GENETIC JOYCE STUDIES - Issue 2 (Spring 2002)



A Second Look at the First-Draft Version of “Finnegans Wake”


Sam Slote


There is always something that comes before. Such a statement is obviously the presupposition behind genetic criticism: before the published text there existed preparatory manuscripts, the study of which can help elucidate (or further complicate) the final work. This statement is also apposite to scholarly research; we are always continuing the work of others before us. While the James Joyce Archive presented reproductions of virtually all the then-known extant Joyce manuscripts, it was not exactly the first work to attempt to represent Joyce’s preparatory drafts. In 1960, Fred Higginson published transcriptions of six intermediary stages of chapter I.8 of Finnegans Wake; in 1961, Thomas Connolly published a (much-flawed but still, nevertheless, pioneering) transcription of the exceptionally large and unusual Finnegans Wake notebook VI.A (otherwise known as “Scribbledehobble”); two years later Robert Scholes and Richard M. Kain made available many of the preliminary notes and manuscripts for Joyce’s Portrait; and in the 1970s Phillip Herring published transcriptions of all the Ulysses notes at the British Library as well as (separately) a selection of the Ulysses notes and drafts at the Poetry/Rare Books Collection of the University at Buffalo.[1]

Unlike the JJA, which aimed at a documentary reproduction of Joyce’s working documents with minimal (albeit very useful) commentary, all these works were attempts to transcribe and edit Joyce’s manuscripts (by different standards) in order to make them more accessible. Apparently, accessibility did not foster acceptance. In his review for the JJQ Phillip Herring called the JJA’s publication “an event of galactic importance”, but also noted that it has largely been ignored by Joyceans who were, perhaps, intimidated by the sheer bulk of material made available and were unaware of much of the work that had already been done on Joyce’s manuscripts.[2] Herring concludes his review by proposing that “it may take another generation before [the Archive] is sufficiently used and appreciated”.[3] His prediction has largely come true and it now finally seems that there is some real interest in genetic matters. Indeed, the “Finnegans Wake” Notebooks at Buffalo series, now in publication,[4] represents a major advance from the notebook volumes of the JJA (which Herring had signalled in his review as being perhaps the most significant component of the series). And so now that we are in the silver jubilee of the JJA, it seems propitious to turn to one of the works that made the Archive possible.

Published in 1963, David Hayman’s First-Draft Version of “Finnegans Wake” was perhaps the most ambitious of all these pre-Archive works. All the works listed above were largely self-evident affairs; for example, Herring’s task of transcribing and editing the Ulysses notesheets at the British Library was a self-contained and self-delimited undertaking (the notesheets exist as discrete, identifiable documents and thus can be represented as such). As Hayman points out, this is not the case with a “first-draft version” of Finnegans Wake. Joyce did not write a complete draft of either the book-as-a-whole or of individual chapters and then make subsequent revisions. Rather, the book was composed piecemeal, and thus the “first-draft version” that Hayman represents is synthetic: “the aggregate of the first-draft versions of all of a given chapter’s parts”.[5] (To some extent, Higginson’s representation of the stages of the evolution of I.8 is also synthetic since his six levels are aggregates or conflations of multiple documents; where Higginson has six levels, Hayman posits 20 in his draft catalogue in FDV). Hayman’s notion of “first-draft” is defined pragmatically: “the earliest available version of every passage in the published work” (FDV 41). For a large part of the text of Finnegans Wake, the earliest available drafts do appear to be first drafts, although, obviously, with a text like the Wake, which depends upon notes and other materials, a rigorous and exacting notion of “first-draft” tends to be somewhat slippery.

To counter the possibly misleading nature of this synthetic representation of the Wake’s early drafts, Hayman provides a lengthy and detailed analysis of the composition of Finnegans Wake in his introduction as well as a comprehensive draft catalogue that corrects many of the (understandable) errors made by T.J. Brown, George Painter, and Harriet Weaver in their catalogue of the British Library’s Wake manuscripts (Hayman’s own catalogue was subsequently emended by Hayman and Danis Rose in the JJA).[6]

The main reason Hayman gives for his critically daring approach is that the reproduction of all Wake manuscripts is “impractical, if not impossible” (FDV 3); in other words, at that time, a project as commodious and comprehensive as the Archive was just a wistful dream (or nightmare). And so, instead of a full archive of Wake drafts, Hayman proffered “a convenient and workable compromise” (FDV 3–4): the First-Draft Version, a proto-Archive.[7]

