The Morality of Strict Liability

THE MORALITY OF STRICT TORT LIABILITY JULES L. COLEMAN* Accidents occur; personal property is damaged and sometimes is lost altogether. Accident victims are likely to suffer anything from mere bruises and headaches to temporary or permanent disability to death. The personal and social costs of accidents are staggering. Yet the question of who should bear these costs has turned the heads of few philosophers and has occasioned surprisingly little philosophic discussion. Perhaps that is because the answer has seemed so obvious; accident costs, at least the nontrivial ones, ought to be borne by those at fault in causing them.' The requirement of fault at one time appeared to be so deeply rooted in the concept of per- sonal responsibility that in the famous Ives2 case, Judge Werner was moved to argue that liability without fault was not only immoral, but also an unconstitutional violation of due process of law. Al* Ph.D., The Rockefeller University; M.S.L., Yale Law School. Associate Professor of Philosophy, The University of Wisconsin-Milwaukee. The author wishes to acknowledge the enormous impact that Joel Feinberg and Guido Calabresi have had on the philosophic and legal ideas set forth in this Article, and to express appreciation to Mitchell Polinsky, Bruce Ackerman, and Robert Prichard for their valuable assistance. This Article constitutes part of the author's forthcoming book, RESPONSIBIITY AND Loss ALLOCATION OF THE LAW OF TORTS (Ed.). 1. The term "fault" takes on different shades of meaning, depending upon the legal context in which it is used. For example, the liability of an intentional or reckless tortfeasor is determined differently from that of a negligent tortfeasor. Use of the term here is limited to determinations of fault in the context of accidents resulting from negligent conduct. Any such fault imputation can be viewed as consisting of two components, one relating to the act and its relationship to an appropriate standard of behavior, and the other relating to the purported responsibility of the actor for the alleged defect in his action. Coleman, On the Moral Argument for the Fault System, 71 J. PHIL. 473 (1974) [hereinafter cited as Coleman, The Fault System]. For a general and helpful discussion of what here is called the fault principle see J. FEINBERG, Sua Culpa, in DOING AND DESERVING 187 (1970). A similar position is taken in R. KEEToN, LEGAL CAUSE IN THE LAW OF ToRTs (1963), which unfortunately is no longer in print. Guido Calabresi recently has put forth a different view. See G. CALABRES1, CONCERNING CAUSE (- ) (to be published by Indiana University Press); Calabresi, Concerning Cause and the Law of Torts: An Essay for Harry Kalven, Jr., 43 U. CH. L. REv. 69 (1975). 2. Ives v. South Buffalo Ry., 201 N.Y. 271, 94 N.E. 431 (1911). 3. In les, the Court of Appeals of New York castigated a 1909 New York workmen's compensation law as "plainly revolutionary" and "totally at variance with the common-law theory of the employer's liability." Id. at 285, 288, 94 N.E. at 436, 437. The court went on to hold the law void as a "taking of property without due process of law," id. at 317, 94 N.E. at 448, which violated both the federal and state constitutions. In justifying its decision, the court stressed that strict liability was a creature virtually totally unknown to the common law: When our Constitutions were adopted, it was the law of the land that no man HeinOnline -- 18 Wm. & Mary L. Rev. 259 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 though Judge Werner's arguments long since have lost whatever appeal they might have had within the legal community, the view persists among philosophers that strict liability is an unjust or immoral standard of responsibility in torts. This Article will identify what that immorality is not and what it might be. I In his great work, THE COMMON LAW,' Oliver Wendell Holmes wrote that a loss ought to lie where it falls, that is, on victims, in the absence of a compelling reason for shifting it to another party.5 One could, of course, maintain precisely the opposite initial liability rule: In the absence of a sound reason for leaving a loss with the victim it ought to be transferred to the injurer.6 Indeed, in products liability as well as in other areas of tort law, this is the initial liability rule we have adopted.' As a general rule, however, losses in tort litigation lie where they fall, thus creating a burden on victimplaintiffs to show why they ought not remain there. Holmes believed that the justification for this initial liability rule is that it is administratively less cumbersome and costly than other alternatives., Thus, the primary liability rule is rooted in considerations of administrative cost avoidance. But losses do not always remain where they who was without fault or negligence could be held liable in damages for injuries sustained by another. That is still the law, except as to the employers enumerated in the new statute ... Id. at 293, 94 N.E. at 439. That the court overstated this point is plain upon consideration of the historical roots of the concept of strict liability. Apparently the early common law of tort liability focused primarily on the undeserved damage suffered by victims rather than on the fault or negligence of injurers. Although it was never universally true that one acted at his peril and was responsible for any and all harm resulting from his actions, liability commonly was imposed without regard to the moral guilt or innocence of the injurer. The common law gradually evolved toward recognition of "fault" as a basis of liability, but it was not until the close of the nineteenth century that this tendency reached the peak epitomized by the Ives decision. See W. PRossER, HANDBOOK OF THE LAW OF Toim §§ 4, 75 (4th ed. 1971). Even in this period, however, the concept of strict liability was by no means a dead issue. See, e.g., Rylands v. Fletcher, [1868] L.R. 3 H.L. 330. 4. 0. HOLMES, THE COMMON LAW (Howe ed. 1963). 5. Id. at 76-77. 6. A general theory of initial injurer liability has been defended by Richard Epstein. See Epstein, A Theory of Strict Liability, 2 J. LEGAL STDIES 151 (1973). Epstein's analysis draws heavily from a prior exhaustive treatment of conceptual and practical aspects of legal liability in tort and other areas of the law in H.L.A. HAm" & A.M. HonoR., CAUSATION IN THE LAW (1959). 7. See PROSSER, supra note 3, §§ 75-81, 98. 8. HOLMES, supra note 4, at 76-77. HeinOnline -- 18 Wm. & Mary L. Rev. 260 1976-1977 1976] STRICT TORT LIABILITY have fallen; nor, apparently, should they. It is not surprising, there- fore, that what has come to be called tort theory is comprised largely of attempts to fashion general principles of compensation popularly termed liability rules.' The following arguments, alone and in combination, have been advanced to justify a system of liability rules. (1) Under a retribu- tive rationale, accident costs are shifted from victims to injurers to penalize blameworthy or faulty injuries.1 On this theory of liability, compensation is likened to punishment or penal sanctions and is said to be required by our notions of retributive or corrective justice. 9. The term "liability rule" is most closely associated with a recent article discussing various theories available to protect "entitlements," i.e., to decide which side should prevail whenever the interests of two or more people conflict. Calabresi & Malamed, PropertyRules, Liability Rules, and Inalienability:One View of the Cathedral,85 HAmv. L. Rav. 1089 (1972). The authors argue that a distribution of initial entitlements to holdings can be protected in one of three ways. An entitlement can be protected by a property rule so that title to it can be transferred to another by a sale at an agreed price. It often is too costly, inconvenient, and occasionally impossible to contract for transfer of an entitlement at mutually acceptable prices, however. Therefore, it may be desirable to protect entitlements by a liability rule, which enables one to effect the partial or complete transfer of an entitlement at an objectively set price. For example, suppose that X operates a polluting feedlot and that Yis a homeowner in the neighboring residential community. After consulting criteria of economic efficiency and wealth distribution we might decide that Y ought to be entitled to be free from unnecessary pollution generated by X's feedlot. To give effect to our decision to protect Y's interest we award him the requisite initial entitlement. Should X's feedlot pollute Y's air space in violation of Y's protected interest, we then would require X to compensate him. In other words, we make X's conduct tortious and hold him liable in damages to Y so that he can pollute only at the price of rendering compensation to Y. The damages X is required to pay Y are set objectively, usually by a court of law. In this way, liability rules protect entitlements. Of course, an entitlement can be protected by both property and liability rules. When an entitlement is protected by what Calabresi and Malamed call an inalienability rule, the holder of it is denied the right or freedom to negotiate its sale. For example, the right to freedom from indentured servitude is one that we might want to protect by an inalienability rule forbidding the sale of one's freedom. The right to exercise one's franchise might be another such entitlement. Id. at 1105-15. The importance of the Calabresi-Malamed analysis is that on the one hand, it details the diversity of protective devices for securing entitlements while on the other, it suggests a fundamental relationship between the foundations of tort and property law. The need to provide legal rules to effect transfers of entitlements also suggests the role of a legal system in overcoming what economists call market failures. Id. at 1109-10 & n.39. Of additional conceptual interest to philosophers in the Calabresi-Malamed scheme is the role of the criminal law primarily as a means for securing the integrity of the market and private law (property, tort, contract) approach to protecting and transferring entitlements. Id. at 1124-27. For a further discussion of Calabresi's position and its consequences in political theory, see G. Calabresi, Torts-The Law of the Mixed Society (.) (unpublished manuscript in Yale Law library). For a general criticism of the Calabresi-Malamed analysis see note 55 infra. 