ABSTRACT
Since the beginning of the new millennium, the use of language corpora and of corpus tools has brought about important advances to the research and teaching of English for Specific Purposes (ESP) in general and of English for Academic Purposes (EAP) in particular. A wide variety of language corpora have been used to describe spoken and written academic genres. Among the methodologies that use corpora for linguistic analysis, we can distinguish those approaches that are considered corpus-based from those that are corpus-driven. A corpus-based approach makes use of a corpus of texts to investigate the use of linguistic features previously identified in the related literature or by the researcher's perception or intuition. A corpus-driven method, on the other hand, uses a corpus to identify linguistic features completely empirically without previous intuitions or preconceptions. This corpus-based/corpus-driven spectrum seems to be a continuum and studies can be placed along this continuum according to the methodological approaches used. At the very end of the data-driven extreme of the continuum, we can find studies of lexical bundles, groups of three or more words that frequently recur in a register (Biber, Johannson, Leech, Conrad, & Finegan, 1999). This chapter focuses on the analysis of lexical bundles in EAP, including the way in which these expressions are identified, their structures, and the functions they performed with a particular emphasis on their use in academic registers, and introduces a study of three-word lexical bundles frequently recurrent in published research articles to complement previous studies of longer corpus-driven expressions in this type of texts.
