22.

Electronic Scholarly Editing

Martha Nell Smith

Though they are not bound together in this multi-authored, multi-edited volume and are not likely to be in any other, this essay is the second in a series I am writing musing on the effects and meanings of the kinds of electronic scholarly editing, work extending well beyond the spatial and typographically fixed limitations of the codex, which I and others have been doing and that has become increasingly visible over the past decade. Johanna Drucker's observations about the nature of writing as she reflects upon the word itself convey well, in order to reflect upon editing, its nature and character. Writing (editing) is "a noun as well as a verb, an act and a product, a visual and verbal form, the composition of a text and trace of the hand"; "letters, words, and pictorial elements all participate in producing a work with complex textual value. At its most fundamental writing is inscription, a physical act which is the foundation of literary and symbolic activity" (Drucker 1998: 3). Editing is a physical as well as a philosophical act, and the medium in which an edition is produced (or an edition's place in the material world) is both part of and contains the message of the editorial philosophies at work. As Adrienne Rich remarked about poetry (in "North American Time"), so editing "never stood a chance / of standing outside history." Indeed, we have entered a different editorial time, one that demands the conscious cultivation of and by many hands, eyes, ears, and voices. While print editions are containers for static objects, artifacts that are by definition unchangeable once produced, the world of digital surrogates practically demands new models for editorial praxes in which editors and readers work together. Such models are encouraged by the fact that in a world with access to photographic copies of texts and images, no one has to bear the burden of forging the perfect linguistic description of the artifact, and by the fact that digital artifacts are by definition alterable once produced. After all, digital surrogates featuring high-quality color images of a writer's manuscripts offer a more ample sense of their textual conditions, including the conditions of the writing scene in which they were produced. Informed more fully about textual conditions, readers can collaborate with the postulating editor in making the editorial artifacts for electronic media in ways not possible when decisions have already been made to exclude or include data and seal the resulting artifact into a printed state.

That this series of my essays examining and hypothesizing about these new editorial praxes does not share the binding of a book's spine is appropriate for a critique of work on scholarly projects that do not have clear end points – the boundaries that variorums and reading editions (such as the printed volume of Emily Dickinson's writings on which I collaborated) do. The first essay – "Computing: What's American Literary Study Got to Do with IT?" – was directed toward my Americanist colleagues who have not embraced new media for their scholarly work. The present essay is directed toward upper-level undergraduate and graduate students and instructors, seasoned scholars and editors, as well as other interested readers, many of whom may have embraced new media for their academic work while others may just be exploring the territory, as it were. Like that first essay, this one is concerned with the changing sociologies surrounding new media critical endeavors, and is focused particularly on the impact of those sociologies on the work and workers involved in electronic scholarly editing and the development of its principles (described in Part Two of this book). The changes evident in digital humanities in how we work in, as, with, and for groups constitute a profound shift in humanities knowledge production. The new editorial praxes made possible, indeed demanded, by the critical environments that new media call into being are pivotal to that shift. Those editorial praxes are not only made visible, but are constituted by some of the new technologies in digital humanities, technologies that have gone under-remarked, even by specialists. These praxes and technologies are this essay's primary foci, because they make possible the digital resources that, as I remarked in my earlier piece, are more than advantages for humanities knowledge production: they are necessities. As such, these praxes and technologies should be taken into account by anyone working or interested in humanities knowledge production.

Technologies bring scientific knowledge to bear on practical purposes in a particular field, and are thus the means by which we accomplish various ends – the tools and devices on which our critical suppositions rely. Some digital technologies, the machines and software with which we produce and consume scholarly editions, are prey to obsolescence (HyperCard, for example, is for all practical purposes already unusable, as are DOS-based machines; the SGML document delivery system DynaWeb has limitations, at least aesthetically, that are much lamented; examples abound). By contrast, all but one of the technologies under review here were already used in some form for book production and are obviously not susceptible to such rapid obsolescence. In fact, in migrating to the electronic realm the technologies have become much more visible and productively exploitable. Constitutive of digital humanities, the technologies are: access, multimedia study objects, collaboration, and increased self-consciousness. Central to the formulations and speculations of this critique is evaluating the effects of a technology implicitly invoked by all four, that of audience. If one considers a sense of audience a technology (with explanation and performance as kinds of knowledge application), then the technology of audience provides analytical perspectives that would not have been obtained had I been writing with only one audience in mind. Those different viewpoints have in turn deepened my understanding of just how profound are the shifts under way in scholarly applications of digital resources to the study of arts and humanities. That I and other enthusiasts have underestimated the degree of that profundity is remarkable. Finally, this chapter makes explicit the importance of standards, as well as the importance of querying the prides and prejudices of those standards and the implications for interpretation made possible in this new editorial work, invoking as a representative touchstone the new Guidelines for Electronic Editions just approved for use by the Modern Language Association (MLA)'s Committee on Scholarly Editions.

The fruits of several evolutions converged to produce a mode of editing suited to the new media: post-structuralist re-inflections of theories of editing; the proliferation of affordable, portable personal computers; the networking, on an unprecedented scale, of homes, individuals, institutions of all sorts (educational, governmental, commercial, religious, and medical), and nations. A fantasy unimaginable as part of the real world just a very short time ago, such networking and the accelerated communication it brings are now facts of everyday life throughout much of the world. Concomitantly, academics have moved beyond worried resistance to the work made possible by new media and computer technologies, and worked through much of the reactionary skepticism so widely and forcefully articulated in the 1990s, in books like Sven Birkerts's The Gutenberg Elegies (1994). By contrast, at the time of writing, under-informed skepticism has been replaced by the realization that critical engagements with new technologies are the best hope for advancing knowledge production in the humanities. That these advances are within our grasp is abundantly clear in the field of scholarly editing.

