Translation, transmutation, transmediation, and transmission in TRANS.MISSION [A.DIALOGUE]


Presented at:
“Translating E-Literature”
Paris 8, Paris, France
11-14 June 2012

This paper examines the operation of translation in the creation and dissemination of computer-generated digital literature through discussion of the web-based narrative dialogue TRANS.MISSION [A.DIALOGUE] (Carpenter, 2011). In this discussion, the term “translation” is situated within a string of variables pertaining to the word trans-: translation, transmutation, transmediation, and transmission. Translated into JavaScript, this string of variables could be written as follows: “var trans=['lation', 'mutation', 'mediation', 'mission'].”

The word trans- is a prefix meaning across, beyond, or through. This prefix may be used in combination with an element of origin: transcontinental, transatlantic. This prefix may be used to imply a state of change: transmit, transfer, transport, translate. And, somewhat more abstractly, this prefix may be used to imply a poetics of coming and going. The word “transverse” applies the prefix trans- to the Latin versus, meaning a turning. Every verse has a re-verse, which is to say, verse has direction. In Greek verse, Strophe sets out from east to west across the stage. Antistrophe replies from west to east. Neither voice is in either place. Both are calling: across, beyond, through.

The word “translation” applies the prefix trans- to the word –lation, which comes from the Latin, latio, meaning borne, as in carried or endured. Traditionally, translation from one form to another implies an equivalency between forms. In the translation of a text from one natural language to another one might expect the meaning, the mood, and perhaps the rhythm of the text to endure. In the translation of a born-digital text from one code language to another, what precisely is borne across, beyond, or through?

The word “transmutation” implies a sudden and/or radical change in form. In a homophonic translation, for example, little or no attempt is made to preserve the meaning of the original text. In the classic example of Luis d’Antin van Rooten’s Mots D'Heures: Gousses, Rames, the phonetic sequence “Humpty Dumpty” endures in the translation “Un petit d'un petit” (1967). Here, the hermeticism of the homophonic translation is counterbalanced by the choice of a “sub-text … so well known as to be recognized by all” (Lecercle, 1985: 21). In the recent spate of remixes of Nick Montfort’s computer-generated poem Taroko Gorge (Montfort, 2009), the contents of the remixed texts as they are displayed on screen may appear to diverge radically from Taroko Gorge, yet these remixes are based on the now familiar sub-text of Montfort’s source code, which endures almost entirely intact in most of the Taroko Gorge remixes, or translations, as we might more appropriately call them. Conversely, the translation of a computer-generated text from one programming language to another may radically alter the source code yet result in little or no change to the content or behaviour of the text displayed on screen, as in the case of Montfort’s own initial translation of Taroko Gorge from Python into JavaScript.

The word “transmediation” refers to movement across, beyond, and through media. Though we may consider languages – natural, code, or otherwise – to be behaviours rather than media, when dealing with code languages we must consider the media used to create and disseminate these languages as integral to their intelligibility. Python files cannot be read in a web browser, for example, and Flash files cannot be read on an iSO device. The translation of a born digital text from one code language to another is most often prompted by hardware and/or software obsolescence. In the example of Judy Malloy’s ground-breaking hypertext Uncle Roger (Malloy, 1986), Malloy has adapted and altered the work a number of times to suit emerging media environments ranging from early newsgroups to BASIC, UNIX, and the World Wide Web. I term this process “transmediation” rather than “remediation” as, particularly in the case of TRANS.MISSION [A.DIALOGUE], this discussion is more concerned with the asynchronous movement of text across, beyond, and through a continuum of forms than with the associative relationship between old and new media forms upon which the discourse surrounding remediation tends to focus.

The word “transmission” refers both to the action of sending across, and to that which is sent. The word “mission”, from the Latin missiō, a sending off, may refer to a group or committee of persons sent to a foreign country to provide assistance, conduct negotiations, establish relations, initiate communications, build fortifications or in any other way translate a strange place into somewhere known. The word “mission” may also refer to an operational task, designed to carry out the goals of a specific program. A computer program, for instance. Thus, a JavaScript may be on a mission, and that mission may be a transmission, a sending across.

