My topic is an aspect of our moral thought that has not to my knowledge received much attention from philosophers. Commonsense morality thinks some acts are right and some are wrong, but it also thinks some wrong acts are more seriously wrong than others. It is wrong to steal a car and wrong to murder, but murder is more seriously wrong than auto theft, which is more seriously wrong than breaking a promise to have lunch. In this paper I examine how one act can be more seriously wrong than another and, when it is, what makes it so. I also ask if there is a parallel concept of more important rightness.
1. Wrongness and Degrees
That one act is more seriously wrong than another is often intuitively compelling in itself; thus it seems self-evident that murder is morally worse than breaking a promise. But judgments about serious wrongness have further implications. If you have acted wrongly you should feel guilt, but you should feel more guilt—more intense or longer-lasting guilt—if your act was more seriously wrong, for example, if it was murder rather than breaking a promise. You are also other things equal more blameworthy for a more serious wrong, and if retributivism is true, you deserve more severe punishment for it. In general, whenever wrong acts call for negative responses, more serious wrongs call for stronger ones. The idea of serious wrongness therefore connects with several other aspects of our moral thought, and this allows further tests of it. To decide whether one act is more seriously wrong than another we can not only consult direct intuitions about the two but also ask whether you should feel more guilt about the first or whether the first makes you more blameworthy or more deserving of punishment.
These tests cannot be applied mechanically, because in each case the effect of more serious wrongness is mixed with others that are not relevant to our topic. Guilt is called for by wrong action, and on at least some views an act's wrongness is independent of its motive (Ross 1930: ch. 1; Scanlon 2008: ch. 2). But more serious wrongs are often done from worse motives, and even if these cannot be the objects of guilt they can prompt the different emotion of shame. Shame about your motivation can then mix with guilt to make for an overall negative response to your act in which the specific role of serious wrongness is harder to see. (If motives are relevant to wrongness, they can prompt guilt as well as shame, but the two can still be hard to pull apart.) Something similar holds for blameworthiness and retribution. On many views you are more blameworthy for a wrong act or deserve more punishment for it if you acted from a worse motive, for example, if you killed from sadistic hatred rather than excessive anger at injustice. Your blameworthiness can also depend on other facts about your mental states, such as whether you were culpably ignorant or acted under duress (for views on which your degree of blameworthiness for a wrong depends both on its seriousness and on facts about your mental states see, e.g., Beardsley [1957: 306–07] and Smith [1983: 556–57]). In all these tests, the effect of serious wrongness on fitting responses is mixed with effects due to your state of mind. Isolating the former effect requires setting these other influences aside.
More serious wrongness may also help to characterize subjective rightness, or rightness relative to your beliefs or evidence. Many philosophers have been persuaded by an example of Frank Jackson that this cannot be done in terms of objective rightness, or rightness relative to the facts; more specifically, the subjectively right act cannot be identified as the one most likely to be objectively right. In Jackson's example you can give a patient one of three treatments. One of the first two will completely cure him and one will kill him, but you do not know which is which; each has a .5 probability of doing either. The third treatment will cure his condition almost entirely and is safe. The subjectively right treatment here is clearly the third, but it is certain to be objectively wrong; one or the other of the first two is right relative to the facts (Jackson 1991: 462–63). But a derivation of subjective from objective rightness need not tell you to maximize your probability of acting objectively rightly or, what is the same, to minimize your probability of acting wrongly. As Peter Graham has argued, it can tell you to minimize your probability of acting seriously wrongly, or to minimize the expected objective serious wrongness of what you do. Since in Jackson's example the act that is certain to be objectively wrong will be only slightly seriously wrong while each of the others has a .5 probability of being horribly so, this yields the desired result (Graham 2010).1 Serious wrongness may also be relevant in cases of moral uncertainty. Imagine that you cannot decide between two moral views and must do either act A or act B, where the first view says A is right and B wrong and the second says the reverse. Andrew Sepielli (2009) has argued that you cannot here consider just the probabilities that the two views are true. If the first says B is only slightly seriously wrong while the second says A is massively so, you should do B even if you think the first view is somewhat more likely to be true.
These last uses of serious wrongness are more controversial. It has been argued that, despite its success with Jackson's example, the proposed account of subjective rightness does not have the implications we want in cases involving permissions, for example, in cases of self-defense or supererogation (Lazar, forthcoming). The account at least needs supplementation to handle these cases. And the account of moral uncertainty requires comparisons of seriousness not only within a moral view, as I will be discussing, but also between moral views, which raises additional difficulties. Nonetheless, these are two further contexts where the concept of serious wrongness may play a role.
Some philosophers have denied that there can be degrees relating to wrongness. Some Stoics, for example, thought that all moral wrongs are equal. Diogenes Laertius reports that they ‘see fit to believe that [moral] mistakes are equal … [so] he who makes a larger [moral] mistake and he who makes a smaller one are [both] equally not acting correctly’ (1998: 199–200). Some present-day philosophers may likewise deny that wrongness admits of degrees. For an act to be wrong, they may say, is for it not to be permitted, and since an act either just is permitted or just is not, it cannot be more or less wrong.