For example, one happy consequence of new media revolutions is that breathtaking advances in networking and communication have enabled editorial praxes previously thought too impractical to enact, whatever their intellectual value. Instead of being bound to representative samples, editions can now include images of all primary documents included in an edition. The surrogates possible in digital humanities are more than just singular photographic examples meant to represent hundreds or even thousands of instances. Thus editorial claims to being "comprehensive" (a staple of the print variorum relying on representative examples) accrue whole new meanings in the extraordinarily capacious digital realm. Ironically, though, electronic textual editors would not be likely to make such claims, realizing just how vast is the information one could choose to include in a "comprehensive" edition. Those same editors would nevertheless demand inclusion of images of all primary documents (manuscripts or corrected proofs, for example) rather than only the analytical descriptions thereof one finds in the editorial apparatus of print translations (iterations mechanically reproduced for public distribution and annotated for scholarly consumption). Also, editions can now implement new kinds of scholarly annotation by including, where applicable, sound and even video reproductions simply impossible to contain as a constitutive part of a book (though a CD might be slipped inside the cover as a companion or a chip ingeniously packaged and placed for a reader's play).

Networking and communication have also made imaginable and agreeable editorial praxes traditionally thought impracticable or undesirable. These practices modify the very nature of scholarly editing itself. For example, if an editor chooses, the much theorized common reader can now provide immediate and actual (in contrast to imagined) critical feedback that might in turn be usefully incorporated into the work of a scholarly edition, making the collaboration between reader and author that is characteristic of reading itself something more than the "creative regeneration" [of texts] by readers invested with and in private intentions such as those D. F. McKenzie describes in Bibliography and the Sociology of Texts (1999). All of these factors will be examined in this essay as examples of under-analyzed technologies, but first some observations about editing and the phrase that provides my title are in order.

Distinguished editing of literary and artistic work is a matter of the heart (emotions), and at least four of our five senses (physical perceptions), as well as of the head (logic). Astute editors take advantage of the acumen offered by these diverse ways of apprehending the world, and though it is in one sense bound by algorithms, digital editing is no exception in terms of its characteristics (of the heart, the physical senses, the head). Reflecting on the individual terms embodied by the phrase, and the traditions they bespeak, deepens and broadens understandings of "electronic scholarly editing", both in its present forms and in the theories about those praxes and the possibilities that inhere in multimedia. Editing makes works (poems, plays, fiction, film footage, musical performances, and artistic and documentary material) publishable (in books, films, television and radio, and recordings) by eliminating unwanted material and organizing what remains for optimal and intelligible presentation to audiences. In other words, editing translates raw creative work into an authoritative (not to be confused with definitive or authoritarian) form. Scholarly editing is editing performed under the aegis of research, learning, sustained instruction, mastery, knowledge building, standard setting. Electronic scholarly editing consciously incorporates phenomena associated with the movement and manipulation of electrons, those indivisible charges of negative electricity, through wires and radio waves onto screens and through speakers. Though it incorporates the principles of conventional scholarly editing (that is, in and for books) into its own methods, electronic scholarly editing is not performed to be read through the machine of the book but via the hardware of computers. Other nomenclatures for this field include "digital scholarly editing", exploiting to the fullest the obvious pun that enables clear contrastive play between the manual and the automated, and those pertaining to editing itself, especially in the world of ubiquitous computers and the connective possibilities of the World Wide Web. Electronic scholarly editing now inevitably involves networking, both among the texts being worked through and the workers performing the editing (including readers viewing the edited works), in ways never realized by that magnificent tool of knowledge transmission that has served so splendidly well for the past several centuries, the book.

Previously I have examined the implications for knowledge production of the fact that in the digital environment, access – both to numbers of study objects and to numbers of audience members – is facilitated on an unprecedented scale. I considered at length the possibilities for literary criticism and theory when the staggering data storage capacities of computers enable so much more visibility of what I refer to as BE-O objects – artifacts that have customarily been viewed By Experts Only. Agreeing that the skepticism with which many mainstream scholars regard "quantitative research methods" was not really misplaced "so long as the computer's primary role in the humanities was, ostensibly, to compute", Matthew Kirschenbaum proclaims that access to a digital surrogate of a "Rossetti painting" or "of one of Emily Dickinson's turbulent manuscripts" erases or at least reverses that skepticism. The visibilities that enable the unprecedented access to images that were previously locked away in library and museum archives for exclusive view by a very few is probably the technological boon with which scholars and readers are most familiar, for that perpetually available and repeatable access is really quite a big deal. The ability to provide artifacts for direct examination (rather than relying on scholarly hearsay) has altered the reception of humanities computing in the disciplines of the humanities so that skepticism is "at least replaced with more to-the-point questions about image acquisition and editorial fidelity, not to mention scholarly and pedagogical potential." These are indeed "questions of representation, and they are eminently relevant to the work of the humanities" (Kirschenbaum 2002: 4). As Ray Siemens's introduction to A New Computer-assisted Literary Criticism? makes plain, enthusiasts of text analysis would take issue with Kirschenbaum's assertion about quantitative research methods, but the point about the power of visibility, of making accessible for as many pairs of eyes as possible, images (rather than their analytical textual descriptions) for critical analysis, cannot be overstated. When editors make as much about a text visible to as wide an audience as possible, rather than silencing opposing views or establishing one definitive text over all others, intellectual connections are more likely to be found than lost.

Though it will not necessarily do so, that access to primary digital artifacts can in turn create extraordinary access to the editorial process itself, especially if editors and readers alike are proactive in taking advantage of these opportunities by building "sound infrastructure, and the organizational and financial structures … essential to consolidate, to preserve, and to maintain" the "new organizational forms" created by electronic media (D'Arms 2000). The standards described by the MLA's Guidelines for Scholarly Editing, print and electronic, are commonsensical and bound to ensure quality: they ask for "explicitness and consistency with respect to methods, accuracy with respect to texts, adequacy and appropriateness with respect to documenting editorial principles and practice." However practical and necessary for meeting principled goals and however derived by scholarly consensus, uses of various media constrain and enable adherence to these benchmarks to different degrees. Indeed, one clear example lies in the fact that explicitness, consistency, and accuracy are all finally matters of faith in the bibliographic realm.

