Identifies some underlying assumptions about language, testing and its use, and second language acquisition (SLA) for those current on the issue of oral proficiency assessment. Suggests that even with its limitations, the assessment procedure can become a source of data for future SLA research. (Author/LMO)

In linguistic analysis of the speech act, the data used to support theoretical conclusions are too often comprised of semantically isolated utterances of the ideal speaker-hearer. In reality, one of the most revealing kinds of data is imperfect data, where the functioning language processes are often unmasked. The study of first language…