First published in Parergon, January 2002

The Canterbury Tales Project, under the general editorship of Norman Blake and Peter Robinson, has been revolutionising Chaucer studies since the appearance of an electronic edition of the Wife of Bath's Prologue in 1996. The aim of the project has been to present every one of the surviving pre-1500 witnesses to the Canterbury Tales, whether manuscript or printed, in a form suitable for advanced textual research. This has meant going beyond the obvious advantages of the electronic form — full transcriptions and digital images of each witness — to provide more analytical materials, particularly databases of word forms and occurrences in each witness, both regularised and un-regularised. All these materials have been extensively interlinked to enable browsing and searching. Software normally employed in genetics has also been used to analyse the textual tradition and to produce stemmata. All these features were well demonstrated in the recent edition of the General Prologue, edited by Elizabeth Solopova (Cambridge University Press, 2000).

The latest addition to this corpus of material takes a rather different approach. It contains the whole of a single manuscript — Hengwrt — rather than many witnesses to a single section of the Canterbury Tales, and uses new software called Anastasia to present the text. At the heart of this edition are colour images of each page of the manuscript, accompanied by diplomatic transcriptions of Hengwrt and of the Ellesmere manuscript for comparison. A range of supplementary materials is also provided, including a detailed description of the manuscript (by Daniel Mosser), an essay on the language of the scribe (by Simon Horobin) and a study of Hengwrt (by Estelle Stubbs).

The power of this edition comes from the sophisticated ways in which the corpus of interlinked materials can be displayed and navigated. Any given section of the text can be viewed in various forms: as image, as image with transcription, or as parallel Hengwrt and Ellesmere transcriptions. Even the underlying SGMLbased text encoding (using TEI) can be seen if required. Images can be displayed as a single page, or as an opening (facing pages), or as conjugates (two pages on the same side of a sheet). They can be viewed at sizes from 25% to 200%, with 50% as the default, and a magnifying glass can be used to enlarge sections of a page. The user can browse, either by going sequentially page by page, or by picking from a list of sections and tales or from a list of quires, or by going directly to a specific page number. The Hengwrt transcription can be searched for words or phrases, and searches can be limited to specific areas of the text. The presentation and navigation of this edition have managed to outstrip even the high standards set by the previous Canterbury Tales Project publications. Not only are the images marvellously clear and the transcription very accurate, but there are also numerous unobtrusive ways in which the materials are made easier to use. I particularly liked the visual assistance provided: the diagram for each page showing its place in a quire, the colour-coded analysis of the different inks used, the highlighting of variants in red in the parallel transcriptions, the 'keyword in context' display of search results, and the colour-coded listing of the different orders of the tales in the two manuscripts. The different navigation panels for the whole edition and for each page are clear and concise and easy to follow. The ability to search and analyse the text is satisfactory, but less sophisticated than in the previous publications. Though this is mainly the result of the different aims of this edition, there are a few significant limitations: only the Hengwrt transcription can be searched (not the Ellesmere), while the restriction of wildcards to right truncation makes it difficult to allow for spelling variants (such as 'lycour' versus 'licour'). Limiting a search to a specific area of the text requires an understanding of the SGML encoding.

The Hengwrt Chaucer comes in two versions, the Research Edition (reviewed here) and a simplified Standard Edition. The latter contains only the key elements of the edition: the same images but at lower resolution (100 dpi), the transcription of the Hengwrt manuscript and the editorial materials. It does not include the collation with the Ellesmere manuscript, and does not have the search engine and the navigational tool.

The Research Edition runs on both Macintosh and Windows computers. It uses Web browser software as the interface — either Netscape or Internet Explorer (5.0 is recommended), though there are some minor problems with Netscape. PCs must have Windows 95 or a later version, but some (and possibly most) Windows 95 users will have to download a 'Winsock2' upgrade from the Web and install it before the CD-ROM will run. Macintosh users must configure their machine to run as a local Web server. The instructions for doing this are provided with the CD-ROM and are fairly straightforward; the Remote Access/ PPP system software must have previously been installed. System requirements include a 1024x768 screen, 24-bit colour and 64 MB of RAM. The Hengwrt manuscript is generally regarded as the earliest — and with Ellesmere the most important — witness to the text of the Canterbury Tales. As Estelle Stubbs notes, it is possible that parts of Hengwrt predate Chaucer's death and that he may have had some involvement in its compilation. While this edition cannot give a definitive answer to the many questions raised by this crucial manuscript, it provides the fullest available corpus of materials about Hengwrt, in a remarkably powerful and usable form.

Toby Burrows