That there is a concept of wrongness that does not admit of degrees does not mean there cannot be one that does. But I have chosen to avoid this issue by speaking not of one act's being more wrong than another –I will concede that that is not possible–but of its having the related but different property of being more seriously wrong, which I understand as follows. Because of the supervenience of moral properties, any act that is right or wrong has other properties that make it so. But if these properties admit of degrees, or if their tendencies to make acts right or wrong do, we can use this fact to define a derivative property of serious wrongness that likewise admits of degrees. Compare the properties concerned with height. There is an initial property of tallness that admits of degrees. By making a cut on the scale of tallness we can introduce a property that does not admit of degrees, such as being over six feet tall in the sense of having some height or other above that. We can then combine these two properties to yield a third that again admits of degrees, that of being more than six feet tall in the sense in which someone who is six feet ten is a lot more than six feet tall, whereas someone who is six feet one is only a little more than that height. I think of the right- and wrong-making properties as analogous to tallness, wrongness as analogous to being at least six feet tall, and being seriously wrong as analogous to being more than six feet tall in the sense that admits of degrees. Being seriously wrong combines underlying properties that can be present to differing degrees with a supervening one that cannot to yield a third property that again can (for similar remarks see Berman and Farrell 2011: 1083–86, 1090–93).
We therefore need a moral view in which the properties that make acts wrong admit of degrees, and the paradigm such view is W. D. Ross's (1930: ch. 2). That view contains a number of prima facie duties, or duties other things equal, each of which identifies a property that tends to make acts right, for positive duties, or wrong, for negative ones. When these duties conflict, they have to be weighed against each other, so an act's rightness or wrongness on balance depends on the comparative strengths of the duties on the two sides, and these strengths can differ. That an act keeps a promise counts a little in its favor, whereas its being a murder counts a great deal against it; if these were the act's only relevant properties, it would be all things considered wrong. But crucial to Ross's view is a concept of prima facie duty that makes acts right or wrong and admits of degrees; it provides the needed material for judgments of serious wrongness.
2. Absolute Strength versus the Gap
The concept of prima facie duty allows a straightforward account of serious wrongness in simple cases, where an act violates one or more prima facie duties and so has one or more wrong-making properties, but has no right-making ones. Then how seriously wrong the act is depends just on the strength of the duty or duties it violates. If one act breaks a promise while another commits murder and neither fulfils any positive duty, the second act is more seriously wrong.
But in more complex cases a wrong act does fulfill some duty, or have some, albeit outweighed, right-making property. Then there are at least two ways of determining its seriousness as wrong. On what I will call the absolute strength view this depends just on the strength of the prima facie duty—for now assume it is just one—whose violation makes the act wrong. If it is wrong because it breaks a promise or commits murder, how seriously wrong it is depends just on the strength of the duty not to do that; no other duties, and in particular none the act fulfils, are relevant. An alternative view, which I will call the gap view, looks at the difference in strength between the duty whose violation makes an act wrong and the duty—again assume it is just one—that the act fulfils. The issue now is how much stronger its wrong-making tendency is than its right-making one, so duties other than the one it violates are relevant. These two views yield identical verdicts in simple cases, but they can come apart in complex ones. Imagine that one act violates a prima facie duty of strength ten and fulfils one of strength six, while another violates a duty of strength six and fulfils one of strength one. On the absolute strength view the first act is more seriously wrong, since ten units of strength are more than six. But on the gap view the opposite is true, since the gap between six and one is larger than the gap between ten and six. My question is which of these two views is preferable. Is it better to assess the seriousness of wrongs by absolute strength or by gap means?2
Both views require cardinal and even ratio-scale measures of the strengths of prima facie duties, but I think the intuitions that enable us to rank duties against each other also give us at least rough measures of these kinds. Surely we can say the duty not to murder is not just stronger than the duty not to break a promise but a lot stronger and even more than ten times stronger. Moreover, a view like Ross's already needs cardinal measures at two points. One is when you are not certain whether an act will fulfill a certain duty—say, to keep a promise or save lives—and can only assign some probability to its doing so; that requires a cardinal measure to multiply the probability by. The other is when several duties tell for or against an act, and you have to add their strengths to determine their combined or overall weight. But if a view already has at least rough cardinal measures, the absolute strength and gap views can come apart. Which should we then prefer?
Consider first a case involving just consequentialist duties, or duties to promote the good, one where you do not fulfill your strongest such duty but do fulfill a weaker one. Imagine that you ought all things considered to save ten people from drowning, but from a selfish desire not to miss the start of a TV show you save only six. You act wrongly, but it would surely be a mistake to make the seriousness of your wrong depend just on the number of people you should have saved, as an absolute strength assessment would. To note only that you did not save all ten people is to ignore the fact that you did save six. This is especially counterintuitive if the six were part of the ten, so you started saving the ten but broke off when your show was about to start. Then you clearly did fulfill part of the duty you violated. But it is also counterintuitive if the two groups were separate, so instead of saving ten people in one lake, you chose to save six in another lake because that would take less time. Here, too, the absolute strength view ignores the fact that, though you did not fulfill all your duty, you did fulfill some of it. We can put this argument against the absolute strength view as follows. To make the seriousness of your wrong in consequentialist cases depend only on the strength of the duty you violate is to make your act when you save six people rather than ten no less seriously wrong than if you had saved none; it is to equate a partial failure with a complete moral failure. The gap view avoids this result, since by subtracting the number of people you did save from the number you should have saved it gives you some credit for acting partly as you ought. It therefore seems preferable in these cases.
This conclusion may seem obvious, and some may want to apply the gap view across the board, so the seriousness of all wrongs depends on the difference between the strengths of the duties whose violation makes them wrong and the strengths of any duties the acts fulfill. This is certainly a defensible view, but to me the gap view is more questionable when the duty you violate is deontological, or not a matter of promoting the good.
