Distant reading, a digital humanities method in wide use, involves processing and analyzing a large amount of text through computer programs. In treating texts as data, these methods can highlight trends in diction, themes, and linguistic patterns that individual readers may miss or critical traditions may obscure. Though several scholars have undertaken projects using topic models and text mining on Middle English texts, the nonstandard orthography of Middle English makes this process more challenging than for our counterparts in later literature.

This collaborative project uses Gower’s Confessio Amantis as a small, fixed corpus for analysis. We employ natural language processing to reexamine the Confessio’s themes, adding data analysis to the more traditional close reading strategies of Gower scholarship. We use Gower’s work as a case study both to help reduce the potential variants across textual versions and to more deeply investigate the corpus than distant reading normally allows.

Here, we share our initial findings as well as our methodologies. We hope to share resources that will allow other scholars to engage in similar types of projects.