Monterey Peninsula College

Abstract

Forty years ago, when Bernard Williams wrote his classic discussion of moral conflict, "Ethical Consistency", the analysis of moral conflict and deliberation was carried on in terms of conflicting obligations or "oughts". An agent finds herself with two obligations, the performance of either of which excludes performance of the other, and chooses to act in accordance with one of them. What happens to "the ought not acted upon?" Williams famously asked. His answer, of course, was that the ought not acted upon lingers as a moral remainder of the conflict, and that the wrongness of not acting on the ought not acted upon can result in appropriate regret and even further obligations of apology or reparation.

In the intervening decades, the discussion has shifted from talk of obligations to talk of reasons. T. M. Scanlon, for instance, in his influential book What We Owe to Each Other, uses reasons as the basic, unanalyzable building blocks of his analysis. Even Kantians these days are as likely to talk about reasons as they are about duties. Moral conflicts are now more often seen as being between conflicting moral reasons—reasons to perform incompatible actions. The task of the hapless agent facing such a moral conflict is to judge, if possible, which of her conflicting moral reasons is more compelling, and then to act on that more compelling reason. In this broad picture, the less compelling or second-best reason can continue to be a reason for the agent; but basic rationality requires the agent to act on the best reason she has, not the second best. If she acts against the moral reason she appropriately judges to be the best reason she has, then she is subject to some degree of moral sanction.

The greatest strength of the conflicting-reasons picture over the conflicting-oughts picture, it seems to me, is that it makes the defeasibility of moral considerations much less mysterious. If incompatible moral considerations appear in one’s deliberation as conflicting oughts, then we must account for the ought not acted upon. Does it persist as a morally troublesome residue of the conflict, or does it disappear? Each answer has it problems—moral remainders seem to imply inescapable guilt and blameworthiness, while disappearing defeated obligations cast doubt on whether there was a real moral conflict in the first place. Much can be said, of course; my point is that much needs to be said. A conflict between moral reasons, on the other hand, can clearly be a real conflict even if one reason is more compelling than the other, and it is hard to see how guilt and blameworthiness could attach to acting for a more compelling reason rather than a less compelling one. In short, reasons are more straightforwardly defeasible than obligations.

I will assume in this paper that the conflicting-reasons picture of moral conflict has this advantage. Where, then, does wrongness fit into this new improved theoretical world of reasons rather than obligations? One possibility would be to say that wrongness grounds reasons—that the fact that an action is wrong gives me a reason not to perform it, and that negative moral reasons, thus grounded, can appear in my moral deliberations along with positive moral reasons, otherwise grounded. But bringing in wrongness as a prior element that grounds reasons would complicate the straightforward defeasibility of reasons, and thus defeat the advantage of the conflicting-reasons picture over the conflicting-oughts picture. The only other possibility, if wrongness is not a prior element in moral conflict, is to place wrongness (in our account) after the conflict. We must say, that is, that an agent acts wrongly if she acts against a moral reason she appropriately judges to be the best reason she has.

But if this is our criterion for wrongness, what will be our analysis? I said previously that an agent (in our reasons account of conflict) is subject to some degree of moral sanction if she acts against the moral reason she appropriately judges to be the best reason she has. Acting against the best moral reason, that is, is sufficient for moral "sanctionworthiness". Now it appears that the same condition is also sufficient for acting wrongly. This suggests (though it does not imply) the following analysis:

However natural it may seem thus to connect wrongness and sanctionworthiness, this analysis, as it stands, seems subject to fatal objections. In the balance of this paper I will defend this analysis of wrongness. Whether or not it turns out to be ultimately defensible, it can, at least, survive the initial salvos.

1. Act-Consequentialism

Act-consequentialists can easily put forward examples of wrongdoing that does not seem sanctionworthy. I decide to help an elderly friend by driving him to his doctor’s appointment. As I safely and vigilantly drive along, a pedestrian darts in front of my car. In spite of the perfect condition of my brakes and my best efforts to avoid hitting the pedestrian, I do hit him, and he is killed. My elderly passenger is so shocked by the accident that he succumbs to a fatal heart-attack; and in the following months I fall into a deep depression, turn to alcohol, lose my job, and end up homeless. By any standard, the consequences of my deciding to help my friend were suboptimal; therefore, according to act-consequentialism, my doing so was wrong. But one is hardly inclined to blame me for deciding to help my friend. The usual thing to say about such cases is that while my action was objectively wrong because it had suboptimal consequences, it was not subjectively wrong, since I had no way of knowing what the consequences would be, and is therefore not blameworthy. What such examples are supposed to show, then, is that we cannot link objective wrongness, at least, to sanctionworthiness.

