I almost wrote “the idiot of digital literature” in the title of this essay. Idiot that I am, I ask: is not such a slip of the pen, the wrong word on the page, or the wrong fingers on the wrong keys, is not such an almost written and almost taking place precisely the domain of the literary? Or equally literary: the right words and the right keys, the pen in exactly the right place, and yet still a slippage in precisely this, an almost not taking place yet written down. How can you know? What is the literary literally? The otherness of the literary is inextricable from “it could be otherwise,” subject to Theodor Adorno's judgment in the essay “Commitment” on the work of art as entirely artificial and constructed artifact, as “instructions for the praxis they refrain from: the production of life lived as it ought to be”  [Adorno 1992, 93].

(I wonder: how can digital humanities deal with the almost written and the almost taking place or almost not taking place yet written down? I find nothing in digital humanities – even amidst the poetic turns and explorations of figurality – that engages with this potentiality of the literary.)

The idiot of digital literature: would this not be a way of entitling myself? Are not all titles, of and in whatever genre, to a greater or lesser degree “signed” by the author? As in TEI P5: Guidelines for Electronic Text Encoding and Interchange by the TEI Consortium Originally edited by C.M. Sperberg-McQueen and Lou Burnard for the ACH-ALLC-ACL Text Encoding Initiative Now entirely revised and expanded under the supervision of the Technical Council of the TEI Consortium? All this appears on the front matter of the TEI P5 Guidelines. I admit that I am unclear and cannot discern or read whether this is a single title, a title with authorship attributions, or if it is several titles, or partially not a title (or something else)? Or is such a question of the author's name really only the case within a certain kind of very literary author function? Is this only the case for literary works, where a title adheres to the writer's name? As in: Shakespeare's Hamlet or Mark Twain's Huck Finn? But it is this very particular and very literary effect that interests me when it comes to examining the literary in digital humanities: this stickiness of the name, this effect of reference in the formal feature of the title. The most unreadable scrap of text is signed by its author, and this is part of what qualifies it as literary.

In this sense, “the idiocy of digital literature” is named for me; I am announcing myself as the idiot. Such abjection is strategic, of course. Here are my strategies: I mean to use “idiot” in the ancient Greek sense of an individual, an ordinary person, a non-professional person, a private person. Towards this, my inquiry poses simplistic and particular questions about digital humanities projects, about popular accounts of digital humanities, about historical affinities between digital humanities and literary studies, and about core protocols of digital humanities.

But also, I mean to bring to digital humanities a sense of the idiot in the Dostoyevskian or Sartrean sense of saintliness through naive questioning. Or even the Iggy Poppian sense of the idiot, bringing a bit of snarling punkiness to digital humanities. At the least, this is a question of style, of an inquiry wandering in and out of the space of academic discourse.

Finally, and most particularly, I mean idiocy in the sense that Clément Rosset wrote of “the idiocy of the real”  [Rosset 1989, 111]. Idiocy as inescapable and yet obdurate; as singular, immediate, and useless. Idiocy as a passion for reality. I find this passion in characters on the screen, in the literalness of what the computers displays, and in the way this literal is played out in digital humanities practices. Is this passion not a description of the literary?

In what follows, I ask idiotic questions of digital humanities, seeking the occluded but productive literary core of this field. At first glance, this still-emerging field is systematized around a priori sets of categories for objects and discourses that presume literature without the literary. By this, I mean the literary to signify a domain of excess and difference that is institutionalized as creativity or innovation. At the least, digital humanities is uncomfortable with this sense of the literary. Digital humanities handles and presents this domain but, at least explicitly, does not participate in it. It assumes the productivity or poetics of the literary, and builds on this assumption. One might say: it follows the trajectory of the literary but is not literary itself. It insists that it is methodological. Such insistence lets digital humanities operate on and make a project out of the literary.

By literary, I also refer to the legal norms of an archivable work with an author, as well as the modern profession that this work implies. Digital humanities is expert at dealing with the literary in this sense. It employs diverse, complex methods and means of studying literature. To be clear: literature is not (and never was) literary, and for this reason literature can be a subject of digital humanities. Literature is a noun, literary an adjective. Literature is a work and an archive, literary an activity and process. Literature is not literary first means that the institution of literature is bereft of the literary. Every work of literature can be said to cite the literary but any account of it (the literary) falls short: it locates it in this or that feature, this or that theory, which is not the literary but a particular formalism, a form that is repeatable and extractable from the text (by definition).

Literature is bereft of the literary means literature refers to, tells the story of the literary. Literature is the institutional memory of the literary. Literature means the literary has left the building. Literature can be a subject of digital humanities for this reason, but the literary cannot. The many monuments of the digital humanities canon already include archives and tools built around literature but not around the literary. Can we imagine a meta-digital humanities on the literary? We cannot. Not in any current imagining of digital humanities, not without rethinking both the digital and the humanities.

