Abstract

The problem of moral luck is rooted in the following pair of conflicting intuitions. On the one hand is the idea that the basis of moral assessments should be limited to factors under agents’ control. This intuition can be formulated as the Control Principle: (CP) We are morally assessable only to the extent that what we are assessed for depends on factors under our control. On the other hand is the fact that there seem to be a plethora of cases in which moral assessments depend (seemingly correctly) on factors beyond agents’ control. This intuition can be formulated as the principle of moral luck: (ML) moral luck occurs when an agent can be correctly treated as an object of moral judgment, despite the fact that a significant aspect of what he is assessed for depends on factors beyond his control. It is important to clarify the kind of control at issue in the Control Principle. The following example will help. Wesley decides, while riding his well functioning bicycle, to turn right. Wesley decides to turn his bicycle to the right and he does so. Additionally, Wesley was also able, at that time, to decide to turn left instead of right and he would have succeeded had he tried to turn left. Wesley has standard control insofar as he turns to the right but he also has the power at that time to turn to the left instead. Wesley would have curtailed control if he turned right but lacked the power at that time to turn left instead. I take it that the kind of control at issue in the Control Principle is at least curtailed control. Whether standard control is the kind required, which would commit us to requiring that there were alternative possibilities open to the agent at the time of the action or event, is not something I will try to answer: trying to do so would take us too far afield. Further distinctions can be made with respect to curtailed control. In addition to requiring that an agent have, at least, curtailed control, we might also require that the agent actually exercise this control over the event in question; or we may require that he simply have the capacity to exercise this control. Further, if we require that agents actually exercise control in order to be responsible, we may require that the control exercised be active control. We can distinguish between direct and indirect control where the former requires an exercise of active control and the latter requires a capacity to exercise control or perhaps a connection to a previous exercise of active control. An agent has direct control, for example, with respect to his decisions and indirect control with respect to the consequences of his decisions. Whether an account of responsibility accepts or denies the Control Principle will turn on whether the Control Principle is read as requiring that agents exercise active control over the event in question or is read as requiring that agents simply have a capacity to exercise control, though not necessarily active control. If we read the Control Principle as requiring that agents exercise active control, then the Control Principle is likely to be accepted (and moral luck denied ) by those accounts of responsibility that require that agents exercise active control. Accounts that require that agents have a capacity to exercise control, but do not require that they exercise this capacity, will likely deny the Control Principle and accept moral luck. I am supposing here that proponents of these accounts will likely give the Control Principle the “exercise of control” reading. An alternative for the latter accounts is to accept the Control Principle but claim that it does not actually require that agents exercise control. In this case, proponents would give the Control Principle the “mere capacity to exercise control” reading. The account of responsibility for which I argue requires that agents exercise active control in order to be morally responsible. The account I initially argue for, which is similar to Michael Zimmerman’s account, takes an “exercise of control” reading of the Control Principle and aims to deny moral luck altogether. As it turns out, luck may not be eliminable from moral responsibility. I ultimately conclude that a variation of my initial account, which is able to accept an “exercise of control” reading of the Control Principle while accepting some luck (i.e., the kind of luck that is not responsibility undermining), is the correct account of moral responsibility. I carve up the responsibility terrain in the following way. I distinguish assessability, one type of moral responsibility, from accountability, a second type of moral responsibility. Assessability makes appropriate certain judgments about agents in virtue of their character traits, actions, omissions, attitudes, intentions, or decisions. I split moral assessability into two competing accounts of what assessability is - attributability and appraisability. We can be morally assessed, in the attributability sense, on the basis of things over which we do not exercise control, though we must still have the capacity to exercise control or the capacity to make judgments and then hold or maintain attitudes based on these judgments. Appraisability, however, requires that we exercise control. We are directly appraisable only for things over which we exercise direct active control, which is limited to our decisions, and indirectly appraisable for the consequences of our decisions. Limiting appraisability in this way is a strategy for avoiding the moral luck that influences attributability. The second type of moral responsibility I discuss is accountability, which concerns whether certain treatment of agents is appropriate. Accountability is closely related to assessability: What makes certain treatment appropriate is the same fact that makes certain judgments appropriate. Accountability has a control condition, without which certain treatment is not appropriate. I argue that accountability, like appraisability, requires that agents exercise control. If this is the case, then for this reason, appraisability, but not attributability, is the kind of assessability that can ground accountability. What makes appropriate the judgment that an agent is appraisable also justifies certain treatment of her, though the fact that this judgment is appropriate may not be sufficient to warrant treatment because we may have other (nonmoral or non-desert-based) reasons to forego treatment. Ultimately, I argue for appraisability over attributability. Originally, I thought that immunity from luck was one of the advantages appraisability has over attributability. As it turns out, appraisability must accept luck at some places but not others, which is problematic. If an account must accept some luck, it looks like it must accept all kinds of luck because there is no principled place to draw a line between kinds of luck. If this is right, and we follow an entirely luck-free account of moral responsibility to its logical end, we end up with responsibility being impossible and no one being responsible for anything. The result of this finding is that no plausible account of moral responsibility can be entirely free of luck and immunity from luck is not one of the distinguishing features of appraisability. There are, however, other significant differences between appraisability and attributability. Appraisability requires an exercise of active control and is tightly connected to accountability (the other type of responsibility). Attributability does not require an exercise of control (active or otherwise), which leaves the account open to problems of manipulation. By not requiring an exercise of control, attributability is not connected to accountability so the moral judgments we make do not enable us to justify treatment, which is some of the work we want and need moral judgments to do. I contend that these differences make appraisability the correct account of moral responsibility. My original account of appraisability must undergo significant revision, though, in order to avoid problems, including the strategy used to minimize the influence of luck