**Comprendre la physique quantique : concepts et conceptualisation dans la pratique de la science**

An interpretation of a physical theory indicates what the mathematical apparatus of the theory is about. It so happens that classical physical theories can be interpreted as a representation of causal processes embedded in space and time that constitute the domain of application of the theory. Orthodox quantum theory, which happens to be one of our most experimentally adequate theories, is notoriously resistant to being interpreted as a representation of causal processes in space and time.

This situation has given rise to roughly two different interpretive views of quantum theory. One response is to adopt a quasi-instrumentalist point of view. The formalism of quantum theory is taken to represent our knowledge of phenomena, rather than phenomena. Many physicists and philosophers do not content themselves with this instrumentalist point of view. If, like most philosophers, we endorse scientific realism, to interpret a theory properly is to say what the world is like if the theory is true. This line of thought has found its climax in the so-called "experimental metaphysics" that developed after the experimental violation of Bell inequalities was observed.

The true meaning of the violation of Bell inequalities has been the subject of much philosophical investigation. It is commonly held that this discussion has reached a satisfactory conclusion. It is widely believed that Bell-type experiments force us to accept a form of non-locality (or action at a distance): the world cannot be accurately described by any theory that gives a representation of the phenomena as local and perhaps causal processes in space and time.

It is a striking feature of the Bell literature that locality is defined in terms of restrictions on conditional probability distributions over events. Arguably, a definition of locality should be intimately related to the space-time structure that events are embedded in. Prima facie, restrictions on conditional probability distributions do not alone indicate how events should be embedded within a space-time structure. The focus on probability distributions seems to be a remnant of Reichenbach's analysis of causation in terms of probability distributions and a challengeable association of local and causal processes.

In my dissertation, I examine the relationship between conditional probability distributions for events, which some, following Reichenbach, have taken as representative of causal relationships, and locality as defined by the causal structure of space-time, from the perspective of the experimental violation of the Bell inequalities. I challenge the construal of conditional probability distributions as mirroring objective (causal) relationships in space-time. I suggest that the Reichenbach-style analysis of causation should be conceived as an ideal for understanding of the phenomena, rather than an objective description of causal processes in space-time. Thus, the standard interpretation of the violations of the Bell inequalities is challenged. I shall defend a pragmatic and perspectival conception of probabilities and their relationship to space-time structure, which leads to a new understanding of quantum probabilities.