Considering an actual instance will make clear the importance of enhanced visibility and the more critical reading it enables. My examples will be drawn from the editorial resources with which I have been most involved, both as a user and a maker, those of the writings of the nineteenth-century American poet Emily Dickinson. Because the practices involved in their making are so widespread in the production of scholarly editions, and similar principles to those employed in producing Dickinson are so widely accepted for making accessible the texts of other authors, these specific instances serve well as exemplifications of general rules and procedures for scholarly textual production. R. W. Franklin spent scrupulous decades producing a three-volume variorum edition of The Poems of Emily Dickinson for Harvard University Press, a publisher highly esteemed for producing study objects meeting the highest standards. Franklin's stated goal is to give a "comprehensive account" of all of Dickinson's texts, as well as any texts directly bearing on her compositional practices (1998: 28). In a comprehensive account, one would expect to see all omissions, additions, alterations, emendations of which the editor is aware marked in some way for his readers. Thus any ellipsis, any omission of data, would be signaled by familiar punctuation (…) and probably explained in editorial notes. Yet this is not the case, and readers cannot readily know when an omission has occurred or a change silently made because their access to the objects edited is limited and the evidence of explicitness, consistency, and accuracy cannot be seen. Instead, readers have to trust that the printed texts they see are explicit, consistent, and accurate about all textual data.

Thus when Franklin erases 74 words from his representation of an epistolary manuscript involving the composition of Dickinson's poem "Safe in their Alabaster Chambers" and does so without commentary and without signifying punctuation, readers of his print variorum have no way of knowing that the omission has been made, much less of evaluating its import (FP 124C). Besides being misinformed about the nature of this text because of omission of information, access to the editorial process itself is limited, the implicit contract with the reader (that such treatment of documentary evidence would be signaled in some readable way) having been violated. Franklin's change might, one could argue, simply be chalked up to bad editorial work. Yet he silently enacts other emendations as a course of standard, widely accepted editorial practice. Much ink has been spilled by Dickinson and American poetry scholars about her use of punctuation marks that most have labeled "the dash." Strong positions have been taken about whether the shorter en- or the longer em-mark most faithfully translates Dickinson's mark into print, and Franklin has resolved the matter with the authoritarian stance that neither the en- nor the em-suffice: according to him, the shorter-than-either hyphen best conveys Dickinson's practice of using a horizontal or angled mark rather than a comma. The problem with staking out a hard and fast position on this matter (as many a Dickinson critic has done) is the "one size fits all" perspective practically required by print production. Examining almost any poem written during Dickinson's middle years shows that within individual poems, indeed within individual lines, she did not consign herself to a single sort of mark but used marks of varying lengths and angles (see, for example, "The name -of it -is 'Autumn'___" (FP 465; facsimile available in Franklin's Manuscript Books of Emily Dickinson, p. 494). Editing nearly 1,800 poems, Franklin makes the decision for readers that a hyphen will suffice to represent Dickinson's diverse, varied marks, and thus also decides for readers that the differences among her punctuating inscriptions are not poetically significant.

Trust in any print edition, then, whether it be that made by Franklin or any other editor, including myself, is necessarily faith-based, for readers cannot adequately see the documentary evidence that determines everything from genre to suitability for inclusion in a scholarly edition. Indeed, such matters of representation are more than eminently relevant to humanities; they are central to humanities knowledge production itself. Opportunities for analysis of who made the scholarly objects we study and for what purposes, and of which parts of those objects are constitutive and worthy of study, are proportionate to the visibility of foundational documentary evidence enabled by the medium of representation.

In contrast to the constrained visibilities of book representations, access to questions of editorial fidelity and therefore to the editorial process itself is much more obtainable in an electronic edition featuring images of all documents edited as well as their translations into typography. In such a realm, 74 words cannot simply be excised without commentary and marks translated with a "one size fits all" authoritarian stance and go unnoticed, because a digital image of the document is available for readers to view, assess, compare with editors' representations. Representation via digital facsimiles of original documents changes access to the foundational materials of scholarly editions, the contours of expertise, and even standard-setting itself.

Consciously foregrounding the ontological differences between electronic and bibliographic scholarly editions is necessary for understanding ways in which this access can work, for learning how electronic editions made might best be used, and for planning and developing electronic editions that are more and more well made (in that they are designed for optimum usability as well as optimum incorporation of relevant but scattered data and commentary). Introducing Harvard University Press's latest variorum edition of The Poems of Emily Dickinson, Franklin claims that although that three-volume edition "is a printed codex", it is practically electronic, for the edition "has an electronic database" and the "poems are in bits and bytes." He then flatly states that "other outputs are possible, including other printed editions, organized or presented differently. Dickinson and Hypertext may well be matched, and images are particularly useful with an unpublished poet who left her poems unprepared for others" (1998: 27–8). Yet electronic editions are by their very constitution markedly different, both from the print edition Franklin made and the hypertext one he imagines. For one thing (which he acknowledges), the bits and bytes to which he refers do not contain images and thus do not provide multifaceted viewpoints, the visibility of digital surrogates of original manuscripts presented in an environment where the distance that separates the libraries in which they are held presents no obstacle for direct comparison of texts within each of those collections spread across oceans and nations, separated by geography, national boundaries, and disparate laws. Collected into an electronic edition, Blake or Dickinson manuscripts logically related to one another but dispersed across various collections are gathered together so they can be studied and synthesized to make new knowledge; collected into the same electronic edition, they no longer lie like scattered pieces of a puzzle, their connections splintered, difficult to see as possibly yoked, and thus next to impossible to imagine. Another difference unacknowledged by Franklin is that the typographical bits and bytes to which he refers are those of word processing, which only encode the textual data found in Dickinson's works. A scholarly electronic edition would likely be prepared in XML or SGML (probably following the guidelines of the Text Encoding Initiative), and would capture logical, structural, and artifactual aspects of the original, as well as its textual content. An electronic edition also might include databases, Flash, video, and sound presentations. Different editions are not simply "outputs" of the same bits and bytes, which Franklin's declaration assumes. The statement makes plain that he envisions electronic editions as typographically edited texts with illustrations that can be accessed with the click of a mouse. But electronic scholarly editions provide much greater opportunities than texts with links to illustrations and to one another, enabling a range of manipulation not possible for those bound to the printed page.

