Investigating grammatical complexity in L2 English writing research: Linguistic description versus predictive measurement

We argue in the present paper that the (socio)linguistic description of grammatical complexity provides a necessary complement to predictive omnibus measures as an analytical approach for the study of student writing proficiency and development. That is, while omnibus measures can be effective for predicting student performance, we argue that a comprehensive grammatical description is required to fully understand and interpret the linguistic characteristics of written texts produced by students. The logic of our argument is simple: Descriptions of grammatical complexity in student writing must be based on linguistically-interpretable analyses of grammar (including syntactic differences). We develop this argument from several perspectives, including a survey of the structural/syntactic features relating to the construct of grammatical complexity in English, an overview of corpus-based research showing that these distinctions truly matter for the description of academic writing, and a critical evaluation of the descriptive adequacy of omnibus measures when considered from this linguistic perspective. In summary, although we recognize the utility of omnibus complexity measures for purely predictive purposes (e.g., to assess L2 writing proficiency), we argue that a comprehensive linguistic description of grammatical structures and uses is required to fully understand the characteristics of student texts and the nature of student writing development.