Consider first a case suggested to me by Brad Cokelet (2013). You steal someone's money, which you ought all things considered not to, but you do so in order to pay a debt you owe a third person, and then pay the debt. On the gap view the fact that you fulfilled your duty to pay the debt should make your act of stealing less seriously wrong than it would otherwise be and by a considerable amount. After all, your duty to pay the debt, though outweighed, is substantial—it may be just slightly weaker than the duty not to steal—and your act contributes to fulfilling that duty. But intuitively, at least to me, the fact does not have this effect, or not this large a one. Unlike in consequentialist cases, there is some intuitive pressure here to look only at the absolute strength of the duty you violated, so your stealing is no less wrong despite the debt-payment it led to, and some may endorse that verdict. Or consider some familiar cases from the literature. You are a doctor with five patients who need organ transplants, but there are no organ donors. So you kidnap a homeless person, kill him, and divide his organs among the five, thereby saving their lives. Or you push a fat person in front of a trolley to stop it from hitting five. Since your duty to save the five has considerable weight in these two cases, that you fulfilled it should, on the gap view, make your killing the one considerably less seriously wrong.3 But that again does not seem right; there is again some appeal in saying your act is no less seriously wrong. And there is further support for this view in the additional tests I described. Should you feel significantly less guilt for killing the one when you saved the five? To me it seems not. Are you significantly less blameworthy or deserving of punishment? Again, arguably not.
As I said, these tests can involve confounding factors. If you killed the one from benevolent concern for the five, your motive was better than in most killings, and you therefore have some less ground for shame. But your ground for guilt, which concerns only your act's wrongness, seems not much diminished. Also because of your motive, you may be somewhat less blameworthy or deserving of punishment, but the reductions again do not seem large. In a modern legal system, even one that is entirely retributively based, you would surely receive the standard punishment for murder, with only minor mitigation. All this suggests that what determines the seriousness of your wrong in these cases may be only or mainly the absolute strength of the duty not to kill and not, or not so much, the gap between that duty and some other duty you fulfill.
Other cases that resist a gap treatment involve tragic moral conflicts as discussed by Bernard Williams (1973) and illustrated by Agamemnon's dilemma in the play by Aeschylus. Agamemnon has a strong duty as a father not to kill his daughter and, at least in the story, an equally strong duty as a military leader to enable the Greek expedition to Troy to sail. But given the demands of a hostile goddess, he cannot fulfill both duties and so, no matter what he does, will do something all things considered wrong. He chooses to kill his daughter, and though a gap assessment may say his act is wrong, it cannot find it seriously wrong, because the act fulfills a duty roughly equal in strength to the one it violates. But this is not Williams's or Aeschylus's view. They think Agamemnon's act is horribly wrong, and in doing so must be looking mainly or only at the absolute strength of the duty he violates.
I will not place much weight on these cases, which are in two ways controversial. Not only do some philosophers deny that tragic conflicts are possible, but others may argue that such conflicts arise only in moral views very different from any we could accept today, views where some duties make acts wrong directly and without having to be weighed against other duties, so what is true in these conflicts does not generalize to ones we might face. Still, the cases are at least suggestive.
The absolute strength view also seems implicit in a claim Aquinas made while arguing, against the Stoics, that not all sins are equal. Since his concept of sin was not purely deontic, one claim was that an act can be more sinful if it is done from a worse motive (1981: I–II, q., 73, aa. 6,8). But another claim concerned only wrongness; it said the gravity of a sin varies with the subject-matter of the duty it violates, so a sin about property is less serious than one affecting human life, which is less serious than one concerning God (1981: I–II, q. 73, a. 3). But in making this claim Aquinas considered only the duty an act violates, with no mention of any duty it may fulfill, for example, of any good it may, however wrongly, promote. And the examples he used to illustrate his claim were all deontological: he said theft (not failing to protect another's property) is a lesser sin than murder (not than failing to save another's life), and murder a lesser sin than blasphemy (not than failing to positively honor God). He made what look like absolute strength claims and made them about deontological wrongs.
We can also consider cases where what is outweighed is a permission, beginning with a consequentialist case. Imagine that you have a prima facie duty to promote the happiness of all persons impartially, but also have an independent prima facie permission to pursue your own happiness.4 The result of weighing the two against each other is that you are permitted to prefer a unit of your own happiness to as many as five units for other people, but not to more than five, but you give yourself one unit of happiness rather than ten to others. Here it seems wrong to make the seriousness of your wrong depend just on the fact that you did not give the ten to other people, as an absolute strength assessment would. We should instead apply the gap view and subtract the five for others you could permissibly have forgone, leaving the five beyond that threshold that you failed to give. Again, in a consequentialist case, what seems to matter is the difference between the duty whose violation made your act wrong and a contrary factor, now a permission, that it outweighed.
A parallel deontological case involves self-defense. You are attacked by an aggressor and in response are permitted to cause him somewhat more harm than he threatens you with if that is necessary to defend yourself. But despite knowing this is wrong, you cause him significantly more harm than that, say, twice as much. Is the seriousness of your wrong affected by the fact that you were permitted to cause the lesser harm? Our judgment here may be affected by mental-state factors, because if the attack aroused you emotionally and limited your ability to control yourself, that may reduce somewhat the blame and punishment you deserve. But we need to set these factors aside and ask whether there is a further reduction for your permission to defend yourself, so we subtract from the harm you did cause the amount you were permitted to cause. That seems more questionable. Certainly, if you were not at all aroused but caused the excessive harm in cold blood—say, from malice—your wrong does not seem much if at all diminished. Again, and unlike in the parallel consequentialist case, there is some intuitive pressure toward an absolute strength assessment of a deontological wrong.5
3. A Mixed View
This appeal is limited, however, because a pure absolute strength treatment of these wrongs also has counterintuitive implications. They emerge when we consider, not just one deontological wrong, but pairs or sets of them.
