Diana Mertz Hsieh
Department of Philosophy
University of Colorado, Boulder
7 December 2007
The problem of moral luck is best understood as a clash of common beliefs about moral responsibility and moral judgment. On one hand, people commonly think that a person cannot be justly praised or blamed for his actions unless he controls them. So if a man releases a critical pulley rope on a construction job due to a sudden heart attack, leaves the scene of an auto accident because he’s spirited away by kidnappers, or breaks a vase when knocked over by a strong gust of wind, his lack of control over his bodily movements should absolve him of any moral blame. The same is said of character traits and the outcomes of actions: a person may be justly praised and blamed only to the extent that he exerts control over them. On the other hand, ordinary moral judgments of persons routinely vary based on the actual goods or evils caused by the person, even when partly or wholly beyond his control. For example, the drunk driver who kills two pedestrians is blamed far more than the drunk driver who merely collides with a telephone pole, even if their driving was equally reckless. The only difference in what they’ve done is due to luck, yet they are blamed unequally by themselves and others. Similarly, the coward in Hitler’s Third Reich who betrays his Jewish neighbors to the authorities is responsible for contributing to genocide, whereas a man of identical character in America today might never be guilty of worse than failing to defend his wife from his sniping parents. These two men differ radically in their moral records solely based on the accidental circumstances of their births.
The problem of moral luck is the apparent conflict between the idea that a morally responsible agent must control his actions and the standard practice of blaming people more simply for causing worse results. As developed most clearly and forcefully by Thomas Nagel, the proposed category of moral luck attempts to highlight a range of cases in which “a significant aspect of what someone does depends on factors beyond his control, yet we continue to treat him in that respect as an object of moral judgment.”1 Matters of luck arguably influence all that a person is morally judged for, not only his choices and actions but also his character. Consequently, Nagel claims, “ultimately, nothing or almost nothing about what a person does seems to be under his control,” meaning that the consistent application of the principle that responsibility requires control threatens most if not all our ordinary moral judgments.2
My thesis, in brief, is that the problem of moral luck stems from a faulty understanding of the conditions of moral responsibility. A person need not solely determine all of that for which he is morally judged, as Nagel supposes. Instead, a person is properly held responsible for his voluntary acts. When a person acts voluntarily, (1) he has the power to act or not and (2) he knows what he’s doing. That Aristotelian understanding of the conditions of moral responsibility is not only consistent with standard intuitions but also grounded in basic facts about human capacities and about the purposes and demands of moral judgment. When developed in sufficient detail and extended to responsibility for a person’s products and qualities, that theory can effectively solve the puzzling cases of moral luck raised by Nagel and others, such that moral responsibility clearly tracks a person’s voluntary actions, products, and qualities.
The basic structure of this prospectus (and ultimately, of the dissertation) is fairly simple. First, I will describe the basic problem of moral luck as developed by Nagel and others. I will scrutinize the standard attempts to solve the problem of moral luck, as well as Nagel’s implicit understanding of the kind of control required for responsibility. I will also consider why the problem of moral luck as formulated by Nagel seems intractable. Second, I will develop a broadly Aristotelian theory of moral responsibility based on an analysis of the nature, purpose, and demands of moral judgment and the nature of human agency. Third, I will further develop and refine that theory of moral responsibility in the course of applying it in turn to each of the three basic kinds of moral luck: resultant luck, circumstantial luck, and constitutive luck. My analysis will show that the seemingly hopeless clash of intuitions in the various cases of moral luck can be satisfactorily resolved by a proper theory of moral responsibility, albeit perhaps not always quite as expected.
The dissertation will rely on Aristotle’s ethics and moral psychology as a general framework for the development of a robust theory of moral responsibility. While I intend to generally steer clear of substantive moral questions, the work will presuppose a teleological rather than deontological approach to ethics, meaning that “the moral propriety of actions depends on their relationship to [the] overarching end” of the agent’s own flourishing.3 It will also rely on an incompatibilist understanding of free will as the agent’s power to perform or not perform some action, independent of prior conditions.4
In the following sections, I sketch the core arguments of my analysis of the problem of moral luck. I have chosen to do so in considerable detail because that enabled me to develop my account of moral responsibility clearly enough to test it against the core cases of moral luck.
