Knobe's general explanation of the way in which moral considerations influence intuitive judgments goes like this: In judging causation, doing/allowing, intentional action, and so on, people select alternative possibilities to compare with what actually happens and their selection of these possibilities is influenced by their moral judgments. How does this idea explain the data about people's causal judgments? Unfortunately, Knobe offers only the briefest hint in his Note 5, which suggests that moral considerations affect people's causal judgments by influencing which counterfactuals of the form “If event c had not occurred, event e would not have occurred” they regard as true. This suggested explanation doesn't work, however, for his own example in which Professor Smith's action rather than the administrative assistant's is regarded as the cause of a problem. This difference is not reflected in any difference in the counterfactuals people regard as true, since it is true that there wouldn't have been a problem if either Professor Smith or the administrative assistant hadn't taken a pen.

Luckily, Hitchcock and Knobe (Reference Hitchcock and Knobe2009) provide the missing elements of the explanation. Hitchcock and Knobe appeal to the finding in the literature on counterfactual availability that people are very inclined to entertain counterfactual hypotheses about what would have happened if a normal event had occurred instead of an abnormal one; and, by contrast, they are much less inclined to entertain counterfactual hypotheses in which normal events are replaced by abnormal ones. So people are willing to entertain the counterfactual about what would have happened if Professor Smith hadn't taken a pen because it “mutates” an abnormal event into a normal event. By contrast, people are less willing to entertain the corresponding counterfactual about the administrative assistant's action because it does not involve the privileged kind of “mutation.” Finally, by positing that people's willingness to make a causal judgment “c caused e” goes hand-in-hand with their willingness to entertain the counterfactual “If c had not occurred, e would not have occurred,” they explain why people are more inclined to regard Professor Smith as the cause of the problem.

I suspect this explanation cannot be right for two reasons. The first is that the explanation involves an uneconomical hypothesis about the capacities involved in causal cognition. The explanation implies that people have an underlying competence for understanding counterfactuals that is linked to their understanding the objective core of the causal concept (the “causal structure” in Hitchcock & Knobe Reference Hitchcock and Knobe2009). This competence is exercised when people understand counterfactuals of all kinds, including the counterfactuals about Professor Smith and the administrative assistant. Sitting alongside this competence, the explanation implies, is a psychological tendency to entertain some counterfactuals as “available,” a tendency aligned to people's propensity to select certain events as salient causes. This hypothesis strikes me as implausible because of its doubling up of capacities involved in causal cognition.

My second reason for suspecting that this explanation can't be correct is that empirical evidence casts doubt on the assumption that people's causal judgments depend on their counterfactual judgments. Mandel and Lehman (Reference Mandel and Lehman1996), Mandel (Reference Mandel2003), and Byrne (Reference Byrne2005) cite experimental data that show that people's causal judgments “c caused e” are dissociated from their counterfactual judgments “If c had not occurred, e would not have occurred”: the former go with judgments about sufficient conditions and productive mechanisms, whereas the latter go with judgments about enabling conditions and preventative mechanisms.

There is another way of developing Knobe's general idea that moral considerations influence people's causal judgments by way of their selection of alternative possibilities. In their classic work, Hart and Honoré (Reference Hart and Honoré1985) argue that the concept of actual causation originates in the situation in which a human action intervenes in the normal course of events and makes a difference in the way these develop. “The notion, that a cause is essentially something which interferes with or intervenes in the course of events which would normally take place, is central to the common-sense concept of a cause” (Hart & Honoré Reference Hart and Honoré1985, p. 29). They argue that our judgments about what constitutes the normal course of events are guided context-sensitively – sometimes by what usually happens, and sometimes by social, moral, and legal norms. Their account readily explains why we regard Professor Smith's action rather than the administrative assistant's as the cause of the problem: for his action makes a difference to what happens normally – that is, in conformity with the prevailing norms – in a way that the administrative assistant's does not.

Hart and Honoré's account of the way our causal judgments are shaped by moral considerations is better than Hitchcock and Knobe's for several reasons: (1) Hart and Honoré's account captures in a seamless fashion the idea that causes are difference-makers for their effects. In contrast, it isn't clear how Hitchcock and Knobe's account captures this idea. Is it through the link with counterfactuals or through the rules about counterfactual availability? (2) Hart and Honoré's account doesn't tie people's causal judgments so closely with their counterfactual judgments, which is a virtue given the empirical evidence dissociating them. If it makes a link with counterfactuals, it is with counterfactuals that are based not on the actual world but on “normalised” worlds that abstract from the abnormal features of the actual world (Menzies Reference Menzies, Price and Corry2007). (3) Hart and Honoré's account provides a more uniform account of the contrastive structure of actual causation. Many philosophers have observed that causal judgments have an implicit contrastive structure: the causal judgment “c caused e” has the implicit contrastive structure “c rather than c* caused e rather than e*.” People typically select as the contrast elements c* and e* events that would normally have occurred if the abnormal actual events c and e had not occurred (Menzies Reference Menzies, Beebee, Hitchcock and Menzies2009). This follows straightforwardly from Hart and Honoré's account, which incorporates the contrastive structure into the semantic content of causal judgments. If Hitchcock and Knobe's account is to explain the contrastive character of causal judgments, it must do so through appealing to pragmatic or non-semantic rules about counterfactual availability.