Self-recordings & oral histories

The method by which a linguist elicits language data is a key dimension of linguistic variation. Traditional variationist methods (Labov 1966; 1972) focus on the differences between the various tasks that comprise a sociolinguistic interview, such as comparing non-scripted, interview speech to scripted, read speech, or comparing read speech from a passage to read speech from a word list. Work that looks at variation according to topic, stance, or other discursive elements reveals further dimensions of variation that can occur within an interview. In this chapter, we focus instead on the linguistic variation observed between a setting where the primary purpose is linguistic research and a setting where the primary purpose is something else. We argue that, despite the range of styles that can be elicited and found between different interview tasks and discourse structures, there is nonetheless an overall reduction in stylistic variation inherent to speech collected in a context where the primary purpose is linguistic research. This suggests that our studies are more or less founded on speech samples representing a reduced stylistic repertoire, which impacts the descriptive adequacy of our research and the scope of our theoretical conclusions.