Far more than display, replications constitute the textual reproductions and representations in scholarly editions such as The Blake Archive, the Rossetti Archive, The Walt Whitman Archive, The Canterbury Tales Project, the Dickinson Electronic Archives. Scholarly electronic editions are not simply a matter of bits and bytes and how they have been encoded to link to one another and to appear in particular browsers. Dynamic rather than static, deeply encoded, they are designed to enable readers to achieve better understanding of texts and open up possibilities for more sophisticated interpretations. Susan Schreibman and others have chronicled the fact that by the mid-1980s it had become clear to numbers of scholars already involved in the creation of electronic resources for the humanities that a standard for encoding texts needed to be developed that was not dependent upon a particular platform or software package (Schreibman 2002a). Achieving platform-independence requires separating the information content of documents from their formatting, which is what led an international group of those concerned with publishing and preserving government documents to develop GML (Generalized Markup Language) in the 1970s, and then SGML (Standard Generalized Markup Language), adopted as an international standard in the mid-1980s. Working with SGML, a group of humanities scholars and librarians launched the Text Encoding Initiative (TEI) in 1987.

Though markup has a performative significance, the TEI is designed to represent already existing literary texts. For electronic scholarly editors, "markup is supplied as part of the transcription in electronic form of pre-existing material…. Markup reflects the understanding of the text held by the transcriber; we say that the markup expresses a claim about the text" (see Allen Renear, in chapter 17, this volume, and in writings elsewhere, and many others, for example Susan Hockey, Michael Sperberg-McQueen, Jerome McGann: see entries in the References for Further Reading). John Unsworth's claim that "any attempt to systematically express one's understanding of a text in the kind of internally consistent, explicit and unambiguous encoding that is required in the creation of a computable SGML or XML edition will produce some intellectual benefit and will ensure some degree of integrity", is indisputable (Unsworth, website). Indeed, the technology of self-consciousness required by computer encoding of texts, especially that such as TEI-conformant XML encoding, produces a healthy self-consciousness about what Bruno Latour and Steve Woolgar describe in Laboratory Life as "black-boxing" – which occurs when one "renders items of knowledge distinct from the circumstances of their creation" (1986: 259n). In black-boxing, critical opinion becomes "fact"; more often than not, amnesia sets in after that factual instantiation, and having been effectively black-boxed, "fact" becomes "truth." As in the case of Dickinson discussed here, wittingly or unwittingly bibliographic editors sometimes remove elements that are (or well might be) constitutive of its poetics when transmitting a text from the messy circumstance of its creation to the ordered pages of print. The challenge for electronic scholarly editors is not to perpetrate similar distortions when enthusiastically embracing the ordering tools of encoding to convey literary expression, which by its very nature often perplexes the orders of expository logic.

Hypertext theorists and practitioners, who have been (sometimes roundly) critiqued for being too exuberant, are sometimes echoed by the encoding specialist's enthusiasm for important developments in XML: "One of the most powerful features of the XML family of technologies is Extensible Stylesheet Language Transformations (XSLT), a language that facilitates transformation of text rather than display (as HTML's Cascading Style Sheet language does)…. The 'infinitely recenterable system whose provisional point of focus depends on the reader' (Landow 1997, p. 36) is made realizable through XML-XSLT technology, and is much closer to the democratized space envisioned by early hypertext theorists" (Schreibman 2002b: 85). Without a doubt, though, new technologies for structuring data, coupled with the guidelines provided by the Committee on Scholarly Editions and the TEI's commitment to work out of a "broad, community-based consensus", are crucial for realizing advances in editorial praxes. Also necessary is the commitment to facilitate difference by evolving "guidelines … to accommodate a greater variety of editorial methods and a broader range of materials and periods" (MLA Committee on Scholarly Editions, website), and by the TEI's "extension mechanism, as well as its consistent insistence… on international and interdisciplinary representation in its governing bodies, its workgroups, its funding sources, and its membership" (Unsworth, website). The collaborative development of these standards is absolutely essential for any demotic ethic valued as a worthwhile goal, and the disambiguation forced upon knowledge workers who thrive on ambiguity often proves critically profound, because texts are seen as never before, both on the surface and in their deep structures.

Yet none of these advances in markup, nor any of these guidelines, is robust enough to accommodate all facets of the actual textual experience of editors working with primary artifacts. Original documents, the raw materials with which editors must work, are by their very nature queer, and must be normalized to some degree in order to be put into an edition. In part, this is a result of the fact that "there are certain basic – irresolvable – philosophical tensions in language – most particularly between its capacity to represent knowledge systematically and its capacity to be knowledge experientially and perceptually – and this tension intensifies in an electronic environment as a tension between machine language and natural language – since it is a reductive hybrid version of the one which can be encoded/encrypted in order to serve as the basis of the other" (Drucker 1998: 219). This characteristic of language itself is much more visible in an environment in which

the more sophisticated we are the more we normalize textual incommensurables. We have internalized an immensely complicated, many-leveled set of semiotic rules and signs, and we control the contradictions of actual textual circumstances by various normalizing operations. We can hope to expose these normalizations – which are themselves deformative acts – by opening the conversation… between analogue and digital readers. We begin by implementing what we think we know about the rules of bibliographical codes. The conversation should force us to see – finally, to imagine – what we don't know that we know about texts and textuality. At that point, perhaps, we may begin setting philology – "the knowledge of what is known", as it used to be called – on a new footing.

(McGann 2001: 207)

In a case such as the Dickinson Electronic Archives, the editions are explicitly designed not to define and normalize texts, as has been the objective of most bibliographic scholarly editions, and has been the practice of many electronic scholarly editions. Many electronic editions are still framed by the good work of the book because the dream of shedding the inertia imported from the bibliographical realm is only gradually finding its modes of practice. The TEI was developed to represent already existing literary texts in electronic media, but in its years of development it has been confronted by queerer and queerer texts and by editors whose commitment to integrity means that they cannot simply normalize those texts to fit a predetermined grid. This, in turn, demands that we grasp the real significance of the truism that editing is a kind of encoding and encoding is a kind of editing, and it also requires that we probe the politics of the encoding standards we are embracing. Readers too

rarely think about the myriad of databases, standards, and instruction manuals subtending our reading lamps, much less about the politics of the electric grid that they tap into. And so on, as many layers of technology accrue and expand over space and time. Systems of classification (and of standardization) form a juncture of social organization, moral order, and layers of technical integration. Each subsystem inherits, increasingly as it scales up, the inertia of the installed bases of systems that have come before.

