The application of textual encoding for a data-driven analysis of the medieval Welsh legal tradition, Cyfraith Hywel

This thesis uses Digital Humanities methodologies to engage with the Cyfraith Hywel (CH) tradition and explore macro-scale comprehension of its contents. The law text CH is amongst the most prolific outputs from the medieval Welsh period, appearing in approxi- mately eighty extant manuscripts and books between the 13th and 18th centuries. The text is widely considered to have been a cornerstone of Welsh identity throughout the period of upheaval caused by the Anglo-Norman invasion. It presents a cross-section of the Welsh social and legal customs throughout, and arguably prior, to the period of composition. Whilst there are certain inherited characteristics that unite the CH tradition, only two manuscripts composed before The Acts of Union (1536) are identical (Jenkins 2000a, p.10). This thesis takes the point of view that each manuscript should be treated as a translation in its own right, bearing unique characteristics of its own spatio-temporal milieu. This holistic perspective has been observed in the literature as the mindset through which the modern understanding of CH might progress. Despite the potential benefits, this approach has been hard to achieve within scholarship due to the inaccessibility of available presentations of CH to digital methodologies. The structural, orthographical and linguistic variations of medieval texts, such as those represented in the CH tradition, renders the accurate application of DH methods challenging without manual intervention. The research in this thesis contributes manual intervention that allows DH methods to benefit CH scholarship. This involves the consolidation of existing organisational and pre- sentational conventions for CH alongside an investigation of how these might be adapted and applied systematically in a DH context. This thesis then applies this analysis to create a standardised corpus of 21 manually encoded manuscripts from the CH tradition. The encod- ing enriches the original text with structural and grammatical metadata for each manuscript, contributing a stepping-stone for both manual and automated examination. It enables users to efficiently access the corpus at multiple levels of granularity and facilitates the fuller ap- plication of DH methods. The thesis further goes on to demonstrate the benefits of this contribution towards broader CH scholarship through the application of select statistical language processing (SLP) tech- niques. The CH corpus is explored to highlight previously unrecognised or uncertain rela- tionships and patterns that exist within and between the encoded manuscripts. In particular this thesis defines the scope of the encoded manuscripts and their contents, investigates the extent of similarity that exists between the manuscripts at a variety of granularities and explores a keyword driven approach to examining recurrent themes within the corpus. The exploration combines quantitative and qualitative data to enhance modern comprehension of CH as well as providing alternative methodological avenues that supplement and advance existing scholarship