Church interpreting - i.e. oral mediation of the sermon into another language than the language used by the preacher - is a routine practice in the growing sector of Evangelical and Pentecostal/ Charismatic churches in Germany and a powerful instrument of their global mission culture. Interestingly, little is known about the emerging Evangelical subculture in Germany and Evangelical practices of language mediation are, too, significantly under-researched in terms of their cultural, performative and linguistic characteristics. Research on the dynamic appearance of new groups in the German religious landscape necessitates an explorative approach and resists methodologies where empirical data just serves as a token to confirm or refute established assumptions about a particular group. Doing justice to these emerging environments means choosing designs that recognise the normative arrangements in interaction (Heritage, 2004) but are also flexible enough to treat context and identity as “locally produced, incrementally developed and [...] transformable at any moment“ (Ten Have, 1999, p. 174). While subtle situational dynamics in the micro-context of interaction are excellently traceable by applying conversation analytical approaches to interpreted interaction, the CA paradigm is, however, clearly suspicious of using larger sociological or ethnographic categories or including data and context knowledge stemming from beyond the interaction itself. Yet, factoring out context knowledge seems neither feasible nor desirable if one seeks to account for the reasons and parameters that lead to the occurrence of a setting-specific linguistic behaviour and to delineate a social practice against its institutional macro-context. In this paper, I explore the possibilities and limits of combining aspects of the interactionist CA paradigm with the action theoretical approach of Grounded Theory and illustrate avenues for integration with sequences from my data.