Toward a Semantic Literary Web: Setting a Direction for the Electronic Literature Organization's Directory

v1.0 January 29, 2007

By Joseph Tabbi

Joseph Tabbi University of Illinois at Chicago


  1. Abstract
  2. Preface
  3. Introduction
  4. E-Lit and the Institutional Avant Garde
  5. On Evaluation as a Pre-Condition for Preservation
  6. Folksonomy
  7. Census
  8. Credibility
  9. Economy
  10. Semantic Web Applicibility
  11. Who is the Directory For?
  12. Works Cited
  13. Notes


Electronic Literature is not just a "thing" or a "medium" or even a body of "works" in various "genres." It is not poetry, fiction, hypertext, gaming, codework, or some new admixture of all these practices. E-Literature is, arguably, an emerging cultural form, as much a collective creation of new terms and keywords as it is the production of new literary objects. Both the "works" and their terms of description need to be tracked and referenced. Hence, a Directory of Electronic Literature needs to be, in the first place, a site where readers and (necessarily) authors are given the ability to identify, name, tag, describe, and legitimate works of literature written and circulating within electronic media. This essay grew out of practical debates among the ELO's Working Group on the Directory, established in the Spring of 2005 and active through the Winter of 2006. The essay offers a set of practical recommendations for development, links to potentially affiliated sites, and an overall vision of how literary form is created in a networked culture. The essay is intended to set a direction for the next phase of Directory development (Fall 2007), central to the ELO's mission of making a place for literary work (and works) in electronic environments. Finally, and as yet tentatively, the essay offers speculations on how this curatorial activity can be coordinated with similar initiatives in the arts and with stakeholders in the current development of a Semantic Web.


Thom Swiss, Professor, University of Minnesota

The quote Joseph Tabbi employs from Don DeLillo for the epigraph to his essay is a helpful one: "You didn't see the thing because you didn't know how to look. And you don't know how to look because you don't know the names." DeLillo's words orient us in the direction of the language-driven, social work that Tabbi argues for in his vision of a semantic literary web.

N. Katherine Hayles opens the aperture more widely and the angle differs slightly as well. Her electronic literature "primer" is a wide-ranging essay that takes the pulse of the e-literature field at this particular moment, reminding us that "literature" has always been a contested category.

Both essays are major contributions to the study of electronic/new media literature — useful, I believe, to those readers new to digital literature as well as those writers, critics and teachers who have helped develop or actively follow and critique the development of literature in a born-digital mode. While both Hayles and Tabbi agree on many points (and cover some of the same territory), there are also some interesting differences between the essays.

While N. Katherine Hayles is largely concerned with defining a field, Joseph Tabbi is concerned more with defining the possibility and conditions of literature's persistence in digital environments. The authors pitch their respective 'approaches' to different audiences; each seems to have a different sense of what needs to be done first — critique digital literary works (Hayles) or define the conditions for the emergence of possible digital literary works (Tabbi). Both strike me as equally important.

And most of us in the Electronic Literature Organization believe that the two critical orientations represented work together. In short, you can't have one without the other, and you can't have anything at all without the pragmatic, writerly tactics employed in the 'Two Bits' essays already online in the ELO library.


There is no larger task than that of cataloging a culture, particularly when that culture has remained willfully hidden to the routine in-gazing practiced by professional disclosers, who, after systematically looting our country of its secrets, are now busy shading every example of so-called local color into their own banal hues. A catalog of poses and motions produced from within a culture may read ... like a form of special pleading, or, at the very least, like a product that must be ravaged of bias by scholars prepared to act as objective witnesses.

(Ben Marcus, The Age of Wire and String)

"You didn't see the thing because you didn't know how to look. And you don't know how to look because you don't know the names."

(Don DeLillo, Underworld 540).

The Electronic Literature Directory (ELD) has been, and needs to be, the first place where people go to learn about born-digital writing. It needs to be extensive, open to structured input, so that readers coming from diverse fields and interests will not only find what they are looking for, but may also notice the ways in which this work of e-literature, the one they sought out or stumbled on, is a part of something larger. Those of us belonging to the Electronic Literature Organization (ELO), the Directory's sponsor, are taking part in a literary movement. Our movement, however, is unlike familiar avant garde transformations. Dadaists with their exploding alphanumerics, the proto-modernist "revolutionaries of the word," Fluxus event scriptors, and Ray Johnson's "New York School of Correspondance" are all potential precursors of electronic literature, but even Johnson cut and pasted his collages on paper, and all avant gardists, prior to the pc, scripted, printed, dictated, or typed their texts. In our case the material environment for reading and writing is itself in transition. A better term than "movement" might be "migration," since we are concerned primarily not with the static transmission of canonical works from one medium, print, to another, bytes, but the transformation of literary forms and values through media that are themselves indifferent to literature. The Directory development is going on even as the ELO itself is undergoing a major transition in its institutional structure and, for similar reasons we are having to reconsider the philosophical premises of literary activity and the political context of the structures of knowledge. In the past, the creators of directories, repositories, collections, and glossaries could assume the stability of their object and a certain agreed taxonomy. This fundamental agreement, a material achievement, allowed for productive disagreements over conceptual issues — over values, for example, over which works should be included or excluded, which audience should be served, and so forth. Categorizations and material standards, however, are precisely what have come under inspection in electronic environments, and an audience for e-lit needs not only to be identified and targeted, but also created.

There can be no one direction for the Directory. It's not as if critics have an agreed definition of what counts as "electronic literature," the presumed object that the ELO hopes at once to produce and make available to various, distinct audiences. The bullet-pointed, working description given by N. Katherine Hayles, echoed in our first published description and elaborated in "Electronic Literature: What Is It?" is a purposely value-neutral, open-ended avoidance of definition: e-lit is writing that is created using electronic components, current or obsolete: "an elegant solution" within the constraints of a given medium. We're not concerned directly or primarily with works printed in books, that much is clear. But still, the trouble with a material definition, even one that allows for technological obsolescence, is that, while we are given a comfortably positive requirement — it must be "electronic" — the definition offers nothing that can guide us either in making selections, distinguishing selected works categorically among themselves, or presenting selected works to audiences with whom the membership has a stake.

