аЯрЁБс>ўџ  РУўџџџжБВГДЕЖЗИЙКСџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџьЅСs ПS“jbjbР Р . ЊkЊkSџџџџџџ]иииииииLLLLL XL/ *€€€€€€€€ьюююююю,Y єM P и€€€€€ €ии€€€€€€€и€и€ььАœАииии€ь€l€ьииьt ŠѕђСLL€ь Philosophy of Law
Brandeis University Web Stite

Corrective Justice and Formalism



Copyright (c) 1992 Iowa University
Iowa Law Review

January, 1992

77 Iowa L. Rev. 449

LENGTH: 37639 words



SYMPOSIUM:: The Moral Foundations of Tort Law +



+ Copyright 1992 Stephen R. Perry. All rights reserved.

NAME: Stephen R. Perry *

BIO:

* Associate Professor, Faculty of Law, McGill University. I would like to thank the participants in the University of Iowa College of Law Conference on Formalism and Corrective Justice for their discussion of the version of the Article that I presented there. I am particularly indebted to Jules Coleman for the many helpful conversations I had with him before, during and after the Conference, and to Ernest Weinrib, who was kind enough to send me extensive written comments on my Conference draft. I have also benefited from comments received when a later version of the Article was presented to a Legal Theory Workshop at Columbia Law School.

SUMMARY:
  ... Theories in the second category are based on individual moral rights, and are often identified by the label of "corrective justice." ... But it is evident that Aristotle thought the principle of corrective justice applied, and hence correlative rights and obligations of reparation arose, even in situations where it could not be said in any ordinary sense that the gainer's gain equalled the loser's loss. ... The argument is noteworthy nonetheless because it suggests that the localized scheme of distributive justice should be expanded so as to take account of the moral worth of actions that in no way causally contributed to the harm of which the victim is now complaining. ... The difficulty that we discovered with Weinrib's characterization of intentional wrongdoing emerges even more strongly here, however, because the risk with which tort law is concerned is risk of harm, in the sense of interference with an interest, and abstract right has no place for harm thus understood. ... But the volitionist argument for strict liability comes closer to forging the necessary connection by recognizing the fundamental importance of outcome-responsibility, and the volitionist/distributive argument for fault builds on that insight. ...  

TEXT:
 [*449]  Introduction

Proposed justifications for the institution of tort law can be divided into two main categories. The first is comprised of theories that look upon civil liability as a means for advancing one or more public policies such as general social compensation, wealth redistribution, loss-spreading, or the attainment of economic efficiency through deterrence. 1 Theories in the second category are based on individual moral rights, and are often identified by the label of "corrective justice." They can be divided into two sub-categories. The first of these is best exemplified by Jules Coleman's "annulment" theory, 2 which takes the fundamental principle of corrective justice to be that wrongful (or unwarranted) gains and losses should be eliminated or annulled. The responsibility of annulment that the theory recognizes does not ultimately belong to specific individuals, such as the person who caused a particular wrongful loss; it is, rather, essentially social in nature. Theories in the second sub-category, by contrast, regard corrective justice as involving a limited moral relationship that holds only between injurer and victim. Under certain circumstances one person who injures another has an obligation owed specifically to the victim to compensate for the harm caused; the victim has a correlative right against the injurer to receive compensation, but no similar right against anyone else. Ernest Weinrib has defended the best known theory of this type. 3

On the annulment theory of corrective justice that Coleman defended for many years, 4 someone who has suffered a wrongful loss at the hands of  [*450]  another is entitled to have the loss annulled by being paid compensation. But the obligation to compensate is not the injurer's, or at least not exclusively: it is a general social responsibility involving more widespread reasons for action. Coleman thus rejected the correlativity of rights to repair and obligations to pay for harm caused, which is the hallmark of theories of corrective justice in the second sub-category. This led him to distinguish between the grounds of recovery and the grounds of liability in corrective justice, and between the grounds and the modes of both recovery and liability. I have argued elsewhere that the core concern of the annulment theory is, in the end, distributive rather than corrective justice. 5 This is not necessarily to say that the theory is wrong, but only that Coleman mischaracterized it. The better general conception of corrective justice, and certainly the dominant view within the tradition of theorizing in this area, is the one presupposed by theories in the second sub-category. Assume that A has causally contributed to an injury suffered by B. The concern of these theories is to establish when and why A owes B an obligation of reparation -- that is, an obligation to compensate B for the injury suffered -- that is correlative of a right B holds against A to recover for the loss. 6 There are a number of quite different accounts of corrective justice thus understood, but it has proven surprisingly difficult to specify the circumstances under which correlative rights and obligations of reparation arise and to say why they are justified.

In this Article I offer a survey, taxonomy, and critique of the most important theoretical attempts to characterize and justify correlative rights and obligations of reparation. After concluding that all these attempts fail, I argue for a different approach. I should say at the outset that the title of the Article is somewhat misleading, in that I focus almost exclusively on moral theories of reparation rather than on the institution of tort law. In my view, the principles of reparation that I develop do constitute the main moral foundations of tort law, but to establish that conclusion in a persuasive way requires more argument than I offer here. The incorporation of pure moral principles into legal or social institutions is an exceedingly complex business, and I do not rule out the possibility in advance that a kind of moral pluralism prevails within the institution of tort law. This might mean, for example, that pure principles of reparation are balanced against, or at least qualified by, norms of economic efficiency. 7 As I say, I do not rule out such pluralism, but I do not argue for it, either. For the most part I simply do not address such issues at all. I mention them here only to avoid giving the impression that the relationship between pure principles of reparation and the law of torts is necessarily a straightforward, one-to-one, correspondance.

 [*451]  Most of the arguments that have at various times been advanced to justify correlative rights and obligations of reparation can be divided into three types. Arguments of the first type attempt to reduce reparation to restitution: A has come into possession of something that belongs to B and hence must give it back. Arguments of the second type start with the fact that a loss has occurred, and are based on a kind of localized distributive justice: B has experienced a loss which is transferrable but which will nonetheless have to be shouldered by someone; as between A and B it is morally preferable that the loss be borne by A, since she is the person who (wrongfully) caused it in the first place; A should therefore be fixed with an obligation of reparation, the effect of which will be to "redistribute" the loss to her. In one version this form of argument offers a justification for strict liability, while in another it supports a principle of fault. Arguments of the third type focus on the normative implications of voluntary action: A has acted, perhaps wrongfully, and as a result of that action B has been injured; one of the appropriate normative incidents of A's (wrongful) conduct is that she should pay compensation to B. As with the second category, this form of argument comes in fault and strict liability versions.

These three categories of argument that purport to justify a principle of reparation are considered in turn. I begin by showing that arguments of the first type are perhaps capable of sustaining some very limited rights and correlative obligations of reparation, but not the entire range of such rights that we intuitively think should be recognized. I then argue that all versions of the second and third types of argument advanced to date are either flawed or incomplete. Finally, I sketch what I hope is a more successful argument for justifying rights and obligations of reparation that has strong affinities with both the type two argument in favor of fault and the type three argument supporting strict liability. The principle of reparation for which I argue is itself fault-based.

I should make clear that most of the proposed justifications for moral rights and obligations of reparation that I consider in this Article presuppose at least some measure of independence between corrective and distributive justice. If one were to assume that corrective justice was wholly ancillary to a simple patterned theory of distributive justice, so that its point was merely to regulate and preserve the pattern, certain obvious difficulties would arise. It would become awkward to maintain that corrective justice is concerned only with interactions between persons, and indeed only with certain types of interactions, as has traditionally been supposed, and not with other ways of departing from the pattern that require, so far as distributive justice is concerned, some kind of rectification. Moreover, since there would be no basis for regarding the local rectification of "corrective justice"-type disruptions of the pattern as having priority over other, possibly incompatible adjustments that might be required to maintain the pattern, the conclusion that the concept of corrective justice should simply be discarded would become very difficult to resist. 8 It is true that there are  [*452]  subtler and more plausible accounts of how corrective justice can be understood as a principle wholly ancillary to distributive justice, a prominent example being the theory of tort advanced by Ronald Dworkin in Law's Empire. 9 Consideration of such accounts is beyond the scope of this Article, although in the end I think they are no more successful than the simplistic view just described. It should also be borne in mind that proponents of the arguments discussed in this Article often make quite different assumptions about the basis and extent of the independence of corrective from distributive justice. This is, again, not an issue that can be considered here. 10

Finally, let me briefly mention an issue of terminology. This Article is concerned with the principles of reparation that underpin tort law. These principles are often labelled "corrective justice," and I sometimes use that expression as well. I prefer to speak of principles of reparation, however, not only in order to avoid confusion with Coleman's annulment theory, but also because the term "corrective justice" is often given different senses even by those who limit its application to correlative rights and obligations between two persons. In the usage of some writers the term refers only to reparation, that is, to principles of responsibility for harm caused of the kind found in tort, but others extend its meaning to include the principles of restitution and contract. In this Article, the context should generally make clear whether the wider or narrower sense of the term is intended.

The Article is long. For those who wish to concentrate on the positive argument, parts I, II(A) and III(A) are less crucial than the remainder.

I. Reparation as Restitution

Aristotle introduced the idea of corrective justice in Book V of the Nicomachean Ethics, 11 where he contrasted it with distributive justice. The latter was said to involve a so-called "geometric proportion," which requires that a good be distributed in accordance with a criterion of merit or desert among a group of persons that is determinate in number, but where there is no antecedently-fixed upper limit on the group's size. Aristotelian distributive justice is, in Robert Nozick's phrase, patterned distributive  [*453]  justice: 12 a person receives a share of the good proportional to his or her desert or merit. While unequal shares are clearly possible, Aristotle saw this form of justice as giving rise to a form of equality through the ratios it yielded. If the distribution has been carried out justly, the ratio of A's merit to B's merit will be the same as the ratio of A's share to B's share.

Aristotle said that corrective justice, by contrast, is a matter of an "arithmetic proportion." This form of justice applies to transactions, both voluntary and involuntary, which take place between one individual and another. Aristotle said that the law "treats the parties as equals," 13 which is a phrase undoubtedly amenable to more than one interpretation but which I shall provisionally assume to mean that both begin with a just distribution of holdings; prior to the transaction they are equal in the sense that each has what is his or her due under distributive justice. As a result of the transaction, one party suffers a gain and the other a loss. The judge "tries to . . . restore the equilibrium," 14 or "restores equality," 15 in the following way: 
As though there were a line divided into two unequal parts, he takes away the amount by which the larger part is greater than half the line and adds it to the smaller. Only when the whole has been divided into two equal parts can a man say that he has what is properly his, i.e., when he has taken an equal part. The equal is median between the greater and the smaller according to arithmetical proportion. 16



Aristotle seems to limit the category of involuntary transactions to what we would today call intentional torts. His examples include theft, adultery, enticement of slaves, bearing false witness, assasination and assault. As regards the nature of the gains and losses to which such transactions give rise, he has this to say: 
When one man has inflicted and another received a wound, or when one man has killed and the other has been killed, the doing and suffering are unequally divided; by inflicting a loss on the offender, the judge tries to take away his gain and restore the equilibrium. For in involuntary transactions we use the term "gain" without any qualification, even though it is not the proper term in some instances (e.g., when a person has inflicted a wound), and we use the term "loss" in similar way when he is the sufferer. But, at any rate, we do speak of "loss" and "gain" whenever the damage sustained can be measured. 17


 
 [*454]  Later Aristotle adds that "the just in involuntary transactions occupies the median between a gain and a loss: it is to have an equal amount both before and after the transaction." 18

This account of reparation is interesting, because on one very plausible interpretation it suggests that an obligation of reparation is to be understood in terms of a limited obligation of restitution. Two persons interact in such a way that one ends up with some of the justly held holdings of the other; since the interaction is involuntary the transfer is not consensual, and is therefore unjust; under these circumstances the appropriate moral response is to require either that the appropriated holdings themselves be returned or, perhaps, that the person in possession make good their value. A limited restitutionary obligation of this sort -- let me henceforth refer to it as an obligation of restoration -- seems to have almost axiomatic status within a system of private property. One of the defining characteristics of private property as a method of distributing material resources is an owner's right of exclusive, although not necessarily absolute, control over an object. 19 If this right is to be meaningful, then a nonowner who comes into possession of the object must have an obligation to return it to the owner. 20 If a particular system of private property is justified in moral or political terms, then the obligation of restoration will also be justified. Some might think it necessary to qualify the obligation and its correlative right in certain ways, but a core obligation of the kind described would seem to be an unavoidable feature of any system of private property.

The proposed justification for a reparative principle which I have here attributed to Aristotle, and which I shall accordingly refer to as the Aristotelian argument, 21 is an instance of the first type of argument for correlative rights and obligations of reparation. The identifying characteristic  [*455]  of such arguments is the claim that reparation can in some way be reduced to restitution. As I noted above, Aristotle seems to restrict the class of involuntary transactions that will give rise to an obligation of reparation to intentionally wrongful torts. He also seems to assume that at least in paradigm cases of corrective justice, the wrongdoer's gain is equal to the victim's loss. Actual appropriations of the sort involved in intentional wrongs like theft or fraud are especially clear instances of such equality. There would seem to be no reason, however, not to recognize the obligation and correlative right in at least some situations where the gainer did not engage in an intentionally wrongful act, as for example where she came into possession of some item of property belonging to the loser unknowingly, or by mistake. This would be in keeping with the restitutionary foundations of the obligation.

The Aristotelian argument takes as its starting point an obligation that will necessarily exist within any system of private property, namely, the obligation of restoration, and then attempts to reconstruct a more general obligation of reparation. 22 It is worth emphasizing that, at least as I am interpreting the argument, it begins with an aspect of the concept of property and not with the concept of distributive justice. 23 It is true that property is assumed to be justly held, and for Aristotle, at least, this was presumably a matter of conformity with a patterned conception of distributive justice. This is a natural reading of the phrase "the law treats the parties as equals," and it also makes sense of the metaphor of the two  [*456]  initially equal lines Aristotle uses to describe the baseline of corrective justice. 24 But the point of corrective justice is not, according to the Aristotelian argument, the maintenance or preservation of a distributive pattern as such; it is the enforcement of the obligation of restoration. If distributive justice enters the picture it is only indirectly, as an element in the legitimation of existing property rights. As a general matter it is not necessary that it enter even to this extent, since there are theories of property that do not make the legitimacy of entitlements turn on whether the requirements of a patterned conception of distributive justice have been met. 25 The Aristotelian argument is thus not subject in any obvious way to the difficulties noted in the Introduction that are faced by theories which make corrective justice wholly ancillary to distributive justice.

The argument does, however, give rise to at least two other difficulties, both of which are concerned more with its scope than its content. The first pertains to the range of interests that the principle of reparation is supposed to protect. Interests other than property entitlements in material resources are ordinarily thought to be subject to claims of reparation. This was clearly Aristotle's view about life and bodily integrity, for example, since he says that corrective justice applies to acts of assault and murder. But if the Aristotelian argument is to apply to such interests it must be shown that they either involve property rights in the strict sense or are governed by general normative relations among persons that in relevant respects are similar to such rights.

Like Aristotle, Richard Epstein regards life and bodily integrity as falling within the scope of reparative principles. As I shall argue shortly, he also attempts to reduce reparation to restitution in the manner of the Aristotelian argument. It is perhaps this combination of positions that has led Epstein to claim that each of us owns his or her own person or body in the same way that we have rights of private property in material resources. 26 But theorists of property have treated claims like this as problematic, or at least as requiring further normative argument. 27 Epstein himself has qualified his position by stating that "if the ownership language does seem artificial, the language of 'personal integrity' can be substituted without change of effect." 28 While this is no doubt true in some contexts and for some purposes, property interests and the interests we have in our physical persons are in many important respects dissimilar. It  [*457]  is plausible to think that the former but not the latter are subject to the requirements of distributive justice, for example. 29 Another difference, which bears more directly on our present inquiry, is that we do not regard rights to be free from interference with our persons as manifesting a concern only with maintaining possession and control of a material resource. Reparation for personal injury seems similarly to involve more than a simple obligation to return something to its owner, and this suggests that the Aristotelian argument has little purchase here: the obligation of reparation is unlikely in personal injury cases to be explicable as, or reducible to, an obligation of restoration. Similar conclusions will apply a fortiori to intangible interests, such as reputation, privacy, and the aspects of emotional wellbeing protected by the tort of assault, that we think are also shielded by general principles of reparation against at least certain kinds of harmful conduct.

The second and, I think, more significant difficulty with the Aristotelian argument is that despite Aristotle's claim to the contrary it does not seem to apply where the gainer's gain is not equal to the loser's loss. The basic situation in which an obligation of restoration arises involves the nonconsensual transfer of possession of the very same object from one person to another. Here gain and loss are equal, at least if we disregard the subjective preferences of the parties. The obligation of the gainer to disgorge the gain by returning the object (or, perhaps, by compensating the owner for its market value) seems clear enough, and it is perhaps possible to explain the obligation of reparation as it arises in certain property-related torts like conversion in these terms. 30 But it is evident that Aristotle thought the principle of corrective justice applied, and hence correlative rights and obligations of reparation arose, even in situations where it could not be said in any ordinary sense that the gainer's gain equalled the loser's loss. He gives the example of inflicting a wound, which certainly bears little obvious resemblance to the paradigmatic case in which a particular item of property or share of holdings belonging to one person is nonconsensually transferred to another. It is clear enough that a wounded person has suffered a loss, and it may also be that the person who inflicted the wound has gained a benefit (satisfaction of some sort, perhaps). But the loss and the gain will not in general be equivalent, as Aristotle recognizes when he says of such cases that "the doing and the suffering are unequally divided," 31 and "the term 'gain' . . . is not the proper term." 32 Repair of loss and disgorgement of gain thus come apart. 33 Since our concern is with  [*458]  reparation, we must focus on the former rather than the latter.

As was noted in the Introduction, the term "corrective justice" has both a wide and a narrow sense. In the narrow sense it refers to rights and obligations of reparation alone. In the wide sense it refers in addition to the rights and obligations that arise under the two other basic principles of private law, namely, the contract and restitution principles. It is evident from his examples and from his distinction between voluntary and involuntary transactions that Aristotle regarded all three principles as falling within the scope of corrective justice. In modern law these principles are clearly demarcated from one another. Aristotle may not, however, have drawn a sharp distinction between restitution and reparation, 34 perhaps thinking that the obligation of restoration could justify claims for both repair of loss and disgorgement of gain. It is far from clear that the obligation of restoration could serve as the foundation for the entire range of restitutionary rights and obligations recognized in modern law, even though it is restitutionary in character itself, 35 but that is not our present concern. The principle of reparation pertains to repair of loss, not disgorgement of gain, and consequently we must ask whether the obligation of restoration can ground an obligation to make good a loss even if the injurer did not make an equal gain, or any gain at all.

In the following passage from his book Takings, Richard Epstein can plausibly be understood as attempting to develop the Aristotelian argument along just these lines: 
Within the private law, the only difference between taking and destroying is that a claim for conversion becomes a claim for wrongful destruction. The destruction makes it very difficult to calculate the benfit obtained by the defendant, so the tort measure -- harm to the plaintiff -- becomes by default the sole basis  [*459]  for recovery. But the resemblances between the conversion of property and its destruction are powerful. Conversion involves the use of force to remove a thing from the possession of the owner. Destruction involves the use of force to work physical changes in things that remain in the possession of the owner. Taking and destroying are close substitutes for each other and each forms an essential part of the law of tort. Just as tort liability goes beyond conversion, so too does prima facie liability of the state under the eminent domain clause. Surely no one would argue that the state does not take private property when it blows up a building, or that thereafter it can condemn the land without paying for the building it has destroyed. 36

Later in the book Epstein says that this discusssion of conversion and destruction shows that "torts themselves are a sub-class of takings." 37

This is an interesting argument. Its central claim is that at least certain ways of causing damage to property are morally indistinguishable from the physical appropriation of property and so should give rise to the same reparative remedy. We can think of the argument as proceeding in the following manner. Assume first that A has knowingly or intentionally appropriated the property of B. The obligation of restoration discussed above would then require A to return the property to B. Suppose next that A intentionally damages the property while it is in his possession. Surely, the claim would be, his obligation of restoration now requires him not just to return the damaged property but to make good the loss he caused while he had control of it. The obligation requires him to return appropriated property in the state it was in when appropriated or, if this is impossible, to pay compensation for the subsequent deterioration or damage, at least when it was intentionally inflicted. If that is so, the argument continues, then it would be morally arbitrary to insist upon an actual appropriation in the first place. Intentionally damaging something after physically appropriating it and intentionally damaging it while it is still in the possession of its owner are morally equivalent, and both should be treated as takings calling for the same remedy of reparation. Destruction is a constructive taking, so to speak. One could then try to push the argument further, as Epstein in effect does elsewhere, 38 by maintaining that there is no morally relevant difference, so far as the assimilation of destruction to takings is concerned, among intentionally damaging property, knowingly subjecting it to a risk that subsequently materializes, and indeed simply causing damage unintentionally or nonnegligently. This is one of the lines of argument Epstein pursues in his attempt to justify a tort regime of strict liability.

I have argued elsewhere that the logic of this argument does not justify Epstein's final move from responsiblity for harm that results from the knowing imposition of risk (subjective negligence) to responsibility for  [*460]  harm however caused (strict liability). 39 But there is a problem with the argument at an even earlier stage. The crucial step is the proposition that the obligation of restoration requires someone who intentionally damages appropriated property to compensate the owner. No doubt an obligation to compensate does arise here, but it cannot be derived from the obligation of restoration: compensation is most plausibly regarded as depending on the intentional, and hence wrongful, nature of the damaging act. To see this, recall that the obligation of restoration seems to arise even where A comes into possession of B's property innocently or by mistake. If the argument under consideration were sound, then one would expect compensation to be owed for damage caused in the course of innocent as well as wrongful "appropriations." Damage inflicted intentionally or negligently during the course of an innocent appropriation may well require compensation, if mistake of fact concerning title is not a morally appropriate defense, but this does not seem to be true of damage caused entirely innocently. 40 This makes it clear, I think, that it is the wrongful character of the damaging act and not the obligation of restoration which leads us to think that compensation is owed. The normative principles that will explain why someone whose wrongful conduct causes loss is required to compensate have not, of course, been stated yet; that is the task of the remainder of this Article. But even so, it is evident that the Epsteinian argument, based on the supposed similarity between taking and destroying, does not succeed in reducing reparation to a limited conception of restitution. 41

The claim that a restitutionary justification for reparation could not be extended beyond cases of equivalent gain and loss was indirectly made some years ago by Jules Coleman. 42 Until quite recently, Coleman held that the right in corrective justice of a victim to recover for her loss was not correlative of an obligation on the part of her injurer to pay compensation. But he did recognize such correlativity where the injurer's gain and the victim's loss were themselves correlative, or, as Coleman also put it, where they were not "logically distinct." 43 The examples he gave of correlative gain and loss were fraud and theft, with respect to both of which the gain on one side is, leaving aside any subjective preferences of the parties, equivalant to the loss on the other. This situation was contrasted with that found in negligence cases, where any gain accruing to the negligent actor  [*461]  will be in the form of ex ante savings that result from his not having taken appropriate precautions; he secures no additional gain if his negligence also happens to injure someone else. Coleman suggested a restitutionary principle as the basis of the correlativity of right and obligation in fraud and theft cases. But since he clearly did not envisage the principle as extending to all wrongful gains the retention of which would constitute unjust enrichment, 44 it is reasonable to suppose that he had in mind the limited restitutionary obligation I have labelled the obligation of restoration. It is evident from the fact that Coleman did not regard the scope of the obligation and its correlative right as extending beyond for wrongs like fraud and theft that he did not think it laid the foundation for a general account of reparation. The arguments of this Part suggest that in this he was right.

II. Reparation as Localized Distributive Justice

The second general type of argument that has been put forward to justify correlative rights and obligations of reparation is based on what I shall call localized distributive justice. A causes B loss, perhaps as a result of wrongful conduct. The claim then is that as between these two, it is preferable that A should bear the loss. (The characteristic phrase used to introduce such arguments is "as between these two.") An argument of this sort is based on a claim of distributive justice because it focuses initially on the loss, which is regarded as a burden to be distributed among a specified group of persons. It is an argument of localized distributive justice because the group is limited to the victim and her injurer (or injurers). There are two main versions of this form of argument. According to the first, the fact that A acted, thereby causing B loss, is sufficient reason to shift the loss to A; this is an argument for strict liability. According to the second, there is only reason to shift the loss to A if A acted wrongfully; this is an argument for fault liability. These two versions of the distributive argument are considered in turn.

A. The Distributive Argument for Strict Liability

Holmes stated the distributive argument for strict liability very succinctly: 
Every man, it is said, has an absolute right to his person, and so forth, free from detriment at the hands of his neighbors. In the cases put, the plaintiff has done nothing; the defendant, on the other hand, has chosen to act. As between the two, the party whose voluntary conduct has caused the damage should suffer, rather than the one who has had no share in producing it. 45

 
It should be emphasized that Holmes did not think this argument was correct. He was merely stating the strongest theoretical case he thought could be mustered for strict liability so as to be able to refute it. The argument emphasizes the activity of an injurer in producing harm and the  [*462]  passivity of the victim in suffering it. The fact of having voluntarily performed an action is regarded as an appropriate basis for redistributing any loss that the action causes. This hints at a powerful idea, but the Holmesian formulation of it is incomplete. 46 Why should a voluntary decision to act have the normative significance the argument attributes to it?

