La théorie poppérienne de la confirmation scientifique

Abstract

The epistemological problem of scientific confirmation is about determining or comparing the degree of confirmation of scientific hypotheses (i.e., the level of epistemical value we can give them) with respect to empirical evidence. This problem is mainly discussed from an inductivist and probabilistic point of view, particularly by the means of Bayesian theories of confirmation. In this paper we suggest that Popper’s theory of corroboration should be considered as an alternative theory of scientific confirmation, which would reject both induction and probabilities. More specifically, in opposition to some interpretations of Popper, this paper will support the idea that Popper’s concept of corroboration can be understood as a form of confirmation. We will try to establish what could be the rules of this kind of Popperian theory of confirmation. We hope to be able to compare Bayesian and Popperian approaches to the problem of scientific confirmation, and highlight their different methodological contributions to scientific research.