Sociology now finds itself in a position where it can make major contributions to the analysis of the production and control of risks. The breakdown of the Great Western Paradigm founded by Descartes has led to the development of new theories in many areas of knowledge. This paper seeks to locate the space that might be available for sociological theory in the development of the reconstruction of knowledge about industrial safety and health. At the macro level some recent work on risk by Giddens and Beck will be reviewed. These authors virtually ignore most sociological work carried out on the actual production, which occurs in workplaces, of the very risks that lead Beck to talk of risk society. Much sociological work has been carried out within three traditions: functionalist, Marxist and phenomenological. Work carried out in each of these traditions makes different types of contributions to understanding the production of risks. No single perspective has achieved dominance. Form within the walls of the disciplines that have traditionally dealt with the control of risks: safety engineering, industrial medicine, psychology and ergonomics, new approaches to the analysis and control of risks are sought, some talk of the necessity for a new paradigm. Here much reflection points towards the development of understandings of work and people at work that are compatible with the phenomenological tradition and the sociology of work. Sociology now finds itself in a position where it can make important contributions to the analysis of the production and control of risks. Three particular areas are identified: the diagnosis of cause, the development of a clinical role and, in a more speculative vein, a theoretically and methodologically sophisticated sociology of risk may result in the development of predictive models. Thus interesting prospects for theory development and, probably, new job markets for sociology graduates are opened up. Introduction Sociology now finds itself in a position where it can make major contributions to the analysis of the production and control of risks. Fundamental in this area have been a series of notions from the sociology of work, they play important roles in diagnosing cause, and in the future they promise to play a clinical role complementing that of technicians and, who knows, if the science comes right sociologists may be tempted to develop predictive models. This perspective opens up very interesting prospects for theory development and, probably, a promising job market for sociology graduates.