Imagine that a doctor with five patients needing transplants has two alternatives to letting them all die. He can kill one innocent person and use his organs to save all five, or kill a different person with fewer healthy organs and use them to save two. On the absolute strength view these two acts are equally seriously wrong. Both violate the same duty not to kill, and though the first better fulfils the duty to save lives, that is irrelevant to its seriousness as wrong. But is killing one to save five, though very seriously wrong, not at least somewhat less so than killing one to save two? For a more extreme case, imagine that to be permitted, an act of killing an innocent person must save at least twenty lives, and a doctor kills one to save nineteen. The absolute strength view says his act is no less seriously wrong than if he killed one to save five, two, or even none. That is hard to accept.
A defender of the absolute strength view may try to explain these intuitions away. He may say a doctor who kills one person to save five acts from a better motive than one who saves two, which is why we think less badly of his act. A doctor who kills one to save nineteen has an even better motive, as well as the excuse that it is hard to know exactly where the threshold for permissible killing lies. But we can stipulate that the doctor in this latter case knows his killing is wrong, and we can neutralize the effect of motives by imagining that all three doctors kill only in order to get paid. In any case, none of the explanations touch the intuition, which I at least have, that these acts of killing differ to some extent just deontically, or in their seriousness just as wrongs, the agents’ mental states aside.
One response to these cases is to revert to the gap view and apply it across the board, and I have said that is defensible. But another response is to adopt a mixed view of deontological wrongs and assess their seriousness by both absolute level and gap means, with some weight for each. If you kill one to save five, the seriousness of your wrong then depends partly on the strength of the duty not to kill on its own and partly on the difference between its strength and that of the duty to save five, so your act is less seriously wrong than on pure absolute strength grounds but more seriously so than given just the gap. To illustrate: if killing an innocent person is permitted only if it will save at least twenty lives, the duty not to kill is equal in strength to the positive duty to save twenty; this must be so if they neutralize each other. The absolute strength view then says killing one to save five is as seriously wrong as failing to save twenty, while the gap view says it is as seriously wrong as failing to save twenty minus five, or fifteen. On a mixed view the act's seriousness is somewhere between these two: it is equal to failing to save seventeen and a half lives if the two factors have equal weights or to not saving a few more or less if one or the other counts for more. But rather than just one of the two measures, a mixed view uses both.
This type of view seems to fit many of the intuitions about deontological cases I cited above. By giving some weight to the gap, it agrees that killing one to save five is less seriously wrong than doing so to save two, and killing one to save nineteen is less seriously wrong still. But by also caring about absolute strengths, it captures the judgments that counted against the pure gap view. In proposing these I did not say the fact that a killing saved five lives does nothing to make it less seriously wrong. I said only that it does not make it significantly less serious, or less serious by as much as a pure gap assessment would. That is captured if gap considerations are combined with absolute strength ones, so the difference in seriousness from the gap is reduced by the lack of difference in absolute strength.6 And the resulting view can be seen to reflect a more abstract idea.
If you kill one person to save one life fewer than would make this permitted—say, nineteen rather than twenty—the gap between your act and one that would be right is one life saved. But on a mixed view the absolute seriousness of your wrong does not depend just on this gap. It also depends on the absolute strength of the duty not to kill and is therefore greater than the difference in seriousness between killing one to save three and doing so to save two, and is also greater than the absolute seriousness of failing to save one when you should do that. Though the gaps in these last cases are the same as in the first—one life—the resulting difference in seriousness is smaller.
This is because, on a mixed view, the gap that separates right from wrong in deontological cases has a significance not shared by other equally sized gaps. Moving across this gap, for example, by saving nineteen rather than twenty when you kill one, introduces a consideration, here the absolute strength of the duty not to kill, that is present on both sides of the comparison between killing one to save three and doing so to save two, and therefore cannot distinguish these acts, and that is not relevant to a case of just failing to save one. So when an act violates a deontological duty, its being wrong matters in itself or makes an independent difference. This is not because the act's wrongness is an additional reason not to do it, over and above the factors that make it wrong, as some philosophers have recently argued (Scanlon 1998: 11; Scanlon 2007; Darwall 2010). The reasons against the act are or can be just those given by the prima facie duties it violates, so the effect we are considering is only on the separate issue of its seriousness as wrong. Nor is it because wrongness always makes the same contribution to the seriousness of deontological wrongs. If you break a promise in order to do a little less good than would make that permissible, what is introduced is the absolute strength of the duty not to break promises, which is much less than that of the duty not to kill. Wrongness therefore contributes indirectly rather than directly to the seriousness of deontological wrongs by making another factor relevant, which can have more or less weight. In a recent discussion Derek Parfit has said that killing one to save one fewer life than would make this permissible is only ‘slightly wrong’, so there is only a ‘weak moral reason’ not to do it (2017: 346, 341), which presumably implies that you should feel only slight guilt about doing it. This seems to apply something like the gap view and to give the act's wrongness no independent weight. In contrast, a mixed view does give wrongness such weight, or takes it to play a special role in deontological cases.