10. For a detailed discussion of the limits of retributive theories of torts, see Coleman, The Fault System, supra note 1. HeinOnline -- 18 Wm. & Mary L. Rev. 261 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 (2) Under a compensatory rationale, faulty injurers are liable for their victims' losses, not to insure that those at fault are punished, but to see to it that fault-free victims are compensated for what are undeserved or otherwise unjustifiable losses. On this view, principles and ideals of compensatory justice rather than those of retributive justice are the touchstone of a system of liability rules." (3) Under a risk-spreading argument, accident costs ought to be allocated either to spread them maximally over persons and time or to insure that those individuals best able to bear them do so.12 (4) Under a deterrence argument, accident costs should be allocated to provide incentives for the party best able to avoid the accident to do so, or, if we are unaware or unsure of who is the best or cheapest accident-avoider, the costs should be borne by that party best able to decide if the accident ought to be avoided; that is, if it is worth its costs.' The traditional view is that whereas the retributive and compensatory rationales appear to require both fault-based liability rules and a case by case determination of liability, the risk-spreading and deterrence justifications for a system of liability rules require nei11. For a general discussion of the relationship between compensatory claims and tort law see Coleman, Justice and the Argument for No-Fault, 3 Soc. THEORY & PRAC. 161 (1974) [hereinafter cited as Coleman, Justice]; Nickel, Justice in Compensation, 18 WM. & MARY L. REv. 379 (1976). 12. The "risk spreading" analysis of tort law is well presented in the writings of Fleming James. See, e.g., James, Accident Liability Reconsidered: The Impact of Liability Insurance, 57 YALE.L.J. 549 (1948). The general argument is canvassed and analyzed carefully in G. CALABRESI, THE COSTS OF AccIDENTS, A LEGAL AND ECONOMIC ANALYSIS 37-67 (1970). 13. The cost-minimization theory of torts, namely that the primary goal of tort law is to allocate losses as incentives to minimize the sum of the costs of accidents and the costs of avoiding them, has been given its most thorough treatment in the works of lawyers trained in economics. The essential essays include Calabresi & Hirschoff, Toward a Test for Strict Liability in Torts, 81 YALE L.J. 1055 (1972); Posner, Strict Liability: A Comment, 2 J. LEGAL STUDIES 205 (1973); Posner, A Theory of Negligence, 1 J. LEGAL STUDIES 29 (1972). Although both Posner and Calabresi agree that cost avoidance (or efficiency in the economic sense) is the fundamental goal of tort law, they differ with respect to whether such a result is best fostered by a system of strict liability. Calabresi appears optimistic that a strict liability system is more likely than a negligence system to minimize the sum of accident costs and accident avoidance costs. Calabresi & Hirschoff, supra, at 1084. Posner, on the other hand, has concluded that economic theory provides no basis for preferring strict liability to negligence. Posner, 2 J. LEGAL STUDIES, supra, at 221. For an essential introduction to the economic model of law, see Coase, The Problem of Social Cost, 3 J. LAW & EcoN. 1 (1960), and R. POSNER, ECONoMic ANALYSIS OF LAW (1972). Mitchell Polinsky has written an exemplary review of Posner's book and of the foundations of the economic analysis of legal rules generally. Polinsky, Economic Analysis as a Potentially Defectiue Product:A Buyer's Guide to Posner's Economic Analysis of Law, 87 HA~v. L. REV. 1655 (1974). For another useful review of both the book and the general methodology, see Baker, The Ideology of the Economic Analysis of Law, 5 PHIL. & PUB. AFFAIRS 3 (1975). HeinOnline -- 18 Wm. & Mary L. Rev. 262 1976-1977 1976] STRICT TORT LIABILITY ther. That is, to be liable, the party with the deepest pocket need not be the individual responsible for the accident, and the individual best suited to avoid the accident or best able to judge if it ought to be avoided need not be the party at fault in causing it to be liable for its costs. For example, factory' owners, by introducing safety measures, might prove to be better and cheaper accidentavoiders than their careless and negligent employees. It is entirely possible that the goals of deterrence, cost-spreading, and wealth distribution may involve fault-based liability rules, for example, when the party at fault is also the best risk-spreader or costavoider. However, unless one equates being at fault with being the cheapest cost-avoider, as Guido Calabresi appears to do," the goal of accident cost-avoidance does not require that liability for the costs of accidents be determined by the criterion of fault. The same can be said of the other instrumentalist goals of tort law. I have argued at length elsewhere that we need not adopt a fault system or a case by case determination of liability to secure the goals of penalization and compensation in torts.'" In short neither considerations of justice nor utility require that liability be based on fault. This argument that justice does not require fault-based liability rules has two thrusts. On the one hand, if there is sufficient fault on the injurer's part, principles of retributive justice may require that some appropriate penalty be imposed upon him. However, standards of retributive justice do not require that the penalty take the form of the faulty party being held liable for the costs of particular accidents caused by him." On the other hand, compensatory justice requires the elimination of undeserved or otherwise unjustifiable losses caused by the fault of another. Thus, a victim is entitled to recompense as a matter of justice for those losses caused by his injurer's fault. 7 However, those considerations that ground an individual's right to recover need not coincide with those that root an obligation to provide compensation. Although the presence of causally efficacious injurer fault may suffice to ground a victim's right to compensation, to be just the obligation to annul or eliminate the victim's loss need not be discharged by the party at fault. '" In short, although principles of corrective justice may justify penaliz-14. 15. 16. 17. 18. Calabresi & Hirschoff, supra note 13, at 1084. See Coleman, The Fault System, supra note 1; Coleman, Justice, supra note 11. See Coleman, The Fault System, supra note 1, at 475.83. See Coleman, Justice, supra note 11, at 173. See id. at 176-78. HeinOnline -- 18 Wm. & Mary L. Rev. 263 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 ing wrongdoers and principles of compensatory justice require elimination of wrongfully inflicted losses, neither standard of justice requires that the costs of individual accidents be borne by those at fault in causing them. These considerations demonstrate not only that fault-based liability rules are not required to secure whatever punitive or compensatory aims we might have, but perhaps more importantly, that the role of fault in a system of tort liability may not best be explained by reference to these ideals of justice." If the fault principle is not required to satisfy either punitive or compensatory ideals, how might an accident law that would secure these goals be constructed? First, all conduct that is at fault without regard to the presence or absence of harmful causal consequences might be lumped together. 0 Once the category of conduct that is at fault is considered generally, appropriate penal-like sanctions could be implemented, measured by the degree of culpability or dangerousness in faulty conduct of different sorts rather than by the standard index of harm actually caused. 