The Problem of Moral Luck
Nagel’s case for pervasive moral luck begins with a brief survey of “the ordinary conditions of moral judgment,” particularly the “control condition” for moral responsibility.5 Appealing to the primitive intuition that “people cannot be morally assessed for what is not their fault, or for what is due to factors outside their control,” Nagel observes that “the appropriateness of moral assessment is easily undermined by the discovery that the act or attribute, no matter how good or bad, is not under the person’s control.”6 So “a clear absence of control, produced by involuntary movement, physical force, or ignorance of the circumstances, excuses what is done from moral judgment.”7 The problem of moral luck arises from the attempt to consistently apply that control condition in our everyday moral judgments. When we look closely, Nagel claims, we find that “what we do depends in many more ways than [commonly thought] on what is not under our control,” yet the “external influences in this broader range are not usually thought to excuse what is done from moral judgment, positive or negative.”8 So the problem of moral luck is that our ordinary moral judgments routinely violate the control condition: people are praised and blamed for matters beyond their control.
Nagel classifies the various cases of moral luck as resultant, circumstantial, or constitutive luck—based on that which is affected by luck.9 In cases of resultant luck, a person is morally judged partly based on the outcome of his action despite his lack of control over that outcome, such as in cases of inherently risky action (e.g., inciting bloody revolution), failed attempts (e.g., shooting someone but not killing him as intended), and negligence (e.g., text messaging while driving).10 In cases of circumstantial luck, the person’s moral record depends on accidental circumstances, as when a person faces a difficult moral test to which others are never put or when a would-be wrongdoer finds his deed already done for him by others.11 In cases of constitutive luck, a person is praised or blamed for aspects of his moral character imposed upon him by his upbringing or his genes, for example.12 These cases seem to show that our standard moral judgments of a person—whether for his products, his choices, or his character—are often substantially based on accidental factors outside his control.
Notably, the problem of moral luck does not merely present us with a limited set of puzzling cases about moral responsibility.13 Luck is a pervasive influence in human life. No one controls the particular family, culture, nation, or era of his birth. No one controls his genetic endowments. Few people have any significant power to influence the economic conditions, political institutions, or moral climate that shape their lives. Our actions often have far-reaching, unexpected, and unpredictable effects in the world. Such external forces seem to influence the thoughts, actions, qualities, and products for which a person is morally judged. If that’s true, then the problem of moral luck undermines attributions of moral responsibility generally, not just in a few select cases. That’s why Nagel claims that “if the condition of control is consistently applied, it threatens to erode most of the moral assessments we find it natural to make.”14
Three Attempted Responses
If true, Nagel’s conclusion—that most ordinary claims of moral responsibility cannot survive the consistent application of the principle that moral responsibility requires control—would be a bitter pill to swallow. Accepting it would spell the end of ethics as a practical discipline. If people aren’t responsible for their actions and characters, the task of differentiating right from wrong and virtue from vice would be an intellectual exercise without further purpose. Consequently, most philosophers commenting on the problem of moral luck attempt to retain our practices of moral judgment in some form by developing alternative accounts of the relationship of morality and luck. That has proven more difficult than expected, in that neither the attempt to exclude luck from morality nor the attempt to include luck in morality seems to produce a plausible general theory of moral responsibility. Both approaches seem to work well when applied to some cases, then lapse into absurdity in others.
The standard strategy for eliminating the influence of luck in moral judgments is simple: remove any and all luck-based differences from moral judgments and punishments, so that cases differing only in matters of luck are judged and punished equally.15 The resulting moral equivalence is undoubtedly plausible in cases of attempted murder foiled after the attack, as when the rapidly exsanguinating victim of an assault is rescued from death by a neighbor who happens to be a trauma surgeon stopping by to return a borrowed garden hose at that very moment. Contrary to current legal practice, the fortuitous rescue of the victim doesn’t seem to be legitimate grounds on which to weaken our moral condemnation of the violent act or its perpetrator, nor lessen his punishment. Similarly, the judge who would accept a bribe if offered seems worthy of equal condemnation (even if not equal punishment) as the judge who did accept such a bribe when offered. However plausible those analyses, the attempt to directly remove the influence of luck from all moral judgments yields the implausible conclusion that people with radically different moral records and moral characters should be blamed equally via a series of cases differing only in luck.16 Here’s how:
Step 1: Eliminate differences in moral judgments due to resultant moral luck. Imagine that two people, Adam and Bonnie, voluntarily drive themselves home from a party, even though seriously impaired by alcohol. At various times, both drivers swerve into oncoming traffic, run red lights, and drive erratically—but only Adam strikes and kills a pedestrian. Since the presence of the pedestrian at the intersection was not in Adam’s control, the death is a matter of bad luck for Adam. To eliminate the effect of luck, Adam and Bonnie must be judged and punished equally for their drunk driving alone, since that’s what they did control.
