Digital Scholarly Editing

Susan Schreibman

Over the past twenty years there has been an evolving body of scholarship exploring
the standards, theories, and methodologies of digital scholarly editing. Scholarship
from the early-to-mid-1990s maintained a bifurcated focus. On the one hand, many
textual scholars found themselves in the slightly unusual position of writing primers,
guidelines, and documentation laying the groundwork for basic digital tasks, such as
Peter Robinsons The Digitization of Primary Textual Sources, published in 1993,
and The Transcription of Primary Textual Sources Using SGML, which followed a
year later, or the largely uncredited scholarship created by the many hands that went
into the guidelines of the Text Encoding Initiative (TEI), first officially published in
1994. Editors and staff members from many scholarly editing projects spent a great
deal of time in documenting encoding practices. This documentation proved to be an
essential resource to ensure consistent encoding and to help those entering the field
understand how a standard like the TEI was put into practice.1

On the other hand, as the decade progressed there was an ever-growing body of
research exploring what this new medium could bring to the study of the
transmission of texts, and thus editorial practice normative to this environment
began to develop. Norman Blake and Peter Robinsons The Canterbury Tales Project:
Occasional Papers I, published in 1993, was one of the earliest, but by the
time The Canterbury Tales Project: Occasional Papers II was published in 1997 other
collections were appearing, such as Richard J. Finnerans The Literary Text in the
Digital Age. Published in 1996, it was one of the first collections of essays devoted to
electronic scholarly editing. Many of its chapters are still required reading, such as C.
M. Sperberg-McQueens Textual Criticism and the Text Encoding Initiative, John
Lavagninos Completeness and Adequacy in Text Encoding, and Robinsons Is
There a Text in These Variants?

The same year, a conference was held in Ann Arbor exploring the semantics of the
page in print, in manuscript, and on screen. The results of that meeting appeared in
the 1998 publication edited by George Bornstein and Theresa Lynn Tinkle
entitled The Iconic Page in Manuscript, Print, and Digital Culture. In that volume,
Kevin Kiernan explores the use of new technologies to restore what is invisible to the
eye in medieval manuscripts (18), and Martha Nell Smith argues that new
developments in technology have the potential to create an entirely new reading
environment for a poet like Emily Dickinson, many aspects of whose work are still
contested. An electronic edition, Smith argued, could be designed so that readers
could explore multiple orders of the poems and choose between contrasting
representations, as well as provide a library of secondary sources within the same
reading space (Corporealizations 215).

In 1997 Kathryn Sutherlands Electronic Text appeared. Like Finnerans collection, it

explored the methods and theories of digital textuality. In Sutherland, Jerome
McGanns The Rationale of Hypertext first appeared, as well as Allen Renears Out
of Praxis: Three (Meta)Theories of Textuality, in which the OHCO (ordered
hierarchy of content objects) theory was explored alongside other theories of textual
representation.

All the authors cited above were also practitioners who created scholarly electronic
editions. These first-generation digital editions did not appear, at least on the
surface, as scholarly or critical in the ways that the textual editing community had
come to understand the terms for print publication. Frequently these early editions
did not include the apparatus, critical notes, and contextual material that, by the late
1990s, had come to signal the apotheosis of scholarly editing. Moreover, claims about
technologys potential, such as those made by Smith, while common in theory, were
difficult and thus rarely implemented in practice (Schreibman 83).

Until the end of the 1990s, there was only one software program, DynaWeb, that
allowed texts encoded in standard generalized markup language (SGML) to be
displayed on the World Wide Web. SGML is the language from which the now more
ubiquitous extensible markup language (XML) and the language of the Internet,
hypertext markup language (HTML), were derived. DynaWeb was a commercial tool
developed for early adopters of SGML (including the pharmaceutical and defense
industries) to search, display, and organize large textual corpora. When the software
was owned by INSO Corporation and Electronic Book Technologies, it was possible
to apply for a free version of the software, and severalincluding Oxford University;
University of California, Berkeley; Duke University; and University of Virginiatook
advantage of this opportunity. It thus became possible for academics to present their
digital scholarship within a database environment that allowed for dynamic searches
and transformations to HTML for delivery.

With the development of XML and associated standards, it became easier to realize
some of the earlier theoretical goals claimed for electronic editions. By 2006, when
the Modern Language Association published Electronic Textual Editing, edited by
Lou Bernard, Katherine OBrien OKeeffe, and John Unsworth, a confidence
reflecting over a decade of practice was evident. John Bryants somewhat earlier The
Fluid Text (2002) and Peter Shillingsburgs From Gutenberg to Google (2006), for
example, explore the tensions between traditional editorial practice and this new
medium. Moreover, the Blackwell Companion to Digital Humanities (2004;
Schreibman, Siemens, and Unsworth) and Companion to Digital Literary
Studies (2008; Siemens and Schreibman) included essays (Smith, Electronic
Scholarly Editing; Price) that were dedicated to electronic scholarly editing and
examined aspects of scholarly editing practice, reflecting its centrality to digital
scholarship.

