The sociological analyses of the technological accidents in the last years, such as Diane Vaughan's work, emphasize how important is it to take into account not only the technological, but also the organizational dimensions of these events. That means that sociologists have to enter areas of production in which the engineers are the only ones professionals engaged in risk analysis and managing risk by technical means. As James F; Short explained it in 1984, "perhaps the chief risk for sociology, if we do not make a major effort to engage the discipline more fully in this area, is that we will increasingly be seen as irrelevant to an area of extreme importance to all of human kind - an area in which all the major institutions, all societies, and all nations are daily affected by analyses and decisions which remain largely uninformed by sociological theory and analyses." The aim of this paper, grounded on an empirical research in French chemical enterprises (Dupré/Le Coze, 2006), is to debate about the difficulty to enter as sociologists such enterprises where the knowledge recognized as valuable is "driven by scientific and technical considerations based on physical and biological sciences and applied mathematics." It will also question the social theory as long as it is unable to build a bridge between technology and social actions, or between physical and social sciences.