Hayman acknowledges that, since it combines texts written between 1923 and 1938, FDV is hardly a “uniform first draft of Finnegans Wake or even a fairly uniform text” (FDV 8). One example of this heterogeneity among the “first drafts” is that the earliest drafts (from 1923–24) are written in a basically English style with few linguistic distortions and almost no foreign elements whereas later first drafts (especially from the 1930s) were already initially drafted in the convolute Wakean form (see FDV 8–12 for specific examples). In this way, FDV obviates the temporal dimension to the Wake’s evolution. A further instance of this genetic heterogeneity is the organisation of FDV around the final text’s four book structure—a structure that had not been in place when Joyce began writing the book. The final form of the book thus has a determining effect upon the presentation of its first drafts. Likewise, the logic of the JJA’s organisation of Wake drafts is also determined by the structure of the final text.[8] Hayman does include various elements that did not exactly make it to the final text, such as “The Revered Letter” (the original kernel for I.5, which was much later incorporated into book IV), “The Delivery of the Letter” (the seed for book III), and “Tristan and Isolde” (which was atomised into the “Mamalujo” vignette to form II.4[9]). In this way, the shape of FDV is not completely dictated by the final text of Finnegans Wake.

An example of how FDV’s text is composite comes with how Hayman treats BL 47471b, the infamous red-backed copybook in which most of book I (chapters I.2–5 and 7–8) was initially drafted and redrafted. The complex and non-sequential intercalation of first- and second-draft passages makes representing this crucial document especially difficult. Hayman is not entirely consistent with how he handles this copybook. One could in fact imagine a “first-draft version of Finnegans Wake” (or “proto-Finnegans Wake”) that would be a transcription of this one copybook (and perhaps augmented with transcriptions of the early vignettes).[10] However, such an enframing would obviate the quadripartite, Vichian structure of the final text. Therefore, FDV obviates many of the peculiar aspects of this copybook by arranging its texts into the structure of the final text (the JJA treats this copybook in an analogous manner). Obviously, a hypertext representation of this copybook would be ideal in that once could alternate between various modes of enframing the document and its text (at the “Genetic Joyce Studies” conference in Antwerp in 2001, Dirk Van Hulle and Edward Vanhoutte demonstrated an ingenious SGML representation of this copybook[11]).

As an illustration of the problems of treating this copybook, and of Hayman’s resolutions, I will turn to his treatment of I.3. Immediately after the first draft of I.2 (which lacked Hosty’s ballad), Joyce drafted the first two units of I.3: the plebiscite (FW 57.16–61.27) on folio 3 and an account of the fate of the Rann makers (FW 48.01–50.32) on folio 2v.[12] The redrafted version of this chapter is scattered across this copybook, mostly written on unused versos. The version Hayman presents comprises the original drafts of the first two subunits augmented by the new material that appeared in the redrafted version. This is an unusual composite since he could have presented the first drafts of the first two subunits along with a complete draft of the second draft (which, while duplicating the first two subunits, would be an integral, rather than composite, draft). Indeed, when it comes to the first part of I.7, Hayman includes a transcription of both the first draft and the redraft that immediately follows it in the red copybook.

Ultimately however, the decision to present a hybrid I.3 is perfectly consistent with the logic underlying FDV: “as my intention is to present as complete a first draft as possible, I am including here the earliest drafts of the other parts of the chapter regardless of chronology” (FDV 69 n.2). The problem with this frame of presentation, that privileges the most primitive versions, is that it limits the usefulness of FDV. One can turn to FDV for an example of the earliest version of any given passage, but it becomes less reliable when one wants to relate different passages since, while the individual passages may be presented in their earliest state, these configurations could well derive from distinct documents or levels. (But then, such carping is faulting FDV for not being the JJA.) In short, then, FDV is a transcription of a text that never did exist until its constitution within the pages of the First-Draft Version; the text of FDV is an imitation without an original.

