One of the presentations I gave at the TAS Seminar a few weeks ago was on moral luck, via Rosalind Hursthouse’s discussion of “the repentant racist” in her 2001 book, On Virtue Ethics. I’ve outlined that argument in an earlier post, so for the next few posts, I’ll focus in a broader way on moral luck as an issue at the intersection of Objectivism and analytic moral philosophy.
When I was a kid, I was a devout Muslim and took great pride in keeping the Ramadan fasts. According to Islamic tradition, a Muslim is obliged during the month of Ramadan to abstain from food, drink (including water), and sex from sun-up to sun-down. (One of those abstentions was easy enough.) Of course, as a soccer player, I took soccer about as seriously as I did religion. So I insisted on playing soccer while I was fasting—stupid, but hard to resist.
Since the Islamic calendar runs on the lunar calendar, and the lunar calendar changes each year with respect to the solar, one year Ramadan fell during the hottest days of the summer. One day during one such summer, I came back from playing soccer, and without realizing what I was doing, poured myself a glass of water and drank it. A few minutes later, I realized with horror what I’d done, and collapsed immediately into a heap of sobbing, praying repentance. AC/DC’s “Highway to Hell” suddenly took on vivid and literal meaning. I was twelve.
Eventually, my mother came home and asked, in effect, what the hell was going on. I told her that I had sinned, and explained how it had happened. “But did you mean to break your fast?” she asked, earnestly. “No,” I sobbed. “It just happened. I don’t know how.” And I really didn’t.
She brightened with the evident relief of a mom who (a) knows that there’s a way out of this one, and (b) is glad it wasn’t something serious, like masturbation. “It isn’t a sin if it was out of your control,” she said, in soothing tones of motherly axiomaticity. “It all depends on your niyat,” she continued, using the Urdu word for “intention.” (Actually, niyat is an untranslatable species of intention—it refers to a religiously sincere intention in contexts of specifically divine scrutiny. In fact, its untranslatability indicates the coarse-grained character of the catch-all English term, “intention.”) “If your niyat is good, the action is good. If your niyat is bad, it’s a sin.”
“But then how did it happen?” I asked. “How could I just do something where it was totally out of my control?”
“Oh,” she rejoined. “That was your guardian angel looking after you. God knew you were dehydrated and needed water, so to protect you, He had your guardian angel make you drink.”
Well, of course! My guardian angel supplied the necessary causal connection between lack of intention and uncontrolled action. One tends to forget that possibility.
Like all maternally-based explanations of complex moral phenomena, this one had the right mixture of plausibility and absurdity about it to make for years of uneasy acceptance. Eventually, the angels dropped out of the mix, and I was left with: “To be blamed for something, I have to control it. No control, no blame.” Had I generalized the principle, I might have come up with something like this principle, dubbed (by Dana Nelkin of UCSD) CP, or the Control Principle.
(CP) We are morally assessable only to the extent that what we are assessed for depends on factors under our control.
My commitment to CP lasted until graduate school, when I encountered the literature on moral luck. Somewhere in the 1990s, I read Thomas Nagel’s and Bernard Williams’s seminal papers on moral luck, along with Martha Nussbaum’s heavily Nagel- and Williams-influenced Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy, etc. The basic idea is that moral luck is a phenomenon whose existence entails that CP is false. It implies that we are morally assessable on the basis of things we can’t control. It was a paradoxical-sounded idea that I didn’t ever properly digest. I ended up thoroughly confused about it.
What gives moral luck its apparent plausibility are hard or puzzling cases in which it looks as though “our” moral assessments “require” the rejection of CP. “We” feel the need to assess an agent or action on moral grounds but can’t readily identify what it is the agent controls. In such cases (say the proponents of moral luck), what seems to matter is not the fact or locus of the agent’s control over the object of moral judgment, but the imperative of passing moral judgment regardless of our knowledge of the facts about agent-control.
Personally, I don’t buy this; I once did, but no longer do. If we can’t be sure of the fact that the agent controls an action, or the locus of that control, how can we be sure of the need for a specifically moral assessment? I don’t think we can, but I’ll get to that in later posts. For now, as a preliminary to that longer discussion, I want to start by defining moral luck. To that end, consider a few classic formulations, not all of them originally intended as attempts at a formal definition of “moral luck,” but all of them helpful en route to a definition. I’m belaboring this preliminary issue because one problem with “the problem of moral luck” is that it’s not always entirely clear what “the problem” really is. So here’s a quick discussion of four statements of “the problem”—Nussbaum’s, Williams’s, Nagel’s, and Dana Nelkins’s. As I see it, there is a problem with the statements of the problem itself.
