The question this article tries to deal with is: how does mental health talk differ from dialogue in other public service areas? The contribution starts with a description of the organization of interpreting in health care and the level of professionalization of the field in the Netherlands, which is really not all that different from that in other western countries. Then a description and discussion of some major publications concerning interpreting in (mental) health care and their methodological backgrounds will be presented. However, only a few publications are based on a systematic investigation. Summarizing the findings from these research projects, we can conclude that the interpreter influences the content of the communication in several ways. Firstly, because of language pair dependent changes and, secondly, because of changes that are dependent on the interpreter: his1 (lack of) knowledge of the type of communication in hand; his (lack of) knowledge of the institutional setting in which the communication takes place and the specific linguistic characteristics of that type of talk; his (un)conscious personal ideas, values and norms which make him diverge from the original text. Finally, the author summarizes where we stand as far as the knowledge about interpreting in mental health is concerned, and the contribution closes with some proposals for further research. [Based on abstract in journal]