Since the classical authors of the nineteenth century, the explanation of macro-social phenomena has been considered as the essential epistemic achievement, hence the "raison d'etre," of comparative analysis in the social sciences. In practice, however, the claims of comparative social enquiry for providing convincing explanations are not easily kept. Their realization depends upon quasi-ontological understandings of causation, and on varying conceptions of social theory. The article resumes and tackles these issues while trying to avoid the pitfalls of both positivist orthodoxy and historicist methodology. In so doing, the essay draws heavily on Max Weber's model of "singular causal explanation." It is based on a dynamic and "triadic" scheme of causal relationships--and of causal analysis--that deals in "courses" of events, and in "divergences" between alternative paths and outcomes. The model is both reinterpreted in the light of more recent epistemological debates and illustrated by examples taken from the author's own contributions to comparative social and intellectual history. (Contains 1 figure and 1 note.)