In order to discuss how this string of trans- variables may be operate as compositional and structural elements in the creation and dissemination of narratives native to networked environments, we will turn now to the specific example of TRANS.MISSION [A.DIALOGUE]. This web-based work is a computer-generated dialogue, a literary narrative of generations of transatlantic migration performed in the form of a conversation, an encoded discourse propagating across, beyond, and through long-distance communications networks. TRANS.MISSION [A.DIALOGUE] is literally a transmission. One JavaScript file sits in one directory on one server attached to a vast network of hubs, routers, switches, and submarine cables through which this one file may be accessed many times from many places by many devices. And TRANS.MISSION [A.DIALOGUE] is literally a dialogue. As with the Latin trans-, the word “dialogue” also refers to a crossing. Stemming from the Greek dialogos dia-, meaning: across, and logos, meaning: a word, saying, speech, discourse, thought, or reckoning – akin to légein, meaning to choose, gather, recount, tell over, or speak. The mission of this JavaScript is to generate another sort of script, a script for live performance. The call “function produce_stories()” produces a response in the browser. As JavaScript is a procedural language, in so far as it must be written and read in a certain order, we may say that the browser is performing the JavaScript. The result of this performance, i.e. the text which appears on screen, is a narrative dialogue intended to be read aloud in three voices. These voices may be called, alternately: Call, Response, and Interference; or: Strophe, Antistrophe, and Chorus; or Here, There, and Somewhere in Between.

TRANS.MISSION [A.DIALOGUE] was written in Python and then translated into JavaScript in autumn 2011. With the exception of a single page of handwritten notes in a passport-sized notebook, the entire compositional process from sentence construction, to variable string population and layout establishment was conducted inside the Python and JavaScript programming languages. The Python iteration of TRANS.MISSION [A.DIALOGUE] is a transmutation of a 1k narrative generator called written in Python by Nick Montfort in 2008. Both the act and the resulting textual fact of the translation of TRANS.MISSION [A.DIALOGUE] from Python to JavaScript are, in a broad sense, translations of Montfort’s own translation of into a JavaScript version known as The Two (Montfort, 2008). It must be noted here, that Montfort aided and abetted in this circuitous translation process through email correspondence. Further, although the translation of natural languages is not my focus here, it must also be noted that both and The Two have been translated in to French, Spanish and Russian. Although the source codes were not radically altered in these translations, the selection of the variables and the structure of the sentences had to be significantly altered in response to these languages’ handling of gender, which, as we shall see, performs a critical operation in these works.

Structurally, and The Two strip the traditional literary form of the short story down to its most fundamental elements: beginning, middle, and end. As Montfort explains in a post to the collective blog GrandTextAuto, “A sentence is chosen from a pool of beginnings. A middle is generated by joining ‘He’ or ‘She’ to a verb or other middle section and concluding that with ‘he’ or ‘she’. Then, an ending is chosen from a pool of endings” (Montfort, 2008).

The police officer nears the alleged perpetrator.
She berates her.
Six years later, neither one remembers the incident.

Given the power dynamics set out in the first sentence, we may be surprised to learn in the second sentence that both the police officer and the alleged perpetrator are female. Why wouldn’t we be? Movies staring females in the roles of both protagonist and antagonist are rare indeed. How differently would we interpret the story if instead it read:

The police officer nears the alleged perpetrator.
She berates him.
Six years later, neither one remembers the incident.


The police officer nears the alleged perpetrator.
He berates her.
Six years later, neither one remembers the incident.

In the above mentioned GrandTextAuto post, Montfort cites Nanette Wylde’s “minimal and clever programs” as inspiration for Certainly we can see the influence of her Electronic Flipbook, about so many things. These “flipbooks” were created in Director for specific installation contexts between 1998 and 2006. They were not available online until 2012. In order that I might view them, they were transmitted to me by Wylde via the post on a CD. As Wylde describes in the booklet which accompanies this CD, and on her website, “about so many things randomly displays the activities of ‘He’ and ‘She’ without bias to gender. That is, the activities are drawn from the same pool of possibilities.” Anything he can do, she can do. Similarly, both and The Two capitalize on the variability of gender assumptions by making gender a variable: “var heshe=['He',' She']”. Although the source codes of and The Two are not literally translations of about so many things, the gender variable is born across. And although the nature and form of Montfort’s narrative were substantially transformed in the creation of the Python version of TRANS.MISSION [A.DIALOGUE], and then further altered in the translation from Python into JavaScript, the operation of this gender variable endures as a central narrative imperative. The string “var heshe=['he','she']” is copied directly from Montfort’s source code in that of TRANS.MISSION [A.DIALOGUE], and a similar string, “var hisher=['his','her']” is added. We can see the operation of this gender variable in the following sentence:

