Internet Shakespeare Editions

A Most Rare Vision: The Internet Shakespeare Editions

Michael Best, 2003.05.15


[** opening screen:]

Back in 1996, I had an ambitious vision: I wanted to create a website with the aim of making scholarly, fully annotated texts of Shakespeare's plays freely available in a form native to the medium of the Internet. A further mission was to make educational materials on Shakespeare available to teachers and students: using the global reach of the Internet, we wanted to make every attempt to make our passion for Shakespeare contagious.

This seemingly simple set of objectives becomes interestingly complicated as we look at what is involved. What do we mean by a "scholarly" text? What is "full annotation"? And, most challenging of all, what kind of text is "native to the medium of the Internet"?

A "scholarly" text

[** next screen]

The complexity of the process of integrating an advanced academic structure with the new medium was such that I chose to designate myself the "Coordinating" Editor: one whose job was to focus on providing connections between the academic and the technical. In the design of the academic structure of the Editions, I followed the lead of other early scholarly sites, like Early Modern Literary Studies: to ensure that the texts were scholarly, as measured by established standards, I created an Editorial Board, [** open "Editorial Board"] with representation from various flavours of Shakespearean editorial traditions, as well as some members whose expertise was in the area of electronic texts. Those of us working in this field know all too well the suspicion that greets publication on the Internet; as a result, we have developed structures of peer review that are if anything more rigorous than obtains for print publications. [** close]

The basic academic design of the site is unchanged after seven years, and has stood the test of experience. There have been, however, some significant developments. The ISE now has a General Textual Editor, a distinguished scholar, Eric Rasmussen, and under his leadership we have added some of the most distinguished editors currently working in the field to our Editorial Board. [** open "transcriptions"] On the site itself, Shakespeare's plays are now transcribed in spelling and format as initially printed; several plays are represented by more than one text, since they were initially published in one or more Quartos before the appearance of the First Folio in 1623. The range of texts has also been expanded to include seven plays rightly or wrongly attributed to Shakespeare. [** close]

"Full annotation"

[** next screen]

As with any major editorial project, each work will be edited by an individual scholar, or by a team of scholars in collaboration. Aspiring editors are asked to submit a detailed proposal, outlining their editorial principles, the way they see the electronic medium changing the way they edit, and a summary of the kinds of supporting materials that the edition will access -- contemporary historical or literary documents, performance materials, and so on. The proposals are rigorously vetted by the Editorial Board.

Editors will contribute the traditional essays on the transmission of the text, on the critical reception of the play, and on stage history. The texts themselves will provide both painstakingly accurate transcriptions of the originals, and a modern-spelling version -- and here I am beginning to discuss the concept of "full annotation," since modern spelling is a kind of annotation of the most basic kind, as editors make decisions about which modern words and word-forms should replace the originals. Another kind of annotation is the collation: the ISE texts collate the major extant editions, at the same level as the well-known Arden series, but use a more expansive and comprehensible format, as the electronic space allows for a more self-explanatory information than the compact runes forced by the printed page.

The electronic medium gives new meaning to the concept of annotation, as it provides ready access to graphic, sound and video, and a far greater space for discursive comment. The problem here is the need for edges: copia may be a respectable renaissance rhetorical strategy, but even in e-space more is not always better. To make the most of the opportunities the medium provides, while controlling the urge of editors to pass on everything they know, the ISE Guidelines provide for three levels of annotation, designed to allow for users either to select the degree of commentary they want, or to see at glance which kind of note they are invoking as they click, or mouseover:

A significant further development of the site will be in the area of performance. Shakespeare is a perfect vehicle for experimentation in the development of multimedia annotation, since his plays are filmed and performed with such frequency that a whole discipline of criticism depends on discussing the insight into the plays that performance provides. One of the great challenges in developing a performance database of this kind is copyright. For this reason, it is unlikely that much material from popular films will be accessible; the "workaround" is to turn to the remarkably rich source of staging documents of various kinds created by the inventive work of the many professional companies that perform Shakespeare around the world, especially in open-air forums. A parallel challenge is to create a data model that will reveal patterns within the materials as they are added, and allow for both simple and advanced searches by students, scholars, and actors. This section of the ISE has the potential to become a major research tool in its own right, and will in due course require that a further specialist editor be appointed to oversee the process of ensuring quality in the kinds of materials stored in the database.

Texts "native to the medium of the Internet"

[** open ISE home page from icon of swan]

This discussion of multimedia annotation is a reminder of the aim of the site to make the editions "native to the medium of the Internet." The overall design of the site itself is structured to make the most of the expectations of the Internet for informal discourse, while at the same time sending a clear signal concerning the academic credibility of the materials we publish. Using the metaphor of a library, I have divided the site into a "Foyer" for introductory and explanatory materials, the "Library" itself, where only fully peer-reviewed materials are published, a "Theater" from which performance materials will be accessed, and an "Annex," where draft texts, discussions, and other less formal materials can be published. This structure allows a combination of a quality "firewall" around refereed materials, while at the same time permitting the publication of more informal work.

[** "Back" on browser]

[** next screen]

At a more granular and technical level, the ISE encounters the familiar problem facing those of us who work with Early Modern texts: the challenge marking up texts where the physical display of the text is not consistent in terms of conceptual markup, and where there will be overlapping hierarchies between tags that describe physical attributes of the text, and tags that define the literary organization of the work. Let me take a moment to remind you why the physicality of the text is often more important to scholars of this period than the kind of conceptual markup emphasized by the Text Encoding Initiative. In drama, the conceptual structure we are used to -- division of the play into act, scene, and line units -- is in most instances a creation of later editors. The early quartos of plays by Shakespeare and most other writers of the period make it clear that they conceived of their works as consisting of a single conceptual element: the uninterrupted <play></play>. Modern readers, however, expect to find their way around through the conceptual structure, so the text must in some way signal modern divisions.