This is demonstrably true both in the large scale configuration of the digital humanities — such as the mission and practice of the National Endowment for the Humanities Office of Digital Humanities — and in paradigmatic digital humanities literature projects, such as those developed and exhibited at sites such as Institute for Advanced Technology in the Humanities at University of Virginia (IATH) and the Maryland Institute for Technology in the Humanities at University of Maryland (MITH). My strategy is to invert the problematic to ask: what is literary about digital humanities? What if digital humanities is a “field” only in that it is literary?[1] Digital humanities may take on or remediate a poetic function in its premises and claims, but this function is the outcome of the literary and not literary itself. Models of a discrete domain of the literary, particularly Jakobson’s “defamiliarization” – though Fry’s “non-construction” would be equally applicable – repeat and extend rather than resolve the problem of situating the literary vis a vis digital humanities.[2] In the end, I turn to text and character encoding as fictional sites of the trajectory of the “character” and the mis-directed wandering of the literary mark. The character is one direction for reading the singularity of the literary in the digital humanities. I return to the digital humanities with an understanding of the literary in its projects, but — unfortunately — with no proposal for reconciling the system of digital humanities with the singularity of the literary.

A bit more on the passion of the digital literary. It is a matter of characters on the screen. My problem is: should I read or should I look? I look at the screen — which one? this one, every one? — and see characters turning to images, characters destined for figural and literary, and not communicative, ends. In short, I see only mojibake. If you read in Wikipedia or elsewhere you learn that mojibake refers to incorrectly rendered characters, a case where the characters in a computer document are incorrectly tagged, or where the system displaying the characters does not support the encoding, and the resulting display or rendering is described as incorrect and as gibberish. Similar terms deal with the concept in other languages, but the Japanese “mojibake” sticks, seemingly for the proximity of the pictoriality of the problem and of Japanese graphemes; and for the exoticism involved, the suggestion of the problem as fundamentally other. Mojibake is a problem of interoperability, a breakdown of a system intended to encode, decode, represent, and display every possible character. A breakdown, therefore, in a universal system of communication based on direct and correct presentation of characters. In the breakdown, the display does something else. It fails in its universality and interoperability, and in failing produces an image or set of images.

I feel unclear and idiotic on the notion of incorrect character encoding. I look at a screen and it may appear beautiful, it may appear horrendous, it appears in many ways, and — idiot that I am — I find it difficult or impossible to insist that a display is correct or not. I insist that no method can account for what the screen displays. Let me add: I worked hard to arrive at this insistence. It is the same way that I find (or seek to find) it impossible to distinguish any writing on the net from spam.[3] In the face of a universal and interoperable technics of the written character, all writing is spam and all characters mojibake. It seems to me that correctness in rendering is not self-evident but partakes of the same desires for universal interoperability as digital humanities: to move from displays and images to abstraction and knowledge, to make everything into projects (fundable, publishable, tenurable, etc.).[4] Is any project possible without correct character rendering?

I started with the question of “should I read.” I could add, should I write: What if characters displayed or rendered on screen are not read; instead, they matter, precisely in the way literature matters? Every should I? and every I should is a literary question, a question of permission and of otherness. “The literary” is and must be asserted and discovered. The ethical should is inextricable from the literary and remains secreted in the digital. The matter of characters displayed or rendered on screen is intensely human and humanistic, an intensity that goes to that last paragraph of Adorno's Aesthetic Theory, which tells of the suffering that is art’s “expression and which forms its substance. This suffering is the humane content that unfreedom counterfeits as positivity”  [Adorno 1997, 260]. What if the positivity of the character displayed or rendered, their thingness, is the intensification of the world's suffering? Even in the most incorrectly rendered characters? The literariness of the digital as this expression. What if? Keep in mind that every “what if” is also a literary question, a question of latency of the world, and of the poetics or productive latency of the world’s narrativity.

I should write of rendering, as in rendering characters on the screen. To “render” is to express, to represent, to give, to produce, to surrender, to narrate, to vomit, to melt down and clarify the fat from an animal, and to extract by means of heating. The last brings us back to the phenomenology of characters, as inscribed intentional glyphs, however technically occluded the agency and pragmatics of intentionality; as surfaces intensified through graphematic marking; and as necessarily given to the eyes for reading subjects, however distant and mechanized those readers may be. In other words, rendering brings me to an organic continuity, an epigenetic landscape of the organism within the technical display, a flesh infrastructure. And to narrate: every rendering is a tale, a story of the flesh. All this is found in the problem of literature.

The passion and problem of the literary takes place today in digital humanities. If digital humanities exists — I am not sure it does, though I wish it did for the possibility of what it might be but is not yet — if it exists, it is as a miasma or screen covering deep desires and even deeper problems. It certainly would not be what it purports to be now: a set of tutorials for the latest tools. If it existed, it would deepen the desire and the problematic, and name both clearly: the name of the desire and the problem is literature.

First the miasma. “Digital” is a widely used term that remains loosely defined in principle and clarified as needed in practice. Since digital humanities claims to be a method or set of methods, it implies clarity and definition (just look at the many articles staking positions on “what is digital humanities”). Fortunately, there is a clear one: “digital” is defined as dealing with discrete information and the processes around the storage, transmission, and reception of such information. The definition is self-referential, since the modern understanding of information following Claude Shannon is premised on the discrete. To write “information” is necessarily to engage with the digital.

