Wrongs and Faults
Wrongs and Faults
Abstract and Keywords
This chapter begins with a discussion of ‘the elementary moral distinction’. It then considers lives and wrongs, people and faults, fault-anticipating wrongs, and the fault principle. It argues that moral philosophy has lost sight of the need to rely on the deservedness of punishment to explain punishment's meanings and consequences. In the process it has lost sight of the need to explain, in a way that does not depend on punishment's meanings or consequences, why the only actions that deserve to be punished are both wrongful and blameworthy (two different and only very obliquely related properties).
1. The Elementary Moral Distinction
The ultimate objects of moral assessment are people and their lives. I will call this ‘the elementary moral distinction’. Many today seem to have lost sight of it. How often are we told that we should show respect for other people, only to discover that what we are actually being asked to show respect for is how those other people live?1 Although these two possible objects of respect are connected, their equation should be resisted. We do not always respect a person by respecting how she lives; sometimes quite the reverse. If someone is wasting her life but still deserves to be respected, the default way to show her the respect that she deserves is to do something that improves the way she is living—shake her out of it, block her path, change her incentives, shield her from further exploitation, etc. Sometimes, of course, there is no action open to us that will yield any improvement in how she lives, while on other occasions the only things we can do are disproportionate. In such cases we have to tolerate her continuing to live as badly as she does. But toleration is one thing and respect is quite another. Toleration is the moral virtue of those who appropriately curb their wish to eliminate what they do not respect. One cannot respect the way someone is living and tolerate it at the same time.2
In philosophy, the contemporary neglect of the elementary moral distinction owes much to Kant. I am not thinking here of Kante’s much-advertised (and much-misrepresented) doctrine of respect for persons. In so far as Kant said (p.52) anything of note about respect for persons, his views were consistent with those I have just sketched.3 Rather, I am thinking of Kant’s more distinctive doctrine that a morally perfect person cannot but lead a morally perfect life. This doctrine is now often remembered, thanks to a famous exchange between Bernard Williams and Thomas Nagel, under the heading of ‘moral luck. Kant is cited by both Williams and Nagel as the philosopher who most sweepingly rejected the possibility of moral luck.4 But on closer inspection Kant did nothing of the kind. He merely argued that morally perfect people cannot be morally unlucky in their lives.5 Thanks to the nature of morality, he said, they cannot live lives falling short of the morally perfect lives that they deserve to live. But Kant did not deny (and did not show any reason to doubt) that morally imperfect people can live lives that are morally worse, or indeed morally better, than those that they deserve to live. Nor, for that matter, did he deny (or show any reason to doubt) that whether someone is a morally perfect or a morally imperfect person could itself be a matter of luck.
Some parts of this chapter, in earlier draft, were delivered as a Brendan Francis Brown Lecture at the Catholic University of America in Washington DC on 21 September 2003. Thanks to Bill Wagner for leading the discussion on that occasion, and for various other contributions. Later drafts have profited from the detailed and insightful criticisms of Andrew Simester and John Tasioulas, to whom I am very grateful. I should also mention my gratitude to Doug Husak and Antony Duff, who together talked me out of a bad mistake.
So Kant certainly did not abolish or attempt to abolish the elementary moral distinction. But it is true that a decline of philosophical sensitivity to that distinction has been among Kant’s most enduring philosophical legacies. Kantian thinking, philosophical and popular, has simplified and radicalized Kant’s own views on the subject of moral luck. So much so that even a retreat to Kant’s own more modest views is sometimes perceived as a bold anti-Kantian move. Consider, for example, the group of contemporary moral philosophers who march, albeit not always in an orderly fashion, under the banner of ‘virtue-ethics’. Claiming to revive a pre-Kantian tradition of ethics traceable back to Aristotle, many of them favour ‘virtuously’ as an answer to the question ‘How should one live?’6 Ironically, this was precisely Kant’s answer to the same question, and it was one that Aristotle explicitly rejected.7 One should of course be a morally virtuous person. That much is analytically true and accepted by Aristotle and Kant alike. But no amount of moral virtue, on Aristotle’s view, ensures that one leads a morally perfect life. The morally perfect life, rather, is the life that a morally perfect person would want to live. Owing to bad luck, even a morally perfect person may live a morally imperfect life. This is the aspect of the human condition known as tragedy. Any large undeserved suffering or loss is nowadays casually (p.53) referred to as tragic. But tragedy, in the stricter classical sense illuminated by Aristotle himself, is not just any large undeserved suffering or loss. It is undeserved moral downfall. It is moral failure out of proportion to, or in extreme cases without, moral failing. It is the fate of all the great heroes from Oedipus to Othello.8 Oedipus is an extreme case (ruled out by Kant) of moral failure not owed to moral failing. Othello is a less extreme case (never ruled out by Kant, but implicitly ruled out by many Kantians) of moral failure out of proportion to moral failing. Othello is obnoxiously jealous, but Iago exploits this moral failing to drive Othello to murder, a moral downfall more spectacular than he deserves.
Has modern theatre, like modern moral philosophy, turned its back on tragedy? Some argue that it has.9 But our primary interest here will not be in how the elementary moral distinction has been interpreted and exposed in the arts. We will be concerned, in the main, with how it has been interpreted and exposed in the law. This may strike you as an improbable and unpromising project. For although legal systems are necessarily in the business of moral assessment, a decent legal system rarely makes moral assessments either of people or of their lives. Instead it mainly makes moral assessments of people’s actions taken one at a time. Isn’t this a further and different task? Aren’t there really three ultimate objects of moral assessment, namely people, their lives, and their actions taken one at a time? For reasons that I hope will become clearer as we go along, this suggestion should be resisted. The elementary moral distinction is the distinction between people and their lives. Actions have moral importance on both sides of this distinction. They matter morally because both people and their lives matter morally, and both people and their lives are partly constituted by their actions. But not always at the same time, by the same actions. The two can come apart. In particular, as the great tragedies teach us, some actions morally blemish a life without morally blemishing, or more than they morally blemish, the person who lives it. And this point, I will argue, is very clearly reflected in the law. The law does not need to make moral assessments of either people or their lives in order to mark the elementary moral distinction between people and their lives as objects of moral assessment. It marks the elementary moral distinction by assessing actions, taken one at a time, in two different dimensions or under two different aspects. It assesses actions, taken one at a time, as life-constituting on the one hand and as person-constituting on the other.
(p.54) 2. Lives and Wrongs
My life may go well or badly because of things that happen to me. I may be injured in a train accident or inherit a fortune from a long-lost aunt. With limited exceptions10 things that happen to me do not directly affect how well my life goes. They have an indirect effect. They affect how well my life goes by affecting what I do with my life. That I inherit a fortune does not make my life any better unless it means that I live better, doing more rewarding things with more interesting people, making better use of my talents, cultivating better tastes, and so forth.
One way to assess how well I live my life is morally. A morally better life is, ceteris paribus, a better life. The importance of living a morally better life is typically both exaggerated and underplayed. On the one hand moral success is often portrayed as the highest kind of success. A morally better life is held up as a better life come what may, not merely a better life ceteris paribus. On the other hand moral success is often portrayed as a self-contained kind of success, such that one can succeed in more ordinary ways (e.g. in one’s marriage or career) without any moral success. In attempting to correct these two complementary errors one may easily be drawn into a tiresome demarcation dispute. Which successes and failures count as moral ones? This question seems more pivotal, the more wedded one is to the two errors just mentioned. As one leaves them behind, the classification of a certain success or failure as moral loses its most dramatic implications. For present purposes we do not, in any case, need to have a complete picture of what the moral assessment of a life includes. We need only agree on this limited, and I hope ecumenical, proposal. Whenever a life is blemished by the wrongdoing of the person living it, that blemish is a moral one. Notice that this proposal does not entail that all wrongdoing is moral wrongdoing. It only entails that when wrongdoing is not moral wrongdoing (e.g. it is wrongdoing only according to an immoral law), it leaves no blemish on the life of the wrongdoer.
The proposal may be ecumenical, but it is also ambiguous. For there are two quite different things we might mean by saying that someone acted wrongly. We might mean that he did something unjustified. Or we might mean that he did something in breach of duty (a.k.a. obligation). This is another fundamental distinction that sometimes gets lost in modern thinking. Many modern philosophers, indeed, have tried to establish the co-extensiveness of the two classes of (p.55) wrong actions. Utilitarians of Benthamite persuasion have claimed that if an action is unjustified, it must also be in breach of duty. Indeed one’s only duty, on a rigidly Benthamite view, is the duty not to perform unjustified actions. Kant and his followers, meanwhile, had the converse idea: if an action is in breach of duty, it must also be unjustified. A duty is a reason that can never be defeated by countervailing reasons; not even countervailing duties. Duties, said Kant, are incapable of conflicting with each other.
Both of these views overgeneralize. There can be actions that are in breach of duty partly because they are unjustified, and actions that are unjustified partly because they are in breach of duty.11 But breach of duty and absence of justification do not, in general, go hand in hand. Many actions are unjustified even though they are not in breach of duty. (Today I failed to bring my umbrella out with me even though rain was forecast—silly me!) Many other actions are justified even though they are in breach of duty. (Yesterday I was late for work because I stopped to assist someone who had just been mugged.) To put it more paradoxically, there are many actions that are wrong without being wrongful, or wrongful without being wrong. The emphatic word ‘wrongful’, and similar terms like ‘wrongdoing’, ‘wronged’, and ‘a wrong’, are normally used to denote a breach of duty, whether or not it is unjustified. Whereas if we simply say that someone did the wrong thing, that normally carries the implication of an unjustified action, whether or not it is in breach of duty. At any rate, this is the linguistic convention that I will be adopting when a linguistic convention is needed.
The distinction between doing the wrong thing and doing something wrongful is of pervasive importance in most developed legal systems. The famous tort case of Vincent v. Lake Erie Transportation Co. illustrates its importance in the common law.12 When an unexpectedly violent storm made it too perilous to set sail, a captain kept his ship moored to somebody else’s pier without permission. By this action he saved his ship and crew, but damaged the pier. The pier owner sued for damages in trespass. The Minnesota Supreme Court ruled that the captain had not done the wrong thing (he had acted with ample justification), but he had acted wrongfully (he had breached a duty owed to the pier-owner not to moor his ship to the pier without permission). Because this ruling sounds paradoxical, many lawyers are reluctant to swallow it whole.13 A common reaction to Vincent is to explain it away as something other than a genuine tort case. Perhaps the (p.56) ‘damages’ awarded by the court for the ‘tort’ of ‘trespass’ were really more akin to a reimbursement for services rendered under the law of unjust enrichment? Perhaps the captain wronged nobody in using the pier without permission in an emergency, so long as he paid for any damage he caused in the process? Perhaps the court was merely enforcing this payment condition? It is easy to see what makes this reinterpretation appealing. But it is unnecessary. Vincent can be read unproblematically as a tort case. It isolates a general proposition of central importance to the common law: the mere fact that one was justified (= not wrong) in acting wrongfully does not mean that one did not act wrongfully, and does not by itself14 block one’s liability to pay reparative damages to those whom one wronged.
The same proposition is at stake in the long-running debates about ‘efficient breach’ in the law of contract. Some writers, especially in the Benthamite wing of the ‘economic analysis’ tradition, say that since damages for breach of contract are awarded in many cases in which the breach was amply justified, we should not persist in thinking of ‘breach of contract’ as the name of a legal wrong.15 Any ‘damages’ awarded by the court for ‘breach of contract’ should really be evaluated on some other model, e.g. as a reimbursement for services rendered under the law of unjust enrichment. One wrongs nobody by abandoning the contract without the other party’s permission so long as one pays for the other party’s expenses to date. All the court does in a standard ‘breach of contract’ case is enforce the payment by which one licenses one’s rightful abandonment of the contract. Those who resist this conclusion, especially those with broadly Kantian sympathies, have often answered that the breach only seems to be justified in such cases. Really when the indomitable moral force of the contract is properly acknowledged, the breach was not justified and hence remained wrongful.16 The two sides here are driven to their polarized conclusions by a shared error: the failure to see that, barring special cases, wrongfulness (= breach of duty) is one thing and wrongness (= absence of justification) quite another. Pace the Kantians, one need not deny that the breach was justified in order to insist that it was a breach of duty. Nor, (p.57) pace the Benthamites, need one deny that it was a breach of duty in order to insist that it was justified. The point recurs throughout the common law of tort and contract. The reason why Vincent brings it out more graphically than countless other tort and contract cases, and hence belongs on every student tort syllabus, is that in Vincent the captain was not only justified in mooring his ship to the pier, but would also have been unjustified in failing to do so. He was therefore in a strictly tragic position: the only justified action open to him was a wrongful action (one that violated the pier-owner’s rights). We may be reminded of one of Aristotle’s own most famous examples of tragedy, in which a sea-captain’s only justified option is to jettison his cargo.17
Why do these cases count as tragic cases? How do wrongful actions that cannot justifiably be avoided contribute to lives that are morally less perfect than the people who live them? In the next section I will explain why justification matters in the assessment of people. But first let me explain why wrongdoing matters in the assessment of lives. Notice that this is not the same as asking why wrongdoing matters full stop. Wrongdoing is the breach of a duty, and a duty (or to be more exact, the fact that one has a duty) is a reason with a doubly special categorical and mandatory force (more on which in a moment). So, for any rational being—any being to whom reasons apply—wrongdoing is something especially to be avoided. That is why it matters. But why wrongdoing matters full stop is not the question. The question is why wrongdoing matters in the assessment of lives. Why is it the case that, when wrongdoing is not avoided, it leaves an imperfection—what I have been calling a ‘blemish’—on the life of the wrongdoer? Why is my life damaged, in extreme cases destroyed, by my breaches of duty?
To get to the answer, one needs to begin by grasping a general truth about reasons. Reasons await full conformity. If one does not fully conform to a reason—if one does not do exactly what it is a reason to do—the reason does not evaporate. It does not evaporate even though one was justified in not conforming to it. It does not evaporate even though it is now too late fully to conform to it. Instead it now counts as a reason for doing the next-best thing. And failing that, the next-best thing again. And so on. Suppose, to borrow an example from Neil MacCormick, I promised to take my children to the beach today, but because of some emergency I have to cancel the trip.18 Without further ado, the reason I had to take the children to the beach today—namely the fact that I promised—now becomes a reason for me to take them to the beach tomorrow, or failing that the next day, or failing any of this to provide them with some alternative treat or privilege, etc. Of course the details may vary. Perhaps, if it was a birthday treat, a trip to the beach or an alternative treat tomorrow does not really count as second-best. Perhaps it only adds insult to injury. Perhaps the best I can do is apologize. Perhaps not even that. (p.58) Perhaps there is nothing I can now do by way of even minimal conformity. Even then the reason does not evaporate. It still makes its force felt as a reason for me to regret that I did not do as I promised.19 Regret is the rational response to any measure of non-conformity with any reason, and the reason for the regret is the very same reason that was incompletely conformed to (coupled, of course, with the fact of incomplete conformity to it). So, in MacCormick’s case, the difference between what I promised and what I did is cause for regret as soon as I postpone the outing to the beach even by one day. It is then cause for additional regret if I postpone the outing again, and for still more regret if I have to substitute a different treat, until the point of maximal regret at which my non-conformity with the original reason is total.
Here we have another important truth that has largely been expunged from modern moral philosophy. It has become the accepted wisdom, most conspicuous in the Benthamite tradition but equally taken for granted among Kantians, that at every moment we start again from tabula rasa, rationally speaking. There must always be a new reason for us to take an interest in an old reason. Regret, apology reparation, remorse, atonement, punishment: all this retrospectivity is irrational unless it now commends itself afresh, as a way of (say) reducing future suffering, or expressing renewed respect for oneself or others. But retrospectivity is in fact built into the bricks of rationality. True, there are invariably new reasons both for and against dwelling on the past. These should not be dismissed lightly. Yet equally one should avoid the opposite error. One should not dismiss the old reasons, the unconformed-to reasons that are still hanging around waiting for conformity. These reasons still have their force and to the extent that they remain unconformed-to their residual force confers prima facie rationality, without further ado, on regretful attitudes towards the path of one’s own life.20
This view may strike some as having scary implications. After all, whatever we do there is always something else we have some reason, however slight, to be doing instead. Doesn’t it follow that over time the old unconformed-to reasons will tend to pile up and overwhelm the new ones, and leave us rationally doomed to a life of (p.59) little else but regret? Of course it can be so. We have all read books or seen movies about lives consumed by rational regret. But normally things are not quite so bad. Among the many reasons that we do not conform to, there are many that are non-categorical. We have them by virtue of our personal goals and we no longer have them when our goals change. Furthermore, many are non-mandatory. They simply weigh in the balance of reasons and do not exclude any countervailing reasons from consideration. When we do not fully conform to a non-categorical and/or non-mandatory reason, the reason that remains to haunt us still is non-categorical and/or non-mandatory Such a reason is therefore permanently vulnerable to the abandonment of old goals and/or to defeat by the new reasons that militate powerfully in favour of getting on with our lives. When things are not so easy is when we are left with old unconformed-to reasons that are both categorical and mandatory, i.e. when we had duties that we failed to perform, and hence acted wrongfully. In such cases the reason left over and still awaiting conformity does not surrender to a change in our personal goals. And it continues to exert mandatory force such that at least some conflicting reasons (some of the otherwise powerful reasons that we have to get on with our lives) are excluded from consideration and cannot suffice to defeat it. Wrongful action, in short, leaves us with regrets that are hard to expunge and the repression of which is hard to justify. That is how my acting wrongfully may damage, and in extreme cases destroy, my life. Of course I am not thinking here of the mere experience of regret, the psychological haunting of the wrongdoer. No doubt one may avoid this by various displacement activities, self-deceptions, etc. Rather I am thinking of the continuing force of the reason which makes the regret rationally appropriate, and which it is now too late completely to eradicate. It is the enduring presence of this reason which, to a greater or lesser extent, constitutes the damage to one’s life.
Was the captain in Vincent doomed to such a damaged life? In some measure, yes. After he had kept his ship moored to the pier without the owners permission, albeit with full justification, it was too late for him fully to perform his duty not to do so. The pier-owner’s rights were violated; there was no going back. But a partial conformity, a second-best option, was still available. By paying reparative damages for the violation and its consequences the captain could still imperfectly perform his duty and leave himself with a less damaged life. You may have been wondering how all my discussion of regrets and blemishes relates to the law, and now you have your answer. In the law we have two duties, a primary duty that we violate when we commit a tort or breach a contract, and a secondary duty to pay reparative damages that is brought into being by the law when and because we breach the primary duty. But part of the case for bringing the secondary duty into (legal) existence is that (morally) it is already there. It is the same primary duty that one violated when one breached the contract or committed the tort. It continues to bind one and to press for second-best conformity. When one pays reparative damages, one imperfectly performs the same duty that, earlier, one failed perfectly to perform. That is what is captured in the common lawyer’s saying that the (p.60) purpose of reparative damages, in tort as in contract, is to put the plaintiff, so far as money can do it, back in the position that she would have been in had the wrongful action not been committed against her.
This being so, there is no need to seek an independent rationale for the law’s secondary duty. We do not need to build it up rationally from tabula rasa. There is no need to argue in Benthamite vein that reparative damages are optimally deterrent, or in Kantian vein that they are a way of re-establishing respect for persons. Of course considerations of deterrence and respect for persons may be relevant in assessing whether and when we should go to all the trouble of making the duty to repair enforceable through the law. The point is that they are not needed to explain why we should want to do so. We should want to do so because when a primary duty is breached, a next-best performance of the same duty is automatically called for without further ado. Often—often enough to dictate the common law’s standard remedies for wrongdoing—the payment of reparative damages counts as such a next-best performance.
In these remarks I have explained why the fact that I acted wrongfully matters in the moral assessment of my life, which also affords an explanation of why the fact that I acted wrongfully often matters in the determination of my legal position. Some have thought that this kind of explanation is back-to-front. Remember the moral philosophers who offer ‘virtuously’ as an answer to the question ‘How should one live?’ Why do these writers fondly imagine that they are at odds with Kant? Partly because they think (mistakenly) that Kant gave a different answer to their question. But also partly because they think that Kant asked a different question. Instead of asking ‘How should one live?’, they think, Kant asked ‘Which actions should one perform?’ And they want to restore the former question to the primacy which they think it had in Aristotle’s work.21 For some such writers, restoring the former question to primacy means this: that one needs to determine which lives are morally imperfect in order to determine which actions are wrongful.22 The picture I have just painted reverses this order of determination. According to my picture, one needs to determine which actions are wrongful in order to determine which lives are morally imperfect. My picture was the one endorsed by Aristotle and Kant alike. Neither believed in the primacy of the question ‘How should one live?’ over the question ‘Which actions should one perform?’ if by ‘primacy’ one means that the first question is to be answered first. On the other hand Aristotle did stress, in a way that Kant did not, the primacy of the question ‘How should one live?’ over the question ‘Which actions should one perform?’ if we mean something else by ‘primacy’. For Aristotle stressed, as I have stressed, that the moral importance of wrongful actions, independently identified, lies primarily in how they affect people’s lives for the worse. In the examples that matter for the law, there are always at least two (p.61) affected lives. There is the life of the wrongdoer and the life of the person wronged. Lawyers are used to thinking of the person wronged as the main person whose life was made worse by the wrong. He can’t use his pier, for example, or he can’t use his legs. Here I have emphasized instead the moral damage to the life of the wrongdoer. Lawyers may find this emphasis counterintuitive. But I have tried to suggest that, even if counterintuitive, the same emphasis is reflected in a great deal of familiar legal doctrine.
3. People and Faults
In casual conversation we might say that the damage to the pier in Vincent v. Lake Erie Transportation Co. was the captain’s fault, or that he was to blame for it. We sometimes use these expressions merely to pick out a person who failed in her duty to prevent, or not to bring about, some eventuality (also sometimes known as ‘the person responsible’). The captain fits this description. But when we are watching our words more carefully, we should avoid saying that the damage to the pier in Vincent was the captain’s fault or that he was to blame for it. For the captain was not at fault and should not be blamed. His action, being wrongful, blemished his life. Yet, being justified, it did not reflect badly on him. In particular it did not show him up as cowardly, imprudent, lazy, mean-spirited, irresponsible, or otherwise morally at fault. People with different moral faults differ in respect of which reasons they overplay and which reasons they underplay, and hence in respect of which actions they are over- or underdisposed to perform. Cowardly people overplay the importance of their own safety, mean-spirited people underplay the importance of other people’s feelings, imprudent people underplay the importance of longer-run consequences, and so on. But what they all have in common, and what constitutes their moral fault, is that they all end up acting for defeated rather than undefeated reasons. A justified action, meanwhile, is one performed for an undefeated reason. It follows that so long as I do not perform any unjustified actions, I remain a person free of all moral faults.
Here I am already challenging a familiar classificatory scheme used by some writers on criminal law as well as some moral philosophers. It is sometimes said that holding an action to be justified or unjustified is an instance of act-assessment, whereas holding a person to be of good or bad moral character is an instance of agent-assessment.23 This is misleading. It encourages a distorted view of justification as well as a distorted view of moral character. If we must talk of acts and agents, justification depends on the agent as much as the act, and moral character depends on the act as much as the agent. But for reasons that will emerge I would rather not put the point this way at all.
(p.62) The most distracting and irrelevant thought that the contrast between act-assessment and agent-assessment brings to mind is the thought that moral character is something that endures in people while their actions are fleeting. This thought encourages us, in Humean vein, to think of someone’s moral character as standing in a contingent relation with her actions: character causes action, action evidences character.24 But this is a blunder. People’s moral characters are constituted, and not merely evidenced, by what they do. Someone who has never done anything dishonest in her life is not a dishonest person, even if she is often tempted to act dishonestly. Rather, she would have been a dishonest person if only she had not had so much self-control.25 Conversely someone who acted dishonestly just once, even though this was the only occasion on which he ever felt tempted, is to that extent a dishonest person. Of course, we might say of such a person that he acted ‘out of character’. In this phrase we juxtapose the fact that he is normally honest with the fact that today he was dishonest. But it does not alter the fact that today he—not just his action, but he—was dishonest. This dishonesty constitutes a blemish on his moral character.
So ‘out of character’ cannot be interpreted to mean ‘not constitutive of character’. The thought that it can comes of the running-together of two questions. One is the question: ‘What is character?’ The other is the question: ‘How is character formed?’ When we think about the formation of character we cannot but think about the cultivation of lasting dispositions or tendencies. There is no way, even in principle, to cultivate occasional honesty or occasional dishonesty. But it does not follow that there is no such thing as occasional honesty or occasional dishonesty. A discussion of character traits that centres on the question of formation inevitably plays up the idea of character as an enduring condition.26 This explains why the fact that a certain action was ‘out of character’ can be a consideration relevant to the aptness of some putatively rehabilitative or reformative reactions to it. But when we are simply interested in assessing people—as opposed to deciding how we might improve them—we should not similarly sideline their occasional abberations.27 I don’t mean, of course, that we should treat an isolated occasion of dishonesty as somehow obliterating a fine record of honesty. Rather we should think of it as blemishing that record. The record now reads: normally very honest, but on one infamous occasion extremely dishonest. This record is a record of (p.63) character itself, not a record of conflicting evidence about character, nor (obviously) a record of reformative plans for character.
This is the main way in which it is misleading to think of the assessment of moral character as an agent-assessment as opposed to an act-assessment. But how is it misleading to think of holding a certain action justified as an act-assessment rather than an agent-assessment? It is misleading in that whether one acts with justification depends not only on what one does, but also on why one does it.28 If there is an undefeated reason to φ, then φing is justifiable. But φing is justified only if the agent φed for that undefeated reason. If one claims that the tyrannical behaviour of Saddam Hussein towards his own people justified one’s pre-emptive strike on Iraq, one claims not only that the tyrannical behaviour of Saddam Hussein towards his own people was an undefeated reason for a pre-emptive strike on Iraq, but also that this was the reason why one launched the strike. And if one asks whether the risk of attack from a burglar justified the use of lethal force against the burglar by the householder, one asks not only whether the risk of attack was an undefeated reason for the householder to use lethal force against the burglar, but also whether that was the householder’s reason for using such force. In justifying an action, in short, it is not enough that there were undefeated reasons for that action unless that actor also acted for at least one of those reasons. Why is this? Why is it not enough, from the point of view of justification, that someone conforms to an undefeated reason without also acting for it?
To get the answer straight one needs to begin by thinking carefully about what justification is for, why justification matters, what is the point of justification. You may think this a strange question. What justification is for, you may think, is surely to make the world a better place, to fill it with better actions rather than worse ones. But in the only sense in which this proposition is true, it is question-begging. Why are justified actions better actions? Some moral philosophers, mainly in the Benthamite tradition, tried to show that they are better actions quite apart from being justified actions, and that their betterness explains why they are justified actions. They are rationally better because they are better tout court. But this project foundered as Bentham’s heirs, from J. S. Mill onwards, gradually rediscovered the deontic aspects of practical thought. Some actions are better actions, or not-worse actions, only in virtue of being justified actions. They are better tout court only because they are rationally better, not vice versa. When we ask ‘What is justification for?’ we want to know how this can be so. Why, when our independent tests of betterness run out, does making the world a rationally better place still continue to make the world a better place? The answer, in brief, is that people are (inter alia) rational agents, and rational agents necessarily aspire to excellence in rationality. Excellence in rationality means excellence in seeing reasons, in using them, and in negotiating conflicts among them. To act for a defeated reason, any defeated (p.64) reason, is to come unstuck as a negotiator of rational conflicts. Subject to an important proviso to be entered in a moment, this reflects badly upon one (constitutes one’s fault) as a rational agent. And that is why justification matters. It matters because, as rational agents, people are the worse for acting without it. And a world with worse people in it is, ceteris paribus, a worse world.
I have simplified this explanation in various ways. Most significantly, I have bracketed a special class of unjustified actions that do not reflect badly on their agents. Sometimes one lacks an adequate justification for what one does, yet one has an adequate justification for the beliefs or emotions on the strength of which one does it. In such a case one’s action is excused. The simplest excuse of all is the justified mistake excuse. Suppose that the captain stays moored to the pier, in a case otherwise akin to Vincent, only because a malicious hoaxer broadcast a false storm warning as if it came from the coastguard. In such a case the captain lacked an undefeated reason to stay moored. Yet he had an undefeated reason to think that he had an undefeated reason to stay moored. (His excuse is ‘How was I to know it was a hoax?’29) We would not expect this imaginary storm to relieve the captain of tort liability for trespass any more than the real storm did. On the other hand, we would expect both to be equally available as defences to any criminal charge against the captain that might arise from the incident. That is because both the justification and the excuse equally extinguish the captain’s fault. In this respect an excuse is every bit as good as a justification. True, any rational agent, given the choice, would rather be justified than excused. It is better to act for an undefeated reason than to be drawn, even for an undefeated reason, into acting for a defeated one. One kicks oneself when one realizes that one acted for a defeated reason. One looks back on it with extra regret. But that is because, like a wrongful action, an action on the strength of a mistaken belief leaves a blemish on one’s life. It is not because it reflects badly upon one as a rational agent. It reflects badly on one as a rational agent only if (a) the mistaken belief is unjustified or (b) the action would not have been justified even if the mistaken belief had been true. Something very similar, albeit a little more complex, is true of actions performed in anger, fear, frustration, desperation, etc. Roughly, they reflect badly upon one as a rational agent only if (a) one’s anger, fear, frustration, desperation, etc. was itself unjustified or (b) one’s anger, fear, frustration, desperation etc. was insufficient to explain one’s action, assuming an acceptable level of self-control.30
The thought that rational agents would rather be justified than excused is not the only possible source of doubts about the fault-negating power of excuses. Another is the thought that at least some excuses are available as ‘concessions to human frailty’, in the words favoured by some criminal lawyers. The same thought may be conveyed less grandiloquently by saying that people only need (p.65) excuses because of their limitations. They do not have unlimited reserves of patience, courage, charity, attentiveness, insight, etc. and that is why they did something unjustified and need to fall back on an excuse. Isn’t the display of a limitation also the display of a fault? So doesn’t it follow that excuses, or some of them, must perform some function in the assessment of people other than the negation of their fault?31 The mistake here begins with the misleading suggestion that one’s limitations explain why one did something unjustified. This is mistaken if it is taken to mean that one displays no limitations in one’s justified actions. Even one’s justified actions are not always exemplary. They do not always exhibit the highest measure of virtue. Here is one proof of the point. The highest pinnacle of virtue is often exhibited in supererogatory actions. That I do not perform such actions whenever they are there to perform often (not always) displays one or more of my limitations. Yet it does not follow that I am unjustified in failing to perform supererogatory actions. How could it follow? It is part of the very idea of a supererogatory action that one is permitted not to perform it. One is permitted not to perform it not merely in the weak sense of having no duty to perform it, but in the stronger sense that the reasons to perform it (although they may be undefeated) lack the ability to defeat their opponents. So any reason not to perform a supererogatory action remains undefeated. So one is always justified in not performing it.32 Hence one cannot be at fault in failing to perform it. It follows that the mere fact that one does not exhibit perfect virtue in one’s action does not entail that one exhibits a fault.
A fault, in the sense that matters here, is not just any shortfall of virtue. A fault is a shortfall of virtue that consists in the performance of actions that are both unjustified and unexcused. Any other shortfall of virtue is a mere limitation. It does not reflect badly upon one as a rational agent. Rather, it reminds us what life is inevitably like for a rational agent. Even undefeated reasons pervasively conflict among themselves. If one is to have a character at all one cannot but lean towards reasons of some types, at the expense of reasons of other types. It follows that it is impossible to have a character which is such that one excels even-handedly in one’s relations with all undefeated reasons. In other words, nobody, not even a saint, can be without her limitations. Every excellence has some limitation as its flip-side. The more perfectly charitable, the less perfectly just; the more perfectly frank, the less perfectly diplomatic; and so on. This is an inevitable feature of the human condition. But all of this is consistent with always acting for an undefeated reason, or at any rate for a reason that one has undefeated reason to treat as undefeated. It (p.66) is consistent with always being justified in what one does, or at least excused. So it is one thing to think of an excuse as a ‘concession to human frailty’ in the sense of a recognition that human beings inevitably have limitations (imperfect virtues) as the price of their excellences (more perfect virtues). It is another thing to think of an excuse as a ‘concession to human frailty’ in the sense of a recognition that human beings inevitably have their share of bad character, i.e. their faults or vices. This is just not true. Perhaps all human beings do have their faults or vices, but there is nothing in the human condition (or at any rate in our predicament as rational beings) that makes this inevitable, and there is no case for ‘concessionary’ extension of justifications or excuses to fault-constituting actions. How could there be? It is the very fact that actions are unjustified and unexcused that makes them fault-constituting.
Here I once again take sides on an issue much discussed by contemporary ‘virtue ethicists’. That it is rational to exhibit moral virtue is analytically true and accepted by Aristotle and Kant alike. But are morally virtuous actions rational because morally virtuous, or morally virtuous because rational? Kant famously defended the idea that they are morally virtuous because rational. Some contemporary writers in the ‘virtue ethics’ school defend the opposite view and claim thereby to be Aristotelian revivalists.33 But on this point as on so many others Aristotle is widely misunderstood. Aristotle and Kant agreed, rightly, that virtuous actions are virtuous because rational.34 In the terms I just introduced, it is the fact that actions are justified or excused that entails that they are not fault-constituting, not the other way around. Why is Aristotle mistakenly associated with the opposite view? Perhaps because he repeatedly and rightly emphasized that one’s moral virtues—one’s qualities as a person—make an intrinsic as well as an instrumental contribution to the quality of one’s life.35 If the morally virtuous person lives by that token a rationally better life (a better life for a rational being), surely the moral virtue must be what makes it rational? So how can the fact that it is rational also be what makes it morally virtuous? Actually the answer is straightforward. That one is a virtuous person brings extra value to one’s life above and beyond the ordinary value, the rational pursuit of which constitutes one’s virtue. First there is the rationally salient value that one pursues, and successful pursuit of which makes one morally virtuous. And then there is the extra value that lies in one’s successfully pursuing the rationally salient value.36 This extra value is not rationally salient. It is the value of rationality itself. Pace Kant, pursuing rationality (or indeed morality) itself, for its own sake, is normally self-defeating. One does (p.67) not exhibit moral virtue in acting with the intention of exhibiting moral virtue. One exhibits moral virtue in acting with other aims, themselves rationally defensible (i.e. sufficient to justify or excuse one’s action). Thereby one makes one’s life better in two ways at once. A life is better for the fact that it was a life of greater conformity with reasons. A person without faults is better equipped, instrument-ally, to live this life (even though tragedy may still strike). But being a person without faults also contributes intrinsically to one’s quality of life. A life is better for the very fact that it is lived by someone with fewer or lesser faults. Virtue, as they say, is its own reward: one’s success in seeing reasons, in using them, and in negotiating conflicts among them is an instrument of better living, but also a constituent of it. When tragedy strikes, one may still console oneself with the second aspect. One may say: a life blemished, but at least not blemished for having been lived by a blemished person.
4. Fault-Anticipating Wrongs
Let me mention one particular way in which a life lived by a better person can be, by that token, a better life. The avoidance of fault sometimes entails the avoidance of wrongdoing. Being at fault, to put it the other way round, sometimes contributes constitutively, and not just instrumentally, to the wrongfulness of one’s actions. Some wrongs are ‘fault-anticipating’. They are wrongs that are committed only in the absence of justification or excuse—in other words, only in the presence of fault.
A common view is that all wrongs are fault-anticipating. If there is no fault, there is no wrong. But on more careful reflection the very opposite conclusion seems more tempting: there can be no such thing as a fault-anticipating wrong.37 If one committed no wrong, surely one has nothing to justify or excuse? If one has nothing to justify or excuse, then surely one neither has nor lacks a justification or excuse? Doesn’t it follow that one must commit a wrong to lack a justification or excuse? That being so, how can it ever be that one must lack a justification or excuse to commit a wrong? How are fault-anticipating wrongs possible? The answer is startlingly simple. Fault-anticipating wrongs are always parasitic or (p.68) secondary wrongs. One commits them only if one lacks justification or excuse for something else one does in committing them. In some cases the ‘something else’ is not a wrong at all—not a breach of duty—but merely a failure to conform to an ordinary reason for action. One has a weighty reason not to take one’s children mountaineering, for instance, but one’s only duty is not to do so negligently. What is wrongful, because a breach of duty, is faulty (unjustified and unexcused) non-conformity with the weighty reason. In other cases, however, the ‘something else’ is the commission of another wrong. One violates a duty whenever one spreads gossip about one’s colleagues, for example, but one violates an additional and more stringent duty when one spreads the same gossip maliciously or dishonestly. The fault-anticipating wrong, the more heinous wrong, lies in the faulty commission of another less heinous wrong that is not fault-anticipating.
Elsewhere I have argued, against Kant, that the primary or basic type of moral wrong is one to which the endeavours of the wrongdoer make no constitutive difference.38 The basic moral question is what you did or didn’t do (Did you kill? Did you cause offence? Did you keep your promise?) never mind what you were trying to do or trying not to do. I argued that wrongs that are constitutively sensitive to what the wrongdoer was trying to do or trying not to do (Did you murder? Did you cheat? Did you conspire?) are wrongs of a secondary or parasitic type, relative to those that are constitutively insensitive to what the wrongdoer was trying to do or trying not to do. To defend this strong conclusion I made a long argument. What I offered just now was a much shorter argument for a much more modest conclusion. The more modest conclusion is that wrongs that are constitutively sensitive to the wrongdoer’s fault are wrongs of a secondary or parasitic type, relative to those that are constitutively insensitive to the wrongdoer’s fault. This conclusion is more modest because, while every fault-anticipating wrong is constitutively sensitive to what the wrongdoer was trying to do (in that it is constitutively sensitive to her reasons for doing as she did), many wrongs that are constitutively sensitive to what the wrongdoer was trying to do are not fault-anticipating. Many wrongs can only be committed intentionally, for example, and yet can still be committed without fault. We know that they can be committed without fault because we can conceive of circumstances in which their commission can be justified or excused. Intentional wounding can sometimes be justified as an action in self-defence, or excused as an action under duress. Not so reckless wounding. To conclude that I wounded recklessly (or likewise negligently, stupidly, unjustly, callously, in a cowardly way, etc.) is already to conclude that I had no justification or excuse. If there is a distinct wrong of reckless wounding or cowardly wounding or callous wounding, it is a fault-anticipating wrong. A distinct wrong of intentional wounding, on the other hand, is not.
Writings about the law, especially the criminal law, often get into a muddle on this front. Consider strict liability. Some criminal lawyers think of (and object to) (p.69) strict liability as liability irrespective of fault. Meanwhile some criminal lawyers think of (and object to) strict liability as liability for wrongs that have no mens rea constituent (i.e. wrongs that can be committed without the wrongdoer’s intending or being aware of any of the wrong-making features of her action). Some seem to think that these two ideas (and hence the two objections) are interchangeable. But if strict liability is liability irrespective of fault, then there may in principle be strict liability for intentional crimes and other crimes of mens rea. A wrong partly constituted by the intention or awareness of the wrongdoer may nevertheless be committed faultlessly, with adequate justification or excuse. If the law does not recognize any such justification or excuse, then it holds the intentional or knowing offender strictly liable, in the sense of liable irrespective of her fault. Conversely, if strict liability is liability for wrongs that have no mens rea constituent, then there may in principle be strict liability that is not liability irrespective of fault. For the law may recognize justifications and excuses as defences to a wrong, and hence allow some of the wrongdoer’s reasoning at the time of the wrong to be relevant to her liability, even though the wrong itself has no mens rea constituent (and nor for that matter a fault constituent).
This muddles matters because the success of any objection to strict liability depends on the defensibility of the principle that strict liability is said to contravene. And there are clearly two different principles that are being advanced here. One, which we could call ‘the mens rea principle’, is a principle requiring criminal wrongs to have certain constituents, namely the wrongdoer’s intending or being aware of at least some of the (other) wrong-making features of her action. The other, which we could call ‘the fault principle’, is a principle regulating the conditions for the imposition of criminal liability, rather than the constituents of criminal wrongdoing. Criminal liability should be imposed only for wrongs that are faultily committed, whether or not the fault in question is a constituent of the wrong (i.e. whether or not the wrong in question is fault-anticipating). Sometimes the law may be able to satisfy both principles at once by treating a certain wrong as fault-anticipating—as partly constituted by, say, the dishonesty or the malice of the wrongdoer. But just as often the mens rea principle and the fault principle place different demands upon the criminal law and the criminal law needs to satisfy them separately.
Our main interest here will be in the rationale for the fault principle. But for contrastive purchase it is worth saying a few words, first, about the rationale for the mens rea principle. There are many wrongs that, even apart from the law, are partly constituted by some intention or awareness on the part of the wrongdoer. When these wrongs are recognized by law the relevant intention or awareness naturally carries over into the legal constituents of the wrong. Cheating, lying, manipulating, coercing, torturing: these are wrongs of mens rea even before the law gets its hands on them. But as a general principle extending beyond such wrongs, the mens rea principle has a largely institutional rationale. According to the ideal known as the rule of law, those of us about to commit a criminal wrong (p.70) should be put on stark notice that that is what we are about to do.39 The criminal law should not ambush us unexpectedly. Of course, to avoid unexpected ambushes we all need to know what the law requires of us. For that reason, criminal laws should be clear, open, consistent, stable, and prospective. They should also forbid specific actions (not courses of action, activities, ways of life, etc.). Even all this, however, is not enough to ensure that those of us about to violate the criminal law are put on stark notice that we are about to violate it. For we may know the law and yet have no grasp that what we are about to do might constitute a violation of it. That is because often we have no idea which actions we are about to perform. I make a light-hearted remark and (surprise!) I offend one of my guests. I turn on my oven and (surprise!) I blow all the fuses. The mens rea principle is the principle according to which such actions—the self-surprising ones—should not be criminal wrongs.
More precisely: according to the mens rea principle, criminal wrongs should be such that one does not commit them unless one intends or is aware of at least one wrong-making feature of what one is about to do, such that (assuming one knows the law) one is also alerted to the fact that what one is about to do will be of interest to the criminal law. The principle does not extend similarly to wrongs under private law, such as torts and breaches of contract. The main reason is that the mens rea principle is a systematically pro-defendant principle. It is one thing for the law to be systematically pro-defendant in the criminal process, and quite another for it to be systematically pro-defendant in a civil dispute. Criminal investigators and prosecutors have much more extensive, and potentially oppressive, power than do private-law plaintiffs. The mens rea principle is one of several principles by which, inasmuch as we live under the rule of law, we are protected against the oppressive use of such power.40
5. Defending the Fault Principle
Like the mens rea principle, the fault principle is one that extends to criminal law but not, or not generally, to private law. Fault can certainly be relevant to liability in tort and contract, but that is because some torts and breaches of contract are fault-anticipating wrongs. Private law cares about wrongdoing. It cares about the wrongdoers fault if and only if his fault is a constituent of his wrong. Criminal law, by contrast, cares about fault even when fault is not a constituent of the (p.71) wrongdoer’s wrong. It gives independent importance to the wrongdoer’s fault. It does this by offering her various possible justificatory and excusatory defences to non-fault-anticipating criminal wrongs. At any rate, that is what the fault principle would have it do.
H. L. A. Hart famously assimilated the fault principle to the mens rea principle in respect of rationale. He argued that the criminal law should admit justifications and excuses for the same reason that it should require mens rea: to ensure that wrongdoers have a fair opportunity deliberately to steer clear of criminal liability in keeping with the ideal of the rule of law.41 No doubt the criminal law’s adherence to the fault principle, like its adherence to the mens rea principle, can help with its adherence to the ideal of the rule of law. But this does not get to the bottom of the fault principle’s rationale. The fault principle is not primarily an institutional principle. It does not apply to the criminal law because the criminal law is part of a legal system and legal systems need to regulate the potentially oppressive power of their own officials. Rather it applies to the criminal law because the criminal law exacts punishments. Punishments are subject to the fault principle irrespective of whether they are exacted by a legal system and irrespective of whether there is any potential for abuse of official power. If I punish one of my friends for wrongdoing (e.g. by not sending him an invitation to my party) it is no cause for complaint on his part that he had no idea and indeed no way of knowing that I might take an interest in his actions. As a friend I am not bound by the rule of law, nor (hence) by the mens rea principle. It is not my job to put people on notice that they are about to get into trouble with me. But I am bound by the fault principle. My friend does have serious cause for complaint if I punish him for a wrong that he committed faultlessly. This is the same complaint that the captain in Vincent v. Lake Erie would have had if, in addition to expecting him to pay reparative damages for his violation of the pier-owner’s rights, the law went on to punish him for what was, even in the law’s eyes, a faultless wrong. Reparation for faultless wrongs is one thing; punishment for them, whether by the criminal law or otherwise, is quite another. The fault principle is a principle governing punishment, and it applies to the criminal law because criminal liability is a liability to be punished.
One should proceed carefully here, because while criminal liability is a liability to be punished, it is normally a liability to be punished only at the court’s discretion.42 The criminal court is normally at liberty (at the ‘sentencing stage’ of the trial) to attach some non-punitive measure to the wrong, such as probation, hospitalization, or conditional discharge, instead of a punitive one, such as imprisonment, community service, or a fine. Can’t the criminal law deal with fault at this sentencing stage, as (p.72) a matter to be considered in the exercise of judicial discretion? Not with equanimity. When things go well for the law, the determination of criminal liability is a determination that a punishment for the wrong is deserved. And a punishment for the wrong is deserved only if the wrong was faultily committed. Judicial discretion at the sentencing stage exists primarily to enable consideration of factors relevant to the justification of punishment other than those that bear on whether the punishment is deserved: whether to make an example of the defendant, whether to give her credit for other respects in which she is of good character, whether to pursue her rehabilitation or reform, and so on. So the question of fault should not, ideally, be left to figure in the judge’s exercise of discretion. But things do not always go so well for the law. Even with the best will in the world, the law’s determination of criminal liability cannot always be a satisfactory determination of the deservedness of criminal punishment. The rule of law, by insisting on clarity, stability, prospectivity, etc., often prevents the law from showing full sensitivity to the differential moral merits of every wrong that is committed and every credible justification or excuse for its commission. Some questions bearing on fault, normally but not always relating to the fine-tuning of fault, are inevitably left over to be dealt with at the sentencing stage. So the fault principle is a principle to be honoured at the sentencing stage to the extent that, alas, it was not fully honoured in the determination of criminal liability. That is a reason for the law to retain a separate sentencing stage complete with judicial discretion to punish or not to punish. But it is not a reason for the law to disregard the defendant’s fault in determining his criminal liability and hence in determining his exposure to that same judicial discretion.43
These remarks already lead us a little closer to the rationale for the fault principle. A punishment, I proposed, is deserved only if the wrong being punished was faultily committed. The proposition is easy to agree to and yet remarkably difficult to vindicate in a satisfying way. Some ascribe the difficulty to the very idea of desert, which they claim to find mysterious. But it is hard to do without the idea of desert. Replacing ‘deserved’ with ‘justified’, for example, falsifies the proposition we are interested in rather than simplifying its vindication. As I just mentioned, various considerations apart from the deservedness of punishment are relevant to the justification of punishment.
In any case, there is nothing remotely mysterious about the idea of desert. I used it unremarkably near the start of this chapter, in elucidating the classical idea of the tragic. Tragedy, I said, is undeserved moral downfall, moral failure out of all proportion to moral failing. The idea of desert that I invoked in saying this is (p.73) straightforward. It is a variant on the idea of aptness or fittingness. What I deserve is simply what is apt or fitting for me because of some excellence or deficiency of mine. As the fastest runner in the race, I deserve to be declared its winner. As the least beautiful contestant, I deserve to come last in the beauty pageant. These latter examples are, of course, examples of non-moral deserts, in the sense that the excellences and deficiencies at stake are not moral excellences and deficiencies. Because of the widespread tendency to think of moral success as the highest kind of success, many people nowadays find such examples disquieting. They tend to think that the non-morally deserving are not really deserving at all.44 The ones who really deserve to win the race are the most dedicated or the most courageous runners, the ones who show the most moral fibre, never mind their short legs and poor co-ordination. Even the element of skill in athletics makes one a deserving winner, in the view of some, only because and to the extent that the cultivation of skill makes demands on moral character. As for beauty pageants, the thinking continues, they have no plausible connection with desert at all, and should be abolished. But all of this is absurdly moralistic. In tests of speed and beauty, the speediest and the most beautiful respectively are those who deserve to succeed. Whether there should be any tests of speed and beauty is another question altogether. To put it another way: one may readily agree that beauty pageants should be abolished, without agreeing that truly beautiful people do not (in respect of their beauty) deserve our admiration and recognition.45
So the problem is not to understand what it is to deserve something. The problem is not even to understand what it is to deserve something morally. (There is no harm in thinking of a human life as, in part, a moral test, and in thinking of the morally virtuous person as the person who deserves to pass the test.) The problem, rather, is to understand how punishment can ever be the thing that is deserved, and how faulty wrongdoing comes to be the basis for deserving it. An instructive but fallacious attempt at a solution begins with the observation that there is a conceptual connection between punishment and wrongdoing.46 If suffering or deprivation is deliberately inflicted but not as a response to (supposed) wrongdoing, that is not a case of punishment. And if the wrongdoing is only supposed, and not real, it is a case of mistaken punishment. This is all true, but inconclusive. It does not establish that punishment is ever deserved as a response to wrongdoing, because it does not establish that punishment is ever deserved. It only establishes that punishment can (p.74) sometimes be mistaken. More importantly, it does not help to explain why, for punishment to be deserved as a response to wrongdoing, the wrongdoer must also be at fault. For there is no relevant conceptual connection between punishment and fault. Punishing a faultless wrongdoer is punishing undeservedly. But it is still punishing. If it were not punishing it would not be regulated by the principles that regulate punishing, such as the fault principle. So we could not condemn it for violation of those principles. Those deliberately inflicting suffering or deprivation on the faultless could always say: ‘Since I’m not punishing, why should I care which punishments would be deserved?’ The conceptual connection between punishment and wrongdoing is certainly important, but for the opposite reason. It is important because those who respond to (supposed) wrongdoing by deliberately inflicting suffering or deprivation are thereby automatically made answerable to the principles that govern the infliction of punishment, such as the fault principle. They cannot avoid such principles by claiming not to be punishers. But nor can the same principles be vindicated—shown to be sound—merely by pointing out that an infliction of suffering or deprivation without a supposition of wrongdoing is not punishment.
Those looking for a better defence of the fault principle may see some hope in what I called the elementary moral distinction, and in particular in the proposal, implicit in my characterization of tragedy, that better people deserve better lives. We know that there are tragic cases of lives that are badly damaged through no (or disproportionately little) fault of those who live them. Equally, there are converse cases of lucky people whose serious faults inflict disproportionately little damage on their lives. Isn’t punishment deserved simply as a way of bringing such people’s lives closer to the lives they deserve? There are two related objections to this hypothesis. First, if someone deserves to be punished, he has committed a wrong. Necessarily his life has already been damaged by his having done so. Indeed he may now have, morally speaking, exactly the life that he deserves, or even worse. So why should we want his life to be further damaged by the deliberately inflicted47 suffering or deprivation of a punitive measure? As if it were not bad enough that we deliberately bring additional suffering and deprivation into the world, we seem to be doing it by kicking someone who is already down. Herein (p.75) lies the first mystery of deserved punishment. Why would we want to inflict yet more damage on what is, ex hypothesi, an already damaged life? Secondly, just as there are many wrongful actions that are not faulty, so there are many faulty actions that are not wrongful. If the point of deserved punishment were to give vicious people the lives they deserve, then one would expect punishment to be deserved on the strength of fault alone. One would expect the commission of a wrong to be irrelevant. But it is not. If luckily no wrong was committed in spite of one’s fault, then luckily no punishment is deserved. The fault principle captures a necessary but not a sufficient condition for being deserving of punishment. Herein lies the second mystery of deserved punishment. Why isn’t fault alone sufficient, never mind wrongdoing?
I believe, although I cannot develop the position in any detail here, that one can tackle both mysteries together by returning to another proposal that I made earlier in this chapter. In § 2 I explained how wrongful action leaves us with regrets that are hard to expunge and the repression of which is hard to justify. As I said there, not everyone actually experiences the relevant regrets. My point was that experiencing them is rationally appropriate. Our lives should be blemished subjectively because and to the extent that they are blemished objectively. More generally, our lives should feel as good, but only as good, as they are.48 This is not as it stands a doctrine of desert. It is not because of some deficiency on the part of the wrongdoer that she should be burdened with regrets for her wrongs. Many wrongdoers, like the captain in Vincent v. Lake Erie, emerge without any stain on their characters. But the doctrine that lives should feel as good, but only as good, as they are is nevertheless a doctrine of aptness or fittingness. It is apt or fitting that people, including those with unblemished characters, should suffer for their wrongs. What they should suffer, in the normal case, is regret for the very fact that they committed a wrong, for their incomplete fulfilment of their duty.
But the case of a wrongdoer who was at fault in committing a wrong presents special problems. Such a person has already shown himself insufficiently responsive to the reasons against doing as he did. That he was at fault means that he committed the wrong for defeated reasons. He did not even have undefeated reasons to treat the reasons for which he acted as undefeated. In failing to see that the reasons for which he acted were defeated, he was insufficiently responsive to the force of the reasons on the other side, those that made his action wrongful. This rational underresponsiveness also militates against his experiencing, without intervention, the apt measure of regret. The reasons for his regretting his wrong are, after all, the very same reasons to which he was not, when he committed the wrong, sufficiently responsive. They are reasons the force of which, thanks to his fault, he underestimates. Depending on which fault he exhibits, he either notices the (p.76) reasons yet underestimates their force, or underestimates their force such that he fails to notice them. Either way, the faulty wrongdoer is prone to live a life that feels better than it is. In fact, doubly so. His life is blemished by his wrong and (as we saw at the end of § 3) further blemished by his fault in committing it. His fault tends to make him oblivious to both blemishes. Punishment, assuming it is successful, serves to correct this oblivious misalignment of the subjective with the objective. It gives people who do not regret their faulty wrong (or do not regret it enough) something extra about their faulty wrong to regret. To the extent that they do not regret committing their faulty wrong qua faulty wrong, they regret it instead qua bearer of unwelcome consequences for them. They regret the (other kinds of) suffering and deprivation that constitute their punishment. And here we have, at last, a relevant principle of desert. It is only because of the wrongdoers deficiency—the fault she displayed in committing the wrong—that the deliberate infliction of such suffering or deprivation upon her in response to the wrong is apt. Had she been free of fault, the apt measure of suffering (the suffering of apt regret for the wrong) would have been self-administering and would not have called for deliberate infliction by anyone. No punishment, in other words, would have been deserved.
This line of thought provides, or at any rate promises upon adequate elaboration to provide, an integrated solution to our twin mysteries about deserved punishment. It indicates why one would want to inflict yet more damage on what is, ex hypothesi, a life already damaged by wrongdoing. And it indicates why wrongdoing, and not only fault, figures in the equation of deserved punishment. Yet it also creates new puzzles of its own. In particular, we may worry about faulty wrongdoers who act ‘out of character’. I do not mean those who possess many virtues (say, kindness, generosity, courage, loyalty) but one vice (say, dishonesty). There may be a case for punishing such people less when they commit a vicious wrong, but the case for punishing them less is not that they deserve less punishment for the vicious wrong. Rather, I am thinking of those who act ‘out of character’ in the sense identified in § 3: those who are normally honest, say, but who were on this one occasion dishonest. These people are not normally underresponsive to the reasons to which they were, on the occasion of their wrong, underresponsive. So, depending on how ‘out of character’ their faulty action was, they may well recognize, in the wake of their wrongs, the full force of the reasons that they underestimated in committing those wrongs. They may well see their wrong and their fault in the sharpest relief. If so, they are not prone to suffer less regret than is apt. Do they or do they not deserve to be punished?
I tend to think that they represent an intriguing borderline case. Here, briefly, is why. The apt response to faulty wrongdoing, thanks to its doubly blemishing impact upon the life of a wrongdoer, is not the ordinary regret of the faultless wrongdoer. It is a special kind of self-critical regret known as remorse. The remorseful give themselves a hard time for their wrongs, a hard time which they hold themselves to deserve on the model of punishment. In the wake of their remorse, they deserve to be punished less, or in extreme cases not at all. Suppose, (p.77) for simplicity’s sake, we are looking at a case of remorse in which no punishment is now deserved. How do we explain the fact that no punishment is now deserved? One way to do so is to say that the remorseful person never deserved punishment, because without intervention she reacted entirely fittingly to what she did. The alternative way to explain the same result is to say that she did deserve to be punished, but that she was punished enough by the hard time she already gave herself. Both explanations are intelligible, and both are sometimes given, because remorse lies at the borderline between regret, suffered spontaneously, and punishment, inflicted deliberately. But whichever explanation we prefer, the consideration of remorse undoubtedly belongs at the sentencing stage of the criminal trial. It is true that being at fault and deserving (further) punishment at the law’s hands somewhat come apart in the case of the remorseful wrongdoer. But the quantification of the (further) punishment deserved by such a wrongdoer cannot, consistently with the rule of law, be built into the law’s determination of criminal liability. To make a deduction from punishment for remorse (or indeed for punishments exacted extrajudicially, such as being abandoned by one’s spouse) the court must first determine that a criminal wrong has been committed without a legally recognized justification or excuse. Only then can the law legitimately authorize a prima facie quantification of the punishment deserved for the offence to which the relevant ‘remorse’ deduction (and other adjustments) can be applied.
6. Beyond Meaning and Consequence
Some say: only those who are at fault (= blameworthy) should be punished for their wrongs, because punishment expresses or communicates blame. I agree. I also agree with those who say, as Kant says, that faulty wrongdoers have an interest in being punished for their wrongs because the punishment expresses or communicates the fact of their being responsible agents, meaning agents who are capable of having and offering justifications and excuses for what they do.49 But these proposals, sound though they are, leave us hanging in mid-air. We still want to know: why does punishment express or communicate these things? Is it an accident of social history? Doffing one’s cap, except when done in irony, expresses or communicates deference. But one can readily imagine a culture in which it is a gesture of contempt. Is punishment similar? Could it imaginably, in some place or at some time, be a communication of forgiveness or exoneration rather than blame, or a denial rather than an assertion of responsible agency? I tend to think, and many seem to agree, that the expressive link between punishment and blame (and hence between punishment and responsible agency) is less contingent than that between hat-doffing and deference. It is logically possible to punish those (p.78) whom one does not blame—to make a scapegoat of an innocent person—but at the same time the social meaning of a punitive act as an act of blaming is not just an accident of social history, an ethnographical curiosity, a symbolism grafted onto the logic of punishment by local convention. Rather, this social meaning is somehow the proper social meaning for punishment, one that could only awkwardly or perversely be defied by a dissident civilization.
