Excerpted: Reading, Scholarship, and Hypertext EditionsSkip other details (including permanent urls, DOI, citation information)
This work is protected by copyright and may be linked to without seeking permission. Permission must be received for subsequent distribution in print or electronically. Please contact firstname.lastname@example.org for more information. :
For more information, read Michigan Publishing's access and usage policy.
I wrote this essay in August 1993. Rereading it in August 1997, only four years later, it seems to me in some ways an artifact of an age now past, and in others a description of a future that still hasn't arrived. I had thought the essay would date quickly, and indeed made the time to write it in 1993 because I believed the points I was stating would soon become common assumptions; but the way the essay has dated is quite different from what I expected.
The essay is an historical artifact.
It never mentions the World Wide Web. Today every piece of printed matter has a URL on it somewhere, but there aren't any in this essay; the bibliography tells you at one point how to get a particular file by electronic mail from a list server, and that (or gopher) was how you did things in 1993. I didn't think that distribution of electronic editions would be especially easy or that we would be able to distribute anything without providing software (and getting into the complexities associated with it).
In the first paragraph, I say that the community of scholarly editors is in the vanguard in using computers. At that time we were.
I thought it was necessary at the start to explain hypertext, because so few people had any experience of it at the time. Some people in the world of literary studies talked about hypertext, but they were very few. And they, by and large, did not want to hear about computers from me — or anyone else.
Today there is a feeling that the Web, hypertext, and easy distribution of electronic editions is more than a little interesteing, and in any case is clearly going to take over the world. And we still have our doubts and fears about what it all means.
Yet this essay looks to the future because most of its suggestions about things we need to be able to do with texts have not been implemented.
The projects I mention in the essay have in some cases produced a lot and have become highly valuable resources, but there has been little advancement in facilities for working with texts. What we've seen is a great deal of effort directed towards getting texts online in the first place; we still have only the simplest sorts of viewing and searching.
The important efforts I write about aren't all we need. Of course, that work has to be done first, and it's not as easy as it looks. To concentrate on the basics is certainly an attractive option, especially when such a huge audience is ready to welcome our creations. I continue to think, though, that we'll need to move beyond this model to one that makes more use of the possibilities for interaction with and analysis of texts. In that respect my essay still seems to me to be quite current.
This paper was published in TEXT: Transactions of the Society for Textual Scholarship, volume 8. TEXT is published for the STS by the University of Michigan Press.
The paper is made available here by permission of the STS, an organization devoted to the interdisciplinary study of textual theory and practice. Membership dues are $15 per year; to join or to obtain further information, write to Patrick Curran, STS Secretary-Treasurer, Ph.D. Program in English, 33 West 42 Street, New York, NY 10036-8099.
To the extent that textual scholars have an image in the larger world of literary studies, it's as stodgy traditionalists more concerned with technical matters than with innovative ways of thinking about literature. But we are the avant-garde when it comes to the use of hypertext. Other literary scholars assemble hypertexts that are intended only as teaching aids, seldom as research tools; literary critics often talk about hypertext, but very few have taken it seriously enough that they've abandoned print and started publishing their work in hypertextual form only. That is exactly what a number of textual scholars are seeking to do, though: efforts are currently under way to create hypertext editions of a wide range of materials—editions that are intended from the start to work as hypertext, and which will never exist in any printed form at all. Such editions are already in progress for the works of Langland, Chaucer, Dante Gabriel Rossetti, Charles Sanders Peirce, Yeats, and Joyce, and proposals for other hypertext editions have become commonplace.
"Hypertext" can sometimes appear to be a confusing or simply fuzzy concept, because of the variety of existing hypertext systems, but it has one essential feature: linking. A hypertext system is one in which links may be specified between different places in the text. A conventional scholarly monograph is like hypertext in this sense: there will be links from the main text to footnotes, and the footnotes may themselves point you to other footnotes or back to the main text, or into appendices or a bibliography. What makes a hypertext system on a computer different from such a book is the provision of mechanical aids to following such links: with a book you have to leaf back and forth to get to the places where the links point, but in a computer hypertext the machine can take you there at once. There is often much more to computer hypertext systems—for example, the ability to add new links to the text—but linking and provisions for following links are the basic requirements.