The most compelling answer to this question, and the one given by another 19th century jurist, Baron Bramwell, is that if the loss were not redistributed to the actor then she would take the actual or potential benefits of her activity while shifting the costs to someone else. 47 In modern economic terms, she would be permitted to externalize her costs. On this view the normative implications of cost externalization are not limited to its effects on efficiency, since there seems to be something fundamentally unfair about forcing another person to bear some or all of the costs of your voluntary actions where only you stand to benefit from them. This suggestion, the intuitive appeal of which is clear enough, gives the distributive argument for strict liability the following cast. A caused B harm. A was active and stood to benefit from her activity, whereas B was simply the passive recipient of the effects of A's action. As between these two it is fairer that A rather than B bear the loss, since otherwise A would take the actual or potential benefits of her activity but would not bear all the costs, at least some of which would be passed on to B. 48

 [*463]  The main difficulty with this argument can be stated briefly. 49 There is no simple distinction to be drawn between the parties to a harmful interaction such that one of them can be labelled the "active" injurer and the other the "passive" victim. Suppose, for example, that A drives into and damages B's parked car. It is true that there is a sense in which A was active and B passive at the time of the accident. But, contrary to what Richard Epstein has sometimes claimed, 50 it is not possible to determine which of the parties should bear the loss by temporally limiting the reparative inquiry to the time at which harm actually occurred. In the example given it is, I think, obvious that the inquiry must consider actions performed by both parties -- A's act of driving, on the one hand, and B's act of parking her car, on the other -- even though one of these actions took place prior to the harmful interaction. Thus, one must consider not only whether A was driving recklessly, say, but whether B had parked her car in a dangerous location where A, even if driving normally, could not see it. 51 It is necessary, in other words, to treat both parties as active, even if the relevant actions might have taken place at different times and, possibly, well before the actual injury. This is true not just of the particular example but of harmful interactions generally, since a victim of a harmful interaction can, except in unusual circumstances, be said to have made a choice to be where she or her property was located when the harm she suffered occurred.

The claim that there is no meaningful distinction to be drawn between an active injurer and a passive victim has two distinct aspects, each of which bears upon the distributive argument for strict liability. The first is that one cannot say that A -- the supposedly active injurer -- was the cause of the harm while B was simply passively suffering the effects of A's activity. 52 Actions of both were causes of the injury, even if those actions did not occur simultaneously. This means that a theory of reparation that tries to tie loss allocation to causation and nothing else -- that is, a theory of general strict liability -- will necessarily be indeterminate. This conclusion follows whether one analyzes causation in terms of strong or weak necessity, these being the  [*464]  two predominant approaches to the subject in the modern legal and philosophical literature. 53 A version of the strong-necessity theory, which is essentially the venerable but-for test, has been defended by John Mackie. 54 Richard Wright has argued vigorously in favor of the weak-necessity approach in the form of the so-called NESS test (necessary element in a set of sufficient conditions). 55 In the example of the parked car struck by another car, both the but-for and the NESS tests would designate the actions of both A and B as causes of the harm. In the case of the NESS test this is because both actions would be conditions required to complete a set of actually existing conditons that together were jointly sufficient to produce the harm.

I have argued elsewhere that Epstein's attempt to produce a theory of causation that justifies the conclusion that one or the other of the parties to a harmful interaction was the cause of the injury fails. 56 He has not presented a pure theory of causation at all but rather an account of individual moral responsibility disguised as a theory of causation: the theory is shot through with normative considerations that are drawn, for the most part, from a distinctive conception of fault. 57 Epstein thus avoids the Scylla of indeterminacy only by surreptitiously embracing the Charybdis of fault. A similar fate, I believe, awaits any theory of causation that claims to be able to identify a particular action as the cause of a given harm. 58

The second important aspect of the claim that there is no meaningful distinction to be drawn between an active injurer and a passive victim concerns not the determinacy of a theory of general strict liability but its justification. The distributive argument for strict liability assumes that an active agent, in pursuit of her own goals and interests, causes harm to a passive bystander. It concludes that strict liability is justified because the harm represents an externalized cost of the agent's activity that she, as the party who stands to benefit, should in fairness be required to bear. But once  [*465]  it is conceded that both parties to a harmful interaction must in general be regarded as actively in pursuit of their own ends and interests, where the activity of each is causally implicated in the harm one of them has suffered, this conclusion becomes untenable. The simple idea that A is benefitting from, while B is bearing at least some of the costs of, a single action or series of actions performed by A can no longer be accepted. A harmful interaction generally results not from a single choice to act but from the intersection of two such choices. The fact that it is B rather than A who suffers injury is simply a contingent fact which tells us nothing about whose activity the cost "really" belongs to. A theory of strict liability that redistributed loss from the person who suffered it to whomever else happened to be causally involved in its production would avoid the indeterminacy problem, but it would also be morally indefensible. Once again we seem to be naturally driven towards normative criteria, such as notions of fault, for determining who among the persons causally contributing to a loss should ultimately be required to bear it.

Ronald Coase relies upon the idea that a loss cannot be uniquely attributed to one of two conflicting activities in his classic critique of A.C. Pigou's thesis that the economically appropriate way to deal with an externality is to place the cost, through governmental action of some sort, on the party who caused it. 59 Coase argues not only that private market transactions can, under certain conditions, achieve allocative efficiency without governmental intervention, but also that the Pigouvian conception of an externality is itself fundamentally flawed: a cost that is incurred when activities conflict cannot be assigned to one of those activities on causal grounds alone. Coase is sometimes accused of "causal nihilism," 60 or of having made a conceptual error in his discussion of causation. 61 But these criticisms are directed towards his claim that because "we are dealing with a problem of a reciprocal nature" we must ask "should A be allowed to harm B or should B be allowed to harm A?" 62 For example, if a rancher's cattle can roam unrestrained, they will damage the neighboring farmer's crops. But if the farmer is granted an injunction requiring the rancher to put up a fence, it is the latter's interests that will be harmed. Coleman notes that "the 'harms' society causes by making one rather than another entitlement decision are conceptually different from the harms two competing activities may or may not cause one another." 63 This is certainly true, but it is far from clear that the "problem" Coase had in mind here was the nature of causation or harm, as opposed to the general economic problem of what to do about the decreased value of overall production that has resulted from a conflict of activities. Surely it is innocuous to say that the  [*466]  relevant economic question is whether A should be allowed to cause harm to B (in one sense of "causing harm"), or whether B should be allowed to cause harm to A (in another sense of "causing harm.")

Whatever Coase may have meant by reciprocal harm is ultimately not very important, however, because he does not rely on the idea when he directly addresses the role causation plays in harmful interactions. In his discussion of the old nuisance case of Bryant v. Lefever, 64 in which the plaintiff's chimney smoked whenever he lit it because the defendant had built a wall nearby that kept the air from circulating freely, Coase says this: "The smoke nuisance was caused both by the the man who built the wall and by the man who lit the fires. . . . Eliminate the wall or the fires and the smoke nuisance would disappear. . . . If we are to discuss the problem in terms of causation, both parties cause the damage." 65 These passages do not involve the kind of conceptual mistake discussed by Coleman, since Coase clearly presupposes an understanding of causation based simply on strong or weak necessity. It has already been mentioned that in contemporary theorizing about causation, these two types of account predominate. It is therefore not possible to dismiss Coase's rejection of the idea that costs can be determinately assigned to particular activities on causal grounds alone with a simple claim that he was employing an eccentric or obviously mistaken conception of causation.

Coase's critique of the claim that a cost can always be uniquely assigned on causal grounds to one among a number of conflicting activities undermines both the determinacy and the justification of economically oriented theories of strict liability that call for the internalization of costs to the activities that caused them. So far as determinacy is concerned, Coase's critique is essentially the same as the objection offered above to the distributive argument for strict liability. This should hardly be surprising, since both the distributive argument and the economic theories conceive of strict liability as a mechanism for internalizing costs, differing only in the reasons that they view such internalization as desirable. For the distributive argument internalization is justified on grounds of fairness. For the economic theories it is justified by considerations of economic efficiency and the proper functioning of the market. Both are vulnerable, however, to the objection that internalization simply cannot be determinately achieved.

In an early article Guido Calabresi defended an economic theory of strict liability on the ground that "'tort' costs should be borne by the activity which causes them." 66 He adopted an explicitly Pigouvian approach to the externalities problem: "The function of prices is to reflect the actual costs of competing goods, and thus to enable the buyer to cast an informed vote in making his purchases." 67 Subsequently, however, Calabresi moved away  [*467]  from this account of tort liability and adopted instead a theory of general or market deterrence, which requires that accident costs be allocated to the actor or activity able to avoid them most cheaply. 68 At the same time he acknowledged that the problem of determining which costs should be allocated to which activities cannot be solved by a formula, let alone by "a metaphysical search for ultimate causes." 69 This abandonment by an eminent tort theorist of the Pigouvian approach to externalities reinforces the conclusion that the distributive argument for strict liability, based as it is on a similar concern for internalizing costs, is seriously flawed. 70

B. The Distributive Argument for Fault

The essence of the second general type of argument claiming to justify correlative rights and obligations of reparation is reliance on a localized conception of distributive justice. The argument for strict liability just considered took the basis of redistribution to be the voluntary performance of an action that subsequently caused harm. The argument of localized distributive justice discussed in this subpart regards the wrongfulness of action as the appropriate basis for redistribution. When the principle of  [*468]  reparation this argument is supposed to justify is translated into a tort regime, it generates a standard of liability based on fault.

The following passage from an article by Jules L. Coleman, in which he defends what he elsewhere calls a principle of "weak retributive justice," offers a good example of the distributive argument for fault: 
In torts we are dealing with activities and their accident costs. The question is always who should bear these costs. There is, in other words, a loss -- one which in a sense will not go away. Someone has to bear it. The common law approach to the problem of deciding upon the incidents of such a loss is almost always to restrict that decision to a choice between the injurer and his victim. If the choice is between a faultless victim -- that is, one whose conduct fails to contribute causally to the harm or one whose conduct, though it contributes to the occurrence, in every way complies with community ideals -- and a faulty injurer, one whose conduct not only contributes to the occurrence but falls below our ideals as well, the loss ought to be imposed on the party at fault. 71

 
Coleman's argument is distributive because it focuses primarily on the victim's loss, treating it as a burden that must be borne by someone. The form of distributive justice it envisages is localized because the group of potential loss-bearers is limited to the injurer and the victim. It should be noted, however, that Coleman regards this restriction as an imposition of the common law and not as a morally fundamental feature of the distributive problem. This is an important point, to which I shall return.

Coleman's assumption, then, is that a loss resulting from a harmful interaction must for extraneous reasons be distributed only between the parties to the interaction. He argues that the appropriate criterion of distribution looks to their relative moral shortcomings, so that if the choice is between an innocent victim and a faulty injurer, the loss should be placed on the injurer. In appropriate cases this would justify the recognition of correlative legal rights and obligations of reparation. An essentially similar argument was earlier relied upon by William Prosser to defend Andrews' view over that of Cardozo in Palsgraf v. Long Island Railroad Co. 72 Like Coleman, Prosser maintained that, given a choice between an innocent plaintiff and a defendant who had acted faultily, the loss should be placed on the latter. 73 Arguments of this kind also appear with great frequency in judgments in tort cases. 74

 [*469]  Coleman modestly limits the application of the distributive argument for fault to a well-established legal context, where existing institutions ensure that the loss cannot be placed on anyone but the injurer or the victim. But it seems at least initially plausible to suppose that the argument also justifies correlative rights and obligations of reparation between injurers and victims at a more fundamental moral level. I shall call this the immodest version of the distributive argument. There are a number of writers whose views on the moral justification of reparation should probably be understood in these terms, even though they do not state the argument in a particularly clear or perspicuous way. George Fletcher, in a well-known article, argues that an obligation of reparation arises where harm occurs as a result of the imposition of a nonreciprocal risk, that is, "a risk greater in degree and different in order from those created by the victim and imposed on the defendant." 75 Fletcher says that, depending on the circumstances, the obligation could be one of negligence or strict liability, but it is clear that he is not speaking of general strict liability based on causation alone, of the sort discussed and criticized in the preceding section. Nonreciprocal risk imposition defines what amounts to a standard of care, and hence Fletcher's theory avoids the indeterminacy to which general strict liability is subject. The reciprocity idea gives rise to other difficulties, 76 but the important point for our purposes is that it does nothing more than define a standard of conduct vis-a-vis others; it does not tell us why the appropriate remedial response to a violation of the standard which leads to injury is reparation by the injurer.

Fletcher's answer to that question comes in a short passage where he advances the following, Rawls-inspired principle of equal security: "We all have the right to the maximum amount of security compatible with a like security for everyone else." 77 Reparation for harm that occurs as the result of subjecting another person to an unfair (that is, nonreciprocal) degree of risk which violates this principle is justified because "compensation is a surrogate for the individual's right to the same security as enjoyed by others." 78 Fletcher's principle of equal security is best understood as a distributive principle. 79 The justification of an obligation of reparation then implicitly depends on a subsidiary distributive principle that reassigns losses so as to maintain the expected level of well-being guaranteed by the initial distribution of security. This is, I would argue, an implicit application of the distributive argument for fault, with "fault" defined as nonreciprocal risk imposition. Neil MacCormick's justification of rights and obligations of  [*470]  reparation, which also relies on a principle of equal security, is subject to a similar analysis, except that he accepts a more traditional understanding of fault. 80

The immodest version of the distributive argument encounters one real and one apparent difficulty. The apparent difficulty, which also represents a potential stumbling block for Coleman's more modest legal version of the argument, begins with the idea that the extent of the loss a person should be required to bear ought to be proportional to the degree of moral shortcoming that he or she has exhibited. But because the magnitude of accident losses is usually fortuitously determined and beyond human control, congruence between loss and fault will in general be impossible to attain. Prosser said that where proportionality of this kind could not be achieved, "there is no justice to be had." 81 Others have gone further, listing the supposed difficulty with proportionality as one count in an indictment of the entire fault system. 82 Prosser did not regard the problem as robbing the distributive argument of all its force, however, since even if an ideal proportionality was beyond reach it still seemed preferable to him that the faulty injurer rather than the innocent victim should bear the loss: "Essentially the choice is between an innocent plaintiff and a defendant who is admittedly at fault. If the loss is out of all proportion to the defendant's fault, it can be no less out of proportion to the plaintiff's innocence." 83

Does the general impossibility of attaining congruence between fault exhibited and loss borne represent a true difficulty for the distributive argument for fault? There are two different ways in which one might view a lack of proportionality as problematic. The first is premised on an analogy with a principle of retributive justice, since it is generally accepted that criminal punishment should, so far as possible, fit the crime (i.e., there should be a just proportionality between penalty and blameworthiness). But as a number of writers have pointed out, this would be an entirely inappropriate principle to apply in the context of tort law, since, in  [*471]  Coleman's words, "there is no criminal law analogue of the 'loss' that must be borne by someone." 84 A principle of proportionality is workable and just in the criminal sphere, where the penalty is both created and shaped by the court, but in the case of tort law the court is faced with a pre-existing loss that will have to be absorbed by somebody. 85 The second way in which lack of proportionality might be viewed as problematic looks to the claim that the concept of distributive justice itself demands congruence between loss borne and degree of shortcoming exhibited. But such a claim would be based on a misunderstanding. Distributive justice does not require that the share of a benefit or burden allocated to a person be proportional in any absolute sense to that person's need or merit or fault, but only that there be a uniform relative proportionality between the need or merit or fault of each of the members in the distributive group and the respective shares they receive. A demand for absolute proportionality would in general make no sense, since both the size of the distributive group and the amount of the benefit or burden to be distributed will ordinarily be determined by independent, contingent factors. If we assume with Prosser that the injurer in a particular case of harmful interaction was at fault and that the victim was entirely innocent, then the only proportionality that can rationally be sought is met by placing the entire loss, whatever its size, on the injurer alone. 86

This brings us to the true difficulty with the immodest version of the distributive argument for fault, which is that the localized nature of the distributive scheme is arbitrary and unjustified; there is no basis for limiting the group of potential loss-bearers to the injurer and the victim alone. One way to see the difficulty is by considering a solution proposed by H.L.A. Hart and Tony Honore to the supposed problem with proportionality just discussed. They suggest that we might be able to come closer to achieving a just proportionality in, for example, a negligence action if we were to take account of previous occasions on which the defendant had behaved negligently: 
I may drive at an excessive speed a hundred times before the one occasion on which my speeding causes harm. The justice of holding me liable, should the harm on that one occasion turn out to be extraordinarily grave, must be judged in light of the hundred other occasions on which, without deserving such luck, I have incurred no liability. 87

 
 [*472]  I leave aside the question of whether this could constitute an adequate general solution to the proportionality problem, since as we have seen the problem is not a real one. The argument is noteworthy nonetheless because it suggests that the localized scheme of distributive justice should be expanded so as to take account of the moral worth of actions that in no way causally contributed to the harm of which the victim is now complaining.

Such expansion seems inevitably to be called for by the following general principle, which can plausibly be taken to underlie the immodest version of the distributive argument for fault: social burdens should be distributed among persons in accordance with the degree to which they have exhibited moral deficiency or shortcoming. I shall accordingly speak of this as the deficiency principle, the complement of which would be a principle stating that social benefits should be distributed in accordance with degrees of moral virtue. Any attempt to restrict the moral shortcomings that should be taken into account in applying the deficiency principle to those manifested in actions that causally contributed to a particular loss seems arbitrary. Hart and Honore implicitly concede at least a limited version of this point. But if, as they suggest, the defendant's previous acts of negligence are to be taken into account, then why not those of the plaintiff? And why restrict the relevant class of moral shortcomings to negligent acts? 88 More importantly, once it is acknowledged that the actions to which the deficiency principle applies are not limited to those that causally contributed to a given loss, why should the group of persons to whom the loss might be distributed not include individuals who have exhibited moral shortcoming but played no part in causing that or perhaps any loss? And is there any reason not to extend the category of distributable losses to include those that are the result of, say, natural disaster and disease? The deficiency principle, taken by itself, would in response to each question recommend the most expansive possible approach. Realizing this enables us to uncover the chief difficulty with Fletcher's proposed justification for rights and obligations of reparation, which is that the logic of the implicit distributive argument offers no basis for reassigning losses only to the nonreciprocal risk-imposers who caused them. There is no reason why everyone who subjected others to nonreciprocal risks, thereby violating the principle of equal security, should not help to make good the losses that result from such violations, regardless of whether they actually caused such a loss themselves. 89

It is worth repeating that Coleman does not regard the argument from weak retributivism -- in my terminology, the distributive argument for  [*473]  fault -- as justifying correlative rights and obligations of reparation that possess an independent, pre-institutional moral status. As he puts it later in the same article, the argument is rather this: "Given the common law tradition of deciding upon the incidence of a loss between respective injurers and victims only, it would be morally justifiable to impose liability upon the injurer." 90 Coleman thinks that there are reasons independent of corrective justice for adopting a social institution like tort law, 91 but for him there is nothing inevitable about this; a more general social distribution scheme might for a variety of reasons be preferable. Such schemes are straightforwardly compatible with the annulment theory of corrective justice Coleman held until recently, since the theory does not insist that a victim's right to recover for a wrongful loss be correlative of an obligation on the part of his injurer to pay. The reasons for action to which corrective justice gives rise are, on this view, more diffuse, and the possible sources of compensation are more numerous.

I have elsewhere criticized the annulment theory on the grounds that it is, at bottom, really a conception of distributive justice. 92 But even though Coleman mischaracterized his earlier theory by calling it an account of corrective justice, he nonetheless followed out its distributive implications in a manner that was, for the most part, quite consistent. 93 In the absence of an argument justifying such a limitation, it is arbitrary for a distributively oriented theory to restrict the group of persons who should be considered as potential bearers of a loss to those who causally contributed to it, and Coleman clearly recognized this. Given his initial premises, he was correct to draw the distinctions he did between the grounds of recovery and the grounds of liability, and between the grounds and the modes of both recovery and liability. It was part of Coleman's general thesis that correlative rights and obligations of reparation cannot arise at anything other than a superficial institutional level -- that they cannot have a more fundamental, pre-institutional moral force -- and in this I think he was wrong. 94 But the positive argument that would show this has not yet been made.

Interestingly, one of the alternative arrangements to tort law that Coleman proposed can be regarded as the natural culmination of Hart and Honore's argument that the law should take account of a negligent injurer's previous acts of negligence. The system Coleman envisaged applies a  [*474]  nonlocalized conception of distributive justice to automobile accidents. 95 Drivers would accumulate demerit points corresponding to the degree of fault they had exhibited while driving. These would be assigned without regard to whether an accident had occurred. At the end of each year, the total sum of compensable losses caused by driving would be divided by the total number of demerits. The resulting figure would be multiplied by the number of demerits each driver had acquired over the year, and the driver would then be required to contribute that amount to the so-called "at-fault" pool. Persons who had suffered compensable losses would then claim directly from the pool. Since full compensation is assumed, proportionality is relative only, not absolute: 96 the burden that an individual would have to bear might not be directly proportional to the fault he or she had exhibited, but the fault/burden ratio would be the same for everyone. We saw earlier that proportionality for a localized scheme of distributive justice is likewise relative rather than absolute, but since in a general scheme the fault/burden ratio would be the same for everyone and not just for particular pairs of injurers and victims, it seems reasonable to think that the general scheme is fairer in this respect than the localized one. It is fairer, at least, so long as there is no reason to limit the scope of the distributive argument for fault to the parties who are causally implicated in harmful interactions. So far, we have discovered no such reason. 97

III. Reparation as an Incident of Voluntary Action

The third general type of argument for correlative rights and obligations of reparation is primarily concerned with the voluntary action of the injurer rather than, as in the case of the second type, with the loss of the victim. The general claim is that a principle of reparation is a necessary normative incident of, or at least is very closely linked to, the concept of voluntary action. I shall label this type of argument volitionist. As in the case of localized distributive justice, it comes in two main versions. According to the first, we are responsible in some fundamental, premoral sense for all the sufficiently proximate outcomes of our actions, whether these are good or bad, and hence for those outcomes that constitute losses to others. It is then claimed that this responsibility, perhaps together with certain other normative considerations, yields an obligation and corresponding  [*475]  right to repair. This is an argument for strict liability. According to the second version of the volitionist argument, the "normative structure" of action is such that we are responsible for those losses that result from voluntary action which is, in some sense to be specified, wrongful. This is an argument for fault liability. These two versions of the volitionist argument will be considered in turn. It will be convenient this time to begin with the argument for fault.

A. The Volitionist Argument for Fault

The leading exponent of the volitionist argument for fault is Ernest Weinrib. As Weinrib develops it, the argument is one aspect of a larger theory in which purely normative elements are conjoined, in a sophisticated and complex way, with considerations of institutional structure. The theory offers an account not only of reparation in a pure moral sense, but also of normativity as such, of the institution of tort law, and indeed of law generally. It is part of Weinrib's thesis that the various elements of this theory constitute an integrated and coherent whole, and this means that it is not always easy to isolate one particular aspect of it, such as the volitionist argument for fault, from the rest. It is not possible to consider the larger theory in its entirety here, but it will assist us in coming to grips with Weinrib's thought on reparation if we at least have before us a sketch of the more comprehensive theory.

1. Weinrib's Theory of Law

It is useful to think of Weinrib's theory of law as resting on three main theses, which I shall label rationalism, formalism, and integrationism. Rationalism is a broad concept in moral philosophy, but the often quite disparate theories to which the term can appropriately be applied share one important feature: they all contend that general moral judgments or propositions are capable of being shown to be true or false through the exercise of reason alone, without appeal to empirical knowlege, intuition, or undemonstrable first premises. Weinrib's version of rationalism, which is very much in the tradition of Kant and Hegel, maintains that normativity is, and can be seen by the operation of reason to be, a necessary incident of the concept of free choice or agency. Thus he says that "the meaning of normativity is precisely the determination of free choice in accordance with its own nature." 98 Practical reason, which provides norms that limit the exercise of free choice, does so by "making explicit the normativity implicit in purposiveness." 99 This leads Weinrib to characterize practical reason as "the necessity appropriate to freedom." 100 He states that "this meshing of freedom and necessity imparts normative force . . . to the entire idea of reason," 101 and that "the integration of free choice and practical  [*476]  reason contains all the normativity there is." 102 It is important to emphasize the conceptual nature of Weinrib's rationalism, which is brought out when he says, for example, that "it is a conceptual necessity that free purposiveness conform to its own nature as a causality of concepts." 103

Weinrib's formalism is a thesis about law considered as a social institution. The claim is that law manifests a specifically legal mode of rationality that is both normative in character and immanent to -- that is, represents the viewpoint already to be found within -- existing legal practice and doctrines. 104 The features of rationality, normativity and immanence are said to constitute the form of law, or, in a different terminology that Weinrib also uses, its essence. 105 Formalism is, for Weinrib, really a species of essentialism whose domain is institutions, or at least that particular institution we know as law. 106 He also regards certain specific areas or doctrines of law as themselves possessing "essential features," which are characteristics relied upon by what he refers to as legal experience to identify a given area of law as "a distinct mode of ordering." 107 It is clear, I think, that this is simply Weinrib's formalism in a scaled-down version, one which focuses on the essence of specific legal regimes rather than on that of law as a whole. In the case of tort law, two features are said to be singled out by legal experience as particularly salient. The first is "the bipolar procedure that links plaintiff and defendant," which in practical terms means that "the plaintiff sues the defendant and, if successful, is entitled to the defendant's performance of a remedial act." 108 This bilateral process is contrasted with a system of disbursements out of a central fund, which Weinrib says would not be an instance of tort law. The second feature of tort that legal experience identifies as particularly significant is the causation requirement.

