Mental Causation: Compulsion by Reason

The standard paradigm for mental causation is a person’s acting for a reason. Something happens she intentionally φ’s the occurrence of which we explain by citing a relevant belief or desire. In the present context, I simply take for granted the following two conditions on the appropriateness of this explanation. First, the agent φ’s because she believes/desires what we say she does, where this is expressive of a causal dependence.1 Second, her believing/desiring this gives her a reason for φ-ing: recognizing that she has this belief/desire makes her φ-ing intelligible as rational in the light of her other attitudes and circumstances. A further condition must be met, though, if this is to be a genuine psychological explanation, a case of her acting for the reason in question. Consider the following example of Davidson’s (1973, p. 79). An exhausted climber is desperate to rid herself of the weight and danger of holding her partner on a rope; and her sudden realization that simply letting go would achieve this so unnerves her that her grip loosens slightly and he falls. Her releasing him causally depends upon her having this belief and desire, which provide a reason for doing what she does. But this is not why she does it: it would be at best misleading to say that she dropped him, intentionally, because she was fed up with holding his weight, or because she thought that she might otherwise fall. Her letting go does not depend upon her having these reasons in the right way. The reason-giving relation is causally irrelevant. If we are to explain a person’s acting for a reason, then her doing