Jean-Louis Lebrave wrote that transcriptions are inherently “both richer and poorer than the original manuscript from which they are drawn”.[13] The organisation of FDV, synthesising and assembling a “first-draft” out of various documents, is an obvious example of how transcription can be “richer” than the manuscript page. But, of course, many of the idiosyncrasies of the manuscript page are invariably lost in transcription. Hayman provides copious footnotes that indicate many of these features, so even though he is transcribing a text (that is, the text of the virtual first-draft version), he maintains an awareness of the documents from which the text derives. Because of the complexity of Joyce’s drafts, Hayman acknowledges “that perfect accuracy and faultless judgement are out of the question” (FDV 45). Indeed, it is not surprising to discover that there are errors in Hayman’s transcription. Hayman cautions that serious scholarship should depend upon an examination of the original document rather than rely upon the transcription (the same caveat also applies to the reproductions in the JJA).[14] Indeed, Almuth Grésillon claims that the “goal of transcription is not perfection but perfectibility”.[15] A transcription can never supersede the document. The disproportionate gusto of Jack Dalton’s scathing critique of the accuracy of Hayman’s transcriptions[16] says more about Dalton’s misguided faith in the possibility of a perfect transcription than about the quality of FDV.

The most immediate problem with arranging a transcription is deciding upon a rationale of presentation. While Hayman’s scheme is relatively simple, it produces a text that appears quite complex.[17] Obviously, Hayman’s transcription is easier to follow than the original, but with all its typographic machinations, it is hardly intuitive.[18] Typographic clarity is also a kind of obfuscation. The very concept of transcription is almost antithetical to the nature of the documents purportedly represented. A draft document is inherently a work in progress, that is, it documents a work that remains unfinished, or incomplete, or inachieved. A draft manuscript is imperfect (which is not necessarily to imply perfection to the final text). Now, if the draft document is imperfect and dynamic (in that it builds or points towards a final configuration of the text that had not yet existed), the transcription of that document is necessarily static and transitive. Simply put, a transcription is a freeze-frame of a past, imperfective textual state; it is a typographic spatialisation of a temporal process that, in so doing, perfects an imperfective textual state. The goal of the transcription scheme should be that it is transparent enough to allow the reader to infer the underlying textual dynamics (this also implies that the reader must be motivated enough to do a little work). As an example, I take Hayman’s transcription of the earliest extant version of the Wake’s first paragraph (which was not the first thing written):[19]

brings us to
Howth Castle & Environs! Sir Tristam, viola d’amores, had not encorepassencore arrived rearrived on a merry isthmus from North Amorica to wielderfight his peninsular war… (FDV 46; BL 47482a 83; JJA 44: 3).

One can extrapolate from this transcription the earliest version of this text on this document, that is, the text as it existed prior to being revised: “Howth Castle & Environs! Sir Tristam had not encore arrived from North Amorica…”[20]. The most primitive configuration of this text on this draft presents a static state, exemplified in the fixed (albeit vaguely delimited) location Howth Castle and Environs.[21] The revisions on this page turn this away from a static place to a dynamic motion: brings us to Howth Castle and Environs. Furthermore, Sir Tristam had not just not yet arrived, but he had not yet rearrived. The revisions add a sense of motion and cyclic continuity rather than the stasis of the first level just as the act of revision itself renders the textualisation on this document dynamic. Yet, the transcription (necessarily) conflates both textual states, the static and the dynamic, into one textual state differentiated only through typography. These problems are, of course, not limited to FDV but are, rather, endemic to the task of transcription.

FDV also records a number of passages that did not make it into the final text of Finnegans Wake (many of which are recorded in the “Lost and Found” section of Genetic Joyce Studies). One such passage, from I.5, admirably (if inadvertently or, even, unconsciously) provides a commentary on the genetic condition. In the conclusion to the initial discussion of the Letter, we read (in the red copybook):

Wonderfully well this explains the double nature of this gryphonic script and while its ingredients stand out with stereoptican relief we can seepeep[22]tour beyond the figure of the scriptor into the subconscious writer’s mind (FDV 87; BL 47471b 25v; JJA 46: 238).

This paragraph was not recopied into the redraft of this passage (also in the red copybook [BL 47471b 49; JJA 46: 246]). Ostensibly, this omitted paragraph discusses the duality of graphic and phonetic language, “the sound sense sympol” (FW 612.29). The key-term for such a reading would be the collocation “gryphonic script”, which suggests a script that amalgamates the grapheme with the phoneme, much like the griffin (or gryphon) is a monstrous miscegenation of an eagle and a lion. From this duality, a double perspective is elicited “with stereoptican relief”. However, such a double nature is not a perfect duality since the combination of the graphic and the phonic is itself scriptive in the “gryphonic script”.