The third main thesis that can be discerned within Weinrib's general theory of law I have labelled integrationism. This is not a term Weinrib himself employs, although related words like "integrate" and "integrative" abound in his writings. It is nonetheless possible to identify in Weinrib's work a distinct thesis that focuses on a notion of integration and that represents both a claim about the nature of law and a methodological postulate that Weinrib thinks legal theory (and, perhaps, social and moral theory generally) should embrace. The mark of integrationism is the search for a theoretical unity that arises from some kind of internal connectedness or coherence among a set of constituent elements. Thus law as a whole is  [*477]  said to be an "idea of reason," a notion which is in turn defined as "an articulated unity of parts in a conceptual whole." 109 In a similar vein Weinrib says of the three features that together make up the essence of law as he conceives it, namely rationality, immanence, and normativity, that they "are not disjointed attributes contingently combined, but mutually connected aspects of a single complex." 110

As one might expect with a theory that lays such emphasis on coherence and unity, Weinrib's integrationism both overlaps with his rationalism and formalism and combines them into a larger theory that is itself to be understood as an integrated whole. The key to this unity is the normativity said to be inherent in agency, since that same conception of normativity also lies at the core of the mode of rationality that constitutes, according to Weinrib, the essence of law. 111 As he says at one point, "the coherence of law as a Kantian idea of reason is grounded in the will's integration of free choice and practical reason." 112 Weinrib describes a progression of conceptual steps, each of which is supposed to be compelled by those preceding it, that he says leads from the existence of free will to the necessity of law. The major steps focus on free will, action, normativity, interaction with others, corrective justice, the need for a third-party arbiter, the justification of judicial decisions, and the authorization of public coercion. 113 This progression, which in accordance with Weinrib's integrationism is itself regarded as an "articulated unity," 114 shows that for him rationalism and formalism are really two sides of the same coin.

Again as one would expect, the integrationist dimension of Weinrib's general theory of law is reproduced in his theory of tort. We have already seen that Weinrib regards the law of tort, like law generally, as having an essence. As in the case of law generally, his integrationism leads him to say that the characteristic aspects of tort are "intelligible only through the integrated whole that they form as an ensemble." 115 Tort is thus said to be an "intrinsic ordering." This is to be contrasted with a conventional ordering, the constituent elements of which reflect "the contingencies of social practice and linguistic usage" rather than "an internal necessity of  [*478]  their own intelligible nature." 116 Economic analyses of tort are prime examples of theories that see tort law as a conventional ordering. As in the case of law generally, Weinrib's theory of tort involves a progression or sequence that begins with the normativity inherent in agency and ends with a particular institutional structure. He sometimes refers to this sequence as the unity or correlativity of doing and suffering. It has both a physical aspect, which involves the causation of harm by an action of the defendant, and a normative aspect, which focuses in the first instance on the wrongfulness of the defendant's action and in the second instance on his duty to compensate the plaintiff for the harm caused. 117 At each stage of the normative sequence there is a correlative right in the plaintiff: first a right that the defendant not act wrongfully towards her, and second, a right to be compensated by the defendant. The sequence culminates with the institutional structure of tort law: features of the "bipolar procedure" such as the two-party litigational form and the remedy of damages are "a reflection of what is normatively implied in the defendant's having injured the plaintiff. The causation of injury has a normative structure parallel to the procedural structure of tort law." 118

Weinrib's general theory of law is an impressive intellectual accomplishment that combines originality in conception with a rich and subtle elaboration of detail. Although the preceding sketch of the theory is too brief to do it justice, it should serve as a sufficient background for assessing Weinrib's solution to the problem of how correlative rights and obligations of reparation are to be justified. I should note here that in my opinion the theory gives rise, despite its sophistication and initial appeal, to a number of fundamental difficulties that it will not be possible to discuss in the present Article. Criticism of Weinrib's views will for the most part be limited to what he has to say about reparation at the level of pure principle. 119

2. Critique of Weinrib's Volitionist Argument for Fault

Let me begin by assuming that we all have duties not to act wrongfully towards other persons, and that correlative of these duties are rights not to be wronged. 120 For the moment I wish to leave unspecified the exact sense of the terms "wrongful" and "wrong." Weinrib maintains, in accordance  [*479]  with his Kantian brand of rationalism, that rights and duties of this kind are a necessary feature of the normativity inherent in agency. While I do not think that he succeeds in demonstrating this, since he asserts rather than demonstrates that normativity of the requisite kind is inherent in agency, the existence of some such rights and duties is hardly a controversial matter. Our concern is with the volitionist argument for fault liability, which in Weinrib's version claims that accompanying the primary duty not to act wrongfully, and similarly derivable from the normativity inherent in agency, is a secondary duty to compensate for harm that results from one's wrongful conduct. 121 This secondary duty, like the primary one, is associated with a correlative right, in this case a right to receive compensation for the harm suffered. The difficulty with Weinrib's argument is that he tells us very little about how to get from the primary duty to the secondary duty. He apparently assumes that a rationalist justification of the former will necessarily constitute a justification of the latter as well, but that is far from obvious. This gap in the argument arises, moreover, even if one thinks that Weinrib has succeeded in justifying the primary duty in rationalist terms.

At one point, during the course of a critique of Coleman's theory of tort, Weinrib suggests that if wrongful conduct has led to injury then the wrongdoer can be said to have made a wrongful gain equivalent to an amount that would be required to undo the injury; the victim's wrongful loss and the wrongdoer's wrongful gain are thus said to be "correlative to each other." 122 This sounds very much like an attempt to reduce reparation to an obligation of restoration along the lines discussed in Part I above. However that may be, Coleman is surely correct to respond that wrongful gain and wrongful loss would only be equivalent in the way suggested if the wrongdoer had a duty to compensate, and a duty to compensate must be distinguished from a duty not to act wrongfully. 123 Some further normative account is required to show how we get from the latter duty to the former, and this Weinrib does not provide.

In another paper Weinrib addresses in the following terms the relationship between the rights that correspond to the primary and secondary duties distinguished above: 
The bipolar procedure transforms the victim's right to be free from wrongful suffering at the actor's hand into a remedy whereby the actor undoes, so far as is possible, the injurious consequences. The bipolarity of doing and suffering matches the bipolarity of the procedure. The conceptual structure of negligence law enables the court to trace the progression from doing and suffering and to reverse it by making the wrongdoer compensate the victim of the wrong. 124

 
 [*480]  Note first that Weinrib cannot be taken here as trying to justify correlative rights and obligations of reparation by relying on the bipolar procedure that he says characterizes tort law, because the explanatory and justificatory sequence described earlier proceeds from agency and normativity, via the two levels of correlative rights and obligations, to institutional structure, and not in the opposite direction; normative principles are, for Weinrib, conceptually prior to institutional practice. He therefore cannot rely on his formalism, through an appeal to the essence of a social practice, to establish the normativity of a principle.

Perhaps the idea mentioned in the quoted passage of reversing the progression of doing and suffering is the key to justifying the secondary duty of reparation. That idea is obviously tied in with Weinrib's notion of the correlativity of doing and suffering, which in turn is a reflection in the tort context of his integrationism. This might be thought to suggest that a duty to compensate is the natural remedial response to the infliction of wrongful loss, for the reason that both the initial duty not to act wrongfully and the secondary duty to repair are constituent elements of a larger and presumably coherent whole, namely, the correlativity of doing and suffering. But it is far from clear what this talk of unity and coherence actually means or why it is important, 125 and in any event we once again encounter the difficulty that to assert such a connection is not the same as establishing it: Weinrib cannot hope to show that the duties in question cohere or form a unity of some kind without saying a great deal more about why. Even if we suppose that the notion of wrong which is said to derive from the normativity inherent in agency carries with it some requirement that the wrong itself be reversed or annulled, Weinrib nowhere demonstrates that any loss which results from the wrongful conduct must also be reversed. Perhaps punishment, which in the Kantian tradition is conceived as an annulment of wrong, is the sole appropriate remedial response. It is significant, I think, that elsewhere Weinrib speaks indifferently of tort law as annulling both "wrongful loss" and "wrongfulness" without any explanation of why annulment of the former is required and not just of the latter. 126

So far we have seen that none of Weinrib's rationalism, formalism, or integrationism assists him in bridging the gap between the primary duty not to act wrongfully and the secondary duty of reparation. The difficulty with Weinrib's volitionist argument for fault is not confined to the existence of this gap, however. Not only does he not show us how to get from the primary duty to the secondary duty, but the conception of normativity that he says is inherent in agency turns out to preclude that very move. To see why this is so it will be helpful to look at the arguments Weinrib makes in his paper "Right and Advantage in Private Law," 127 since that is where he  [*481]  offers his most detailed discussion of the relationship between the normativity supposedly inhering in agency and the correlative rights and obligations that are recognized in tort.

Weinrib draws a distinction in that paper between right and advantage. To understand what he means by right we must first begin with the Hegelian concept of abstract right, which refers to "the structure of relationships" that arises out of the normativity presupposed, according to Hegel, by freely willed activity. Abstract right is said to constitute "an intrinsically normative sphere, where the obligations incumbent on the actor reflect the nature of action." 128 This, it should be clear, is simply an Hegelian formulation of Weinrib's rationalism, according to which normativity is inherent in agency. Weinrib says that the demands of abstract right can be summed up in Hegel's phrase, "Be a person and respect others as persons." 129 A person, in Hegelian terminology, is an actor conceived from the standpoint of the abstract universality of free will. That standpoint is in turn derived from the capacity of the free will "to abstract from any particular object of choice." 130 Abstract right involves abstraction not just from particular choices, however, but also from what Weinrib calls "the particularity of advantage." 131 An advantage is defined as "something that contributes affirmatively to the contingent level of welfare that someone enjoys at a relevant time," whereas a disadvantage "is something that diminishes that level." 132

The quoted definitions suggest that advantages are to be understood instrumentally. On this view, their point would be the sustenance or advancement of independently-specifiable human interests that have intrinsic or ultimate value. Weinrib does sometimes conceive of advantages in an instrumental fashion, but more often he regards them as being themselves interests or aspects of human welfare or human well-being. 133 Thus he speaks of abstract right as abstracting not just from "the particularity of advantage," but also from "particular interests" 134 and from "all notions of well-being." 135 Similarly he refers, in his discussion of a rival view of rights,  [*482]  indifferently to "the maintenance of [a person's] welfare level" and "maintaining an initial level of advantages." 136 An instrumental conception of advantage, if it were the understanding Weinrib had principally in mind, would lend a certain plausibility to his claim that "whereas rights are morally relevant because of their conceptual connection with free will, advantages have no independent normative significance: it takes further argument to give them moral force." 137 But on the noninstrumental understanding that in fact predominates in Weinrib's account, acording to which an advantage is an aspect of human well-being, that claim becomes deeply suspect: human well-being itself is described as having no independent normative significance. Weinrib must make this implausible claim, however, because otherwise abstract right would not be able to abstract from all particularity of choice. There would be choices to act, namely those forwarding or intended to forward human well-being, to which it would have to accord a privileged status.

It will be helpful at this point to say something more about Weinrib's concept of right. He quotes Hegel as defining right (and, it would seem, a right) as "an existent of any sort embodying the free will." 138 Free will is embodied in one's own body and can be embodied in a material thing, which then becomes property. 139 We have already seen that abstract right requires respect for persons, or, as Weinrib sometimes puts it, respect for personality. Quoting Hegel once again, Weinrib says that "personality essentially involves the capacity for rights." 140 He adds that "in abstract right the significance of particular rights consists solely in their being actualizations of this capacity and not in their contribution to the satisfaction of the rights holders' particular interests." 141 This, together with his statement that "every right crystallizes certain advantages," 142 helps to clarify Wenirib's meaning when he says that advantages have no independent normative significance. Such normative significance as they do have clearly derives from the embodiment of personality and not from the fact that they are aspects of (or, on the instrumental understanding, that they help to forward) human well-being. Given his claim that "the integration of free choice and practical reason contains all the normativity there is," 143 Weinrib would appear to be committed to the position that human well-being as such has no normative or moral significance at all.

 [*483]  Weinrib asserts that rights and advantages are "constituents of different justificatory paradigms" and that "it is impossible for both to be basic elements in a single coherent theory." 144 By issuing this decree of divorce between the concept of right and human interests Weinrib is denying, among other things, the possibility of both an interest theory of rights and Rawls' thin theory of the good. On their face the quoted statements might seem to leave room for a normative account of well-being that was completely independent of abstract right but in light of the conclusion of the preceding paragraph this is doubtful. In any event our concern is with reparation, which for Weinrib falls entirely within the realm of abstract right. Since it is clear that abstract right denies any normative significance to well-being as such, Weinrib's theory of tort is confronted by two problems. First, it is by no means clear that it can characterize wrongful conduct in the way that tort law does, namely, as action that the actor knows or could reasonably foresee will interfere with a protected interest. Second and relatedly, Weinrib seems to have no way to justify a secondary duty of repair (together with its correlative right), as opposed to a primary duty not to engage in wrongful conduct. 145 This is because rights and obligations of reparation are concerned with loss, and a loss in this context is nothing more than detrimental interference with well-being. It is, in Weinrib's terms, a disadvantaging. In order to avoid ambiguity in the discussion of these two problems that follows, I use the term "interest" to refer to advantages in the noninstrumental sense: an interest is to be understood as an aspect of human well-being. Protected interests in tort are clearly a sub-set of interests thus defined.

Let us consider the first problem in greater detail. I earlier left unspecified the content of the primary duty not to engage in wrongful conduct, and we are now in a position to see that there are two main possibilities here. Concentrating for the moment on intentional wrongdoing, the first is that wrongful conduct of this kind involves deliberate interference with the embodiments of another's personality, that is, it consists of an act that deliberately causes harm to person or property. The second possibility is that intentional wrongdoing is characterized by deliberate interference with personality tout court, that is, it is characterized by deliberate interference with the capacity for rights rather than with rights as such. Weinrib usually, but not always, 146 employs something like the first formulation of intentional wrongdoing, 147 which in essence is also the understanding adopted by tort law. The difficulty, however, is that his general characterization of both abstract right and particular rights requires the second formulation. The concern of abstract right is with disrespect for personality rather than with interference with particular rights or the interests they represent.

 [*484]  It is of course possible to show disrespect for personality by interfering with its embodiment, and Weinrib gives a good example of this: "My murdering [someone] would be inconsistent with his having the capacity for rights, since I would be treating him as a thing that was available to my purposes and, thus, not an embodiment of free will." 148 The point is, however, that disrespect for personality -- in Weinrib's alternative formulation, disrespect for the capacity for rights -- must, given his premises, be regarded as the essential characteristic of intentional wrongdoing. Moreover, one can show disrespect for the personality of another, thereby presumably wronging him, without interfering with an embodiment of his personality. Consider the following scenarios. I attempt to murder someone but fail. Surely I have shown just as much disrespect for his personality as if I had succeeded. Or suppose I use a sleeping person as a repository for my coat because there is nowhere else to put it except on the dirty floor. Using someone as a coat rack is clearly "treating him as a thing . . . available to my purposes and, thus, not as an embodiment of free will." But, we may suppose, my using him as a coat rack did not harm him in any way; it did not actually interfere with an embodiment of his free will. We may further suppose that it was clear to me from the outset that using him as a coat rack would do him no harm, so that in this case the disrespect that I have shown for personality is not even a function of the possibility of causing harm.

Weinrib might want to reply at this point that tort law is only concerned with insults to personality that also happen to interfere with an embodiment of personality. The difficulty with this response is that interference with an embodiment of personality must be understood as involving harm to person or property. But harm is just the degradation of an interest, and as we have already seen interests as such have no significance for abstract right. Abstract right does not, in other words, possess the resources that would permit it to attribute a special status to those intentional insults to personality with which tort law is particularly concerned, namely, the ones that cause harm. This has consequences at the remedial stage that we will come to in a moment. First, though, we must say something about unintentional wrongdoing, the characterization of which gives rise to even greater difficulties for abstract right.

Weinrib wants to say that abstract right regards contravention of the objective standard of care in negligence law as wrongful conduct. His argument in support of this conclusion focuses on the concept of risk: 
Risk allows us to conceive of the doing and suffering of an accidental harm as the maturation of a single process, since it captures the potential for harm present in the defendant's act that materializes in the plaintiff's injury. Risk is the relational concept that connects the active and passive aspect of injurious conduct, so that what the defendant did and what the plaintiff suffered are not regarded as two discrete happenings. Thus risk supplies for unintended injuries the unity that, for intentional harms, is found in the identity of the consequence that the actor desires and the  [*485]  victim suffers. 149

 
Here we see Weinrib's integrationism coming to the fore, since risk is said to be the glue that binds doing and suffering into a normatively significant unity. The difficulty that we discovered with Weinrib's characterization of intentional wrongdoing emerges even more strongly here, however, because the risk with which tort law is concerned is risk of harm, in the sense of interference with an interest, and abstract right has no place for harm thus understood. Weinrib makes the problem explicit by stating that "abstract right deals with the doing and suffering of a harm," 150 since this is, surely, precisely what abstract right does not deal with. No doubt abstract right can treat the advertent imposition of risk as wrongful, since such conduct can plausibly be construed as showing disrespect for personality. But there does not seem to be any way for the theory to characterize inadvertent risk imposition as a wrong. In order to do so it would have to ascribe a normative significance to harm as such -- in Weinrib's terms, to disadvantaging -- when it lacks the resources to do this. Thus the supposed unity of doing and suffering that Weinrib's integrationism leads him to perceive is not only unsupported by argument, it also forces together elements that by his own lights are incompatible.

There is a further problem with Weinrib's characterization of unintentional wrongdoing, and it is worth a slight digression to make clear what this is. Weinrib describes risk as though it were a physical process or potentiality that necessarily accompanies action. By conceiving of risk in this way, he is able to bring his understanding of unintentional wrongdoing into line with the assumption of negligence law that only the creation of a substantial risk, and not any risk whatsoever, gives rise to potential liability. He argues that abstract right regards as wrongful only the creation of risks above a certain level because "risk is an unavoidable concomitant of all action." A duty not to impose risk at all would deny the possibility of action, and, hence, "would be incompatible with the status of the actor as a person." 151 The difficulty here is that risk is not properly regarded, from the point of view of abstract right, as an unavoidable or necessary concomitant of all action. Risk, or at least the risk with which tort law is concerned, is not a physical process of some sort, or a physical potentiality capable of "maturation," that can be regarded as inextricably bound up with the concept of action. It is an epistemic notion that has to do with the actor's state of knowledge about the processes he is setting in motion. An action can only be said to be risky where (1) the actor is ignorant about whether those processes will in fact result in harm, and (2) she knows enough to be able to say that there is a certain probability of harm. 152 The concept of probability in question here is, it should be noted, itself  [*486]  epistemic. 153

It is true that there is a sense in which action is always accompanied by risk, since we are always ignorant about many of the future consequences of our actions while at the same time knowing that there is a probability they will result in harm. But this is simply a particular fact about human beings, not a fact about action or agency as such, and there is no obvious reason why abstract right should be more concerned about this brand of particularity than it is about any other. An omnisicient being could act without imposing risk. Thus even if abstract right were entitled to characterize as wrongdoing the imposition of risk understood by reference to an attributed level of knowledge rather than the actor's actual epistemic state -- this being what adoption of the objective negligence standard effectively amounts to -- it is not clear why it should attribute to the actor the level of knowledge that would be possessed by a reasonable person rather than, say, omniscience. In legal terms, the result of attributing omniscience would be, of course, a form of strict liability, a consequence hardly likely to appeal to Weinrib.

We have already seen, however, that abstract right is only entitled to characterize wrongdoing in subjective, not objective, terms. Intentional wrongdoing must be understood as involving an intentional display of disrespect for personality, not an intentional harming of interests, although it is true that intentionally harming an interest will ordinarily manifest disrespect for personality as well. 154 Abstract right can also treat the advertent imposition of risk to an interest as wrongful, since this "manifests a failure to respect the persons who are within the ambit of the risk's effects." 155 But, contrary to what Weinrib suggests, the same cannot be said of inadvertent risk imposition. This brings us by a natural route to the second of the two main problems that confront Weinrib's theory of torts. The difficulty here concerns the remedy to wrongdoing that abstract right can appropriately require, my claim being that it is not able to justify an obligation (and correlative right) of reparation.

We can begin our consideration of this question by noting the following two points. First, as was just remarked, the wrongdoing that concerns abstract right is subjective in nature. It involves either intentionally wrongful conduct or culpable negligence, both of which presuppose blameworthy states of mind. Second, conduct can be wrongful in the sense required by abstract right whether or not it causes harm to anybody's interests. In the case of intentional wrongdoing this was illustrated by the attempted murder and coat rack examples above. So far as advertent negligence is concerned, knowingly subjecting another to a substantial risk of harm obviously displays a lack of respect for the other as a person even  [*487]  if the risk does not subsequently materialize. Let me assume for the sake of argument that abstract right demands that wrongfulness must be annulled. If that is so, surely it must be annulled whether or not it has resulted in harm. Since the wrongdoing that falls within the ambit of abstract right is always culpable wrongdoing, the obvious remedy is criminal punishment. This point can be illustrated by reference to Weinrib's discussion of "takings." He says that a taker "attempts to assert dominion over something that is already a physical or proprietary embodiment of another's personality," thereby signalling through his actions "that he does not recognize the categorical difference between persons, who have a capacity for rights, and other entities, which lack free will and therefore do not have that capacity." 156 Weinrib states that "takings are wrongs that constitute general denials of the validity of right," 157 but surely this denial is present whether or not the attempt to assert dominion succeeds. According to Weinrib, liability in tort "juridically nullifies" the denial, but liability presupposes harm 158 and so cannot serve as a general mode of nullification. The remedy that naturally serves this role is, again, criminal punishment.

It remains to consider whether the remedy of damages is an appropriate method of nullification in those cases where wrongdoing does result in harm. The answer to this question must be no, for the reason already considered. Compensation is for loss and loss must be understood as a degradation of an interest, but interests as such are accorded no normative significance by abstract right. Weinrib does say that "because every right crystallizes certain advantages, the infringement of a right is usually remedied by a damage award that quantifies the value of the advantages of which the victim has been deprived." 159 But this is to treat advantages as having normative significance of their own, as interests, that is, as aspects of human well-being. Such treatment is irreconcilable with Weinrib's position that "in abstract right the significance of particular rights consists solely in their being actualizations of [the] capacity [for rights] and not in their contribution to the satisfaction of the rights holders' particular interests." 160

In the following passage Weinrib himself makes clear the implications for tort law of the austereness of abstract right: 
Abstract right . . . does not guarantee any specific condition or value to what is owned. Abstract right accounts for ownership in a way that makes the condition of particular owned things irrelevant. It thus reflects the obvious truth that a particular  [*488]  owned thing remains the property of its owner whatever its condition or value. Although your injuring something that is mine may reduce the satisfaction I derive from it, what was injured nevertheless remains mine. Therein lies the abstractness of abstract right. 161

 
Weinrib is concerned to show here that damage by itself cannot constitute a violation of abstract right, so that strict liability could not be the appropriate tort regime. In fact, the quoted passage shows a great deal more than that. It demonstrates that damage or loss is simply irrelevant to abstract right, which means that abstract right cannot justify any general tort regime, whether based on strict liability or fault. At most it can accommodate a very limited obligation of restoration: a person who appropriates the property of another and then intentionally damages it would be required to do no more than return the bits to the owner. Abstract right by its very nature rules out the possibility of correlative rights and obligations of reparation. It is, at best, a theory of criminal liability. Contrary to what Weinrib suggests, abstract right makes tort law impossible.