Theories of Digital Textual Scholarship: New Norms, New Paradigms,

New Modes of Analysis

In his monograph on the history, nature, and practice of textual

studies, Bibliography and the Sociology of Texts, D. F. McKenzie grappled with the
relevance of traditional bibliographic practice in the light of the wider field of literary
studies. If bibliography, he argued, was simply the practice of describing or
enumerating texts, it would have little relevance to contemporary theoretical
concerns:

The problem is, I think, that the moment we are required to explain signs in a book,
as distinct from describing or copying them, they assume a symbolic status. If a
medium in any sense effects a message, then bibliography cannot exclude from its
own proper concerns the relation between form, function, and symbolic meaning. If
textual bibliography were merely iconic, it could produce only facsimiles of different
versions. (10)
McKenzie defines text as verbal, visual, oral, and numeric data, in the form of maps,
prints, and music, of archives of recorded sound, of films, videos, and any computer-
stored information, everything in fact from epigraphy to the latest forms of
discography (13). He goes on to suggest a new definition of bibliography as a
practice that not only recognizes but also seeks to describe how the form of the text
affects meaning. Moreover, it is not simply the material embodiment of the text that
informs meaning but also the social process of transmission, [its] physical forms,
textual versions, technical transmission, institutional control, . . . perceived
meanings, and social effects (13).

First given as a series of lectures at Oxford University in 1985, some seven years
before the advent of the World Wide Web, McKenzies writing seems strikingly
prophetic. Here, McKenzie meticulously develops a vocabulary to describe a not-yet-
invented medium:

In terms of the range of demands now made of it and of the diverse interests of those
who think of themselves as bibliographers, it seems to me that it would now be more
useful to describe bibliography as the study of the sociology of texts. If the principle
which makes it distinct is its concern with texts in some physical form and their
transmission, then I can think of no other phrase which so aptly describes its range.
(13)

In many ways, the great enterprise of the last decade and a half to digitize our
cultural heritage has been an exploration of the sociology of texts. It has fostered new
theories of editing and new modes of editorial practice. Some of these theories share
much with their print counterparts; others exist only in the digital realm. Textual
scholars were among the first to explore the visual, computational, and navigational
possibilities offered by this new medium. By creating a docuverse that flattened all
cultural artifacts to binary code, the materiality of the original object was brought
into sharp relief. Deciding what was essential to re-present provided the distance
that textual scholars needed to reconceive the fields theories, methodologies, and
practices. As McGann, one of the fields most eloquent critics, has written, it was only
when textual scholars had the opportunity of editing in a medium other than the
book that they were able to realize the constraints that the medium imposed on
them:
This symmetry between the tool and its subject forces the scholar to invest analytic
mechanisms that must be displayed and engaged at the primary reading levelfor
example, apparatus structures, descriptive bibliographies, calculi of variants,
shorthand reference forms, and so forth. The critical editions apparatus, for
example, exists only because no single book or manageable set of books can
incorporate for analysis all of the relevant documents. (Radiant Textuality 56)

There are, of course, many shared goals between print editions and first-generation
digital editions: above all, the creation of new works by means of a re-presentation of
the works of the past. Issues of authority, textuality, and representation are of
concern in both mediums: editors must decide to what level of fidelity the linguistic
codes (the linguistic elements of the text and paratext) are maintained and whether
and if so, to what degreethe bibliographic codes (the material aspects of the text:
the typography, advertisements, illustrations, decorations, etc.) are captured. What
editors of first-generation Web-based editions discovered was that a plethora of new
intangibles also preoccupied them, such as whether HTML was expressive enough to
create a digital scholarly edition and, if it was not, what other markup scheme might
be appropriate; how best to encode structural divisions of texts, such as paragraphs,
titles, footnotes, and lines of verse; and how closely a digital surrogate of a print
publication should, or indeed could, capture and make clear to users essential
qualities of the material object.