Literary definitions have never been value-neutral; they tend to be open-ended but not so relativistic that any writing, everything that includes alphabetical characters in some language printed somehow, counts as literature. "It must give pleasure"; "It must change"; "It must be abstract": coming from a working poet, Wallace Stevens, these are the sorts of requirement, resistant to positivistic, objective definitions, that a literary scholar can respond to and work with. Dee Morris for example, in a paper delivered at the November 2005 meeting in Chicago of the Modern Studies Association, produced a detailed, concise, and nicely descriptive account that is notable for its consistent negativity: by running a search on the word 'not,' one discovers that the poems Morris looks at are not hypertexts; they are not interactive, at least not necessarily so. Least likely of all, is for traditional, born-print poems to sustain readability when transferred to electronic environments, and it is unlikely that the terms scholars and writers have developed, in the print environment, will be in every instance the terms authors want in the digital realm and, when they do get carried over, the terms are not likely to carry the same meaning. This non-transferability is suggested in Morris's epigraph from Marshall McLuhan, who perceived already in the Sixties that "To treat [new media] as humble servants...of our established conventions would be as fatal as to use an X-ray unit as a space heater." Most of the celebrated archives of conventional lit, in Morris's estimation, from UbuWeb to PENNSound to the extensive literary archives at SUNY Buffalo and the University of Virginia, may be placed under the category of x-ray space heaters.

We might pause for a moment over the space-heater metaphor and remark that the arrangement of selected literary objects, rather than the identification, revision, and active circulation of working concepts, conveys a misleadingly spatial arrangement when what we are talking about is a field that develops in time. The power, but also the limitation, of a database is that no one object housed there is in itself recognizably different from any other object; only the algorithm for accessing the objects can be set to distinguish values and differences and here — at the level of algorithm and machine readable terminology — is the place where literary value has a chance to enter the picture. What is needed, at the algorithmic level, are not only, or not primarily, descriptors of poems, novels, essays, even authors — although there is of course every reason to give this information, along with titles, publication dates, and names. But bibliographical housekeeping, while fundamental, cannot by itself orient the Directory toward a potential literary development and make that development known within a developing network of readers. The means of accessing a Directory of objects under development needs itself to develop and change over time. The interface that allows access to an E-Lit Directory should, in other words, demonstrate in itself the history of the objects it represents.

In a time of transition, more generic and more qualitative terms are needed: narrativity or fiction more generally than novel, poesis more generally than poem, conceptual writing more generally than essay. Within this more generalized terminological environment, new discriminations might arise — as they once arose in print when, for example, the epistolary novel emerged out of the Eighteenth-century practice of writing letters, even if these were not intended at first to be literary. A blog may, or may not, engender a literary form; the exchange of emails, some confined to an intranet, others more freely circulating on the Web, is not inherently literary but can be made so, as Rob Wittig made literature out of these a corporate in-house chat room in his serial fiction, Friday's Big Meeting (a narrative that itself circulated over the course of five days in the year 1999 via email installments, with graphic elements sketched by the author). No machine can recognize such transformations, when and if they occur. We can safely predict that, when a new literary genre does emerge, its works will be multi-modal, combining visual and verbal elements as well as technical capacities offered by various media, some of which will persist and most of which will not. The appearance and disappearance of these multiple modalities are precisely what need to be tracked, and the Directory allows a means for such tracking, through categories and terms that are themselves evolving, linking together, and creating new kinds of distinctions.

A multi-modal environment requires not positive or fixed descriptors but rather a highly differentiated language, which is one way of understanding the consistent negativity in Dee Morris's paper on electropoetics. Morris does not avoid giving a distinct characterization of the objects under consideration. But when the description comes it, too, is cast within mostly negative terms: "The new media poems I look at here do not tell stories; they have no beginning or end; in fact, in a real sense, they do not have any development, thematically, formally, or otherwise, that organizes their elements into a fixed sequence." The poems by Cayley and Giles Perring, the lexical and graphic elaborations offered by Talon Memmott, Brian Kim Stefans, and Brazilian new media artist Giselle Beiguelman are "not lyrics remediated for the net — super space heaters, as it were — but x-rays that illuminate another way of thinking." Specifically, the "other way" involves thinking about how we think about machines, how we might think with machines, and how (if we're not attentive) we are thought by machines.

E-Lit and the Institutional Avant Garde

An opportunity, but also a risk, of making terminological innovation the primary engine for a literary migration, is that such terms are being generated in any number of fields as they come into verbal contact with the Web environment. Looking beyond the practice of literary criticism to research in the arts generally, one finds a similar emphasis on terminology, along with a similar inability to reference an agreed canon or to offer a positive definition of "works." Nonetheless, the digital arts have achieved a cultural presence and institutional recognition that has so far eluded electronic literature and the discourse on preservation in the Arts is at present more mature and more widely circulating. Richard Rinehart's thoughts on the elusiveness of new art forms can be applied to works of e-lit, and the "Archiving the Avant Garde" project for "Scoring Works of Digital and Variable Media Art" might serve as one model for our own database- and keyword-driven Literary Directory:

[New media] art forms have confounded traditional museological approaches to documentation and preservation because of their ephemeral, documentary, technical, and multi-part nature and because of the variability and rapid obsolescence of the media formats often used in such works. In part due to lack of documentation methods, and thus access, such forms do not often form the foundation of research and instruction. In many cases these art forms were created to contradict and bypass the traditional art world's values and resulting practices. They have been successful to the point of becoming victims to their own volatile intent. "Archiving the Avant Garde: Documenting and Preserving Digital/Media Art."