Nussbaum’s Fragility of Goodness opens with what Nussbaum describes as an Aristotelian form of anti-Kantianism about moral luck heavily influenced not just by Aristotle but by Nagel and Williams. She writes: “For the Kantian believes that there is one domain of value, the domain of moral value, that is altogether immune to the assaults of luck” (p. 8). Her discussion, she says, focuses on three questions:
1. “How much should a rational plan of life allow for elements such as friendship, love, political activity, attachment to property or possessions, all of which, being themselves vulnerable, make the person who stakes his or her good to them similarly open to chance?” (p. 6)
2. “Do [the components of a rational plan of life] coexist harmoniously, or are they capable, in circumstances not of the agent’s making, of generating conflicting requirements that can themselves impair the goodness of agency?” (p. 7)
3. Finally, how much should a rational person give to the expression in her life of “the appetites and emotions,” given that their expression depends on resources and persons outside of the agent’s own control? (pp. 7-8)
Nussbaum describes her topic in the book generally as follows:
Human cognitive limits circumscribe and limit ethical knowledge and discourse; and an important topic within ethical discourse must be the determination of an appropriate human attitude toward those limits. (p. 8)
Despite the impression she gives in the book, I don’t think Nussbaum’s way of framing the issues has much to do with the problem of moral luck.
Take question (1). It’s a good question, but it has nothing to do with the existence or non-existence of moral luck. Suppose that a rational plan of life should allow plenty of room for friendship, love, political activity, and the rest. Yes, all of those things depend on an openness to vulnerability, and yes, all of them leave the agent open to the contingencies of luck, including bad luck. But the question of moral luck concerns our moral assessment of the agent’s moral agency, not the overall trajectory of his life and well-being. The agent in question could still be morally assessed insofar as he controlled his conception and implementation of the plan, but not morally assessed for the success or failure of his projects (insofar as that was out of his control). So no answer to Nussbaum’s question establishes the existence of moral luck. We could recommend that people stake their well-being on love, admit that love can go wrong, admit that love’s going wrong leads to heartache, admit that its doing so can be partly a matter of chance, and still be none the wiser about moral luck.
Or consider question (2). This is closer to the mark, but something of a loaded question. Suppose that the components of a rational life all coexist harmoniously (as per Ayn Rand’s conception of the harmony of rational interests). There is in that case no special reason to think that CP is threatened, so set it aside. Now suppose (pace Rand) that the components of a good or rational life generate conflicting requirements. Nussbaum’s implication seems to be that this supposition somehow threatens CP. I don’t see why. Since the fact of conflict (its metaphysical basis) is ex hypothesi out of the agent’s control, and that fact leaves the agent incapable of rational decision-making, why assume that it impairs the goodness of the agent’s agency? It has precisely nothing to do with the agent’s agency. It certainly screws up the agent’s life. But it leaves him morally intact. If it’s not his fault that the moral dilemmas arise, why think that he’s to be morally assessed when they do? If there is no rational decision to make in a case where all directives are equally unjustifiable, no praise or blame can arise for whatever decision he does make (if he makes one at all). Yes, it would be tragic if life constantly (or even occasionally) led to dilemmas not themselves explainable by failures of human reason or virtue. But it wouldn’t falsify CP.
Question (3) is just like question (1), and so is the response to it. Maybe an agent should give plenty of free rein to appetites and emotions, but at the end of the day, he is responsible for controlling the free rein he gives (or the tight control he maintains). The relevance to moral luck remains unclear.
Finally, consider Nussbaum’s description of her topic. The description is fair enough as the description of a topic well worth discussing, but if the “determination of an appropriate human attitude toward those limits” can be within an agent’s control, as can compliance with the limits, we don’t even have prima facie reason for rejecting CP or affirming the existence of moral luck. And nothing about the existence of human cognitive limits suggests that we should be morally assessed even when the basis of the assessment is literally beyond our cognitive limits. Suppose that you somehow discovered with 100% certainty that the people of Gaza City or South Waziristan were so traumatized by their predicament that they were literally and utterly incapable of cognizing the requirements of justice. Now suppose that some of them joined Hamas or the Taliban and killed thousands of innocents. Could they be morally assessed for doing so? My answer: no. I find this “no” so obvious that I literally do not understand how anyone could say “yes” to the question. (Incidentally, I don’t happen to think that the preceding supposition is true. I just mean that if it were true, all moral bets would be off.)
For these reasons, it’s somewhat unclear, despite her invocation of the topic, what relevance Nussbaum’s book bears to the topic of moral luck.
