Date created: | Last Updated:

Category: Project

Description: The writing of Science, Technology, Engineering, and Mathematics (STEM) research papers is highly structured. Numerous discourse analysis studies have investigated this structure within and across disciplines, with various models proposed to describe its features. Traditional corpus-based methods can also be usefully applied in the understanding of STEM writing, but they tend to target bottom-up, sentence level features of language use rather than top-down discourse level features. In the presentation, I will introduce some important innovations in corpus methodology that allow for the large-scale analysis of STEM discourse features, including tools for the automatic annotation of research paper section information as well as move and step divisions. I will also discuss how artificial intelligence (AI) methods can be integrated with corpus methods to provide novel insights on discourse-level language use and suggest future directions for tool development in this area.