Precisely because works are designed to "contradict and bypass" traditional values and practices, their documentation needs to accommodate the emergence of new values and practices along with the terminology for describing such values and practices. Programmatically, such convergence is in the interest of projects like the ELO Directory and other digital initiatives based in literature, even while our respective efforts at funding are carefully articulated with respect to affiliated projects. Literature and the Avant Garde in the Arts don't want to be working at cross-purposes, not at this moment in history when the audience for "literature" could well be merging with, or emerging out of, the audience for visual and new media art. At the same time, the consistency of our aims with the aims presented in the "Archiving the Avant Garde" project, should also give us pause: for if literature is defined along the same lines as the arts, with many of the same terms made available through similar channels and portals, can literature as such be said to exist at all in the new electronic environments? The danger of a negative definition, as evinced in Morris's article, is that by the time a distinctive characterization for poetry does arrive, it might not look much different from any number of conceptual arts projects (or, for that matter, pornography or sports statistics):

Instead of narrative, emotional, or dramatic texture, instead of an isolated human author, character, and/or reader, we have a creative cultural practice through applied technology whose importance is calculated, like other ambient practices, by its intermittent captures of our searching, browsing, exploratory attention. (Morris)

Instead of literary works, we are asked to consider a generalized literariness that circulates as a "creative cultural practice" in search of "exploratory attention," which is fair enough as a description of online reading. But neither acts of attention nor "ambient practices" are available for reference in a Directory (except in the trivial sense of accounting for hits and recording the duration of visits to a site). A Directory, ultimately, references pages, paratexts, metadata, and semantic settings, not creative processes, so here again we are led to look for what is literary not in familiar genres and gestures but at the level of code and algorithm. This is the place where literary textuality needs to be embedded in a document, if the document is going to have a chance of being recognized and recovered as a work of literature. Much depends on authors being willing to include metadata themselves, during the act of composition: authors need to know at what level of writing they are writing, and what kinds of self-descriptions and semantic representations will be accessible to readers beyond one's own in-groups, networks, and communities. And so the Directory, by making available an ongoing, never definitive, specifically literary set of metatags, descriptors, and keywords, can offer authoritative examples of the kind of tagging that will be developed not only in retrospect, but invented by individual authors. (The Directory, then, can help to reconnect the notion of authorship to the social construction of "authority" as something processual and in the hands of many.)

On Evaluation as a Pre-Condition for Preservation

One goal of this paper, written in consultation with the Working Group on "Preservation, Archiving, and Dissemination (PAD)", is to discover and specify ways that the Directory (ELD) can be made to develop in concert with the revived PAD and X-Lit initiative. Neither the Directory nor the work on Preservation and Archiving should be considered as works of conservation, attractive though the language of cultural preservation may be to archivists and grant-giving institutions. Promoters of e-literature should avoid sounding too disappointed about the "loss" of established works of e-lit whose platforms now are outdated. Important works need to be recovered, but what is recovered will not always be works in their initial setting, but rather descriptions of works and representations of its operation in settings that may have since become obsolete. The recovery is itself a social and political act, since the only sure criterion for a work's importance is that someone, some group, cares enough to recognize and recover the work. This is a collective critical act that, like the recognition and reproduction of works by the dead, goes on after an author has completed a work. An author can of course take precautions along the lines indicated in Acid-Free Bits, and surely authors will be consultants when it comes time to recover or revive a work, even as filmmakers sometimes (but not always, and not always productively) consult authors whose books they are treating. But basically, the "preservation" of e-lit requires not just the author but a network of dedicated programmers, scholars, and writers — namely, us.

In an environment of ambient practices and networked agency, the likelihood that works will be lost is a relief, not a cause for regret. And by this one does not intend to pass judgment on individual works; rather, we might suggest that the model of an individual genius working in isolation, toward the creation of an object that (against all odds) lasts, is a mentality that authors working in networks can no longer afford. Within the ELO, ebr, and the ALT-X Digital Arts network, my own observation of careers and works over the past decade suggests to me that an e-lit author needs to enter a network actively, from the moment of a work's conception through the acts of composition, circulation, and site maintenance, or authorship as such, in this environment, will be denied.

A clear and explicit recognition of this field defined exclusively through multi-modal constructions and differences, might free both PAD and the Directory to become something more than curatorial projects, an aspiration that can be discerned in the following formulation from Born-Again Bits:

...the fight against electronic literature obsolescence must ultimately occur in a wider framework. Seen in a larger perspective, the problem is not the preservation of old or aging e-lit per se. It is the description and representation of electronic literature of any vintage in a neutral, open source, standards-based format — one capable of maintaining the essential experience of a work while allowing its presentation to adapt to evolving hardware and software channels through understood, regular, and automated methods of transformation.

The database "representation" of works must be "neutral" and its source code should be open and consistent with Web standards. But the act of preserving - which is to say, usefully categorizing — any e-lit object is not value neutral. Each of these activities involves "methods of transformation" in bringing the source code to a generalized, available platform. In other words, whenever an electronic object is ported to the Directory or captured through X-Lit that object "must change." Certainly it is possible in theory to "preserve" every work ever posted to the Web, literary or otherwise, by reducing the text to neutral code — but without further description keyed to searchable concepts, the preserved work is as good as lost. Preservation needs to be oriented as much to the future as to the past: we must account for the conceptual form of the literary object's creation and circulation within specific networks and among specific readers at two different moments: its digital origin and the moment of its recovery for the Directory. That recovery "must be abstract," because it cannot replicate a past reading experience or duplicate ("preserve") a concrete object; and "it must give pleasure" because why would anyone go to the trouble of recovering or listing a work that nobody, among ELO readers or membership, had enjoyed at one time or another?


Why go to such trouble, for that matter, if the Directory is to be yet another resource for yet another special interest group, when what the Web offers is a unique opportunity for recovering a mass audience for literary work. If experimental music and, later, improvisational jazz, lost their audience with the development of pop-oriented mass media, literary writing lost its audience a generation later, with the conglomeration of publishing in the Eighties and Nineties. Both of these declines occurred with the rise of broadcast media, which by their nature reach a critical mass when content is homogenized and distributed from a centralized source. The Web, by contrast, is a reception medium, and the distribution of content is inseparable from the development of collectivities of people in touch with one another. Hence it is of the first importance that any Directory development should create clear and, to some degree, automatic procedures for the creation of reader participation. This is not a question of sending out personalized invitations and collegial notices. It is rather a question of organization, through levels of trust and accreditation that offer incentives for participation and confidence that the labor of reading and note-taking will be preserved along with the texts under investigation.