Many first-generation digital editions explored whether it was more apposite to

realize a projects goals through a unique encoding scheme rather than to use a
standard like DocBook or the Text Encoding Initiative Guidelines. Moreover, issues
that had heretofore been the preserve of publishers and typesetters preoccupied
many literary scholars, such as how to represent characters for nonstandard text
(e.g., accented characters or special symbols) or how to deal with edition-specific
typographic features (e.g., running headers, font, and pagination). Moreover, this
new medium allowed an exploration of textuality beyond the printed word, creating
editions of other cultural objects, such as images (still and moving) and audio. The
digital environment leveled the playing field for multimedia and text, creating a
holistic environment within which to seamlessly navigate between primary objects
and the contextual, between the visual and the aural, engendering a reevaluation of
the social and material ontologies of the text (Loizeaux and Fraistat 5).
Much of this exploration has gone on within a new genre of scholarly production, the
thematic research collection (TRC). One might argue that TRCs became the
framework within which editors conceived, explored, and realized these new
editions. In 2000 Unsworth set out essential characteristics of TRCs. Above all, they
are electronic, contain heterogeneous data types, are extensive but thematically
coherent, and are structured but open-ended. They are designed to support research,
are written by at least one if not many authors, are interdisciplinary, and are
collections of digital primary resources. Carol Palmer further refined the definition,
distinguishing collections created by libraries or other cultural heritage organizations
from those created by scholars:

In taking a thematic approach to aggregating digital research materials, they are

producing circumscribed collections, customized for intensive study and analysis in a
specific research area. In many cases these digital resources serve as a place, much
like a virtual laboratory, where specialized source material, tools, and expertise come
together to aid in the process of scholarly work and the production of new
knowledge. (34849)

TRCs take many forms. Many early digital editions explored the notion of
uneditingthat is, reproducing the text in documentary form, typically in the form of
facsimiles. In print this was a fairly expensive undertaking reserved for the most
canonical of authors. It worked particularly well where there was one copy of a
manuscript, such as T. S. Eliots The Waste Land, in which Eliots typescript was
edited by Ezra Pound, Valerie Eliot, as well as Eliot himself. The more recent
facsimile editions of Windham Lewiss Blast capture the anger and arrogance of the
original typography as no other representation could.

Unediting is, in many ways, fairly trivial in digital form. Digital images are relatively
inexpensive to create and store. Many projects choose to take this route rather than
to transcribe and encode text. But current technology can also make these editions
clunky. It can be difficult to ascertain the level of engagement one needs to make
with a text published as a series of PDFs, a problem not encountered with my print
copy of Blast. Too many projects use unimaginative strategies for browsing (such as
displaying twenty or even one hundred thumbnails to a page) or simply provide a
page-turning strategy ( la Google Books).
Other facsimile editions, such as The William Blake Archive, are works of extreme
editing. The William Blake Archive was one of the first projects of the Institute for
Advanced Studies in the Humanities at the University of Virginia. It explored the
technical, editorial, and legal issues surrounding the creation of an image-based
digital scholarly edition. The goal was to make publicly available Blakes nineteen
illuminated manuscripts. Begun in 1992, it is among the earliest of the TRCs. The
archives adherence to strict digitization standards, to capturing the fidelity of the
original artifacts, and to creating a vocabulary that would allow unprecedented
access to the complexity of the rich visual vocabulary that Blake employed has set the
gold standard for image-based electronic editions.

The more recently published In Transition: Selected Poems by the Baroness Elsa
von Freytag-Loringhoven, edited by Tanya Clement and available through the
University of Maryland Libraries, seamlessly melds the facsimile tradition with
rigorously edited, TEI-encoded text that is surrounded by scholarly apparatus.
Clement argues that the electronic re-presentation of the twelve poems in different
versions charts the texts composition history through a textual performance in an
electronic environment. She calls the presentation of networked text (networked
through material space, reception, and theme) in a networked environment textual
performance theory.

Other TRCs are based on alternative theoretical approaches. George Landows The
Victorian Web was an early electronic edition that used hypertext theory as its
philosophical underpinning. Hypertext theory describes an ideal textuality, in
which content items (text, images, audio, etc.) are linked through multiple paths,
chains, or trails in an open-ended, perpetually unfinished textuality (Landow 3).
This notion of textuality was viewed as an embodiment of poststructuralist theory, in
which readers could navigate between inter- or intra-textual lexias and engage in
multisequential reading (Schreibman 78). The Victorian Web, unlike the other
SGML and XML projects discussed here, is not database-driven but maintained as a
series of static HTML pages interlinked through the ubiquitous HTML <a href> tag.
The Creditspage of The Victorian Web demonstrates just how different the
challenges are for digital scholarly editors than for those who edit for print
publication. This page, with its litany of standards, protocols, and software,
documents the changes in technology, audience expectations, and skills needed to
create and maintain the site.