Many projects for literature-oriented hypertexts focus on annotation rather than on textual problems; they exploit the similarity of linking to the operation of traditional footnotes. Hypertext editions of the kind I'm going to discuss have a specifically textual focus because they're all prepared to deal with more than one version of a work—with variants of individual words and passages, or with whole variant texts. Other hypertext systems assume that you've previously established a sin gle text without such variation. I'll be concentrating on the way in which variation is to be handled in hypertext editions: many are also intended to include facilities for annotation (not only traditional explanatory notes, but also pictures, sound clips, and video clips) and other common hypertext constructs, but these constructs wouldn't behave very differently in any hypertext edition that I've heard of, so I won't say much about those things.
As textual scholars, we are enthusiastic about hypertext because it promises solutions to the problems we've encountered in dealing with variation. There are two clusters of such problems: one has to do with widely-discussed questions about the nature of literary works, and the other has to do with rarely-noticed questions about how we read and study texts. These problems have conditioned the shape of projected hypertext editions, and while they have helped steer us towards hypertext they've also kept us from seeing the full range of hypertext's possibilities.
The appeal of hypertext
In the last few decades, many textual scholars have come to believe that classical texts and modern texts have very different kinds of textual problems and constitute different kinds of literary works.
Texts from classical antiquity have great textual problems: any manuscript that has survived to our day of such texts is the product of a long sequence of copyings and recopyings, so that it's likely to be full of errors in transmission that need to be corrected. These are errors on such a scale that the works are often simply unreadable without editorial correction. But for modern texts, the body of surviving evidence is very different. For texts circulated since the invention of movable type, and particularly for texts written since the seventeenth and eighteenth centuries, the problem of mistransmission is less and less imposing. The texts have been copied only a few times prior to the creation of our sources, rather than many times, and we often have many more of the sources, sometimes going back to the author's own drafts. Error will always be present, and is still sometimes a great problem, but it ceases to be the central problem.
What we have for many modern works is not a shortage of reliable information, but an excess: often there is far too much textual information to include in any printed edition. We find, for example, cases in which a writer made extensive revisions over a span of many years, so that there may be a number of versions that were all produced by the same person and that all have good claims to our attention; but which one should be the text that a scholarly edition prints?
That depends on the focus of our studies: since every version is potentially interesting from some perspective. We are increasingly interested in the way literary works are disseminated to the public, too, so that we now often want to study the physical presentation of the original editions: even a version that doesn't change the wording at all may be of importance for the sake of its appearance alone. And when the emphasis of textual scholarship moves to versions in this way, the importance of editing becomes much greater. Before, an editor would substitute a corrected text of a work for a faulty text, but there was still only one text. Now textual scholars say that critics—and earlier editors—had the entire object of study wrong: they thought a work was a single thing, when it's really a whole range of texts and their relationships.
But it has been difficult to change the form of printed editions in response to the interest in multiple versions. Hypertext has seemed, to many people, a good answer to the two principal problems with representing in printed form works that exist in multiple versions.
The first problem is that it's very expensive to print books that contain all these versions. There are six manuscript versions and two printed versions of Middleton's Game at Chess.
"Our screens are too small, and less pleasant to read from than books are."
The traditional approach in printed editions to presenting this information is the scholarly apparatus. A second reason for hypertext editions, though, is that it is very difficult to actually "read" a version that is presented in this way. It involves constantly moving back and forth between text and notes, and careful study of the notes to pick out the variants for the version you care about from all the others listed there. There is little evidence that apparatus of this traditional sort gets used very much by literary scholars today. What happens instead is that there is one version at the center of the edition—the version whose text is the main text—and other versions are subordinated to this central version by being relegated to the notes. Moreover, the apparatus is typically incomplete for many modern works: editions of Dreiser and Conrad, for example, will tell you that there are just too many variants in the surviving versions to include, given the constraints of paper publication. So versions other than the main text may not be completely recoverable from the data presented in the edition.
