The Nun's Priest's Tale on CD-ROM
This CD-ROM edition of Chaucer's Nun's Priest's Tale is the fourth in a series of single-tale CDs produced by the Canterbury Tales Project, the others being the Wife of Bath's Tale (1996), General Prologue (2000), and Miller's Tale (2004). The project has also produced CD-ROM editions of two important witnesses to the complete Canterbury Tales, the Hengwrt Manuscript in the National Library of Wales (2001) and the pair of editions by William Caxton, from copies in the British Library (2003). The series of single-tale editions is an interesting experiment in a computer-based approach to presenting a text of which multiple versions exist. A commonplace of late twentieth-century critiques of traditional editorial practice states that, while a critical edition does represent all witnesses with its lists of variants, nevertheless the editor's choice of a base manuscript governs and mediates the presentation of all the other witnesses, and the eclectic "correction" of the base text with preferred variants from other manuscripts means that the critical edition offers a text that is new and idiosyncratic. In the name of reconstructing the author's "final intentions," a new and non-authorial version of the text is authored by the editor, and, because the other versions are reduced to lists of variants (printed in small type at the bottom of the page or at the back of the book) the user's access to the versions that the editor does not prefer is limited. In many cases, such as when there are a limited number of witnesses to a text, varying only in a limited number of readings, such a critical edition may be adequate; however, in cases of complex textual transmission the inadequacies of this approach multiply exponentially with the increase in divergencies among witnesses to the text. Such is the case, of course, with respect to the Canterbury Tales, for which there are some 70 fifteenth-century witnesses to all or parts of the text, with substantial variability.
In cases where witnesses are so divergent that the editor cannot readily reconcile one copy of the text with another, the approach suggested by Tanselle ("The Editorial Problem of Final Authorial Intention," Studies in Bibliography, 29 [1976], 192) is to treat such divergent copies of the text as separate "versions" and to edit [End Page 128] them separately and in parallel editions (as is commonly done with the A, B, and C versions of Langland's Piers Plowman). Peter Robinson and his Canterbury Tales Project (of which the publisher of this CD-ROM, Scholarly Digital Editions, is an offshoot) extrapolate from this idea of separate editions of "versions" an idea of a computer edition which incorporates full transcriptions of every "version," which transcriptions are accompanied by computational means of collation and comparison, allowing the user to study each version separately or in relation to one, several, or all other versions. While such comprehensiveness opens up all sorts of possibilities for studying the text in its many versions, the quantity of information quickly becomes overwhelming. For someone with a lifetime or two to dedicate to the study of the variants in the Nun's Priest's Tale, this is the CD to buy. And for any Chaucerian concerned with the manuscript witnesses, and for graduate students being trained to be aware of the editorial decisions that go into the creation of a critical edition such as the Riverside Chaucer, having such a tool on one's computer for periodic consultation is most desirable.
There is an impressive array of materials on the CD, beginning with complete transcriptions of all the fifteenth-century witnesses to this tale, including more than 50 manuscript versions and four incunabular prints. Besides the transcriptions, every page of each witness is photographically reproduced in 250 dpi images (high enough resolution that one can zoom in on details); most of the images are blackand white, some are scanned from microfilm copies, but some are new photographs and some are in color. While there is, then, some variability in the quality of the images, they do serve to give the user reasonably unmediated access to the witnesses, so that one can question a reading in the transcription or puzzle over some crux in the manuscript. The images allow the user a level of access to primary materials which previously, in the tradition of critical editions, was reserved for editors.
These images and transcriptions are accompanied by computer programs that display the texts with an emphasis on the variant readings. The default display is "text and image," with the image of a particular manuscript page in one window and the transcription of the text on that page in the adjacent window. One can move about the pages in the collection of images by folio number or by line number within a given witness. Using the search box to look for a given word produces a listing of all occurrences of that word in all of the witnesses, with contexts ("KWIC" concordance style). Further, the transcriptions are linked to full collations in both regularized (with abbreviations expanded and spelling "normalized") and unregularized forms. The collations are given both for individual words and for complete poetic lines of the text: moving the cursor over a word gives one a list of variants for that word; clicking on the line gives one a list of all the variant forms of the line with the variable words highlighted in different colors. There are also "variant maps," which are tree-like representations of how the witnesses cluster according to the agreement and disagreement among them: for any word in the text, one can examine the variant map (which has the appearance of a stemmatic tree, though it does not carry the same assumptions about the historical relationships between witnesses) with the sigla of the various witnesses color-coded according to their reading at this particular point in the text. The software also includes a tool for searching the database of variants for various types of information, including searches based upon measures of variability.
The images, transcriptions, and programs are also accompanied by a series of introductory essays and a bibliography of secondary sources. The sections of the introduction cover such topics as the project and its procedures, guidelines for the transcriptions, the clustering of witnesses by variants, the scribes of manuscripts of [End Page 129] the Canterbury Tales (including a recognition of Linne Mooney's recent identification of "Scribe B" as Adam Pinkhurst), a formal catalogue of the manuscripts and early printed editions, and instructions for navigating through the texts and using the software tools. The introduction to the transcriptions of the Nun's Priest's Tale is by Paul Thomas, the editor of this CD.