In the production of this new knowledge space, many literary scholars have found
themselves not only assembling and editing the content but also building the tools
and software that enable this scholarship. For example, The Versioning Machine,
which I developed with colleagues over many years at several institutions, is a
framework for creating electronic scholarly editions of texts that exist in various
versions. It was available to Clement and allowed her to develop her theories of the
text. While The Versioning Machine provides for features typically found in critical
editions, such as annotation and introductory material, it also takes advantage of the
opportunities afforded by electronic publication to allow for the comparison of
diplomatic versions of witnesses and the ability to easily compare an image of the
manuscript with a diplomatic version.2 Like much software created by the digital
humanities community, it is freely available and open-source, so that Clement was
able to download it and make changes to it to represent her theories of
performativity.

Technologies and Standards

Before the visual capabilities of the World Wide Web, most of the work in the literary
studies that used computation was in the service of text analysis. Linguists and
literary scholars used software for the creation of concordances and for text retrieval.
Some of the issues facing these early adopters have relevance todaymethods of
alphabetization, the size and range of context units, and the treatment of ambiguous
symbols (Hockey 49). Computers traditionally read text as a sequence of characters,
and thus when plain (unencoded) ASCII text is processed, many ambiguities can
arise. For example, software may not be able to disambiguate the Roman
numeral I from the personal pronoun I. In the 1980s the artificial intelligence
communitys expert systems and more recently the Semantic Web communitys
linked data have sought to develop powerful algorithms to allow software to
construct meaning through context. But much of this context still relies on structured
text. Also known as markup or encoding, this extra intelligence allows computers to
more effectively locate and process semantic textual units. For example, the now
fairly ubiquitous <p> sign represents a paragraph. By making explicit that a block of
text functions as a paragraph, the computer can be instructed to style it using generic
style sheets or to search for a specific term only within that unit as opposed to
another such as <title>.

Individuals as well as communities of practice developed encoding schemes and the

systems to support them to mark, search, and display features of text that their
disciplinary area deemed important. By the mid-1980s, however, the academic
community realized this cacophony of signs, symbols, and standards did not serve
the individual scholar (who frequently had to invent a system) or the broader
scholarly community (since these schemes were typically incompatible). All too
frequently work was undocumented, making it impossible for other scholars to build
on. In other cases systems were so specific to platforms and standards that when
those systems were no longer usable, neither was the scholarship embedded in them.
Other scholars chose to work in proprietary formats, thinking their work had better
chances of longevity, but too frequently their scholarship was locked in systems that
proved to be nonmigratable when the company went out of business or abandoned
the software.

Thus in November 1987, a group of scholars from the humanities, information

studies, libraries, and computer science came together at Vassar College in
Poughkeepsie, New York, to discuss the idea of creating an open, nonproprietary
standard that would be created by and sustained by the academic community for
which it was developed. At the close of the conference, nine design principles, known
as the Poughkeepsie Principles, were articulated. These became the intellectual
foundation for TEI Guidelines.

These design principles have served the scholarly text-encoding community well,
although in places they show signs of their age. For example, principle 6 established
subcommittees to draft guidelines for text documentation, text representation, text
interpretation and analysis, and metalanguage definition and description. The
implicit understanding was that text was the only media practicable to work with in
1987. However, other principles, such as the first four, are still relevant to the
scholarly editing community:
The guidelines are intended to provide a standard format for data interchange in
humanities research.

The guidelines are also intended to suggest principles for the encoding of texts in the
same format.

The guidelines should

define a recommended syntax for the format,

define a metalanguage for the description of text-encoding schemes,

describe the new format and representative existing schemes both in that
metalanguage and in prose.

The guidelines should propose sets of coding conventions suited for various
applications.

An early theory emanating from the textual editing community was the concept of
texts as an ordered hierarchy of content objects (OHCO). This theory of textuality
was informed by and informed the development of the TEI Guidelines and was
developed by Allen Renear, Elli Mylonas, and David Durand in the late 1980s and
went through several revisions through the early 1990s. According to it, text is
composed of nesting objects, such as chapters, sections, paragraphs, lists, and so
forth. Like a set of Chinese boxes, these content objects fit neatly into one another,
from the smallest (a letter or a word) to the largest (a book or monograph), with a
myriad of other nested units in between (sentences, paragraphs, chapters, sections,
etc).

While this theory was instrumental in the development of the TEI, it never fully
accounted for the problem of overlapping hierarchies. Overlapping hierarchies
breaks the neatly nesting pattern described above. For example, a metaphor in a
poem may cut across two or more lines (marked by the tag <l>). It might seem like a
purely technical issue that a language like XML requires one element to close before
another opens, as in the following:
<l>text text text</l>

<l>text text <metaphor> text text</metaphor></l>

as opposed to the following:

<l>text text <metaphor> text text</l>

<l>text text </metaphor></l>.