Unfortunately, the vast majority of digital humanities projects in no way require digital information. They are not digital in any significant or necessary way. I will make this case with several arguments, some better and some worse. I see these arguments as obvious, even idiotic. Let me be clear: as I claim below, protocols and standards such as the Text Encoding Initiative (TEI) do engage with the digitalization of text, and do so precisely as a displaced engagement with the literary, but in ways intended to lead to projects that, as “projects” (or intentional works), are not digital. I argue idiotically, two incompatible positions: digital humanities is not at all digital and only exists as a warmed over version of the institution of literary criticism; and if digital humanities is in fact digital, it is only because it is focused on the problem of the literary (and not on the institution of literary criticism). A program is at work here, a tropology of inscription and displacement. The several arguments I offer separate projects from encoding practices to focus on the poetic core of digital humanities, a core that is the contained and muted (discrete) problematic of the literary.[5]

As a start, you must admit that the broad field of visualization and multimedia presentations — undoubtedly useful re-imaginings of humanities work beyond the page and book, and the examples most vividly invoked in popular accounts of digital humanities in the New York Times or elsewhere — do not require a computer. Or, if we (you and I) insist on the computer in this case, it is for its ability to display many media forms, no doubt an ability enhanced by flexible digital storage and processing, but an enhancement and ability not in any way impossible without such storage and processing. If I create a humanities project with paper, with images cut out of magazines, with film, with cassette tapes, the result may be retro but is not in any substantial way different from what appears on my computer. Digital humanities projects are undeniably prettier and more exciting than the paper and scotch tape project I described. (Is that what we mean when we say it is an exciting new field? Digital humanities is pretty humanities?) Let me add: I see nothing wrong with this, as long as we come clean and admit it.

Social media or crowdsourced or semantic web projects are no better. If we grant that such Web 2.0 concepts differ from Web 1.0, and this is not self-evident, it must also be granted and truly self-evident that these again do not require digital information. Crowdsourcing is perfectly implementable outside of and prior to digital encoding, as are “friends” of various sorts. Even more: the very impulse to recreate a social network in “digital media” seeks out and leverages not “the digital” but the hiding of the digital aspects of the medium. Visualizing a social network uses the computer as a medium to picture relationships that are not digital. Such applications go to great effort not to be “digital.”

Things are equally vague with the seemingly similar notion that digital humanities involve humanities “beyond print.” This is a repetition of arguments such as Marshall McLuhan's in The Gutenberg Galaxy and also does not relate to digital encoding. This “beyond” could be addressed in terms of radio, film, and so on. This is not to say that computers do not fit in the sequence of media moving beyond the printed page, but the computer fits in a sequence of media and not of analog to digital encoding. Or rather: the new threshold of media using digital encoding would be part of the trajectory of media described by McLuhan. We should talk of media humanities.

What exactly do we get in insisting that this work is digital? The work under consideration is certainly displayed on computers and using new media devices, but such devices are not “digital.” The phrase “digital media” seems to describe the computer, but that object is plastic and wires, glass and tubing, and only the data stored in it can be called digital. Digital data could just as well be stored and computed in other devices, including abacuses and fingers (i.e. digits). Digital refers to the domain of the discrete and computable, but it does not equal a computer. Discrete computation can be on paper, and indeed the characters we use for writing and enumeration are already discrete elements. A computer can just as well be implemented out of Twinkies or rocks, Lincoln Logs or atomic particles. To speak of “digital media” or “digital technologies” is to refer to the fact that such media or technologies encode or carry digitally encoded information. We easily move from this to the medium itself through a synecdochal rhetorical trick, the encoding extended to the whole, moving from digitally-processed information to the content and semantics of that information, as if the fact of the encoding spread across the whole, allowing for a marketing style boosterism of the all that takes place on the computer as “digital,” a labeling that suggest a novelty and innovation that is not evident but is everywhere.

At this point, you may be annoyed or even infuriated, and you may insist that it is clear that the powers of display, transmission, storage, and so on, associated with these media only became possible with the development of digital information processing. Perhaps, but such a counter-argument means we deal with by-products of media using digital encoding, and not with discrete digital processing. The speed, storage, and so on are part of a lengthy history of speeding up, a “logistics of perception,” in Paul Virilio's useful phrase, of which computers with their digital encodings are only the latest version. Such a history could be centered on the discrete mark, on the digital as a long-range phenomenon. Any such explanation must recognize a shift in literacy that includes the impact of television, film, paperbacks, the printed book, and — why not — papyrus. We might be better to speak of modern humanities or faster humanities, or anything but digital humanities. Perhaps we should talk simply of humanities, since this speeding up is part of the era of knowledge that makes humanities possible. Or, in a rigorous move, cast the digital back into the fundamentals of discrete elements (e.g. Stiegler 2009). This is a persuasive but complex argument, and destroys “digital humanities” as an emerging field. Television, film, print, papyrus, all are digital, taking us into the very fact of the graphic mark as digital, to arrive at the question of the letter, that is, of the literary.

If the majority of digital humanities projects are not digital, arguably what remains is a minority of projects that are digital but are not humanistic. The use of computational models such as hidden Markov chains, Backus-Naur forms, rewriting systems, and so on, can treat humanities materials as formal languages in the computer science sense. The premise of discrete alphabets of digital characters is precisely the condition of such languages.

I will deal with these models elsewhere, but I argue that encoding practices such as TEI are ways of treating text as aggregates of discrete characters. Character encoding builds on the already discrete qualities of written traces, and in a way that poetically extends the non-identity of these qualities and the literariness of the inscribed mark. Early discussions of the roots or fundaments of language focused on “stochoie,” which Max Müller described as degrees or steps making up a whole. The term came from physics, naming basic elements of nature, but was applied to the letter of the alphabet by the Greeks. In Latin it was translated as “littera” or letter. As I will elaborate below and elsewhere, formal languages and encoding are engaged with that long history of the literal, a history inextricable from a material and prosthetic exteriorization. At which point the humanities no longer deal with human utterance and expression, even if such applications are applied to the archived traces of natural language.