(Bowker and Star 1999: 33)

Besides asking who made our objects of study, generative questions about standard setting, who made them, and for what purposes should be posed, and in ways, as McGann argues, not previously imagined. "There is more at stake – epistemologically, politically, and ethically – in the day-to-day work of building classification systems and producing and maintaining standards than in abstract arguments about representation" (Bowker and Star 1999: 10). Standards are of course crucial for realizing reliability, sustainability (both in terms of being intellectually substantive and in terms of residing in a preservable medium), and interoperability among different works and even different systems. Those editing for new media are carrying on this responsibility by working with and helping to develop the new international standards and guidelines. In this ongoing work, we must self-consciously pose questions about the consequences of standardizing, classifying, and categorizing. In previous scholarly editions, classifications or the qualifying criteria dictating how an entity would be classified have been invisible. One benefit of encoding is "to understand the role of invisibility in the work that classification does in ordering human interaction" and thus to keep an eye on the "moral and ethical agenda in our querying of these systems. Each standard and each category valorizes some point of view and silences another. This is not inherently a bad thing – indeed it is inescapable" (Bowker and Star 1999: 5). Standards and categories become problematic when they are insufficiently critiqued, and when "a folk theory of categorization itself" prevails. That folk theory "says that things come in well-defined kinds, that the kinds are characterized by shared properties, and that there is one right taxonomy of the kinds" (Lakoff 1987:121). Both the TEI Consortium and the MLA's Committee on Scholarly Editions have said that principled accommodation is necessary, and that's as it must be, if electronic scholarly editing is to realize its potential by pre-empting this compulsion toward standard setting as a kind of intellectual police force. The community has established standards that are flexible and adjustable, adopting what in the builder's trade is called a lesbian rule – a mason's rule of lead, which bends to fit the curves of a molding; hence, figuratively, lesbian rules are pliant and accommodating principles for judgment (OED). Commitment to such principles for judgment must be vigilantly maintained, with editors relentlessly asking: "What work are the classifications and standards doing? Who does that work? What happens to the cases that do not fit the classification scheme? What happens to cases in which several classification schemes apply?" After all, automatons do not make literary texts: people do.

Understanding the poetics and principles of electronic scholarly editing means understanding that the primary goal of this activity is not to dictate what can be seen but rather to open up ways of seeing. The disambiguating codes are tools to understand texts, not to define them. Though consensus may develop around certain elements of what is seen, no consensus need be developed on how to read those elements. In fact, the different perspectives offered via the encoding, image-based digital surrogates, and direct critiques from "common" readers that are possible in electronic scholarly editing create a climate of possibility for interpretation that is precluded by the invisible controls placed on literary works and their interpretation in bibliographic editing. The editorial goal of electronic scholarly editing is not to find the proper literary container for "poems" (or whatever genre is being produced) but to find the medium that transmits more, rather than fewer, of the techniques inscribed and found on the literary page. Propriety (which inheres in the notion of the proper container) is not the issue, hence the TEI's extension mechanism. The issue is textual pleasure and access to greater understanding, and enabling audiences to avail themselves of as many aspects as possible of the scriptures under study.

So in scholarly editing, the challenge of using these electronic tools that create so many advantages for storage of data (including sound and images), retrieval, and searching is to develop them so that editorial praxes themselves are truly advanced and the hieratic ethic of editing for books is not simply imported into a new, more proficient medium (as was originally the case with the TEI). Derrida's deconstruction of the meanings of "archive" (Derrida 1996) and its connotations of "commandment" and "commencement" help clarify distinctions between hieratic and demotic and show why the latter is an advance when it comes to editorial approaches. By tracing his emphasis on a hierarchy of genres, one can see that commandment – stressing authority, social order – describes the guiding principle of Franklin's variorum. Thus in spite of his emphasis on versioning in his iteration of 1,789 Emily Dickinson poems, order is imposed on that which is otherwise unruly – the messy handwritten artifacts of poems, letters, letter-poems, scraps, notes, fragments. The idea of "poem" disciplines and contains views and the version represented of Dickinson's writings so that they conform to social order and literary law, whatever the material evidence may suggest. According to this principle of commandment, the material evidence, the manuscripts, contain the idea of "poem" and an editor's job is to deliver that idea in a container that makes "poem" extractable. Here textual boundaries are clear, commanded as they are by the ideas that demarcate genres for books. The temptation in disambiguating markup is to repeat that strategy of containment, for it is much easier simply to claim that a text is prose or poetry rather than to acknowledge that it is much more complicated than that, a blend of the two, and requires innovative extensions of the predetermined TEI markup schemes.