The creators of the OHCO theory concede that this may be more than a technical
issue and that it may point to some of the thorniest issues surrounding text encoding
as an intellectual endeavor. Text encoding, like any other area of textual scholarship,
is not theory-free. It is subjective, theoretical, and interpretative. Texts, particularly
literary texts, have competing hierarchies, all of which may have equal claim to being
represented as they express different views of the text. For example, the hierarchy
that SGML and hence TEI most eloquently expresses is what one might term the
editorial or bibliographic; that is, representing the text in terms of sentences,
paragraphs, chapters, front and back matter, and so on. This is not surprising given
SGMLs roots as a language written to publish documentary texts in electronic form.
From this point of view, one might deduce that the documentary view of text can be
read as its onlystructure.

Yet there are many textual features that do not conform to this hierarchy. As
mentioned previously, metaphors may span many lines or stanzas of verse. Narrative
events may span many paragraphs and indeed may overlap. Verse drama contains
dialogue lines (speeches), metrical lines, and sentences. But these sentences and
metrical lines overlap in the case of enjambment or when a character begins talking
and another interrupts (Renear 11921). All these hierarchies have equal claim to
representation.

The TEI has risen to the challenge of accommodating alternative views of the text.
Since its establishment as a consortium in 2000, it has created opportunities
(particularly through its chartering of special-interest groups) for new communities
of practice to propose additional tags, as well as to inform and influence its
intellectual growth. For example, the Manuscript Special Interest Group has
proposed a methodology to create a genetic edition of a text. Within this view of
textuality, the editor not only identifies what is on the page but also attempts to
reconstruct the process by which those linguistic and bibliographic codes came into
being.

Even if the cultural objects of an edition are primarily multimedia (rather than full-
text), searching, browsing, and to some extent display are performed on metadata
attached to the objects as opposed to the objects themselves. Some media-based
projects use the TEI Header to encode bibliographic information for nontextual
objects. Others use standards, such as VRA Core (developed by the Visual Resource
Association for the cultural heritage community), a rich encoding scheme for image-
based material. Like the TEI, VRA Core can be used as an interchange format, can be
integrated with other XML standards through the use of XML name spaces, and can
also be mapped onto a less-expressive standard, such as Dublin Core, which has
become the de facto interchange format for basic metadata.

But the encoding standard used is only one piece of the framework for electronic
scholarly editions. The encoded text or other digital object typically resides in a
database that enables the sophisticated searching and browsing we have come to
expect from online editions. Some editions simply use scriptsa combination of PHP
or JSP with XSLT to transform texts to HTML for Web delivery. Open-source, XML-
aware databases, such as eXist, are frequently used for single-author or themed
editions. An enterprise-level solution can be found
in FedoraCommons (Fedora stands for Flexible Extensible Digital Object
Repository Architecture), an architecture for storing, managing, and accessing
digital objects. FedoraCommons is media-agnostic: its architecture defines a set of
abstractions for expressing these objects, for asserting relationships among digital
objects, and for linking behaviors (i.e., services).FedoraCommons provides a
framework for multiple projects to be housed within one repository, allowing for
greater reusability of code, but even more important is that editions within a
common framework do not exist as digital silos.
Possibly the most important lesson from the past twenty years of digital scholarly
editions is that it is necessary to separate content from display and to present the
objects of our contemplationthe full-text files, the images, the audio, and moving
imagesaccording to well-established standards. There is no doubt that the editions
we create today will be migrated into new platforms and formats in the future. There
are also more opportunities for derivative works to be created by harvesting objects
from a variety of sites into new compilations in which the objects original context is
lost. Knowing that this type of reuse is possible, it is even more important for editors
to ensure that essential information, such as copyright, reuse statements, and
provenance, is attached to every object as opposed to having that information reside
in an About the Project Web page.

Digital editions are typically never complete: there is frequently no point that marks
the end of a project, as with book publication. Rather, digital editions tend to be
open-ended. The relative ease in adding new material as well as in correcting old
content allows for an expansionary editorial model, much like hypertext itself, with
shifting centers and nodes. Editors of digital editions need to be cognizant of systems
and applications that no longer function in newer Web environments and keep
abreast of standards and protocols to decide when and how to migrate resources.
Moreover, in recent years we have been more aware of the fragility of the works we
create: how easy it is, for example, for a server to be turned off, resulting in years of
work disappearing into the ether or in the migration of a custom-made encoding
scheme and Web application into a newer environment being too great an effort, no
matter how valuable the resource.

Digital scholarly editions are not simply the encoded text that can be glimpsed
through a browsers reveal codes or, indeed, that can be hidden from view completely
as the text is transformed from the database to HTML for Web delivery. Digital
scholarly editions are a myriad of standards and transformations, of scripts and
software that are brought into existence when a reader comes to a site and makes a
request. On a server, possibly halfway around the world, a series of commands are
called into play that seemingly instantaneously deliver the results of a query. But the
decisions that enable this Internet alchemy are practical and rooted in the theoretical
and philosophic concerns of three disciplinary areas: computer science, information
studies, and the humanities. This triad of expertise contains within it the
underpinnings of the most successful digital scholarly editions.