21 While culpability is being penalized through penal-like sanctions, compensation of faultless victims could be secured through a system of private insurance contracts or from general tax coffers by which individual victims protect themselves against losses they might suffer at the hands of the faulty. This form of protection might be limited only to faultless victims. In that way, although everyone who engages in an activity would be free to purchase an insurance policy against activityrelated losses, the protective device would be triggered only when the victim could establish the unjustifiability of his loss. For reasons of administrative and economic efficiency, the insurance could be extended either to the class of faultless victims who are injured through no one's fault or to the general class of persons engaged in 19. I do not mean to suggest that the fault system is not the best compromise, but to argue only that justice in tort liability does not require the fault criterion. For an argument that fault is the best compromise, see Blum & Kalven, The Empty Cabinetof Dr. Calabresi,Auto Accidents and General Deterrence, 34 U. Cm. L. Rav. 239 (1967). 20. See Coleman, The Fault System, supra note 1, at 484-86. 21. This view, or at least one similar to it, was considered and rejected in Wagon Mound I. Overseas Tankship (U.K.) Ltd. v. Morts Dock & Eng'r. Co., [1961] A.C. 388 (P.C.). The court held that a ship owner was not liable for damage to a wharf that resulted when oil spilled from the ship ignited and burned, and stressed that such damage was not a foreseeable consequence of the lack of care that resulted in the spillage. [Tihere can be no liability until the damage has been done. It is not the act but the consequences upon which tortious liability is founded. Just as. . . there is no such thing as negligence in the air, so there is no such thing as liability in the air. Id. at 425. HeinOnline -- 18 Wm. & Mary L. Rev. 264 1976-1977 1976] STRICT TORT LIABILITY the activity whether they are at fault or free from it. Although such an expansive approach would not be required by ideals of compensatory justice, insurance schemes of this sort nevertheless are consistent with such ideals. Thus, although principles of compensatory justice do not confer an entitlement to compensation on individuals who have suffered losses through their own fault or through no one's fault, these considerations do not bar a faulty party from indemnifying himself against the costs of his own misdeeds. This is not to suggest that compensatory justice comports with every possible no-fault allocaton of losses. Principles of compensatory justice are best viewed as equilibrium or "steady-state" principles of justice in that they protect a distribution of wealth against unjustifiable gains and losses by requiring the annulment of both. Thus, justice requires that those at fault forfeit the gains imputable to their misconduct; a no-fault allocation of losses in ivhich there is wrongful gain or advantage would not be compatible with this ideal of compensatory justice. The maxim that one ought not profit from one's own wrongdoing and corresponding prohibitions against unjust enrichment are therefore elements in a completely satisfactory account of compensatory justice. This brief summary of the relationship of fault to justice in tort liability leaves a number of questions unanswered. Foremost among them are the following: If fault-based liability rules are not required to secure the goals of justice in tort law, what justification exists for a fault system? If shifting losses on the basis of fault is not required either to compensate victims or to penalize injurers, what could possibly be unjust about a system of strict liability that imposes liability without regard to fault? The question is less paradoxical than it appears at first. It does not follow from the conclusion that our principles of justice do not require the fault system either that every non-fault alternative to the fault system comports with our notions of justice, or that the fault system is not morally superior to certain non-fault alternatives. Even if the fault system is not required by considerations of justice, the question remains whether it is nevertheless morally superior to the rule of strict liability. II Apparently, most of those unfamiliar with the case law and many of those familiar with it believe that tort liability is based on the criterion of fault. According to the fault principle, the costs of an HeinOnline -- 18 Wm. & Mary L. Rev. 265 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 accident ought to be borne by that party whose fault caused it."2 One can then distinguish among four categories of limitations on the scope of fault as a basis of tort liability. In cases of the first sort there is defendant fault, but the victim nevertheless is denied recovery because of some real or imputed shortcoming in the victim's conduct. Thus, a plaintiff may be denied recompense for his losses caused by the defendant's fault if he is contributorily negligent" or if the contributory negligence of another is imputed to him.24 In cases of the second sort, the defendant is at fault but the plaintiff fails to recover either because the plaintiff assumed the risk"5 or because the relationship between the defendant's fault and the victim's injury is insufficiently proximate to justify compensation.26 In these cases there is no defect, real or imputed, in the plaintiff's conduct, yet he fails to recover either because the defendant's fault fails to reach far enough or because of some contractually assumed or otherwise accepted level of risk on the plaintiff's behalf. Cases of denied recovery because of victim status, rather than victim misconduct, also could fall into this category. For example, a defendant's fault may not reach far enough to protect the innocence of a trespasser though the trespasser's loss 22. See notes 1-2 supra & accompanying text. 23. See PROSSER, supra note 3, § 65. 24. Although it was a viable concept in older common law, imputed contributory negligence is all but nonexistent in modern tort law. Today it survives principally as an adjunct to "automobile consent" statutes in those states where courts have interpreted such statutes as mandating the imputation of a driver's contributory negligence to an automobile owner, thus barring his action for damages. See id. § 74. 25. See, e.g., Brown v. San Francisco Ball Club, 99 Cal. App. 2d 484, 222 P.2d 19 (1950); Hunn v. Windsor Hotel Co., 119 W. Va. 215, 193 S.E. 57 (1937). See generally PRossER, supra note 3, § 68. 26. See, e.g., Palsgraf v. Long Island R.R., 248 N.Y. 339, 162 N.E. 99 (1928); GreenWheeler Shoe Co. v. Chicago, R.I. & P. Ry., 130 Iowa 123, 106 N.W. 498 (1906); In re Polemis, [19211 3 K.B. 560 (C.A.). Given that the injurer is at fault and that his fault is a cause of the victim's injury, the question in such cases is whether it is sufficiently proximate to entitle the victim to recovery. See generally PROSSER, supra note 3, §§ 41-44. Hart and Honors have attempted to construct a coherent analysis of the concept of legal cause in torts and elsewhere. Their general analysis involves two steps. First, in order for liability to be imposed, a minimal causal relation must exist between the defendant's conduct and the alleged harm. Once the minimal condition has been established, the next step is to determine if any conditions exist that would defeat or negative the attribution of legal responsibility based on the first-step causal inquiry. Their paradigm of a negating condition is the intervening voluntary action of a third party. In the absence of such conditions, liability is justly imposed. See HART & HONORI, supra note 3, at 126-30. Compare their analysis with J. FEINBERO, Causing Voluntary Actions, in DOINO AND DESERVING 152 (1970) and with Calabresi, Concerning Cause and the Law of Torts: An Essay for Harry Kalven, Jr., 43 U. CHI. L. Rv. 69 (1975). HeinOnline -- 18 Wm. & Mary L. Rev. 266 1976-1977 1976] STRICT TORT LIABILITY would not have occurred but for the defendant's fault. 7 Cases falling into a third category differ from those in the first two in that the defendant is immune from liability either to the instant plaintiff or to all possible plaintiffs. Examples of defendant status immunity include family2' and charitable2 immunities. These categories of exceptions to the fault principle overlap and might be put in other terms. Generally, however, exceptions to fault-based liability within the fault context are grounded either in plaintiff or defendant status, contractual agreement, plaintiff misconduct, or in the insufficient reach of defendant fault. In each of these categories, the defendant is at fault and his fault is causally responsible for the victim's losses, yet the victim goes uncompensated and the defendant escapes liability. As an empirical matter, these limitations on the reach of the fault standard ought to shed some doubt on the widely held view that the fault system aims primarily to compensate the victims of fault or to punish the perpetrators of it. Each of these limitations on the availability of compensation to victims presupposes a basic commitment to the criterion of fault as a standard of liability in torts. They represent limitations on the extent to which satisfaction of the fault requirement determines 27. The common law distinction among licensees, invitees, and trespassers provides perhaps the best example of cases in which the right to recovery for accidental injury has been determined by the status of the victim rather than by the culpability of the defendant. In general terms, a landowner has been held to a duty of ordinary care in maintaining his property in a safe condition when injury has occurred to an invitee. A less stringent standard of care has been applied when the victim was a licensee, and the landowner has been held to owe virtually no duty of care to a trespasser. The approach to tort law exemplified by these distinctions is one that analyzes fault in terms of a failure to discharge a duty of care, and incorporates in doctrinal terms the duty owed to various categories of victims. Recent cases have made significant inroads into the doctrinal invitee.licensee-trespasser trichotomy. Courts that have rejected the common law categories generally have done so when they have been presented with the issue in an inviteelicensee context. See, e.g., Keimarec v. Compagnie Generale Transatlantique, 358 U.S. 625 (1959); Smith v. Arbaugh's Restaurant, Inc., 469 F.2d 97 (D.C.C. 1972); Rowland v. Christian, 69 Cal. 2d 108, 443 P.2d 561, 70 Cal. Rptr. 97 (1968). 