This page was created by Andrew Nevue.  The last update was by Kathleen Zoller.

Cherchez le texte: Proceedings of the ELO 2013 Conference

Translating afternoon, a story by Michael Joyce, or How to Inhabit a Spectral Body.

Arnaud Regnauld1

From now on, [the living present] bears death within itself and reinscribes in its own immediacy what ought as it were to survive it. It divides itself, in its life, between its life and its afterlife, without which there would be no image, no recording. There would no archive without this dehiscence, without this divisibility of the living present, which bears its specter within itself.

Jacques Derrida

As a work of electronic literature, Michael Joyce’s afternoon, a story exists on various levels, the ones experienced by the average reader (as opposed to those critiqued by the metareader, or the translator for that matter who both have to peer into the code as well) being parcellary by definition. The underlying layers of code underpinning the text-to-be-seen remain obfuscated below the visible surface of the screenic interface, offering only a partial view of the text which comprises both lines of code and strings of English interspersed with a few graphic elements. To further complexify the reader’s relation to the surface text, the fragmentary nature of hypertextual lexias prevents the recomposition of an ever elusive and paradoxically impossible totality since some narrative lines are simply incompatible with one another unless one should decide to embrace a form of quantum textuality. The inaccessibility of the work as a whole etches out a ghostly body of text, a hazy halo that haunts the margins of each lexia notwithstanding the underlying layers of code whose presence is to be felt, if not actually seen or read, in the performance of the text. What is to be translated then? Or, to put it differently, is the code to be deemed part of what is usually designated as “the text” as if it could be abstracted from its medium? It ensues that the relation between the underlying code, or layers of code, to natural languages must be interrogated. This even though “the code is not the text,” (Cayley 2002) to quote John Cayley, and yet clings to the body of text from which it cannot be dissociated, nor repressed along the translating process even though code does not translate per se.

As the reader navigates through the maze of afternoon’s hypertextual architecture, a sequence of competing narrative variations seems to emerge from a chaos of innumerable possibilities without ever allowing her to draw the contours of a clearly delineated figure. As she explores the work further, each lexia haunts her memory with its ghostly presence while pointing to yet unseen combinatorial configurations, always lurking in the periphery of her field of vision, always there at least in potential. Or in Michael Joyce’s own visual imagery :

Contours are the shape of what we think we see as we see it but that we know we have seen only after they are gone and new contours of our own shape themselves over the virtual armature, the liminal form, the retinal photogene (after-image) of what they have left us (1995, 222).

Whenever she chooses to click on a word that may or may not yield, following Michael Joyce’s metaphor, which incidentally poses a challenge to the French translator striving to combine the erotic charge of the verb with a more utilitarian sense, she is always anticipating the meaning of invisible connections which actually make more sense in retrospect even though the act of clicking is obviously based on a combined sense of extrapolation and expectation, or in Susana Pajares Tosca’s analysis grounded on the linguistic theory of relevance: “Every link communicates a presumption of its own optimal relevance.” (2000, 80) Should the translator then focus on the shifting rhetorical value of each and every connection? Privileging such a level of granularity would probably enhance the disjunctiveness of hyperlinks, thus threatening both form and meaning. One of the issues encountered here consists in striking the right balance between the innumerable interruptions of the narrative flow initiated by a network of 948 hyperlinks working against structure while preserving a general architectural coherence deceivingly loose as it may seem due to its intrinsic complexity.

Not only do hyperlinks virtually rupture narrative syntax, but they also rip sentences apart given that a single word may yield and disrupt the syntagmatic flow projecting prose into the tabular realm of poetry. The hyperlinking process on which electronic hypertext relies defies totalization as it keeps fraying a textual fabric that is bursting at the seams while begging for an endless recomposition that points to the seriality inherent in the concept of translation (Littau 1996). Each reading is akin to a versioning of a text that remains ungraspable as a whole, be it only for the cognitive overhead any attempt at holding all the threads in one hand would most certainly cause confusion for the translator/reader. No (human) translator can ever commit to memory the totality of the work’s multiple forking paths and recurring loops, even though she still needs to build a mental overview of the general architecture, somewhat intuitively. Now, the maps available for the metareader fail, however, to account for the complex topology of the work. The issue is to preserve, or at least to try to preserve, the frayed isotopic threads that maintain the coherence of the text as a whole. As Jean-Pierre Balpe puts it:

Hypertext enacts the same upheaval at the level of narrative syntax as the poem at the level of the sentence syntax. Similarly to the way poetry unbinds words from the fettering linearity of the syntagmatic axis to project them into a network of thematic, phonetic, metaphoric etc. correspondences which delineate a pluri-isotopic configuration, hypertext frees narrative sequences from the yoke of traditional narrative grammar to usher them into the multidimensional space of an entirely new and open structure (n.d.).

