Organizational learning and industrial change: the example of robotization of automotive sheet metal work.

Summary For a long time, technical progress was associated with economic progress. However, in recent years, it has become apparent that the efficiency of production systems also depends on the adequacy of organizational models to new technical constraints. We are interested in a particular technical breakthrough: the robotization of automotive sheet metal units. Formalizing an industrial mutation implies a theory of organizational change. We have chosen an interactionist approach to organizational change, as the evolutions take place in a context of incomplete knowledge and imperfect information. We will apply the notion of organizational learning to the case under study and explore this concept further. The thesis is based, on the one hand, on an intervention research in a sheet metal factory and, on the other hand, on a comparison of the evolutions of five other French sheet metal factories recently robotized. Several aspects are examined: the evolution of technical systems, structures, skills, instrumentation and performances. The results obtained in the thesis can be articulated in three main points: the perspective of the learning theories presented in the literature, the characterization of a learning process, the construction of three ideal-types of learning. In order to clarify the existing definitions of learning, we insist on the notion of learning perimeter which allows us to identify the actors involved in the construction and memorization of new knowledge. We also emphasize the organizational dimension of evaluation, which often relies on management instruments that have already been established. Our analysis grid of the learning processes includes three dimensions: organizational deployment, spatial dimension, and temporal dimension, each of them gathering different variables specified in the thesis. From the comparison of the itineraries of different units that have experienced the same technological leap, we identify three ideal types of learning. We evaluate these three types of learning from the results of the sheet metal works whose dynamics are similar to a pure type. The conditions under which each of the types is adapted will be emphasized.