Let me consider, secondly, in what ways digital humanities is literary. It is hard to know how to begin to answer this question, given the still uncertain status of digital humanities. The 2011 conference at Stanford billed itself as “Big Tent Digital Humanities,” emphasizing heterogeneity and lack of a single method or view. I began earlier with the literariness of the act of naming, and “Big Tent Digital Humanities” said almost too much by publicly airing the problem and poetics at the core of digital humanities. As a phrase referring to political gatherings where participants forsake their philosophies for the sake of the overall movement, the ideological concealment of “big tent” is perfect. (Think of the connotations: the carnival, the religious gathering, the Sunday picnic, and so on.) The fuzziness at the core is both a problem and the true method of digital humanities. A program is at work, or is executed, in the broadly sketched opposition between the chatter and glamour of digital media as faster and pretty forms of humanities projects, on the one hand, and the inhuman calculability of formal language, on the other. The digital is a poetic tropology posited at the core of digital humanities: on the one hand, encoded information paratactically referenced in screen outcomes; on the other, deep computational structures beyond human comprehension. This spectrum is the occluded problematic of the literary. There appears to be a “field” of digital humanities, given the availability of project and infrastructure funding, such as the vanguard NEH program of Digital Humanities Start-Up grants; the establishment of an essential journal such as Digital Humanities Quarterly; and the many job offerings now featured in the Chronicle of Higher Education (at the same time as openings in other fields dwindle). Of course, remarking on this rapidly developing yet unclarified and uncodified state of digital humanities is itself part of the self-understanding and heroic emergence narrative of the field: to discuss it is to comment on its newness and innovation.

At first glance, there is little literature in digital humanities. Consider the showcase NEH Start-Up grants. A search done in early October 2011 of grants awarded by this program finds that out of 251 funded projects, the keyword “literature” appears less than ten percent of the time (16 times) and “literary” even more seldom (13 times). When these terms are used, literature refers to a disciplinary rubric (a project might appeal to “literary scholars”) or to generic distinctions for identifying certain types of texts as literary texts, (e.g. as verse or as a novel). These distinctions are then typically opposed to other clearly demarcated classes of scholar and text, such as “historians” and “historical texts.” To broaden the search to uses of terms such as “narrative” includes a wider range of projects but also deepens the question.

In general, no digital humanities project directly deals with literature as a topic, or with literariness. Based on the descriptions in the NEH library of funded projects, the only project that directly engages with the “literary” as a research question is the “Electronic Literature Directory: Collaborative Knowledge Management for the Literary Humanities.” This project asks what role literature, in its specificity and as against other forms and practices, can play in the digital humanities. The project concludes that literature can play a critical role in the traditional sense of literary criticism as humanistic discernment. I will address the question of “electronic literature” at length elsewhere. It stands in its own problematic and troubled relation to digital humanities, or to “digital” and “humanities.”

What is clear is that all projects in the grant program, and indeed all projects appearing under the rubric of the humanities, are concerned with textuality in some way. All the projects are proposed in a document and lead to a variety of texts, most typically a “white paper.” At one extreme, “literature” means “it is written.” At this second glance, digital humanities seems deeply literary. It is clear that the crucial monuments of digital humanities are built on literature. IATH, undoubtedly the template for digital humanities centers nationwide, incubated well-known literature-oriented projects such as “The World of Dante,” “Uncle Tom's Cabin and American Culture,” “The Walt Whitman Archive,” “The William Blake Archive,” “The Rossetti Archive,” and “The Dickinson Electronic Archive.” All are famous examples of digital humanities. Projects such as the Rosetti Archive are not simply excellent examples of digital humanities projects but function as paradigms. The application of digital humanities to these canonical literary works produced canonical digital humanities projects. The Rossetti Archive is an exemplary and long-lived digital humanities project — begun in the pre-web days of 1992 and conceptualized even earlier — and now functions as a touchstone for scholarly discussion of the development of digital humanities. Jerome McGann uses it to frame his consideration of Radiant Textuality: Literature After the World Wide Web. If the project is at some level “simply” a place for storing and retrieving documents by and about Rossetti, it is understood as far more: a prototype and reference point, an argument in the form of an archive. “The Dickinson Electronic Archive” is the same, and also notable for its role as a switchpoint of energy and interest from Virginia to Maryland, where its creator and curator Martha Nell Smith built on the success of the IATH project to found the Maryland Institute for the Humanities. Inseparable from the archival is the foundational, institution-building quality of these literary projects. Of course, I could go on and list other projects at other centers, many of which are paradigms of digital humanities scholarship.