28. The common law rule barring personal tort actions between spouses still is followed in many jurisdictions. PROSSER, supra note 3, § 122, at 861-62. However, there is a growing trend toward outright rejection of spousal immunity and a concomitant allowance of an action by either spouse for personal torts, negligent or intentional, committed by the other. Id. at 864. The common law parental immunity from personal tort actions by children remains more prevalent than spousal immunity, but in many jurisdictions it too has been eroded by exception or abrogated outright. Id. at 866-68. 29. The common law tort immunity of charitable organizations and enterprises has been abrogated in a majority of jurisdictions and has been limited in one way or another in almost all others. Id., § 133. HeinOnline -- 18 Wm. & Mary L. Rev. 267 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 liability even in a fault context. In contrast there is an everexpanding body of tort law in which the incidence of a loss is said to be determined without regard to the criterion of fault. This category of exceptions to the fault standard commonly is referred to as strict liability. It includes the law governing workmen's compensation," abnormal or ultrahazardous activities,3 nuisance,3 and products liabilty.3 3 First party, no-fault accident law is a special kind of strict liability in which each litigant is strictly liable for his own losses, not for those of another. The categories of cases now decided on a strict liability basis are so diverse that the most interesting current literature" on tort law has been devoted to trying to establish a family resemblance among them. These efforts hopefully will be enhanced by drawing certain distinctions that bear on the application of strict liability theory. First, the procedural and substantive aspects of strict liability can be distinguished. On the one hand, strict liability is primarily a theory about evidence, burdens of proof, and standards of recovery; on the other hand, it is a liability rule about where losses ought to fall. As a procedural matter, strict liability dictates that it is not 30. A large portion of the field of law governing employer liability for injuries to employees has been preempted by workmen's compensation acts, which have been enacted in all states. These acts impose a form of strict liability in that the employer, aided by a system of compulsory liability insurance, is made liable for injuries arising out of his business, irrespective of whether the accident was caused by his own negligence, an employee's, or no one's. Id., § 80. 31. Strict liability generally is imposed when a defendant damages another by a thing or activity that is unduly dangerous and inappropriate to the place it is maintained. Id., § 78, at 508. This rule was derived from the famous English case of Rylands v. Fletcher, [18681 L.R. 3 H.L. 330, in which defendants were held strictly liable when water from their reservoir escaped through abandoned coal mining shafts and flooded an adjoining mine. American courts have applied the rule in a variety of contexts. See, e.g., Exner v. Sherman Power Constr. Co., 54 F.2d 510 (2d Cir. 1931) (quantity of explosives stored in city); Britton v. Harrison Constr. Co., 87 F. Supp. 405 (W.D. Va. 1950) (blasting); Green v. Gen. Petroleum Corp., 205 Cal. 328, 270 P. 952 (1928) (drilling oil well in residential area). 32. Liability for nuisance may be based on any of three types of conduct encroaching upon the interests of another: an intentional encroachment, a negligent one, or one resulting from conduct that is abnormally dangerous and out of place in its surroundings. It is to this last. type of conduct that strict liability applies. PRossER, supra note 3, § 87. See, e.g., Whittemore v. Baxter Laundry Co., 181 Mich. 564, 148 N.W. 427 (1914) (quantity of gasoline stored near plaintiff's residence); Rider v. Clarkson, 77 N.J. Eq. 469, 78 A. 676 (1910) (vicious dog). 33. The development of products liability law has involved an interplay of three conceptual bases of liability: negligence, warranty, and strict liability. PsossER, supra note 3, §§ 96-98. Strict liability presently is applied in this area by some two-thirds of the courts. Id., § 98, at 658. 34. See, e.g., Calabresi & Hirschoff, supra note 13; Epstein, supra note 6; Posner, 2 J. LEGAL STUDIES, supra note 13. HeinOnline -- 18 Wm. & Mary L. Rev. 268 1976-1977 1976] STRICT TORT LIABILITY relevant to the plaintiff's claim for recovery that he establish the fault of the defendant in causing his injury. Although the defendant may well be at fault, the plaintiff's right to secure compensation from him does not depend on showing to the satisfaction of a judge or jury that he is. Substantively, the strict liability rule is that even if the defendant is not at fault, he is responsible for the victim's loss. So, on the one hand, strict liability is a series of rules about the kind of cases a plaintiff-victim must make out in order to trigger the application of a liability rule; on the other hand, it is the liability rule that is triggered thereby. In those categories of cases governed by the strict liability rule, cases in which there is in fact defendant fault and those in which the defendant is free from fault are indistinguishable from the point of view of standards of proof and threshold requirements of liability. Yet there is the obvious distinction that in some cases the defendant is at fault and in others the defendant is not at fault. At first blush, it seems somewhat illogical to treat some cases in which there is defendant fault less like other cases in which there is defendant fault and more like cases in which the defendant is faultless. Of course, if one has reason to believe that in areas of the law governed by the strict liability rule the goals of tort law are advanced by ignoring the requirement of fault, then abandoning fault is a perfectly plausible strategy. The point can be put more generally. The law develops by categories. If a category of cases can be delineated that are best dealt with on a no-fault basis and if a case falls within the boundaries of the category, it makes perfectly good sense to apply the strict liability rule to it even if in the instant case the criterion of fault is or can be satisfied. But to ask why, in certain kinds or categories of cases but not in others, the presence of defendant fault is absolutely necessary to trigger the application of a lossshifting liability rule is simply to raise in another form the question of what the goals of tort law are. Although the procedural aspect of strict liability compels us to. ignore the criterion of fault even when it is present, the substantive or liability aspect of it may require us to impose liability upon individuals who cause harm through no fault of their own. The substantive aspect of the strict liability rule therefore raises this difficult question: Why is it that failure to find defendant fault constitutes a complete bar to recovery in some areas of tort law, whereas in other cases absence of defendant fault is no bar to recovery at all? Indeed, when strict liability is the rule, not only does the absence of fault fail to bar liability, but failure to find even a causal HeinOnline -- 18 Wm. & Mary L. Rev. 269 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 relationship between the defendant's conduct and the plaintiff's injury may not be a defense against liability. 