This reply to Erik Weber's commentary agrees that mechanisms are important for causal inference in social science, but argues that Weber makes the mistake that was the main focus of my original essay: inferring that since a problem cannot be solved without mechanisms, it can be solved with them. As it stands, this inference is invalid since the problem might be unsolvable with or without mechanisms. Any claim about the usefulness of mechanisms for some purpose requires an adequate account of how mechanisms can actually fulfill that function, which Weber has not provided with regard to the issues he discusses.
This is a forthcoming book entitled Across the Boundaries: Extrapolation in the Biological and Social Sciences.
2.
The claim that causal inference based on mechanisms is fundamentally distinct from causal inference from statistical data is, however, clearly expressed by George and Bennett (2005, 207).
3.
The only publication devoted to this topic that I know of is Danks (2005).
4.
Danks, David.2005. Scientific coherence and the fusion of experimental results . British Journal for the Philosophy of Science56: 791-808.
5.
George, Alexander, and Andrew Bennett.2005. Case studies and theory development in the social sciences . Cambridge, MA: MIT Press.