This is in contrast with consequentialist cases, where wrongness plays no such role. If the seriousness of a consequentialist wrong depends just on the gap, the difference between saving twenty when you should and saving just nineteen, which is the absolute seriousness of saving just nineteen, is the same as the difference between saving three and saving two when you should save twenty and also the same as the seriousness of failing to save one when you should save one. If all that matters for consequentialist duties is the amount of good you produce, equal differences in that are equally significant; in consequentialist cases, wrongness does not matter in itself.
All this suggests that mixed views may reflect the more abstract idea that when you act wrongly by violating a deontological duty, your act's being wrong has a kind of significance it would not have if you violated a consequentialist duty. To me this idea has some appeal: wrongness does seem to matter for deontological duties in a way it does not for consequentialist ones. And the idea is captured if their being wrong affects the absolute seriousness of deontological wrongs as it does not affect that of consequentialist ones.
I am, then, proposing a partly asymmetrical view of serious wrongness. When an act is wrong because it fails to produce as much good as it should, its seriousness is determined by just the gap. But when it violates a deontological duty, its seriousness depends, differently, on both the gap and that duty's absolute strength. And this asymmetry can affect the use of serious wrongness to characterize subjective rightness or deal with moral uncertainty. More specifically, it can favor deontological conclusions in both these contexts.
Imagine, to take a case of moral uncertainty, that by killing one innocent person you can save five others. You know a consequentialist moral view says doing so is right and a deontological one says that it is wrong, but you are uncertain which view is true and find each equally credible. Despite this, I think some will have the intuition that you should care more about avoiding the deontological wrong and therefore not kill the one person (this seems to be part of the thrust of the titular principle, proposed for cases of moral uncertainty, in Guerrero [2007: 78–87]), and an asymmetrical view of seriousness gives some support to this verdict. If consequentialism is true and you do not save the five, the seriousness of your wrong is mitigated to some degree by the fact that the one is still alive. More specifically, that person's being alive counts fully in favor of your act or is fully subtracted in an assessment of its seriousness. But if deontology is true and you kill the one, there is only partial mitigation; that five lives were saved counts only partly against the seriousness of your wrong, because the gap is only one consideration. There is therefore a kind of thumb on the deontological side of the scale, a factor that favors deontological over consequentialist conclusions in cases of moral uncertainty. I do not assume everyone will think you should prefer avoiding the deontological wrong in these cases nor that an asymmetrical view gives the only or best explanation of why you should do so. But if there is full mitigation of consequentialist wrongs but not of deontological ones, that does something to favor deontological conclusions in cases of incomplete knowledge.
I have connected a mixed approach with certain intuitions, but these intuitions may not be universally shared. The idea of more serious wrongness connects with other aspects of our moral thought, but it is not as central to that thought as simple rightness and wrongness and we do not as often make judgments about it. As a result, our intuitions about it may have less weight. In addition, the view I have proposed is theoretically untidy. Not only does it treat the two types of wrong differently, but its account of deontological ones is mixed rather than simple. Those who greatly value theoretical unity may therefore prefer a pure gap view, and I have said that is a defensible position. My aim has been less to identify the one true view about serious wrongness than to describe some possible views about it. I think there is some intuitive pressure toward an absolute strength treatment of deontological wrongs, and a partly asymmetrical view accommodates that. But if some prefer to resist that pressure and apply the gap view across the board, I have no decisive argument against them.
4. Wrongness and Degrees, Again
The views about serious wrongness I have discussed can all arise in a view like Ross's, whose wrong-making properties admit of degrees in a thoroughgoing way. But not all moral views share this feature. Some include wrong-making properties that do not admit of degrees or that do so only in some contexts but not in others. In these views some judgments about serious wrongness cannot be made.
Consider, for example, the first formulation of Kant's categorical imperative, especially in the ‘contradiction in conception’ version that is supposed to yield perfect duties, such as those against making lying promises and suicide. This version says an act is wrong if the universalization of its maxim cannot even be thought, because the situation in which everyone acts on the maxim is self-contradictory (1998: 33). On a straightforward reading this property does not admit of degrees, since a situation either just is self-contradictory or just is not: there cannot be more or less contradiction. It seems to follow that, according to this part of Kant's view, all acts that violate perfect duties have the same wrong-making property to the same degree, and none can be more seriously wrong than any other (Calder 2005). The same is true of the second formulation of the categorical imperative, which grounds the perfect duties in a prohibition against treating others merely as means (Kant 1998: 38), where this again is something you either just do or do not; to treat another even a little as an end is not to treat her only as a means. On both formulations making a lying promise to have lunch seems to be just as seriously wrong as murder and to deserve the same punishment, which for Kant presumably is death.
A second such view has deontological wrong-making properties that admit of degrees more than Kant's but less than Ross's. Whereas Ross avoided the absolutist view that killing an innocent person is always wrong by positing an independent duty to promote the good that can sometimes outweigh the duty not to kill, this alternative view does so by building exceptions into its duties, saying it is wrong to kill an innocent person except when this will save, say, at least twenty lives, and wrong not to save up to nineteen lives except when this will involve killing. The resulting duties do not conflict, as Ross's do, but instead are formulated so they harmonize. Henry Sidgwick thought this is the form a coherent deontology must take (1907: 341), and it was defended, against Ross, by Alan Donagan (1977).
Unlike Kant's, this second view can distinguish more and less serious wrongs in what I have called simple cases. It can say the duty not to kill, with its exceptions built in, is stronger than the duty to keep promises, so its violation calls for more guilt, blame, and punishment. But it cannot make differentiating judgments in more complex cases, such as those that counted against a pure absolute strength treatment of deontological wrongs. A doctor who kills one person to save five and a doctor who kills to save two both violate the duty not to kill except to save at least twenty, and they do so equally. Since no other duty bears on their choice, since the exception in the duty to save lives makes it not applicable, on this view their acts come out as equally seriously wrong. The judgment that the second doctor's wrong is more serious cannot be made.