5 Although the issues raised by the procedural and substantive elements of the strict liability rule are very similar, they are not identical. The procedural component raises the question of why we would abandon fault as the criterion applicable in some cases but not in others when fault is present in both. The substantive element of the strict liability rule asks us to justify the imposition of liability in the absence of fault. If strict liability always involved only an abandoning of fault when, as a matter of fact, it was present, the strict liability rule doubtlessly would prove less troublesome to those who object to it from the moral point of view. But as a matter of fact, adherence to the strict liability rule means that at least some of the time liability will be imposed upon individuals who are absolutely free from fault. Thus, strict liability may require individuals to bear the costs of accidents that are not their fault. We must now distinguish between the two senses in which liability may be imposed strictly. There are two fundamental components in all fault judgments. To be at fault, an actor (1) must fail to satisfy a standard of reasonable conduct, and (2) in some sense that failure must be imputable to him as his doing. The distinction between these elements in fault attributions is illustrated best by the positive defenses one can offer to them: justifications and excuses. Briefly, a justification is proffered to establish that the conduct in doubt indeed meets required standards. An excuse is offered not to contest the failure to measure up, but instead is offered to deny responsibility for it. An appropriate justification defeats a fault attribution because a justification unveils that the attribution has been mistakenly grounded in an allegation of failure to measure up to standard. An excuse defeats an ascription of fault by pulling the rug from under the claim of responsibility implicit in it. 6 35. There are at least two kinds of cases in which liability need not require a causal connection between the conduct of the party held strictly liable and the harm: (1) liability on the model of workmen's compensation statutes; and (2) liability for results (not actions). Liability for results is most prevalent in criminal law in which criminal penalties have been imposed despite the prosecutor's failure to prove criminal intent, most frequently when statutes impose sanctions for offenses against public health or safety, see Morissette v. United States, 342 U.S. 246, 256-60 (1952); United States v. Dotterweich, 320 U.S. 277, 281 (1943). and including, recently, violations of environmental protection laws, see State v. Arizona Mines Supply Co., 107 Ariz. 199, 484 P.2d 619 (1971). 36. See H.L.A. HART, Prolegomenon to the Principlesof Punishment, in PUNISHMENT RESPONSIBILITY 1, 13-14 (1968). HeinOnline -- 18 Wm. & Mary L. Rev. 270 1976-1977 AND 1976] STRICT TORT LIABILITY This distinction demonstrates that there are two senses in which liability in torts can be said to be strict, that is, imposed without regard to fault. Liability is strict if it is not defeasible by an excuse. That is, an actor may be held liable when, in Hart's words, he "could not help doing what he did""7 (though lack of capacity constitutes but one sort of excusing condition). In strict liability of this sort a defendant may be liable for the costs of accidents caused by his conduct even though he could show to the satisfaction of the trier of fact that he lacks substantial capacity to comply with required standards of care. This is liability for excusable conduct. On the other hand, liability without fault may be liability for conduct that is reasonable or that satisfies standards of due care. In other words, in strict liability of this sort, an actor may be accountable for injuries caused by his perfectly acceptable conduct. This is liability for justifiable risk-taking. This Article is concerned only with strict liability in the latter sense and with the accusation that liability for the harmful causal upshots of reasonable conduct is unjust." I want to argue that this criticism of strict liability is at best overstated and perhaps is fundamentally misconceived. 37. H.L.A. HART, Legal Responsibility and Excuses, in PUNISHMENT ANDRESPONSIBILITY 28, 39 (1968). 38. Although liability without fault, in the sense of liability in the absence of culpability or responsibility, presents interesting philosophic issues, an adequate analysis of the role of excuses in a theory of tort liability ought to be discussed within the context of a general appraisal of the objective character of fault judgments, for it is common in tort law that even when liability is imposed on the basis of fault, the fault required need not be a moral one. Thus, a defendant may be at fault even if his failure to exercise due care is not his doing-if he has an excuse-provided the reasonable man could have exercised proper care and foresight. This is a lesson to be derived from the landmark case of Vaughan v. Menlove, 132 Eng. Rep. 490 (C.P. 1837), in which the court made its classic statement on the objectivity of the standard to be applied in determining negligence: Instead, therefore, of saying that the liability for negligence should be coextensive with the judgment of each individual, which would be as variable as the length of the foot of each individual, we ought rather to adhere to the rule which requires in all cases a regard to caution such as a man of ordinary prudence would observe. Id. at 493. See also J. Coleman, Fault Defeasibility and Blame (Dec. 6, 1976) (unpublished manuscript in The University of Wisconsin-Milwaukee Library). Consequently, the standards of both fault and strict liability may impose the obligation of recompense in the absence of moral responsibility. For that reason a discussion of the alleged injustice of strict liability for excusable conduct is best left for an overall appraisal of the role of responsibility as a condition of tort liability. Ultimately, the issue boils down to a jurisprudential one: Is tort liability fundamentally a problem in the theory of responsibility, in the economic theory of loss allocation, or both? HeinOnline -- 18 Wm. & Mary L. Rev. 271 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 m The simple truth is that even in the fault system individuals may be required to bear the costs of accidents that are not their fault. This is an important consequence of those limitations on the scope of fault discussed in some detail above. 9 But the non-fault element in the fault system extends beyond these technical limitations on the scope of the criterion of fault. To see that this is true, we have to reconsider the initial liability rule of the fault system that the costs of accidents ought to lie on victims. This liability rule enables a particular victim to have liability for his losses transferred to his particular injurer provided he can establish the fault of his injurer in causing the harm. What that implies of course is that if the victim is unable to establish that the criterion of fault has been satisfied, he bears the costs, even if he is entirely free from fault. This is not merely a matter of evidence, but of substantive law. If neither litigant is at fault, the fault system is designed so that the faultless victim bears the costs. That is a direct consequence of the initial liability rule. The upshot of the initial liability decision, rooted, if we are to follow Holmes, in considerations of administrative cost avoidance, 0 is a far-reaching no-fault component in the fault system. This no-fault element requires no more and no less than does its counterpart in the strict liability rule. In a system of strict liability, the opposite initial liability rule operates. The injurer bears the costs of accidents unless he can show some defect in the victim's conduct sufficient to free him from liability. If he is able to do so, the costs of the accident are transferred to the victim. Failing that, however, the injurer must bear the costs, even if he is entirely without fault. Thus, in both strict and fault liability, an individual may be required to bear the costs of accidents that are not his fault. This point should not be overstated. It is very plain, but overlooked, that in both a fault and a strict liability system there are circumstances in which an attribution of liability cannot be defeated by the positive defense of freedom from fault. At the same time, the point should not be underestimated. What has been shown is that both the fault system and the strict liability rule are committed at their core to liability without fault.'" Thus, the charge that 39. See notes 23-29 supra & accompanying text. 40. See note 8 supra & accompanying text. 41. See notes 22-32 supra & accompanying text. HeinOnline -- 18 Wm. & Mary L. Rev. 272 1976-1977 1976] STRICT TORT LIABILITY what is peculiarly wrong with strict liability is that it does not permit an individual to free himself from the burdens of liability by the defense of faultlessness may be simply misplaced or at least overstated. Although absence of fault is not a defense in any area of strict liability law, different categories of strict liability may be distinguished from one another on the basis of the sorts of plaintiff-related defenses a party held strictly liable may offer to defeat an ascription of liability. In some cases, the presence of victim negligence or con- tributory fault is a defense to strict liability."2 In other areas of tort law, such as workmen's compensation, plaintiff negligence is not a defense, though wilful misconduct by a plaintiff may be. 3 In pro- ducts liability, misuse of product or unforeseen usage may be a defense," though plaintiff fault may not be.", Yet in other kinds of 42. A plaintiff's mere lack of ordinary care is generally not enough to constitute the degree of contributory fault required to bar his recovery in a strict liability case. But if the plaintiff voluntarily and unreasonably has exposed himself to a known danger, he will be deemed to have been contributorily negligent, or to have assumed the risk. See, e.g., Hughey v. Fergus County, 98 Mont. 89, 37 P.2d 1035 (1934) (entering field with bull); Worth v. Dunn, 98 Conn. 51, 118 A. 467 (1922) (plaintiff's conduct increased danger from blasting). PROSSER, supra note 3, § 79. 