It seems to me, though, that unless one is in the grip of act-consequentialism, one will feel very little inclination to call my benevolent and blameless action wrong in any sense, subjective or objective. The phenomenology of the situation supports this intuition. Suppose that the day after the accident my friends find me sitting in Joe’s Bar and Snooker Parlor, drinking straight scotch and intermittently sobbing. My friends will try to console me by telling me that I did absolutely nothing wrong; and when they tell me this, they will feel that they are telling me the truth. This suggests that the idea that I did something objectively wrong is a revisionist feature of act-consequentialism; it is not really part of our pre-theoretical intuitions. So even if act-consequentialists must accept the objective/subjective wrongness distinction as a cost of holding their theory, there is little reason for the rest of us to accept it.

2. Crisp’s Objection

Another objection arises in the context of act-consequentialism. Suppose someone acts wrongly—he acts, that is, in a way that does not optimize value—but morally sanctioning him under the circumstances would not optimize value either. Would not such a case falsify our analysis of wrongness, under which wrongdoers are "subject to some degree of moral sanction"? Roger Crisp generalizes this objection beyond a consequentialist context:

Crisp’s point seems to be this. If we assume our analysis of wrongness as sanctionworthiness, and assume that there are principles which tell us what we ought to do in terms of rightness and wrongness, we end up with, first, a specification of what we ought to do in terms of rightness and wrongness, and, second, a definition of rightness and wrongness in terms of what we ought to do (praise or blame or punish). We thus have no independent notion either of rightness and wrongness or of what we ought to do. It follows (according to Crisp) that if we are to have independent moral principles and understand what they mean, we should reject our analysis of wrongness as sanctionworthiness.

This objection does have some force in the context of act-consequentialism—a context in which every action, including morally sanctioning someone, is directly subject to moral assessment. But outside act-consequentialism, Crisp’s objection is easily answered. When we say that a wrongdoer is "subject to some degree of moral sanction," we need not take this to mean that sanctions are morally required. We may interpret this to mean, instead, that sanctions are rationally licensed. If we do so, then we are not basing our analysis of wrongness on a prior notion of moral requirement, as Crisp charges, but rather on a prior notion of what is rational. This presents its own challenges, of course, but it at least saves us from Crisp’s charge of circularity.

One might object that we do in fact view the requirement to sanction wrongdoers (and only wrongdoers) as a moral requirement. I do something morally wrong when I sanction someone I know to be innocent of wrongdoing, and I exhibit a moral failing (of perhaps a different sort) when I fail to sanction someone who deserves it. But neither of these facts forces us to conclude that the requirement to sanction wrongdoers and only wrongdoers is itself a moral (rather than merely a rational) requirement. If I sanction someone I know to be innocent of wrongdoing, I am gratuitously harming that person, and the gratuitous harm is a moral wrong. And if I fail to sanction wrongdoers, I display either a lack of moral judgement or a willful lack of concern about wrongdoing, both of which reflect a serious lack of character. Thus, there is no need to conclude that the moral disvalue involved in inappropriately sanctioning or failing to sanction comes from a moral requirement to sanction wrongdoers and only wrongdoers. The moral disvalue can have other sources—the wrongness of gratuitously harming others, or the moral disvalue of a flawed character.

So it seems that we can answer Crisp’s objection to analyzing wrongness as sanctionworthiness, provided, that is, that we are not act-consequentialists, who view all consequential actions in terms of moral requirement. But this, I think, is good enough. Recall that our analysis of wrongness as sanctionworthiness is motivated by a conflicting-reasons picture of moral conflict; and note that act-consequentialism is not very compatible with that picture. (An act-consequentialist deliberator, after all, is not weighing competing reasons, but calculating the value of the available alternatives.) So an act-consequentialist may have reason to reject our analysis of wrongness; but if one is attracted to the conflicting-reasons picture, one has good reason not to be an act-consequentialist.

3. Global Consequentialism

Consequentialists agree that we are morally required to bring about more value rather than less. We can accomplish this by our actions, but also by what rules we internalize and what motives we act on—and probably in other ways as well. Local consequentialism, in Philip Pettit and Michael Smith’s coinage, picks one of these evaluands as the focus of moral requirement, and then defines the rightness of the others in terms of it. So, for instance, act-consequentialism specifies right actions as those that produce optimal consequences, and right motives as whatever motives produce right actions. Or, for another instance, motive-consequentialism specifies right motives as those that produce optimal consequences, and right actions as actions produced by right motives. Pettit and Smith argue that any form of consequentialism that privileges one evaluand and defines the rightness of the others in terms of it will have unconsequentialist implications, and is therefore incoherent. What consequentialists must do, they conclude, is adopt a global consequentialism, in which all categories of evaluand—actions, motives, rules, dispositions, "or whatever"—are directly optimized.

