The current thesis is concerned with constructions and perceptions of what it means to be ‘a competent language user’ in the context of a language programme in basic Swedish called Swedish for Immigrants (SFI). A particular focus is given to the testing and assessment of oral interaction. The prevailing communicative approach to language teaching and testing makes it relevant to investigations of both language use and reflections on communicative experiences. The thesis is based on three studies. Drawing on insights from linguistic anthropology, multimodal interaction analysis, phenomenology and social theory, the three studies address different sociolinguistic perspectives on language testing and assessment. Whereas Studies I and II investigate paired speaking tests in the final national exam in SFI as a speech event, Study III builds upon focus group discussions with SFI participants with the aim of exploring the participants’ reflections on communicative experiences.

       Drawing on linguistic anthropological performance theory, Study I makes the case that the paired speaking tests can be analysed as staged institutionalized performances that put speaking and ideologies on display. Study I draws on an analysis of sequences in the test data where the participants expressed beliefs on language learning, language use and language competence. One important resource for the test takers to maintain the discussion in front of the examiners was to draw on dominant discourses on language and integration, such as stating the importance of learning Swedish, speaking only Swedish, attempting to find Swedish friends and taking responsibility for one’s learning, making testing practices an important site for the reproduction of such discourse. The orientation to being ‘a competent language user’ was performed by indexing other images of being ‘a good student’ and ‘a good immigrant’.

      Study II takes an interactional practice in the paired speaking tests, word searching sequences, as its starting point. Word searches tap into two aspects of communicative language testing: vocabulary knowledge and the ability to negotiate meaning and solve interactional problems. The test takers drew on different embodied semiotic resources to negotiate participation and meaning or to display an avoidance to participating in the fellow test taker’s word search. Overall, the participants prioritized the progressivity of talk over lexical precision. By avoiding using languages other than Swedish during the test, the test takers sustained and constructed a monolingual orientation to language competence.

     Study III discusses how the SFI participants’ lived experience of language constituted their understanding of what it means to be ‘a competent language user’. Accordingly, the participants’ comments primarily constructed a view of competence as made relevant through and being shaped in social interactions, making language competence a primarily relational construct. Corroborating the relational construction of language competence was the importance given to language assessments, both those made by others and internalized self-assessments. In the focus group discussions, overall, being ‘a competent language user’ was oriented to as a desired, but yet unstable and vulnerable subject position.

      Taken together, the three studies address ideological, embodied, emotional and relational perspectives on language and language competence. By contrast, language testing practices are built upon a view of language competence as an individual and objective ability that can be measured. The main conclusions drawn in the thesis are that testing and assessment practices constitute a social practice where perceptions and constructions of language competence are constructed and regimented metapragmatically as well as interactionally. Furthermore, embodied experiences of language assessment made in institutional and everyday practices make competence a powerful concept influencing L2 users’ self-perception and agency.