The 2003 PAD committee was right to recognize the necessary limits to the "possible volunteerism of the open-source community" (PAD report 2003), although the ELO, as of this writing (Summer 2006), has not yet tested those limits and we will need to establish a simple form, accessible onsite, that allows readers and Board members to submit works as well as tags for describing works. Many of us in the ELO have near at hand one important and electronically well connected network of participant-observers - namely, students in universities. Those on the Board who are teaching classes in e-lit, and who know of classes being conducted by others, might be asked to develop syllabi in such a way that works selected may be submitted to the Directory, and works discovered by students during the course can be submitted to the Directory for convenient reference. Those scholars and writers who review works of e-lit, might be asked as a matter of course to submit keyword descriptors of these works under review. The vetting Board might also devise enticements for general readers, for example a front page of the Directory indicating "10 most recent additions" and an RSS feed for these. Then the act of entering a piece in the Directory comes with an immediate reward — the person doing the inputting is featured until 10 more entries are added.

These are only suggestions, and there may be other ways to develop a community of Directory users, the "folk" who will create, and be supported by a keyword-driven e-lit database (that is, the folksonomy or distributed creation of communal and institutional knowledge). The wiki model, which exists precisely to organize chaotic bodies of material provided by Internet working groups, could be of use in our own efforts to build a network of authors, scholars, and readers around the Directory. Wiki is a Web-technology that allows users to add and edit material on a Website continuously by using a simplified markup language based on a number of agreed conventions. Its advantage is versatility, flexibility, soft security, and (not least) a familiarity among an existing mass audience that could make the transition to reading literature less awkward in networked environments. Most wiki systems are free and open source software, so reconfiguring the system to our purposes should not be difficult, although we must be clear about these purposes: namely, even as we build an audience of e-lit readers, reader input needs to be channeled through a flexible system of peer review, reader feedback, and collaboration.

As this essay goes into production, we can be confident that by getting out the word, getting a better Directory, and getting out the Electronic Literary Collection ( ELC, published Winter 2007), more works will come our way in a more steady and systematic way. It will be desirable to set up or recruit authors in the Directory to read — and describe — and evaluate the new entries. That might yield a tier of candidate texts that a smaller rotating vetting board might read - perhaps the same as the rotating ELC editorial group who could then recruit from these entries for the annual collection. The creation of an Electronic Literature Repository (ELR) would be desirable as a complement to the more developmental, folksonomical activity I am recommending for the Directory. But no work can "repose," and no librarian can curate, without this prior work of collectively setting "directions" for the semantic description and recovery of literary works. Together with a long-term Repository (ELR) and yearly Collection (ELC), I submit for consideration an Electronic Literature Wiki (ELW) to accommodate hourly and daily use by readers, who presumably will have ideas for describing works as they browse the Directory (ELD).


The Electronic Literature Collection keyword list, like other lists that we hope will be contributed over time by various working groups and individuals, is not in itself definitive or fixed. Speaking for myself, had I participated in the ELC discussions, I might have objected to the inclusion of separate categories for "Non-English" works, "Authors from Outside North America," and works by "Women Authors" on the grounds that the appearance of such categories by themselves, unaccompanied by such terms as "Works in English" or "Male Authors," implies the continued dominance of the unmentioned categories. My own inclination would have been to leave out all such identity-based signifiers and do away with Writing School jargon such as "Creative Nonfiction." But the key point here, is not what I, personally, might prefer; nor is it a matter exclusively of what the ELC editors offer, or what casual readers suggest over the transom. The ELO should be soliciting as wide a range of categories as possible, that all suggestions should be vetted, and that a mechanism is implemented for monitoring the frequency of keyword use by readers and editors alike. This way, Electronic Literature can aspire to an alternative to the age-old, vexed question of naming as a form of empowerment for some and disempowerment for others. Instead of proceding semi-consciously, as happened (for example) in the development of racial categories in 19th century national censuses, we can create a field where naming, and its consequences, can be observed and debated, and where weightings are not given by authorities alone, but in widespread use (as observed by a board of editors).

What will be common to all ideas for setting literary value, whether based on wikis or custom programming, is that each implementation will place considerable responsibility on the ELO Board, since nobody would want unvetted contributions to be fully automated. In the layered organization of Web structures, running from XML documents through content identification to statements about documents to validation, there needs to be a trust layer, that is, a high-level set of observations controlling quality. In Web services offering recommendations to clients, this layer usually implies accreditation of some sort, and assurances of security. In our case, readers need to be assured that what they are viewing is, in fact, regarded as literature by trusted and informed agents. The "layer cake" model of Semantic Web organization is discussed by Grigoris Antoniou and Frank van Harmelen in their Semantic Web Primer (17-18). This layered approach might be reflected in a working arrangement where Directory submissions are sent to a que and picked up in a reasonable time, by a select group of responsible Board members (the reconstituted and expanded Working Group, if my suggestion is followed). As more people use the Directory, its keywords and descriptors will become more widely known and (in turn) more widely used. One needs a certain degree of faith, that a wide readership will adopt our terminology, but there are ways of providing a kernel that will get such adoption started.

Folksonomies, after all, are simply emergent ontologies, a kind of communal and institutional knowledge produced by individually casual, community-contributed tags. Folksonomies are taxonomies created by non-experts, and in this sense they are no different from the metadata framework I have recommended for modeling concepts about resources (in this case, works of electronic literature). That said, they are often put forth by their proponents as a critique of high-level, demagogic Semantic Web models: in contrast to folksonomies, high-level ontologies impose top-down, official, authorized representations of knowledge. For our purposes, these caricatures have limited use, especially at a time when the only people with really good terms happen to be those who are doing the high-level and specialized work of creating e-lit. These efforts, to take hold, need to circulate more widely and, eventually, bring in fresh concepts from a non-expert (but active) readership. So I would argue that we should be looking closely at hybrid or multi-modal systems, with simple community-contributed tags and high-level authorized models applicable to different problems. We might think of the ELD as containing two primary metadata models, one based on works, the other on "discourse" about works, which describes the semantics of each individual work.