But there is a form of print presentation that's rather popular, and that's the parallel text. With a parallel text, you can read a single version without going to the mental effort of "reconstructing" it, and you can still make comparisons to other versions at any point to see what they say. Printed parallel-text editions still have to select the versions they'll display: they are rarely able to print more than two. But a hypertext edition could allow you to select any two.
What a number of scholars have imagined a hypertext edition would be, then, is a system that would store both electronic texts and images of all the versions of the works in question, and offer the ability to display parallel texts of any two versions, as either images or electronic texts. Every hypertext edition in progress will do more than this, but this is the core of them all, and is what I will mean in the rest of this essay when I refer to hypertext editions.
The most common objection to these systems, or at least to the idea of these systems, is that they will make reading difficult. For the most part these objections are not responses to hypertext as such, but to the limitations of present-day computer hardware: to the fact that our screens are too small, and less pleasant to read from than books are. As literary scholars we mainly use paperbacks, and the paperback is friendly: it's cheap, you can read it in bed or on the beach, and it requires no batteries. The computer is expensive and hard to use, and even laptops are too heavy.
But of course these hypertext systems are not meant to substitute for paperbacks (which are unlikely to go out of existence, any more than writing with pen and paper was superseded by the typewriter); they are substitutes for large-scale scholarly editions, which have typically been published in large volumes too heavy and too expensive for good beach reading. They are meant for study, not for reading. Vannevar Bush, the inventor of hypertext, intended it as a mechanical aid to scholarly labor, not to ordinary reading: the associative motion of the link, rather than the straightforward linear motion of conventional reading, is how scholars work, Bush thought, and something that print tends to resist (see Nyce).
But the actual focus of hypertext editions has not generally been on scholarship. People who work on hypertext from other perspectives tend to find these editions rather strange for just this reason: they are inclined to say, "Why do they call this hypertext?" The focus of these editions is not on moving from one text to another in an associative manner: it's on having complete versions, single documents that we can read without having to move off to other texts—just the kind of reading that we normally do with paperbacks.
The focus, in other words, is on something that is rarely mentioned in any kind of literary scholarship: on reading as an involving process, not as interpretation or decoding. It is reading as an experience and not as mere collection of data: it can lead to interpretation, but only by way of generating reactions that we subsequently seek to describe or explain, not by being the only way we can collect all the references to fish for the footnote on page 37. A fundamental condition of such reading is that it
"Decisions made early on can make it impossible to do certain things later on without starting all over again."
We will, soon enough, become accustomed to reading from the computer screen, or will get something better than computer screens, and then we'll be able to use these hypertext systems as they were intended: but that intention is largely oriented towards reading as performance. Hypertext editions have sought to make it possible for readers to perform all versions of a work, an action that was not possible with most printed editions. Most projections have given much less attention to the question of what's needed for scholarship. We must consider whether these hypertext systems are readable but not usable: performable in a way that conventional scholarly editions are not, but not usable for study in the ways that conventional editions are, or in the new ways that will be necessary with works constituted of multiple versions.
We can see the nature of the problem in Roger Laufer's account of the hypertext edition he and Antoine Bertholon created—the first hypertext edition ever created, so far as I know. Laufer's work originated in an incident which is emblematic of the dissatisfaction with printed editions: he looked at his edition of Le Sage's Diable boiteux some time after its publication—his own edition—and he found it unreadable. Hypertext, Laufer saw, was a way to make all the versions of such a work available in a readable form: conventional apparatus was a "pis-aller detestable" for that purpose (119). Accordingly, he and Bertholon set about constructing a hypertext edition of La Rochefoucauld. But they discovered that, once they could read the various texts of the Maximes, they then wanted to study them, and to do that they needed apparatus. Or, at any rate, they needed something to serve some of the functions that apparatus serves in printed editions: they wanted summaries of how the texts differed, for example, and not just full texts. Parallel presentation for reading, the heart of their system, was indeed essential, but it was also not enough.