The visibility of this paradigm is displayed or staged even more intensely as digital humanities becomes the star of the humanities in public discourse. Take a look at the “Humanities 2.0,” a series of articles by Patricia Cohen in the New York Times, begun in November 2010, where literature plays a crucial but somewhat mysterious role. Cohen states: “A history of the humanities in the 20th century could be chronicled in ‘isms’ ― formalism, Freudianism, structuralism, postcolonialism ― grand intellectual cathedrals from which assorted interpretations of literature, politics and culture spread.”  [Cohen 2011] She then adds her claim: “The next big idea in language, history and the arts? Data.” While Cohen introduces digital humanities as a whole, the majority of her examples are projects dealing with literature, and the status of literature in these projects is troubling for all concerned: it is an exemplary display of the success and danger of the method. One researcher describes being “excited and terrified” by the way digital humanities is “actually shaping the kind of questions someone in literature might even ask.” The “potential of electronic tools to reduce literature and history to a series of numbers, squeezing out important subjects that cannot be easily quantified” hovers over these articles.

At the same time, we are repeatedly told that literature remains beyond the method of digital humanities, despite re-shaping and reduction to quantities. Anthony Grafton declares to Cohen: “I don’t believe quantification can do everything. So much of humanistic scholarship is about interpretation.” The exemplary status of literature as the unclarified core of the humanities is never clearer: it is the text that digital humanities never tires of, it is the perfect crash text dummy for the new methodology that always emerges unscathed.

My point is that Cohen's suggestion that digital humanities replaces the humanistic focus on interpretation with data processing methodology is too simple, or rather too little as an explanation. Instead, you should recognize the literary tropism of digital humanities: on the one hand, it re-energizes the study of literature in a way that goes beyond the various “isms” of theory and is rigorously grounded in method. Of course, digital humanities does this for a variety of fields, not just literature; but literature surely was and remains the apex of debates about “isms.” The bloodiest battles of the “ism” wars were fought in literature departments.

On the other hand, and more than this: literature is not just one case among others. If it is returned to again and again as exemplary, this repetition goes beyond the fact that literature is resistant to data-driven methods. “New digital tools are bringing new ways to teach humanities courses, even Shakespeare” writes Cohen. The “even” is supposed to indicate both the supposedly transgressive move entailed in applying digital humanities to the Bard, but also to indicate the expected and inevitable place of Shakespeare in this argument, where “even Shakespeare” means that the argument could not be complete without Shakespeare. Cohen's examples belong to an existing motif of understanding media through the lens of re-situated literature monuments, a motif that includes “Shakespeare on Film” — in all its configurations, from books to films such as Shakespeare in Love or the recent Anonymous, a film that cloaks pseudo-questions of Shakespearean authorship in a core re-assertion and intensification of the survival and of humanities through interest in literature — or similar versions of this motif, such as Jay Clayton's Charles Dickens in Cyberspace or Jerry Flieger's Is Oedipus Online? (testifying to Freud as literary author). In such works, the literary functions as a topos mobilizing well-read cultural materials to understand new forms. The anxiety around the danger of the new is a way of re-asserting the literariness of literature. Indeed, the exemplarity of literature rubs off on the other projects Cohen looks at, so that history or art become equally rich but also inexhaustible subject for digital humanities. Literariness circulates in the description of other fields, without any clear conceptualization or without even being named (no need to state that digital humanities treats history and the rest as literary).

In all this, there is no need on Cohen's part to read literature, no requirement to cite texts by Shakespeare (or other authors). I would argue that the projects at IATH and elsewhere discussed above possess similar citational value. Literature “itself” is the citation, an invocation that adds literariness to the digital humanities overall. What is the point of this citation? Firstly, proof of the survival value of literature, “even” in the digital era. Literature is the survival value of texts, it is the name of what lives on in texts, in this and every era. Literature is stored up and hidden in every digital humanities project.

And so, secondly, the renaissance of digital humanities is a literary renaissance, a renewal through the “project” (through every digital humanities project) that finds its paradigms in self-declared literature projects such as the Rossetti Archive. Moreover, the citation of literature implies “depth” in digital humanities projects, a deepening of projects that ensure their value and return on investment. Literary texts bear repeated reading, including repeated critical readings. (How many books are there on Hamlet?) So too digital humanities projects will bear re-use and add something, some specific “digital” something to the work. The point is not what is added but that literariness provides a place for this value. Does it matter if the projects are used or is the point simply that they exist? Surely the frantic desire to build or “start up” monuments of digital humanities exceeds any possible use of the projects? Unreadability is the point, unreadability of a literary “core” or “origin” to all products, unreadability as the guarantee that we need new projects.

What can we make of this unreadability? How do we take it? Perhaps the deep ambiguous unreadability of the literary is the name for the satisfaction of the digital humanities project. Is this really so? How can we tell? It is troubling to say that literature, with all its fictionality and delusiveness, stands for the satisfaction of the desire named digital humanities. Literature is necessary to the humanities — if the humanities is not literary it is nothing else. Literature is the promise of the humanities' survival. But is the survival of the humanities anything but this promise? Is the survival anything but a story told, a figure of language, a bit of writing? The literary in digital humanities is an occluded tropology promising monuments and “projects.” What is a project, that core object of digital humanities, other than a version of the literary work? At the same time, the literary in digital humanities is a productive, or more precisely, a poetic momentum.

The project is a monument built on the momentum of poetics.

It is remarkably difficult to distinguish literary texts from other texts. Close scrutiny of so-called non-literary texts always brings to the foreground problems of the poetic turn, authorial voice, or figural language. Where text declares itself to be literature, announces its literariness, it remains equally difficult to define the literariness given in the text, beyond the fact of this announcement. No end of definitions are possible, but the sheer ongoing existence of literary criticism is a tribute to the non-conclusion of this project. You can decide that a text is literary because of metaphor or tropological play, but the only clear thing is that such a decision guarantees that anything other than these qualities determine the literariness of the text. The distinguishing characteristic of the literary is our inability to distinguish its character. Literature is always something other than itself.