When I speak of "value," I should perhaps indicate that I am not talking about some canonical or otherwise institutionally determined set of "major" works. Such determinations will be made, and, once made, they will be challenged: reading lives are limited, and the vast majority of past works are fated to be lost not through acts of critical judgment but through neglect: negativity, not advocacy, is the condition of preservation in a canon, even as forgetting is a condition of memory.

One of the arguments made by N. Katherine Hayles in the opening remarks to her 2002 presentation on "The Future of Electronic Literature," was that the obsolescence of particular technological platforms was making so many signal works unavailable that our movement was failing to build the canon necessary to the development of a "field." But a number of authors, Michael Joyce, for example, Stuart Moulthrop, and Shelley Jackson, are cited frequently enough, their work is anthologized and sufficiently cross-referenced in various media, for us to observe that e-lit writers, when they are celebrated, are treated very differently even from their near contemporaries in print. The Michael Joyce "papers" have been acquired by a library and a scholarly edition of Afternoon might be expected eventually, with obsolescent works "re-scored," if deemed worthy of being read on current platforms. The numerous obsolescent hard- and softwares that "our" Joyce must have worked with are not less available, surely, than James Joyce's working notes for Exiles. "Our" Shelley, having patched her work into Mary Shelley's Frankenstein, has gone on to create intricate signifying networks in the arts world, but Jackson has not gathered around herself the kind of scholarly industry, among writers, that is devoted to her self-selected Romantic predecessor. Moulthrop is himself one of the best critics of Thomas Pynchon. At the end of American literary history, Victory Garden may be to the Gulf War what Gravity's Rainbow has been to the literary imagination of the Second World War. But even as the V-2 differs from a smart bomb, so has the critical response for these two writers, born less than two decades apart, taken an entirely different form. If the function of authorship differs markedly in electronic environments, the function of and reasons for referencing works will also differ. These works of e-lit have instantiated a field, only the terms of field creation and single-author scholarship are themselves as yet unformed in electronic environments and (we might as well admit) the print formations don't seem very likely to re-form in the current generation of literary production, no matter what the medium.

What critic would write a scholarly essay on The Unknown, when it is so much simpler to enter into the network, to attend the same conferences and hang out at the same parties attended by the authors, and be written about? Who would imagine that the network of artists and folk, each with his or her own tattooed word from Skin, could be transformed into an audience for some future Jackson Notes?

Why comment on a work, when you are the work?

If there is a technological subtext to the development of a field, or its failure to develop, that may have less to do with "lost" works and technological obsolescence than with the inability, at the current moment, for the mix of print and electronic documents, employed by all potential authors, to cohere at the level of composition. Each work whose form of bibliographic citation scholars are not sure about, or authors hesitate to sample, is a work lost to literary history. Each file on every author's desktop or on the Web that doesn't get transferred to the notebook, Blackberry, or Internet cafe terminal where the author happens to be working, limits the range of the composition under way at any given time. Each decade added on to the term of copyrighted protection, detaches a generation of contemporary literature from a tradition centuries in the making. So the problem of field construction needs to be addressed, not at the level of endangered "works" but rather at the level of incompatible media ecologies and as yet unformed circuits for the reception and recirculation of texts.

Canonization as such, then, is not within our provenance: rather, what this Organization is uniquely capable of bringing to the process, with the set of talents and considerable 'aura of legitimacy' that was earned through the participation of several pioneering e-lit authors early in the Organization's development, is the ability to offer clear descriptions of works, which presumably should be keyed to metadata and tags that someone, some critic, scholar, or writer, needs to create. The achievement of credibility means that, if the ELO is going to feature works in a Directory, the Organization should be responsible for the works featured — and responsibility means that each work should have been read by at least one person on the Board, and an account of the reading should be made public (in the form of descriptors and metatags: these are the items that need to be generated from the Directory and reported on). Activity and the quality of engagement with literary works, not their number, not their audience defined in advance or separately from this activity, can be taken as the measure of a successful Directory. The curatorial and the creative, in this way, can be held together by a common interest in a developing language accessible electronically.

I recognize that my plea for an overt evaluative activity runs into objections on two fronts, one practical and one political. The practical objection is that a vetted directory is unlikely ever to be comprehensive, that is, capable of serving all those groups and individuals with a potential interest in reading and writing e-lit. The institutional objection has to do with authority and is, in some ways, easier to address than the practical concerns about comprehensiveness and sustainability. We need to ask ourselves, why a set of experts who speak more or less with a single voice merits our respect and our credence? We give credence to specialists largely on the basis of three assumptions: that they are trained by creditable institutions, and they are reasonably disinterested, and they are sufficiently dispersed geographically and culturally so as to represent a reasonable diversity of literary activity. Maintaining this credibility and literary authority can be accomplished by ELO's opening of the vetting process as well as the metadata vocabulary creation to input from many sources and all users. Selection criteria should also be public and debatable. Any solution to the practical problem of comprehensiveness, can be addressed only after the institutional authority of the Organization and Directory is clearly in place, because only then will we be in a position to interact with similar Organizations involved in similar activities - most recently, the Directory of E-Literatures in French, NT2 : Nouvelles textualités, nouvelles technologies. If an E-lit Directory is set up in a way that is recognizable by those at work on similar developments in the arts and library sciences and, more broadly, by those developing standards enabling the widespread adoption of Semantic Web practices, then comprehensiveness will follow (without compromising the stringencies of our own selective processes).



By establishing itself, in the first place, as an evaluative enterprise, the Directory will then be an engine enabling the Electronic Literature Organization meaningfully to work toward the goal stated in the 2003 Preservation, Archiving, and Dissemination report, namely, the creation of "a permanent, centralized distribution point (a "one stop shop") for PAD deliverables." Still, we need to avoid a one-way conception of distribution: we don't want the Directory to remain a 'silo' (to use the tech jargon). That is, we should avoid building a solitary project that does not interact with potentially affiliated sites through RSS feeds and content sharing arrangements. Further, since nobody to my knowledge has ever paid to read an archived work of e-lit, we must be careful to specify that the "exchange value" of the works we present is given in the activity of sharing data across the networked community as a whole. The directory is not a "store" where people shop or browse, but rather a fully developed marketplace where literary value is created, in communication with similar values, and often similar vocabularies under development in affiliated disciplines. 

