Cognitive Sociolinguistic Variation in the Old Bailey Voices Corpus: The Case for a New Concept-Led Framework

The current paper contributes to a greater understanding of concepts in the context of cognitive sociolinguistic variation. We show how concepts, which we root in thesauri-based units, co-occur with other concepts and how they vary between and within groups of speakers as represented in historical transcripts from the Old Bailey court. In order to carry out distributional sociosemantic analysis at a conceptual level we have developed a tool which combines word sense disambiguation with dependency analysis, which we refer to as Conceptual Dependency Corpus. A case study of gendered concepts from the Old Bailey Voices Corpus demonstrates that there are differences in the usage of concepts both within and between sociodemographic groupings of speakers. In particular, the co-occurrence distributions of the concepts MAN, WOMAN and CHILD vary according to a speaker's gender and role in the court. By introducing methodological innovations, we empirically elicit socioconceptual polysemy, that is different speakers using the same concept in different ways and propose a framework to analyse and interpret this phenomenon. The results also provide support for engaging computational semantics in researching cognitive and sociolinguistic phenomena.