By contrast, commencement – physical, historical, ontological beginning – describes the guiding principle of the Dickinson Electronic Archives' production, including our markup. Unpersuaded that "poem" is an "idea" easily separable from its artifact, the editors of the electronic archives feature images of Dickinson's manuscript bodies in their multiple sizes and shapes, in all their messiness. Though our markup designates verse, letter, verse-letter, letter with embedded verse, and letter with enclosed verse, what constitutes a "poem" and poetic meanings is left up to the reader (http://jefferson.village.virginia.edu/dickinson/nosearch/documentation/). A work might sport a stamp or a cutout from Dickens placed by Dickinson into the literary scene that the reader deems part of a "poem", and the editors of the electronic archives refuse to bind those elements as extra-literary but put them on display and include them in the claims our markup makes about the text. Finger smudges, pinholes, paste marks, coffee stains, and traces of ribbons, flowers, or other attachments offer a view into the manuscript circulation and exchange so central to Dickinson's literary world, and we likewise put them on display and include questions about such signs in our editorial submission forms so that markup and notes recognize these materialities as part and parcel of literary production and circulation. Also featured are images of the printed pages, the bodies that have transmitted Dickinson's writings to the world, and in these are stories of Dickinson's history as a poet whose writings are read and enjoyed by a wide audience. The tidy organizations of those printings bound into The Poems of Emily Dickinson and The Letters of Emily Dickinson juxtaposed with the not fully intelligible bindings of manuscripts (into the manuscript books found in her room or her correspondences to 99 recipients) by Dickinson herself, as well as with her many writings unbound (single sheets, notes, drafts, fragments, scraps), renew ontological questions about the identities of these many writings. Textual boundaries are not clear, hence on our submission form for co-editors, the word "genre" in the question asking them to select one for encoders is in quotation marks to underscore the fact that such denotation is forced by the encoding scheme. Underscored both in the markup and the display of original documents is the fact that though an ideal idea of "poem" or "letter" must dominate for the writings to be neatly divided by bibliograph-ically determined genre, and though some denotation must be made for the meta-information of markup, electronic editing enables a much more flexible approach to understanding Dickinson's writings and her manipulations of genre. With many of the documents, readers cannot help but begin to ask "what is this?" "What is this writer doing?" The electronic archives are designed to enable much more extensive considerations of such questions, to confront the philosophical tensions in language rather than smoothing them over and masking them via editorial "clean-up."

Thus if editors self-consciously work beyond the inertia inherited from bibliographic editing, the editorial environment made possible by these technologies of publication can have profound implications for the interpretation of texts and for knowledge production. Though editing might pretend to be objective, it is always someone enacting his or her critical prerogatives. Presenting artistic and critical works can never be done without some bias being conveyed. Though the fact that editing is always interpretation was realized long before their incorporation into humanities work, the new media, and new and newly recognized technologies facilitate accommodation of this fluidity of authorial and editorial/authorial intentions. Editorial theorists as diverse as Tanselle, McGann, Bornstein, Bryant, Shillingsburg, McKenzie, and Grigely in one way or another acknowledge the philosophical tensions in language in which these fluidities inhere and point out that "writing is fundamentally an arbitrary hence unstable hence variable approximation of thought" (Bryant 2002: 1). In the context of these recognitions, all of these values are profoundly affected: authority, literariness, authenticity, sociology, access, reproductivity, original/originary texts/moments, editorial responsibilities, authorship, intention. Electronic editions are most valuable in their capaciousness, not only in their ability to offer different versions but also in their ability to facilitate varying views and valuations, both of what is seen and of how it is comprehended, as text or as extra-textual.

Though major editorial projects have often involved collaborators, the collaborations enabled by electronic editions can be something different than previously seen in humanities research, as are the editions themselves. Because websites can be mounted by individuals and because most have not been vetted, peer review has been a contentious issue in the first generation of digital humanities production. Publishers, libraries, and other institutions invested in quality control routinely vet their digital publications in ways akin to those used for print publications: two or three readers' reports by experts in the field are presented to a board who then votes to accept or reject. This is important and valuable, but critical review need not stop there. In electronic publishing, tools such as dynamic databases enable a more sustained and collaborative critical engagement of reader with editor than has previously been seen. The Dickinson Electronic Archives has, for example, opened a critical review space <http://emilydickinson.org/review/deareview.php>, making user feedback an integral part of the editorial process. In this virtual space users have the opportunity to post commentary, reviews, or critical analysis of a particular section(s) or of the site as a whole. Because comments are stored (either anonymously or with signature) in a searchable database, users have the option of reading and responding to previous commentary. Such a space enables interactions with the editors and with readers and provides a generative, previously unavailable, context for critical responses. The dynamic interplay of the audience, the original writer who inscribes the marks, and the editors communicating those marks to posterity is thereby more likely to open what Dickinson herself would call "doors and windows of possibility." In turn, these myriad perspectives can enable a much more sustained reflection on how our critical findings are produced and then transmitted, on the mechanisms of authorization and critical review, and on the criteria for authenticity. These assessments can then begin to penetrate critical mystiques in ways likely to expand rather than restrict knowledge, and to focus attention more on the knowledge itself than on the individual responsible for bringing it to the fore. Access to such knowledge can in turn foster a variety of new co-editorial collaborations among authors, editors, and readers. Digital surrogates alone make definitive analytical descriptions, on which readers of scholarly editions have depended, neither possible nor desirable. Instead, such analytical descriptions accompany images that readers can examine, make judgments about, and then use to assess the editorial commentary. The Dickinson Electronic Archives is complementing the Open Critical Review database by importing and making available to users the Virtual Lightbox <http://mith2.umd.edu/products/lightbox/>, a software tool via which images can be compared and evaluated online, and the Versioning Machine <http://mith2.urnd.edu/products/ver-mach/index.html>, a software tool designed by a team of programmers, designers, and literary scholars for displaying and comparing multiple versions of texts. Both products are open source and have been produced and maintained at Maryland Institute for Technology in the Humanities (MITH) <http://www.mith.umd.edu/>. The database and the sophisticated online tools for comparing images and texts make it possible for users not only to examine the editorial process, but even to become part of that process. Such collaborations are ways of turning the editorial process inside out in order to advance its integrity, its scope, and its reach, as is the choreography of guest co-editors working with a group of general editors.

On multivolume editions of a major author's work, guest co-editors have worked under a rubric framed by (a) presiding editor(s), but electronic editions provide new opportunities for more extensive knowledge transfer and diverse yet coherent knowledge production in electronic scholarly editions. As do their print predecessors, electronic co-editors act as second and third readers, the first readers being the writer at work on creating text to be disseminated and the general editor conceiving of an edition appropriately designed for the electronic field. Since not simply the originating writer but also editors are creative agents, coordinating groups of qualified editors to work on a vast body of work rather than relying on a single master editor or master pair or trio of editors to bring a scholarly edition into being practically insures, as do the disambiguating codes of TEI-conformant XML, greater integrity in textual production.