I will not rehearse the complex history of the discourse on “the literary.” Instead, I will briefly examine TEI, the Text Encoding Initiative, surely one of digital humanities' great accomplishments. My argument is in three parts.

First, literature functions as the core example for TEI. By “core,” I mean to highlight that TEI begins its self-definition by invoking literary texts. Literature is necessary to the definition of TEI.

Second, digital humanities scholars already recognize the literary potential of TEI. I focus on Julia Flanders’ provocative arguments about text encoding as a “world-generating” performance that opens the possibility for “poetic” markup.

Third and finally, TEI builds its momentum on the poetics of the letter, on the literary mark. Character rendering is a literary performance, and we follow the figural turns of this performance when we speak with and engage with the computer and its displays.

Literature functions as the core example of TEI. As originally set out in 1987 and formulated over the following years, TEI principles are oriented towards humanities texts. The principles are well known: they make “recommendations about suitable ways of representing those features of textual resources which need to be identified explicitly in order to facilitate processing by computer programs.” Specifically, TEI provides tags added to text to “to mark the text structure and other features of interest.” Encoding text leaves the text in place but adds a surplus coding of marks. The monuments of digital humanities I described above are variously enabled by TEI tagging.

The primary examples in the TEI guidelines are drawn from literary texts, both hypothetical and actual. The guidelines begin with a markup of William Blake's “The Sick Rose.” The example concludes:

The standard adds that this “is an XML comment and is not treated as part of the text.” The example of the boundary between TEI and text is exemplified by the boundaries of a poem — in this case the canonical Blake poem familiar from Norton anthologies — and the larger unit of a hypothetical anthology of poems. The burden of marking “text structure and other features of interest” is carried by the example of a poetry anthology. For our purposes: poetry is the analog boundary of discrete digital tags.

The guidelines continue with many more literary examples. Of course, the presence of these examples no doubt indicates the preference and background of the authors. Guidelines such as TEI are protocols, in the sense of the Internet Request for Comments (RFCs) that provide recommendations for the operation of the Internet. Such protocols describe the behaviors of a community. They are prescriptive and binding but not absolute. They constitute a crowd. In this case, the crowd of digital humanities is again a community of poetry readers, who agree in principle on recommendations for marking up. Once again, a program is executed and at work.

Digital humanities scholars already recognize the literary potential of TEI. Julia Flanders and Jacqueline Wernimont, in a presentation given at the Digital Humanities 2011 conference at Stanford, proposed approaching TEI markup as a practice of poetics [Flanders and Wernimont 2011]. Flanders and Wernimont proposed to adopt the concept of “possible worlds” from philosophy, and from its deployment in narratology, to “describe how authorial markup might leverage the formal tools of a structural model in order to enact a generative or poetic mode of markup.” My short summary will not do justice to their argument. To define what they mean by “authorial” markup, they begin from “long-standing conceptual tension within the markup (and especially the TEI) community between two models of markup.” The first is rooted in “mimesis and surrogacy,” while the second is “is more concerned with meaning creation and the domain of annotation, interpretation, authoring.” The authors focus on the second model, which they propose is suspended between representation and production. “The most common examples in the present day include annotation, interpretive markup such as the association of themes and keywords with spans of text (e.g. using the TEI @ana and <interp> mechanism), and the creation of new documents such as articles using an XML markup language as an authoring system.” Flanders and Wernimont conclude that “the markup itself becomes a world-generating mode of knowing that must carry several registers of meaning arising from different kinds of scholarly agency.”

While Flanders and Wernimont set out an argument about two different forms of markup, the ground of their claim is the “performative and illocutionary qualities of markup.” Which is to say: all markup is performative, all markup must be authorial and world-generating. The most “mimetic” markup, devoted to the straightforward representation of the text would surely be a great achievement of authorship, where the registers of meaning would be cunning, almost ironic in their silent overlay. The momentum of this poetics is such that there can be no way to determine whether markup is a representation of the text or in a fictional relation to the text.

TEI builds its momentum on the poetics of the letter, of the literary mark. It is not that the text turns poetic and thus digital humanities is large enough to accommodate this turn. No, the computer is not textual, or rather is only textual after the fact. It is not that we read the screen and possibly discover literary texts in our reading. On the contrary, we read the screen because it is literary from the first. The literariness of the computer is the condition of our reading.

Much more could be said. There is always more idiocy. What do you make of Saussure’s remarks on the written sign that the “actual mode of inscription is irrelevant, because it does not affect the system. […] Whether I write in black or white, in incised characters or relief, with a pen or a chisel — none of that is of any importance for the meaning”  [Saussure 1983, 118]? Or for that matter Freud's remark that “the single letters of the alphabet […] do not occur in pure nature”  [Freud 1965, 261]? The trajectory of the problem of the literary as digital can be unread, tracked, allegorized, and lost through a much more complex history that casts the discrete back into text encodings that include Morse and ASCII and FIELDATA, but also Viète and Bacon's ciphers. Still, you want the literary. You want me to address the literary in digital humanities, whereas all I do in this essay is speak to its absent efficacy.

