Discourse and interaction: quantitative methods

In discourse studies, the term interactional language broadly refers to language used by speakers and writers to establish and maintain interpersonal relations (e.g. first- and second-person pronouns, discourse markers, and attitude markers). In recent years, the methodologies of corpus linguistics have greatly expanded opportunities to study interaction in discourse through the analysis of relatively large quantities of electronically stored language. Corpus tools can automatically retrieve items of interest and generate quantitative data which is used as empirical evidence of how interaction is reflected in discourse. Although corpus linguistics is often associated with large-scale language research, its techniques are also useful for analyzing smaller corpora in a two-phase operation: a preliminary computerized processing phase followed by the researcher’s interpretive phase. This corpus-assisted approach is particularly effective when working with corpora of specialized language with a specific research objective in mind. This article provides a step-by-step description of how corpus methods can be applied to investigate interactional features of spoken academic discourse discourse, specifically the adjectives used by a professor of economics to express attitudes toward content during a lecture. The research phases include accessing and recording the lecture, transcribing the lecturer’s speech, preparing the data, processing and editing the data, displaying the results, and interpreting the findings. The two corpus tecniques implemented in the analysis, namely part-of-speech (POS) tagging and concordancing, are illustrated and explained. The article concludes with suggestions of sources for readers who would like to experiment with quantitative text analysis.