Responsibility, Tracing, and Consequences

Some accounts of moral responsibility hold that an agent's responsibility is completely determined by some aspect of the agent's mental life at the time of action. For example, some hold that an agent is responsible if and only if there is an appropriate mesh among the agent's particular psychological elements. It is often objected that the particular features of the agent's mental life to which these theorists appeal (such as a particular structure or mesh) are not necessary for responsibility. This is because there appear to be cases in which an agent acts at an earlier time which causes her to lack the appropriate psychological features at some later time and yet, intuitively, she is responsible at that later time. Instead, it is thought, we must adopt a tracing principle that allows us to account for responsibility in these cases by tracing back from the later consequence to some earlier action. My project here is twofold. First, I want to suggest that there is a plausible response to this objection. Second, I want to suggest that the tracing strategy is not only an unnecessary component of a theory of responsibility, but that there are independent reasons to reject tracing in general. This is because the tracing approach is a strategy employed in order to account for an agent's responsibility for the consequences of her actions but, I will argue, agents cannot be responsible for the consequences of their actions.

Moral responsibility, as I am using the term, is taken to be the extent to which an agent is blameworthy or praiseworthy. It is a continuum upon which blameworthiness and praiseworthiness are poles. According to this conception, a difference in blameworthiness just is a difference in responsibility due to the fact that blameworthiness is simply a form of responsibility. I will focus primarily on blameworthiness in the remainder of the paper yet the point applies equally to praiseworthiness.

Consider Harry Frankfurt's influential account of freedom of the will.¹ Frankfurt distinguishes between an agent's first and second-order desires. A first-order desire is simply a desire for some object or state of affairs. At any given time we may have a host of different and often conflicting first-order desires. But one of these desires will be effective in the sense that it will move one to act. For Frankfurt, the first-order desire that moves one to action is one's will. Additionally, we often have desires that take as their object another desire, what he calls an agent's second-order desires. I may, for example, desire the desire to work long hours. Of an agent's second-order desires we can distinguish those that are not merely the desire to have some desire but the desire that some particular first-order desire be effective. These he calls an agent's second-order volitions; an agent's desire that his will be a certain way. On this view, an agent is responsible if and only if there is the proper alignment between the agent's will and his second-order volition.

Gary Watson, in his famous critique of Frankfurt, proposes his own account that also grounds moral responsibility in the agent's psychological architecture.² Drawing on a Platonic conception of moral psychology, Watson distinguishes between an agent's motivational system and an agent's evaluational system. On Watson's view an agent is responsible if and only if there is an appropriate mesh between these two systems; when one's motivations are in line with one's values.

On both of these views, responsibility is to be explained by appeal to structural features of the agent's psychology at the time of action. They both hold that responsibility requires that the agent's psychology exhibit these purely structural features.

Many have objected to the claim that some such psychological structure is a necessary condition of responsibility. The drunk driving case is the stock example used in this context. Consider an agent who freely and knowingly becomes inebriated at a party and then proceeds to drive home...