The translator conveys her encouragements.
The administrator relays his congratulations.
The pilot broadcasts her explanations.
The receptionist transmits his salutations.

In the source code of the Python iteration, the syntax for the code which calls variables is as follows: “choice(heshe).” “choice()” is part of Python’s “random” module, which also contains “shuffle()” and “random(),” which aren't all grammatically consistent. “choice()” is not built into the JavaScript programming language, but rather, is created as a function. In translating to JavaScript, Montfort created a function called “choose().” Although “choose()” performs in exactly the same way as the built-in option of “choice()” does in Python, Montfort’s choice of the word choose in writing his own seems to imply a rather more imperative emphasis on variable selection.

In either case, it must be stressed that we are not dealing with particularly difficult code here. The source code for is 26 lines long; the file is less than 1k. The encoded assumptions about gender alluded to by the stories generated by The Two are far more complex than the JavaScript source code which generates them. The deceptive simplicity of Montfort’s generators would seem to undermine Roberto Simanowski’s argument that, “the internal problem of this genre of digital literature is its poetics of technology, which replaces a language juggler with a crafter of code” (2011: 91). TRANS.MISSION [A.DIALOGUE] further thwarts this argument, in so far as the source code was not entirely crafted by me. Technically, less a craft than a crude life raft, my code is a transmutation, a wilful mutilation, a hack. The decision to hack rather than craft code anew is a deliberate one. In A Hacker Manifesto, McKenzie Wark argues, “[t]o hack is always to produce the odd difference in the production of information… by transforming in some way the very process of production” (Wark, 2004: 222). Something of the uncanny twinning of characters at work in Wylde’s about so many things and Montfort’s The Two underpinned my process production; my hack transforms Montfort’s source code into a code medium of sorts. TRANS.MISSION [A.DIALOGUE] sends and receives dialogue on and through source code and associated media haunted by generations of past usage.

This haunting may be understood, in part, to be the result of an operation of memory. In a programming language like C, “var=” refers to a specific location in memory. A location is always both a place and the act of locating that place. Thus, a location always exists before it is located. Yet, rather than assigning variables to a specific location in memory, JavaScript distributes the operation of processes including memory across networks and devices. Until it is referred to, the location of a variable may be anywhere. Once it has been referred to, through a process known as garbage collection, a variable may disappear. Or, the reference to it may disappear. This mode of dispersed, temporary, and transitory memory allocation suits the performance of a narrative text of place and displacement nicely. Though a function such as “choose()” might be called upon to select from a string of trans variables: “choose(trans)”, for example, only one result will be returned: transmission. The rest remain in memory, as potential selections for possible future translations. Thus, this text performs the act of selection from memory. Every 80000 milliseconds, a new instance of the text – one of an infinite number of possible translations – is displayed on screen. The text displayed is doubly a translation, performed in the first instance by the JavaScript and in the second by the browser, which translates the source code into the text we see on the screen.

Cybertext theory distinguishes between these two instances of the same text with the terms textons and scriptons. Textons are strings of signs as they are in the text, i.e. the source code. Scriptons are strings of signs as they appear to readers/users. The mechanism by which scriptons are generated from textons is termed a traversal function. In Espen Aarseth’s typology of textual communication, transiency is listed as a variable of traversal functions. “If the mere passing of the user’s time causes scriptons to appear, the text is transient…” (Eskelinen, 2012). TRANS.MISSION [A.DIALOGUE] is in every sense a transient text. The mere passing of time causes scriptons to appear. These scriptons spell out stories of transmission, of transience, of transit:

Why shouldn't the wanderers dream of clearer manuals?
The passage from Cornwall proved cruel.
Ancient migrants described itineraries. Three were from the Hebrides.

