Corpora and phonological analysis

Abstract

AbstractThis chapter discusses the use of corpora in phonological analysis, both historically and with an eye to the future. A corpus can be generally defined as any collection of linguistic data, such that most empirical phonology is in some sense ‘corpus-based’ phonology, but the term is most commonly associated with work that involves answering a research question using pre-existing data of an appropriate type. Although corpora vary widely, they offer a number of advantages for phonologists. In particular, they often provide large, accessible bodies of relatively naturalistic data on which many different types of analysis can be run and replicated, and lend themselves particularly to statistical investigations of frequency, probability, and variation in phonology. This chapter discusses both advantages and limitations of corpora, providing examples of how they can and have been used for doing phonological analysis as well as information on tools and resources available for conducting such work.