Conversation analysis and talk-extrinsic data in research on speech and language therapy.

The present article addresses the issue of combining conversation analysis (CA) with talk-extrinsic data, with a specific focus on speech and language therapy and education. Even though the use of CA together with other data sources is now generally accepted, particularly in the field of 'applied CA', it is still important to discuss how this is to be done, what pitfalls to avoid, and what methodological problems might remain unresolved. The procedures from a Swedish project on interaction in speech and language intervention that combines CA with interviews and video-based retrospection sessions are used as a basis for discussion. It is stressed that the fruitfulness of combining CA with talk-extrinsic data is strongly related to what research questions are asked, and what values are ascribed to different empirical materials in relation to the overall aims of a project. The article also addresses the potential risks of a methodology such as CA when it is turned into 'just a tool' for detailed transcription. Unless continuous attention is paid to CA's origin and its use for applied purposes, it might become 'bleached', in the sense that it becomes solely a technique, and its theoretical and methodological underpinnings get forgotten.