Further underling the traversal function of transiency, the reader can never quite reach the end of this transmission. Mid-way through a reading, a new version is generated. The sentence structures stay the same, but all the variables change. In a very long sentence in The Order of Things, Michel Foucault describes the classical sentence as a signification engine; a mechanical construction which performs the task of linking otherwise disassociated elements together. He writes, “in a single continuous sentence it is possible to indicate relations of time, of consequence, of possession, and of localization” (Foucault, 1994[1966]: 100). In TRANS.MISSION [A.DIALOGUE] these relations shift as time passes, so that we might have immigrants now, where once we had explorers; a persistent tap eclipses a strange whir; a message instead of a passage; Nova Scotia in place of Scotland; a submarine cable replaces a shipping network. If we were to think of translation merely in terms of equivalencies, we would not likely consider a submarine cable a suitable substitute for a shipping network. We might avoid replacing the word passage, with its double implications of a passage across the Atlantic and a passage of a larger text, with the word message and its more singular meaning. But by situating translation within a string of trans- variables we arrive at a somewhat different understanding of how these “otherwise disassociated elements” are indeed linked together.

TRANS.MISSION [A.DIALOGUE] is a mechanical construction, a sentence engine performing the programmatic function of associating suspended variables with syntactic signification that they might travel through networks and emerge intact as narrative units. The dialogue generated by this engine is both technically and topically inflected with the syntax and grammar of code language. Some variable strings contain nothing but codes. “var receiving=” for example, reproduces shorthand gleaned from logs kept at the Glace Bay Marconi Station, circa 1911 (now kept in the Marconi Archive at the Bodleian Library, Oxford, UK):

var receiving=['40 words local paper', '30 words local paper', '100 words special news', 'a few scraps of a private message', 'distinguishable dots', 'dots only', 'heavy traffic', 'something again', 'atmospherics', 'last message from ship', 'repeated \"are you there\"', 'repeated \"where are you\"', 'request to repeat', 'several distinct dashes', 'something from another station', 'a weak signal', 'no answers to our enquiries', 'no answer', 'weak readable signals', 'no signals', 'no signals received, probably not sending', 'strong readable signals, sending fast', 'medium strength readable signals', 'some static', 'lightening all around'];

Jay David Bolter and Richard Grusin term the representation of one medium in another “remediation” and argue that “remediation is a defining characteristic of the new digital media” (1999: 45). Yet it is of little significance that the above cited variables were once printed text and are now digital textons. Trans- seems a more specific prefix than re- in re-lation to the pre- digital multi-media ecology referred to by this work. TRANS.MISSION [A.DIALOGUE] performs the transmediation of texts from archival sources, but these “texts” have already have passed across, beyond or through the code mediums of wires, switches, signals, air, ears, hands, paper.

TRANS.MISSION [A.DIALOGUE] externalises a poetics of technology. Codes, their creators, the modes through which they operate, propagate, and communicate, and the confusion they instigate are one of the main topics of the dialogue TRANS.MISSION [A.DIALOGUE] generates. Simanowski suggests that, “because absurdity, weirdness, and illogicality are the default modes of text generators, mastery is only proven by overcoming such characteristics” (2011: 91). This generator aims not to overcome but rather to embrace such characteristics. Absurdity, weirdness, and illogicality are the default modes of long-distance communication, migration, displacement, and difference. And so, TRANS.MISSION [A.DIALOGUE] generates cacophony, liminality, atemporality and asynchronous exchanges of mixed messages pertaining to miscommunications and network failures.