The “double nature” described in this paragraph is also appropriate to genetic criticism. The geneticist has a double perspective over the text: viewing the drafts proleptically from the basis of the final text and the final text analeptically from the avant-textes. >From the vantage point of the geneticists’ hindsight, the text and its avant-textes are mutually engendering.[23] And so, from this perspective, the double nature would be not a combination of the graphic and the phonic but, rather, a combination of what is written and what is erased. Each new draft stage writes over, or deletes, the previous stage even as it inherits and propagates the text from that level (Joyce’s habit of crossing out old drafts after they had been recopied illustrates this principle). This genetic dual perspective is, of course, what renders the manuscripts’ textualisation dynamic. Because of this, we do not so much see what the text is but, rather, we tour around the texts that have been written and the texts that have been erased.

A further complication with this passage, and its treatment in FDV, comes with Hayman’s own second look at this (omitted) first draft paragraph. In The “Wake” in Transit, Hayman cites this passage, along with a citation to the FDV transcription, but with a silently modified (or edited) reading: “Wonderfully well this explains the double nature of this gryphonic script and while its ingredients stand out with stereoptican relief we can see beyond the figure of the scriptor into the subconscious editor’s mind”.[24] The manuscript is somewhat murky and messy here, but the second, revised, and edited reading is probably the correct one. The subconscious writer is now an editor. This little lapsus between writer and editor, apperceived après coup, highlights the condition of transcription. The double nature of the task of the transcriber is suspended between writing and editing: the mediation of the transcriber means that what is on the page is invariably edited. In this way, the transcriber is not unlike Hayman’s notion of the arranger (conceived as an explanation of narrative enframing in Ulysses): “a significant, felt absence in the text, an unstated but inescapable source of control”.[25]

A transcription is a representation of a text but it is also an imitation of a manuscript.[26] These are not exactly isomorphic functions since the manuscript is, as Daniel Ferrer states, a “protocol for making a text”.[27] That is, the manuscript represents the text at a specific, inchoate level of its evolution. And while the manuscript does exist independently of its transcription (unless it had been subsequently destroyed), the text does not (and can not) exist apart from any material incarnation. Therefore, both manuscript and transcription represent a textual state that has no independent status. By imitating the manuscript, the transcription could be said to represent the text as indicated on a specific manuscript. An example of the transcription’s mediated refraction of the text is Hayman’s inferred erasure of the word “peep” in the above example. Joyce did not cross out this word on the manuscript and so, by crossing it out, Hayman is not adhering to the manuscript. Instead, Hayman has inferred a text that is implied by, but not actually present on, the manuscript page.

The text is thus like the “subconscious… mind” inferred out of “stereoptican relief” from the manuscript and its transcription. And, coming from the editor’s, rather than the writer’s, mind suggests that the text that is made manifest is not even original but, rather, secondary and mediated. The original is made inaccessible precisely because of the mediation of other intervening figures. Such an analogy between (mediated) text and subconscious is not new (and, hopefully, I hardly need to cite Lacan here). A different reading, and perhaps more apposite to this particular case, comes with Derrida’s reading of Freud’s model of the Wunderblock (mystic writing pad) as a model of memory in “Freud et la scène de l’écriture”. Freud uses the Wunderblock (a kind of precursor to the etch-a-sketch) as a model or metaphor of the psychic apparatus. In other words, the psychic apparatus can only be described indirectly through a model or representation. The Freudian model of the Wunderblock depends upon and performs a metaphor, a form of writing.