43. The general role of workmen's compensation laws is that misconduct of the employee, whether negligent or wilful, is immaterial because the basic test for recovery is whether the injury is related to the employment, not whether one party or the other was responsible. Exceptions to the rule exist in those jurisdictions whose statutes provide for "wilful misconduct," or some similar terminology, as a defense and in those whose statutes provide for more narrow grounds of defense, such as "wilful violation of a safety regulation." The most common type of statute provides for no such affirmative defenses based on employee misconduct in the course of employment. 1A A. LARSON, THE LAW OF WORKMEN'S COMPENSATION § 30.10 (1952). 44. Generally the seller is entitled to anticipate only normal uses of his product and is not liable when it is put to abnormal uses. See, e.g., Zesch v. Abrasive Co., 354 Mo. 1147, 193 S.W.2d 581 (1946) (grinding wheel used improperly); Dubbs v. Zak Bros. Co., 38 Ohio App. 299, 175 N.E. 626 (1931) (shoes intentionally worn two sizes too small). PROSSER, supra note 3, § 102. There is a subtle but important distinction between the contributory fault and victim misuse defenses from liability, though both require establishing a defect in the victim's conduct. One can imagine an emergency in which it would be reasonable (i.e., not faulty) for a person to "drive" a power lawn mower down a busy street and through a dangerous intersection. For example, suppose one's car fails to start after a guest at one's house is suddenly stricken by a heart attack and must be taken to a hospital a mile or so away. Clearly it may be a misuse, and certainly an unforeseeable use, of the mower to employ it in this fashion. Should an inherent defect in the mower cause an accident resulting in injury to the operator of the mower, his stricken passenger, or some other third party, the mower company, normally liable under products liability law, could escape liability, not because the injury was caused by the driver's fault, but because it resulted from product misuse. 45. As with strict liability generally, see note 42 supra, mere lack of ordinary care is insufficient to constitute contributory fault to bar a plaintiffs recovery in a products liability HeinOnline -- 18 Wm. & Mary L. Rev. 273 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 cases, no plaintiff-related defense will succeed, at least once certain minimal conditions of liability are met." It would be fruitful, therefore, to distinguish four general systems of liability: (1) the fault system; (2) strict injurer liability with the defense of plaintiff (victim) fault; (3) strict injurer liability with no defense of plaintiff misconduct; and (4) strict victim (plaintiff) liability with no defense of injurer wrongfulness. Categories (3) and (4) above may be labeled absolute injurer and absolute victim liability respectively. The first thing to note about these categories is that (2) is merely (1) in reverse. This is a point already well made by Guido Calabresi. 7 Substantively, the two theories of liability differ only with respect to whom liability will be attributed initially, and consequently, where it will fall if the fault-based liability provision cannot be met. When the fault system decides in both cases against the victim, the system of strict liability with the defense of plaintiff fault decides against the injurer. Secondly, under both an absolute victim liability system and a fault system, the victim would bear the loss if there were no fault on the part of either litigant. But in the same circumstance, the injurer would bear the loss under a strict injurer liability system; this would be so whether or not the system was one that allowed victim misconduct as a defense. Thirdly, in both absolute injurer and absolute victim liability systems, not even misconduct by a victim or injurer, respectively, would suffice as a defense. In that respect, the absolute liability case. Thus, that a plaintiff has failed to discover the danger in a product is not a viable defense to strict liability, see, e.g., Brockett v. Harrell Bros., 206 Va. 457, 143 S.E.2d 897 (1965) (failure to discover shot in ham); nor that he has failed to take precautions against its existence, see, e.g., Dagley v. Armstrong Rubber Co., 344 F.2d 245 (7th Cir. 1965) (negligent driving on defective tire). However, when the plaintiff voluntarily and unreasonably encounters a known danger, he will be barred from recovery on a contributory negligence or assumption of risk theory. See, e.g., Youtz v. Thompson Tire Co., 46 Cal. App. 2d 672, 116 P.2d 636 (1941) (negligent driving on tire known to be defective); Gutelius v. General Elec. Co., 37 Cal. App. 2d 455, 99 P.2d 682 (1940) (washing machine used after discovery of dangerous defect). PaOSSER, supra note 3, § 102. 46. This would be true, for example, if a state "dog bite" statute provided for owner liability for all injuries caused by his dog unless the person injured was committing trespass or a tort. See, e.g., Nelson v. Hansen, 10 Wis.2d 107, 102 N.W.2d 251 (1960); Ingeneri v. Kluya, 129 Conn. 208, 27 A.2d 124 (1942). Similarly, no plaintiff-related defense will succeed in a jurisdiction whose workmen's compensation statute provides for no defenses based on employee misconduct. See note 43 supra. 47. Calabresi, OptimalDeterrenceand Accidents: To FlemingJames,Jr., 84 YALE L.J. 656 (1975). HeinOnline -- 18 Wm. & Mary L. Rev. 274 1976-1977 19761 STRICT TORT LIABILITY systems differ most significantly from their non-absolute counterparts. We now can begin to evaluate the claim that strict liability is an immoral theory of responsibility in torts. Several questions are involved here. First, is there a moral reason for preferring the fault system to strict liability with the defense of plaintiff fault? As both systems of liability transfer losses to the party at fault if there is fault, they differ only with respect to who bears the loss if there is no culpable litigant. The question ultimately becomes whether there is a moral reason for favoring one no-fault decision over another, that is, for systematically favoring the no-fault decision against victims over the same decision against injurers. Following Holmes,4" one might argue that the no-fault decision against injurers is the administratively costlier alternative of the two. Because administrative costs are wasteful and lead to an inefficient allocation of resources, and because waste and inefficient resource use is morally undesirable, the fault system is to be preferred to a system of strict liability. I am not prepared to deny that inefficiency in the economic sense is a moral shortcoming in a system of liability rules. 9 Yet it hardly follows from the fact that trans48. See note 8 supra & accompanying text. 49. According to the prevailing wisdom of neo-classical welfare economics, efficiency is the primary standard for evaluating an allocation of resources. An allocation of resources is maximally efficient when it is Pareto Optimal. To say that resources are allocated in a Pareto Optimal fashion is to claim that any further voluntary (costless) transfer (contract or sale) of resources will make someone better off only by making at least one other person worse off. So a Pareto Optimal distribution exists when no one can be made better off without making someone else worse off, judged by each individual's standard of welfare or satisfaction. The distinction between Pareto Optimality (maximal economic efficiency) and both Pareto Superiority and Maximal Product Output must be kept clear. An allocation scheme is Pareto Superior to an alternative if and only if in the Pareto Superior distribution no one conceives of himself as worse off and at least one person conceives of himself as better off. A distribution of goods may be Pareto Superior without being Pareto Optimal. Interestingly, it does not follow from a policy-making standpoint that a Pareto Optimal allocation is preferable to a Parento Superior one. The weaker policy-making implications of Pareto Optimality largely have been ignored by lawyers and ought to be examined closely by philosophers concerned with the foundations of institutional rule making. Imagine a two person universe and an all-purpose resource, manna. Suppose that the manna is distributed so that A receives all of it. Any redistribution of the resources (without compensatory provisions) would make A worse off. Thus, this apparently unfair allocation scheme is Pareto Optimal. Although there seem to be good reasons for adopting a Pareto Superior distribution scheme to an alternative rule because, by hypothesis, it makes no one worse off and at least one person better off, the same cannot be said always of a Pareto Optimal distribution. A Pareto Optimal distribution need not maximize output. A Pareto Optimal resource HeinOnline -- 18 Wm. & Mary L. Rev. 275 1976-1977 WILLIAM AND MARY LAW REVIEW (Vol. 18:259 ferring losses from victims to injurers is costly that doing so is invariably not worth its costs. Surely the question