Genetic Joyce Studies - Special Issue JJA (Summer 2002)
The James Joyce Archive from an Archival Perspective
Luca Crispi

I am a member of that generation of Joyceans Philip Herring predicted would come to use and fully appreciate The James Joyce Archive.[1] In fact, I am one of an increasing number of scholars who find it insufficient to read Joyce’s works without relying on and referring to the materials reproduced in The Archive . For genetic readers, the published editions of Ulysses and of Finnegans Wake, for example, are mere moments in a much richer and more complex reading experience that is founded on the texts’ extensive pre-history as is manifest in their manuscripts. This more expansive reading experience is based on the textual layers that are represented in The Archive.  

Finnegans Wake II.2, the “night studies” chapter, for instance, had an extremely complicated and convoluted textual history that preceded the imposition of its distinctive schoolbook format. In fact, its principal tale, “The Triangle” (282.05-304.04), appeared twice in print without it.[2] Furthermore, the chapter as a whole was significantly restructured again just a year before the Wake was published. Seen from this perspective, the transformations Joyce forced on what were already fairly stable texts account in large measure for its disrupted narrative and help to explain why reading this chapter can be such a disorienting experience.  

The Archive was immediately recognized as an intellectual and scholarly triumph, the pinnacle of forty years of Joycean textual and archival work. It was also an incredible practical achievement by its editors (as well as by its publisher, Gavin Borden of Garland Publishing).[3] But today, twenty-five years after its publication, at least two intertwined questions still persist. The first and most important is why are there still so few scholars and readers willing and able to undertake archival textual research? The Archive affords a monumental opportunity to explore the creative evolution of some of the most important texts of the twentieth century, from their earliest notes and drafts, through all of the texts’ various draft levels, to the works as published. In spite of the greater access permitted by this productive resource, genetically-informed publications have remained relatively few. Nonetheless, after all is said and done, The Archive had its intended impact: it multiplied the number of Joycean manuscript and textual readers around the world. Today scholars in Sydney, Tokyo, Antwerp, and Paris study Joyce’s manuscripts as do others in London, Buffalo, Dublin, New Haven, and Austin. But their ranks remain disproportionately small compared to the number of scholars who study Joyce from various theoretical perspectives.


The second persistent question is what impact did the reproduction of the manuscripts in The Archive have on the use and relevance of the original documents in the actual, physical archive? In his review in the JJQ, Herring claimed: “it is today rarely necessary for Joyceans to visit Joyce manuscript collections.”[4] Others argued precisely the opposite: their reproduction would encourage, in fact necessitate, scholars who used The Archive to consult the original material in the archive. Herring’s prediction proved more accurate. Internal debates here at University at Buffalo (and elsewhere I am sure) addressed the impact of reproduction on the scholarly value and the material preservation of the manuscripts.[5] Some believed that the reproduction of our holdings diminished (both the financial and scholarly) value of the manuscripts. Others maintained that their reproduction would ensure their preservation because fewer people would need to handle the fragile and priceless originals, stored safely in vaults, in archival boxes, in folders, nestled in acid-free paper.


For twenty-five years there had been a steady stream of researchers to our Joyce Collection, augmented by those who requested microfilm copies of the manuscripts for study around the world. Although there were never as many visiting scholars as one would predict given the scholarly and cultural importance of this material, our records show that in the years immediately following the publication of The Archive there was a sharp decrease in their numbers.[6] This situation can be attributed to many factors, including the basic fact that so many of these same, committed researchers were also the editors of The Archive itself: Michael Groden, Hans Walter Gabler, David Hayman, A. Walton Litz and Danis Rose. As for most other readers and scholars whose interest should have been encouraged by this vast publication, they may simply have been overwhelmed by the field as represented by those sixty-three volumes and not pursued the matter any further. There will probably always be few scholars who have the patience and endurance to do archival work, especially when it involves the conflicting demands such work makes on their professional and personal lives, as well as the inconvenience and the expense of travel to many distant holding libraries. On the other hand, The James Joyce Archive had its most profound impact on the burgeoning community of Finnegans Wake manuscript (and especially notebook) scholars around the world. The recent publication of the first three volumes of the Finnegans Wake Notebooks at Buffalo edition (Brepols 2001- ) is the direct result of the dissemination of The Archive around the world and the groundbreaking work in the field it provoked.