Il n’y a pas de texte écrit et présent ailleurs, qui donnerait lieu, sans en être modifié, à un travail et à une temporalisation… qui lui seraient extérieurs et flotteraient à sa surface. Il n’y a pas de texte présent en général et il n’y a pas même de texte présent-passé, de texte passé comme ayant été-présent. Le texte n’est pas pensable dans la forme, originaire ou modifiée, de la présence. La texte inconscient est déjà tissé de traces pures, de différences où s’unissent le sens et la force, texte nulle part présent, constitué d’archives qui sont toujours déjà des transcriptions.[28]

This passage is remarkably congruent with the brief, elided paragraph from I.5 concerning the Letter’s “double nature”. The text is never “there”, never present (not even in the sense of having been present); rather, the text, such as it is, is constituted as a differential pattern emerging from the various strata of textualisation. It is always already a virtual text (not unlike a “first-draft” version of Finnegans Wake), viewed from a doubled perspective. The implication of all this is that it is not just the transcriptions of FDV or the reproductions in JJA that are secondary, even the Wake manuscripts themselves (housed and preserved in the British Library) are secondary documents, they are already transcriptions of a text bereft of an absented original. All text is tracing.

This can be seen in an interesting example of textual evolution, one which inadvertently (or subconsciously) endorses Hayman’s rationale in assembling the FDV, to “illustrate the nature of Joyce’s inspiration and the direction of his thought” (FDV 4). The first draft of a passage describing the Letter reads: “The original document was what is known as unbreakable script tracery” (FDV 89; BL 47471b 42v; JJA 46: 302). In writing out the subsequent draft, Joyce reversed the emendation of “script” to “tracery” by restoring the word “script”: “The original document was what is known as unbreakable script” (BL 47471b 48v; JJA 46: 312; FW 123.31–33). But this purported “script” by which the original document may be known is itself a product of the palimpsest of textual deviation and differentiation between the drafts that constitute and are constituted by the text. The script is (still) tracery lying between two apparently contradictory possibilities (hanno o non hanno): “The haves and havenots: a distinction” (FW 295.L4–6). And, of course, by the time we get to the final text, the word “trace” has itself been erased, that is, the distinction can no longer be distinguished.


