NO ˆ US 51:3 (2017) 484–520 doi: 10.1111/nous.12202 Explanation and Manipulation 1 ALEXANDER PRESCOTT-COUCH Massachusetts Institute of Technology Abstract I argue that manipulationist theories of causation fail as accounts of causal struc- ture, and thereby as theories of “actual causation” and causal explanation. I focus on two kinds of problem cases, which I call “Perceived Abnormality Cases” and “Ontological Dependence Cases.” The cases illustrate that basic facts about so- cial systems—that individuals are sensitive to perceived abnormal conditions and that certain actions metaphysically depend on institutional rules—pose a challenge for manipulationist theories and for counterfactual theories more generally. I then show how law-based accounts of causal structure can answer such challenges. The moral of the story is that the basic manipulationist idea that our interest in causal structure is driven by our interest in manipulating our environment faces decisive problems in a central domain of application, the social sciences. 1. Introduction Good scientific theorizing tells us more than what happens; it also tells us why things happen. Science does not only describe; it also explains. Accounts of scientific explanation seek to elucidate these truisms. Traditionally, an account of the explanatory/non-explanatory distinction was thought complete if it successfully answered the question “what do we want to know when we ask ‘why p?’” A successful account need only inform us of the sort of information explanations provide and the form that information takes. For example, take the deductive-nomological theory– the view that explaining p consists in providing a sound deductive argument for p one of whose essential premises is a law. While the DN-account is false, 2 it is at least ostensibly complete. It answered, one might think, all the questions a theory of explanation should answer. But did it? Say that every explanation did in fact consist in a sound deductive argument one of whose essential premises is a law. One might wonder—why care about that? If we cannot discover a reason, then even if we’ve accurately captured an aspect of scientific practice, we would lack an understanding of its rationale. Even worse, in trying to understand this important aspect of scientific activity, we might have accidently undermined it by showing that perhaps we have little rea- son to care about it after all. Of course, undermining certain explanations may be desirable, since there are many pseudo-explanations lurking around. But if we can’t come up with a rationale for good explanation, what basis do we have for C  2017 Wiley Periodicals, Inc. 484