In his critique of “the vagueness of remediation as a concept” Markku Eskelinen argues, “the heuristic question may no longer be what a medium is, but what a medium does and is used for” (2012: 20). Whether read by a network, by a machine, by software, or by a human eye; whether read as textons or as scriptions in either a fixed or generative instantiation, or spoken by the mouth, or experienced by the ear; what TRANS.MISSION [A.DIALOGUE] does is generate a dialogue about the translation from one place to another, and what it is used for is a script for live poly-vocal performance. TRANS.MISSION [A.DIALOGUE] has been performed in Amsterdam, NL; Bristol, UK; Banff, CA; and Oxford, OH, USA. Each instance constitutes a new translation, or transmediation, into a new and unique configuration of performers, audience, acoustics, and spatial arrangements. In “Dramaturgy and the Digital,” an article written by Barbara Bridger after having participated in one of these live performance, Bridger comes to a conclusion uncannily close to Eskelinen’s, though couched in very different terms:

One of the central characteristics of this work is its interrogation of its own modes of operation: an approach that is less concerned with deciphering the meaning of a piece of work, and more interested in the structures that allow this meaning to be transmitted.” (2013)

The most basic, most fundamental of these structures is the dialogue. The figures of Strophe and Antistrophe represent the most basic communication network - call and response. TRANS.MISSION [A.DIALOGUE] begins with a call: Begin! Followed by a response: How? With a question. What emerges from a question? Distant shores, to lure us. Location, location, location. Jacques Derrida observes, “Site, this land, calling to us from beyond memory, is always elsewhere. The site is not the empirical and national Here of a territory. It is immemorial, and thus also a future” (1978[1967]: 66). The act of locating a distant shore provides a context for the fact of our present position, which is always already in the past, already behind us. In her long poem, “The Fall of Rome: A Traveller’s Guide,” Canadian poet and classicist Anne Carson writes: “A journey …/ begins with a voice / calling you name out / behind you. / This seems a convenient arrangement. / How else would you know it’s time to go?”(1995: 75).

And so Strophe sets out from east to west on a treacherous mission, across high seas and frozen wastes, in search of a Northwest Passage, in hopes of trade routes, and fountains of eternal youth. And Antistrophe returns from west to east with scurvy, captive natives, and furs. Neither ever arrives. Both only just barely finish leaving. Through generations of transatlantic migration, characteristics of one place become trans-posed upon another. Another trans- word, transposition re-places. In the case of the call “choose(place)”, “var place=” refers both literally and figuratively a location in memory.

The furthest sea shores are reminiscent of those of England.
The neighbouring vistas compare to those of Cornwall.
The nearest lands could easily be confused with those of Nova Scotia.

Although the translation of natural languages is not my focus here, the inextricable association between language and nation necessitates the question: Were this work translated into another language, such as French, for example, would the location of memory also be translated, or re-placed, to reflect generations transatlantic migration from France to Nouvelle France? Would Cornwall be replaced with Bretagne, Nova Scotia with Acadie? In the interest of soliciting a response to this question, in April 2012 a single output of TRANS.MISSION [A.DIALOGUE] was posted to Vertaallab (TranslationLab), an ongoing translation experiment edited by Rozalie Hirs on the Dutch blog Ooteoote, in which, translators are invited to post translations as comments to the featured works (Hirs, 2012). There were two responses to TRANS.MISSION [A.DIALOGUE]. The first, posted by Ludy Roumen-Bührs, translated the text from English to Dutch. The English place names were retained. The line: “Eleven were from England.” became “Elf kwamen uit Engeland.” The second response was posted by @netwurker, born Mary-Anne Breeze, aka Netwurker Mez, a pioneering author of digital literature known for developing and writing in the hybrid code-poetry language “mezangelle.” To Vertaallab, Mez posted a portion of TRANS.MISSION [A.DIALOGUE] translated into mezanglle:

_TRANS.MISSION [A(hhh).DIA(multimodal)LOGUE]_

be[en there, done that, a]g[a]in[:out(re)] Transmission.
with a[hhhh] quest[].
wot_merges 4rm?
have ARGs + Augments been[+/or]gone, yet?
y Kant [u.c]?….
low nrg_lvling.
Eleg[ant]raphic.[s|w]Itches, here. .[knot.....*here*].