is not whether a liability rule involves incurring administrative costs but whether the rule is worth its costs. Alternatively, one could argue, as Holmes appears to,5" that the injurers are the doers, the industrialists, the opportunity seekers, and the providers. Consequently, on this view they deserve the advantage created by an initial liability rule favorable to them. The argument rests on the dubious initial premise that as a class injurers are doers, and that victims, as a class, are merely passive agents. More importantly, absent a moral accounting of the entire lives of categories of victims and injurers, past, present, and future, there is absolutely no reason to believe that injurers rather than victims have earned or merited a favorable initial liability rule. Perhaps the most intriguing agrument for preferring as a general rule that the no-fault decision favor the injurer rather than the victim relies on the premise that in strict liability, but not in the fault system, an act of compensation is involved; the defendant must make good the victim's loss if neither is at fault.5 On the other hand, the no-fault element of the fault system requires only that the victim make good his own losses. Though there is a no-fault component in both theories of liability, only in strict liability does the nofault decision translate into an act of compensation. Indeed, in the fault system the no-fault decision against the victim requires neither compensation nor an attribution of liability. When the victim distribution may involve transforming all the manna into five pairs of shoes that are then awarded to A. If it turns out that A's welfare is maximized by owning five pairs of shoes, then any other resource use and allocation would make him worse off than he is now. This would be true even if the manna could have been transformed into large quantities of books, works of art, food, shelter, and the like. It is important for philosophers venturing into the economic analysis of legal rules to recognize the limited role of Pareto Optimality in policy making. For a useful and altogether readable discussion of these basic concepts, see B. ACKERmAN, ECONoMIC FOUNDATIONS OF PRoPERY LAw (1975). 50. HoLMEs, supra note 4, at 77. 51. Arguments based on the condemnatory or attitudinal aspects of punishment have been advanced to shed doubt on the moral propriety of strict criminal liability as a general theory of criminal responsibility. For a discussion of these arguments and their limits, see J. FEINBERG, The Expressive Functionof Punishment, in DoING AND DESERVING 95 (1970). For reasons I suggest below, see notes 52-65 infra & accompanying text, arguments based on the expressive elements in tort liability or compensation do not weigh nearly as heavily against strict tort liability. My arguments do not hold (nor should they) when tort liability involves so-called punitive damages. HeinOnline -- 18 Wm. & Mary L. Rev. 276 1976-1977 1976] STRICT TORT LIABILITY is unable to establish the fault of the defendant, we do not characterize his having to bear his own losses as his being held liable for them, though we might describe him as being liable for them. Compensation is required as a matter of justice when it is necessary to annul or to eliminate an unjustifiable loss; that is, to make good a loss caused by the fault of another. 52 But compensation also may be a vehicle for expressing public attitudes of disapproval or, in extreme cases, of condemnation toward the conduct of the offending party. More likely, compensation may serve as a private law remedy by which injurers take responsibility for, or declare as their own, actions that cause harm to others. Or perhaps compensation is a useful social instrument through which injurers express their regrets and apologies to the victims of their misdeeds. If that is so, strict liability may be an immoral standard of liability because it requires a morally inappropriate act. By requiring an injurer to compensate his victim, strict liability may compel an injurer to offer apologies when he has done nothing for which he ought to apologize. What may be wrong with strict liability as a standard of tort responsibility is that the act of compensation it compels is a source of "undeserved responsive attitudes and unfair judgments of blame." 3 There are really two closely allied arguments here. First, strict liability is immoral because it requires the injurer to express morally inappropriate feelings of regret, shame, guilt, fault, or the like. Secondly, because attributions of liability imply judgments of blameworthiness, strict liability involves making formal, official pronouncements of culpability in a judicial context. It requires us, in short, to impute blame to blameless conduct. Arguably, however, only infrequently does tort liability invoke feelings of contrition, regret, shame, guilt, remorse, or the like on the part of the party required to render compensation. For the most part, compensation as a remdy in torts is impersonal and usually is transacted through insurance mechanisms. To illustrate the diverse nature of compensation as a tort remedy, consider the case of a polluting feedlot, a simple nuisance. In the absence of an injunction, the polluter may be free to operate his nuisance provided he compensates (often on an ongoing basis) those individuals whose property values are adversely affected by the nuisance. In cases of this sort, the compensation rendered by the offending party more accur52. See note 17 supra & accompanying text. 53. FEINBERO, supra note 51, at 115. HeinOnline -- 18 Wm. & Mary L. Rev. 277 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 ately represents an operating cost of the feedlot or the recompense required under a "takings-like" doctrine54 than it represents the vehicle through which the wrongdoer is made to express his apologies to his victims. The truth is that in a number of nuisance cases, 54. The fifth amendment to the Constitution provides: "[Nior shall private property be taken for public use, without just compensation." U.S. CONSr. amend. V. This section of the amendment has come to be called the "takings clause." It is a well of philosophically interesting problems. First, its implementation requires a model for determining if a taking of property is a taking in the constitutional sense, i.e., one that requires compensation. Such'a formula is needed if consistent answers are to be given to the practical questions that arise under the clause. For example, are condemnation of property and consequent transfer of title necessary ingredients in the takings formula? Does drastic diminution of value brought about by government action fall within the formula? Are zoning changes and consequent limitations of owner prerogatives remediable under the clause? Next, the takings clause requires formulation of a workable distinction between private and public uses without begging the entire question. Third, it demands a well formulated theory of fittingness or appropriateness in compensation. Ultimately, the entire takings question leads one to an independent analysis of the concepts of property and ownership. One recent approach to understanding tort liability analogizes polluting one's neighbor to a taking of his property by diminishing its value sufficiently to require compensation. Whether compensation is owed to someone, in this case the neighboring resident, is not at issue. The issue is how the compensation rendered is conceived. The advantage of the takings analogy is that it subsumes compensation in tort law under a property law model. Specifically, compensation rendered by injurers to their victims in certain tort cases is based on the model of eminent domain. Thus conceived, compensation in torts does not center around expressions of regret, shame or apology; it requires no element of culpability. Instead, compensation is a cost one must accept in order to pollute or in some sense use the property rights initially conferred upon another. The takings analogy is of questionable validity. At'the least, in order for it to hold, the injurer (tortfeasor) must be an agent of the state or must satisfy the state action requirement in some other way. After all, the fifth amendment enumerates limitations or restrictions on state, not private,.action. The analogy would be less questionable, for example, were the polluter a public utility. Nevertheless the takings analogy sometimes is employed to explain holdings in a variety of tort cases. For example, in Vincent v. Lake Erie Transp. Co., 109 Minn. 456, 124 N.W. 221 (1910), the court employed takings analogy reasoning to hold the owner of a steamship that was moored to a dock during a storm liable for damages to the dock. Although the master of the steamship had "exercised good judgment and prudent seamanship," id. at 458, 124 N.W. at 221, in keeping the vessel fast to the dock, the court found liability because "those in charge of the vessel deliberately . . . held her in such a position that the damage to the dock resulted, and, having thus preserved the ship at the expense of the dock, . . . her owners are responsible to the dock owners to the extent of the injury inflicted." Id. at 459, 124 N.W. at 222. Compare this reasoning with that of the court in Holmes