One objection that we can anticipate - indeed it has arisen internally whenever someone in our Organization raises the question of value — involves a resistance to any person or group, even those possessing considerable credibility, who offer to make judgments of any kind. Such objections appear to have gained their legitimacy by virtue of social arrangements and status anxiety rather than through conceptual or material force. Anyone (the author of this essay, for example) who raises the issue of value is likely to be asked, "Who are you to say, what is valuable and what is not? How do you know that the values you were trained to see, aren't being transformed along with the transformation in the culture generally?" The fact that the questioner's refusal to state value is then ceded to other, usually commercial, authorities, is rarely addressed. Nonetheless, without setting ourselves up as generalized literary authorities, we should not deny or under-represent the powers we do have, and we should accept that the network of literary scholars, as it is now constituted institutionally and technologically, certainly limits but does not entirely do away with the assertion of categorical value by individuals on the basis of a limited and continually vetted authority. The Directory can be conceived as a way of channeling and making visible this continuous, hitherto largely hidden, evaluative activity that is at the core of literary scholarship and professional authorship. And we should note also that the free and contentious advocacy for and against works is one area where criticism and traditional scholarship converges with everyday Web practice: evaluation is the core of reading and it's all over the Web. Without a gift economy, there would be no scholarship, and neither would there be a

Most ELO members are asked to recommend works, to write references and blurbs, and such promotional work is a necessary contribution to the non-commercial support of works and authors at any given moment. This is a responsible use of literary authority, in the short to middle term. To keep works alive past the moment of their initial, professional reception, however, requires an appeal to something larger than an individual, something better defined and less easily romanticized than "the literary community." I would suggest that, the entity that might give legitimacy to the values attributed to works of e-literature, should be the collaborative electronic network itself. And by this I don't mean the computers, servers, transmission systems, gateways, and other material structures that support the Internet. Rather, I am thinking of current efforts to create, or re-create using broad-based XML tagging, a "Semantic Web" (SW). The SW conception of internetworking is relevant to the Directory project because it, too, is about managing knowledge that is not only 'stored' in a repository, but that circulates through minds in communication with one another (but whose communication, by virtue of the network, is now recorded, machine readable, and reproducible). 

Semantic Web Applicability 

Experience with ebr and its database-driven strategy for describing electronic literature shows that, along with the visual representation of essays and their relations to one another, a number of authors working at different levels need to be involved in the creation of semantic and conceptual representations. Professors, graduate students, and (not least) contributors who remain engaged beyond the date of their work's publication are all engaged in keeping the archive current, and keeping their own works alive and connected to other, later works (and also earlier works whose relevance is recognized later). The Directory and ebr each, in different ways, have something specific to contribute to a literary semantic representation - and one suspects that many affiliated projects can find ways to integrate with the Directory, more meaningfully than by suggesting an obligatory 'link' between and among otherwise independent sites. Hyperlinks join data strings. Tags, wraps, gatherings, glosses, and other semantic software applications, if so designed, are capable of identifying and producing knowledge.

Long-standing collections under development by Board members — Wordcircuits, The Voice of the Shuttle, TIR Web, Grand Text Auto, and so forth — could equally be cited as potential partners in the generation and circulation of Semantic content through the developing Directory network. ELO is well advised to develop its own network presence by making further arrangements with similar "lit-friendly" projects around the Web — Rhizome, for example, and any of the sites mentioned in the Rhizome list discussions: Variable Media, Archiving the Avant Garde, and the Canadian DOCAM projects, to name a few. Creators need to be in touch with work going on in e-lit, if only because their complementary work in the arts and new media put them in a unique position to recognize the emergence of born-digital literature as a field unto itself.

The problem faced by those who build the Directory is at once technological and social. To build a federation of sites concerned with the creation of literary value, requires the interaction of Board members with the "folk" (in "folksonomical") through protocols that are non-trivial and need to be articulated. Total control over terminology is neither desirable nor possible - and neither is it productive to seek "comprehensiveness" by throwing brute-force string searches at the largest textual and image storehouse ever built, the Web. The new way to be a cultured person, even one who consults the Harvard Classics "shelf" at, is to know what to ignore, while selecting works of value using adequate types of search, some of them automated, none claiming to be comprehensive. The Semantic Web (SW) is a possible solution to the problem of constructing rich metadata models and for sharing these models among systems, allowing the perception of what these systems might have in common — a concern with literary value and invention.

To realize any (social) ideal of a confederation of literary sites, the ELO needs technologists who can devise ways that the Directory might "communicate" with these other projects that are similarly committed to the creation and identification (through metadata) of literary works together with a discourse about works. A confederated literary network needs to create not hyperlinks, but a shared database that receives normalized data from its affiliated sites. If a new essay is added to ebr, for example, then its metadata (author and title, along with conceptual identifiers) should be recognizable by the Directory database, and accepted into it. If a new essay is added to ebr, then the system likewise should inform the ELO Directory, which can create a new entry for the essay. Basically, this is a question of registering "listeners" (or "subscribers") on one side, and of "syndicating" content on the other. This is a two-way reciprocal process: ebr signs up with ELO as a subscriber to new ELO content, and ELO signs up with ebr as a subscriber to new ebr content. RSS is perhaps a way to create an interoperability between these two (and one would hope many other) literary "silos," using a consensual, machine interpretable format.