Why? This is where the deservedness of punishment, as I have explained it, enters the story. It is not true, as many claim, that the reason why only the blameworthy deserve to be punished is that punishment expresses or communicates blame. The truth is the opposite: the reason why punishment expresses or communicates blame is that only the blameworthy deserve to be punished. Admittedly, the fact that punishment expresses or communicates blame is itself a complete reason not to punish the blameless, and a reason with considerable force. Possibly it is this reason that does most of the hard work in making it the law’s duty not to attach criminal liability to faultless wrongdoing. Nevertheless it is a reason that takes its shape from another, namely the reason we have (quite apart from what punishment expresses or communicates) to punish only those who deserve to be punished. So even if you are not seduced by my sketchy and tentative proposals above in defence of the view that faultless wrongdoers do not deserve to be punished, there is still this message to take away with you: some such proposals are needed. We need some relatively independent explanation of the deservedness of punishment to explain why, from the Iliad to the Koran and the French Revolution to the Cultural Revolution, punishment expresses or communicates blame.
At least from Kant and Hume onwards, modern thinkers across many disciplines lost sight of the need for such proposals and lacked the resources to make them. They lacked the resources to make them because they endorsed as an article of faith the tabula rasa view of rationality, thought of as progressive and superstition-busting, according to which one always needs some further (new) reason to dwell on (old) reasons that were not perfectly conformed to at the right time. The old reasons are water under the bridge, the thinking goes; the time for conforming to them is over; we are no longer interested in them unless there is something else that can now be achieved, expressively or instrumentally, by attending to them. Without this ‘something else’ we cannot even get started with an argument for regret, reparation, remorse, punishment, and similar retrospective responses. Thus modern thinkers came to blows, mostly, about what the ‘something else’ in question might be. Reduction of net suffering? Restoration of respect? Reform of bad character? All of these and many other considerations can, of course, be relevant to the justification of regret, reparation, remorse, punishment, and so on, and especially to the costly and difficult institutionalization of such responses in the law.50 Nevertheless they are all, in one way or another, (p.79) parasitic considerations. One cannot provide adequate foundations for the expressive or the instrumental significance of any of these responses to wrongdoing unless one first understands how and when they are apt responses quite apart from their expressive or instrumental significance.51 In particular, one cannot make a philosophically satisfying expressive or instrumental case for punishment unless one first makes proposals for explaining how and when punishment is deserved.
As well as lacking the resources to make such proposals, many modern thinkers, taking their cue from Kant and Hume, lost sight of the need to do so. Underestimating or even denying the central place of tragedy in the human condition, they saw little or no space for lives that are morally better or worse than the people who live them. Experientially better, yes. So one may indeed ask whether someone has a better life than she deserves, meaning an experientially better life. But morally better, no. Either lives are not objects of moral assessment at all, but only of experiential assessment (in the Hume–Bentham tradition); or else the moral assessment of lives closely tracks the moral assessment of the people who live them (in the Kantian tradition, to which most contemporary ‘virtue-ethicists’ belong malgré eux). Either way, the elementary moral distinction takes a back seat. The idea that a life is often morally blemished through no fault of the person living it becomes philosophically alien. Alienated in the process is the idea that the primary or basic type of wrongdoing is faultless wrongdoing, and the associated idea that the primary or basic responses to wrongdoing (regret, apology, reparation) are apt responses irrespective of fault. Faultless wrongdoing and the strict liability to repair that is associated with it come to be thought of, mistakenly, as moral oddities in need of special explanation. Meanwhile fault-anticipating wrongs, and liability dependent on fault, come, mistakenly, to be thought of as the normal or default case, in need of no special explanation.52 Hence the need for a special explanation of the fact that punishment expresses or communicates the blame-worthiness of the wrongdoer goes increasingly unnoticed. Since punishment (conceptually) is for wrongdoing, and wrongdoing is by default blameworthy, (p.80) what else would punishment express or communicate but blame? At this point, the main question left over is about the relative importance of expressing or communicating things. Should we deliberately inflict suffering and deprivation merely to express or communicate something? Or should we be prepared to do it only to reduce other suffering and deprivation? In this debate the claim that punishment for faulty wrongdoing is deserved loses its independent role. The deservedness of punishment becomes the output rather than the input of endless arguments about punishments meanings and consequences and the relative justification importance of the two. Moral philosophy has lost sight of the need to rely on the deservedness of punishment to explain punishment’s meanings and consequences. In the process it has lost sight of the need to explain, in a way that does not depend on punishments meanings or consequences, why the only actions that deserve to be punished are both wrongful and blameworthy (two different and only very obliquely related properties).
(1) A good example: R. M. Dworkin, Taking Rights Seriously (London 1977), 272 ff.
(2) Of course tolerating a persons unrespectable way of living may involve respecting other things of hers, e.g. her rights. This was one of the themes of a paper that I wrote with Stephen Shute, ‘The Wrongness of Rape’ in Jeremy Horder (ed.), Oxford Essays in Jurisprudence: Fourth Series (Oxford 2000).
(3) Kant reminds us: ‘The object of reverence is the [moral] law alone…. All reverence for a person is properly only reverence for the law (of honesty and so on) of which that person gives us an example.’ Groundwork of the Metaphysic of Morals (trans. Paton, New York 1964), 69 n.
(4) B. Williams, ‘Moral Luck’, Proceedings of the Aristotelian Society, Suppl. 50 (1976), 115; Thomas Nagel, ‘Moral Luck’, Proceedings of the Aristotelian Society, Suppl. 50 (1976), 137.
(5) Groundwork, 62 ff.
(6) See R. Crisp’s introduction, ‘Modern Moral Philosophy and the Virtues’ in Crisp (ed.), How Should One Live: Essays on the Virtues (Oxford 1996), 5–6.
(7) Compare Kant, Groundwork, 62, with Aristotle, Nichomachean Ethics 1153b 16–21. Kant holds that perfectly good character is necessary and sufficient for a perfectly good life; Aristotle that it is necessary but insufficient.
(8) Poetics 1453a1–17. For further discussion see M. Nussbaum, The Fragility of Goodness (Cambridge 1986), 378 ff.
(9) See Arthur Miller s 1949 New York Times essay ‘Tragedy and the Common Man’, repr. in Robert Martin (ed.), The Theater Essays of Arthur Miller (New York 1978), 3–7. Miller diagnosed a loss of interest in tragedy associated with the modern democratization of theatre. In response he argued, and demonstrated in Death of a Salesman (1949), that tragedy is not an undemocratic genre. That classical tragedians used grandees as heroes was not essential to their art. Grandees make a bigger splash when they fall, which can be good for dramatic impact, but those whose heroism is the heroism of ordinary life may fall just as far, and just as disproportionately to their failings, and hence conform just as fully to the classical idea of the tragic.
(10) The main exceptions are pleasures and pains. They are capable of having a direct as well as an indirect effect on the quality of my life. Among pleasures and pains there are raw or sensory pleasures and pains, that straightforwardly happen to us. But there are also pleasures and pains reflecting an awareness of value, such as the joy of falling in love and the pain of losing a loved one. Arguably the latter belong (in non-pathological cases) to the active side of our lives, and so are not ideally represented as things that happen to us. On the other hand they are also misrepresented as things that we do. For discussion see Harry Frankfurt, ‘Identification and Externality’, in his The Importance of What We Care About (Cambridge 1985), and Joseph Raz, ‘When We Are Ourselves: The Active and the Passive’, in his Engaging Reason (Oxford 1999).
(12) 124 NW 221 (1910).
(13) For different degrees of indigestion see Ernest Weinrib, The Idea of Private Law (Cambridge, Mass. 1995), 196 ff., and Arthur Ripstein, Equality, Responsibility and the Law (Cambridge 1999), 118–21. The following thought often deepens lawyers’ worries about Vincent. The captain is justified in keeping his ship trespassorily moored to the pier, but since it remains an actionable trespass the pier-owner is also justified (under the common laws doctrine of self-help) in using the minimum necessary force to cut the trespassory ship adrift. One may therefore imagine a situation in which the pier-owner uses minimum necessary force to cut the ship adrift, but the captain responds in like kind by using minimum necessary force to keep the ship moored. The situation escalates, and third parties join in on each side, all claiming justification. Whose side does the law take? Who is the criminal aggressor and who is the legitimate self-defender? I agree that this is a potential problem for the law. But it is not a philosophical problem. There is no reason to doubt that the two sides could both be justified simultaneously even though the justified action of one impedes or frustrates the justified action of the other. There is only the practical problem of how we are going to prevent such situations proliferating and escalating to the point at which everybody is acting unjustifiably rather than justifiably This is an ordinary co-ordination problem to which there are a number of rival solutions, more than one of which is in all probability a justifiable solution. This being so, different legal systems may justifiably choose different solutions. In my view they may also justifiably remain silent on the problem until a relevant case arises and even then only solve it very locally without anticipating all the possible variations on the scenario.
(14) It may block ones liability when taken in combination with other facts, such as the fact that what justified one in doing as one did was the need to protect or rescue the very person whom one thereby wronged.
(15) For a good sketch see R. Craswell, ‘Two Economic Theories of Enforcing Promises’, in P. Benson (ed.), The Theory of Contract Law (Cambridge 2001).
(16) e.g. D. Friedmann, ‘The Efficient Breach Fallacy’, Journal of Legal Studies 18 (1989), 1.
(17) Nichomachean Ethics 1110a8.
(18) MacCormick, ‘The Obligation of Reparation, in his Legal Right and Social Democracy (Oxford 1982). MacCormick gets into a muddle with efficient breach—he imagines that a justified breach of duty cant be a breach of duty—and so ends up inverting the lesson of the example.
(19) In ‘Moral Luck’, Bernard Williams introduces the label ‘agent-regret’ to designate regret with (something like) this rational structure. Unlike Williams, I tend to think that all regret is agent-regret, in the sense that it reflects (what the regretter takes to be) incomplete conformity with reason. Some regret is, however, vicarious agent-regret. It is regret on behalf of others, reflecting the continuing force of some reason that they did not completely conform to. Vicarious regret clearly needs some justification beyond the existence of the unconformed-to reason. We need to know why one should ever respond to another’s non-conformity to a reason as if it were one’s own. I will not give any attention to this question here, nor indeed to any of the closely associated puzzles of vicarious liability in law.
