Handling Abbreviations: HTR, TEI, and Evolving Hands

The development of HTR (Handwritten Text Recognition) and its adoption by archives seeking to create full text outputs from images of their handwritten collections provides an opportunity not only to produce plain text for archival searching, but also to export this text in TEI (Text Encoding Initiative) format. This enables the text to be further encoded, either manually or by up-conversion scripts, to add value to these archival texts. In the production of digital scholarly editions from HTR outputs, there is an editorial question of how to handle abbreviations as part of the HTR and TEI workflows. This paper will investigate these issues using a case study from the AHRC/NEH Evolving Hands project. At its very core, the issue of how to handle abbreviations comes down to the editorial approach: is it hyper-diplomatic (almost a facsimile), a diplomatic edition, a normalised edition, or some combination of these? If such editorial approaches represent interpretative distance from the material object, then this should dictate how an editor, or crucially the technical workflow they are relying on, should handle abbreviations. The same is true with graphemes that are no longer in current usage – for example should the Middle English letter ‘þ’ (thorn) be normalised to ‘th’. Similarly, the handling of abbreviations is controlled by the editorial principles for normalisation. This spectrum of choices ranges from not expanding abbreviations at all (but representing, or not, their abbreviation markers), silently expanding the abbreviations, indicating the presence of supplied text in an expanded form, to fully and completely recording both the original abbreviated and the suggested expanded form with the supplied letters noted. In both print and digital outputs, expanded abbreviations often use font-based changes (such as italics) or additional characters (such as parentheses) to indicate the presence of supplied text in the expansion. For example, the abbreviated phrase ‘corpꝰ chrī’ (where the ‘ꝰ’ is used to indicate a ‘us’ abbreviation and the ‘ī’ stands for a preceding ‘ist’) is intended for reader to mentally expand this to ‘Corpus Christi’. In outputs this is usually indicated with italics ‘Corpus Christi’ or parentheses ‘Corp(us) Chr(ist)i’ or sometimes choices to toggle between or use tooltip popups to provide both readings. Archives rightly question whether there is sufficient interest in abbreviated forms for them to undertake more problematic HTR conversion just to preserve these abbreviation/expansion distinctions in the texts they create. However, HTR coupled with TEI gives a number of options with regard to these issues, including direct silent expansion (recognising the ‘Chrī’ pattern simply as ‘Christi’) or representing supplied characters with pseudo-markup for later conversion into more sophisticated TEI for providing a choice between abbreviation and expansion. This paper argues that, where feasible, good practice should be that which preserves more interpretation (such as encoding both the abbreviation and expansion) and uses a case study from the Evolving Hands project on the Records of Early English Drama collections to examine this issue.