Pushing the Boundaries of Digital Philology

In Winter 2025, I completed an independent study that sought to push beyond the traditional boundaries of digital philology. Developed in collaboration with PhD candidate Davide Pafumi, the project combined computational experimentation with historical linguistic inquiry to investigate how digital tools can support, challenge, and extend philological practice.

The first quartar of the project focused on applying survival analysis, which was adapted from health science and medical research, to study the lifespans of Old English words. We examined lexical emergence, persistence, and disappearance across the poetic corpora of Old English, while also considering other factors which could be factors in survival such as syllable count and word length. This study revealed interesting results regarding the type of words which survived and died, specifically when considering the lexical classifications given by the historic Thesaurus of English. An unexpected result from this experiment was that there were implication that the way we seperated words in the corpera could be used to examine the themes in the poetic codices. This work was presented at the 2025 Congress of the Canadian Society for Digital Humanities (CSDH) conference.

The second part of the independent study focused on digital collation. Using the Textual Communities platform, we collated all the manuscript witnesses of Cædmon’s Hymn to experiment with how textual variation can be computationally visualized. This phase explored how editorial decision-making normally hidden or implicit—can be made legible through interface design and data structuring, and how multiple textual traditions can be represented simultaneously rather than hierarchically.

The third major experiment focused on the sonification of codicological data from The Canterbury Tales manuscripts. In this phase, we explored the transformative potential of data sonification—the process of converting data into sound—as a means of interpreting and experiencing manuscript features such as page dimensions, script density, and layout. Drawing on quantitative codicology, we aimed to create immersive soundscapes that render these patterns audible. This method not only broadens accessibility for visually impaired audiences but also offers a novel way of engaging with medieval material culture. We hope to present this work in the coming year under the title Can You Hear the Sound of Manuscripts? We hope to present this experiment some time in early 2026.

These three experimental strands culminated in our final paper, “Can We Push the Boundaries of Digital Philology? The Limits of Screwmeneutics for Old English”, which critiques the infrastructural and epistemological limitations of current digital tools. While playful methods like “screwmeneutics” offer new modes of exploration, we argue that without a baseline of technical reliability, such approaches risk becoming incoherent. Drawing on our lemmatization work and the broader failures of NLP systems applied to Beowulf, we call for minimalist but validated workflows grounded in transparency, documentation, and collaboration. We build on models from Dombrowski (2014), Krauwer’s BLARK (1998), and Unsworth’s ‘scholarly primitives’ (2000) to outline a sustainable vision for historical NLP. This paper will be presented in September 2025 at Hermeneutica in Practice: Honoring the Work and Legacy of Stéfan Sinclair, a conference focused on text analysis, tool-building, and critical digital humanities.