Syntactic-semantic capture of historical texts as a platform for source-critical analysis: telling the story of a premodern heresy trial with Computer-Assisted Semantic Text Modelling (CASTEMO)

Abstract Data capture from texts is both at the heart of digital humanities research and a key point of tension with established interpretive approaches within the humanities at large. The targeted extraction of information often seems at odds with more fluid and contextually interwoven ways of reading, not least the source-critical approach that historians have long emphasized. By comprehensively capturing not only text content, but also its deep contextual embedding, the close modelling of source texts as a series of syntactic-semantic data statements can help bridge this gap. This paper demonstrates the application of such an approach—Computer-Assisted Semantic Text Modelling (CASTEMO)—to the earliest surviving record of a medieval heresy trial (that of Bernard-Oth of Niort and his family), and its advantages for the inclusion of source criticism within formal data analysis. The data capture process itself is shown to provide a facilitating form of close reading: systematically reconstructing every clause of the text as a syntactic-semantic data statement reveals otherwise hard-to-see patterns in textual discourse and its context (i.e. the responses of the witnesses and the way those witnesses are characterized in the text). Subsequent data analysis of the relationships between these features serves to unravel the text’s conditions of production and even tell the human story behind the record.