Making scholarship possible in hypertext editions
Hypertext editions as currently projected are incomplete; as Bertholon and Laufer discovered, the display of reading texts needs to be supplemented by the analytical tools that scholarship will demand for working with multiple versions. On the one hand, this is not an imposing problem, since a hypertext edition can always be extended: as most have perceived, an electronic edition needn't ever stop growing and changing (though it is also true, though less widely recognized, that decisions made early on can make it impossible to do certain things later on without starting all over again). Getting the images and texts into the computer in the first place will have to be the central task in the early stages of such an edition, no matter how it's done, and we could simply assume that analytic tools will follow in later stages.
On the other hand, these analytical tools are in general harder to create than a basic system for hypertextual display, and less likely to be provided by commercial vendors, who are not going to recognize on their own why they are necessary or perceive any large markets for them. Good commercial hypertext systems exist, suitable for use in creating hypertext editions, but almost all of the worthwhile text-collation software, for example, has been written by literary scholars; and writing such software is a costly task, not one that can be slipped in around the edges. It is also going to be essential to have these analytical tools early on, not just as a later stage of refinement: given the difficulty of dealing with many versions of a work, with handling the vast body of data that a hypertext edition can include, scholars will need these aids in order to get anywhere at all, and a system that does not include them may fail to attract enough supporters to keep its funding going until they can be incorporated. Yet it is striking how many proposals for hypertext editions fail to mention even the rather ordinary function of text searching, although, mundane as it is, it is one of the most valuable things that can be done with electronic texts; and in a world where thinking about works as composed of multiple versions is not widespread, a hypertext edition needs to offer as much help as it can to the scholar—going beyond simple access to the texts and the analytical tools, like text searching, that are common today.
"Once you have multiple versions you want to compare them; and once you want to compare them, you discover the limitations of parallel presentation."
Working with a hypertext edition, we will move back and forth between two distinct activities, reading and scholarship: between reading versions of a work in their entirety, and studying those versions and their relationships. Facilitating reading is an important achievement, but, as Robin Cover observed, it is also not something that particularly calls for hypertext; the real advantage of a hypertext edition will lie in its tools for scholarship. We have yet to finish learning just how to study multiple-version works, but four tasks will certainly be necessary: selecting versions to look at; comparing versions; constructing new and possibly more representative versions of the text on the basis of the information available; and integrating all this study with other scholarship and criticism.
Some projects for hypertext editions have retained the notion of a main text, a central version; while others would follow Bertholon and Laufer and do without a center. But either way, they would not properly represent our understanding of the works in question. The reason for making every version of a work available is that each version has particular characteristics of its own that make it interesting, or not, from a given perspective. They aren't all of identical significance, nor do they form the same array of significance from every point of view. What we want is some way to reorder or recategorize the versions based on our interests. But the projected systems don't do anything to make them different, and consequently leave us in the dark when we ask, Which version should I read? What they provide us with is the texts themselves: but the internal information in the texts is not enough for the purpose of selection. We will also need external information about the circumstances surrounding the production and dissemination of each.