In other words, confronted with a fragmentary and elusive complex of textual fragments that sometimes may wear down to a single word, let alone a mere letter, the translator will have to attune herself to the numerous echoes and resonance phenomena specific to what may be called afternoon’s polyphonic melody (a metaphor incidentally central to another of Michael Joyce’s digital works, Twelve Blue (1996) ) to reveal the work’s hidden systematicity. In contradistinction with the formalisation of binary code, the task of the translator therefore remains somewhat intuitive and yet constrained by the interface logic. For instance, the functionalities of the Storyspace Reader enable the metareader to perform a search on keywords which ties the emergence of isotopies to a machinic form of parsing. The Storyspace Reader interface brings fiction into the realm of scholarly editions, the search function playing a role analogous to that of an index highlighting recurrences and variations while limiting the reader’s interpretive possibilities, subsequently fettering the translating process.

The paradigmatic shapes the syntagmatic into interlaced and yet shifting configurations that cannot be dissociated from the invisible hypertextual architecture, and more generally from its software environment as the code remains embedded within the verbal weave. The translator is therefore caught within a double bind: either focus on the letter of the text, importing the foreign hypertextual and verbal syntax into the translating language to the detriment of meaning, or conversely remaining hypothetically “true” to the spirit of the so-called original while repressing the hypertextual drift at work in afternoon. Since hyperperlinks appear as part of a metagrammar ruling over the behavior of the work, a mere transposition of the general hypertextual architecture combined with the translation of the verbal contents of each lexia may end up not doing justice to the intricacies of an already hybridized, and as such foreignized language embedded with code.

Should one follow Antoine Berman’s critique of Benjamin’s essay, “The Task of the Translator,” it appears that form, i.e the letter of the text, should prevail over meaning, therefore distorting, or foreignizing the translating language to let the so-called original shine through the translucent surface of the translated text according to Berman retranslation (2008, 168) of Maurice de Gandillac’s previous attempt. Interestingly enough, Steven Rendall translates durchscheinend into English the same way as Gandillac’s rendering of the word in French: “True translation is transparent, it does not obscure the original, does not stand in its light, but rather allows pure language, as if strengthened by its own medium, to shine even more fully on the original.” (1997, 162) Harry Zohn has also opted for the same term in English: “A real translation is transparent; it does not cover the original, does not black its light, but allows the pure language, as though reinforced by its own medium to shine upon the original all the more fully.” (1968, 21) Now the adjective durchscheinend literally means “shining through” in German, and it may be interpreted as transparent or translucent, a semantic choice which either invisibilizes or reveals the act of translation while keeping the original at a proximate distance. This is no small matter considering the politics of the translator’s repressed authority.

Now, before pursuing this specific path, it might prove useful to examine a quotation in which Michael Joyce describes the experience of reading in an electronic hypertextual environment in quite similar terms. Joyce highlights the paradigmatic thrust of digital hyperfictions while he incidentally offers a theoretical commentary upon the cybertranslator’s task: « the text becomes a present tense palimpsest where what shines through are not past versions but potential alternate views.” (1995, 3) In other words, there is no original as such, but mere potentialities waiting to be actualized by the reader, a process analogous to the relation a translation bears to the translated text. The palimpsestic nature of any text, and more particularly digital texts embedded with code, calls for an act of translation that may reveal, or at least make partially visible the various layers it comprises. However, as aptly remarked by John Cayley in his critique of digital utopianism: “there is no mutual transparency or translatability of code and language.” (2002) The underlying code is not primarily meant for the reader but actually addressed to the sole machine unless it is codework, which could lead to the conclusion it completely eludes the translator’s task. This is, however, more complex than it actually seems.