I realize that my experience with The Archive is atypical, differing in several fundamental ways from those of my colleagues. I encountered The Archive alongside my principal assignment of revising and augmenting the long outdated catalog of the Joyce Collection at Buffalo.[7] My first task was to bring the bibliographic and textual information concerning our holdings in accord with the research embodied in The Archive. My work as a bibliographer, cataloger and genetic critic would have been that much more difficult if I were not able to continually consult the virtual archive of Joyce’s manuscripts in The Archive. In so many different ways, what I know of the Joyce manuscripts at Buffalo and elsewhere has been shaped by the introductions and arrangements of the documents in that series.


Not only did The Archive reflect accumulated scholarship, it continues to encourage further research. For example, Spielberg’s ordering of Cyclops manuscripts (V.A.6 and 8) had been incontrovertibly reversed by Groden’s research and the new Buffalo catalog will have to indicate that. Similarly, the catalog will take into account the work done on the revised chronology of the Finnegans Wake notebooks by Roland McHugh, David Hayman, Danis Rose and recently and more conclusively by the editors of the Finnegans Wake Notebook edition. Given the limits of what was then known concerning Joyce’s manuscripts and writing practices, Spielberg’s catalog was remarkably accurate. The same is true of The Archive: very few errors of commission in The Archive have come to light over the years.[8] Only recently has it been ascertained, for example, that the loose pages at the start of Finnegans Wake notebooks VI.B.10 and B.6 had been reproduced in a reversed order. This was revealed by internal bibliographical information and confirmed by external, textual considerations.


In 1999, several decades after the last major new Joyce manuscript had come to light, I naively wrote of the Buffalo Joyce Collection that its re-cataloging would “[…] mark an end of an era in some ways because it is unlikely that any other assemblage of Joyce material of this scope and depth remains to be uncovered.”[9] Since then two spectacular Ulysses manuscripts have been unveiled, one of the Circe and the other of the Eumaeus chapters.[10] Even though Joyce himself had documented the Circe manuscript in a letter to John Quinn, its existence seemed to surprise most of us (LIII  40). On the other hand, on textual and theoretical grounds, Gabler had postulated that such a Eumaeus manuscript must have existed, though there was no collateral, external evidence of it.[11] Therefore, when it surfaced (and with such an unexpected provenance), it should have made it clear that other “lost” manuscripts would probably continue to be unearthed. These and other, similar experiences have taught me (and I am sure others) an important lesson: count on the fact that new and often unexpected evidence will revise or overturn determined historical and textual records. In fact, these newer manuscripts will inevitably upset and redefine our current interpretive models. Textual and archival scholarship is an ever-developing process of discovery and insights.


The editors of The Archive did an outstanding job of accounting for all of the then-known manuscript material. Furthermore, they established a system of genetic coding for Finnegans Wake that allowed for textual material to find its place within a well-articulated structure.[12] With this system a particular text can be synchronically and diachronically defined. If we take as our example the first set of galley proofs for II.2 (BL 47478, fs. 329r-355r), dated February 1938, that is coded as 1.13/2.11/3.13/5.3-6.5/7.4/8.14/9.12, we note that as published this chapter is composed of nine sections, each of which developed in different stages from three to thirteen drafts, as follows.[13]


The earliest drafts of the first and last sections (§§1-3 and 9) were written in 1934 and went through as many as ten different stages of textual development, including several holograph drafts, a fair-copy and up to four typescripts and then proofs before they were published as the “opening and closing pages” of II.2 in transition 23 in July 1935: §§1.9/2.7/3.9/9.8. Over the next two and a half years, these texts went through several more draft stages and were published together once again as Storiella as She was Syung in February 1938, at the same time that Joyce was reformatting and reworking the other section of II.2 to conform to the chapter’s new layout.