The Futures of Digital Scholarly Editing

In the last decade and a half, digital scholarly editing has matured as a field. No
longer does it seem herculean to create an electronic edition, although the barriers to
entry are still high. But as we enter a period in which the born-digital artifact is more
frequently the literary or cultural artifact, and as more derivative works are created
without reference to an analog original, scholarly editing may not solely or typically
be about migrating the analog into the digital or about re-presenting print norms in
digital format. In closing, I will discuss just a few of these new modalities and the
issues they raise for new genres of digital scholarly editions.

Electronic Literature

The Electronic Literature Organization (ELO) defines electronic literature as works

with important literary aspects that take advantage of the capabilities and contexts
provided by a stand-alone or networked computer. The ELO also identifies a
number of forms of practice, including hypertext fiction and poetry, kinetic poetry
presented in Flash or other software, computer-art installations with prominent
literary aspects, interactive fiction, and collaborative writing projects that allow
readers to contribute to the text of a work.

The ELO has created an online anthology of electronic literature. Like more
traditional anthologies, contextual informationsuch as biographical information
and a short prcisis included for each work. Also included is information such as
the software and plug-ins needed to run the work. What would an electronic
scholarly edition of one of these works look like? How would variants be presented?
the works genetic history? What metadata and code are important to capture about
these works when the software and hardware they were created for no longer
function? Will textual editors become forensic scientists examining the palimpsests
of hard drives as assiduously as they now examine watermarks? The ELO has begun
to tackle these questions in their Acid-Free Bits: Recommendations for Long-
Lasting Electronic Literature (Montfort and Wardrip-Fruin) and Born-Again Bits: A
Framework for Migrating Electronic Literature (Liu et al.).

Crowdsourcing and the Social Edition

Arguably, other disciplines have been more creative than literary studies in engaging
the public in large community-based projects. Projects such as Galaxy Zoo, Old
Weather, and Foldit allow anybody with an Internet connection to help solve
questions of contemporary science. The Australian and Finnish national libraries
developed tools to facilitate correction of OCR (optical character recognition) errors
in newspaper conversion projects, while, to date, over one million dishes have been
transcribed for the New York Public Librarys Whats on the
Menu? project. Transcribe Bentham is one of the first projects to engage the public
in contributing to a scholarly edition. By customizing MediaWiki, people with no
experience in editing, editorial theory, or scholarly transcription are provided with a
platform to easily transcribe the letters of Jeremy Bentham adding (unbeknownst to
them) light TEI encoding. Over 2,975 letters were transcribed by volunteers during
the initial seven-month period of the project (Causer, Tonra, and Wallace 126).

We have already seen how successful distributed editorial projects, such as

the Perseus Digital Library or Romantic Circles, can be. Robinson goes further in
his 2010 article Editing without Walls to suggest a new model in which the
traditional idea of the scholarly editor overseeing the production process is replaced
by a distributed workflow in which anybody can participate according to their skills,
interests, and abilities. This type of edition, much like Wikipedia, relies on the
intelligence of the crowd to correct errorsas opposed to the erudition of the editors.
As holding institutions make images available for reuse, as the tools to make editing
easier are developed, and as we engage the imagination of the public, the number of
(scholarly) editions will increase: new models will proliferate, and exigencies, beyond
the economics of the canon, will guide their creation.

Virtual Worlds

Increasingly, the scholarly community is investigating virtual worlds as research and

teaching spaces. Virtual worlds are immersive environments that can model,
annotate, and stage critical encounters with temporal and spatial realities. These
modeled worlds create their own ecosystems that provoke and encourage evolving
thought about the material, aesthetic, and cultures of the real-world events they
simulate.

Spatial reconstructions are common in virtual worlds such as Second Life. One can
visit a model of the Globe Theatre or of Great War battlefields. But how could the
immersive power of a virtual world be harnessed to provide new insights into works
of literature? Just as architectural historians have used the power of virtual-worlds
modeling to test theories of building construction and use, literary scholars could
manipulate the spatial and temporal aspects of the material evidence surrounding a
works genesis and reception to test their assumptions. TRCs often bring together a
wealth of documentary evidence from disparate sources, but these environments, like
the codex, flatten time and space onto a one-dimensional reading surface. Immersive
environments provide a venue to raise issues of how the phenomenology of place and
space can be used to design a new language of scholarly editions, one that has the
ability to model sensorial experience lost because of technological and evidentiary
constraints.