It isn't adequate just to have a prose description of the nature of each version attached to it. This is certainly a good idea, but there will often be too many versions for that to be manageable as our only tool for selection: there are dozens of Chaucer manuscripts, for example, differing by date and place of creation, scribe, hand, paper, and contents—to name only the most obvious traits. Without computer-aided facilities for selection scholars are likely to fall back on familiar forms of selection such as sequence, and simply look at the first or the last version. What we need are ways to specify our interests at this particular moment and get a ranking or categorization of versions on that basis, and a display of the version that is most important from our present perspective. Where a printed edition cannot easily represent more than a few perspectives, this sort of selection tool will allow the reader to adopt a new perspective every day and get the equivalent of a new edition each time; a hypertext edition with this capability will represent the work not in just one way, as a decentered (and incomprehensible) heap of undifferentiated texts, or as a central text with peripheral perturbations, but in many ways—as many ways as there are questions that can be asked. Jack Stillinger's work on Coleridge's revisions makes productive use of one such form of selection: it is "a kind of archaeological excavation, as it were, in which versions actually existing at the same time can be connected to form layers of Coleridgean textual history" (138). Looking at the revisions that Coleridge made to a group of his poems in a particular period, rather than at the apparently haphazard changes to individual poems across his entire career, suggests intentions behind Coleridge's practices that are not otherwise visible. Stillinger didn't use a computer; he made a chart of Coleridge's texts by hand, which was feasible since he studied only seven poems. But a hypertext edition should be able to reconstruct such archaeological layers for us, from the perspective of any date in the author's career.
The Canterbury Tales Project is already including one important selection feature of a very different kind: it will include computer-generated analyses showing possible stemmata for the entire corpus of versions, which could be used to help you decide which are worth taking a look at (Robinson). This is a much more sophisticated version of the familiar sequence criterion of selection; and like Stillinger's approach, it focuses not just on selecting a text or texts to read, but on investigating how different versions are related. That is the form of question that the study of multiple-version works must involve.
As Bertholon and Laufer discovered, once you have multiple versions you want to compare them; and once you want to compare them, you discover the limitations of parallel presentation. Comparing passages of any length, or more than two texts of any work, remains a laborious task even with a parallel display; in those cases, we soon find we want the apparatus back again. Apparatus may be a "pis-aller detestable" for providing reading texts, but it is very advantageous for comparative study: it summarizes in a concise way the differences between texts. We cannot do without that in studying multiple versions of works.
But we can ask for an apparatus that is more suited to our particular needs, since it need not be static like a printed apparatus. We want a collation facility at hand that will compare the texts we select—taking any one of them as base text—and tell us about the particular features we care about. We may not want to see differences in punctuation or spelling, or they may instead be exactly what we are studying; we should not have to pick the differences we're studying out of the list ourselves, as is necessary with a printed apparatus—the collation should include only the kind of variants we asked to see. It should give us a very specific report on particular questions we have about the relations of a particular body of texts.
We needn't display such a report in the traditional manner. It would be useful to have a graphical display showing a measure of basic similarity, perhaps broken down by chapter or book; or showing the variants longer than a certain size. It would be useful to retain the parallel display but add highlighting of the passages that differ, as John Hazel Smith did in his parallel-text edition of Chapman's Bussy D'Ambois. Because such highlighting gets in the way of reading, it is rarely used in printed editions; but in a hypertext edition it needn't be there all the time, only when we have asked for it.
The Canterbury Tales Project has done something about this too: it includes databases of
"Print editions tended to give you a readable text of the editor's corrected version and no other; but now we can give you all the original data and let you choose."
The results of selection and comparison may well lead us to want to create new texts. The establishment of new texts, however, is one use that is ruled out by many hypertext editions.
This is part of the reaction against the limitations of print that's behind these editions: print editions tended to give you a readable text of the editor's corrected version and no other; but now we can give you all the original data and let you choose. And yet, is this all that readers want? Because in fact one would very often like to know that a scholar who has spent years working with the texts thinks there's a problem at a particular point; the communication of ideas about the texts is, surely, one goal of editions, even hypertextual ones. And, if we want to communicate such ideas, is apparatus the best way to do it? Not according to other ideas at work in the same systems. If the uninterrupted text is a good way of presenting a version, it is also a good way of presenting an edited text.