Two instances of codework based on LISP may be found in afternoon (see screen shot reproduced below), metafictionally pointing to the digital nature of the work while obliquely raising the issue of remediation even though the code proper has ceased to be operative. What’s more, these strings of code call up one of the central questions arising from various narrative paths: did the main character, Peter, actually kill his wife and son in a car accident that particular morning? The issue of scientific truth as a fictional construct (Regnauld, 2013) is one of the themes running through the work. Now, breaking down the actual function which is more complex than the following pared down version, it should be noticed, for instance, that the sequence “(LAMBDA (TRUTH))” returns any input value that is submitted as “truth,” ascribing a double meaning to the variable depending on the narrative context. Now, a value may certainly be “true” or “false” within a computer program, but the labeling of a list of values under such a conceptual category as “TRUTH” proves too radical a move not to bear some ironical flavor which should not escape the translator’s attention. What is lost in translation, though, is the possible interplay between LISP as a coding language and the English language from which it has been partly derived, a form of estrangement that could be partially compared with the difference to be found in proximate languages that have evolved from the same linguistic core with one crucial distinction: programming languages are not meaningful but purely performative. Here the ambiguity lies in the status of codework for the translator who may decipher instructions as actual words such as “car,” meaningless though the syntax may appear, which stands as quite a confusing element in the chain of signifiers considering the general narrative context. More specifically the seminal event of the car accident in which Peter may or may not have seen his wife and son die that morning.

WCodework has a doubly foreignizing effect on the translating language as the proximate relationship of English with higher level programming languages is irremediably lost, unless one should invent a new version of LISP, for instance, in which one would derive the coding instructions from the translating language. The reference to the original language known by programmers would not be entirely erased as the non-linguistic signs and general syntax and non-linguistic characters could remain unchanged while the connotative value of code as code would still be perceived by less computer-literate readers.

The foreignizing effect of translation on the translating language is akin to Benjamin’s understanding of the concept of aura in A Short History of Photography as “a peculiar web of space and time: the unique manifestation of a distance, however near it may be.” (1931) Code always casts a long shadow on the observer as it cannot be observed in the here and now of its performance. It lines the visible and readable surface of the work while remaining at a distance, withdrawn from the reader’s gaze. What’s more is that the performance of the work itself, as a text-to-be-seen that is, always leaves an ungraspable spectral body unread and yet radiating, or shining through. This could provide us with a redefinition of aura in the digital age as an aspect of the work revealed and possibly destroyed in the very act of translation: the absolute singularity of aura as defined by Benjamin would paradoxically lie in the iteration pertaining to the archival nature of such digital works placed under the translator’s authority as the one who ensures the texts’ afterlife. Indeed iteration implies alteration, and not a mere reproduction of the same on an industrial scale as deplored by Benjamin. Paraphrasing Derrida’s comment on the task of the archivist, I contend that:

“The [cybertranslator] produces more [digital] archive, and that is why the archive is never closed. It opens out of the future. How can we think about this fatal repetition, about repetition in general in its relationship to memory and the archive? It is easy to perceive, if not to interpret, the necessity of such a relationship, at least if one associates the archive, as naturally one is always tempted to do, with repetition, and repetition with the past. But it is the future which is at issue here, and the archive as an irreducible experience of the future.” (1995, 38)

In the context of digital literature what is “a peculiar web of space and time” then? This metaphor possibly reflects the way the relationship between natural language and code affects the auratic quality of digital works, as both natural language and code fade out at radically different paces. The accelerated obsolescence of programming languages results in the necessary porting and regular remediation of such works as afternoon which already counts numerous versions 2 as if the editorial process of this most unstable work could never actually stop. Digital works beg for remediation in the same way as print literature begs for translation but on a radically different scale, further complexifying the task of the translator while reinforcing the auratic quality of the work as a digital archive. Following Derrida’s analysis of the spectral structure of the archive, it appears that its untranslatable idiomaticity preserves it from the destructive power of repetition:

“… the archive always holds a problem for translation. With the irreplaceable singularity of a document to interpret, to repeat, to reproduce, but each time in its original uniqueness, an archive ought to be idiomatic, and thus at once offered and unavailable for translation, open to and shielded from technical iteration and reproduction.” (1995, 50)

