ABSTRACT
This chapter discusses the evolution of the use of corpus-based methods from traditional learner corpus research towards more theory-informed and methodologically more sophisticated second language acquisition work within cognitive/usage-based linguistics. The chapter outlines some of the main challenges that usage-based research towards communicative competence is currently trying to overcome, exemplifies corpus methods on both a coarse- and a fine-grained level, and offers a few recommendations for current practice. It concludes by discussing desiderata on three levels: data (e.g., the need for more and better [annotated] data), theory (e.g., the recognition of the important role played by notions such as dispersion and entropy), and statistical analysis (e.g., the adoption of more powerful statistical tools, especially for observational data).