Here again, the dissatisfaction with printed editions that hypertext editions are attempting to remedy is a response to real problems—but the solution is to solve the problem and nothing more, whereas if we consider what's actually possible with hypertext we can see ways to retain the merits of print editions while alleviating their problems. Having an edited text on offer is a very different matter when it is easy to display the originals and offer justification for emendations. It no longer displaces the original documents, and so it has a much less authoritative presence. Providing an edited text along with electronic texts of the original documents, and facsimiles of them, puts the reader in a position not usually available with a printed edition: that of being able to check, very easily, any editorial decision. But to provide nothing but the originals will, with many works, mean that the reader is not assisted but is given a rather sizable burden: that of editing every work as well as reading it.
What proponents of these systems have imagined is that they would be transparent: they would not interpose an editor between the sources and the reader. Yet that implies that these sources themselves are always transparent, are never concealing something that scholarship can help us to perceive. This idea, that we require no form of help with original documents, is not really very different from the idea that literary criticism is unnecessary because our untutored reactions to literary works are more authentic, and those reactions are likely to be repressed or distorted if we hear any discussion of what the texts mean. To refrain from editing is an easy way to alleviate our nagging professional worries about being wrong; but it also means that we lose the opportunity to be right about anything, and to give other readers the benefit of our perceptions.
A hypertext edition can go beyond the inclusion of edited texts: it can attempt to bring more of its readers into the process of editing, by including tools for doing editorial work. Much of an editor's work is based on the kinds of collation that I've suggested will be needed in a hypertext edition; for the construction of new edited texts, facilities will be needed for recording decisions about which variants to choose. The Electronic Peirce Consortium's plan includes a further mechanism tailored to the peculiar nature of the Peirce papers: it is a formidable editorial problem just to establish the sequence of manuscript pages for a work from the many radiating drafts that Peirce generated (Neuman). The task is so huge that the Consortium has assumed that the only way to get it done is to distribute the work among many scholars and to create ways to share editorial findings; their solutions are far more ambitious than the kinds of extensions to hypertext editions that I suggest, but those solutions point to some of the possibilities of hypertext that other projects have not yet investigated.
Integration is the fundamental challenge of any attempt to transform scholarly editions. I've described the principal challenge, as seen within the field, as being that of dealing with multiple versions of texts. But the problem of integration with what other literary scholars are doing is really a bigger andmore intractable problem. Literary scholars,
"It's possible for hypertext to constitute merely an innovative way to be provincial."
It's possible for hypertext to constitute merely an innovative way to be provincial. The particular provinciality that we indulge is the belief that, if we could only remove the editorial presence from the transmission of texts, readers would then have a true and complete perception of texts. The actual limitations of that view, and the limitations that most hypertext editions proposed so far would suffer—except for those like the Chaucer and Peirce projects that have broader goals in mind—are exemplified in The Complete King Lear, put together by Michael Warren. "Complete" according to his idea of how texts should be presented (rather than edited in the conventional sense) means that this publication provides unbound facsimiles of the two Quartos and of the Folio text, and a separate presentation in parallel of the Folio and First Quarto texts, again in facsimile. In other words, it does with print pretty much what has been proposed for hypertext editions: it makes the original documents available, and it provides a parallel presentation of them.
