Abstract
the sense of observation is not only useful but indispensable in sociology. This capacity is first and foremost a practice of ordinary knowledge, but it can be refined and improved. French sociology has recently come closer to ethnography and has reversed observation in the light of criticisms about the validity and alleged universalism of classic formulas. Of the four main sociological methodologies of questionnaire surveys, interviews, study of documents and sociological observation, the first three have their history and their instrumental inventory is well described, but not the observation. However, in sociology, several observation currents have cohabited. The article focuses on tracing them, detailing their conditions and rehabilitating direct or indirect observation, spontaneous or prepared, presenting or slightly retrospective observation as an essential tool for understanding social worlds far from its own, all those which are not closely known through direct contact.