Agentivity as a determinant of lexico-grammatical variation in L2 academic writing

This paper examines novice writers’ strategies in the (non-)representation of authorship in academic writing drawing on data from the Corpus of Academic Learner English and a native-speaker control corpus. The analysis focuses on the quantitative and qualitative use of pronouns, subject placeholders, as well as verbs and inanimate nouns that frequently occur in academic writing. The findings indicate that even advanced learners are insecure about the (non-)representation of authorship in academic texts, but lack the resources to report events and findings without mentioning an author-agent. The learner data evidence a significant overrepresentation of first person pronouns and subject placeholders as default strategies to suppress the author-agent. This imbalanced clustering is argued to be due to a significant underrepresentation of constructions with inanimate nouns as subjects that are preferred reporting devices in abstracts and research articles in the humanities. The paper concludes by addressing implications for language teaching, testing and assessment.