I feel about The Complete King Lear just the same way I do about many hypertext editions: it's a wonderful thing, but I wish it did more to convince people outside textual scholarship of its wonderfulness. And Warren himself had noticed that the simple availability of facsimile texts didn't tend to change anything: in 1985, writing about why he wanted to create a Shakespeare edition of this kind, he had begun by talking about an earlier wave of facsimile publishing (from the 1940s to the 1960s), the great claims made for those facsimiles, and the general lack of effect they had on literary criticism. His essay suggested that an analytic commentary could be included with such a facsimile, though there is none in The Complete King Lear; subtleties of the original printing and spelling, which his essay discusses, are apparently just left for the reader of his book to rediscover, though without some sort of guidance few readers will even know what to look for. Since this work is already very costly, the familiar economic restrictions of print may have imposed this omission. His argument for such a commentary remains important for hypertext editions, though:
In cases such as these where the relation of bibliographical interpretation to critical interpretation is becoming recognized as crucial, indeed avoidable only at great risk, there will soon be no escaping the need for a new format that will enable a critical argument to be presented and the undertaking of proper study in consequence. That argument in all its complexity will need to appear as separate commentary, preferably in a separate volume, clearly subordinate to the object of study, not impinging upon its identity, thereby making prior claims for the editor's position and so prejudicing the reader. ("Textual Problems" 36)
Many creators of hypertext editions have also tended to take this view of the prejudicial effects of editing, but have not thought enough about the other side of the question—that "proper study" requires a knowledge of textual matters, not just unmediated access to the originals (or rather to facsimiles of them). With printed editions, these imperatives have conflicted; but they don't have to in a hypertext edition, which can provide textual commentary of unprecedented size and detail, and yet keep it clearly separated from the texts themselves, if that's desired. But to facilitate the development of this way of thinking about texts, we need more than that: a good textual commentary can suggest directions for further research, but without computer assistance in pursuing such research it will not become a more common activity. Hypertext editions as currently planned will provide the access that we will require to all the versions of each work; but to progress to thinking about the work as encompassing all those versions, rather than merely performing the same critical operations upon a larger number of single texts, is the task that really requires computers, and it is the edition that can help us to perform that task that will be fruitful.
Bender, Todd K. "Literary Texts in Electronic Storage: The Editorial Potential." Computers and the Humanities 10 (1976): 193-199. [doi: 10.1007/BF02426307]
Conrad, Joseph. The Secret Agent: A Simple Tale. Ed. Bruce Harkness and S. W. Reid. The Cambridge Edition of the Works of Joseph Conrad. Cambridge: Cambridge University Press, 1990.
Cover, Robin. "Encoding for Textual Parallels and Critical Apparatus." Text Encoding Initiative Working Paper, February 21, 1990. Available as part of file TEI-L LOG9002 from the Listserv program at uicvm.uic.edu on the Internet, and at uicvm on the Bitnet.
Delany, Paul, and George P. Landow, eds. Hypermedia and Literary Studies. Cambridge: MIT Press, 1991.
Dreiser, Theodore. Sister Carrie. Textual ed. James L. W. West III. The Pennsylvania Edition. Philadelphia: University of Pennsylvania Press, 1981.
Duggan, Hoyt. "Velum to Video: The Piers Plowman Electronic Archive." MLA Convention. Hyatt Regency Hotel, San Diego. 29 Dec. 1994.
Faulhaber, Charles B. "Textual Criticism in the 21st Century." Romance Philology 45 (1991): 123-48.
Finneran, Richard, and George Bornstein. "Towards a Hypermedia Edition of the Poems of W. B. Yeats." Colloque international Consensus ex Machina? Association for Literary and Linguistic Computing/Association for Computers and the Humanities. Paris — 19-23 avril 1994. Paris: n.p., 1994. 73-74.
Gabler, Hans Walter. Response to panel "Editing Joyce: The Implications of the Ulysses Debate Ten Years Later." Society for Textual Scholarship Conference. Pierpont Morgan Library. 8 Apr. 1995.
Groden, Michael. "Ulysses in Hypermedia." Society for Textual Scholarship Conference. CUNY Graduate Center, New York. 7 Apr. 1995.
Laufer, Roger. "Edition critique synoptique sur ecran: l'exemple des Maximes de La Rochefoucauld." Les editions critiques: Problemes techniques et editoriaux. Actes de la Table ronde internationale de 1984. Ed. Nina Catach. Annales Litteraires de l'Universite de Besancon 370. Paris: Les Belles Lettres, 1988. 115-125.
McGann, Jerome J. "The Complete Writings and Pictures of Dante Gabriel Rossetti: A Hypermedia Research Archive." TEXT: Transactions of the Society for Textual Scholarship 7 (1995): 95-105. Also available as file ROSSETTI TXT from the Listserv program at [formerly iubvm.ucs.indiana.edu] on the Internet, and at iubvm on the Bitnet.