B. The Volitionist Argument for Strict Liability

Weinrib's volitionist argument for fault looks upon rights and obligations of reparation as flowing from normative principles that govern interaction, where the requisite normativity is said to be implicit in the concept of agency. The norms of interaction are concerned not with loss or harm as such but with conduct that is wrongful because it displays disrespect for personality in the abstract, Hegelian sense. If the argument were successful it would justify a fault-based principle of reparation. The argument fails, however, because the concept of normativity on which it rests abstracts from all human interests and so cannot attribute normative significance to the notion of loss that must constitute the primary concern of any principle of reparation. It might be thought that the appropriate response to this failure would be to direct our attention from the outset to the fact of loss, considered as a normatively significant interference with human well-being, and to ask what ought to be done about it. Taking this tack would, without more, lead to one or the other of the two distributive arguments already considered.

There is another possible response to the failure of Weinrib's argument, however, which is to retain his volitionism but to reject the understanding of normativity that he thinks flows from agency. The idea would be to establish a normative connection, of a sort that Weinrib's theory forbids, between the concept of action and the losses action can cause. But a connection between action and loss only would be too narrow. A more promising avenue is to argue, first, that action necessarily gives rise to responsibility, in some sense to be specifed, for all the outcomes that result from our choices to act, or at least for all the outcomes that meet some general test of proximity; and second, that this responsibility justifies,  [*489]  perhaps together with certain other considerations, an obligation of reparation when the outcome of an action is a loss for someone else. A volitionist argument along these lines is suggested by the theory of individual responsibility advanced by Tony Honore, 162 although as we shall see it is not quite the argument that Honore himself accepts.

Honore's theory concerns a normative but premoral notion that he refers to as outcome-responsibility. Because this notion focuses in the first instance on the outcomes of action rather than on action as such or on interactions between persons, the associated conception of normativity is different from Weinrib's. It is natural to think that a conception of responsibility that emphasizes outcomes might generate a principle of strict liability rather than one of fault. Honore's own argument does purport to justify strict liability, although he thinks that outcome-responsibility can also show why it is legitimate for the law to rely in some circumstances on the objective negligence standard. In his view, that standard has more in common with strict liability than it does with a true fault principle, since it holds persons responsible for shortcomings they are subjectively incapable of avoiding. Given that Honore thinks objective negligence and strict liability stand or fall together and both must be distinguished from a true principle of fault, it will be convenient to ignore negligence and refer to his argument as one supporting strict liability.

Honore begins his characterization of outcome-responsibility with the analogy of a wager. Choosing, he says, is inescapably like betting. In making a decision to do X rather than Y "we are choosing to put our money on X and its outcome rather than Y and its outcome." 163 In opting to attempt a U-turn rather than to go on to the next roundabout, for example, we implicitly bet that we will arrive at our destination more quickly. But we might lose the bet by getting involved in an accident, say, or by discovering that going via the roundabout would have been the faster route after all. It is important to see that Honore is not at this point saying anything about reparation. The payoffs of the bet are credits and discredits in a social ledger of some kind. Thus in the U-turn example we receive a credit if we are successful in getting to our destination quicker, but "if we botch it, have an accident, or mistake the route, that is chalked up against us." 164 Moreover, we receive the discredit even if the botch or miscalculation was not our fault. Choosing to act is said to be an implicit bet on outcomes, where neither the stakes nor the social payoffs are necessarily known in advance.

This understanding of outcome-responsibility as a social system for allocating credits and discredits is supplemented later in Honore's article by a strikingly different one. According to the second understanding, "we could not dispense with outcome-responsibility without ceasing to be persons." 165 The idea here seems to be that personhood is inescapably bound up with agency, but we cannot be agents capable of choice without  [*490]  accepting, in some very fundamental sense, that we are responsible for the consequences of the choices we make. Honore also says that "outcome-responsibility is crucial to our identity as persons," 166 by which he apparently means that it is crucial to our identities as individuals and not just to our having the general status of personhood. The sense of this claim is perhaps clarified by his later statement that "it is outcomes that in the long run make us what we are." 167 Outcomes of actions are what make up a life, and so long as we have a minimum range of options to choose from -- that is, a minimum degree of autonomy -- we are, in making our choices, part authors of our own lives. 168 Our sense of identity as individuals is at least partially determined by the lives we create for ourselves through the choices we make.

There is an obvious tension between Honore's two characterizations of outcome-responsibility. The second of the two understandings, according to which outcome-responsibility figures prominently in our sense of our own agency, is a fundamentally important notion for both the theory of action and moral theory. It is, however, very difficult to reconcile with the first, social understanding. Honore says that "by allocating credit for the good outcomes of actions and discredit for bad ones, society imposes outcome-responsibility." 169 But the idea that outcome-responsibility is socially imposed is completely at odds with the claim of the second understanding that we could not dispense with the concept without ceasing to be persons, since if this latter statement is true then outcome-responsibility is simply not the sort of thing that society could impose; it is too fundamental a notion to be the subject of a social decision.

Moreover, if a social decision to impose outcome-responsibility were possible we would be entitled to ask whether it was justified, and this would require us to offer moral reasons for accepting a supposedly premoral concept. Honore does say that the "system" of outcome-responsibility must be shown to be fair: "[It] must in its operation be impartial, reciprocal and over a period beneficial." 170 He asserts that these conditions are in fact met for most people, but whether that is so or not is irrelevant to the status of outcome-responsibility understood in the second sense; when Honore asks whether outcome-responsibility is fair he is posing a question that cannot, so far as the second understanding is concerned, properly be posed. He makes a similar point himself when he says that outcome-responsibility could not rest on a social contract because "the exchange would not even, properly speaking, be in our interest; for to be responsible is part of what it means to be a person and hence to have interests." 171 If the idea of exchange presupposes interests then so do general notions of fairness, which means that the principles of fairness to which Honore is appealing are rooted in the very notion of outcome-responsibility that he says they  [*491]  justify. Honore's insistence that outcome-responsibility is a social system also undermines his apparent acceptance of the idea -- a very natural one on the second, agency-oriented understanding -- that outcome-responsibility applies universally to all actions and all outcomes, since there is no reason to think that social conventions assigning credit and discredit will be comprehensive in this way. It is highly implausible to suppose that "society" cares one way or another about whether I get to my destination more quickly by making a U-turn, or that my achievement in doing so will be rewarded by a credit of some sort in a social ledger.

It is the social rather than the agency-oriented understanding of outcome-responsibility that Honore employs in his proposed justification for strict liability. But before we examine that argument it is important to consider how, on either understanding, outcome-responsibility can be said to be a normative notion. To begin with the social understanding, it is clearly normative in the straightforward sense that most people have a reason to act in ways that will invite credit and discourage discredit. This reason may be outweighed by other reasons, of course, and acting on it may be difficult where specific outcomes cannot be foreseen, but the normativity of social outcome-responsibility is not problematic. The agency-oriented understanding of outcome-responsibility does not, by contrast, necessarily give rise to prudential reasons for action, nor is it obvious that it always creates reasons of some other kind. Honore characterizes the principles permitting the ascription of personhood and outcome-responsibility as normative, 172 but he appears to have in mind a broad sense of the term in which there need not be any necessary or immediate implications for reasons for action. To help understand what is at stake here, let us return once again to the U-turn example.

What, on the agency-oriented understanding of outcome-responsibility, might be the analogues of the credit we supposedly receive from society for a successful outcome, and the discredit that is said to be chalked up against us for an unsuccessful one? The obvious answer is that the appropriate analogues are the benefits and harms to ourselves that might accrue from the various possible outcomes. Getting to our destination more quickly is a benefit and a desirable outcome, while losing our way or becoming involved in an accident, even if there is no question of paying compensation to anyone else, is a harm and an undesirable outcome. But outcome-responsibility cannot just be a matter of having to live with and accept the beneficial or harmful consequences of our own actions for ourselves, since that would be to view persons primarily as experiencing rather than as choosing beings. Personhood necessarily involves a capacity for self-reflective cognition, experience and choice, but outcome-responsibility is primarily concerned with the normative significance of choice. We must equally live with and accept the consequences of disease and other events that just happen to us, quite apart from anything we choose to do. A more promising suggestion is that agency-oriented outcome-responsibility involves the retrospective evaluation of action in light of its outcome, initially by the agent herself, and that it is in this sense  [*492]  that it can be said to be a normative notion. Thus we view the making of the U-turn as somehow a better action if the outcome was beneficial to us, because we arrived sooner at our destination, and as somehow a worse one if it was harmful to us, because we lost our way or got into an accident.

This idea that actions are subject to evaluation in light of their outcomes, after the fact and independently of how the agent could have antecedently expected things to turn out, has been explored with great sensitivity and insight by Bernard Williams and Thomas Nagel under the rubric of moral luck. 173 Although Honore does not emphasize it, the idea should, I think, be regarded as the core of his agency-oriented understanding of outcome-responsibility. So far we have been talking of the ex post evaluation of action in a premoral sense because it is the agent herself who is doing the evaluating, in light of her own ends and goals, so that the action is spoken of as better or worse from a self-interested or prudential perspective. But the ex post evaluation is not just a retrospective projection of the agent's assessment of the value of the outcome (to her) onto the action that produced it. It is a matter of determining whether the agent can justify (to herself) the doing of the action, or whether she must judge it to have been wrong (again, from her own perspective). There is perhaps something mysterious about this kind of evaluation of action, but it seems to be an unavoidable feature of agency as we experience it. As agents we do not simply act and produce effects. We possess an awareness of ourselves as being capable of making a difference in the world, and this awareness colors our sense of our own agency in the following ways. First, we identify with both the actions we perform and the outcomes they produce: they are our actions and outcomes. 174 Second, we act with the knowledge that the justification or otherwise of our actions depends at least partly on the particular way that they make a difference in the world: the worth of our actions is, in our own eyes, partly determined by the actual outcomes they produce. If the action turns out badly, we at least sometimes regret not just the outcome but the action itself, a feeling Williams refers to as agent-regret. 175

Honore's agency-oriented understanding of outcome-responsibility begins, like Williams' notion of agent-regret, with the agent's attitudes towards her own past actions; this is why the label "premoral" is appropriate. The social understanding, on the other hand, begins with relationships between persons, and hence has moral overtones from the outset. Honore clearly has in mind the social understanding when he states that outcome-reponsibility involves making a bet with other members of our community  [*493]  and only indirectly with ourselves. 176 If the wager analogy is to be applied to agency-oriented outcome-responsibility, it is best regarded as a bet that at least initially we make with ourselves and not with the community. But agency-oriented outcome-responsibility does have a bearing on moral questions that arise from interaction between persons. As Honore notes, if I run you over, then even if the accident was entirely your fault it is appropriate that I feel and express regret for the harm I caused you, even though I cannot be blamed for what happened. 177 What Honore does not mention is that I also have reason to express regret, in Williams' sense of agent-regret, for the action and not just for the harm to which it led. This is an instance of outcome-responsibility in which an agent's evaluation of, and attitude towards, her own action is affected by its outcome for another. What is of particular interest here is that when our actions harm others, outcome-responsibility is normative not just in the weak evaluative sense we have so far been considering, but also in the more robust sense of affecting our reasons for action. Honore implicitly takes note of this when he says of the running-over example that simply by virtue of having hurt you I have a responsibility to take certain steps. I have, in other words, reasons for action that a bystander would not have, including in particular an obligation to obtain assistance.

Agency-oriented outcome-responsibility for harm to others affects not just our attitude as individuals to those of our actions that produced the harm. It also affects the attitudes that others do and should take to those actions, and, perhaps, their reasons for action vis-a-vis the actor as well. At this stage we are talking about moral attitudes and reasons for action, both as regards ourselves and others. It is important to emphasize, however, that outcome-responsibility need not be regarded as the source, or at least the sole source, of the moral elements in play here. Actions can be morally evaluated apart from their outcomes, and harm to human well-being is a morally bad thing about which other people have reason to be concerned, even where the harm is not the result of anyone's action. There is thus no reason to assume that the normativity associated with outcome-responsibility is, in Weinrib's phrase, all the normativity there is, or that other manifestations of normativity are somehow derivable from it. Nor does Honore assume either of these things. He clearly regards outcome-responsibility as just one factor that enters, together with others like fault and desert, into overall moral judgments. 178

The interplay between outcome-responsibility and other moral factors comes out in Honore's discussion of examples taken from criminal law. Murder, he rightly notes, is judged more severely than attempted murder, and causing death by dangerous driving is appropriately regarded as more serious than dangerous driving alone. 179 The bad outcome leads us to judge the action that caused it, together with the agent, more harshly than we otherwise would (although, as was mentioned earlier, Honore is not clear  [*494]  that it is the action itself which is being judged and not just the outcome). Outcomes matter in the moral assessment of action, and this is so quite apart from any consequentialist considerations. But other things matter too, as these examples make clear. An act of attempted murder or of dangerous driving is wrong even if, as it turns out, there is no bad outcome. Outcome-responsibility is only one element among others in the moral universe.

This brings us, finally, to the question of the relationship between outcome-responsibility and moral responsibility to compensate for harm caused. One approach to the justification of rights and obligations of reparation that might seem promising is an argument that outcome-responsibility in the agency-oriented sense automatically creates a moral obligation to compensate, thereby justifying a regime of strict liability. This is a genuine volitionist argument: it claims that the normativity which, according to the agency-oriented understanding, is necessarily associated with agency, always gives an injurer a reason to compensate for any harm he causes. The promise of the argument soon withers, however, because it encounters at least two difficulties. The first is that even if outcome-responsibility affects our moral reasons for action, there are no grounds for thinking that it always gives rise by itself to a generally-applicable reason to pay compensation (let alone an obligation to do so), as opposed to a reason to call an ambulance, express regret, and so on. Further argument would be needed to establish this, and it is by no means clear what it would be.

To understand the second difficulty we begin with the observation that in cases in which A causes harm to B, both A and B may turn out to be outcome-responsible for the injury, and not just A alone. We have not yet discussed the relationship between outcome-responsibility and causation, but even if, as seems plausible, it turns out that the two are not simply to be equated, the likelihood that both parties to a harmful interaction will be outcome-responsible for the harm one of them suffers seems high enough to ensure that the pervasive indeterminacy of causation-based systems of general strict liability would also be encountered if outcome-responsibility were made the sole basis of reparation. Consider yet again the U-turn example. Honore seems to assume that the U-turner is outcome-responsible for the consequences of a collision with another car in exactly the same way she would be for getting to her destination more quickly, but this is not so. In the collision case the accident is, for the reasons discussed in Part II, the consequence of two choices to act, not one; the actions of both parties would be causes of the harm. Moreover, both made bets in Honore's sense, and even if we ultimately wish to say that causation is not equivalent to outcome-responsibility, it seems almost inevitable that in a case like this each of the parties would be regarded as outcome-responsible for the consequences. 180

Although Honore does not expressly recognize that both parties to a harmful interaction could be outcome-responsible for the injury just one of them suffers, his proposed justification for strict liability avoids the indeterminacy problem anyway because it relies on a different argument from  [*495]  the one just considered. For one thing, it employs the social rather than the agency-oriented understanding of outcome-responsibility. Civil liability, which is effectively seen as a form of penalty, is said to specify "an extra sanction to be imposed on a person who has anyhow lost a bet and who will in consequence incur discredit." 181 Honore does not say that being outcome-responsible for a harmful outcome should automatically lead to the imposition of a legal duty to compensate because "an extra element is needed to ground the legal sanction." 182 This could be fault in a moral sense, but in the case of strict liability it is the existence of a special risk of harm that the type of conduct in question creates. Honore thus conceives of strict liability in terms of limited doctrines like the rule in Rylands v. Fletcher 183 or the American rule for ultrahazardous activities rather than as a general form of liability based simply on causing harm (or being outcome-responsible for it). This prevents his argument from falling prey to the indeterminacy objection, but he does not insist on an "extra element" in the justification of strict liability in order to avoid indeterminacy. Presumably he does not move directly from outcome-responsibility to general strict liability because civil liability is viewed, on the social understanding of outcome-responsibility, as a legally imposed discredit, and as a matter of wise policy a legal system should no doubt be wary of subjecting people to essentially penal sanctions on too indiscriminate a basis.

Although it is not vulnerable to the indeterminacy objection, Honore's argument does not succeed on its own terms. The special risk that serves as the extra element in the justification of strict liability brings into play, he says, "a consequentialist argument for avoiding serious harm [that] reinforces the non-consequentialist arguments for imposing outcome responsibility." 184 But it is not clear what this consequentialist argument is. Honore states that the strict liability which attaches to specially risky conduct is not fault-based because there may be good reasons, social or otherwise, for the actor to engage in such conduct. In that case, however, it is far from clear what consequentialist benefits could accrue from imposing legal liability once injury occurs. Deterrence does not seem to be at issue, and so far as compensatory or loss-spreading goals are concerned, there appears to be no basis for treating the "special risk" situation any differently from other types of cases. Moreover, even if we ignore this gap in Honore's argument it still could not justify correlative rights and obligations of reparation, since from the outset civil liability is viewed as a sanction and not as a basis for compensating for harm. If it succeeded the argument would simply show tort law to be a species of criminal law.

The failure of Honore's own argument for strict liability does not mean that outcome-responsibility is irrelevant to the justification of rights and obligations of reparation. It is true that the social understanding of outcome-responsibility is unlikely to be of much assistance here, since an obligation of reparation does not seem reducible to a mere discredit in  [*496]  Honore's sense, but the general normative connection that the agency-oriented understanding establishes between the evaluation of actions on the basis of outcomes, and reasons for further action in light of those outcomes, is worth more detailed scrutiny. We have seen that a genuine volitionist argument that tried to move directly from the agency-oriented understanding of outcome-responsibility to general strict liability would not succeed, but there may well be a role for that understanding in a different argument. The possibility of developing such an argument is discussed in the following section, where unless otherwise indicated the term "outcome-responsibility" will be used to refer to the agency-oriented understanding only.

IV. Outcome-Responsibility as the Basis of a Distributive Argument

A. The Basic Argument

Let us take stock of the discussion so far. The Aristotelian argument, which tries to found a principle of reparation on a limited restitutionary feature of the concept of property, namely the obligation of restoration, fails because it is not comprehensive enough; at most it can justify very limited rights relating to conversion. The distributive argument for general strict liability fails because the principle that costs should be internalized to the actor or activity that caused them is pervasively indeterminate: a sound understanding of causation shows us that actions of both (or all) the parties to a harmful interaction are in general causes of the harm. The distributive argument for fault does not succeed in justifying correlative rights and obligations of reparation because there is no reason for limiting the distributive considerations it invokes to a localized scheme of victim and injurer alone: a more generalized distributive mechanism seems to be called for in which victim and injurer have no special status vis-a-vis one another. Weinrib's volitionist argument for fault, interesting and sophisticated as it is, cannot justify a principle of reparation because the concept of normativity it yields is unable to ascribe any moral significance to the notion of a loss. Honore's argument for strict liability goes wrong because it does not back up its consequentialist claims, and also because it is the wrong sort of argument: in relying on the social understanding of outcome-responsibility and treating civil liability as a form of penal sanction, it does not even try to justify rights and obligations of reparation. Finally, a genuine volitionist argument for strict liability, which takes as its starting point the agency-oriented understanding of outcome-responsibility, fails for two reasons. First, even though outcome-responsibility affects our reasons for action, there are no grounds for thinking that it always and necessarily gives rise by itself to a reason to compensate. Second, the resulting scheme of strict liability would be subject to pervasive indeterminacy because both parties to a harmful interaction can be, and, there is reason to think, often are, outcome-responsible for the loss.

Two of the arguments we examined turned out to be not so much wrong as inconclusive or incomplete. The distributive argument for fault focuses on the normative significance of loss, regarded as a degradation of some aspect of human well-being, and asks if there are moral reasons for redistributing the interference with interests that the loss represents to  [*497]  anyone else. If the argument were to justify correlative rights and obligations of reparation some basis would have to be found for restricting the group of potential loss-bearers to victim and injurer, and nothing plausible was suggested by the versions of the argument we examined. But we discovered no reason to conclude that a basis for restricting the distributive group does not exist. We also have no grounds for thinking that the distributive approach should not be applied to losses caused by human agency.

The other argument that turned out to be more inconclusive than wrong is the volitionist argument for strict liability that was suggested by Honore's discussion of outcome-responsibility. The argument tries to move directly from outcome-responsibility to an obligation to compensate for any harm caused, and this is problematic for the reasons we have considered. But outcome-responsibility establishes a close and normatively significant connection between actions and their outcomes that can affect an actor's subsequent reasons for action, so it remains plausible to think that it has a role to play in justifying rights and obligations of reparation. By itself it is insufficient, but together with other considerations it might yield a generally-applicable reason to compensate and not just a reason to express regret, provide assistance, et cetera. (The failure to generate such a reason was the first of the two difficulties encountered by the volitionist argument for strict liability.) At the same time those considerations might narrow the resulting liability rule so as to avoid the indeterminacy to which general regimes of strict liability are subject. (The threat of such indeterminacy was the second difficulty.) Everything depends, of course, on what these other considerations are taken to be.

I would like to make the following suggestion. The localized distributive argument for fault and the agency-oriented understanding of outcome-responsibility are complementary. Each completes a gap in the other that prevents it from constituting an adequate justification, standing on its own, for correlative rights and obligations of reparation. Taken together, though, they form a single, coherent, justifying argument. Outcome-responsibility ensures that the distributive argument for fault can be nonarbitrarily limited to victims and injurers. Fault is the further consideration that supplements outcome-responsibility so as to produce a general obligation to compensate. Precisely because it is fault-based, the obligation also avoids the problem of pervasive indeterminacy.

Let us examine the various elements of the proposed argument, beginining with outcome-responsibility, to see how they fit together. A person who is outcome-responsible for a loss has a normatively significant connection with it that is capable of affecting her subsequent reasons for action. She has a special responsibility because there is a sense in which the outcome is hers, although perhaps not exclusively so. It resulted from an exercise of her capacity to act, and that capacity is closely bound up with her status and sense of herself as a person. She necessarily acted with the intention of making a difference in the world, and even if the difference that resulted is not one she intended or foresaw, she has a special connection with it all the same.

 [*498]  There may be more than one person who is outcome-responsible for a given loss, and this group may include the victim. Even if it does not, he has his own normatively significant connection with the loss just by reason of having suffered it. If there is a sense in which a loss belongs to those who are outcome-responsible for it, there is also an obvious sense in which it belongs to the person who suffered it. As we saw when discussing Honore's theory, personhood has an experiential as well as an agential dimension. It is an unavoidable aspect of being a person that I experience what happens to me rather than just perceive it as another event in the world, and that I have to live with and accept the consequences for myself. Losses can, in a sense to be explained, be transferred, but harmful or injurious experiences cannot. The loss is the victim's simply by virtue of his having experienced a harm or an injury.

The term "injury" I take to refer either to an event that interferes with an interest, or to the actual setback to an interest that results from such an event. "Loss" refers more abstractly to the type or magnitude of a setback to an interest, but without supposing that the extent of the interference is necessarily quantifiable in any exact way, either monetarily or otherwise. A "harm" is, strictly speaking, an injury that affects future well-being (although I sometimes use it more loosely, as a synonym for "injury"). Harm in the strict sense ordinarily takes the form of a setback to personal autonomy, which is the interest a person has in being able to continue to pursue current projects and to have available different options and opportunities in life. 185 Personal autonomy is the most important interest with which principles of reparation are concerned because interferences with it can represent especially serious and long-lasting effects on well-being, but the principles encompass other interests as well. 186 Losses are capable of being shifted or distributed in the following sense. If A pays B compensation the augmentation of B's resources makes up for, even if it does not exactly eliminate, the original setback to B's interests by increasing the options and opportunities he can pursue; it advances his interest in personal autonomy. A, on the other hand, suffers a degradation of her autonomy because the resources available to her have been depleted. Losses are shifted, in other words, by first being converted into a common currency of autonomy, which is the most fundamental human interest to which principles of reparation apply.

A loss is of moral concern because it represents an interference with human well-being. Together the victim of a loss and any persons who are outcome-responsible for it -- the victim may, to repeat, be one of these -- define a group whose members have, in different ways, a close and normatively significant connection with the injury the loss represents. But there is an asymmetry between the situation of the victim and that of the  [*499]  other members of the group. In the nature of things an injured person must continue to bear his loss unless someone else has a reason to relieve him of it. But a person who did not experience the initial injury must, if she is to assume the loss, have a reason for doing so. Outcome-responsibility is reason-affecting, but, as we have seen, there are no grounds for thinking that it is sufficient by itself to justify shifting a loss. The question to be asked, then, is whether there is some other factor that, together with outcome-responsibility, could serve as an appropriate justification for redistribution.

The present suggestion is that the distributive argument for fault points to just such a further factor: among those persons who have a normatively significant connection with a given loss, it is morally preferable that it be borne by whoever acted faultily in producing it. If no one was at fault, then for reasons already considered the loss remains with the person who suffered the injury. If more than one person was at fault -- the victim might be one of these -- then the loss is proportionally shared among them. The basic question is whether the loss should be shifted from the victim to one or more persons who are outcome-responsible for it, and the essential characteristic of outcome-responsibility is the fact of having voluntarily performed an action or actions that causally contributed to the outcome in question. 187 This suggests that in the present context the distributive argument should look to whether those actions were, under this description, to be faulted or not. It is not, in other words, fault in the air or general moral deficiency that matters, but fault in bringing about the injury the victim suffered. This approach seems particularly apposite when it is considered that autonomy is the interest of principal concern for the theory of reparation, because an action that is faulty due to its intended or foreseeable effects on someone else represents an improper or excesssive exercise of autonomy at that person's expense. Thus the redistribution of losses envisaged by this version of the distributive argument does not turn on an unrestricted application of the deficiency principle discussed in Part II, and so avoids the difficulties of expansiveness to which that principle can give rise.