Here, the syntax and grammar of the code languages engaged in enacting this born digital literary text have heavily inflected, or, we might say, infected its translation. The resulting text is a transmutation in the order of “un petit d’un petit.” TRANS.MISSION [A.DIALOGUE] serves as a subtext from which to digress into a systematic punning which echoes and extends my own use, in the title of the work, of square brackets, periods, slashes, plus signs, and other punctuation marks common to programming languages. These devices divide phonetic sequences into complex parenthetical segments, in which, new words appear. The first line – “Begin Transmission” – becomes: “be[en there, done that, a]g[a]in[:out(re)] Transmission.” The second line fuses and confuses all of the possible (w) variables ['why', 'where', 'how'] into one impossible word: “[w]H[y]ow[l]?” Mez’s transmutation reflects generations of migration – not across the Atlantic, but rather, into an online networked game space, in which, in mezanglle, at least, the binarisms of he or she and here or there might collapse, into [s.he] and []. The potential of the hybrid s[t]he[re] space is proposed in the line: “[Br]Av[e][s+divided.Queens].”

In 2013, TRANS.MISSION [A.DIALOGUE] was translated into French by Ariane Savoie, a PhD student at Université Québec à Montréal, and was published as TRANS.MISSION [UN.DIALOGUE] in a special translation issue of bleuOrange, a Montreal-based online journal of “littérature hypermédiatique,” which launched at the Electronic Literature Organization conference Chercher le texte in Paris 24 September 2013. In personal correspondence, Savoie shared certain thoughts on her process, which I will now synthesise here. A strict translation of all the English variables into French equivalents would have resulted in subject-verb gender disagreements, the resolution of which would require considerable modification to the source code which, Savoie felt, would have diminished the variability of the generator and the structure of the piece. Instead, Savoie elected to respect the structure of the source code. Gender conflicts were avoided by the population of strings with variables from only one gender, letting go of any variables that didn't have the exact equivalent in that gender in French. Initially, this resulted in an eradication of the gender variable altogether. Eventually, a compromise was reached in which two versions of certain variable strings were created, that both masculine and feminine proper nouns might be called at different points in the script.

Although the string “var heshe=['he','she']” is not carried over into Savoie’s translation, something of the either/or binarism of Wylde and Montfort’s “var heshe=” endures, both through the introduction of gender variables through other means, as cited above, and through the variable string “var place=”, in which, the location of each place named is either on one side of the Atlantic or the other: Canada or England, Acadie or France, the new world or the old, home or away. Through the operation of this variable, here and there become doppelganger of one another. In The Uncanny, Sigmund Freud defines doppelganger as persons who have to be regarded as identical because they look alike (2003[1919]: 141). If we re-place the word “persons” with the word “places” in Freud’s definition we arrive at a similarly uncanny conclusion. Between places inextricably linked by generations of immigration “[t]here is the constant recurrence of the same thing, the repetition of the same… features, the same characters, the same destinies, the same misdeeds, even the same names, through successive generations” (id.: 142).

It could be argued that we are not speaking of translation at all here. Perhaps what we are seeing is simply a case of influence, of resemblance – an uncanny recurrence of code processes carried across from one generation of computer-generated text to the next. Perhaps. Text generation is the oldest form of literary experimentation with computers, after all. Etymologically, the word “generation” so heavily implies regeneration that it would be difficult if not impossible for a second, or third, or fourth generation of generator generators to not be influenced by previous generations of generators of generators. But I have framed the process of creating and disseminating TRANS.MISSION [A.DIALOGUE] in terms of translation all the same. Translation, transmutation, transmediation, and transmission have played a central role in the creation and dissemination of this text. Traces, phrases, structures and functions form the source codes of and The Two endure in its textons. The results of the operation of variables such as gender are borne across into its scriptons. The question of what is borne across, beyond, and though each new generation of this text is reposed every 80000 milliseconds.

There is no repose for the questions this text poses. I will close with one of an infinite number of possible endings proposed by TRANS.MISSION [A.DIALOGUE]:

Is the delivery mechanism functioning?
Some of us believe it’s working.
Please try again.


Les systèmes sont-ils présents?
Les autorités imaginent qu'ils sont brisés.
Veuillez s'il vous plait réessayer.

Pour citer
Carpenter, J. R. 2014. « Translation, transmutation, transmediation, and transmission in TRANS.MISSION [A.DIALOGUE] ». Dans Traduire l’hypermédia / l’hypermédia et le traduire. Cahiers virtuels du Laboratoire NT2, n° 7. En ligne sur le site du Laboratoire NT2. <>. Consulté le 21 février 2017.