Mass Digitization

The large-scale mass digitization projects currently under way may be shifting some
of the locus of scholarly activity in literary studies from the creation of relatively
modest, finely crafted TRCs to developing methods and models to answer the
question, What do we do with a million books? To be sure, the creation of TRCs, as
well as the tools and software to create and present these collections, are as robust as
ever. But mass-digitization projects, beginning with Google Books in 2004
(then Google Book Search), have created opportunities for literary scholars to engage
with more books than any one person could read in a lifetime. What does this mean
for our field? What services and software are needed to engage with such massive
data sets?

Mass digitization seems, in many ways, the antithesis of scholarly editing. But can we
envision a new type of interactive edition that harnesses the distant reading theories
of Franco Moretti and others to create dynamic variorum or genetic editions? We will
need new editorial structures, ones that do not rely on centralized control but instead
are algorithmically generated to exploit, study, and analyze these corpora. Editors
working in this environment will not only work with but also design new analytic
tools and displays to allow users to engage with an ever-shifting, unbounded textual
field. They may include secondary materials impossible to convey in print form, such
as interactive maps and geographic data sets, moving images, and audio, as well as
other born-digital derivatives (e.g., folksonomies, Twitter feeds, crowdsourcing).
This superset of interlocking multitexts and services will allow deeper and wider
corpora to be generated and studied than previously available or even imaginable,
fostering new forms and theories of textual and bibliographic scholarship.

Digital scholarly editors work in an environment of abundance; they are no longer

bound by the constraints of the codex and the economics of print publication. New
dialectics, such as the mutual pressure of the alphabetical, figural, and aural within a
single representational space (Flanders), have taken the place of material and
economic limitations. Digital editions also frequently embed in them a dialectic
between what we might consider the more traditional and intuitive bases of literary
interpretation and the disambiguating premise of stylometrics, attribution studies,
and other statistical methodologies common to computational and algorithmic
processing (Drucker 687). But it is the very provocations that these encounters
provide that enable new forms of analysis, meaning, and insights. Next-generation
digital scholarly editing may well provide some of the most exciting theories for our
discipline by melding disciplinary concerns and practices of fields such as computer
science, information studies, and human-computer interaction with traditional
theories in literary scholarship.

Notes

1. Early examples of this are my own documentation from The Thomas MacGreevy
Archive, the Victorian Womens Writers Project, and DALF: Digital Archive of
Letters in Flanders, an extension to the TEI Guidelines, to encode modern
correspondence.
2. Diplomatic editions are those in which, as far as possible, the marks on the page
(any of type of textual witness beyond a typeset page with no emendations) are
represented typographically. Emendations, such as additions and deletions, are
represented to mirror the edited page. Thus the goal in diplomatic editing is not to
present the reader with a finished text but to give the reader insight into the revision
process. A witness is one of n number of documents in a texts composition history.
For example, a poem might exist in five states: three manuscript drafts, an editors
proof, and a final published text. Each of these five versions would be considered a
witness, a chain of physical documents that demonstrates the authors evolution of a
single text.

Works Cited

Bernard, Lou, Katherine OBrien OKeeffe, and John Unsworth, eds. Electronic
Textual Editing. New York: MLA, 2006. Print.

Blake, Norman, and Peter Robinson, eds. The Canterbury Tales Project: Occasional
Papers I. Oxford: Office for Humanities Communication, Oxford U Computing
Services, 1993. Print.

. The Canterbury Tales Project: Occasional Papers II. Oxford: Office for
Humanities Communication, Oxford U Computing Services, 1997. Print.

Bornstein, George, and Theresa Lynn Tinkle, eds. The Iconic Page in Manuscript,
Print, and Digital Culture. Ann Arbor: U of Michigan P, 1998. Print.

Bryant, John. The Fluid Text: A Theory of Revision and Editing for Book and
Screen. 2002. Ann Arbor: U of Michigan P, 2005. Print.

Causer, Tim, Justin Tonra, and Valerie Wallace. Transcription Maximized; Expense
Minimized? Crowdsourcing and Editing The Collected Works of Jeremy
Bentham. Literary and Linguistic Computing 27.2 (2012): 11937. Print.
Clement, Tanya, ed. In Transition: Selected Poems by the Baroness Elsa von
Freytag-Loringhoven. U of Maryland Libs., n.d. Web. 8 Jan. 2010.

Design Principles for Text Encoding Guidelines. 9 Jan. 1990. Web. 17 Sept. 2012.
<http://www.tei-c.org/Vault/ED/edp01.htm>.

Drucker, Johanna. Theory as Praxis: The Poetics of Electronic

Texuality. Modernism/Modernity 9.4 (2002): 68391. Print.