Step 2: Eliminate differences in moral judgments due to circumstantial moral luck. Imagine a third party-goer Clive, who intends to drink and drive exactly like Adam and Bonnie, but who stumbles on a hidden rock while walking to his car, bumps his head, and passes out in the bushes. Since Clive’s intended drunk driving was prevented by the mere accident of an ill-placed rock, Adam, Bonnie, and Clive should be blamed and punished equally, based on their equal intention to drive drunk. Next, imagine David: he would have drunk to excess and driven home but was precluded from doing so solely by his work schedule. Perhaps he even attempted to switch shifts with his fellow workers but found no takers. Since David’s drinking and driving was precluded by the mere accident of his work schedule, he should be blamed and punished equal with Adam, Bonnie, and Clive based on a willingness to drive drunk. The only differences between them are the product of luck, so to judge them differently would be unjust.
Step 3: Eliminate differences in moral judgment due to constitutive moral luck. Imagine that Ernie would have driven home drunk from the party, had his best friend not been killed by drunk driver last year. Instead, he drinks club soda and lime then drives himself home safely. Ernie might seem to be a morally better person than the others since he refuses to drive drunk. Yet he would not always drive sober if his friend hadn’t been killed—and that experience was purely a matter of luck. We can even imagine that Adam, Bonnie, Clive, and David would not drive drunk if something similar happened to them. Consequently, Ernie should be judged and blamed equal to Adam, Bonnie, Clive, and David.
In sum, the attempt to purge luck from moral judgments would require declaring Adam, Bonnie, Clive, David, and Ernie morally equivalent and treating them equally—despite the often-vast differences in the harms caused, the actions taken, the intentions formed, and the characters enacted by each person.17 Ultimately, the most vicious criminals might need to be judged morally equal to the most virtuous saints, on the grounds that some chain of purely luck-based differences connects them, just as Adam is connected to Ernie. Here, luck in the circumstances that shape a person’s basic character is of particular concern. After all, the fiery abolitionist who helped slaves escape north to freedom in the 1850s might have defended and practiced brutal forms of slavery if unlucky enough to have been raised in a slaveholding southern family. And the 20-year-old thug who shoots a frightened bystander in a robbery might have been quietly studying in his college dorm if his parents had been more concerned with his education than with their next fix.18 If those counterfactuals are true, then the abolitionist must be judged the same as the slaveholder and the thug the same as the student in order to eliminate the effects of luck from our moral judgments. Even if those counterfactuals are merely probable or possible, then to praise one person and condemn the other seems to make the moral judgment unfairly dependent on luck. Yet equalizing the moral judgments to eliminate the effect of luck also seems deeply unfair, if not absurd.
In addition, the equalization of judgments renders the proper judgments and deserts of persons totally mysterious.19 Does justice demand excusing everyone on the grounds that they might have driven sober with Ernie’s good luck or blaming everyone on the grounds that they might have killed a pedestrian with Adam’s bad luck? Absent some luck-free baseline for moral judgments, any answer seems arbitrary.20
Philosophers seeking to exclude luck from moral judgments deny that Ernie is the moral equal of Adam, however. They often accept that Adam, Bonnie, Clive, and David should be judged and punished equally given their equal moral character, meaning based on the counterfactual claim that all would have acted the same under the same conditions.21 The same cannot be said of Ernie, however. On this approach, moral judgments are not made of a person’s luck-infected choices, actions, or outcomes; instead, only a person’s character is subject to moral judgment.22 So then what is said of the problem of constitutive moral luck, i.e., of luck in moral character itself? Rescher offers the standard reply, arguing that very concept of constitutive moral luck is logically incoherent because a person must be someone in particular in order to be subject to luck.23 He claims that “one cannot meaningfully be said to be lucky in regard to who one is, but only with respect to what happens to one.”24 That’s why it’s not sensible, for example, to ask what kind of person you’d be if you were unlucky enough to be born to starving Somali refugees, medieval Germanic peasants, or abusive drug-addicted celebrities. In such cases, you wouldn’t exist at all; someone else would.
This attempt to avoid moral luck is subject to serious objections. Personal identity may well entail that some aspects of a person’s character are essential to him, such that those aspects cannot be coherently attributed to luck. Yet the elimination of all luck in character would require the far stronger thesis that any and all aspects of a person’s character are essential to his identity. Surprisingly, Rescher seems to endorse that strong view in describing “one’s inclinations, disposition, and character,” whether within one’s control or not, as “a crucial part of what constitutes oneself as such”—without any qualification or distinction between essential and non-essential traits.25 That view of personal identity would render even the most minor changes to a person’s character impossible, meaning that Ernie would be a different person before and after his friend’s accident.26 In fact, such inessential changes are not only routine but also often inspired, influenced, or facilitated by accidental and unexpected external forces, i.e., by luck.