Now, as the use of LISP within the text-to-be-seen serves as a metafictional gesture that takes us back to the 1960s when this language was invented, the question of updating the reference for the contemporary reader becomes as problematical as that of whether archaisms should be left as such or translated into a more modern version. John Cayley is right when he concludes his essay claiming that “it must be possible to recompile the codes as operative procedures, as aspects of live-art textual practice. The code is not the text.” (2002) It ensues, however, that keeping the code alive, or rather, preserving its “afterlife” 3 is also part of the cybertranslator’s (most likely collective) task (Regnauld and Vanderhaeghe, 2014). The question remains as to whether this should take place through a fully destructive gesture of remediation or not. This is not a mere issue of computer logic as there is a definite aesthetics proper to the Storyspace interface that may be affected by any reprogramming of the work. The remediated excerpt of afternoon (Joyce, 1987) as made available on the Web by Norton is, for instance, parasitized by supplementary parergonal elements. Remediating afternoon on the Web has not only transformed the underlying code into HTML and other web-specific programming languages, but also the visual reception of the lexias. Their online configuration does not direct the reader’s gaze in the same way as the original rectangular windows; in addition, there is a different set of colors and graphics absent from the Storyspace version of the work whose semiotic value needs to be taken into account. Any reprogrammation, or porting of a work onto a different platform involves the destruction of the initial programming. Now, the question remains whether it should allow for a more or less analogous emulation of what may be called the initial version of the work. Or is there such a thing as an initial version? Bear in mind one generally remediates the latest version of the code while always going back to the so-called original when translating from a natural language.

What about a tactile version on a current tablet considering the haptic dimension of afternoon which does not allow the user to read from the map? Doesn’t the introduction of a help button (as is the case for Steve Tomasula’s iPad version of TOC: A New Media Novel) making instructions readily accessible in the iPad version of the work run counter to the blindly groping heuristics specific to navigation in afternoon? What’s more, there are at least two levels at which the work can be apprehended, that of the reader and that of the metareader who can peek behind the scenes, which raises another aesthetic issue: should the code enabling authors to create conditional links (see screen shot below), for instance, in the Storyspace environment be emulated for historical reasons, or remediated using an entirely different set of interfacial buttons in a dedicated menu for instance? This question may not apply to other works using different authoring systems, but it so happens that the development of Storyspace cannot be dissociated from the writing of afternoon, Michael Joyce being one of the conceptors of the computer program.

Considering the general aesthetics of afternoon from an editorial point of view, it must also be noted that contrary to the codex, we do not have a conventional or stable model of reference as to what an electronic scholarly edition should be despite George Landow’s Dickens Web or Jerome McGann’s Rossetti Archive (the William Blake Archive). Storyspace as an interface remains bound to the print model even though it is also akin to Hypercard and the Macintosh GUI aesthetics (Matt Kirschenbaum extends the list of influences to “literary experimentation and postmodern theory — notably that of Eco —, but also interactive computer fiction, artificial intelligence and story generators, word processing, desktop publishing and the then-new GUI conventions of the Mac, hypertext system research, and interactive videodisc technology” (2008, 177) ). The set of instructions clearly invites the reader to “turn pages” by clicking the Return key while the writing spaces, that is the textual boxes that pop up on screen, are highly reminiscent of the page format. From a technical point of view, the Cd-Rom on which afternoon has been marketed so far tends to disappear as a material substrate for electronic works as it is gradually relocated in the clouds of remote servers from which they can be downloaded. As a matter of fact, the current version of afternoon cannot be read under the latest operating system for Macintosh or under the 32-bit version of Windows 7, a technical parameter which certainly shifts the task of the translator who can no longer work on her own, unless she develops programming skills as well.

The code is what produces the text, insists John Cayley in his critique of the very notion of “translation” as a process that carries across some shared thing among languages: we should concentrate instead “on the ‘translation’ of concepts and corresponding methods of textual production rather than on the texts themselves.” (2014) A digital artwork is not a text proper, but a dynamic, generative system, and this machinic assemblage must be translated into the target culture as well. The half-life of digital works entails their constant dislocation and relocation, which is all too often conducive to “the erasure of the memory of defamiliarization devices that accompanies the displacement of these translations,” (Joyce 2014) a homogenizing process apparent in the different versions of afternoon, a story. Now, if we are to follow Walter Benjamin’s notion that a translation always begs for a retranslation as part of the afterlife of any literary work, it would seem that it has been submitted to a temporal shift, a technological acceleration involving a whole chain of production. The work exists, or remains alive through the multiple versions it comprises.