Section 4, “Scribbledehobbles,” had a much more complicated textual history.[14] When he started writing this piece in 1932, Joyce intended it to segue from II.1 and introduce the children “at their pensums.” But after drafting, fair-copying and having it typed twice, revising it extensively at each stage, he abandoned it in 1934 in favor of the newer pieces he had written (§§1-3). Joyce only returned to this piece in November 1937, cannibalizing a copy of its second typescript (4.5') to refashion a new piece, §5. This narrative also underwent several drafts before it was set in galley proof for Finnegans Wake where it appeared in print for the first time.


Sections 6 and 7 were first written along with §§1-3 and 9, but were set aside along with the “Scribbledehobbles” piece in 1934 and only taken up again and reformatted like the newer sections, along with §5, for this setting of the galley. Joyce wrote §8, “The Triangle,” in the summer of 1926. It was the first section of II.2, in fact of Book II, to be written and was always the centerpiece of the chapter. Like the other sections, it was redrafted, typed, copied and, in this case, even printed two times before it joined all the other sections here for the first time in this earliest setting of the galley proof. Only in February 1938, almost twelve years after it was begun and just sixteen months before Finnegans Wake was published, did II.2 become the chapter that most readers of the work would recognize. The Archive’s draft code allows readers to follow the temporal development of the individual sections of texts as they emerged individually and then coalesced, becoming parts of a more integrated chapter and work.


Rose’s intricate and detailed draft code for Finnegans Wake makes accounting for newly uncovered draft material a relatively simple matter. For example, among Beach’s previously uncataloged business correspondence at Buffalo there are more than fifty textual fragments that document the composition and evolution of Finnegans Wake (Buffalo MSS. VI.E–I). Most of this material consists of abandoned bits and pieces of paper that Joyce gave Beach as tokens of thanks for her innumerable duties and invaluable service in getting “Work in Progress” into print. Because they were uncataloged at Buffalo at the time, none of these textual fragments were reproduced in The Archive.


One example can stand in for an entire set of broadly similar instances. It is difficult to describe the excitement I felt when I found a heavily revised copy of transition. The text made it obvious that it was a copy of a “Continuation of a Work in Progress” in transition 8 (November 1927), but there were nonetheless many genetic issues that still needed to be resolved. Fortunately it was a relatively easy endeavor to answer all of these questions conclusively by consulting the archive and The Archive. The quality of the paper and the fact that this document was printed and paginated in signatures indicated that it was either page proofs for publication in transition or a revised copy of the text as printed in transition for another publication. I compared the base text with a copy of transition (Joyce’s own in this case) and they were identical. I then compared the revisions and emendations to the two other marked sets of transition 8 as reproduced in The Archive  (Yale 7.7, fs. 17-35 and BL 47474, fs. 248r-257r; JJA  48.218-37). In due course, I determined that this manuscript may have been used to set the galley proofs for the Crosby Gaige edition of Anna Livia Plurabelle (29 October 1928), but was certainly used to verify the setting of the text on the galleys. This copy contains the identical revisions and emendations as the Yale and British Library manuscripts, but here they are in Beach’s hand in black ink as are the notations in the margins that indicate page breaks in the galley proofs. Therefore, following the JJA’s draft code I was able to designate this document Buffalo MS. VI.E.2 as I.8§1.11"+.[15]


Precise and comprehensive as the JJA’ s draft code is, some of the newly uncovered “Work in Progress”/Finnegans Wake manuscripts make it necessary to contort the genetic arrangement as laid out in The Archive. As though I needed any further reminders that there are always new and unexpected discoveries to be made, in November 2001, while researching questions concerning the editing and publication of the Letters in the Ellmann papers at the McFarlin Library at the University of Tulsa, Lori Curtis, Head of Special Collection, showed me several sets of Finnegans Wake page proofs.[16] These were the copies of the sets Paul Léon retained after working alongside Joyce on their revision from July 1938 through January 1939.[17]


The existence of these manuscripts at Tulsa should not have been as surprising as it seemed at the time given Joyce’s long-standing habit of presenting his manuscripts and books to his friends and associates and I should have expected that Léon would have preserved these considerate and valuable gifts. In fact, the Buffalo Joyce Collection is founded on materials saved by Léon and Beach. They each played parallel roles in Joyce’s life and similarly aided in the production of Joyce’s works during his almost twenty years in Paris. If nothing else, my work on Beach’s Joyce archive should have led me to expect that Léon would have had a similar personal collection.


The JJA reproduced all of the proofs that were then-known to be extant, but these manuscripts were acquired by Tulsa after its publication. Furthermore, because the Joyce/Léon Papers were unavailable for research when The Archive was being prepared, it was not known just how complicated the final revision process of the proofs had been. Reading the correspondence in the Joyce/Léon Papers looking for information about the proofs, I should have been able to surmise that several sets had been produced and revised.[18] In fact, I could have “discovered” these new manuscripts without leaving home, simply by consulting the University of Tulsa’s on-line finding aid.[19] Nonetheless, it was an accident that I came across this material when I did.


Generally, in comparison to the galley proofs not as many additions and emendations to the text were made on the page proofs, there are still some spectacular textual transformations. Just to take the first paragraph of Finnegans Wake as an example, on the first set of page proofs, it reads:


riverrun past Eve and Adam’s brings us by a commodius vicus of recirculation back to Howth Castle and Environs, Sir Tristram, […] ringsome on the aquaface.  

Joyce had Léon emend the text (in green ink, as usual on these proofs) to read:  

riverrun^,^ past Eve and Adam’s by ^from swerve of shore to bend of bay,^ brings us by a commodius vicus of recirculation back to Howth Castle and Environs^.^ Sir Tristram, […] ringsome on the aquaface.  

On the second setting in proof, the first paragraph is still fourteen lines long, but it is only now that Joyce instructed the printer to divide the paragraph and it became as it appears in Finnegans Wake:  

riverrun, past Eve and Adam’s from swerve of shore to bend of bay, brings us by a commodius vicus of recirculation back to Howth Castle and Environs. ^¶^
Sir Tristram, […] ringsome on the aquaface.


The Archive’s draft code will be disrupted in all those cases when it allocated a place for only one set of page proofs when, as was usually the case, several sets (and multiple copies of these sets) were printed and revised. Fortunately, because they fall so late in the genetic history of Finnegans Wake , their impact will be limited to the single level that follows them: Joyce’s errata to the published work. (But this is not the case with other manuscripts that have already been uncovered, nor will it likely be the case with those that still remain to be unearthed.)


II.2 can once again serve to illustrate the implications that these newly documented manuscripts will have on the Archive’s draft code.[20] An incomplete and unmarked set of page proofs for II.2 was reproduced in The Archive (53.401-13) and its draft code is 1.14/2.12/3.14/5.4-6.6/7.5/8.15/9.13.[21] This manuscript consists of signatures “R-T,” dated 23 September 1938, by the printer. Léon had sent this particular copy of this set of proofs to Weaver and it is in the British Library (47477, fs. 319v-325v), along with all of Weaver’s Finnegans Wake Collection.


A collation of the text as set on this proof and that published in Finnegans Wake indicated that there must have been another, missing, duplicate set; this one revised by Joyce and Léon and used by the printer to emend the text. This other copy of the set of proofs was postulated by the Archive’s editors and appropriately coded as such with a plus sign at each section. This manuscript, the printer’s copy of the page proofs, is now at Tulsa.


Because of the unexpected complications the printers faced setting the marginalia in this chapter, the preceding galleys were oddly set, with the marginalia “boxed in the text” (JJA 53.305, BL 47478, fs. 329r-355r; see JJA 53.307-34). Joyce instructed the printers to correct the layout and they were reset with the marginalia as it appears in Finnegans Wake (BL 47478, fs.357r-390r). But this is not all Joyce did, as with all the other galleys, he made numerous intrusive additions to the text here. Whereas the printers managed as best they could to integrate his emendations in the other chapters, the placement of the marginalia and the numbering of the footnotes here made the changes more troublesome, necessitating their re-placement and shuffling every time Joyce made a significant change or addition. Still another problem surfaced with the page proofs: here the right- and left-side marginalia were reversed on the even numbered pages. Therefore, at the start of the chapter (FW  260) there is a note in Léon’s hand.[22] Most of the instructions for revisions are corrections of mis-settings of the text, the alteration of some punctuation and the addition of many commas, but in general these are all minor emendations.


Besides this copy of this set, there is also an identical (although complete) second copy of first set of (unmarked) page proofs Tulsa, and so, following the established draft code it should be: II.2§1.14'/2.12'/3.14'/5.4'-6.6'/7.5'/8.15'/9.13'. On the other hand, there was also another setting of the text that was not accounted for by The Archive’s draft code structure. This time, signature “R” is dated 9 December 1938 and the two other signatures are not date stamped by the printer. Four copies of each of the signatures were printed, but only one of each was stamped “REVISE” and corrected in Léon’s hand (following Joyce’s instructions), this time in purple and black inks and pencil. These copies were all sent to Joyce and Léon. The printers themselves also checked and corrected some typographical issues on the revised copy of this set. The emendations indicated on the first set were incorporated here (including, most obviously, the reversion of the marginalia on the even numbered pages), and several more instructions for revisions were given here, but the primary issues that still needed to be resolved were the correct alignment of the marginalia vis-à-vis the central text (this was problematic throughout, but especially in the long list of great personages at the end of the chapter) as well as the leading of the footnotes.[23] These formal layout issues must have caused the printers more headache than the innumerable corrections, revisions and emendations Joyce had already made.


According to the logic of the draft code, this set should be II.2§1.15/2.13/3.15/5.5-6.7/7.6/8.16/9.14 (and the three other copies would be appropriately coded with prime signs). But this code has already been assigned to the errata Joyce and Léon compiled in the summer of 1939 using Louis Gillet’s unbound press copy of Finnegans Wake (Buffalo MS. [VI.H.4.a.] VI.J.1.),[24] which Maria Jolas typed (Buffalo MS. [VI.H.4.b. and c.] VI.J.2.a. and b.) and took to America for eventual publication and subsequent incorporation in future editions of Finnegans Wake .


So, what do to? Do we alter The Archive’s draft codes when necessary or do we find some other means of accounting for new manuscripts and inconsistencies? And whatever is done, how do we make the changes evident to users of The Archive  and our readers? I leave these as open questions.  



[1] Phillip F. Herring, Review of The James Joyce Archive , JJQ 19.1 (Fall 1981), 85-97. 


[2] Finnegans Wake II.2§8 (282.05-304.04) first appeared in transition 11 in February 1928 and then again under the newer title “The Muddest Thick That Was Ever Heard Dump” in Tales Told of Shem and Shaun (Paris: Black Sun Press, June 1929). See JJA  52 and 53.


[3] For a more extensive account of the process of getting the James Joyce Archive published, see Michael Groden’s “Perplex in the Pen—and the Pixels: Reflections on the James Joyce Archive, Hans Walter Gabler’s Ulysses , and ‘James Joyce in Hypermedia,” Journal of Modern Literature 22.2 (Winter 1998/99), 225-31.


[4] Herring, “James Joyce Archive ,” 86.


[5] See G. Thomas Tanselle, “Reproductions and Scholarship,” Studies in Bibliography , 42 (1989), 25-54; reprinted in Literature and Artifacts (Charlottesville: The Bibliographical Society of the University of Virginia, 1998), 59-88.


[6] The publication of the Gabler edition sparked a renewed but brief interest in Buffalo’s Ulysses holdings.


