Data Collection

This chapter discusses the various ways that data can be collected for linguistic analysis. It begins by discussing “introspection,” a method of data collection that was common in generative approaches to linguistic analysis and involves individual judgments about the acceptability or grammaticality of sentences. It goes on to discuss more “empirically” based ways of obtaining data. For instance, in elicitation experiments, individuals are asked directly to make judgments about the grammaticality of individual instances. In the corpus‐based approach, large linguistic databases (“corpora”) of different types of spoken and/or written language are compiled and investigated to study how language is actually used. A central issue in this section is “representativeness”: how can we ensure that a corpus is maximally representative of the language or sub‐language that it samples? The chapter concludes by discussing the pros and cons of the recent use of the Internet as a source of linguistic data.