What is Context?

The notion of context should not be understood as descriptive but as analytical, in the sense that the package of features characterizing a context is dependent on the question the sociologist wants to solve. Also, the features evoked in a context should be empirically observable. This excludes introducing unobservable dispositional features in the context: a principle Weber, Durkheim and many of their modern followers endorse. The fathers of sociology have practiced since long contextual analysis in this sense. Examples drawn from their work illustrate the powerfulness of their explanation of numerous macroscopic puzzles. They illustrate how to solve the micro-macro link question. They suggest that the solution of this question is dependent on the nature of the macroscopic facts to be explained. The paper discusses the question indirectly through the accumulation of examples illustrating the powerfulness of contextual analysis. A major point of the article is also that social action includes always beliefs. Except in the cases where explaining the beliefs raises no question, it is a challenging point for social analysts. Rational Choice Theory is efficient in the cases where beliefs raise no question. If they do, Rational Choice Theory is doomed to introduce controversial notions as frame or bias that do not correspond to any observable reality. Because of its instrumental view on rationality, Rational Choice Theory is also unable to explain the goals different categories of individuals follow, while contextual analysis can, thanks to its broader conception of rationality.