Anthony Wilden, in his short essay on “Analog and Digital Communication,” which after forty years still remains little-read and yet one of the few useful texts on the topic, respects the complication of distinguishing the analog and digital. In fact, he sees such distinction-making as itself implicated in the boundaries of the analog and digital. For Wilden, the negation of the literal is what orders the analog and digital. In short, any communicational “ensemble” may be understood as analog or digital depending on the boundary drawn and the position of the observer vis a vis the communication. In terms of character encoding, Wilden's argument about negation is important: he argues that there is only negation in the digital order, and that negation corresponds to relational meta-organization of information. The presence of zero as the simple flip of the digit is enough to negate the letter.

Encoding sites the letter. Encoding grants the letter presence, while a different encoding negates it and puts another letter in its place. For Wilden, this structural fact is introduced into the analog in a way that negates the material substrate bearing or carrying a mark, and poetically and punctually produces the digital mark. The digital mark is information only as it negates or re-marks what bears it (178-188).

The informational and energetic actuality of the discrete mark is a problem of boundaries and systems. It is, in short, the problem of the literal, which itself is another name for the problem of the literary. (The literary: a problem with multiple names, already a story at work, a story of the letter and its literary literalness.) We deal with the remainders of system-generating literary inscription, with the guarantee of a narrative resting on the meaning and sense-making around sites of literal inscription. It is written: literature specifies topics of contention and debate, including specifying conditions of authorship and of the nature of the work, and of topics that are carried out as discourses identified as digital. Open source or file sharing are examples of such topics: part of the institution of literature that programs in advance the debates and contentions. Literature names the discreteness — the digitalization, if you want — left by the literary. Discreet and discrete: literature’s fictional narrative is poetic, productive, but not at all truthful, and can tell us nothing of the literary except that it is written.[6]

Let me stick close to the concerns of TEI and digital humanities to focus on the literal inscription on the screen and in the computer. As an XML document, TEI-encoded text relies on the Unicode standard. Unicode guarantees that all XML documents, and thus all TEI markup, “whatever languages or writing systems they employ, use the same underlying character encoding (that is, the same method of representing as binary data those graphic forms making up a particular writing system).” The guidelines continue: “Unicode provides a standardized way of representing any of the many thousands of discrete symbols making up the world's writing systems, past and present.”

Following this statement of the importance of Unicode, the TEI Guidelines immediately turn to “textual structures,” stating:

While TEI will foster an awareness and understanding of the complex relation between text and its units, the fact of this differentiation is fundamental and unquestioned, a fact that is given literally as a datum through the literalness of Unicode.

The story I am telling you might climax here, here is my climax: Unicode as the literal, defining digital writing. On the Unicode Consortium's “Acclaim for Unicode” page, James J. O’Donnell, the classicist, digital humanist, and provost of Georgetown University declared: “Unicode marks the most significant advance in writing systems since the Phoenicians.” The significance and advance need to be followed and understood. There is only one code on the net and it is Unicode. Read the following carefully: it is “the universal character encoding standard for written characters and text.” The Unicode Consortium’s “What is Unicode?” webpage, providing their central definition, begins with the following mantra: “Unicode provides a unique number for every character, no matter what the platform, no matter what the program, no matter what the language”  [Unicode]. From most points of view, the capacity of Unicode is enormous. “The majority of the common characters used in the major languages of the world are encoded in the first 65,536 code points,” but the sixteen-bit encoding has the capability to encode up to 1,114,112 code points. Unicode even encodes fictional writing systems such as Elvish or Klingon. Its codespace includes any writing system whatsoever, without regard to whether this writing was ever employed by human culture. “It provides the capacity to encode all characters used for the written languages of the world.” What an achievement! To encode all humanities writing systems, past, present, and fictional! More than this: since streams of ASCII are the basis of all file transfers on the net, since ASCII is now a subset of Unicode, and since Unicode provides a structure for exchange and storage of data, then we must recognize that this encoding is the fundamental writing of all that is on the net.

Brian Lennon's excellent discussion of Unicode focuses on the limits of its claims to totality: “As a system defining universally efficient transmission, it must emit redundant waste in the form of merely local variation, precisely to claim its place as system”  [Lennon 2010, 170]. As with the ASCII standard, Unicode is easily critiqued for the way it inevitably re-maps geopolitical concerns and contentions. But it is the distinction between character and glyph, central to the logic of Unicode, which accounts for the literariness involved. The enormous range of encodings, on the one hand, and the flexibility in transmission, storage, and display of characters, on the other, all come down to this distinction. Joe Becker’s original 1988 draft proposal for the Unicode standard states that a “clear and all-important distinction is made between characters, which are abstract text content-bearing entities, and glyphs, which are visible graphic forms.” This distinction is carried on in the Unicode standard as implemented and maintained to this day. The distinction is in operation within all encodings and implies a philosophy of the screen and of the object. The current standard Unicode 2.0 Core Specification released on February 17, 2011, states: “Characters are the abstract representations of the smallest components of written language that have semantic value.”

What is meant by “abstract”? Character refers to the “abstract meaning and/or shape, rather than a specific shape.” What is meant by “semantic” here? Not that a given character is meaningful. The character “a” may or may not be meaningful. Rather, as Korpela’s Unicode Explained argues, it “would be better to say that a character has a recognized identity and it may be sometimes used as meaningful in itself”  [Korpela 2006, 11]. Meaning and semantics are part of the recognizable identity of the character within the system of Unicode. “The character identified by a Unicode code point is an abstract entity, such as ‘latin capital letter a’ or ‘bengali digit five.’ The mark made on screen or paper, called a glyph, is a visual representation of the character” (5).