As we have seen, new materialities of editing, the fact that multimedia representations make it possible for readers to examine documentary evidence previously hidden from their view, that dynamic databases make it possible for critical readers to record and store their feedback in a manageable format, and that the electronic work produced by editors has been instantiated in forms much more manipulable than the rigid freeze of print make innovations in the work of editing itself possible. Teams of editors, rather than a solitary master with her assistants, can work on projects as never before possible. However, multiply edited electronic editions run the risk of inducing a kind of interpretative vertigo in the perspective of readers. After all, though the objective is no longer to produce the definitive, immutable text, electronic editions do aim to produce greater understandings. As readers' critical responses are managed through an online database, so editors' work can be managed via a database-driven editorial submission form such as that provided by the Dickinson Electronic Archives <http://jefferson.village.virginia.edu/dick-inson/nosearch/admin/submissions.php>, with its guidelines for encoding, statement of editorial principles, and documentation with links to regularization databases, lists of hands identified on the Dickinson manuscripts, lists of manuscript repositories, and so forth.

Just as any editor has to take into account authorial intentions, and any reader is wise to consider both the writer's and subsequent editors' authorial intentions, multiple editors all working on parts of the same textual body must take into account the intentions of their co-editors and collaborators, those readers willing to commit time and energy and to abide by the principles established for scholarly textual production. Such forced engagements with the sociologies of intention that frame and inhere in any and all textual productions are immensely valuable for creating editorial environments that are not only more trustworthy but that are bound to advance critical understandings, teaching scholars to ask questions heretofore unimagined. As I noted in an earlier essay, Lucy Suchman makes insightful observations about the conditions necessary for optimizing knowledge production. Instead of viewing the "objective knowledge" proffered by a critical editon "as a single, asituated, master perspective that bases its claims to objectivity in the closure of controversy", "objective knowledge" in the production of a dynamic critical edition online can more easily be seen as "multiple, located, partial perspectives that find their objective character through ongoing processes of debate." Since critical vision is parallactic rather than unidimensional, the processes of comparing and evaluating those different angles of seeing as one compares and evaluates different images or different perspectives of the same images is essential in order to see more clearly and accurately. The locus of objectivity is not "an established body of knowledge… produced or owned by anyone", but "knowledges in dynamic production, reproduction and transformation, for which we are all responsible." By contrast, the hieratic models of the master editorial perspective do not acknowledge how "layered and intertwined" are the "relations of human practice and technical artifact" and how such individualistically driven productions can tend to obstruct rather than facilitate intellectual connections, treating editorial and critical works as "finished… achievements" rather than as ongoing research activities and part of a "process of accretion" of editorial technique and knowledge, part of midrash, as it were (Suchman, website).

Lessons learned by the makers and users of electronic scholarly editions can help answer the call of the late John D'Arms and others to "move the scholarly monograph into the second generation of the digital age", as the electronic scholarly edition has moved into the first generation. As President of the MLA, Stephen Greenblatt issued a letter to the membership about the plight of scholarly publishing in the humanities and a call to action to address the "systemic, structural, and at base economic" crises confronting twenty-first-century humanities scholars committed to the free flow of information that enables knowledge building ("Call for Action on Problems in Scholarly Book Publishing" <http://www.mla.org/>); as President of the Association for Computers and the Humanities and Chair of the TEI Board of Directors, John Unsworth answered Greenblatt's call in presentations such as "The Emergence of Digital Scholarship: New Models for Librarians, Scholars, and Publishers" <http://www.iath.virginia.edu/mu2m/dartmouth.02> and in his work on the MLA Committee on Scholarly Editions and the TEI.

In order to move the scholarly monograph into the digital realm, humanists need to embrace the new technologies developing in digital humanities communities of practice, technologies making not only new work but new ways of working possible, especially those that will not become obsolete (access, multimedia study objects, collaboration, self-consciousness, audience).

Thus digital humanities needs to move from producing primarily the scholarly archive (first generation of electronic literary projects) to also producing digital monographs or multiply (and to varying degrees) authored works – polygraphs. There is no reason that tightly focused critical inquiry cannot be produced by a group. Revealing and practical questions about reproductive fidelity – such as "what degree of imaging accuracy is needed for critical inquiry?" and "who should be responsible for maintaining and making available the highest quality copies?" – are crucial for such editions and have expanded group efforts. Initiatives such as the LEADERS project, which "aims to enhance remote user access to archives by providing the means to present archival source materials within their context", ensure that archivists best trained to maintain both the actual and the surrogate archival copies assume the responsibilities for doing so and work with scholarly editors best trained to organize the data critically. That they can work remotely, as do multiple guest co-editors discussed earlier in this chapter, can be imagined, and perhaps achieved, only in the digital realm of information organization and exchange. Similarly, to develop digital monographs the most basic question – "what lines of critical argument are possible only in a digital monograph?" – needs to be posed repeatedly. Certainly, there are adept ways of using sound and image, as well as linguistic codes, that are only possible in the digital realm, but that is only the beginning. Moving from editor (author surrogate) and author to reader (including editor and author), from enacting definitude (editing for the static printed page) to enacting fluidity (the dynamic screen), is enabling profound innovations in editorial praxes, changes demonstrating how vital are "recent moves to reframe objectivity from the epistemic stance necessary to achieve a definitive body of knowledge, to a contingent accomplishment of dynamic processes of knowing and acting" for enriching our intellectual commons (Suchman, website). Acknowledging the fluidity of texts instead of insisting upon single-minded, singularly-oriented texts, "learning the meaning of the revision of texts", as well as the revision of our editorial practices, creates an environment in which a "new kind of critical thinking based on difference, variation, approximation, intention, power, and change" can flourish and work for the common good. If we are to shed the inertia bequeathed from the bibliographic world, editorial integrity and fidelity created within and upon the "shifting sands of democratic life" demand a "new cosmopolitanism" in scholarly editing (Bryant 2002: 177), adopting the "lesbian rule" of principled accommodation for digital humanities.

References for Further Reading

Bernstein, Charles (2000). The Art of Immemorability. In Jerome Rothenberg and Steven Clay (eds.), A Book of the Book: Some Works and Projections about the Book and Writing (pp. 504–17). New York: Granary Books.

Bornstein, George and Theresa Tinkle, (eds.) (1998). The Iconic Page in Manuscript, Print, and Digital Culture. Ann Arbor: University of Michigan Press.