——— . "Which Yeats edition?" Times Literary Supplement 11-17 May 1990: 493-4.
Middleton, Thomas. A Game at Chess. Ed. T. H. Howard-Hill. The Revels Plays. Manchester: Manchester University Press, 1993.
Miilumaki, Risto. "The Prerelease Materials for Finnegans Wake: A Hypermedia Approach to Joyce's Work in Progress." Conference Abstracts of the 19th International Conference of the Association for Computers and the Humanities and the Association for Literary and Linguistic Computing, June 16-19, 1993. Washington, DC: Georgetown University, 1993. 122-123.
Nelson, Ted. Literary Machines. Swarthmore, Pennsylvania: Ted Nelson, 1981.
——— . "A New Home for the Mind." Datamation March 1982: 169-171, 174, 178, 180.
Neuman, Michael, Mary Keeler, Christian Kloesel, Joseph Ransdell, and Allen Renear. "The pilot project of the Electronic Peirce Consortium." ALLC-ACH '92 Conference Abstracts and Programme. Christ Church, Oxford. April 1992. 25-27.
Nielsen, Jakob. Multimedia and Hypertext: the Internet and Beyond. Boston: AP Professional, 1995.
Nyce, James M., and Paul Kahn, eds. From Memex to Hypertext: Vannevar Bush and the Mind's Machine. Boston: Academic Press, 1991.
Robinson, Peter M. W. "Manuscript Politics." The Politics of the Electronic Text. Ed. Warren Chernaik, Caroline Davis, and Marilyn Deegan. Oxford: Office for Humanities Communication, 1993. 9-15.
Smith, John Hazel, ed. Bussy D'Ambois: A Tragedie. The Plays of George Chapman: The Tragedies, with Sir Giles Goosecappe. A Critical Edition. Gen. ed. Allan Holaday, assisted by G. Blakemore Evans and Thomas L. Berger. Cambridge: D. S. Brewer, 1987. 7-263.
Stillinger, Jack. Coleridge and Textual Instability: The Multiple Versions of the Major Poems. New York: Oxford University Press, 1994.
Taylor, Richard, ed. Variorum Edition of "Three Cantos" by Ezra Pound: A Prototype. Bayreuth: Boomerang, 1991.
Warren, Michael, presenter. The Complete King Lear, 1608-1623. By William Shakespeare. 4 vols. Berkeley: University of California Press, 1989.
——— . "Textual Problems, Editorial Assertions in Editions of Shakespeare." Textual Criticism and Literary Interpretation. Ed. Jerome J. McGann. Chicago: University of Chicago Press, 1985. 23-37.*
1. For information on Langland, see Duggan; for Chaucer, see Robinson; for Rossetti, see McGann "Complete Writings"; for Peirce, see Neuman; for Yeats, see Finneran; for Joyce, see Gabler, Groden, and Miilumaki.
2. For a general technical account of hypertext, see Nielsen; for a survey of literary applications, see Delany. I omit a discussion of the history of hypertext editions; the writings of Bender and Nelson contain the earliest proposals for hypertext editions, and McGann's writings on the subject have been the most influential. See also Cover and Faulhaber for good current articles on the practical aspects of creating hypertext editions.
3. The Cambridge edition of The Secret Agent does not list variants in accidentals (xii, 328); the Pennsylvania edition of Sister Carrie does not list all variants or even all emendations (587). The editors of both volumes have sought to make this information available in other ways—through library deposit, for example.
4. "...la reprenant peu d'annees apres l'avoir terminee, j'ai constate qu'elle etait pour moi-meme illisible. Pourquoi? Parce qu'une telle edition n'est que la materialisation d'un patient travial de releve, fait a grande fatigue des yeux, et non un dispositif a lire. Je suis persuade que personne n'en a jamais lu critiquement plus d'une ou deux pages d'affilee, et sans doute trois pages tout court" (118).