An introduction to the development of "cooperative information systems" is given by Grigoris Antoniou and Frank van Harmelen in their Semantic Web Primer. Particularly relevant is the premise, in this text, that "metadata will play a crucial role" in both describing and integrating data sources (xv). The recognized need for the Directory to change, and go on changing, by design, is consistent with the conclusion these authors draw from the evolutionary nature of "information resources and systems" - namely, that any "infrastructure" needs to support "not only development, but also evolution of software" (xvi). The success of the Semantic Web, according to Antoniou and van Harmelen, depends "on the proliferation of ontologies and relational metadata" (211). Such success cannot be imposed by fiat or by oligopolistic fixing of a single standard or source code. Rather (like the adoption of HTML standards early in the development of the Internet), the Semantic Web construction will depend on the early adoption of XML and metadata by a core set of users. There is a constructivist, self-creating element to Web development, a tendency for structures to build their foundations not in advance, but in use: A standard gets adopted because many users adopt the standard. And that is why interoperability is so important among the Directory and databases under construction elsewhere, by other organizations for other (at times overlapping) audiences.

What the Semantic Web offers, essentially, is a social and technical context for building a federated metadata database from a set of independent databases.

Because all such cooperating, values-based Web projects depend on the proliferation of keywords and descriptive, "relational metadata," project developers will always be faced with the problem of "a knowledge acquisition bottleneck" (A Semantic Web Primer 211). Because literary works are highly differentiated and original in form and expression, descriptive keywords need to be largely "hand-created," not automated. Where automation might help to alleviate the bottleneck, however, is in the collecting and inputting of hand-generated data, and this is the place where the Directory can make its greatest contribution. Certainly input should be obtainable from as large and diverse a pool of e-lit readers as possible. Those readers whose entries are successful and frequent, might very well ascend in influence, on a "Most Valuable Poster" (MVP) model. However, we should not think of this readership as pre-existing the Directory development. Rather, our audience will emerge with each reader's activity in suggesting works and identifying semantic and conceptual content.

We should recognize that when we speak of 'descriptors' and 'metatags,' we are normally talking not about works as such, but about how works are talked about: the conversation comes first, and only then will readers, participants in the conversation, have the means of accessing "works" as such (which would reside, not in the Directory itself, but in the Electronic Literature Collection [ELC] and perhaps in the Electronic Literature Repositiory [ELR]). This distinction, between "works" and "discourse," suggests that the Directory will need to find a way to negotiate between two kinds or levels of metadata: one from editors or authors that will be more-or-less bibliographical/technical (e.g., a core description of a work suitable for entry in a library system); and one that is "folksonomical" and extends to topics, genres, influences, and so forth. It is recommended that the latter, discursive development should be formalized in a substantial Electronic Glossary of Literary Terms (ELG), edited by selected Board members but open to readerly input acquired through the Directory.

One problem with generating descriptors, is that they are recognized only as a string of characters (e.g., 'hypertext') not as a set of related concepts ('hypertextuality'). Initial, as of this writing highly tentative, advice from programmers suggests that many of the limitations of a statistical keyword search (like the one that is now implemented in the ebr system and in the Directory) could be overcome by providing the search engine indexer not just with arbitrary, editor-created keywords but with a semantic model that would include a thesaurus (one that includes all variations on a string, so that not just nouns, but adjectival and adverbial forms are given). Similar modeling could be developed through the Directory descriptors: the set of all possible topics for electronic literature is vast, so such a model could only reasonably be applied to a limited ontology of electronic literature and cyberculture: this technical limitation is another reason why the ELO is advised, at the start, carefully to establish the Directory as a resource for presenting not only works but also discourse on e-lit (which is to say, works that are describable by terms developed systematically and specifically, with a view toward constructing  a field).

In the above presentation, mention has been made of an "limited" ontology. The SW is not a technical problem but an institutional one. The establishment of common semantics among stakeholders in multiple institutions is far from easy. But that, precisely, is what the development of metatag and field descriptors is for. And it is why the ELO Directory Working Group has emphasized this aspect from its establisment in the Spring of 2005: the ELO needs to do what its members can to make available an accepted set of tags and identifiers, accessible through the Directory, so that these may circulate and reinforce themselves in affiliated projects. Affiliation, itself a relational attribute, is in many ways consistent with the relational nature of metatag descripion, and in any case actively negotiated affiliations have a substantial, political, and conceptual basis, whereas a simple 'hyperlink' has none at all.

Who is the Directory For?

If the Directory must account for the uncertainties of disciplinary knowledge, that does not mean that its builders can be unsure about its database construction. In practical discussions started in the Summer of 2005 and continuing over the following year, the necessity of providing "exact" descriptions was often mentioned, meaning that any proposal has to be given in such a way that it can be implemented by a programmer working at the level of databases and tables. A semantic description that might produce five different Directories, if submitted to five different programmers, is unacceptable. The need to define an audience was also voiced: who, exactly is the Directory for? Is it for the extensive but rather specialized network of recognized and aspiring e-lit authors? for university students and professors designing courses in e-lit? or for a more general, even popular, gathering of readers offering folksonomical descriptions of works? A clear sense of user requirements is thought to be a prerequisite to any decision concerning the Directory structure, its interface, and its maintenance. However, a too-clear definition of audience runs the same risk as the use of recognized literary categories in describing works: subject positions can be fixed in advance; structures of evaluation can be imported, unreflectively, into electronic and institutional environments where such structures might not make sense.

Rather than define a target audience explicitly, the Directory development should be driven by the language of users, regardless of their subject positions. The value of debating keywords rather than works, is that we institute not a fixed bibliographical resource, but rather a way for the discourse to emerge, keyword by keyword, category by category - and it is understood that, even as personal freedom allows for more than one subject position, a work of e-literature can occupy more than one category: the categories need to be permeable. We grow keywords, not works, and we offer this development as a reflection of the emerging field of Electronic Literature.

Today not even tenured professors are immune to the distributed and flexible nature of labor in the world-economy: our literary movement is nomadic, geographically disbursed, and seeded among disciplines. Our work is powered by international conferences and supported through Organizations and granting institutions both in Europe and the Americas and (one hopes) increasingly in the East. Nomadism is especially common when a researcher makes new media not just a subject for study, but a tool for the composition of works and the construction of para-disciplinary interest groups. The moment a professor or a writer stops regarding the computer as an enhanced typewriter, ceases to treat "techies" as service personel uninterested in literature, and seriously seeks to locate literary concerns and create works in the new media environment, that potential e-lit author no longer enjoys the implicit support of a discipline. Authors working in electronic environments soon find themselves subject to stringencies of corporate and commercial enterprises that have their own, not always compatible, social structures and values set on knowledge production, description, and location. In the academy, lip service has long been paid to interdisciplinarity: those who work seriously and well between disciplines will be, in the worst case, tolerated as mavericks who are working on 'cutting edge' theories. In the best case, trans-disciplinary researchers are respected for adding something 'new,' a value again wholly consistent with the world-economy's commitment not simply to innovate, but to require that innovation should be endless.

