the textual evidence for his decisions. Some of those decisions may have been wrong, some of the execution of the project may have been flawed, but the evidence for these decisions is certainly there.


[7]  Bruce Arnold, “ Review: Literary Fiction: The Restored Finnegans Wake James Joyce – Edited by Danis Rose and John O’Hanlon”, Irish Independent, 19 May, 2012.


[8]  See


[9]  Gabler makes this handicap quite explicit in his contribution: “It is regrettable that an ‘apparatus’ comes neither here, with the reading text, nor separately in electronic form” (Booklet, page 14) and again, “While we await information from the data bank in support of the reading offered in this case, we may take it on general trust from the editors that modifications such as this one are not extraneous emendations.” (Booklet, page 15)


[10] This dearth of information contrasts greatly with the quite detailed introduction to Rose’s “Reader’s Edition” of Ulysses, which, whatever the work’s other flaws, does provide a full rationale, even if not a full apparatus, for the controversial choices made. See James Joyce, Ulysses, edited by Danis Rose (London: Picador, 1997).


[11] Finnegans Wake edited by Danis Rose and John O’Hanlon, Dublin: Houyhnhnm Press, 2010, Booklet: page 36. Van Mierlo also rightly draws attention to this important passage.


[12]  See, for instance, David Hayman, A First-Draft Version of Finnegans Wake (London: Faber and Faber, 1963), pages 46, 47, 58 and virtually passim. To take just one instance, Joyce initially writes “upjack & hackums” (FDV p. 49). This becomes in the final text “upjock and hockums” (p. 7). Rose in this case does not retain the ampersand: his text (p.7) also reads “upjock and hockums.” Rose’s claim, incidentally, that ‘Joyce, it seems, wanted to avoid the distancing effect of the word “and” in this context’ is highly debatable; it is hard to see what the grounds for the assertion are.


[13]  The current archivist in the Guinness company confirms that in the late nineteenth century, after incorporation, the formal title was “Arthur Guinness Son and Company Limited”.


[14]  Booklet, page 7


[15]  Finnegans Wake, translated by Robbert-Jan Henkes and Erik Bindervoet, (Amsterdam: Athenaeum-Pollak and Van Gennep, 2002). Finnegans Wake, edited by Finn Fordham, Robbert-Jan Henkes and Erik Bindervoet, (Oxford: Oxford World’s Classics, 2012).  

[16]  This passage is in the final Finnegans Wake page proofs held in the McFarlin Library, University of Tulsa, and does not appear in the James Joyce Archive. It is discussed by Robbert-Jan Henkes and Erik Bindervoet, The Tulsey Town Treasure Trove: Preliminary Report on the Rediscovered Finnegans Wake Proofs, Genetic Joyce Studies, Issue 3, Spring 2003. Their conclusion is much the same as my own.

17  We confront here the question of whether “plausibility” can ever be a criterion for an editor of Finnegans Wake. Is this book so special, so unique, that normal editorial considerations of likelihood etc cannot apply? This issue is discussed in two earlier essays in Genetic Joyce Studies: Robbert-Jan Henkes’ and Erik Bindervoet’s ‘Finnegans Wake: The Corrected Text’ (GJS, Issue 4) and Sam Slote’s Soundbite against the Restoration (GJS, Issue 1). Van Mierlo also addresses these essays in his piece. In a sense, both these earlier essays have been overtaken by events: whether a new edition of Finnegans Wake should or could be done or not, it has been done; and it cannot be argued, I believe, that the edition is such a disaster that it proves the case of those earlier arguments. I do believe that “plausibility” is a relevant criterion here, having regard also of course to the textual evidence.

[18] The JJA reproduction is faulty here but the word is “raven”.


[19]  The incomplete word is “dirtby”.


[20] The incomplete word is of course “expiating”.


[21] Booklet, p. 36


[22] It follows from van Mierlo's statement that syntactical coherence in Finnegans Wake is not a shibboleth. We can take one example which has figured largely in this discussion, namely Joyce's use of full stops. Dirk Van Hulle, in “The Lost Word: Book IV” in Luca Crispi and Sam Slote (editors), How Joyce Wrote “Finnegans Wake” (Madison: University of Wisconsin Press, 2007) pp. 454-5, has drawn attention to the placing of full stops after the definite article on at least three occasions, each of which is followed by a reference to a closing door. These are pages 20.17, 257.27 and 334.29-30. (Some of these full stops, as Van Hulle notes, were inserted at a very late stage, indeed one at least at the very last possible point, being marked in the unbound copy of the book in which Joyce/Paul Léon entered corrections in the summer of 1940.) The function of the disruptive full stop in these instances is quite clear: it is to strengthen the sense of closure, though also, as Van Hulle suggests, with a hint of a new beginning (no full stop in Finnegans Wake is ever final, especially since the only one that would be, the one that would end the book, does not exist.) The best way to be sure of the status of these disjunctive full stops is to follow the textual evidence.