Representing culture in OpenCourseWare Lectures: A Corpus-based Semantic Analysis

This paper explores the representation of culture in a corpus based on fifteen OpenCourseWare (OCW) lecture transcripts available from the Massachusetts Institute of Technology and Yale University. The corpus is divided into three components of five lectures each from the broad disciplinary areas of humanities, soft sciences, and hard sciences, thus allowing for a contrastive analysis of cultural references across different domains of specialized knowledge. The analytical approach is based on the concept of a “cultureme”, i.e., a unit of analysis for a culture-specific phenomenon and its linguistic expression. The corpus was processed with the semantic annotation tool of Wmatrix that automatically assigns lexical items in a corpus to pre-established semantic domains. The lexical items in domains associated with human cultural experiences (e.g., education, religion, history, food and drinks, sports, the media, entertainment, geographical names, proper names) were then examined to identify culturemes. Extensive follow-up cross-domain analysis was necessary to tease out culture-specific meanings across the corpus. The results indicate that, although not present in high frequencies, a number of culturemes were dispersed throughout the corpus, with most referring to the domain of education, followed by government & politics and entertainment. The paper concludes with some reflections on the pedagogical implications of the findings in the context of helping L2 learners cope with the comprehension challenges of culture-specific meanings in lecture discourse.