Rather than attempt to create a literary movement along what are now entirely conventional avant-garde lines, producing ever new, ever more specialized knowledges, the ELO Directors can better advance the cause of electronic literature through acts of discernment — specifically, by discerning what our work shares in common with established forms of knowledge production worldwide to identify affiliated projects in the arts and computer sciences, and to express this commonality with reference to the unique mix of technologists, authors, librarians, program directors, and humanities professors who would not otherwise be in communication with one another. The role of the Directory is to provide, not only a centralized source of Works, but the language now under development for describing a collective and transdisciplinary literary endeavor whose outcome is uncertain. World-Systems theorist Immanuel Wallerstein speaks of professional nomadism when he cites economic pressures on Academia, not only to increase class sizes, office hours, frequency of meetings and so forth, but (more fundamentally) to instantiate a "creeping flight of scholars, especially the most prestigious ones, to positions outside the university system, scholars who are likely thereby to find themselves in structures that ignore existing disciplinary boundaries" ( The Uncertainties of Knowledge 31). Scholars who make this move gain "no advantage," and perhaps run risks in renouncing disciplinary loyalties. Nonetheless, the flight continues, and the work of such scholars, despite their frequent eminence within a speciality, does not reproduce disciplinary categories.

Hence it becomes crucial that the Directory should serve this emergent, para-institutional network of trans-disciplinary scholarship and corporate activity. Rather than gather these energies and talents under such weak headings as 'innovation' or the 'cutting edge,' the Directory needs to establish itself within a stabilizing context, consistent with the commitment to preserving, archiving, and disseminating works of e-literature. Scholars who build and use the Directory are not necessarily producing more of the 'new'; we are rather investigating uncertainties within a delimited field, defined by what is known about modes of literary transmission, about the material conditions necessary for creative collaboration and the perpetuation of works written in different, inevitably obsolescent platforms. Instead of reproducing disciplinary categories (that evolved to some degree with the stabilization of print and the enforcement of print culture in the juridical sphere), scholarship needs to work among those programmers and Web designers who are setting standards and creating ontologies — that is, communal and institutional knowledge. Here is where the general X-Lit initiative is so important, as the standards plug-in that unifies various elements under development by our group — the Directory, the Collection, and the eventual Electronic Literature Repository.

Such standards, crucially, will be developed in the awareness of similar developments Web-wide, and the general interest in categorization and tagging that unites many different fields. Traditionally, humanities scholars defined such activity in metaphysical terms; ontological inquiry was a matter of defining "the kinds of things that actually exist, and how to describe them," what categories to assign objects based on shared properties (Antoniou and van Harmelen, A Semantic Web Primer 10). If that word has been taken over by computer scientists, if instead of "ontology," researchers develop "an ontology" for each application and a specialized vocabulary for each network of repeat users, that is itself an indication of the way that certain keywords and metags might develop, in the context of the Directory development. We might take the inflection of the term, "ontology," as an instance of the kind of material and verbal transformation both reported and generated by our movement. What had been a fundamental, but also rather specialized and technical term in Humanities research, now circulates in the vernacular — and this evolution is also something that needs to be recorded and made available automatically, so that literary scholars can analyze the development of a Field through the evolution of its language. (If our programmers add a date-tagged field, for instance, someone could use the Directory database to trace the origin and evolution of our terms.)

If what we are involved in, as an Organization and corporate enterprise, is a continuous project not of knowledge production primarily but its contemporary redescription, then we stand to benefit in this instance by following the programmers. The terminological migration, from a generalized "ontology" to specific, multiple "ontologies" that can be made operational, implies on the one hand a diminished concern with propounding timeless, eternal truths, and, on the other hand, a certain indifference about remaining on the edge of the production of the perpetually new. Ultimately, the Directory is for those migrant literary workers who are harvesting terms and hybrid strains from the languages in use and continually being produced in electronic environments. If people can find works here, that they might not find elsewhere, that is all to the good. But searching and selecting is not the only role of the Directory: rather, it needs also to be a place where readers and scholars can gather to debate about the forms and features of electronic literature, not to enforce generic categories but to realize a phenomenology of electronic literature.

Works Cited

  1. Antoniou, Grigoris, and Frank van Harmelen, A Semantic Web Primer. Cambridge: The MIT Press, 2004.
  2. Jarrett, Michael. "On Hip-Hop, a Rhapsody." in Illogic of Sense: the Gregory L. Ulmer Remix. Edited by Darren Tofts and Lisa Gye. Alt-X Critical E-Book Series, 2007.
  3. Morris, Dee. "Thinkertoys: Wetware, Hardware, and the Algorithmic Imagination." Paper given at the November 2005 Modern Studies Association Conference, Chicago.
  4. Rinehart, Richard. "A System of Formal Notation for the Scoring of Digital and Variable Media Art." Archiving the Avant-Garde project. Retrieved [24 april 2006].
  5. Wallerstein, Immanuel. The Uncertainties of Knowledge. Philadelphia, Temple University Press, 2004.

ELO logo

The Electronic Literature Organization

Colophon · The template for the Web edition of this document was marked up by Nick Montfort in valid XHTML 1.1 with a valid CSS2 style sheet. It is screen-friendly and printer-friendly; a style sheet for printer output is provided which browsers should use automatically when users print the document. To cite a specific part of this document, give the section number (such as 3.2); it's also possible to link to specific parts of this document by using the links at the top, under the heading "Contents."Creative Commons License This work is licensed under a Creative Commons License. You may reproduce Toward a Semantic Literary Web noncommercially if you credit the authors and the Electronic Literature Organization. To reprint this work in a commercial publication, contact the ELO.