[1] Fred H. Higginson, Anna Livia Plurabelle: The Making of a Chapter, Minneapolis: University of Minnesota Press, 1960; Thomas E. Connolly, Scribbledehobble: The Ur-Workbook for “Finnegans Wake”, Evanston: Northwestern University Press; Robert Scholes and Richard M. Kain, The Workshop of Daedalus, Evanston: Northwestern University Press, 1965; Phillip F. Herring, Joyce’s Notesheets in the British Museum, Charlottesville: University Press of Virginia, 1972; Phillip F. Herring, Joyce’s Notes and Early Drafts for “Ulysses”: Selections from the Buffalo Collection, Charlottesville: University Press of Virginia, 1977. This list omits critical works that are descriptive of Joyce’s practices of composition.
[2] Phillip F. Herring, Review of The James Joyce Archive, JJQ 19.1 (Fall 1981): 85–97, 86, see 87–88.
[3] Herring 1981, 97.
[4]The “Finnegans Wake” Notebooks at Buffalo, eds. Vincent Deane, Daniel Ferrer, Geert Lernout, Turnhout: Brepols, 2001–.
[5] David Hayman, A First-Draft Version of “Finnegans Wake”, Austin: University of Texas Press, 1963, 41. (Hereafter FDV.) This text, as well as other pioneering works of Joyce criticism selected by Hayman, is now available online:
[6] I have attempted to refine Hayman’s admirable, and still-indispensable, account of the compositional history of Finnegans Wake, taking into account recent discoveries and controversies, in the introduction to “Finnegans Wake”: Genetic Essays, eds. Luca Crispi and Sam Slote, forthcoming.
[7] Jane Lidderdale and Mary Nicholson’s biography of Harriet Weaver records that shortly before Hayman had applied for permission to publish the early drafts, Weaver had turned down a similar request to publish a “fair-copy” version of Finnegans Wake on the grounds that such a volume might be taken by prospective readers as a substitute for the finished text. Hayman’s project, on the other hand, did not seem to her to present such a danger (Jane Lidderdale and Mary Nicholson, Dear Miss Weaver, New York: Viking, 1970, 431).
[8] Such rearranging is inevitable in genetic criticism. See Pierre-Marc de Biasi, “Vers une science de la littérature: L’analyse des manuscrits et la genèse de l’œuvre”, Symposium: Les Enjeux (Encyclopædia Universalis), vol. 2, Paris: Encyclopædia Universalis, 1993, 924–37, 930–32.
[9] See FDV 36 and Jed Deppman’s essay in “Finnegans Wake”: Genetic Essays for a description of how these two vignettes were merged.
[10] It is worth noting that Hayman titled his work with an indefinite article: A First-Draft Version of “Finnegans Wake”, as opposed to the definite article used in naming The James Joyce Archive.
[11] Dirk Van Hulle and Edward Vanhoutte, “The ‘Guiltless’ Notebook”, unpublished paper presented at “Genetic Joyce Studies”, Antwerp, March 30, 2001.
[12] There is some ambiguity as to which subunit was composed first; see Hayman’s Introduction in JJA 45: xi and Bill Cadbury’s essay in “Finnegans Wake”: Genetic Essays.
[13] Jean-Louis Levrave, “Déchiffrer, transcrire, éditer la genèse”, Proust à la lettre, eds. Almuth Grésillon, Jean-Louis Lebrave, Catherine Viollet, Tusson: Du Lérot, 1990, 143–62, 143. Translation mine.
[14] Bill Cadbury, among others, cautions that “authoritative assertions about the documents must be backed by inspection of the originals” (Bill Cadbury, “The Development of the ‘Eye, Ear, Nose and Throat Witness’ Testimony in I.4”, Probes: Genetic Studies in Joyce, eds. David Hayman and Sam Slote, Amsterdam: Rodopi, 1995, 203–54, 245). 
[15] Almuth Grésillon, Éléments de critique génétique, Paris: PUF, 1994, 129. Translation mine.
[16] Jack P. Dalton, “I Say It’s “Spinach”—“Watch!”, A Wake Newslitter, old series, 17 [November 1963]: 1–14.
[17] Hayman’s scheme is easier than the one Gabler devised for his synoptic text of Ulysses, but Gabler reckons with multiple draft levels simultaneously whereas FDV concerns just one.
[18] Shortly before I moved to Wisconsin to attend graduate school, I showed a copy of FDV to a friend and said that the person I was going to be studying with wrote this book. Upon seeing the transcriptions, my friend replied that Hayman had managed to make Finnegans Wake even more complicated.
[19] I follow Hayman’s transcription scheme: additions are in italics; deletions are struck-through; and substitutions are in bold (FDV 44–45).
[20] One problem with this extrapolation is that Hayman’s transcription scheme, like most other linearised (as opposed to diplomatic) transcriptions, does not distinguish between currente calamo revision and subsequent revision. However, a quick check at the JJA shows that the extrapolated base version is correct.
[21] Indeed, the most primitive draft of this passage could be considered to be the following passage from notebook VI.B.15: “rHowth Castle and Environs” (VI.B.15: 33). Other items in this passage also derive from this notebook: “isthmus” (VI.B.15: 36); “peninsular battle” (VI.B.15: 37); and “Amory Tristram” (VI.B.15: 57).
[22] Hayman notes that this word is not crossed out on the manuscript even though he crosses it out in his transcription to indicate an intended deletion (FDV 87 n. 26). Hayman’s hollistic transcription does not indicate that the word “peep” is written over the word “see” and that the word “tour”, written in the margin, is keyed in as a substitute.
[23] See Daniel Ferrer, “Clementis’s Cap: Retroaction and Persistence in the Genetic Process”, trans. Marlena G. Corcoran, Yale French Studies 89: Drafts, eds. Michel Contat, Denis Hollier, and Jacques Neefs, New Haven: Yale University Press, 1996, 223–36.
[24] David Hayman, The “Wake” in Transit, Ithaca: Cornell University Press, 1990, 175. See also Dirk Van Hulle, “The Wake’s Progress: Toward a Genetic Edition”, Text 13 (2000): 221–32, 224.
[25] David Hayman, “Ulysses”: The Mechanics of Meaning, revised edition, Madison: University of Wisconsin Press, 1982, 123.
[26] Hayman’s footnotes serve as constant reminders of the material aspects of the manuscripts he is transcribing.
[27] Daniel Ferrer, “The Open Space of the Draft Page: James Joyce and Modern Manuscripts”, The Iconic Page in Manuscript, Print, and Digital Culture, eds. George Bornstein and Theresa Tinkle, Ann Arbor: The University of Michigan Press, 1998, 249–67, 261.
[28] Jacques Derrida, “Freud et la scène d’écriture”, L’Écriture et la différence, Paris: Seuil, 1967, 293–340, 313–14.