Worrisomely, the alternative strategy of abandoning the control condition so as to allow some role for luck in moral judgments is no more appealing in its general results. That approach would allow us to praise and blame a person based on what he actually does, causes, and is like—even when influenced by luck. So Adam should be blamed more severely than Bonnie because he caused more damage, while Bonnie should be blamed more severely than Clive and David because she acted wrongly while they merely intended and hoped (respectively) to do so. Far less plausibly, however, two equally vicious people with identical malicious plans executed in exactly the same way would be judged differently based on the unforeseeable intervention of random outside forces, like whether a bird flew into the path of the bullet intended to kill or whether the trauma surgeon attempted to return his neighbor’s garden hose at just the right moment. That seems manifestly unjust, since the differences between the lucky and the unlucky wrongdoer had nothing to do with anything about them. Their choices, actions, and characters are the same, yet they are blamed unequally.27 As we shall see, that approach would frustrate the basic purpose of moral judgment.
Moreover, the most plausible substitutes for the control condition seem unable to adequately account for our ordinary judgments of responsibility. Presumably, a person’s moral responsibility must be limited to that which he causes; otherwise, car salesmen leading quiet lives in upstate New York might be blamed for a shortage of tea in remote villages in China.28 That causal condition isn’t sufficient, however, as people also shouldn’t be held responsible for the myriad far-flung, improbable, and unforeseeable effects of their actions.29 A person shouldn’t be blamed for telling the time to a stranger at the mall, for example, even if that knowledge helps the stranger kidnap a child. Such false claims of responsibility could be excluded by requiring the outcome to be foreseen or at least reasonably foreseeable.30 Yet that wouldn’t preclude holding a person responsible for some event he foresees and causes but cannot act to prevent. So if clever terrorists kidnap Jack and securely rig him to a bomb set to explode after his heart beats a few hundred times, he causes and foresees the explosion, yet he is not properly blamed for it, most plausibly because he cannot control the beating of his heart. So notwithstanding the puzzling cases of moral luck, control does seem somehow necessary to moral responsibility, as Nagel claims.
For some philosophers, the failure of the attempt to eliminate luck from moral judgment shows that moral luck is inescapable.31 Yet as we’ve seen, the attempt to incorporate luck into moral judgments by rejecting the control condition creates significant problems for a general theory of moral responsibility, just as does the attempt to exclude luck from moral judgments by strictly applying the control condition. Nagel attempts a third approach: the problem of moral luck is the product of an irreconcilable conflict between the subjective and the objective perspectives on persons.32 He claims that we initially think of ourselves and others from a subjective (or first-person) perspective, i.e., as agents in control of and responsible for our own actions.33 Yet as we investigate the external forces that influence a person’s choices, actions, and character, we are forced to assume the objective (or third-person) perspective according to which “actions are events and people things.”34 Then we see the morally responsible agent as “merely… a bit of the world,” such that “the alternatives that he may think of as available to him are… just alternative courses that the world might have taken.”35 So ultimately, Nagel claims, “nothing remains which can be ascribed to the responsible self, and we are left with nothing but a portion of the larger sequence of events, which can be deplored or celebrated, but not blamed or praised.”36 Nonetheless, we cannot abandon our original understanding of ourselves and others as agents, not “even when we have seen that we are not responsible for our own existence, or our nature, or the choices we have to make.”37 Consequently, control still seems necessary for moral responsibility, even though the strict application of the control condition in the proposed cases of moral luck ultimately undermines all our standard attributions of moral responsibility.38 So for Nagel, the problem of moral luck is ultimately insoluble. Like other related problems of autonomy and responsibility, it has “no available solution.”39
This view of the origin of the problem of moral luck raises perhaps more troubling questions than the already-surveyed attempts to forbid or permit some role for luck in moral judgments. Nagel regards the subjective and objective perspectives on human agency as equally compelling and equally necessary—yet hopelessly contradictory.40 In particular, the identification of ethical values requires the assumption of an objective perspective, yet that very perspective precludes regarding ourselves and others as morally responsible agents.41 The implications of Nagel’s willing acceptance of such major philosophic contradictions as beyond our power to resolve are disturbing, to say the least. Presumably, we are unable to rationally determine truth about the nature of human agency and responsibility either because humans are systematically deceived about our place in the natural world or because reality itself is contradictory. Either way, rational philosophy would be in serious peril, if even possible. In fact, however, the source of the Nagel’s irresolvable conflict is probably more mundane: although Nagel repeatedly denies that his arguments presuppose any form of determinism, that is precisely what the objective perspective seems to demand.42 However, neither Nagel’s understanding of moral luck as rooted in conflicting perspectives nor his seeming determinism is necessary to feel the force of his cases of moral luck; few (if any) other philosophers accept his account of the problem’s origins.