[** next screen]

However, to privilege an anachronistic conceptual structure imposed on the plays is clearly to distort them, and to make them less accurate as representations of the originals, where the only divisions were those arbitrarily created by the processes of printing. The same kind of problem is even more acutely illustrated in the tagging of conceptual elements like speech prefixes, for example. [** scroll down] In the originals, a speech prefix is normally italicised, and normally indented by two "m" spaces; but there are exceptions, and these exceptions are of especial interest to Early Modern scholars, since they may reveal information about the compositors, or the copy from which they were working. Even in the Folio texts, which are remarkably regular in the format of speech prefixes, there are some interesting anomalies. In this example from All's Well That Ends Well, the curiously titled Lords "1. Lord E" and "Lord G" are usually normalized to First and Second Lords; in the second of the speech prefixes here, the initial number is omitted, and the curious "E" is left unitalicized. This anomaly reinforces our suspicion that the manuscript was in some way itself anomalous -- clearly a piece of information that will assist the editor in creating the modern version. One theory is that the added letters may possibly be vestigial indications, added by another hand, of the names of the actors who played these "bit" parts.

Because of these kinds of problems, rather than engaging in complex "workarounds" with the TEI-SGML that was the standard when I founded the site, I took the risk of creating a special, simplified tagset for the editions, designed to make the process of tagging sufficiently uncomplicated that it could be used by a scholar more versed in the intricacies of Shakespeare's texts than in TEI-SGML. The tagset simply ignores problems of overlapping hierarchies, allowing them to exist side-by-side, and it provides a full spectrum of tags to deal with the physical appearance of the text as well as its conceptual substance -- and here I was guided in part by the work done by Ian Lancashire on his Renaissance English Texts. As you can see in this example, a speech prefix is tagged as a conceptual entity, but the text also includes tags to indicate whether it is italic, so that the exceptions where it is not italicised will be clear.

[** next screen; ** open sample page "Variable displays"] At the same time, the site adheres to the general principles concerning electronic texts as established by early work in the field by pioneering scholars like Faulhaber and Schillingsberg. Thus ISE-tagged texts can be converted to well-formed XML by a Perl script -- selecting either the hierarchy of physical structure or the hierarchy of act, scene, and line. On the ISE site, visitors may choose to see the transcripts displayed in either of these hierarchies, or as a single, complete file. It is also possible to convert the ISE files to TEI-conformant XML, though the result is far more verbose. There are signs that XML will become more comfortable with multiple hierarchies in the future, and when that happens, the ISE texts will again be capable of transformation to the new syntax. [** close]

The modern-spelling texts will be simpler to integrate with the TEI, since they will employ purely conceptual tags. The most interesting feature of the modern text is the opportunity the medium provides for experimentation. Especially in the plays with a complex textual history, it will be possible to produce texts that highlight the multiplicities of the originals, by providing means of displaying variant readings. Mouseover actions might pop up variant words in plays like Othello where differences measure in the thousands; in even more complex cases like Lear or Hamlet, viewers might be able to view parallel texts interlineated, or in parallel windows. [** open "Animated variants..."] A suggestion I made at this meeting last year that ambiguous readings and imprecise entrances or exits could be indicated by animation has been received with a possibly surprising enthusiasm by the editing community in Shakespeare studies.

A vision for the future

My title for this paper celebrated the vision that initiated the project of putting scholarly, multimedia editions of Shakespeare online. At this stage, it is clear that this vision is becoming increasingly a reality: on an average day, an average of over 20,000 pages will be downloaded from the ISE site; we have just published a CD-ROM as part of our educational mission; the site has been given a number of awards both for scholarship and educational content, and recent paper sessions at the Shakespeare Association of America and the Pacific Northwest Renaissance Society focussed on the work of the ISE. It is appropriate, then, that I end this short paper with my sense of our current vision for the future direction of the ISE.

[** close]

[** next screen]

  1. In technical terms, the site will evolve towards the increasing use of relational database structures. We are increasingly able to link text to images of the originals, and the performance database will become something of a model for the whole site.
  2. Computer-driven tools for critical analysis will become more integrated into the site. The ISE is donating its marked-up texts to TAPoR, the Text Analysis Portal for Research at present under development by a consortium of six universities in Canada. Together with our sister institutions, we look forward to enhancing the kinds of readings we can offer visitors to the site. Here the expertise of Ian Lancashire's group at the University of Toronto will be vital, as further tools for the online analysis of texts, along the lines of TACT and PatterWeb, are developed.
  3. Collaboration will increase in a number of areas.
    • As we work towards integrating multimedia texts with encoded transcriptions and modern, annotated editions, the design of sensitive and inclusive metadata, of the kind under development in TAPoR by the team at the University of New Brunswick will be central.
    • The ISE has developed a close relationship with the Shakespeare Theatre Association of America -- for which we host the home page -- and their collaboration will greatly enhance the range of the performance database as it is developed.
    • We are also currently negotiating with two major publishers, one in electronic media, one in print, for interconnected methods of publication. There is a real possibility for a fertile "remediation" of the texts as they are adapted to other media, without sacrificing the prime goal of creating texts "native to the medium of the Internet."