Bernard Williams’s famous paper “Moral Luck” doesn’t once define the term. The closest we come to a definition are these two formulations on the first page:
The idea that one’s whole life can in some way be rendered immune to luck has perhaps rarely prevailed…but its place has been taken by the powerfully influential idea that there is a basic form of value, moral value, which is immune to luck….
[On this view], [a]nything which is the product of happy or unhappy contingency is no proper object of moral assessment, and no proper determinant of it, either. (p. 20)
So “moral luck” must name the denial of the previous thought. I’d translate Williams’s point this way:
Moral luck is the idea that it is impossible to render one’s whole life immune to luck. There is no such thing as distinctively moral value immune to luck. Some things that are the product of good or bad luck can be proper objects and proper determinants, of moral assessment.
That’s not exactly a definition, but it gets us further than Nussbaum’s discussions.
One thing that remains fuzzy in Williams’s account is “things.” If all of X is a product of luck, is all of X a proper object of moral assessment? If my character is partly a product of luck, is the luck-caused part a proper object of moral assessment? Or are we allowed to divide character up into the luck-caused and volitionally-caused parts? I am tempted to take the first sentence in the excerpt to say that every part of life is wholly subject to luck. If that were true, one’s character would be wholly subject to luck, and there would be no way to distinguish between the luck-caused and volitionally-caused parts of a person’s character. I am not sure Williams intended that implication, but it’s not an uncharitable inference from what he does say.
Despite its classic status, I find Williams’s discussion of moral luck maddeningly opaque. I find Nagel’s (equally classic) discussion a lot clearer. Here’s Nagel’s succinct formulation in Mortal Questions:
Where 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, it can be called moral luck. (p. 26, emphasis mine)
The “in that respect” that I’ve emphasized seems to answer the questions I asked of Williams. Nagel seems to be insisting that the very thing beyond the agent’s control is the thing subject to moral assessment qua non-volitional.
Dana Nelkins is a contemporary specialist on moral luck. Here’s her formulation in the Stanford Encyclopedia of Philosophy:
(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.
I see two problems with this definition. For one thing, it lacks a genus. Second, in saying that “a significant aspect of what he is assessed for depends on factors beyond his control,” Nelkin seems to me to fudge the very issue raised by critics of the coherence of moral luck (and was the point of my hair-splitting about the Williams and Nagel formulations above). The issue is not whether “a significant aspect” of the object of judgment is out of the agent’s control, but whether the object qua-object-of-moral-judgment is out of the agent’s control. This latter point, as I’ll try to explain in further posts, is what I regard as the crux of the issue. I would insist on integrating “the object qua-object-of-moral-judgment” element into a definition of moral luck from the outset, even at the risk of someone’s complaining that my definition was too theory-laden. (All definitions of complex phenomena are theory-laden, and I regard it as an exercise in futility to try to bypass that fact by trying to produce theoretically “neutral” definitions. This issue was hotly debated at the TAS Seminar, and I hope to return to it.)
In my TAS talk on moral luck, I used a revised version of Nelkin’s formulation:
“Moral luck” refers to a situation in which a moral agent’s actions or character can justifiably be given a moral judgment despite the fact that the object of the judgment—the thing being judged—is beyond her control.
Incidentally, most philosophers distinguish types of moral luck (resultant, circumstantial, constitutive, and causal), but I’ve omitted that from my definition on the assumption that they are (or ought to be understood as) species of moral luck. I’ve just given a general definition of the phenomenon itself.
With a definition in hand, most of the issue of moral luck turns on whether particular cases can or can’t be understood compatibly with CP. The case I described at the outset—my Ramadan fast-breaking—seems a very clear instance of one that is clearly compatible with CP. If we assume ex hypothesi that fast-breaking was wrong, then doing so without realizing what you are doing is not to be doing anything blameworthy—[even if it ruins the fast and whatever value fasting is supposed to have]*. I take the fast-breaking case to be a sort of paradigm case. The question is whether other cases—all other (relevant) cases—can be understood in some analogous way. I’ll talk about a couple cases in my next post, and try to outline a (very) general strategy for dealing with them.
PS., August 27, 2013: In retrospect, it occurs to me that the bracketed phrase here is badly worded. What I should have said was: “even if it ruins the successful effectuation of the fast, whatever value its full effectuation was intended to bring about.” Since my point was to distinguish between the parts of “the fast” that are in the agent’s control and the parts that are not, the way I’ve put things sounds as though the lapse I described destroys the fast tout court. But my mother’s point was that it doesn’t. And for once, I’m agreeing with Mom.