Please refer to published version for the most recent bibliographic citation information. If a published version is known of, the repository item page linked to above, will contain details on accessing it.

This article addresses the question of how practices perform power effects within and across communities of practice. It does so by drawing on a study of two medical innovation projects leading to radical changes of practice. Situated learning theory has to some degree acknowledged the asymmetry in power between masters and apprentices. Meanwhile, this study suggests that the novelty of new practices may lead to a contestation of the established master-apprentice relationship and even challenge the basis of the community of practice itself. We therefore argue that innovation processes may highlight the political processes and negotiations already at play in communities of practice. Hence, we investigate how communities of practice tried to control the new practices through mobilizing arguments, marginalizing opponents and building alliances. Consequently, the article argues that changing practices may be highly political.

This paper addresses the challenges that arise when knowledge production occurs in crossdisciplinary settings. To date most studies on communities of practice have focused on knowledge production within communities of practice rather than across communities of practice. We analyse the various professional groups in a medical R&D department as a constellation of distinct, but interconnected communities of practice with different epistemic cultures. The medical R&D case is particularly interesting for this purpose, because it involved creating new cross-disciplinary practices between different pre-existing and well-established communities of practice. In line with our focus on the challenges and processes involved in cross-disciplinary knowledge production, we describe negotiations and tensions during the establishment of the department, as well as in day-to-day practice. In particular, we focus on how the 'machineries of knowledge production', that is, the actual mechanisms by which knowledge is pursued, are different across the various communities of practice. These machineries belong to different epistemic cultures on a national or even international scale, and thus every community of practice is part of a complex web of people, activities and material structures extending well beyond the immediate work context. This networked character of knowing in practice explain why learning on the system level of communities of practice can be challenging. It may lead to path-dependent learning processes, and radical change can become limited if the knowledge required by new and different practices is incompatible with the existing stock of knowledge. Consequently, we suggest that the communities of practice approach could be enriched by looking at diversity and discontinuity in the epistemic cultures and networks that the different communities of practice are associated with.