Any electronic work is characterized by its intrinsic lability and quick obsolescence. Print is evidently subject to a certain number of variations too, be they typographical or editorial, for instance, but we are confronted with an entirely different scale when dealing with the transience of digital works. This involves transformations which may produce effects that lie beyond the aesthetic scope of the text-of-inscription. The fact is the author can never control all the parameters presiding over the execution of a program, starting with the speed of the microprocessor, let alone the wider technological context in which the work is read, which may change the semiotics of the work to the point of corrupting its code to death. Paul de Man’s reading of Walter Benjamin’s essay “The Task of the Translator,” highlights the fragmentation lying at the core of the translating process: the work exists only through the multiple versions it comprises. As claimed by Walter Benjamin in “The Task of the Translator,” a work always demands a translation which is both an alteration and a guarantee of its perpetuation, be it as a ghost:

“… it can be demonstrated that no translation would be possible if in its ultimate essence it strove for likeness to the original. For in its afterlife -- which could not be called that if it were not a transformation and a renewal of something living -- the original undergoes a change.” (2000, 249)

This conduces me to reinterpret the singularity of aura in the light of the “spectral messianicity” (Derrida 1995, 27-28) pertaining to any archive whose revelation, in the photographic sense, is always already deferred, always lying at a proximate distance and yet ungraspable as such, “always struggling painfully out of darkness” (Benjamin 1931, np) (to quote Benjamin's praiseful comment on the light in Hill’s photography). The endless translating/remediating process of both code and text paradoxically ensures the work’s afterlife precisely because it proves to be an impossible challenge. It can never exhaust the potentialities of the archive that remains always open onto the times to come. Besides, technical reproductibility does not automatically entail the destruction of aura whose messianic nature seemed to be already part of Benjamin’s reflection as illustrated in the following comment (the picture of Dauthendey) on photography as an archiving technique opening onto the future:

However skillful the photographer, however carefully he poses his model, the spectator feels an irresistible compulsion to look for the tiny spark of chance, of the here and now, with which reality has, as it were, seared the character in the picture; to find that imperceptible point at which, in the immediacy of that long-past moment, the future so persuasively inserts itself that, looking back, we may rediscover it. (1931, np)


The translating process may be construed as a form of archiving as it involves a selection which is also a destruction of “the original,” a form of erasure especially acute in the case of code and paradoxically meant to ensure the survival of the work as the translated fragments migrate into a new spectral body of text spliced with updated strings of code enabling its performance, or becoming-text. The act of translation is both a performative repetition and a critical alteration which renews former interpretations of the archive it keeps open onto the future of its re-translation. This process can be thought beyond the metaphysics of an ultimate version of the text in the light of what Derrida terms the messianic and defines as follows:

It is a question of this performative to come whose archive no longer has any relation to the record of what is, to the record of the presence of what is or will have been actually present. (1995, 27-28)

Reading/translating afternoon, a story is akin to being caught within an infinite feedback loop which exacerbates the iterability of any textual form in its very performance. Each attempt at translation can be interpreted as a terminating condition which interrupts the potentially infinite loop on which afternoon’s performance is based and thereby offers transient islands of stability in a sea of proliferating and monstrously hybridized and dreamlike possibilities, each time begging anew for a redrawing of the limits of the wor(l)d. Or, as Michael Joyce put it back in 1982 4:

Words seem more like the hive of electronic hashmarks to me than these carbon strikes before me. It seems right to construe language as a series of switchmarks, loads and unloads, zeros and ones. The form it eventually takes is almost incidental; do we always write down dreams? (Kirschenbaum 2008, note 18, p.169)


  1. 1. Subsequent to the ELO 2013 “Chercher le Texte: Locating the Text in Electronic Literature” conference, portions of the initial presentation have been published in a slightly different form in the introduction to Translating E-Literature, Yves Abrioux & Arnaud Regnauld (eds.), 2014. URL : back

  2. 2. for a detailed analysis, see Matthew Kirschenbaum’s analysis in Mechanisms (2008).back

  3. 3. for an in-depth analysis of this particular aspect, see 10 PRINT CHR$(205.5+RND(1)); : GOTO 10 by Nick Montfort et al (2013) back

  4. 4. Personal correspondence from Joyce to Becker, Jan. 7, 1982. Michael Joyce Papers, Harry Ransom Humanities Research Center, U. of Texas at Austin, quoted in Kirschenbaum. back

This page references: