The paper discusses Raymond Boudon’s theory of ordinary rationality as well as his assessment of so-called “analytical sociology.” On the first point, the paper argues that, in order to combine the realism of an unconstrained conception of rationality with the ex-ante facto predictive power of the narrow version of rational choice theory, we should better understand the relation between potentially triggering events and the actor’s “reasons.” Heuristics, social identity, and emotions may help build a broader rational choice theory. On the second point, the paper recalls some factual elements suggesting that Boudon’s assessment of analytical sociology is excessively severe and explains why some analytical sociologists are right in devoting special attention to computational modeling.