What is a glyph? A glyph is perceived. “Glyphs represent the shapes that characters can have when they are rendered or displayed.” A glyph is also technically produced. It is an outcome of technology. “In contrast to characters, glyphs appear on the screen or paper as particular representations of one or more characters.” Glyph and character are part of a single writing technology, the one side abstract, encoded and conceptual; the other side, material, perceived, and undefined.

Obviously, glyphs outnumber characters. There are multiple possible renderings for a character: many sizes, many resolutions, and many forms of visibility, any of which may be recognizable as characters. There is no comparison with glyphs: they are fundamentally unnumbered. They appear and are seen, but are not and cannot be defined by Unicode. Glyphs are recognized for the encoding that they render. Codes are set out in tables and carefully controlled. Appearances are disordered and without accounting. The difference is between wild phenomena and specific forms or characters. The character on the screen, the text that I see and read, is not a body. There is no character there but only an encoding that I read through the appearance. What I see, but not what I read, is an innumerable disordering of appearances. Characters are the logistical construction of “reading” and “seeing” through a systematic technical distribution of the visual and symbolic.

The subjectivity of the writer and reader of character codes is not a function of the so-called gaze of psychoanalysis, a concept that involves the problematization of the object in relation to the eye. Unicode empties out appearances of any subject-object relations. There are no relations in the visible, no subjects and no objects, only a vast and innumerable field of appearances. Code is the truth that appearance is not: all that appears is nothing but the negated world of matter, nothing but flows of intensity, outside the abstraction of code. The gaze, the site of the subject and of all of our anxieties and traumas, is an undefined surplus effect left on the screen, never penetrating or seeping into the code.

Of course, I recognize that code is visible too. Gaze seeps across code and smears its surface. Therefore, the visible field must be controlled. It is paradoxical that Unicode must be represented in charts, not to mention in books and other writings: tables, guides, diagrams, and so on, all displaying what is not possible to make visible. What are these charts? They are depictions of “character” as the gaze on the encoded character. The charts come with a disclaimer: “Character images shown in the code charts are not prescriptive.” Not prescriptive but narrative and poetic. For this there is the “representative glyph,” which is “not a prescriptive form of the character,” and yet “enables recognition of the intended character to a knowledgeable user and facilitates lookup of the character in the code charts.” It is a paradoxical appearance through which the symbol is recognized. The representative glyph is immediately false, not to be taken as the abstract character but as its perfect fiction.

The visual and symbolic churn and turn within this mess of glyph, character, and representative glyph, a mess of allegorical narratives of the literary, narratives that go by the name digital humanities, are all built on markup and method as the poetic turning and returning to (and in relation to) literary inscription. It is a return leaving not the literary but literature as the wayward trajectory of marking. I tell you: the entwined source of the literary and digital humanities is here.

The history of character and glyph are similar. The earliest English usages of both refer to a carving, a cutting, or a marking. In short, usage refers to the problem of rendering. By the seventeenth century, character came to refer to the mental and moral qualities of an individual, while glyph continued to refer to the material and physical engraving. As Marshall McLuhan detailed in the Gutenberg Galaxy, the emergence and dominance of print led to the perceived equivalence of the qualities of individuals and printed characters. I tell you: not just a letter but also any appearance on the screen is a glyph etched or inscribed on that surface. Every screen is a glyph; every screen is the stochoie or littera. The literal is always being screened. Character encoding means I write it, it writes me. To screen is to touch, is to contact the other. Such a terminal screen is never touched by the other; it is never smearing the same fluid on my fingers, on my tongue. How rigid is this “never touched”? What relation exists through this rigidity?

Much more could be told of the theater of the screen, of character armoring, of the gender of the letter, and the satisfaction found through rendering and display. I remind you: the computer is not digital. Computers melt, overheat, shed light and lose data.[7] They exist in a narrowly-controlled thermal range. The website for Intel Corporation, producer of many of the world’s microprocessors, states that “thermal management” refers to “two major elements: a heat sink properly mounted to the processor, and effective airflow through the system chassis. The ultimate goal of thermal management is to keep the processor at or below its maximum operating temperature.” The computer is not discrete and differentiated, not an archive of texts, not an enclosure of circumscribed inscriptions, but a radiating sun in a constant state of decay. Using a computer is handling the decay and dispersal of thermal management — not to mention managing hardware interrupts, breakpoints, and quantum tunneling.

On any given day the amount of energy used to power Facebook pages, recipe searches, news sites and all that the internet entails is about 20 or 30 gigawatts. In terms of greenhouse gas emissions, power consumed daily by the internet is roughly the same as the airline industry.

And yet you say the computer is digital. By “you,” I imagine Julia Flanders, because I admire her work and her argument: cogently, persuasively, correctly, arguing for the world-generating poetics of markup. By “you,” I address digital humanities. The digital is this narrative: deceptive, fictional, a narrative of the otherness of the world, a narrative of project management as management of otherness. Digital means a narrative imagining the other through information. A delirium of data as absence of the world, as mourning this absence, as telling the closure of data. Unicode turns every project into narrated knowledge that can be handled and managed.[8]