Prof. Dennis TAY, Department of English and Communication

Data analytics and the modelability of metaphor in discourse. The Fourth Cultural Linguistics International Conference (CLIC) 2023. College of International Studies, Southwest University, Chongqing, China, 22-24 September 2023.

Abstract
Various kinds of comparative metaphor analysis offer insight into the relationships between language, cognition, and cultural conceptualizations. It is helpful to consider how such analysis can leverage the rapid growth of data analytic techniques and provide more nuanced outcomes for theory, pedagogy, and application. In this talk, I discuss some of my recent work on time series models, association rules mining, and survival analysis as a suite of data analytic tools to model metaphor usage patterns across key comparative parameters (e.g., cultures, individuals, and time). These widely used tools in finance, health, and related fields have understated potential for metaphor analysis. I show how they address questions like the randomness vs. predictability of source domain choices across time, the dis/associative tendencies of different source domains as profiles of individual users, and factors determining how well metaphors ‘survive’ in spontaneous interaction. A key point that cuts across these techniques is the ‘modelablity’ of metaphors in discourse domains ranging from news translation to psychological counseling. Modelability refers to the extent to which metaphors (and other phenomena) are describable and interpretable from quantitative data analytic angles. We could view it as a characteristic property of said phenomena in discourse, as well as an empirical condition for making theoretical generalizations in cognitive and cultural linguistics.