If serious wrongness requires wrong-making properties that admit of degrees, moral views in which this is not true to the same extent as in Ross's may not allow all the judgments of seriousness we want to make. To me this counts significantly against these views.
5. More Importantly Right?
I now turn to a related topic: whether there is a parallel concept that admits of degrees on the side of rightness, one of being, as we can say, more importantly right. This is a less familiar concept than more serious wrongness, and some have argued that there is no room for it: whereas there can be differences of degree relating to wrongness, there cannot be such differences relating to rightness. I will argue, to the contrary, that there is a concept of more important rightness though it plays a more limited role in our thought.7
Graham has put the argument against the concept as follows (2010: 99). Whereas in a given choice situation there can be many acts that are wrong, with some further from being right than others, there is often just one act that is right, which cannot differ from itself in importance. Even if several acts are right, because none is less well supported by prima facie duties than the others, they do not differ from each other with respect either to rightness or to the properties that make for it. So while there can be differing degrees of seriousness among wrong acts, there cannot be similar differences among right ones.
This argument succeeds for ‘within-choice’ comparisons, which concern acts in the same option-set or available to the same person at the same time. But our discussion of serious wrongness did not consider only these contexts. It did not say only that when you can either break a promise or commit murder rather than do some other right thing, murder is the more seriously wrong alternative. It also said that murder chosen from one set of options is more seriously wrong than promise-breaking chosen from another, and so calls for more guilt, blame, and punishment: it made ‘between-choice’ comparisons. And even if there cannot be rightness-related differences of degree within an option-set, there can be such differences between sets. It is possible that choosing to save ten lives from one set is more importantly right than choosing to keep a minor promise from another. This does not mean the first act is more right. Like simple wrongness, I am conceding, simple rightness does not admit of degrees. But just as the prima facie duties that make an act wrong can differ in strength, so can the duties that make an act right. If saving ten lives in one situation fulfils a stronger duty than keeping a promise in another, it can in principle be more importantly right. In between-choice contexts, there is no bar to more important rightness.
What matters, though, is not whether this concept can be defined. It is whether it plays a role in our ethical thought similar to that of serious wrongness, by helping us to choose in various kinds of uncertainty or by warranting stronger responses to right action, where these would now be the positive ones of pride, praise, and perhaps reward. If an act was more importantly right, you would be justified in feeling more pride in it, as would others in praising you more and maybe giving you a greater reward. Does important rightness play either of these roles?
I take Graham's argument to show that it cannot play the first role. If in any choice-situation there is either one right act or a number that are equivalently supported by prima facie duties, there are no differences in importance that could guide any choice. This leaves the second role of calling for stronger positive responses, but here too there is a contrary argument. It simply denies that these responses are ever called for by right acts as right and can be called for more by some than by others. More specifically, it says that whereas you should feel guilt after acting wrongly, it is not fitting to feel pride after acting rightly. Right action is just what is expected, what you simply should do, and pride in it is therefore inappropriate. You likewise do not deserve praise or any reward for a right act as you deserve blame and punishment for a wrong one. A. C. Ewing noted this asymmetry, saying ‘there are many morally obligatory actions which we should not admire a man for performing, though we should blame him for not performing them’, so ‘the consciousness of obligation seems to be more closely connected with the notions of shame and disapproval as fitting “sanctions” if we do not do what we ought than with the notion of approval if we do’ (1947: 169). The recent literature on ‘reactive attitudes’ has likewise focused much more on negative responses, such as resentment and blame, than on positive ones such as praise (Strawson 1962; Wallace 1994).
This argument need not deny that positive responses to action are ever appropriate. It can say that if you save ten people from a benevolent concern for their welfare, you deserve praise and perhaps some reward for your motive. You can likewise deserve praise if you act rightly despite unusually strong temptations or if recognizing what is right in your situation required special moral acuity. It is only rightness as such or what makes for it that warrants no such responses. The argument can also allow a deontic ground for these responses, but say it is supererogation rather than any related to rightness. An act is right, it can say, if it meets a standard defined on the scale of prima facie duties and wrong if it falls below it. Since an act can fall further or less far below this standard, it can be more or less seriously wrong. An act can also exceed the standard, in which case it is supererogatory and can warrant pride, praise, and even reward. Moreover, it can exceed the standard by more or less and be more or less supererogatory. But a supererogatory act is not more or more importantly right; it is beyond rightness and therefore in a different deontic category. So while there is a positive deontic concept that admits of degrees and can justify positive responses, it is supererogation rather than any related to rightness.
Supererogation can indeed justify positive responses, but it does not follow that rightness cannot do so too. An act is supererogatory if it not only fulfils a prima facie duty but also has significant costs for you, so the duty is outweighed by your prima facie permission to pursue your own good; this is why the act is beyond rightness (for this account of supererogation see Hurka and Shubert 2012). As I have argued, however, an act that is merely right, because it involves no or minimal costs for you, can nonetheless have features that make it importantly right, such as fulfilling an absolutely strong duty or one that is much stronger than any it violates. This can in principle warrant pride and praise, and Ewing, for one, did not deny this. He said only that our deontic concepts are ‘more closely’ (1947: 169) associated with negative than with positive reactions, not that they are never associated with the latter. There may still be some occasions for positive responses to rightness.