(20) See J. Gardner and T. Macklem, ‘Reasons’, in J. Coleman and S. Shapiro (eds.), The Oxford Handbook of Jurisprudence and Philosophy of Law (Oxford 2001), 463–4, 467–8. I have benefited from reading two more recent papers, each of which independently defends much the same view in greater detail: M. Henken, ‘No Way Out: Conflict, Regret and Compensation’ (unpublished, 2002), and J. Raz, ‘Personal Practical Conflicts’, in P. Baumann and M. Betzler (eds.), Practical Conflicts: New Philosophical Essays (Cambridge 2004), 172. The latter volume also contains an interesting critique of the same view, Monica Betzler’s ‘Sources of Practical Conflict and Reasons for Regret’.
(21) R. Crisp, ‘Modern Moral Philosophy and the Virtues’, in Crisp (ed.), How Should One Live, 1.
(22) This seems to be Alasdair MacIntyres view in After Virtue (London 1985), 204 ff.
(23) See e.g. Paul Robinson, Fundamentals of Criminal Law (2nd edn., New York 1995), 526; Claire Finkelstein, ‘Excuses and Dispositions in Criminal Law’ (2003) 6 Buffalo Criminal LR 317, at 326.
(24) Hume, Enquiries concerning Human Understanding and concerning the Principles of Morals (3rd edn., ed. Selby-Bigge and Nidditch, Oxford 1975), 98.
(25) Of course this does not make her a positively honest person. As Aristotle explained it makes her an enkratic (self-controlled) person who is to that extent neither honest nor dishonest: Nichomachean Ethics 1145a15 ff.
(26) This is why Aristotle plays up the ‘settled disposition aspect of moral virtue in his discussion of moral education in book 2 of Nicomachean Ethics, but plays it down from book 3 onwards, having turned his mind to moral assessment.
(27) In his proto-utilitarian way, of course, Hume thought that assessing people was mainly a step towards improving them, which helps to explain why he thought that an action being out of character made it irrelevant to assessment. See Enquiries, 97–9.
(28) See further my ‘Justifications and Reasons’ in A. P. Simester and A. T. H. Smith (eds.), Harm and Culpability (Oxford 1996).
(29) In special circumstances this might serve as a justification rather than an excuse (e.g. if the captain undertook in his contract of employment, or as part of his professional accreditation, to heed all credible storm warnings).
(30) See further my ‘The Gist of Excuses’ (1997) 1 Buffalo Criminal LR 575.
(31) A recent reassertion of this view is William Wilson, ‘The Filtering Role of Crisis in the Constitution of Criminal Excuses’ (2004) 17 Canadian J Law and Jurisprudence 387.
(32) I am here borrowing some points from Joseph Raz, ‘Permissions and Supererogation, American Philosophical Quarterly 12 (1975), 161. A common mistake, pointed out by Raz, is to think that the permissibility of not performing a supererogatory action constitutes an excuse rather than a justification for its non-performance. But even if we were to let this mistake stand, the final conclusion still follows: nobody is at fault in failing to act supererogatorily.
(33) ‘A pure virtue ethics … will suggest that the only reasons we ever have for acting or living in a certain way are grounded in the virtues.’ R. Crisp, ‘Modern Moral Philosophy and the Virtues’ in Crisp (ed.), How Should One Live, 7. Crisp rightly points out that a virtue ethics need not be pure, but he wrongly ascribes the pure position to Aristotle.
(34) Nichomachean Ethics 1114b28–9; Groundwork, 102–3.
(35) e.g. Nichomachean Ethics 1098a16–17.
(36) I have defended this explanation in more detail, with the virtue of solidarity as my example, in ‘Reasons for Teamwork’ (2002) 8 Legal Theory 495.
(37) The best and best-known defence of the view that there are no fault-anticipating wrongs is W. D. Ross, The Right and the Good (Oxford 1930). The conclusion is stated most starkly at p. 45. Ross’s argument for it fails, but in the process he makes light work of demolishing the ultra-Kantian view at the opposite extreme that all wrongs are fault-anticipating wrongs. Sadly, Ross’s demolition fell mostly on deaf ears. The ultra-Kantian view continues to exert a hold over many non-philosophers, including many legal scholars. I call the view ultra-Kantian because even in his most hardline moments Kant claimed only that breach of duty is never justifiable. He did not suggest that it is never excusable, and therefore always faulty. The idea that he did reflects, like so much else, an exaggeration of his views about moral luck. From the thesis that perfect people cannot but lead perfect lives, it is tempting but fallacious to derive the thesis that an imperfect life can have been led only by someone who is at fault. This line of thought ignores the existence of limitations, i.e. imperfections of character that are not vices. Ross himself tends to paint Kant, mistakenly, in ultra-Kantian colours.
(38) John Gardner, Obligations and Outcomes in the Law of Torts’, in Peter Cane and John Gardner (eds.), Relating to Responsibility: Essays for Tony Honoré on his 80th Birthday (Oxford 2001).
(39) But they need not have similarly stark notice of their possible justificatory and excusatory defences. For explanation, see George Fletcher, ‘The Nature of Justification, in Stephen Shute, John Gardner, and Jeremy Horder (eds.), Action and Value in Criminal Law (Oxford 1993), 175.
(40) It has become fashionable across the political spectrum to attack the pro-defendant principles of criminal law by reconceptualizing the criminal process on the model of a civil dispute, i.e. as wrongdoer v. person wronged rather than wrongdoer v. the law. I criticized this consumerist trend in ‘Crime: in Proportion and in Perspective’, in A. J. Ashworth and M. Wasik (eds.), Fundamentals of Sentencing Theory (Oxford 1998).
(41) See ‘Legal Responsibility and Excuses’, in his collection Punishment and Responsibility (Oxford 1968).
(42) In my view, Stephen Shute, Jeremy Horder, and I exaggerated the importance of this distinction in our introduction to Action and Value in Criminal Law, and hence went too far in disengaging the principles of criminal law from the principles of punishment.
(43) An alternative (more Aristotelian) conceptualization of the main idea in this paragraph: ideally a determination of criminal liability settles that punishment would be just, while questions about whether it would be humane, merciful, efficient, etc. are dealt with at the sentencing stage. Thanks to the demands of the rule of law, however, some considerations of justice (the particularized considerations of justice known by the name of ‘equity’) are inevitably left over to the sentencing stage. See Nichomachean Ethics 1137b12–1138a2.
(44) The most influential statement of this view is John Rawls, A Theory of Justice (Cambridge, Mass. 1971), 103–4.
(45) The same point is nicely made by Michael Sandel, Liberalism and the Limits of Justice (Cambridge 1982), 135–47. Sandel focuses on the advantages enjoyed by the (innately) academically talented in university admissions. As Sandel shows, if these advantages of talent are undeserved it is not because (as Rawls puts it) the notion of desert seems not to apply to these cases’. One may clearly deserve the rewards of ones talents, e.g. in winning a talent contest. The only question is whether competition for university admissions should be anything like a talent contest, and that depends on the nature and value of universities, not on anything about the concept of desert.
(46) I am here rehashing Anthony Quinton’s ‘On Punishment’, Analysis 14 (1954), 512.
(47) In the sense relevant to an understanding of punishment, suffering or deprivation is not deliberately inflicted unless inflicting it is one of the inflictor s objectives. If no suffering or deprivation materializes, then the punishment fails qua punishment. This explains why making a wrongdoer pay reparative damages, however vast, is not normally a punitive act. Normally the deprivation of the wrongdoer who pays reparative damages is a mere side-effect of getting her to do her duty So the fact that she is insured against the deprivation is a plus from the reparative point of view, even though it would have been a minus from the punitive point of view. The puzzle I am raising in the text is therefore a puzzle about the inflicting of punishment that does not extend to the awarding of reparative damages. It is a heightened version of the more general puzzle: how could the fact that someone escapes suffering and deprivation ever count as a minus for a rational agent? How could bringing more suffering or deprivation into the world ever be a worthy objective, as opposed to an unavoidable side-effect of a worthy objective?
(48) Of course, how a life feels is itself part of how that life is. A life is damaged by the suffering of regrets, even apt regrets. By my doctrine, such a life is itself apt to be regretted by the person living it. ‘I regret that I spent so much time regretting what I did’ would be an intelligible, if sad, deathbed utterance.
(49) This was the theme of my paper ‘The Mark of Responsibility’, Oxford Journal of Legal Studies 23 (2003), 157.
(50) Here I have bracketed this question: granted that punishment or some other response to wrongdoing is called for, why is this the business of the law? Elsewhere I placed the need to control the excessive reactions of others, such as victims and their families, at the centre of the answer. See my ‘Crime: in Proportion and in Perspective’.
(51) By ‘instrumental’ here I mean ‘based on the consequences of punishment’. When punishment is successful (i.e. when the person whom we attempt to punish is actually punished) it has a certain result, namely that the person punished suffers or is deprived and thereby has something else to regret apart from his wrong. A result of an action is an outcome that is partly constitutive of that action. A consequence is an outcome that is not. An action’s results, to the extent that they bear on its value, bear on its intrinsic (non-instrumental) value. That the deservedness of punishment depends on the result of punishment does not show, therefore, that the case for deserved punishment (or for the principle by which it is deserved) is partly instrumental.
(52) Bernard Williams saw the importance of this mistake and made sustained efforts to correct it. See especially his Ethics and the Limits of Philosophy (London 1985), 174 ff., and ‘Internal Reasons and the Obscurity of Blame’ in his Making Sense of Humanity (Cambridge 1995). Alas, Williams misdiagnosed the mistake and ended up reinforcing it. He traced the blameaholics’ emphasis on blame to an emphasis on wrongdoing (breach of obligation), which in turn he criticized. If he had learnt the lesson of his own paper ‘Moral Luck’, he would not have made this connection. Not all blameworthy actions are wrongful and not all wrongful actions are blameworthy. An emphasis on wrongdoing does not naturally carry an emphasis upon blame in its wake.