But this evaluative pluralism creates the possibility of conflict. Derek Parfit cites Clare, who has the right, value-optimizing motive of overriding love for her child. This motive causes her consistently to privilege the interests of her child over the interests of strangers. But now she faces a situation in which she "could either give her child some benefit, or give much greater benefits to some unfortunate stranger. Because she loves her child, she benefits him rather than the stranger." (Parfit, p.32.) So Clare’s right motive causes her to perform a wrong action. If we are disinclined to blame Clare for her wrong action, then we appear to have an example of blameless wrongdoing.

This objection raises many issues. Disentangling them all would be the business of a very long paper. To preserve the shortness of this short paper, I will simply cite what I think is a powerful answer from Elinor Mason. Mason argues that we cannot separately optimize different evaluands, as global consequentialism wants. One’s actions and motives and dispositions interact in complex ways, and interact as well with the actions, motives, and dispositions of others. Value can only be determined in context; consequently, the value of individual evaluands cannot be determined separately. Instead, she says,

Though she does not use the term, what Mason seems to be proposing is a holistic consequentialism, rather than a global one. The upshot for our purposes is that if individual evaluands cannot be optimized separately, as I believe Mason successfully shows, then conflicts that produce blameless wrongdoing cannot arise.

1. The Open-Question Argument

Crisp pushes an open-question style objection to J.S. Mill’s analysis of wrongness. Crisp says:

This objection can be applied to our analysis: I can say that a is wrong, and still leave it an open question whether reason requires that a be punished.

This objection can readily be answered. Open-question arguments, to state the obvious, depend on intuitions about openness. In Crisp’s open question argument, he manipulates our intuitions by using the word "punished" rather than the more neutral term "sanctioned." It may feel open whether a wrong act should be punished (by fines, imprisonment, or public floggings?); but it is a much less open question whether a wrong act should be sanctioned, particularly when we recall that the appropriate sanction might be very slight, as little as a mild pang of guilt.

2. Counterexamples

Even if wrongness as sanctionworthiness can survive these theoretical objections, it might still fall victim to counterexamples. Counterexamples might take either of two forms: They might be cases of sanctionworthiness in the absence of wrongdoing, or they might be cases of wrongdoing which is not sanctionworthy. Examples of the first sort seem, as a class, to be counterintuitive; it is irrational on the face of it to morally sanction someone who has done no wrong. Consequently, I will confine my discussion to alleged examples of wrongdoing which is not sanctionworthy.

Proposed counterexamples might cite instances of wrongdoing by children whom we are disinclined to blame, or by people of diminished capacity—people suffering from mental retardation, or from dementia due to old age or illness. Other proposed counterexamples might cite people who perform wrong acts under duress—either the sort of external duress that produces what Aristotle calls "mixed actions," or the internal duress of temptations too great to bear. There might also be examples of people who act wrongly from a deeply, but involuntarily, flawed character or psychology—people warped, say, by a horrifically abusive childhood. And, finally, there is the problem of whether psychopaths are blameworthy for their wrong actions.

I want to suggest two general responses to these sorts of counterexamples. One response is to point out that moral sanctions vary widely in nature and severity, from execution (if one thinks that judicial killing is permissible) all the way down to the mildest pang of guilt. In between these extremes we find every level and combination of internal and external sanctions. This response, all by itself, can largely neutralize counterexamples citing wrongdoing by people we are disinclined to blame—children, mentally diminished people, people under duress, and so forth. We might be disinclined to shun or publicly denounce such people for their wrong action, but that doesn’t mean that no sanction is appropriate. Some mild internal or external sanction might very well be appropriate in such cases.

The second response covers cases of extreme delusion, diminishment, or duress, in which we are disinclined even mildly to sanction the wrongdoer. In such cases we can question, it seems to me, whether the wrong action under consideration is really the agent’s action. Suppose that I am so delusional that I think the mailman is a ferocious grizzly bear about to attack me, so I brain him with a shovel. Regardless of one’s theory of action, it seems reasonable to say that the action called "braining the mailman with a shovel" was not my action, at least not under that description. My action was defending myself from a ferocious grizzly bear, an action which, under most circumstances, is neither wrong nor morally sanctionable.

Whether or not "wrongness as sanctionworthiness" ultimately survives, it at least isn’t dead on arrival, or so I have argued. If one is attracted to the reasons picture of moral deliberation, wrongness as sanctionworthiness is both an intuitively pleasing and (provisionally) a theoretically viable analysis of wrongness.

Roger Crisp, Mill on Utilitarianism (Routledge, 1997).

Elinor Mason, "Against Blameless Wrongdoing," Ethical Theory and Moral Practice 5 (2002), pp.287-303.

Derek Parfit, Reasons and Persons (Oxford University Press, 1984).

Philip Pettit and Michael Smith, "Global Consequentialism," in Hooker, Mason, and Miller, editors, Morality, Rules, and Consequences (Rowman and Littlefield Publishers, 2000), pp.121-133.

Bernard Williams, "Ethical Consistency," in Williams, Problems of the Self (Cambridge University Press, 1973), pp.166-186.