I do not see any call for these responses when the duty you fulfill is deontological. Not killing, not lying, not breaking a promise—these acts are indeed just expected, and pride in or praise for them would be unfitting. If you acted from an especially good motive or despite great temptations, you may properly feel some of the pride that is the positive analogue of shame, but your act's deontic character warrants no positive response from you or anyone else. For deontologically right acts, the skeptics’ view seems correct.
But things may be different if your act fulfils a consequentialist duty, in particular by doing a great deal of good. Consider people who are especially well placed to do so: political leaders. In assessing a leader's choice-making through her career, we will not credit her positively for not murdering or lying, though we will blame her if she did those things. We will instead count in her favor those right acts that affected other people, mainly but not only her citizens, for good. But we surely will not count all such acts equally. If she was right in a minor administrative appointment and also right on a major issue such as war or national health care, we will count the latter choices for more. Just as a wrong minor appointment matters less for her record than a wrong decision about war, so does a right minor appointment. The political scientist Richard Neustadt reportedly held that though any leader will make many right and many wrong decisions, in judging her we should ask, ‘Was she right or wrong on the major issue of her day?’ This view is exaggerated if it tells us to consider only the one issue, but it is surely correct that we should care more about right decisions on large issues than on trivial ones. There is more to credit in a leader who was right about war or health care but wrong in some minor appointments than in one whose record was the reverse.
It may be objected that these judgments are only instrumental, concerning just the effects of a leader's acts rather than anything related to their rightness. But importantly right acts can also evoke the emotional responses I have mentioned. Former Canadian prime minister Jean Chretien has often expressed pride in his decision, which I will assume was correct, to keep Canada out of the 2003 Iraq War. This decision was not supererogatory for him; if anything, given Canadian public opinion, it benefited him politically. Nor does he describe it as epistemically difficult; he says he had a simple test for participating in war that the Iraq War did not pass. But his pride seems nonetheless justified because of the significant harms his choice prevented, for example, to Canadian soldiers. A politician may likewise rightly feel proud of, and deserves praise for, instituting a national health care program even if that was not costly for her: again the amount of good done warrants the response. Then-US Vice President Joe Biden was surely responding to this factor when he called the signing of the Affordable Care Act a ‘big fucking deal’. He was proud in part because of the bigness.
We also sometimes reward right action. The Nobel Peace Prize is often given for supererogatory acts, like those of Nelson Mandela and Malala Yousefsai. But Mandela shared his prize with F. W. de Klerk, who had a moral duty to end apartheid and was not doing anything supererogatory when he helped to do so. Yet given his act's huge benefits, his prize seems appropriate. The Nobel Peace Prize has also been given to leaders who negotiated the end of a war or serious conflict, such as, albeit controversially, Henry Kissinger and Le Duc Tho in 1973 and Yasser Arafat, Yitzhak Rabin, and Shimon Peres in 1994. Their acts too were not supererogatory; it is any leader's duty to end a conflict that is pointlessly taking lives. Yet, given the great harms their acts were at least thought to have prevented, their prizes again seem fitting.
These claims about important rightness should not be exaggerated. Even when a right act does significant good, it may call for weaker positive responses than the negative ones warranted by a serious wrong, and the call for them may be less strong. Right acts that produce just minor goods may warrant no praise or reward—only ones above a threshold may—and important rightness on its own may be insufficient to do so. If by merely pressing a button you can save a hundred lives and then do, you may have no ground for pride. That may require, in addition to significant benefits for others, at least some effort on your part, either to identify what is right or to do it. My claim is just that when some such effort was exerted, how much good your act did can affect the fitting response. A politician who rightly avoids a war or introduces a major social program may properly feel a degree of pride that would not be justified by a less consequential act even if the latter posed the same epistemic challenge and had the same effects on her interests.
I am, then, proposing another asymmetry. When a right act fulfils a deontological duty, the concept of important rightness is idle because it does not have further implications like those of serious wrongness. But when an act fulfils a consequentialist duty, in particular by doing a great amount of good, there can be some such implications, if not for choice in uncertainty then sometimes for how much pride, praise, or reward is warranted. There is, however, a complication.
When an act is wrong because it violates a consequentialist duty, its seriousness is determined by the gap view, and we might expect the same to hold for important rightness, so what matters is the difference between the good you did and some lesser amount you could have done instead. But this seems not to be so. Imagine that in the case where you can save either ten people from drowning in one lake or six people in another, you rightly save the ten. A gap assessment would say your act's importance is ten lives minus six or less than it would be if there were not the option of saving the six. That does not seem right. It seems irrelevant to the pride or praise your act warrants that you could have done something less good. Or imagine that a politician can do one of two things but not both: introduce national health care, which will do a great deal of good, or reform education, which will do slightly less good. If she rightly chooses to introduce health care, the gap view says her act's importance is the difference between the good it did and what would have resulted from improving education, which is less than if she had had just the health care option. That again seems wrong. Whereas the seriousness of wrongly violating a consequentialist duty depends on the gap, the importance of rightly fulfilling it seems to depend just on that duty's absolute strength or just on the amount of good it by itself did.8
There are, to be sure, difficulties in determining just how much good this is. We need to compare the state of the world after the act with some baseline state, but which baseline state? Is it the status quo before the act, so what matters is the change the act produced? Or is it what would have been the case had no act been performed, for example, if neither you nor anyone else had been present and able to act? The latter sounds more promising, but what about morally right omissions, such as Chretien's decision not to join the Iraq War? Must the relevant baseline here not involve some positive act of participating? And what about overdetermination cases, where if you had not produced a good outcome someone else would have? Does your act still get credit for all the good that resulted, just some, or perhaps none? These are difficult questions, but they are general ones about all consequentialist duties rather than specific to important rightness, and they arise equally for important wrongness. If an act was wrong because it failed to produce some good, that is because it failed to produce an outcome that was better than some baseline, which again must be identified. I will assume that whatever account of this matter works best in other contexts can be extended to important rightness, so the same baseline applies there. My point is just that, in contrast to what grounds serious wrongness, the importance of a consequentially right act seems to depend only on the good the act did relative to this baseline, rather than on a comparison between that good and some lesser amount you could have done relative to the same baseline.
