Collaborative creation of spoken language corpora

Abstract

Analysing authentic interactions at progressively greater levels of complexity is one means of promoting deeper engagement with pragmatic phenomena amongst L2 learners. However, effective analysis often requires a greater amount of data than learners can feasibly gather. It is proposed here that encouraging students to collaborate through the creation of a corpus of spoken interactions is one potentially effective way to help them engage with a much richer set of interactional data than they might normally encounter. Here we report on a corpus created through “crowdsourcing” the collection and transcription of recordings of spoken interactions, the Griffith Corpus of Spoken Australian English (GCSAusE), which was then made available to L1 and L2 students to use in analysing pragmatic aspects of spoken interaction. In this way, the students had the opportunity to be both creators and users of the corpus, and see how it results in the real and ongoing accumulation of knowledge about language use. The degree of engagement of students with the corpus was assessed through their research projects, a written survey, and a focus group conducted with a number of students who took the course.