Analyzing attitudinal stance in OpenCourseWare lectures: An experimental mixed-method approach

This paper describes a two-pronged experimental methodology for analyzing the multimodal expression of attitude in university lectures that combines techniques from corpus linguistics and multimodal discourse analysis. In the first phase, evaluative adjectives are identified in the speech transcripts of OpenCourseWare lecture videos by means of part-of-speech tagging, and then analyzed by Martin and White’s (2005) appraisal model. In the second phase, the videos are viewed to distinguish co-occurring nonverbal features (e.g., prosodic stress, gesturing, gaze direction, facial expressions, body positioning) in the form of multimodal ensembles, which are then displayed and interpreted with the aid of multimodal annotation software. The method is illustrated through the analysis of three multimodal ensembles containing the adjectives biased, (not) dismal, and unrivaled used by three lecturers to convey their attitudes towards aspects of the lecture content. The mixed-method approach enabled a systematic analysis of the multimodal expression of attitude, which would have been much less feasible through video clip observation alone. It could also be applied in instructional settings to develop strategies that promote a more holistic form of learning in which students recognize meanings in multiple semiotic forms.