One of the ways in which language testing interfaces with applied linguistics is in the definition and validation of the constructs that underlie language tests. When language testers and score users interpret scores on a test, they do so by implicit and explicit reference to the construct on which the test is based. Equally, when applied to new understandings of the way people use language, the challenge for language testers is to develop tests that embody these new understandings. This paper puts forward a relatively new theory of spoken language use in face-to-face communication--"interactional competence"--and discusses the implications of the theory for the design and interpretation of tests of oral communication. The theory of interactional competence seeks to explain the variation in an individual speaker's performance from one discursive practice to another. It is concluded that in validly assessing interactional competence, close attention must be paid to exactly what happens during performance assessments so that it is known exactly what configuration of interactional resources apply to the practice of assessment. It is also concluded that such testing provides a principled way of generalizing from performance in the discursive practice of a performance assessment to performance in other non-test contexts. Second language acquisition researchers must pay careful attention to the contexts in which they elicit interlanguage data. (Contains 32 references.) (KFT)