The aim of this study is to demonstrate the usefulness of corpus linguistics as a methodology in grammar studies. By analysing naturally occurring spoken and written samples from various corpora, the author examines the distribution of subjects across the semantic domains and information structure of adverbial clauses and across text types.