Bowker, Geoffrey C. and Susan Leigh Star (1999). Sorting Things Out: Classification audits Consequences. Cambridge, MA, and London: MIT Press.

Bryant, John (2002). The Fluid Text: A Theory of Revision and Editing for Book and Screen. Ann Arbor: University of Michigan Press.

Condron, Frances, Michael Fraser, and Stuart Sutherland (2001). Digital Resources for the Humanities. Morgantown: West Virginia University Press.

D'Arms, John H. (2000). The Electronic Monograph in the 21st Century. Based on Scholarly Publishing in the 21st Century, presented at the American Historical Association, Chicago, January, 2000. Available at http://www.acls.org/jhd-aha.htm.

Derrida, Jacques (1996). Archive Fever: A Freudian Impression. Chicago and London: University of Chicago Press.

Drucker, Johanna (1998). Figuring the Word: Essays on Books, Writing, and Visual Poetics. New York: Granary Books.

Finneran, Richard J., (ed.) (1996). The Literary Text in the Digital Age. Ann Arbor: University of Michigan Press.

Flanders, Julia (1998). Trusting the Electronic Edition. Computers and the Humanities 31: 301–10.

Franklin, R. W., (ed.) (1998). The Poems of Emily Dickinson: Variorum Edition, 3 vols. Cambridge, MA, and London: Harvard University Press. (References to texts in this edition will use "FP" and refer to the poem's number.).

Grigely, Joseph (1995). Texualterity: Art, Theory, and Textual Criticism. Ann Arbor: University of Michigan Press.

Hamill, Sam. "Open Letter" on "Who We Are", Poets Against the War. Accessed spring 2003. Available at http://poetsagainstthewar.org.

Hockey, Susan (2000). Electronic Texts in the Humanities. Oxford and New York: Oxford University Press.

Hockey, Susan, Allen Renear, and Jerome McGann. What is Text? A Debate on the Philosophical Nature of Text in the Light of Humanities Computing Research. Accessed spring 2003. Available at http://www.humanities.ualberta.ca/Susan_Hockey/achallc99.htm.

Kirschenbaum, Matthew (guest ed.) (2002). Image-based Humanities Computing. Special issue of Computers and the Humanities 36: 1–140; his introductory essay (pp. 3–6).

Lakoff, George (1987). Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago and London: University of Chicago Press.

Landow, George (1997). Hypertext 2.0: The Convergence of Contemporary Critical Theory and Technology. Baltimore: Johns Hopkins University Press.

Latour, Bruno and Steve Woolgar (1986). Laboratory Life: The Construction of Scientific Facts, 2nd edn. Princeton, NJ: Princeton University Press. (Originally published 1979.).

Love, Harold (1993). Scribal Publication in Seventeenth-century England. Oxford: Clarendon Press.

McGann, Jerome (2001). Radiant Textuality: Literature after the World Wide Web. New York: Palgrave.

McKenzie, D. F. (1999). Bibliography and the Sociology of Texts. Cambridge: Cambridge University Press. (1986) London: British Library. (1985) The Panizzi Lectures. (1984) The Sociology of a Text: Oral Culture, Literacy and Print in early New Zealand. The Library 6 (December 1984).

MLA Committee on Scholarly Editions. Revision Plan for the Guidelines for Scholarly Editions of the Committee on Scholarly Editions. Accessed spring 2003. Available at http://www.mla.org.www_mla_org/reports/reports_main.asp?. Guidelines for Electronic Scholarly Editions: Available at http://sunsite.berkeley.edu/MLA/guidelines.html. Guidelines for Editors of Scholarly Editions: Available at http://jefferson.village.virginia.edu/~jmu2m/cse/CSEguidelines.html.

Rockwell, Geoffrey (2002). Gore Galore: Literary Theory and Computer Games. In Siemens (ed.), A New Computer-Assisted Literary Criticism? (pp. 345–58).

Schreibman, Susan (2002a). Computer-mediated Texts and Textuality: Theory and Practice. In R. G. Siemens (ed.), A New Computer-Assisted Literary Criticism? (pp. 283–93).

Schreibman, Susan (2002b). The Text Ported. Literary and Linguistic Computing 17: 77–87.

Siemens, Raymond G. (guest ed.) (2002). A New Computer-assisted Literary Criticism? Special issue of Computers and the Humanities 36: 255–378; his introductory essay (pp. 259–67).

Shillingsburg, Peter L. (1996). Scholarly Editing in the Computer Age: Theory and Practice. Ann Arbor: University of Michigan Press.

Smith, Martha Nell (2002). Computing: What's American Literary Study Got to Do with IT? American Literature 74: 833–57. Online. Project Muse. Accessed spring 2003. Available at http://muse.jhu.edu/journals/american_literature/v074/74.4smith.html.

Sperberg-McQueen, C. M. and L. Burnard, (eds.) (1990). Guidelines for Electronic Text Encoding and Interchange: TEI P3. Chicago and Oxford: The Text Encoding Initiative. (Reprinted 1999.) Available at http://www.tei-c.org/Guidelines/index.htm.

Sperberg-McQueen, C. M., Claus Huitfeldt, and Allen Renear (2000). Meaning and Interpretation of Markup. Markup Languages 2: 215–34.

Suchman, Lucy. Located Accountabilities in Technology Production. Published by the Department of Sociology, Lancaster University. Online. Accessed Spring 2003. Available at http://www.comp.lanes.ac.uk/sociology/soc0391s.html.

Unsworth, John (1996). Electronic Scholarship; or, Scholarly Publishing and the Public. In R. J. Finneran (ed.), The Literary Text in the Digital Age (pp. 233–43). Ann Arbor: University of Michigan Press.

Unsworth, John. Electronic Textual Editing and the TEI. Online. Accessed spring 2003. Available at http://jefferson.village.virginia.edu/~jmu2m/mla-cse.2002.html.

Winder, William (2002). Industrial Text and French Neo-structuralism. In R. G. Siemens (ed.), A New Computer-Assisted Literary Criticism? (pp. 295–306).