Let me call this the volitionist/distributive argument for correlative moral rights and obligations of reparation. 188 Its central claim is that fault  [*500]  and outcome-responsibility together create a reason to assume part of a loss proportional to one's relative degree of fault in producing it, compared with the fault of other outcome-responsible persons. Much more needs to be said about the argument, and it will not be possible to say everything in this Article. But in the discussion that follows I would like to place the argument in a more general context of moral justification and discourse, say something further about the nature of outcome-responsibility, and develop the argument in a little more detail, focusing on the kind of fault it involves and the obligatory nature of the reasons for action to which it gives rise.

B. Philosophical Presuppositions

Let me begin by saying something about the philosophical character of the volitionist/distributive argument and the understanding of moral discourse it presupposes. We have already seen that outcome-responsibility has its roots in a set of evaluative attitudes, including most importantly agent-regret, that an agent holds with respect to her own past actions. These attitudes are premoral insofar as they bear on the outcomes of the agent's actions for herself, but they take on a moral quality when the outcome is an adverse effect for another. The attitudes the agent holds towards her own actions are part of a larger network of evaluative attitudes that she has. Included in this network are her attitudes towards the actions of others insofar as they affect her negatively, whether by causing her harm or in some other way -- these are, in the phrase Peter Strawson employs in his classic essay "Freedom and Resentment," 189 her personal reactive attitudes, which Strawson takes to be exemplified by resentment -- and also her attitudes towards the actions of others insofar as they negatively affect third parties -- what Strawson calls the vicarious (or impersonal or disinterested) analogues of the personal reactive attitudes, in this case exemplified by moral indignation. Strawson says that both types of attitude "involve, or express, a certain sort of demand for inter-personal regard. The fact of injury constitutes a prima facie appearance of this demand's being flouted or unfulfilled." 190 He explains that this appearance can be shown to be just appearance by a class of considerations that defuse resentment or indignation  [*501]  by, essentially, calling into question the voluntary nature of the act vis-a-vis the negative outcome but without inviting us "to see the agent as other than a fully responsible agent;" rather they invite us "to see the injury as one for which he was not fully, or at all, responsible." 191

Although Wiliams and Nagel, in their respective discussions of moral luck, 192 do not explicitly adopt a Strawsonian framework, their work suggests that the network of evaluative attitudes is, as Strawson describes it, incomplete. Williams' notion of agent-regret makes clear that the network extends to self-as well as other-regarding attitudes. Williams and Nagel also show how the evaluative moral attitudes that we hold regarding both our own and others' actions are more complex than Strawson in his essay allows. 193 The class of neutralizing considerations Strawson discusses, which includes such factors as lack of intent, lack of knowledge, necessity and duress, clearly defuses responsibility in one sense; in the criminal context, they might negative culpability or blameworthiness. But as Nagel in effect points out, we have a sense that agents are in another way morally responsible for the effects of their actions on others that survives the defusing of responsibility in this first sense and infuses our evaluative attitudes towards both the actions and the actor. This is what Honore refers to as outcome-responsibility. One of the points I urge below is that there is a third level of responsibility, related to but distinct from the other two, which enters into the justification of rights and obligations of reparation by determining what the distributive argument for fault should take "fault" to be. This level of responsibility presupposes outcome-responsibility but does not extend to all of its instances. In tort law it is recognized in the concept of the reasonable person.

Let me turn to the nature of moral justification and argument that is at issue here. In a famous passage in his essay Strawson says that questions of justification are internal to "the general structure or web of human attitudes and feelings" he describes, and that the existence of this framework of attitudes is itself given with the fact of human society: "As a whole, it neither calls for, nor permits, an external 'rational' justification." 194 To  [*502]  put this general idea in somewhat different terms, the network of attitudes of which Strawson speaks is given to us as an unavoidable feature of our sense of ourselves and of others as persons, that is, the sense we have of individuals as agents who are also capable of experience and self-reflective cognition. These attitudes are normative in the weak, evaluative sense that was discussed in the preceding section, but also in a stronger, reason-giving sense. When they take on what Strawson calls a vicarious or disinterested aspect they become moral attitudes, and the reasons become moral reasons. But the important point for present purposes is that moral argument and justification take place within this framework. It is pointless to demand an external justification for these attitudes or for their connection to reasons for action. 195 Moral argument is primarily a matter of clarifying our evaluative attitudes and developing a sense, so far as future action is concerned, of what is appropriate in light of them.

This is a very different picture of moral argument from that associated with Weinrib's Kantian rationalism, since there is no remotely similar assumption that norms governing human relations can be unconditionally derived by means of a purely conceptual argument. 196 I do not think that such derivations are possible, although that is too large a topic to be considered here. But I have argued that Weinrib has not succeeded in producing such a derivation for the principle of reparation he defends. He is correct to say that normativity is inherent in agency, but only in the sense that in our own experience of agency (and of personhood generally) evaluative attitudes of the sort we have been discussing seem inescapable. The understanding of rights and obligations of reparation that I am advancing thus cannot be regarded as being necessitated by the concept of agency, apart from our experience of it. The volitionist/distributive argument is an attempt to clarify our evaluative attitudes regarding actions that produce harmful outcomes, and to show why it is appropriate to conclude, in light of those attitudes, that under certain circumstances reparation is due. I do not think that anything more than this can be said.

 [*503]  C. Outcome-Responsibility Reconsidered

It will be convenient at this point to say something more about the nature of outcome-responsibility, taking as our point of departure the relationship between outcome-responsibility and causation. I have already mentioned that outcome-responsibility is not simply to be equated with having causally contributed to an outcome. As we saw in Part II, causation is usually understood in terms of a requirement of strong or weak necessity, and it is possible that an action can meet the relevant necessity requirement, whatever it turns out to be, with respect to an outcome that in other ways is quite remote from the action. In this subpart I wish to suggest that both outcome-responsibility and the associated attitude of agent-regret are based on a sense of having made a difference in the world. 197 It is quite conceivable that this sense may be lacking, so far as a given outcome is concerned, even though the agent has causally contributed to its occurrence. Interestingly, Hart and Honore explain what they call our "common sense" conception of causation in terms of a sense of having made a difference. Our initial and most basic understanding of causation, they maintain, arises with the observation of changes that directly follow upon our own deliberate interventions in or manipulations of the immediate physical environment, and is then extended to encompass cases where an event -- whether an action or not -- can be spoken of as making a difference because it constitutes a departure from the "normal" course of events. 198 It has been persuasively argued that Hart and Honore's "common sense" account is not a very satisfactory theory of causation, 199 but it does, I think, make a great deal of sense if it is understood as an analysis of outcome-responsibility. Their discussion of the significance of our sense of making a difference offers an important insight into the basis of outcome-responsibility, and what they have to say about the limits of "causal attribution" can be very usefully read as a partial characterization of the limits of outcome-responsibility, that is, of the ways in which it is not simply to be equated with causation properly understood.

Hart and Honore say that we ordinarily refuse to attribute an occurrence to an action as its consequence, even if the occurrence would not have taken place but for the preceding action, where there intervenes either a deliberate human act intended to bring about the occurrence, or a conjunction of events that amounts to a "coincidence." 200 An example of the first type of case is this: A sets fire, perhaps carelessly, to some bracken, but just as the flames are about to die out B, acting independently, deliberately pours petrol on them and burns down the surrounding forest. Hart and Honore say that this kind of intervention by a second human agent "is a paradigm among those factors which preclude the assimilation in causal judgments of the first agent's connection with the eventual harm  [*504]  to the case of simple manipulation." 201 If the phrase "causal judgments" is replaced by "judgments of outcome-responsibility," this is surely correct. B has in effect deliberately used A's action, or at least the immediate consequence of it, as a means to bring about the forest fire, and as a result A may have little or no sense of having made a difference with respect to its occurrence. Even though this particular fire would not have happened but for A's action, there are any number of other imaginable means that B could have used to bring about the end he intended, and perhaps he would have employed one of them to cause a different but similar fire if A's flames had not been to hand. 202 B is outcome-responsible for the forest fire (and responsible in the culpable sense as well), but A is not (which is of course not to say that an agent can never be outcome-responsible for the consequences of another agent's deliberate actions).

Hart and Honore's example of a coincidence is this. Suppose that A hits B, who is bruised by the blow and falls to the ground stunned; at that moment a tree crashes on the very spot where B is lying and kills him. 203 They characterize this as a coincidence because, inter alia, the conjunction of these events was not the result of human design and was in addition "very unlikely by ordinary standards." 204 Hart and Honore conclude that A caused only B's bruises, not his death. But even transposed into the language of outcome-responsibility this is not, I would argue, quite correct. It is certainly true that epistemic considerations of expectability or foreseeability affect our judgments of outcome-responsibility, since there is less of a sense of having made a difference when an outcome is so unusual or unpredictable or freakish as to seem to have nothing to do with the original action, but physical and temporal proximity also has a bearing on such judgments. This latter dimension of proximity is presumably related to the primitive sense of making a difference, described by Hart and Honore, that is associated with direct manipulation of one's immediate physical environment. I would argue that in the death-by-treefall hypothetical not only did A's action causally contribute to B's death, but A would generally be judged as outcome-responsible for it, at least in a pre-legal sense, because of the relatively direct physical and temporal connection between the two events. (Whether A would be judged as morally culpable for the death, or as having a moral obligation of reparation, are obviously separate questions.) As the number of intervening contributing causes grows larger, or as the time span between action and outcome lengthens, the less direct this connection will be, but here there was only one intervening cause and the time span was relatively short. 205

 [*505]  Negligence law has dealt with the proximity that limits outcome-responsibility under the doctrinal heads of duty of care and proximate cause (remoteness, as it is sometimes called). In both areas a test of foreseeability has eventually won out, although a physical/temporal test of "directness" prevailed for a while in the latter area. 206 I have suggested that there is more to proximity than foreseeability, but it will not be necessary to explore that issue further here because the law has correctly sensed that it is proximity-as-foreseeability that is particularly relevant to reparation. The reason is that the existence of fault depends itself on epistemic considerations, in the form of belief in or actual or constructive knowledge of causal regularities, and this gives rise to a natural continuity between fault and proximity-as-foreseeability. Both the basis of and the limitations on outcome-responsibility are determined by the sense of having made a difference, and this is a complex phenomenon. But there is no doubt that it is present where our actions set in motion a foreseeable train of events that conforms to known or partially known causal regularities, since this increases our sense that we could have had a measure of control over the situation, or at least that some agent, perhaps an idealized one, could have had some control. If action generally produced outcomes that conformed to no specifiable regularities, so that we could never or almost never predict what the result of an action would be, then we would have no sense that agency was in any way meaningful, either for ourselves or with respect to its "effects" on others; there would be no sense of making a difference. It is the possibility of control, which depends in turn on the existence of knowable regularities, that creates meaning of both kinds.

This idea of the "possibility" of control need not relate to our own subjective capacities. It is the sense that natural processes are often or regularly capable of being controlled by some agent, imaginable to us, that makes agency a meaningful notion. This suggests that outcome-responsibility for a specific outcome involves, in at least one important dimension, a retrospective evaluation of action that depends on a comparison with what would have been foreseen by an idealized agent to whom has been attributed a certain level of knowledge of the relevant causal regularities. So far as the actual agent is concerned this will often be a necessarily retrospective evaluation, since there is no suggestion that she herself could, on the particular occasion, have foreseen the outcome in question. It is possible, obviously, to attribute different levels of knowledge to this idealized agent. There is a continuum of possibilities here, ranging from omniscience at one end to the actual knowledge and beliefs of the particular agent at the other. An attribution of omniscience would equate outcome-responsibility with causation. Evaluation in light of the agent's actual knowledge and beliefs would equate it with a voluntarist conception of responsibility of the sort that justifies judgments of culpability and  [*506]  blameworthiness. 207

I doubt that there is a single conception of the ideal agent which defines outcome-responsibility in a uniquely correct way for all purposes. Different persons will inevitably judge their own actions against different idealizations, as concerns the effects of those actions for both themselves and others. 208 This means, among other things, that agents may differ in how they think the outcomes of their earlier actions affect their present reasons for action, and in particular may differ on the question of when they have a reason -- as oppposed to an obligation -- to offer compensation to someone whom they have injured. In the much-discussed English case of Bolton v. Stone 209 the defendants, from whose property a cricket ball was hit which injured the plaintiff on an adjacent street, allowed her to keep the damages that had been awarded by a lower court even though they were ultimately held not to have been at fault. Richard Epstein has said that the feeling many people had at the time that this was an appropriate thing to do is an indication of an inconsistency between the fault principle and ethical sentiment, 210 but this presupposes too simplistic a view of both morality and reparation. It is possible for someone to have moral reasons, possibly dependent not just on his own actions and situation but on his personal evaluation of them, that do not rise to the level of obligations. A particular agent's regret and sense of outcome-responsibility for a harmful outcome can provide a moral reason to compensate, even in the absence of an obligation to do so.

It is nevertheless true that general moral principles of reparation are concerned with rights and obligations, not with defeasible reasons, and they must apply to people in a uniform way. The law of torts holds persons to a minimal uniform level of outcome-responsibility, represented doctrinally by the rules on duty of care and proximate cause, which is determined primarily but not exclusively 211 by what a reasonable person would foresee  [*507]  as possible or likely outcomes of a given action. These doctrines can be taken, I think, as reflecting, roughly but more or less accurately, the content of the underlying moral principles. Those principles take as a given that outcome-responsibility is an inevitable dimension of agency in both a premoral and moral sense. They reflect a judgment that, so far as public accountability for outcomes is concerned, it is appropriate to employ a uniform idealization of agency that is determined by reference to some notion of "common" knowledge, or what the "ordinary" person of average mental capacities can be expected (in a non-normative sense) to know. (Public accountability means accountability in public, not accountability to the public.) It is important to emphasize that we are not at this stage talking about either fault or an obligation of reparation. At issue is the threshold requirement that places someone who has causally contributed to a harmful outcome within the group of persons who should be publicly considered as possible bearers of the loss.

D. Rights, Obligations and Fault

We are now in a position to address the following question. Who, if anyone, among those persons properly considered as possible bearers of a loss, should actually be required to bear it? Who, in other words, should be recognized as having a publicly acknowledged moral obligation of reparation? The answer to this question must be located within a more general theory about the nature of rights and obligations. An interest theory of rights is appropriate for this purpose because the losses for which reparation is to be made are setbacks to interests, understood as aspects of human well-being. The general understanding of rights I employ is Joseph Raz's version of the interest theory. According to Raz, a person has a right if, inter alia, his interest or well-being is a sufficient reason to hold someone else to be under a duty or obligation. 212 On this view, rights are grounds of obligations in others, although there need not be a closed list of obligations corresponding to the right. This understanding of rights will apply to each of the two levels of correlative rights and obligations that reparation involves. The primary level concerns the conduct that people are obligated not to engage in because of the loss it will or might cause others by interfering with their interests, where everyone also has a correlative right that no one engage in conduct of that type. The secondary or remedial level concerns correlative rights and obligations of reparation that come into play when a primary right has been violated and harm has resulted.

I shall assume that there exists, independently of principles of reparation, a primary moral obligation (together with the appropriate correlative right) not to intentionally or knowingly harm or attempt to harm the person or property of another, or to intentionally or knowingly subject either his person or property to a substantial degree of risk of harm (that is, to treat his person or property recklessly), without justification for doing so.  [*508]  I shall refer to this as the obligation not to culpably harm another. 213 When loss occurs as a result of a breach of this obligation, so that the interest protected by the corresponding primary right has been damaged, it is necessary to ask if there is sufficient reason to recognize a secondary obligation of reparation (together with a correlative right). The localized distributive argument for fault tells us, I argue, that in these circumstances there is sufficient reason: it is morally preferable that the person or persons who were culpably at fault in causing a loss bear it instead of the innocent victim. 214 But it is important to keep in mind that the conclusion that an obligation of reparation arises here is not, within the framework of Raz's interest theory of rights, capable of demonstration or proof. It emerges from a sense of appropriateness that depends on an appreciation of the normative significance of the interference with the victim's well-being, on the one hand, and of the injurer's outcome-responsibility and culpable fault, on the other. 215 This sense of appropriateness is simply given a more concrete shape by the distributive argument for fault.

I suggested earlier that our evaluative attitudes recognize different levels of moral responsibility. Thus far we have been concerned only with responsibility for acts of culpable wrongdoing and for the losses that flow from such acts, but the law of torts imposes liability for violation of an objective standard of care that need not involve wrongdoing in this sense. We saw earlier that Weinrib's theory of torts could not satisfactorily account for rights and obligations of reparation that are imposed for losses caused by nonculpable wrongdoing. 216 Can the approach to reparation I am defending do better in this respect? It is of course possible that the law does not, in its adoption of an objective standard of care, faithfully reflect the underlying moral principles, but I wish to argue that those principles do sometimes call for reparation where the actor has only exhibited "fault" in a nonculpable sense. The key here is that culpable fault is defined independently of principles of reparation, so that the application of the localized distributive argument to instances of culpable wrongdoing can serve as a paradigm that enables us to extend the argument analogically to  [*509]  cases where fault is not characterized apart from reparative concerns. The paradigm gives us the idea of a comparative inquiry in which it is asked who among the members of a distributive group, defined in the manner described above, should be regarded as the morally preferable bearer of a given loss. The existence of culpable fault in bringing about the loss tips the balance in a fairly decisive way, but the idea of the comparative inquiry can be extended to cases where fault in that sense is not present.

As we have seen, outcome-responsibility involves the retrospective evaluation of action in light of its outcome. Such evaluation can be moral but need not involve the ascription of blame or culpability. It is possible to judge actions as actions, considered simply as the product of agency, without judging the agent herself, although as the author of the action the agent's reasons for subsequent action may be affected; it is, after all, still her action. It is a mistake to think that evaluation of action in general, and moral evaluation in particular, must depend on voluntarist assumptions about what the particular agent was capable of doing in light of knowledge and beliefs she actually possessed. I have suggested that outcome-responsibility is a complex phenomenon, the roots of which lie in a sense of having through one's actions made a difference in the world. I have also suggested that it is outcome-responsibility understood in terms of reasonable foreseeability that is most relevant to reparation, because there is a natural continuity between this dimension of outcome-responsibility and the concept of fault. Culpable fault is assessed in voluntarist terms by judging the agent's actions in light of her actual knowledge and beliefs about the outcome. It is a judgment of the agent herself. Outcome-responsibility, understood in terms of reasonable foreseeability, yields a preliminary judgment of the agent's action from the perspective of a so-called reasonable person, to whom has been attributed common knowledge and beliefs about possible causal connections between actions and outcomes of the relevant kinds. It is a judgment of the action and not the agent because there is no necessary implication that the agent herself was capable of possessing and acting upon the knowledge in question, at least on the particular occasion. There is a kind of objectification of action and of the evaluation of action that is built into outcome-responsibility from the outset.

We have still not arrived at the notion of fault associated with the objective negligence standard, however, because the ex post attribution of common knowledge that treats an outcome as foreseeable does not tell us what should have been done in light of that knowledge; it simply gives the agent a retrospective sense of a certain kind that her action made a difference. My claim is that in retrospectively evaluating actions that have produced harmful outcomes, we sometimes have a sense that the action should be judged morally faulty for the purposes of reparation. Such judgments are based, as I sugggested a moment ago, on an analogy with the distributive argument for fault as it applies to instances of culpable wrongdoing. The argument in that paradigmatic form holds that culpable fault in causing a harmful outcome provides sufficient reason for shifting the loss to the culpable party. The present suggestion is that when common knowledge of the relevant causal regularities would lead an agent of average mental capacities to be aware of a sufficiently high level of risk of  [*510]  harm to other persons, taking account of both the probability and seriousness of the outcome, 217 then the action should be treated for purposes of reparation as faulty because it is more appropriate that the agent whose action is being evaluated should bear the loss than that the victim should.

There is a kind of circularity here, because the conception of fault that it is claimed the distributive argument should employ is being partly determined by the anticipated outcome of the argument, but it is not a vicious circularity. We already have an independent understanding of fault, in the form of culpability, and we have the idea of shifting a loss because the action (or one of the actions) that caused it was faulty with respect to its occurrence. Outcome-responsibility gives us the idea that action can be morally evaluated, in light of its outcome, in an essentially retrospective way. What is being suggested is simply that at a certain point outcome-responsibility for the harm a given action has produced should, so far as a publicly acknowledged obligation of reparation is concerned, be treated like culpable fault. The conclusion that there is such a point, together with the ex post determination of when it has been reached, are not matters capable of rational demonstration within the confines of the distributive argument, nor are they claimed to be. Like that argument even in its paradigmatic application to culpable fault, they depend on a sense of appropriateness that emerges from considered reflection on the normative implications of outcome-responsibility, where the outcomes in question are harmful interferences with human well-being.

The extension by analogy of the distributive argument for fault takes us from a notion of culpable responsibility that applies directly to the agent, to an understanding of fault that is rooted in outcome-responsibility and hence applies in the first instance to the action itself. Jules Coleman has spoken of a breach of the objective negligence standard as a "shortcoming in the doing, not in the doer," 218 and this is an apt way of putting the matter. But even though this objective notion of fault applies to the action, it has normative consequences for the agent whose action it is. This gives us the third, intermediate level of responsibility to which I referred earlier. So far as the particular agent is concerned, the moral judgment of her action will at this level be necessarily retrospective, since there is no suggestion that she could have acted at the time otherwise than as she did; the judgment will reflect an ex post perspective rather than the perspective of the agent at the point of action. Even so, the judgment must no doubt be accompanied by a sense that some agent, imaginable to us, could have acted without imposing the unacceptably high level of risk, and the idealization of  [*511]  agency that tort law has adopted under the label of the reasonable person is presumably meant to suggest that the "ordinary" or "average" person could and would have so acted. (To say that the reasonable person "would" have so acted is to attribute to him or her a moral character, in addition to a certain level of knowledge.)

There ordinarily seems to be no harm, especially in light of the assumption that the reasonable person could have avoided imposing a given risk, in translating the retrospective judgment that a harm-producing action was faulty into a judgment that the agent did what she ought not to have done, or that she breached a primary obligation. 219 Such a translation posits a standard of conduct -- in negligence law, the objective standard of care -- that the particular agent may not have been capable of attaining. But this is not problematic so long as it is remembered that it is the retrospective judgment that the action was faulty which leads us to say that the agent breached a standard of conduct, and not the other way around. The extension by analogy of the distributive argument for fault in effect works backward from the secondary, reparative level of correlative rights and obligations to the primary level, whereas in the paradigmatic case of culpable fault the argument proceeds in the opposite direction. Breach of the objective standard does not justify the attribution of blame or culpability to the agent, but by the same token the standard is not subject, so far as the individual's own capacities are concerned, to the dictum "ought implies can." To the extent that that dictum can apply to a particular agent's capacities, it is limited to subjective standards of conduct, importing mens rea of some kind, the breach of which could support a judgment of blameworthiness. 220 But it is only by reference to an idealization of agency like the reasonable person that the dictum can have any application to the objective negligence standard. Since the moral evaluation of action need not depend on voluntarist assumptions about what a particular agent was capable of doing, this "objectification" of the dictum should not be troubling.

The intermediate level of responsibility associated with the objective negligence standard is, on the view I am defending, a construction of the reparative inquiry. I cannot pursue here the question of how particular levels of risk are determined by that inquiry to be acceptable or unacceptable. But it should be noted that it may not be possible to capture completely the content of the objective standard in the form of one or more forward-looking, action-guiding norms (rules, for short). Weinrib has said that the general standard of permissible risk imposition "is not susceptible  [*512]  of precise measurement and is applied by the trier of fact on a case by case basis," 221 that is, on an ex post facto basis. This may well be correct. If it is, however, that suggests that judgments about an agent's success in meeting the objective standard of care might be retrospective not just in the sense that they do not assume the particular agent to have been capable of acting otherwise at the time, but also in the sense that the content of the standard is not necessarily taken to be explicitly statable in advance.

This possibility should trouble Weinrib, since it does not seem to sit very well with the Kantian notion that one's actions should comply with general maxims that can be formulated by rational deliberation prior to action. 222 It is less troubling for the view I am advancing here, which is not to say that it is entirely unproblematic. But just as it is a mistake to think that the moral evaluation of action must necessarily be constrained by the agent's own knowledge and capacities, so it may be a mistake to think that it must always be carried out on the basis of norms that are generally capable of antecedently guiding action. 223 It is sometimes said of the reasonable person in negligence law that he would not impose an unacceptable risk of harm on someone else. But if it cannot always be specified explicitly in advance which risks are unacceptable, then it is a capacity to make intuitive moral judgments that is being attributed to the reasonable person. It may be, moreover, that at least sometimes these moral judgments are being made from a thoroughly ex post perspective that could not be equated with the perspective of a reasonable person at the time of action, even one to whom has been attributed the capacity just mentioned. 224 These are difficult issues, and they cannot be further considered here.

E. The Distributive Nature of the Volitionist/Distributive Argument

There are two points about the distributive dimension of the volitionist/distributive argument to which I would briefly like to draw attention. The first is that it allows and indeed calls for an apportionment of loss among the various parties to a harmful interaction where more than one has acted faultily. The principles of Aristotelian distributive justice dictate that the apportionment should be made in accordance with the relative degree of fault each person has shown. This is, of course, exactly  [*513]  what happens in tort law under modern schemes of comparative negligence and contribution among tortfeasors. Furthermore, since the possibility of such apportionment is present in every case, it is reasonable to think that the moral inquiry in tort law is, as a general matter, precisely the localized distributive inquiry, as constrained by outcome-responsibility, that I have been describing. It is worth emphasizing that Weinrib seems to have no way of explaining apportionment schemes except by also appealing to localized schemes of distributive justice. If he were to make that move, however, there would be little or no justificatory work left to be done by his own theory of abstract right. Weinrib in fact appears to be committed to rejecting comparative negligence outright. Subjecting oneself to risk is presumably not a wrong so far as abstract right is concerned, and in any event it seems to have no relevance to whatever remedial measures are licenced by abstract right against another person who has acted wrongfully towards one.

The second point is this. One of the implications of the volitionist/distributive argument is that there is no hard and fast dividing line between corrective and distributive justice. Corrective justice, or at least the principles of reparation that fall within its bounds, is to be understood as a localized scheme of distributive justice that applies to losses. The reason losses are regarded as normatively significant, and hence are appropriately subjected to possible redistribution on a limited scale, is that they constititute interferences with human well-being. But it is also concern for human well-being that ultimately underpins distributive concerns at the comprehensive, community-wide level of social justice, and it is losses understood as interferences with human well-being that are the focus of general distributive schemes that take the form of, say, workers' compensation programs, or compulsory no-fault insurance plans. Losses thus understood, and the concern for human well-being that informs distributive justice at the social level, give rise to agent-neutral reasons for action, that is, reasons applicable to everyone, although these reasons are to a large extent coordinated and channelled by state action. The same concern for human well-being lies at the foundations of rights and obligations of reparation, except that here, because of the nature of outcome-responsibility, it generates reasons for action that are agent-relative; they apply only to persons whose faulty actions were causes of the relevant loss.

There is nonetheless a continuity of principle between localized and more comprehensive conceptions of distributive justice, and this may mean that in certain situations the dividing line between the two becomes blurred. It may also mean that localized distributive justice can be replaced by more general distributive schemes, like a compulsory no-fault insurance plan, without violating any fundamental moral rights. Finally, this underlying continuity of principle suggests that localized distributive justice makes certain presuppositions about background distributive justice at the social level. In my view, principles of reparation do presuppose certain facts about the general distribution of material resources, but that is a very large topic and not one that can be dealt with here. 225

 [*514]  Conclusion

I have considered in this Article six main arguments for correlative rights and obligations of reparation, namely, the argument from restitution, the distributive arguments for fault and strict liability, the volitionist arguments for fault and strict liability, and the volitionist/distributive argument for fault. The argument from restitution, based as it is on an incident of private property, is not so much mistaken as simply too limited in scope. The distributive argument for fault goes wrong in focusing exclusively on the normative significance of loss, thereby ignoring the normative significance of action. The volitionist argument for fault makes the reverse mistake. The distributive and volitionist arguments for strict liability both try to establish a normative connection between an action and loss caused by the action, but in each case the proposed connection is, for different reasons, problematic; a shared difficulty is a susceptibility to pervasive indeterminacy. But the volitionist argument for strict liability comes closer to forging the necessary connection by recognizing the fundamental importance of outcome-responsibility, and the volitionist/distributive argument for fault builds on that insight. It constructs a tighter and more adequate normative link between action and loss by incorporating both outcome-responsibility and the principal elements of the distributive argument for fault.

The volitionist argument for strict liability and the distributive argument for fault are, in truth, best regarded as incomplete aspects of a single argument. The resulting volitionist/distributive argument is complex, but that is to be expected if the normative significance of both action and loss is to be appropriately taken into account. The argument also captures what is best about the other arguments that were discussed (apart from the argument from restitution, which is of a different type). The distributive argument for strict liability is right in holding that the costs of an injury should sometimes be internalized to actors who causally contributed to the injury's occurrence, but wrong in trying to make causation alone the basis of internalization. The volitionist argument for fault is right to maintain that normativity is inherent in agency and that this has a bearing on questions of reparation, but wrong in its characterization of the relevant conception of normativity. Like the distributive argument for fault and the volitionist argument for strict liability, each of these arguments has some intuitive appeal. It is, I think, a mark in favor of the volitionist/distributive argument that it acknowledges and incorporates the kernel of truth in all of them.

Legal Topics:

For related research and practice materials, see the following legal topics:
Criminal Law & ProcedureDefensesIgnorance & Mistake of FactReal Property LawInverse CondemnationGeneral OverviewTortsNegligenceCausationCause in Fact

FOOTNOTES:


Click here to return to the footnote reference.n1. The best known theory of this type is Guido Calabresi's, which combines versions of the last three policies mentioned in the text. See, e.g., Guido Calabresi, The Costs of Accidents (1970); Guido Calabresi & Jon T. Hirschoff, Toward a Test for Strict Liability in Torts, 81 Yale L.J. 1055 (1972).

Click here to return to the footnote reference.n2. See, e.g., Jules L. Coleman, Corrective Justice and Wrongful Gain, 11 J. Leg. Stud. 421 (1982); Jules L. Coleman, Tort Law and the Demands of Corrective Justice, 67 Ind. L.J. 349 (1992).

Click here to return to the footnote reference.n3. See, e.g., Ernest J. Weinrib, Causation and Wrongdoing, 63 Chi.-Kent L. Rev. 407 (1987); Ernest J. Weinrib, Understanding Tort Law, 23 Val. U. L. Rev. 485 (1989); Ernest J. Weinrib, The Special Morality of Tort Law, 34 McGill L.J. 403 (1989).

Click here to return to the footnote reference.n4. Coleman has modified his position in Jules L. Coleman, The Mixed Conception of Corrective Justice, 77 Iowa L. Rev 427 (1992).

Click here to return to the footnote reference.n5. Stephen R. Perry, Comment on Coleman: Corrective Justice, 67 Ind. L.J. 381 (1992).

Click here to return to the footnote reference.n6. I borrow the term "obligation of reparation" from Neil MacCormick, who adapted it from Scots legal usage. Neil MacCormick, The Obligation of Reparation, in Legal Right and Social Democracy: Essays in Legal and Political Philosophy 212 (1982).

Click here to return to the footnote reference.n7. Pluralism of this kind is categorically rejected by Weinrib's formalist approach to tort law, briefly described below. For a very good discussion of the relationship between substantive principles and institutional structure, see Jules L. Coleman, The Structure of Tort Law, 97 Yale L.J. 1233 (1988).

Click here to return to the footnote reference.n8. Cf. Larry Alexander, Causation and Corrective Justice: Does Tort Law Make Sense?, 6 Law & Phil. 1, 6-7 (1987). Alexander's argument is limited by the fact that he does not discuss complex patterned theories of distributive justice, like Ronald Dworkin's theory of equality of resources, which define the pattern of distribution diachronically rather than in terms of what Robert Nozick calls a time-slice principle. See generally Ronald Dworkin, What is Equality? Part 2: Equality of Resources, 10 Phil. & Pub. Aff. 283 (1981). Alexander also does not consider the possibility, discussed in Part I below, that corrective justice is ancillary not to distributive justice but to the concept of property.

Click here to return to the footnote reference.n9. Ronald Dworkin, Law's Empire 276-312 (1986). Dworkin's theory is more plausible because it begins with the more complex, diachronically-defined conception of distributive justice defended in Dworkin, supra note 8, and also because it looks upon corrective justice as a means for refining or making precise a set of initially abstract distributive rights rather than as a means for simply preserving or policing an antecedently-given and completely determinate distributive scheme. I criticize Dworkin's theory of tort in Stephen R. Perry, The Relationship between Corrective and Distributive Justice (unpublished manuscript on file with the author).

Click here to return to the footnote reference.n10. The issue is discussed in Perry, supra note 9.

Click here to return to the footnote reference.n11. Aristotle, Nicomachean Ethics (Martin Ostwald trans., 1962) [hereinafter NE].

Click here to return to the footnote reference.n12. Robert Nozick, Anarchy, State and Utopia 155-60 (1974).

Click here to return to the footnote reference.n13. NE 1132a5.

Click here to return to the footnote reference.n14. NE 1132a9.

Click here to return to the footnote reference.n15. NE 1132a25.

Click here to return to the footnote reference.n16. NE 1132a25-30.

Click here to return to the footnote reference.n17. NE 1132a7-14.

Click here to return to the footnote reference.n18. NE 1132b19-20.

Click here to return to the footnote reference.n19. Cf. Nozick, supra note 12, at 171.

Click here to return to the footnote reference.n20. See, e.g., Tony M. Honore, "Ownership" in Oxford Essays in Jurisprudence 107, 114 (Anthony G. Guest ed., 1961) (discussing the right to possess as one of the incidents of "the liberal concept of full individual ownership": "If dispossession without the possessor's consent is, in general, forbidden, the possessor is given a right in rem, valid against persons generally, to remain undisturbed, but he has no right to possess in rem unless he is entitled to recover from persons generally what he has lost or had taken from him.").

Click here to return to the footnote reference.n21. I am not overly concerned with the historical or interpretive question of whether Aristotle did in fact advance this argument, since it strikes me as an important argument whether or not he actually intended to make it. Weinrib offers a different interpretation of Aristotle's account of corrective justice, according to which it is an inchoate version of Weinrib's own Kantian/Hegelian theory, discussed in Part III below. See Ernest J. Weinrib, Corrective Justice, 77 Iowa L. Rev. 403 (1992). Weinrib is much better qualified than I to interpret Aristotle, and his reading may well be the preferable one. I would only make the following two points. First, if Weinrib's interpretation is correct, then Aristotle's argument is both incomplete and difficult to reconcile with the more general Aristotelian understanding of ethics. Second, Weinrib says that Aristotle could not have intended that the equality of corrective justice be that of distributive justice, because this would lead to the collapse of the distinction between them. Id. at 420. But, as is noted below in the text, this collapse will not occur if corrective justice is understood in terms of the obligation of restoration, and hence is not grounded directly in distributive justice. See also Weinrib's own argument against collapse in Ernest J. Weinrib, Legal Formalism: On the Immanent Rationality of Law, 97 Yale L.J. 949, 984 n.75 (1988).

Click here to return to the footnote reference.n22. As will be shown later, one of Richard Epstein's arguments in favor of strict liability also takes this form. Epstein has a number of other arguments as well, one of which is that strict liability is directly entailed by the concept of property; this would obviate the need for an intermediate step in the argument which relies on the obligation of restoration. See Richard Epstein, Takings: Private Property and the Power of Eminent Domain, 97-98 (1985) [hereinafter Takings]; Richard Epstein, Causation and Corrective Justice: A Reply to Two Critics, 8 J. Legal. Stud. 477, 500-01 (1979). As a purely conceptual claim this idea of direct entailment seems highly dubious. If, however, the argument is to be understood as a normative one, then it is no longer clear that it has anything to do with the concept of property. See Stephen R. Perry, The Impossibility of General Strict Liability, 1 Can. J. L. & Juris. 147, 151-52 (1988). Ernest Weinrib also criticizes this claim by Epstein, but then offers what might easily be taken to be -- and is criticized by Coleman as -- a purely conceptual argument intended to show that the concept of property directly entails a negligence standard. See Weinrib, Causation and Wrongdoing, supra note 3, at 425-28, and Jules L. Coleman, Property, Wrongfulness, and the Duty to Compensate, 63 Chi.-Kent L. Rev. 451, 454-60 (1987). In a later article, however, Weinrib makes clear that it is a particular normative conception of property which he thinks is associated with a fault-based understanding of reparation. See Ernest J. Weinrib, Right and Advantage in Private Law, 10 Cardozo L. Rev. 1283, 1289-91 (1989) [hereinafter Right and Advantage]. In fact Weinrib's theory of reparation and his conception of property both seem to flow from a source more fundamental than either, which is a certain understanding of agency and moral personality. So far as reparation is concerned this gives rise to an argument of the third type, and will be discussed as such in Part III below. As we shall see, Weinrib's argument is indeed conceptual, but at the level of agency rather than property; the claim is that normativity is inherent in the concept of agency.

Click here to return to the footnote reference.n23. John Rawls appears to favor an interpretation of Aristotle's writings according to which corrective justice is wholly ancillary to distributive justice. See John Rawls, A Theory of Justice 10-11 (1972). A similar view is advanced in James Gordley, Equality in Exchange, 69 Cal. L. Rev. 1587, 1588-92 (1981). The difficulties to which simple versions of this position give rise are briefly discussed in the Introduction, supra.

Click here to return to the footnote reference.n24. There is no need to understand this metaphor as suggesting that there must initially be an actual equality of holdings.

Click here to return to the footnote reference.n25. See, e.g., Nozick, supra note 12, at 150-52; Weinrib, Right and Advantage, supra note 22, at 1289-91, 1293-97.

Click here to return to the footnote reference.n26. Epstein, Causation and Corrective Justice, supra note 22, at 500.

Click here to return to the footnote reference.n27. See, e.g., Honore, supra note 20, at 129; Jeremy Waldron, The Right to Private Property 33 n.15, 177-83, 361-63, 398-400 (1988). Waldron points out that writers like Nozick who defend a Lockean right of self-ownership do so on the grounds that it is only such a right that can provide protection for individual integrity against various kinds of interference of an essentially tortious character. Id. at 399-400. In that case, however, it would be circular to argue that individual integrity is protected against tortious interference because there is a right of self-ownership.

Click here to return to the footnote reference.n28. Epstein, Causation and Corrective Justice, supra note 22, at 500 n.69.

Click here to return to the footnote reference.n29. Cf. Dworkin, supra note 8, at 301.

Click here to return to the footnote reference.n30. The explanation would not extend to compensation for consequential damages or for the owner's subjective loss due to, say, valuing the object at more than market value.

Click here to return to the footnote reference.n31. NE 1132a8.

Click here to return to the footnote reference.n32. NE 1132a11.

Click here to return to the footnote reference.n33. Aristotle may have thought otherwise, since he says that "the just occupies the median between a gain and a loss: it is to have an equal amount both before and after the transaction." NE 1132b19-20. This suggests, rather implausibly, that the point of corrective justice is to split the difference between gain and loss even when they are distinct and unequal. Modern private law, however, distinguishes clearly between principles of tort, which are concerned with repair of loss, and principles of restitution, which are concerned with disgorgement of gain. The distinction is nicely illustrated by the common law doctrine of waiver of tort, which under certain circumstances gives an injured party an election between obtaining compensation for her loss and recovering the other party's gain. The plaintiff must opt for one claim or the other; she cannot pursue both. The classic discussion of the doctrine is Arthur L. Corbin, Waiver of Tort and Suit in Assumpsit, 19 Yale L.J. 221 (1910).

Click here to return to the footnote reference.n34. See supra note 33.

Click here to return to the footnote reference.n35. Richard Epstein is apparently of the view that the general obligations of both restitution and reparation can be reduced to what I have called an obligation of restoration. After referring to the election between restitutionary and reparative remedies which is offered by the doctrine of waiver of tort, he goes on to say the following: 
Where the loss to the plaintiff exceeds the gain to the defendant, the plaintiff recovers the full extent of his loss under a tort theory. Where the gain to the defendant exceeds the loss to the plaintiff, the plaintiff captures that gain under a restitution theory. . . . This remedial choice is presented where the benefit and loss are of roughly the same magnitude or where they are widely different. Yet the entire debate is intelligible only because it presupposes that the original wrong for which relief is required is the taking of private property.

 
Epstein, Takings, supra note 22, at 36-37. Epstein does not explicitly say as much, but the reason why relief is required for a taking is presumably that the taker has an obligation to restore what she has taken. It is not enough, however, simply to assert that comprehensive obligations of restitution and reparation can be reduced to an obligation of restoration: an argument is required. Epstein's argument with respect to reparation is set out in the passage quoted in the text, infra note 36. He does not seem to have an argument for restitution.

Click here to return to the footnote reference.n36. Epstein, Takings, supra note 22, at 38.

Click here to return to the footnote reference.n37. Id. at 74.

Click here to return to the footnote reference.n38. See Richard Epstein, A Theory of Strict Liability, 2 J. Legal Stud. 151, 158-60. See also Perry, supra note 22, at 148-50.

Click here to return to the footnote reference.n39. Perry, supra note 22, at 148-50.

Click here to return to the footnote reference.n40. Innocent damage caused during a wrongful appropriation is a more difficult case that need not be considered here.

Click here to return to the footnote reference.n41. A similar but somewhat different argument for the same conclusion would begin with the premise that the obligation of restoration conceivably requires no more than that a person in possession of someone else's property return it to the owner in whatever physical state it happens to be in. Further normative argument concerning the relationship between wrongdoing and reparation is thus required, the argument would continue, in order to show why and under what circumstances one is entitled to have property in a certain condition, as opposed to just having it. As we shall see below, Ernest Weinrib argues that the principle of abstract right, which he thinks underlies corrective justice, is indifferent to the physical state of property. He relies on this contention to argue against Epstein's thesis that destruction of property is a taking. Weinrib, Causation and Wrongdoing, supra note 3, at 423-24.

Click here to return to the footnote reference.n42. Coleman, Corrective Justice and Wrongful Gain, supra note 2, at 424-25.

Click here to return to the footnote reference.n43. Id. at 425.

Click here to return to the footnote reference.n44. Coleman's position differed on this point from Epstein's: See supra note 35.

Click here to return to the footnote reference.n45. Oliver W. Holmes, The Common Law 68 (Mark DeWolfe Howe ed., 1963).

Click here to return to the footnote reference.n46. Holmes' own response to the argument was for that reason unsatisfactory. See Perry, supra note 22, at 154-55.

Click here to return to the footnote reference.n47. Bamford v. Turnley 3 B. & S. 66, 85, 122 Eng. Rep. 27, 33 (Ex. 1862): "If, though the wood were their own, [railway owners] still would find it compensated them to run trains at the cost of burning the wood, then they obviously ought to compensate the owner of such wood, not being themselves, if they burn it down in making their gains." This argument is clearly premised on the thought that a person who gains from activity -- or, no doubt, who merely stands or hopes to gain -- ought, in fairness, to bear the costs of that activity as well.

Click here to return to the footnote reference.n48. I have argued elsewhere that something like this argument lay at the heart of the case for strict liability that Richard Epstein was attempting to develop in his early papers, although it is often obscured by a terse and elliptical formulation or by the overlay of other arguments. See Perry, supra note 22, at 154-55. Although it is not expressed very perspicuously, the essence of the argument can be discerned in the following sentence from Epstein's first article on strict liability: "The doctrine of strict liability holds that proof that the defendant caused harm creates [a presumption of liability] because proof of the non-reciprocal source of harm is sufficient to upset the balance where one person must win and the other must lose." Epstein, supra note 38, at 168-69. (I argue in Perry, supra note 22, at 159 & 160, that such phrases as "non-reciprocal source of harm," which appear quite frequently in early Epstein, are simply surrogates for the distinction between activity and passivity.) In subsequent articles Epstein sometimes formulated the elements of the argument in more transparent terms. For example, its character as an instance of localized distributive justice was brought out quite clearly when Epstein wrote that "the only proper question for tort law is whether the plaintiff or the defendant will be required to bear the loss." Richard Epstein, Defences and Subsequent Pleas in a System of Strict Liability, 3 J. Legal Stud. 165, 169 (1974). Similarly, the idea that it is unfair to externalize one's costs is forcefully expressed in the following passage: "Once a defendant is allowed to excuse himself on the grounds that he acted with due regard for the plaintiff, it follows that he will be able to keep the benefits of his own actions even as he imposes their costs upon a stranger." Richard Epstein, Intentional Harms, 4 J. Legal Stud. 391, 398 (1975). Epstein has recently modified his earlier views, although the nature and extent of his shift in position are not entirely clear. See Richard Epstein, Causation -- In Context: An Afterword, 63 Chi.-Kent L. Rev. 653 (1987).

Click here to return to the footnote reference.n49. I have criticized the distributive argument for strict liability in Perry, supra note 22, at 154-59, and consequently will not consider it at length here.

Click here to return to the footnote reference.n50. Richard Epstein, The Temporal Dimension in Tort Law, 53 U. Chi. L. Rev. 1175, 1190 (1986).

Click here to return to the footnote reference.n51. Note how natural it is in this context to employ terms like "reckless" and "dangerous," which have unavoidable connotations of fault. As is remarked below, Epstein's putative account of causation, which is based on "paradigms" of ways to cause harm, is really a theory of responsibility that implicitly incorporates a particular conception of fault. See further Perry, supra note 22, at 161-68. That the theory relies on fault-oriented, normative considerations is especially evident with respect to the so-called paradigm of dangerous conditions, which is one of two paradigms -- the other is the paradigm of force -- that Epstein would say are potentially applicable to the parked car example described in the text.

Click here to return to the footnote reference.n52. Weinrib holds a view of causation that is surprisingly Epsteinian. See, e.g., Weinrib, Right and Advantage, supra note 22, at 1293 ("[Abstract right] employs a transitive conception of causation, in which action originating in one person reaches out to infringe the physical or proprietary rights of another."); cf. Weinrib, Understanding Tort Law, supra note 3, at 517 ("For tort law, one party is active and the other passive as the effects move from their origin in the defendant's act to their resting point in the plaintiff's injury."). Although Weinrib argues strenuously against strict liability and in favor of fault, his understanding of causation is an open invitation to the distributive argument for strict liability.

Click here to return to the footnote reference.n53. See Richard W. Wright, Causation, Responsibility, Risk, Probability, Naked Statistics, and Proof: Pruning the Bramble Bush by Clarifying the Concepts, 73 Iowa L. Rev. 1001, 1020-21 (1988). Wright argues persuasively that a weak-necessity approach is preferable to one based on strong necessity. Id. at 1022-34.

Click here to return to the footnote reference.n54. John Mackie, The Cement of the Universe: A Study of Causation 29-87 (rev. ed. 1980).

Click here to return to the footnote reference.n55. Richard W. Wright, The Efficiency Theory of Causation and Responsibility: Unscientific Formalism and False Semantics, 63 Chi.-Kent. L. Rev. 533 (1987); Richard W. Wright, Causation in Tort Law, 73 Cal. L. Rev. 1737 (1985); Wright, supra note 53. As Wright acknowledges, the NESS test was first formulated by H.L.A. Hart and Tony Honore, who were themselves drawing on the work of John Stuart Mill. See H.L.A. Hart & Tony Honore, Causation in the Law 17, 111-17 (2d ed. 1985).

Click here to return to the footnote reference.n56. Perry, supra note 22, at 159-68.

Click here to return to the footnote reference.n57. Id. Cf. Izhak Englard, Can Strict Liability be Generalized?, 2 Oxford J. Legal Stud. 245, 251-52 (1982); Wright, Causation in Tort Law, supra note 55 at 1750-58; Gary T. Schwartz, The Vitality of Negligence and the Ethics of Strict Liability, 15 Ga. L. Rev. 963, 987-88 (1981).

Click here to return to the footnote reference.n58. A qualified claim of this sort is part of Hart and Honore's "common sense" account of causation. Hart & Honore, Causation in the Law, supra note 55, chs. 2, 3. The account is criticized for importing elements of policy or moral responsibility in Mackie, supra note 54, at 127-29, and in Wright, Causation in Tort Law, supra note 55, at 1745-50.

Click here to return to the footnote reference.n59. See Ronald Coase, The Problem of Social Cost, 3 J. L. & Econ. 1 (1960).

Click here to return to the footnote reference.n60. Ernest J. Weinrib, The Case for A Duty to Rescue, 90 Yale L.J. 247, 249 (1980). See also Weinrib, Right and Advantage, supra note 22, at 1299 ("Causation of harm is identified [by Coase] with the potential reciprocity of effect among economic actors in their competition for a scarce resource.").

Click here to return to the footnote reference.n61. Jules L. Coleman, Efficiency, Exchange and Auction: Philosophic Aspects of the Economic Approach to Law, 68 Cal. L. Rev. 221, 235-36 (1980).

Click here to return to the footnote reference.n62. Coase, supra note 59, at 2.

Click here to return to the footnote reference.n63. Coleman, supra note 61, at 236.

Click here to return to the footnote reference.n64. 4 C.P.D. 172 (1878-79).

Click here to return to the footnote reference.n65. Coase, supra note 59, at 13 (emphasis in original).

Click here to return to the footnote reference.n66. Guido Calabresi, Some Thoughts on Risk Distribution and the Law of Torts, 70 Yale L.J. 499, 533 (1961) (emphasis added).

Click here to return to the footnote reference.n67. Id. at 502. Cf. id. at 514 ("Not charging an enterprise with a cost which arises from it leads to an understatement of the true cost of producing its goods; the result is that people purchase more of those goods than they would if their true cost were reflected in their price."). In the same article Calabresi remarked that automobile accidents are probably a cost both of driving and of walking or living generally, but then added that "I have not, in this article, attempted to probe what influences our decision that a particular 'cost' is caused by one activity rather than another." Id. at 506 n.24. Interestingly, Calabresi has recently said that this footnote is all that remains of a full analysis of "causal reciprocity" along Coasean lines that had been contained in an earlier version of the article. He was convinced to remove that discusssion and replace it with a conventional Pigouvian treatment of causation. See Fred R. Shapiro, The Most-Cited Articles from the Yale Law Journal, 100 Yale L.J. 1449, 1483-84 (1991).

Click here to return to the footnote reference.n68. Calabresi, supra note 1, at 135.

Click here to return to the footnote reference.n69. Guido Calabresi, The Decision for Accidents: An Approach to Nonfault Allocation of Costs, 78 Harv. L. Rev. 713, 725 (1965). Calabresi later acknowledged that there is very little, if any, room in a theory of market deterrence for the concept of causation as it has traditionally been understood both in tort law and by corrective justice theorists. Guido Calabresi, Concerning Cause and the Law of Torts: An Essay for Harry Kalven, Jr., 43 U. Chi. L. Rev. 69, 85 (1975). Michael Trebilcock has recently suggested that there is an "inherent indeterminacy" in the process of internalizing accident costs even when the criteria adopted are economic rather than causal. See Michael Trebilcock, The Social Insurance-Deterrence Dilemma of Modern North American Tort Law: A Canadian Perspective on the Liability Insurance Crisis, 24 San Diego L. Rev. 929, 987-89 (1987).

Click here to return to the footnote reference.n70. It is interesting to note that an argument strikingly similar to the distributive argument for strict liability has sometimes been advanced to justify the complete abandonment of tort law in favor of a general no-fault compensation scheme. I am indebted to Gabrielle Turner for drawing to my attention the following passage from the Report of the Royal Commission of Inquiry, Compensation for Personal Injury in New Zealand 40 (1967). (This report, known as the Woodhouse Report, was the basis for the introduction of the general no-fault regime for personal injury in New Zealand.) 
Just as a modern society benefits from the productive work of its citizens, so should society accept responsibility collectively for those willing to work but prevented from doing so by incapacity. And, since we all persist in following community activities, which year by year exact a predictable and inevitable price in bodily injury, so should we all share in sustaining those who become the random but statistically necessary victims. The inherent cost of these community activities should be borne on a basis of equity by the community.

 
The objections to the distributive argument for strict liability that are discussed in the text do not apply to this collectivist version, although it may well suffer from other difficulties.

Click here to return to the footnote reference.n71. Jules L. Coleman, Mental Abnormality, Personal Responsibility, and Tort Liability, in Mental Illness: Law and Public Policy 107, at 120-21 (Baruch A. Brody & H. Tristram Engelhardt Jr., eds., 1980) [hereinafter Mental Abnormality]. See also Jules Coleman, Justice and the Argument for No-Fault, 3 Soc. Theory & Prac. 161, 174, 180 n.18 (1975) [hereinafter Justice and No-Fault] (introducing the term "weak retributive justice"); Jules L. Coleman, On the Moral Argument for the Fault System, 71 J. Phil. 473, 488-89 (1974) [hereinafter Moral Argument].

Click here to return to the footnote reference.n72. 162 N.E. 99 (N.Y. 1928).

Click here to return to the footnote reference.n73. William L. Prosser, Palsgraf Revisited, 52 Mich. L. Rev. 1, 17 (1953).

Click here to return to the footnote reference.n74. See, e.g., Sindell v. Abbott Laboratories, 607 P. 2d 924, 936 (Cal. 1980) ("The most persuasive reason for finding plaintiff states a cause of action is that . . . as between an innocent plaintiff and negligent defendants, the latter should bear the cost of the injury.").

Click here to return to the footnote reference.n75. George P. Fletcher, Fairness and Utility in Tort Theory, 85 Harv. L. Rev. 537, 542 (1972). Fletcher's views are further developed in George P. Fletcher, Synthesis in Tort Theory, 2 Law & Phil. 63 (1983).

Click here to return to the footnote reference.n76. See Jules L. Coleman, Justice and Reciprocity in Tort Theory, 14 U. W. Ont. L. Rev. 105 (1975).

Click here to return to the footnote reference.n77. Fletcher, Fairness and Utility in Tort Theory, supra note 75, at 550. There are serious problems with equating the impermissibility of nonreciprocal risk imposition with the provision of maximal security, but these cannot be considered here.

Click here to return to the footnote reference.n78. Id.

Click here to return to the footnote reference.n79. Cf. Jules L. Coleman, Moral Theories of Torts: Their Scope and Limits: Part I, 1 Law & Phil. 371, 389 (1982).

Click here to return to the footnote reference.n80. MacCormick, supra note 6, at 217-19. MacCormick assumes that "conditions of relative distributive justice" hold, id. at 217, and also that the obligation of reparation is limited by an obviously distributive qualification based on the injurer's ability to pay. Id. at 218-19.

Click here to return to the footnote reference.n81. Prosser, supra note 73, at 17. Cf. Coleman, Justice and No-Fault, supra note 71, at 167.

Click here to return to the footnote reference.n82. See Peter Cane, Atiyah's Accidents, Compensation and the Law 415 (4th ed. 1987). MacCormick criticizes Atiyah's proportionality argument on the ground that he is relying on a conception of justice that is apposite for criminal penalties but wholly inapposite for securing justice between the parties to a civil action. MacCormick, supra note 6, at 223-24.

Click here to return to the footnote reference.n83. Proportionality arguments are sometimes relied upon in tort cases to limit the defendant's liability. See, e.g., Overseas Tankship (U.K.) Ltd. v. Morts Dock & Engineering Co. (the Wagon Mound No. 1) [1961] App. Cas. 388, 422 (P.C.) (opinion of Viscount Simonds) ("For it does not seem consonant with current ideas of justice or morality that for an act of negligence, however slight or venial, which results in some trivial foreseeable damage the actor should be liable for all consequences however unforeseeable and however grave, so long as they can be said to be 'direct'.") Such arguments are seriously flawed because they ignore Prosser's point that absolute proportionality cannot be achieved in any event, so that if someone has to bear a disproportionate loss better that it be the person who has exhibited at least some fault. This is not to say, of course, that there are not other arguments which support the particular test for remoteness that was adopted in Wagon Mound.

Click here to return to the footnote reference.n84. Coleman, Mental Abnormality, supra note 71, at 125. Cf. Stephen R. Perry, Second-Order Reasons, Uncertainty and Legal Theory, 62 S. Cal. L. Rev. 913, 991-92 (1989). See also supra note 82, summarizing Neil MacCormick's response to Patrick Atiyah's critique of the fault principle. MacCormick, supra note 6, at 223-24.

Click here to return to the footnote reference.n85. As is explained below, a generalized distributive scheme cannot achieve absolute proportionality either, although it can bring about a more consistent relative proportionality.

Click here to return to the footnote reference.n86. Of course, the victim might not be wholly innocent, in which case the general principle of distributive justice would call for a relative apportionment of the loss between the parties in accordance with the relative degree of fault that each exhibited. This is how modern regimes of comparative negligence function.

Click here to return to the footnote reference.n87. Hart & Honore, supra note 55, at 268.

Click here to return to the footnote reference.n88. Cf. Robert E. Keeton, Legal Cause in the Law of Torts 21 (1963) ("But if it is relevant to take into account defendant's fault with respect to a risk different from any that would include the harm plaintiff has suffered, then would it not also be relevant to take into account his other faults as well? And would it not seem equally relevant to consider plaintiff's shortcomings? Shall we fix legal responsibility by deciding who is the better and who the worse person?").

Click here to return to the footnote reference.n89. Fletcher's implicit adoption of the immodest version of the distributive argument for fault may explain why he, like Coleman, accepts that rights and obligations of reparation are not necessarily correlative of one another. See Fletcher, Fairness and Utility in Tort Theory, supra note 75, at 551, 553-54.

Click here to return to the footnote reference.n90. Coleman, Mental Abnormality, supra note 71, at 124 (emphasis added).

Click here to return to the footnote reference.n91. Coleman, Corrective Justice and Wrongful Gain, supra note 2, at 426-27; Coleman, Tort Law and the Demands of Corrective Justice, supra note 2, at 365.

Click here to return to the footnote reference.n92. Perry, supra note 5, at 395, 397.

Click here to return to the footnote reference.n93. The one aspect of the theory of which this seems not to be true is the stipulation that losses only fall within the ambit of corrective justice if they result from human agency. See Coleman, Tort Law and the Demands of Corrective Justice, supra note 2, at 371-72. Of course if this stipulation were dropped it would become very difficult to continue to maintain that the theory was really an account of corrective rather than distributive justice. See Perry, supra note 5, at 397-98.

Click here to return to the footnote reference.n94. Coleman now agrees that his earlier view was mistaken. See Coleman, supra note 4. I discuss the theory of corrective justice Coleman currently defends in Stephen R. Perry, The Mixed Conception of Corrective Justice, 15 Harv. J.L. and Pub. Pol. 917 (1992).

Click here to return to the footnote reference.n95. See Coleman, Moral Argument, supra note 71, at 484-85.

Click here to return to the footnote reference.n96. In Coleman, Justice and No-Fault, supra note 71, at 167-68 Coleman seems to suggest that this is a problem with the scheme, basing his argument on an analogy with criminal law. For reasons already given I think this is wrong; distributive justice cannot, in general, rationally aim for absolute proportionality.

Click here to return to the footnote reference.n97. In Part IV below, I discuss what is, in effect, just such a reason. It should be noted that there are schemes of localized distributive justice that employ criteria of distribution not involving the moral assessment of conduct. Examples include economic theories of tort which hold that a loss ought to be placed on the party to a harmful interaction who is best able to spread it, or on the one with the deepest pocket. Not surprisingly, policy-based theories of this kind are often criticized for arbitrarily restricting the group of potential loss-bearers to injurers and their victims: there are likely to be third parties who have yet deeper pockets, or who are still better placed to spread losses, than either of these two. See, e.g., Coleman, supra note 7, at 1241; Trebilcock, supra note 69, at 984; Weinrib, Understanding Tort Law, supra note 3, at 498-99.

Click here to return to the footnote reference.n98. Ernest J. Weinrib, Law as a Kantian Idea of Reason, 87 Colum. L. Rev. 472, 485 (1987) [hereinafter Kantian Idea].

Click here to return to the footnote reference.n99. Id.

Click here to return to the footnote reference.n100. Id. at 486.

Click here to return to the footnote reference.n101. Id. at 486-87.

Click here to return to the footnote reference.n102. Id. at 486.

Click here to return to the footnote reference.n103. Id. Cf. Weinrib, Right and Advantage, supra note 22, at 1291 ("The exclusion from abstract right of duties to act for the benefit of another is categorical. It is an implication of abstract right as it arises out of the conceptual structure of free will.")

Click here to return to the footnote reference.n104. Weinrib, Legal Formalism, supra note 21, at 953-57.

Click here to return to the footnote reference.n105. Id. at 960.

Click here to return to the footnote reference.n106. The scope of Weinrib's formalism is not clear. At times he suggests that it is a quite comprehensive theory, extending to physical objects like tables. See id. at 958-61. At other times he seems to limit the scope of formalism to that which is "essentially conceptual" or "constituted by thought," as law is said to be. See id. at 961-62.

Click here to return to the footnote reference.n107. Weinrib, Understanding Tort Law, supra note 3, at 493-94.

Click here to return to the footnote reference.n108. Id. at 494.

Click here to return to the footnote reference.n109. Weinrib, Kantian Idea, supra note 98, at 485.

Click here to return to the footnote reference.n110. Weinrib, Legal Formalism, supra note 21, at 955. Understood as a methodological presupposition, integrationism calls for "noninstrumental" rather than "instrumental" understandings of legal phenomena (and perhaps of other social phenomena as well). An instrumental understanding requires reference to an end or goal that lies outside the phenomenon to be understood, whereas a noninstrumental understanding sees juridical relationships "in the light of their underlying forms and thus by reference to themselves." Id. at 964. Weinrib accordingly says that formalism, which is itself said to be an integrative notion, id. at 955, "repudiates analysis that conceives of legal justification in terms of some goal that is independent of the conceptual structure of the legal arrangement in question." Id. at 964-65.

Click here to return to the footnote reference.n111. Id. at 995-99; Weinrib, supra note 98, at 491-500.

Click here to return to the footnote reference.n112. Weinrib, Kantian Idea, supra note 98, at 491.

Click here to return to the footnote reference.n113. Id. at 491-501. Weinrib speaks of "the intricate conceptual progression by which law arises inexorably from the structure of willing." Id. at 491.

Click here to return to the footnote reference.n114. Id.

Click here to return to the footnote reference.n115. Weinrib, Understanding Tort Law, supra note 3, at 495.

Click here to return to the footnote reference.n116. Id. at 495-96.

Click here to return to the footnote reference.n117. Id. at 512-13. Cf. id. at 524 ("[The concepts of negligence law] represent negligence as a single sequence that extends from the potentiality to the actuality of injury. Misfeasance and actual causation capture the physical aspect of this sequence, the movement of energy from the initiation to the reception of effects. Reasonable care, proximate cause, and duty comprise the normative aspect of the sequence, the materialization in harm of unreasonably created risk. The notion of an act has both a physical and normative dimension, since the rooting of external effects in the volition makes the sequence amenable to moral assessment.").

Click here to return to the footnote reference.n118. Id. at 513. Cf. Weinrib, Causation and Wrongdoing, supra note 3, at 430-32.

Click here to return to the footnote reference.n119. For a powerful critique of Weinrib's formalism see Leslie Green, Law's Rule, 24 Osgoode Hall L.J. 1023, 1027-34 (1986). See also Stephen R. Perry, Professor Weinrib's Formalism: The Not-So-Empty Sepulchre, 16 Harv. J. L. & Pub. Pol.     (1993, forthcoming). The general project of Kantian rationalism is sympathetically assessed but persuasively criticized in Bernard Williams, Ethics and the Limits of Philosophy 54-70 (1986).

Click here to return to the footnote reference.n121. The terminology of primary and secondary duties is Jules Coleman's, not Weinrib's. See Coleman, Tort Law and the Demands of Corrective Justice, supra note 2, at 367. See also MacCormick, supra note 6, at 219.

Click here to return to the footnote reference.n122. Weinrib, Causation and Wrongdoing, supra note 3, at 437-38.

Click here to return to the footnote reference.n123. Coleman, Property, Wrongfulness and the Duty to Compensate, supra note 22, at 469-70.

Click here to return to the footnote reference.n124. Weinrib, Understanding Tort Law, supra note 3, at 524.

Click here to return to the footnote reference.n125. See Dennis M. Patterson, The Metaphysics of Legal Formalism, 77 Iowa L. Rev. 741 (1992); John Stick, Formalism as the Method of Maximally Coherent Classification, 77 Iowa L. Rev. 773 (1992).

Click here to return to the footnote reference.n126. Weinrib, Causation and Wrongdoing, supra note 3, at 434-35.

Click here to return to the footnote reference.n127. Ernest J. Weinrib, Right and Advantage in Private Law, supra note 22.

Click here to return to the footnote reference.n128. Id. at 1289.

Click here to return to the footnote reference.n129. Id. (quoting G. Hegel, Philosophy of Right para. 36 (Thomas M. Knox trans., 1952)).

Click here to return to the footnote reference.n130. Weinrib, Right and Advantage, supra note 22, at 1288.

Click here to return to the footnote reference.n131. Id. at 1286.

Click here to return to the footnote reference.n132. Id. at 1284.

Click here to return to the footnote reference.n133. This ambiguity is reflected in Weinrib's statement that theorists who advocate advantage-based accounts of private law differ on how advantages are to be characterized. See id. at 1285. Weinrib holds out as possibilities the following: preferences, utilities, wealth, and basic aspects of the good. Basic aspects of the good and utilities are presumably to be understood as intrinsically or ultimately valuable. By contrast wealth is, on the best understanding, of instrumental value only. See Ronald Dworkin, A Matter of Principle 237-66 (1985). Preferences could presumably be understood in any of these three ways, depending on one's moral theory. On the difference between instrumental, intrinsic and ultimate value, see Joseph Raz, The Morality of Freedom 177-78, 200 (1986). It is perhaps unfair of Weinrib to lump together a disparate group of theorists as defenders of a single "advantage model" of private law, where the general notion of advantage that he invokes is not only not employed by them, but is ambiguous and unclear to boot.

Click here to return to the footnote reference.n134. Weinrib, supra note 22, at 1287. Cf. id. at 1291.

Click here to return to the footnote reference.n135. Id. at 1293.

Click here to return to the footnote reference.n136. Id. at 1297.

Click here to return to the footnote reference.n137. Id. For reasons to be explained below, it is by no means clear what the "further argument" to which Weinrib refers could be.

Click here to return to the footnote reference.n138. Id. at 1287 (quoting Hegel, supra note 129, at para. 29).

Click here to return to the footnote reference.n139. Here I am skipping over a number of potential difficulties facing Hegel, Weinrib's interpretation of Hegel, or both. Weinrib maintains, not uncontroversially, that the Hegelian concept of property has no distributive implications. Id. at 1291, 1293-97. For a contrary view see Waldron, supra note 27, at 343-89. Waldron also offers an illuminating discussion of the various problems associated with the notion of placing one's will in a material thing.

Click here to return to the footnote reference.n140. Weinrib, supra note 22, at 1290 (quoting Hegel, supra note 129, para 36) (emphasis in Weinrib).

Click here to return to the footnote reference.n141. Id. at 1290 (emphasis added).

Click here to return to the footnote reference.n142. Id. at 1284.

Click here to return to the footnote reference.n143. Weinrib, supra note 98, at 486.

Click here to return to the footnote reference.n144. Weinrib, supra note 22, at 1297.

Click here to return to the footnote reference.n145. Cf. Lawrence G. Sager, Righting Rights, 10 Cardozo L. Rev. 1311, 1322, 1325-29 (1989).

Click here to return to the footnote reference.n146. See, e.g., Weinrib, supra note 22, at 1292 n.22 (discussing "infringement on personality").

Click here to return to the footnote reference.n147. See, e.g., id. at 1303-04.

Click here to return to the footnote reference.n148. Id. at 1292.

Click here to return to the footnote reference.n149. Id. at 1304-05.

Click here to return to the footnote reference.n150. Id. at 1304.

Click here to return to the footnote reference.n151. Id. at 1305.

Click here to return to the footnote reference.n152. Cf. Warren A. Seavey, Negligence -- Subjective or Objective, 41 Harv. L. Rev. 1, 5-7 (1927); Perry, supra note 22, at 162-63.

Click here to return to the footnote reference.n153. See Stephen R. Perry, Protected Interests and Undertakings in the Law of Negligence, 42 U. of Toronto L.J. 247, 252-62 (1992); cf. Brian Skyrms, Choice and Chance 15 (1966) ("The epistemic probability of a statement is the inductive probability of that argument which has the statement in question as its conclusion and whose premises contain all of our relevant factual knowledge.").

Click here to return to the footnote reference.n154. Will it always? This is far from clear.

Click here to return to the footnote reference.n155. Weinrib, Right and Advantage, supra note 22, at 1306.

Click here to return to the footnote reference.n156. Id. at 1303 (emphasis added).

Click here to return to the footnote reference.n157. Id. Elsewhere, Weinrib has stated that "criminal law falls under corrective justice," adding that mens rea involves a "general wrong" that only state prosecution and punishment can undo. Weinrib, Legal Formalism, supra note 21, at 982-83 n.73. If, as I have argued, wrongs in Weinrib's conception of corrective justice necessarily involve mens rea, it follows that criminal punishment is always the appropriate remedy.

Click here to return to the footnote reference.n158. For historical reasons damage is not the gist of the action with respect to some torts, such as trespass. Even if we suppose that this is not in principle anomalous, it remains the case that battery (tresspass to the person) requires contact with the plaintiff's body. The defendant's attempted assertion of dominion may not have succeeded even to this extent.

Click here to return to the footnote reference.n159. Weinrib, Right and Advantage, supra note 22, at 1284.

Click here to return to the footnote reference.n160. Id. at 1297.

Click here to return to the footnote reference.n161. Id. at 1302. Cf. Weinrib, Causation and Wrongdoing, supra note 3, at 423-24.

Click here to return to the footnote reference.n162. Tony Honore, Responsibility and Luck, 104 Law Q. Rev. 530 (1988).

Click here to return to the footnote reference.n163. Id. at 539.

Click here to return to the footnote reference.n164. Id. at 540.

Click here to return to the footnote reference.n165. Id. at 531.

Click here to return to the footnote reference.n166. Id. at 543.

Click here to return to the footnote reference.n167. Id. at 545.

Click here to return to the footnote reference.n168. Cf. Raz, supra note 133, at 204, 369-70.

Click here to return to the footnote reference.n169. Honore, supra note 162, at 530 (emphasis added).

Click here to return to the footnote reference.n170. Id. at 540.

Click here to return to the footnote reference.n171. Id. at 544.

Click here to return to the footnote reference.n172. Id.

Click here to return to the footnote reference.n173. Bernard Williams, Moral Luck, in Moral Luck 20 (1981); Thomas Nagel, Moral Luck, in Mortal Questions 24 (1979).

Click here to return to the footnote reference.n174. This may require some qualification. Raz notes that it is autonomous agents who identify with their choices, where autonomy constitutes an ideal to be attained. See Raz, supra note 133, at 382.

Click here to return to the footnote reference.n175. He distinguishes agent-regret from regret that a certain state of affairs has come about. Williams, supra note 173, at 27-28. As Williams notes, the focus of agent-regret is on the possibility that one might have acted otherwise.

Click here to return to the footnote reference.n176. Honore, supra note 162, at 539.

Click here to return to the footnote reference.n177. Id. at 544-45.

Click here to return to the footnote reference.n178. Id.

Click here to return to the footnote reference.n179. Id. at 545.

Click here to return to the footnote reference.n180. Recall that outcome-responsibility is not equivalent to fault.

Click here to return to the footnote reference.n181. Honore, supra note 162, at 530.

Click here to return to the footnote reference.n182. Id. at 541.

Click here to return to the footnote reference.n183. L.R. 3 H. L. 330 (1868).

Click here to return to the footnote reference.n184. Honore, supra note 162, at 542.

Click here to return to the footnote reference.n185. These points about harm are made in Raz, supra note 133, at 413-14.

Click here to return to the footnote reference.n186. Pain, like offense to one's sentiments or dignity, involves interference with an interest other than autonomy. But such injuries are generally only harmful in the strict sense if they do affect autonomy. See Raz, supra note 133, at 413-14. It will not be possible to consider in this Article the general question of which interests are protected by moral principles of reparation. For a preliminary discussion of some of the issues, see Perry, supra note 5, at 404-08; Perry, supra note 153, at 252-70, 289-91.

Click here to return to the footnote reference.n187. As we shall see later, this is not the only characteristic of outcome-responsibility, which consequently cannot simply be equated with causation.

Click here to return to the footnote reference.n188. Judith Jarvis Thomson has proposed an argument which in some respects might be thought to resemble the one advanced here. See Judith J. Thomson, Rights, Restitution and Risk 195-207 (1986). Thomson begins with an overtly distributive principle: "What we are concerned with . . . is not blame, but who is to be out of pocket for the costs." Id. at 197. (She also assumes an initial equal distribution of wealth.) Thomson then argues that causation matters -- that it is, in effect, a proper basis for limiting the distributive scheme -- "because if B did not cause . . . A's injury, then B's freedom of action protects him against liability for A's injury." Id. at 202. We thus cannot call upon someone at random to pay B's costs because that would be to interfere without reason with that person's freedom of action. From this Thomson apparently concludes that if B did cause A's injury then she can be held liable for A's costs, at least if B was also at fault. But it does not follow from the proposition that without causation you cannot impose liability, that with causation you can. While it is true that if the argument were successful it would establish that causation is, in a formal sense, a necessary condition of liability, this is compatible with the complete nonexistence of liability: the argument does not establish a substantive connection between liability and causation. Thomson uses freedom of action to try to forge a link between non-harm-causers and non-liability, when it is a link between causers of harm and liability for the harm that is needed. The argument defended here offers such a link in the form of outcome-responsibility, which derives not from freedom of action but from a related, more fundamental source, namely, our sense of our own agency. A different criticism of Thomson's argument is presented in Weinrib, Causation and Wrongdoing, supra note 3, at 411-14. Weinrib remarks that even if it is allowed that causation singles out a causer of harm as a potential bearer of costs, this does not show why fault should not serve that function also, even in the absence of causation. Related objections to Thomson's argument are put forward in Shelly Kagan's interesting discussion in Shelly Kagan, Causation, Liability, and Internalism, 15 Phil & Pub. Aff. 41 (1986). Thomson replies in Judith J. Thomson, A Note on Internalism, 15 Phil. & Pub. Aff. 60 (1986). See also John M. Fischer & Robert H. Ennis, Causation and Liability, 15 Phil. & Pub. Aff. 33 (1986).