I have not argued that important rightness is as central an ethical concept as serious wrongness, only that it plays a more qualified and limited role. But it is part of a more general phenomenon. Commonsense morality does not just divide acts into those that are right and those that are wrong; it also makes distinctions of degree. It thinks some wrong acts are more seriously wrong than others and, at least in some contexts, that some right acts are more importantly right. A full account of this morality should address these claims about degrees and explain what they involve, what their ground is, and what ethical role they play. This paper has tried to make a start on these questions.
Aquinas, St. Thomas. (1981) Summa Theologiae. Translated by Fathers of the English Dominican Province. Westminster, MD: Christian Classics.
Beardsley, Elizabeth L. (1957) ‘Moral Worth and Moral Credit’. Philosophical Review, 66, 304–28.
Berman, Mitchell N., and Farrell, Ian P.. (2011) ‘Provocation Manslaughter as Partial Justification and Partial Excuse’. William and Mary Law Review, 52, 1027–109.
Brown, Campbell. (2016) ‘The Rightest Theory of Degrees of Rightness’. Ethical Theory and Moral Practice, 19, 21–29.
Calder, Todd. (2005) ‘Kant and Degrees of Wrongness’. Journal of Value Inquiry, 39, 229–44.
Darwall, Stephen. (2010) ‘But It Would Be Wrong’. In Paul, Ellen Frankel, Miller, Fred D. Jr., and Paul, Jeffery (eds.), Moral Obligation (Cambridge: Cambridge University Press), 135–57.
Donagan, Alan. (1977) The Theory of Morality. Chicago: University of Chicago Press.
Ewing, A. C. (1947) The Definition of Good. London: Routledge & Kegan Paul.
Gert, Joshua. (2004) Brute Rationality: Normativity and Human Action. Cambridge: Cambridge University Press.
Graham, Peter A. (2010) ‘In Defense of Objectivism about Moral Obligation’. Ethics, 121, 88–115.
Guerrero, Alexander A. (2007) ‘Don't Know, Don't Kill: Moral Ignorance, Culpability, and Caution’. Philosophical Studies, 136, 59–97.
Hurka, Thomas, and Shubert, Esther. (2012) ‘Permissions to Do Less Than the Best: A Moving Band’. In Timmons, Mark (ed.), Oxford Studies in Normative Ethics, vol. 2 (Oxford: Oxford University Press), 1–27.
Jackson, Frank. (1991) ‘Decision-Theoretic Consequentialism and the Nearest and Dearest Objection’. Ethics, 101, 461–82.
Kant, Immanuel. (1998) Groundwork of the Metaphysics of Morals. Translated by Gregor, Mary. Cambridge: Cambridge University Press.
Laertius, Diogenes. (1998) ‘Lives of the Philosophers’. In Inwood, Brad and Gerson, L. P. (eds.), Hellenistic Philosophy, 2d ed. (Indianapolis, IN: Hackett), 190–203.
Lazar, Seth. (Forthcoming) ‘Deontological Decision Theory and the Grounds of Subjective Permissibility’. In Timmons, Mark (ed.), Oxford Studies in Normative Ethics (Oxford: Oxford University Press).
Lockhart, Ted. (2000) Moral Uncertainty and Its Consequences. New York: Oxford University.
Parfit, Derek. (2017) On What Matters. Vol. 3. Oxford: Oxford University Press.
Peterson, Martin. (2013) The Dimensions of Consequentialism: Ethics, Equality and Risk. Cambridge: Cambridge University Press.
Olsen, Kristian. (2018) ‘Subjective Rightness and Minimizing Expected Objective Wrongness’. Pacific Philosophical Quarterly, 99, 417–41.
Ross, W. D. (1930) The Right and the Good. Oxford: Clarendon Press.
Scanlon, T. M. (1998) What We Owe to Each Other. Cambridge, MA: Harvard University Press.
Scanlon, T. M. (2007) ‘Wrongness and Reasons: A Re-examination’. In Shafer-Landau, Russ (ed.), Oxford Studies in Metaethics, vol. 2 (Oxford: Oxford University Press), 5–20.
Scanlon, T. M. (2008) Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press.
Sepielli, Andrew. (2009) ‘What to Do When You Don't Know What to Do’. In Shafer-Landau, Russ (ed.), Oxford Studies in Metaethics, vol. 4 (Oxford: Oxford University Press), 5–28.
Sidgwick, Henry. (1907) The Methods of Ethics. 7th ed. London: Macmillan.
Smith, Holly M. (1983) ‘Culpable Ignorance’. Philosophical Review, 92, 543–71.
Strawson, P. F. (1962) ‘Freedom and Resentment’. Proceedings of the British Academy, 48, 1–25.
Wallace, R. Jay. (1994) Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press.
Williams, Bernard. (1973) ‘Ethical Consistency’. In Williams, Problems of the Self (Cambridge: Cambridge University Press), 166–86.