In my attempt to evaluate these materials, I did some spot-checking of the transcriptions and found no errors, but there were several coding faults that I encountered: buttons that did not go where they should, line-numbering errors, and one image that never appears because the text for that page appears at the end of the previous page. Similarly, there is evidence of a failure fully to revise material being reused from earlier CDs in the series: a number of pages in the introductory materials appear with The Miller's Tale in the headings. There are also some design weaknesses: the navigation bar is not regularly refreshed to indicate the page that you are viewing, but only the page that you last requested through the navigation bar; and the zoom tool for the images is contained inside the image window and scrolls out of sight as you move around the image. Such annoyances, though, are reasonably few, and one hopes that there will be improvements in future editions. The bigger problem is with the system used for displaying pages, which seems to have some major flaws.
Rather than creating their own display system from scratch, the programmers chose to display pages through whatever web browser is available on the user's computer. In order to do this, a pseudo-server ("Apache") runs in the background, producing HTML versions of the XML pages created by the "Anastasia" software that creates and formats the texts and images. Unfortunately, this marriage of Anastasia and Apache seems not to be a perfectly happy one. Whenever one starts the program, for instance, Apache will declare that it is "unable to start" because another instance of Anastasia is already running and must first be "killed"; this ominous and violent indication of marital disharmony (or revolutionary zeal), however, is always followed by the program starting up without difficulty (Anastasia escapes to be killed another day). The publisher's website indicates that the programmers are aware of this fault and have spent years trying to rectify it, but cannot do so because of some bad code in the Windows operating system. While there may, indeed, be some operating system flaws that they cannot fix, they might consider, at least, mollifying the error message so that one is not having to "kill Anastasia" like a Bolshevik every time one starts or stops the program: perhaps we could "set Anastasia free" when we shut down the program instead of "killing" her.
The false error message whenever the program starts is an annoyance, but one can learn to ignore it after the initial surprise and the discovery that this bug is "benign." There is another kind of error, however, that is a considerably greater problem, which again indicates some mismatch between Anastasia and Apache, or between Apache and the operating system. Testing the CD with several different browsers on several different computers, I regularly, repeatedly, and frequently encountered "time out" and other errors while trying to load pages to display, sometimes of the "apache.exe must close; would you like to send a report to Microsoft?" type, or "connection was reset while loading the page" messages, and sometimes pages displayed in an unformatted state. In such cases, reloading or refreshing the page would usually get past the error, but not always; and having repeatedly to reload a page to get it to display correctly becomes annoying rather quickly, and it seriously undermines the usability of the program. I tried the CD with three different browsers (current versions of Internet Explorer, Firefox, and Safari) and [End Page 130] had similar problems with all of them. Further, I tested the CD on four computers (three different computers running Windows XP Pro as well as one Mac iBook running OS X); curiously, the one computer that ran the program with the fewest errors was my Windows notebook, which happens to have the slowest processor (1. 10 GHz) of the four test machines. Four computers, of course, is an insignificant sample, and correlation does not prove causation, but it could be indicative, and so I was led to wonder whether fast hardware causes "synchronization" problems, and whether the user's best chance for error-free performance is to install the program on an older computer with a relatively slow processor speed.
Such frustrations aside, this is an impressive tool, and it is an interesting experiment in a new kind of editorial practice. Dr. Robinson has asserted the "manifest superiority" of digital editions (see, for instance, his "Current Issues in Making Digital Editions of Medieval Texts—or, Do Electronic Scholarly Editions have a Future?" Digital Medievalist 1. 1 [Spring 2005]); <https://www.digitalmedievalist .org/journal/1.1/robinson/>, but one must ask "superior for what purposes?" For some purposes, clearly, an edition such as this is essential, and it opens up new ways of studying Chaucer's text. As a basis upon which to do a study of the Nun's Priest's Tale that involves comparison of versions and the study of variants, this CD-ROM edition will be invaluable. Further, the description of the manuscripts by Professor Mosser, which appears here as well as on other SDE Chaucer disks, is a major work of scholarship that should take its place beside the catalogues of Canterbury Tales manuscripts by Manly and Rickert, by M. C. Seymour, and by other Chaucer bibliographers. On the other hand, the critical edition in its traditional printed form still, I think, has its uses—for general scholarly study, for instance, or for teaching, or for reading for pleasure—and the digital edition is not yet, as Dr. Robinson seems to suggest, "manifestly superior" for every conceivable purpose.
As a tool for the study of variants and versions of a single tale, however, this is excellent, taking the Chaucer Society's "six text" parallel edition to a whole new level, a "fifty text" edition with software tools. Similarly, the "search" facility makes this something of an interactive concordancing system, which is clearly useful. More generally, this is an experiment in how new tools can create new types of editions, and as such will be of interest not only to Chaucerians but to editors and textual critics more generally. And if such an electronic edition encourages literary scholars to pay more attention to textual variants and the editing process, then it has done a great service to scholarship, and in the hands of a true textual critic it has the potential to be much more than that.