Click here to return to the footnote reference.n189. Peter F. Strawson, Freedom and Resentment, 48 Proc. Brit. Acad. 187 (1962).

Click here to return to the footnote reference.n190. Id. at 201.

Click here to return to the footnote reference.n191. Id. at 193.

Click here to return to the footnote reference.n192. Supra note 173.

Click here to return to the footnote reference.n193. This is not necessarily to say, of course, that Strawson was unaware of possible further complexity.

Click here to return to the footnote reference.n194. Strawson, supra note 189, at 208. Thomas Nagel takes issue with Strawson's claim that it is not humanly possible to maintain indefinitely what the latter calls "a sustained objectivity of inter-personal attitude." Thomas Nagel, The View from Nowhere 124-26 (1986). Strawson points to a second class of considerations, distinct from those discussed in the text, which inhibit the reactive attitudes in a wholly different way. Considerations of this sort, which include insanity and compulsive behavior, inhibit the reactive attitudes in general, as regards the agent in question, and not just on a particular occasion. They promote what Strawson calls an objective view of the individual as someone who is not a morally responsible agent and who simply poses problems of intellectual understanding, control and treatment. Strawson says that it is possible occasionally to adopt the objective attitude even in the case of normal agents, as a matter of intellectual curiosity or to obtain relief from the "strains of involvement," but maintains that we cannot do this altogether or for long. This supports his main thesis that acceptance of some form of determinism would not and could not prevent us from taking up the whole range of reactive attitudes, and so would not lead us to modify our sense of our own freedom. It is this thesis that Nagel wishes to question when he says that once we have become capable of an external view it is not difficult to maintain a "radically external view of human life," one which reduces the idea of free agency to unintelligibility. But Nagel should not, I think, be understood as denying that moral justification and criticism must adopt the internal view, if we are to engage in such justification and criticism at all. "Internal" here means internal to the Strawsonian web of evaluative attitudes, not internal to the point of view of individual persons. Nagel does develop an interesting account of objectivity in ethics, but the objectivity comes from adopting an impersonal standpoint that leaves our sense of agency intact, rather than a radically external standpoint. This is consistent with Strawson's account of justification, since the network of attitudes he describes includes those he calls impersonal, vicarious, or disinterested.

Click here to return to the footnote reference.n195. Cf. Raz, supra note 133, at 288-89 ("Pervasive and unshakeable features of human practical thought need no justification, though they call for an explanation.").

Click here to return to the footnote reference.n196. Cf. Ernest J. Weinrib, Enduring Passion, 94 Yale L.J. 1825, 1840 (1985): 
Kant's focus is not on the discerning of a fundamental human identity which can then be mysteriously seeded with normativity. His focus is instead on the nature of normativity itself, which he regards as the correspondence of action with its own intelligible nature; he is concerned with man only insofar as the conditions of normativity are embodied in him.



Click here to return to the footnote reference.n197. There is a connection here with Coleman's notion of making one's mark on the world. See Coleman, supra note 4, at 443 n.24.

Click here to return to the footnote reference.n198. Hart & Honore, supra note 55, at 28-41. On the importance of actual intervention in the physical world to our understanding of causation, see also Mackie, supra note 54, at 56-57.

Click here to return to the footnote reference.n199. Supra note 58.

Click here to return to the footnote reference.n200. Hart & Honore, supra note 55, at 71-81.

Click here to return to the footnote reference.n201. Id. at 74.

Click here to return to the footnote reference.n202. Cf. Mackie, supra note 54, at 128.

Click here to return to the footnote reference.n203. Hart & Honore, supra note 55, at 77.

Click here to return to the footnote reference.n204. Id. at 78.

Click here to return to the footnote reference.n205. Another example of a coincidence that Hart and Honore give is this. As a result of an earlier act of speeding A, who is now driving his car at a normal speed, is at the exact location where B dashes in front of the car and is killed. Id. at 103. Here I would suggest that A is outcome-responsible for the death so far as his act of driving at the time of the accident is concerned (even though he may not then have behaved faultily in any way), but probably not as regards his earlier act of speeding (even though that act was a faulty one). Much depends, however, on how much earlier the speeding took place.

Click here to return to the footnote reference.n206. Re Polemis and Furness, Withy & Co. Ltd., [1921] 3 K.B. 560, overruled by The Wagon Mound No. 1, [1961] App. Cas. 388.

Click here to return to the footnote reference.n207. Blame is assignable not just where the agent acts with knowledge of fairly specific facts, say that a certain action will or might cause a certain harm. It is also assignable where the agent knows that he ought not to act without first obtaining knowledge of the specific facts (knows that he ought to know, for short).

Click here to return to the footnote reference.n208. There is an important connection here with what Bernard Williams calls the ideal of the mature agent, by which he means an agent who is rational and who deliberates about his own actions and desires, but who realizes that he has not deliberatively constituted his own character and identity as an agent, at least not entirely. See Bernard Williams, Voluntary Acts and Responsible Agents, 10 Oxford J. Legal Stud. 1 (1990). Williams says that the fact that the mature agent knows that his character is not his own deliberative construct does not mean that that he cannot acknowledge actions that flow from it as his, and continues: "He will be able to acknowledge . . . that he can be as responsible for some things that he did not intend as he is things he did intend, and in ways that have nothing to do with the law of negligence. For him, to be responsible is not simply to be properly held responsible by others, by the institutions of control and cohesion, but to hold himself responsible. He acknowledges that in that sense, responsibility can reach beyond the voluntary (and not merely in the ways acknowledged in the law of negligence)." Id. at 10.

Click here to return to the footnote reference.n209. Bolton v. Stone, [1951] App. Cas. 850 (H.L.).

Click here to return to the footnote reference.n210. Epstein, supra note 38, at 170-71.

Click here to return to the footnote reference.n211. Hart & Honore's example of the intervening actor who deliberately starts a forest fire shows that reasonable foreseeability is not a sufficient condition for the existence of outcome-responsibility. So far as reparation is concerned it is probably a necessary condition, although this statement might require some qualification with respect to, for example, the principles underlying the thin skull doctrine in negligence law.

Click here to return to the footnote reference.n212. Raz, supra note 133, at 166.

Click here to return to the footnote reference.n213. "Harm" is being used here in the loose sense I spoke of earlier, as a synonym for "injury." It should be noted that an individual might in one sense have violated the obligation not to culpably harm, but still be found not to meet the minimal requirements of being a morally responsible person. This would be to find that one among Strawson's second class of considerations capable of inhibiting the reactive attitudes applied. See supra note 194. The issue is a difficult one, but there is a case to be made that the distributive argument for fault does not apply under such circumstances, since the individual is properly regarded as someone to be controlled and treated and not as a full member of the moral community. To some extent, at least, this is recognized in tort law by the basic requirement of capacity.

Click here to return to the footnote reference.n214. If the victim was not totally innocent in the production of the loss then he should bear a proportional share in its distribution.

Click here to return to the footnote reference.n215. See further Perry, supra note 5, at 405-06. As I note there, some interests may not be important enough to justify an obligation of reparation even when they are damaged culpably. (The obligation not to culpably harm, as defined in the text, is limited to the interests of person and property.)

Click here to return to the footnote reference.n216. It should be remembered that for Weinrib's theory of abstract right culpable wrongdoing must involve a culpable show of disrespect for moral personality and not, as I am assuming, a culpable disregard for interests as such. The two conceptions of culpability must be kept distinct, even though one way of showing disrespect is to disregard interests.

Click here to return to the footnote reference.n217. This is, essentially, the understanding of the objective negligence standard which is generally accepted in England and Canada. See Bolton, [1951] App. Cas. at 867 (H.L.) (opinion of Lord Reid). Lord Reid refers to unacceptably high risks as "substantial" risks. See also Perry, supra note 22, at 169-71. I shall not here offer an argument in favor of this understanding of the negligence standard, which can be contrasted with the Learned Hand test and similar formulations that rely on a balancing of individual interests. A nonbalancing conception of negligence is adopted in Fletcher, Synthesis in Tort Theory, supra note 75, at 81-82. For some years Weinrib defended the Learned Hand test, but has recently come around to a nonbalancing view as well. See, e.g., Weinrib, Right and Advantage, supra note 22, at 1306.

Click here to return to the footnote reference.n218. Coleman, Tort Law and the Demands of Corrective Justice, supra note 2, at 370.

Click here to return to the footnote reference.n219. Problems do arise in certain kinds of cases, for example necessity cases in which A, in order to save his own life, destroys property belonging to B. Most courts and commentators think that even though A did nothing wrong, he still owes reparation to B. See, e.g., Vincent v. Lake Erie Transportation Co., 124 N.W. 221 (Minn. 1910). There are obvious difficulties in saying of such cases that A did what he ought not to have done when he destroyed B's property. But the volitionist/distributive argument, with its emphasis on a retrospective, comparative inquiry, does not force us to say this, and hence it can handle such cases fairly well. See further Perry, supra note 5, at 403 n.5.

Click here to return to the footnote reference.n220. Even this may be problematic, given the unclarity of the idea that an agent could have acted otherwise than as he did. Cf. Williams, supra note 119, at 175.

Click here to return to the footnote reference.n221. Weinrib, Understanding Tort Law, supra note 3, at 519.

Click here to return to the footnote reference.n222. Weinrib's view of the objective standard is of course quite different from the one advocated here, since for him it constitutes a norm that characterizes certain kinds of action as wrongful independently of the reparative inquiry: that inquiry always moves from the breach of a primary obligation to a secondary, remedial obligation, and not in the reverse direction. See, e.g., Weinrib, The Special Morality of Tort Law, supra note 3, at 409: "The obligation to compensate is the juridical reflex of an antecedent obligation not to wrong." We have already seen that Weinrib has difficulty establishing that a failure to comply with the objective standard constitutes wrongdoing in the sense required by abstract right. The present point is related, in that one would expect the content of an independent norm forbidding wrongdoing to be capable of fairly explicit formulation in advance. Otherwise, how is an agent supposed to be able even to try to comply with the norm?

Click here to return to the footnote reference.n223. Cf. Williams, supra note 173, at 24. Similar themes can be found in Williams, supra note 119, at 174-96.

Click here to return to the footnote reference.n224. The distinction between ex ante and ex post perspectives is discussed in Fletcher, Synthesis in Tort Theory, supra note 75.

Click here to return to the footnote reference.n225. I deal with this question in Perry, supra note 9.






[PHIL 1A] [Syllabus] [Handouts] [Home] [Bio] [CV] [PHIL DEPT.] [E-MAIL]

Send comments to: Andreas Teuber
URL:   http://phils7.dce.harvard.edu/
Last Modified: 01/16/05
Instructor's Toolkit
Copyright © The President and Fellows of Harvard College

 

S“.?FxЇмф§ўџ €ЖРСЯню  >UV[§§§§§§§§§§§§§§§§§§§§§§§ћћћћћћ.?FxЇмф§ўџ €ЖРСЯню  >UV[mŽtŽyŽzŽ{Ž|Ž}Ž†Ž‡ŽˆŽ‰ŽŠŽ‹Ž”Ž•Ž–ŽЛŽдŽїŽ!HŽв_‘’œБВЬB‘T‘U‘Ѓ‘И‘Й‘э‘’N’П’и’й’ю’ “*“@“A“I“Q“R“S“§§§§§§§§§§§§§§§§§§§§§§§§њњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњ R[mŽtŽyŽzŽ{Ž|Ž}Ž†Ž‡ŽˆŽ‰ŽŠŽ‹Ž”Ž•Ž–ŽЛŽдŽїŽ!HŽв_‘’œ§§§§§§§ћћћћћћћћћћћћћћћћћћћћћћБВЬB‘T‘U‘Ѓ‘И‘Й‘э‘’N’П’и’й’ю’ “*“@“A“I“Q“R“S“§§§§§§§§§§§§§§§§§§§§§§§ћАа/ Ар=!А"А# $ %А [4@ёџ4NormalCJOJPJQJmH <A@ђџЁ<Default Paragraph Font4Z`ђ4 Plain Text CJOJQJS. џџџџ^!џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ  џџ џџ  џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ  џџ! џџ" џџ# џџ$ џџ% џџ& џџ' џџ( џџ) џџ* џџ+ џџ, џџ- џџ. џџ/ џџ0 џџ1 џџ2 џџ3 џџ4 џџ5 џџ6 џџ7 џџ8 џџ9 џџ: џџ; џџ< џџ= џџ> џџ? џџ@ џџA џџB џџC џџD џџE џџF џџG џџH џџI џџJ џџK џџL џџM џџN џџO џџP џџQ џџR џџS џџT џџU џџV џџW џџX џџY џџZ џџ[ џџ\ џџ] џџ^/ ŸF(8ƒG%WаfŽv&†Џ•$Ѕ•Д№У.гЛт&ђeу# к/9?šNS^Ёmѓ|MŒс›,ЋjК’ЩГи]шјэO'Е6њE•U e–t „p“=ЃГЌТyвэсyё‘ћo/Ў>SNз]dmы|ПŒд› ЋџЙ­Щъиш4ї'ЅЧ$п3CR0ap'_ŽJ‚ЌQЛ_ЪAй шї($Ё3єBэQJaqp˜–ŠSџџџџџџџџџ џ џ џ џ џџџџџџџџџџџџџџџџџџџ џ!џ"џ#џ$џ%џ&џ'џ(џ)џ*џ+џ,џ-џ.џ/џ0џ1џ2џ3џ4џ5џ6џ7џ8џ9џ:џ;џ<џ=џ>џ?џ@џAџBџCџDџEџFџGџHџIџJџKџLџMџNџOџPџQџRџSџTTuUEfVЬVWyGX€8Y#)Zќ[е \%]S“’[S““•–S“”џџ iMac User( :Users:andreasteuber:Desktop:perry.htmlџ@€ŠŠєХ ŠŠЩ%§ЛS@@GTimes New Roman5€Symbol3 Arial3Times? Courier New"1ˆаhУMВІЧMВІхЭž•^XЧ ^$ЅРДД€0HЁPџџ< iMac User iMac Userўџ р…ŸђљOhЋ‘+'Гй0XˆœЈМШи ьј   ,8@HP'<ss iMac User MacNormale iMac User 1acMicrosoft Word 8.0d@Œ†G@r‘ZШLЧ@ŠžщШLЧ^хЭž•ўџ еЭеœ.“—+,љЎDеЭеœ.“—+,љЎ<ј hpŒ”œЄ ЌДМФ Ь к'Brandeis UniversityЧ XHЁb < Title˜ 6> _PID_GUID'AN{AECCFD01-B891-11DB-AD42-FED9786531D0}  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ЁЂЃЄЅІЇЈЉЊЋЌ­ЎЏАБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯабвгдежзийклмнопрстуфхцчшщъыьэюя№ёђѓєѕіїјљњћќ§ўџ      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ЁЂЃЄЅІЇЈЉЊЋЌ­ЎЏАБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯабвгдежзийклмнопрстуфхцчшщъыьэюя№ёђѓєѕіїјљњћќ§ўџ      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ЁЂЃЄЅІЇЈЉЊЋЌ­ЎЏАБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯабвгдежзийклмнопрстуфхцчшщъыьэюя№ёђѓєѕіїјљњћќ§ўџ      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ЁЂЃЄЅІЇЈЉЊЋЌ­ЎЏАБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯабвгдежзийклмнопрстуфхцчшщъыьэюя№ёђѓєѕіїјљњћќ§ўџ      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ЁЂЃЄЅІЇЈЉЊЋЌ­ЎЏАБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯабвгдежзийклмнопрстуфхцчшщъыьэюя№ёђѓєѕіїјљњћќ§ўџ      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—ўџџџ™š›œžŸўџџџЁЂЃЄЅІЇўџџџЉЊЋЌ­ЎЏўџџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџНўџџџўџџџўџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџRoot Entryџџџџџџџџ РFйnŸLЧП€1Tableџџџџџџџџџџџџ˜WordDocumentџџџџџџџџ. SummaryInformation(џџџџ DocumentSummaryInformation8џџџџџџџџџџџџЈCompObjџџџџXObjectPoolџџџџџџџџџџџџйnŸLЧйnŸLЧџџџџџџџџџџџџўџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџўџџџџџ РFMicrosoft Word DocumentўџџџNB6WWord.Document.8Root Entryџџџџџџџџ РF“_ИЃLЧХ@1Tableџџџџџџџџ˜WordDocumentџџџџџџџџе6 SummaryInformation(џџџџ ˆ‚ƒ„…†‡ˆ‰Š‹ŒŽоџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџ™š›œžŸўџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџ§џџџџџџџџџџџџџџџџџџџџџџџТ§џџџўџџџўџџџЦФўџџџШЩЪЫЬЭЮЯабўџџџгдўџџџз§џџџийклмнпрстуфхвџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџDocumentSummaryInformation8џџџџџџџџџџџџдCompObjџџџџXObjectPoolџџџџџџџџџџџџйnŸLЧйnŸLЧ0TableџџџџџџџџџџџџЧСўџџџ ўџџџ ўџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџ{AECCFD01-B891-11DB-AD42-FED9786531D0}ўџ р…ŸђљOhЋ‘+'Гй0XˆœЈМШи ьј   ,8@HP'<ss iMac User MacNormale iMac User 2acMicrosoft Word 8.0d@^аВ@r‘ZШLЧ@Ъ…ЭLЧ^хЭž•ўџџџџџ РFMicrosoft Word DocumentўџџџNB6WWord.Document.8ўџ еЭеœ.“—+,љЎDеЭеœ.“—+,љЎ<ј hpŒ”œЄ ЌДМФ Ь к'Brandeis UniversityЧ XHЁb < Title˜ 6> _PID_GUID'AN [4@ёџ4NormalCJOJPJQJmH <A@ђџЁ<Default Paragraph Font4Z`ђ4 Plain Text CJOJQJ/. џџџџ^!џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ  џџ џџ  џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ џџ  џџ! џџ" џџ# џџ$ џџ% џџ& џџ' џџ( џџ) џџ* џџ+ џџ, џџ- џџ. џџ/ џџ0 џџ1 џџ2 џџ3 џџ4 џџ5 џџ6 џџ7 џџ8 џџ9 џџ: џџ; џџ< џџ= џџ> џџ? џџ@ џџA џџB џџC џџD џџE џџF џџG џџH џџI џџJ џџK џџL џџM џџN џџO џџP џџQ џџR џџS џџT џџU џџV џџW џџX џџY џџZ џџ[ џџ\ џџ] џџ^I Ъ}(=8ЗGXWgУv]†х•SЅЧД&Фbгѓт\ђ—Y 0m?ХN‹^дm(}{Œœ`Ћ•КХЩшиŠшBј €4'щ6 FШU>eПt;„Њ“sЃIГхТЌв(тЁёФ5Ѓ//ф>‡N ^m"}ђŒџ›=Ћ,КнЩй<шdї[мћ$4DCNRbaDpIŽ}ВЌ‰Л†ЪeйQшJїMKW$Ш3%CRuaЄpа—Š/џџџџџџџџџ џ џ џ џ џџџџџџџџџџџџџџџџџџџ џ!џ"џ#џ$џ%џ&џ'џ(џ)џ*џ+џ,џ-џ.џ/џ0џ1џ2џ3џ4џ5џ6џ7џ8џ9џ:џ;џ<џ=џ>џ?џ@џAџBџCџDџEџFџGџHџIџJџKџLџMџNџOџPџQџRџSџTўtUђeVVW$GX+8Yд(ZЅ[y \].?FxЇмф§ўџ €ЗСТаоя >UVIŠPŠUŠVŠWŠXŠYŠbŠcŠdŠeŠfŠgŠpŠqŠrŠ—ŠАŠгŠ§Š$‹j‹Ў‹№‹;ŒmŒnŒxŒyŒŒŽŒЈŒ01”•Що*Ž›ŽДŽЕŽЪŽшŽ%-.1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!uЏР!uЏР!uЏР!uЏР!uЏР!uЏР!uЏР!uЏР!uЏР!uЏР!uЏР!uЏР!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!ћ1Р!uЏ.?FxЇмф§ўџ €ЗСТаоя >U[IŠ1˜€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜@€€˜€€ž€€˜€€ž€€3 ’[l/ š2 “•–šS“”џџUnknown iMac Userџџ iMac User( :Users:andreasteuber:Desktop:perry.htmlџ@€ББєХ ffБ`b”ј|-Б TkŠ‹./A. @x. A>1 @€1 @82 @Ј @AЦ2 A@@^@@Є& @GTimes New Roman5€Symbol3 Arial3Times? Courier New"1ˆаhУMВІшMВІхЭž•^XЧ ^$ЅРДД€0HЁPџџ< iMac User iMac User="#FFFFEE"> <center> <a HREF="http://www.brandeis.edu"><img ALT="Brandeis University Web Stite" BORDER=0 SRC="new-brandeis-seal.gif"></a> <br><h1>The Moral Foundations of Tort Law</h1> </center> </td> </tr> </table> <hr WIDTH="530" ALIGN="left"> <div class="bodytext" style="margin-left:5px"> <!-- if _printable -->br"/><br class="br"/>SYMPOSIUM:l/ ж/ ц/ 0 0 0 0 0 20 1 1 Œ1  1 Ђ1 О1 к1 ќ1 82 :2 <2 š2 §§§§§§§§§§§§§§§§§ћћћьЅСs ПS“jbjbР Р 6 ЊkЊk/џџџџџџ]ииииииLвввв о$P S*88888888ЄІІІІІІ,}єqPви88888вџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџџ  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€8ии88*8888и8и8ЄьАœАииии8Є8l8ЄииЄ N§ђСP ‚в8Є Philosophy of Law
Brandeis University Web Stite

Corrective Justice and Formalism



Copyright (c) 1992 Iowa University
Iowa Law Review

January, 1992

77 Iowa L. Rev. 449

LENGTH: 37639 words



SYMPOSIUM:: The Moral Foundations of Tort Law +



+ Copyright 1992 Stephen R. Perry. All rights reserved.

NAME: Stephen R. Perry *

BIO:

* Associate Professor, Faculty of Law, McGill University. I would like to thank the participants in the University of Iowa College of Law Conference on Formalism and Corrective Justice for their discussion of the version of the Article that I presented there. I am particularly indebted to Jules Coleman for the many helpful conversations I had with him before, during and after the Conference, and to Ernest Weinrib, who was kind enough to send me extensive written comments on my Conference draft. I have also benefited from comments received when a later version of the Article was presented to a Legal Theory Workshop at Columbia Law School.

SUMMARY:
  ... Theories in the second category are based on individual moral rights, and are often identified by the label of "corrective justice." ... But it is evident that Aristotle thought the principle of corrective justice applied, and hence correlative rights and obligations of reparation arose, even in situations where it could not be said in any ordinary sense that the gainer's gain equalled the loser's loss. ... The argument is noteworthy nonetheless because it suggests that the localized scheme of distributive justice should be expanded so as to take account of the moral worth of actions that in no way causally contributed to the harm of which the victim is now complaining. ... The difficulty that we discovered with Weinrib's characterization of intentional wrongdoing emerges even more strongly here, however, because the risk with which tort law is concerned is risk of harm, in the sense of interference with an interest, and abstract right has no place for harm thus understood. ... But the volitionist argument for strict liability comes closer to forging the necessary connection by recogS“3 .?FxЇмф§ўџ €ЖРСЯню  >UV[§§§§§§§§§§§§§§§§§§§§§§§ћћћћћћ.?FxЇмф§ўџ €ЖРСЯню  >UV[mŽtŽyŽzŽ{Ž|Ž}Ž†Ž‡ŽˆŽ‰ŽŠŽ‹Ž”Ž•Ž–ŽЛŽдŽїŽ!HŽв_‘’œБВЬB‘T‘U‘Ѓ‘И‘Й‘э‘’N’П’и’й’ю’ “*“@“A“I“Q“R“S“§§§§§§§§§§§§§§§§§§§§§§§§њњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњњ R[mŽtŽyŽzŽ{Ž|Ž}Ž†Ž‡ŽˆŽ‰ŽŠŽ‹Ž”Ž•Ž–ŽЛŽдŽїŽ!HŽв_‘’œ§§§§§§§ћћћћћћћћћћћћћћћћћћћћћћБВЬB‘T‘U‘Ѓ‘И‘Й‘э‘’N’П’и’й’ю’ “*“@“A“I“Q“R“S“z. œ. Њ. / l/ §§§§§§§§§§§§§§§§§§§§§§§ћ§§§§§Аа/ Ар=!А"А# $ %А<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML//EN"> <html LANG="en"> <head> <title> Philosophy of Law</title> <meta NAME="generator" CONTENT="BBEdit 5.1.1"> <meta HTTP-EQUIV="Content-Type" CONTENT="text/html"> </head> <body BGCOLOR