Description: It is not rare to employ interpreters in research abroad or in the context of development promotion in southern countries. These intermediaries are at the heart of the relationship between the "expert" trying to get or transmit knowledge, and the target "lay" population. Recent researches (especially in the field of healthcare) show how difficult it is for professionals to manage such a complex situation. They usually feel powerless for a double reason: loss of control on the content of the interaction and loss of control of the interaction itself. The outcome is uncertainty about the knowledge transmitted (either way). Scientific literature is very quiet on this matter, even more when it comes to understand who these go-betweens are, what their background is, how they do their work, and how they feel about it. This paper presents a research project parallel to another one. While enrolled to explore the socio-cultural context of emergence of the disease noma (a gangrenous condition, which starts in the mouth as a benign oral lesion and rapidly destroys both soft and hard tissues of the mouth and face of children) in the Zinder area (Niger), the principal investigator experienced this double powerlessness and wished to understand more about the processes going on in interpreted research and development work. How do interpreters position themselves in the interaction? As neutral helpers or active mediators? What competencies do they develop? What kind of relationship do they keep with "expert" and "lay" knowledge? Methodology: Interpreters were recruited with an announcement through a local radio. Eleven out of 38 people were selected, on the criterion to have at least once been employed as an interpreter. All were males, from 31 to 50. They all speak French and Haussa; nine speak also English, 2 Arabic and 4 an other local language. They have an extended educational background: all went to university (even if they didn't all finish their degree). They worked as interpreters mainly for foreign teams (non governmental organisations) on development promotion projects, 3 of them on research projects with ethnologists. Usually the target population is from villages in rural areas. A first series of semi-structured interviews were done with the 11. After preliminary analysis of the data, 7 interpreters accepted a second interview. This second interview helped specify some of the situations mentioned in the first interview and go deeper into some aspects of their work. Conclusions: Interpreters are not passive transmitters. They play active roles in the construction of the relationship and in the construction of a "politically correct" knowledge for the "lay" population as well as for foreign "experts". Leaving their different roles unquestioned may have a significant impact on development promotion and research activities. The findings of this research could also provide clues to understand the power play in health education work with immigrant communities or research in immigrant communities in Europe and the role of interpreters in these dynamics (as well as what it means to work without interpreters…).