[7] For a preliminary description of the new edition of the catalog, see my “ReCollecting Joyce at Buffalo: Revising and Completing the Catalog,” Genitricksling Joyce,  eds. Sam Slote and Wim Van Mierlo, (Amsterdam: Rodopi, 1999), 13-26.


[8] Besides providing a valuable “Key to The James Joyce Archive  Volume Numbers” and an “Index to Library Collections,” Groden compiled an “Errata for The Archive ” in James Joyce Manuscripts: An Index (New York & London: Garland Publishing, 1980), 161-73.


[9] Crispi, “ReCollecting Joyce,” 19.


[10] See the informative catalog descriptions of the manuscripts: Chris Coover, James Joyce’s “Ulysses”: The John Quinn Manuscript of the “Circe” Episode (New York: Christie’s, 2000) and Peter Selley, “The Lost ‘Eumaeus’ Notebook/ James Joyce, Autograph Manuscript of the ‘Eumaeus’ Episode of ‘Ulysses ’,” (London: Sotheby’s [supplementary catalog] Lot 197, 10 July 2001).


[11] See Gabler’s “Afterword” to the synoptic edition of Ulysses (New York and London: Garland, 1984), 1864-67.


[12] See Bill Cadbury’s essay in this volume of Genetic Joyce Studies for a more comprehensive and detailed description of JJA  draft coding of I.2-4.


[13] See Hayman’s introduction to volume 52 for an account of the textual evolution of II.2.


[14] See David Hayman, “‘Scribbledehobbles’ and How They Grew: A Turning Point in the Development of a Chapter,” Twelve and a Tilly: Essays on the Occasion of the 25th Anniversary of “Finnegans Wake,”  eds. Jack P. Dalton and Clive Hart (London: Faber and Faber, 1966), 107–18.


[15] See The Archive ’s draft analysis for I.8, JJA  48.xvi.


[16] Sidney F. Huttner, former Head of Special Collections, McFarlin Library, provided a preliminary bibliographic description of the Finnegans Wake proofs in the catalog The Paul and Lucie Léon/James Joyce Collection , [University of Tulsa: McFarlin Library, 1985]. 


[17] Finnegans Wake was printed by R. MacLehose and Company (Glasgow, Scotland) for Faber & Faber publishers.


[18] The James Joyce–Paul Léon Papers in the National Library of Ireland,  compiled by Catherine Fahy (Dublin: National Library of Ireland, 1992). 


[19] See <>.


[20] I am currently preparing a collation of these various texts as well as bibliographic descriptions of the manuscripts that will appear in a forthcoming installment of the GJS.


[21] Even at this late stage, the printers still referred to the work as “J. W. P.” (that is, “J[oyce’s] W[ork in] P[rogress]”) and, in fact, Joyce revealed the name of the book to the world at such a late date that the signatures on the first edition still reads “J. W. P.”


[22] Léon’s note reads: “The comments on the margins are all wrongly printed and should be inverted i. e. All comments in [Block] capitals should go on the right margin and all the italics on the left. Therefore all your lefthandside pages are wrongly margined; the righthand pages are correct.” This note is in the same green ink in which most of corrections were made on the proofs and in which Joyce signed the deluxe copies of the first edition.


[23] A double-spaced typed, unsigned, note accompanied the revised proofs. It reads: “Work in progress / Please note that in all cases the first line of side-headings should aline [sic] with the corresponding line of the text. Please take care to see that this is done. / We must emphasize that when leading footnotes all leading should be even, and the same lead should be placed between the lines of a footnote of more than one line as between the footnotes themselves. This applies to all books of ours. Work in Progress is very unevenly leaded as regards the footnotes, and requires correction. There is too much space between the footnotes.”


[24] See Fahy, James Joyce/Paul Léon , 189: Paul Léon to Richard de la Mare, 22 July 1939: “Every time Mr Joyce looks at Finnegans Wake he finds misprints, he would like to start making corrections for the second edition on the sewn sheets sent to Monsieur Gillet.”