In comparative politics, explanations should join social mechanisms and empirical analysis. The more reason to accept the explanatory value of the social mechanisms and the greater the validity of the empirical work presented, the more reason there is to accept the explanation. This general statement modifies the opening line of my chapter in the book's first edition, “In comparative politics, theory seeks explanations for sets of political phenomena” (Zuckerman 1997: 278). Here, I replace explanation with social mechanisms for those with covering laws and simple causal accounts, and I also insist on the need for rigorous empirical work in substantiating explanatory accounts. The change indicates both more modest and more demanding requirements for scientific understanding in comparative politics.

The chapter in the first edition argues for the need to develop new standards for explanation in comparative politics. Drawing on work in the philosophy of science, in that chapter I displayed flaws in the subfield's two most widely accepted modes of explanation–covering law (also called “deductive-nomological” or “hypethetico-deductive”) and causal explanations. Some of these weaknesses derive from the formal properties of these modes of explanation, and these are well known to philosophers of science even if they are less well understood by comparativists. Others, I suggested, result from the clash between the simple ontology of linear and frequently determined relationships, which accompanies each of these explanatory tropes, and the nature of the political world, which includes endogenous, nonlinear, and stochastic relationships.