Eliot, T. S. The Waste Land: A Facsimile and Transcript of the Original Drafts,
including the Annotations of Ezra Pound. New York: Harcourt, 1971. Print.

Finneran, Richard J., ed. The Literary Text in the Digital Age. Ann Arbor: U of
Michigan P, 1996. Print.

Flanders, Julia. The Productive Unease of Twenty-First-Century Digital

Scholarship. Digital Humanities Quarterly 3.3 (2009). Web. 5 Jan. 2010.

Hockey, Susan. Electronic Texts in the Humanities: Principles and Practice. Oxford:
Oxford UP, 2000. Print.

Kiernan, Kevin. Alfred the Greats Burnt Boethius. Bornstein and Tinkle 732.

Landow, George P. Hypertext 2.0: The Convergence of Contemporary Critical

Theory and Technology. Baltimore: Johns Hopkins UP, 1997. Print.

Lavagnino, John. Completeness and Adequacy in Text Encoding. Finneran 6376.

Lewis, Wyndham, ed. Blast. 1914. Foreword by Paul Edwards. Berkeley: Gingko,
1981. Print.

Liu, Alan, et al. Born-Again Bits: A Framework for Migrating Electronic

Literature.Vers. 1.1. Electronic Lit. Assn., 5 Aug. 2005. Web. 5 Jan. 2010.

Loizeaux, Elizabeth Bergmann, and Neil Fraistat. Reimamagining Textuality:

Textual Studies in the Late Age of Print. Madison: U of Wisconsin P, 2002. Print.
McGann, Jerome. Radiant Textuality: Literature after the World Wide Web. New
York: Palgrave, 2001. Print.

. The Rationale of Hypertext. Sutherland 1946.

McKenzie, D. F. Bibliography and the Sociology of Texts. Cambridge: Cambridge

UP, 1999. Print.

Montfort, Nick, and Noah Wardrip-Fruin. Acid-Free Bits: Recommendations for

Long-Lasting Electronic Literature. Vers. 1. Electronic Lit. Assn., 14 June 2004.
Web. 5 Jan. 2010.

Moretti, Franco. Graphs, Maps, Trees: Abstract Models for a Literary

History.London: Verso, 2007. Print.

Palmer, Carol L. Thematic Research Collections. Schreibman, Siemens, and

Unsworth 34865.

Price, Kenneth. Electronic Scholarly Editions. Siemens and Schreibman 43450.

Renear, Allen. Out of Praxis: Three (Meta)Theories of Textuality. Sutherland 107

26.

Renear, Allen, Elli Mylonas, and David Durand. Refining Our Notion of What Text
Really Is: The Problem of Overlapping Hierarchies. N.p., 6 Jan. 1993. Web. 17 Sept.
2012. <http://www.stg.brown.edu/resources/stg/monographs/ohco.html>.

Robinson, Peter. The Digitization of Primary Textual Sources. Oxford: Office for
Humanities Communication, Oxford U Computing Services, 1993. Print.

. Editing without Walls. Literature Compass 7.2 (2010): 5761. Web. Aug.
2012.

.Is There a Text in These Variants? Finneran 99116.

. The Transcription of Primary Textual Sources Using SGML. Oxford: Office for
Humanities Communication, Oxford U Computing Services, 1994. Print.
Schreibman, Susan. The Text Ported. Literary and Linguistic Computing 17.1
(2002): 1787. Print.

Schreibman, Susan, Ray Siemens, and John Unsworth, eds. A Companion to Digital
Humanities. Oxford: Blackwell, 2004. Print.

Shillingsburg, Peter L. From Gutenberg to Google. Cambridge: Cambridge UP, 2006.

Print.

Siemens, Ray, and Susan Schreibman, eds. A Companion to Digital Literary Studies.
Oxford: Blackwell, 2008. Print.

Smith, Martha Nell. Corporealizations of Dickinson and Interpretive Machines.

Bornstein and Tinkle 195221.

. Electronic Scholarly Editing. Schreibman, Siemens, and Unsworth 30622.

Sperberg-McQueen, C. M. Textual Criticism and the Text Encoding Initiative.

Finneran 3762.

Sutherland, Kathryn, ed. Electronic Text: Investigations in Method and Theory.

Oxford: Clarendon, 1997. Print.

TEI: History. Text Encoding Initiative. N.p., n.d. Web. 5 Jan. 2010.
<http://www.tei-c.org/About/history.xml>.

Unsworth, John. Thematic Research Collections. N.p., 28 Dec. 2000. Web. 10 Jan.
2010. <http://www3.isrl.illinois.edu/~unsworth/MLA.00/>.

DOI: 10.1632/lsda.2013.4