The Electronic Beowulf Project is a joint project of the University of Kentucky and the British Library and has made available on CD-ROM high resolution images of the manuscript and its early transcripts, together with commentary, glossary, and searchable transcription and edition of the poem. The edition and the transcript have been encoded in SGML to enable the users to search for editorial decisions such as restorations and emendations in the edition, and for palaeographical features and various types of scribal activity in the transcript. Particular features which are now searchable as a result of encoding the transcript include: scribal abbreviations, additions, deletions and corrections, missing text and characters damaged or concealed by restoration attempts.
The Beowulf manuscript suffered damage from the fire of the Cottonian library in 1731 which destroyed or damaged a significant proportion of the text, making textual restoration a near impossible task. Quite apart from the fire damage the manuscript presents various palaeographical and textual problems, and has been a subject of considerable scholarly controversy in recent years. Individual readings in the poem often depend on complex evidence from different sources including seventeenth and nineteenth-century transcripts of Beowulf, from back-lighting, ultraviolet photography and, more recently, digital image processing, as well as the study of spelling and scribal habits throughout the entire Nowell codex of which Beowulf is a part. The Electronic Beowulf Project has brought together a wide range of textual information about the poem, making it easily accessible for re-evaluation and further research. The textual material has been encoded with SGML which, whilst not strictly TEI-conformant, is indebted to the TEI for both the original inspiration and for individual encoding decisions. The markup records a wealth of minute palaeographical and textual detail in both the transcript and the edition of the poem. This paper will briefly discuss the development of the DTD before looking at a selection of the markup problems encountered in the encoding of the Beeowulf transcript. Finally, I is hoped to demonstrate the usefulness of the encoding for specific searches within the Electronic Beowulf CD-ROM.
The first group of encoding problems is represented by cases where the feature which needs to be encoded is actually smaller than the smallest segment of the electronic text - the character. Examples include: the erasure or deletion of a character's minim by underdotting (or indeed added by the scribe); damaged or partly-obscured letters, especially on the edges of a fire-damaged or partly-restored folio. Any editorial restoration of damaged or concealed characters needs to be accompanied by the evidence for a particular reading, often incomplete and based on a range of sources. The editor's comments referring to damaged letters range from 'only descender survives', to 'part of the letter survives', to 'only traces are preserved'. The TEI Guidelines offer a method for describing such instances within its primary documents tag set. The element <DAMAGE>, for example, has an attribute EXTENT which, according to the TEI Guidelines, can have values such as "half-letter", "minim" etc. However a similar attribute is not available for other elements such as <DEL> (for deletions). The Electronic Beowulf Project developed both descriptive notes and customized elements and attributes to accommodate these features.
Multiple scribal alterations carried out in a particular sequence offer another set of encoding problems. For example, the editorial note which accompanies the reading, 'aet' in the transcript reads, 'Originally "seah", but "e" underdotted, "eah" crossed out, and "aet" written over it in the same hand'. All types of scribal activity mentioned in this context, such as the additions and various types of deletions, are commonly encountered in the manuscript separately and encoded with a corresponding set of elements and attributes. However, multi-scribal activity not only defy conventions which work well for the large majority of simpler cases, but present particular difficulties in the need to record the order in which the alterations took place.
A third group of difficult cases concern the presence of ambiguity in the material itself where the encoding had to reflect the unavailability of a straightforward interpretation. An example of this is described in the following editorial note: 'After "dream", traces of erased or faded letters, sometimes restored as "ic", appear to be bottoms of "h" and "e" under ultraviolet light'. The uncertainty expressed in the editorial note presents an encoding problem, even though elements for both faded and erased text were broadly used in the encoding of the Beowulf transcript. A similar difficulty occurs when a particular feature falls under two or more categories distinguished within the markup system. Thus a common method of deletion in Anglo-Saxon manuscripts is underdotting a letter or a word. On the other hand a point beneath the line is commonly used by one of the Beowulf scribes to indicate the place where additions written above the line were intended to belong. There are cases however, where it is impossible to say whether a point beneath the line is an insertion or a deletion mark for in fact it stands for both.
Finally the paper will discuss the use of cross-referencing and the recording of additional information in the markup to represent the complex evidence for individual readings in Beowulf. Thus, the encoding may record that a particular word is missing from the manuscript whilst also recording that its presence in the edition is a result of an editorial restoration based on the seventeenth-century Thorkelin transcripts. The markup, therefore, also serves to indicate something of the status of the reading in the transcripts in order to supply the reader with as much information as possible. This is particularly important given that some readings in the Thorkelin transcripts appear to be a later additions and so already lost from the manuscript when the transcripts were made, casting doubt on usually reliable source of evidence. The paper will discuss the integration of this essential but descriptive material into the textual markup and the subsequent display of the search results on the CD-ROM.