L'évaluation rétrospective des risques collectifs. Le cas de l'impact sanitaire des essais nucléaires. – RETRORISK

The sociology of collective risks has undergone an important renewal for the past fifteen years. Differing from approaches which considered the perception of risks, the most innovating works in this field have regarded collective risks according to a larger point of view: that of public problems whose emergence, framework and institutional treatment need to be questioned. These works notably attempted to analyze the changes induced by the acknowledgement of the scientific uncertainties regarding certain threats. This research project lies within this renewed context of the sociology of risk. However the goal is to open a new path to research while taking an interest in situations of uncertainty which are not produced by a threatening future but by hazards belonging to the past. Its purpose is to analyse different retrospective assessments allowing one –in these situations– to ‘colonize the past’ and to examine the plans conceived to come up against these particular situations of uncertainty. The empirical case chosen to address the problems and stakes of the retrospective assessment of risks is that of health hazards in nuclear trials. This matter is a subject to controversy in France as well as in other countries such as the United Kingdom and the USA . To analyse these controversies, our research strategy relies on the use of three theoretical and methodological tools: the first is the mobilisation of analogical reasoning which consists in relying on the analysis of situations of uncertainty related to the future to question other situations associated to the past; the second, with respect of a symmetry principle, compels one to treat both scientific risk assessments and other types of assessments (such as testimonies and court trials) in the same way; eventually, the third considers international comparison which has to make it possible to explain the various plans that have been devised to deal with these controversies. The outcome expected from this research is notably to initiate general thinking about a fundamental problem which –up to now, has only but partially been treated in literature dedicated to risks: that of temporality and the way doers invest time in these areas