ABSTRACT
A corpus can be defined as a collection of machine-readable authentic texts (including transcripts of spoken data) that is sampled to be representative of a particular natural language or language variety (McEnery et al. 2006: 5), though “representativeness” is a fluid concept (see Section 7.3). Corpora play an essential role in natural language processing (NLP) research as well as a wide range of linguistic investigations. They provide a material basis and a test bed for building NLP systems. On the other hand, NLP research has contributed substantially to corpus development (seeDipper 2008 for a discussion of the relationship between corpus linguistics and computational linguistics), especially in corpus annotation, for example, part-of-speech tagging (see Chapter 10), syntactic parsing (see Chapters 8 and 11), semantic tagging (see Chapters 5 and 14), as well as the alignment of parallel corpora (see Chapter 16).