v. Mather, L.R. 10 Ex. 261 (1875), in which the court decided that the owner of a horse was not liable when the animal, startled by a barking dog, ran out of control and injured the plaintiff. In reaching this holding the court relied upon its finding that the defendant had committed no wrongful act and therefore could not be held liable. The basic factual situation was similar to that in Vincent, yet the holding was opposite. The literature on the takings question itself is sparse but improving. For useful overviews, see F. BOSSELMAN, D. CAmuEs & J. BANTA, THE TAKING ISSUE (1973) and the forthcoming book, B. ACKERMAN, PRIVATE PROPERTY AND THE CoNsTrrrUoN (to be published by Yale University Press). HeinOnline -- 18 Wm. & Mary L. Rev. 278 1976-1977 1976] STRICT TORT LIABILITY 279 compensation may be simply the price an offending litigant must pay to be free to victimize his neighbors. As long as the neighbor's entitlement to be free from pollution is not protected by an injunction against the polluter, paying compensation on an ongoing basis may be one of the ways in which the polluter may purchase the right to pollute from his neighbor." For the economist, compensation may be required in such cases to insure that the impact of the pollution on resource use is reflected adequately in the price of the manufacturer's goods. Were the polluter not required to compensate his victims, they in turn would be forced to internalize the costs of pollution and the polluter would be free to pollute "costlessly." 55 Because he would not be required to internalize pollution costs, the prices of his goods would give 55. The view that tort liability has a right-purchasing component is suggested by the takings analogy. See note 54 supra. A general analysis of tort liability presented along these lines can be attributed to Calabresi and Malamed. See note 9 supra. Calabresi and I part company at this point. It is a necessary consequence of his notion of a liability rule that compensation in torts be regarded as a vehicle for purchasing partial or entire entitlements from those upon whom they have been conferred initially. For example, even if X is initially entitled to security from Y's harmful conduct, Y may become entitled de facto to harm X at that price by compensating X for his losses. Thus, compensation is merely the price Y must pay to purchase the entitlement from X. Moreover, X has no say in determining the price Y is to pay, a say that he would have were his right protected by a property rule buttressed-by an appropriate criminal prohibition against theft. Ironically, Calabresi and other legal theorists inclined to economic analysis have fostered an unfortunate analogy between tort and criminal liability. By conceiving of compensation primarily as an operating cost of harmful conduct, they have suggested an analogy with the criminal law conception of punishment or incarceration as the cost of harmful conduct. Thus the rational, utilitarian offender is given the option of complying with a set of rules (criminal law), or with an allocation of entitlements to security (tort law), or violating the norms or purchasing the entitlement at a price. Through this analogy, therefore, both criminal law and tort law are seen as pricing systems setting the requisite "costs" should one choose the latter option. This seems altogether misleading. First, neither tort nor criminal law is primarily cost-setting in its goals. The criminal law establishes standards of conduct that one ought to follow; it is not a pricing system for violations of that code of conduct. But that does not mean tort law is also primarily concerned with setting standards of conduct. Although establishing standards of responsible behavior is one function of tort law, its primary goal is to provide a vehicle for protecting victims from unjustifiable or wrongfully absorbed losses. Thus, compensation is a vehicle for protecting or securing entitlements by annulling or rectifying losses that result from their violation. Compensation, at least when fault is involved, is an instrument for reinforcing one's entitlement; it is not a vehicle for transferring it. 56. This is not to suggest that imposing liability on the polluter is the only way to insure that the prices of his goods adequately reflect the impact of their production on resources; nor is compensation through liability rules required to insure that the neighbors are freed from the costs of internalizing the effects of pollution. For example, one might try to tax the polluter and pass the revenue to the affected parties. This could prove quite unsatisfactory, however, where the polluter is a monopoly. See Buchanan, ExternalDiseconomies, Corrective Taxes, and Market Structure, 59 AM. EcoN. Rav. 175 (1969). HeinOnline -- 18 Wm. & Mary L. Rev. 279 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 systematically misleading cues or information to the consumer with respect to the marginal rate of transformation of resources into goods. The price he would charge could be significantly lower than the marginal rate of transformation, thus leading to both an inefficient allocation of resources and to more pollution. To fail to see compensation and liability rules in this broader context is to miss entirely the role of tort liability in pursuing Pareto Optimal" and/or just distributions of resources. But to see compensation this way is to recognize that disputes settled by compensation are not always best framed by and in terms of the interests of the individual litigating parties only. The more often compensation is required by liability rules governing areas like environmental control, in which disputes often are not best formulated by or in terms of the interests of litigating parties, the less likely is compensation to retain the personalized meaning that has come to be associated with the term by virtue of its usage in the non-legal context. The second thrust of the moral argument against strict liability, that it imputes blame to individuals or to conduct unworthy of blame, rests on the premise that the attribution of liability is itself either a judgment of blame or the source of it. When liability is based on.fault, the attribution of liability is often a shorthand for ascribing blame for the faulty conduct. But if our liability rules do not require fault, an ascription of blame need not be implicit in an attribution of liability. To hold someone strictly liable is not necessarily to blame him for what he has done. Instead, the liability decision may be the result of a delicate blanacing of non-fault related considerations, for example, the relative ability of litigants either to bear risks or to evaluate the worthiness of accident causing conduct.58 That moral stigma or blame attaches to the judgment of fault rather than to the attribution of liability may be illustrated by an example from case law. In Western & Atlantic Railroad v. Henderson,9 the Supreme Court struck down, as a violation of the due process clause of the fourteenth amendment," section 2780 of the Georgia Civil Code 57. See note 49 supra. 58. See notes 12-13 supra & accompanying text. 59. 279 U.S. 639 (1929). 60. U.S. CoNsT. amend. XIV, § I states in pertinent part: "No state shall ... deprive any person of life, liberty, or property, without due process of law. . .... Although the Court noted that legislation creating evidentiary presumptions is valid "if there is a rational connection between what is proved and what is to be inferred," the Court found the Georgia statute violative of the due process clause because it created "a presump. HeinOnline -- 18 Wm. & Mary L. Rev. 280 1976-1977 1976] STRICT TORT LIABILITY which created a nearly irrebuttable presumption of fault' against railroads in the event of "any damages done to persons, stock, or other property by the running of the locomotives, or cars, or other machinery of such company, or for damage done by any person in the employment and service of such company.""2 On the other hand, in St. Louis & San FranciscoRailway Co. v. Mathews,6 3 the Court upheld a Missouri statute that made railroads strictly responsible "indamages to every person and corporation whose property may be injured or destroyed by fire communicated, directly or indirectly, by locomotive engines in use upon the railroad owned or operated by such railroad corporation."" Liability based on an irrebuttable presumption of fault, at least at the turn of the century, was thought to violate the due process clause of the fourteenth amendment. Except for the decision in the anomalous Ives case, 5 strict tort liability never has been thought to constitute an unconstitutional violation of due process. Yet the only meaningful element that seems to underlie the different holdings in Henderson and Mathews is that in the former case the statute provided for liability to be imposed on the basis of fault tlat was irrebuttably presumed to exist. In Mathews no judgment of fault was required to establish liability. When the cases were decided, they tion that is arbitrary or that operates to deny a fair opportunity to repel it." 279 U.S. at 642. 61. Commenting upon the construction given to the statute by the Georgia court, the Court wrote: Upon the mere fact of collision and resulting death, the statute is held to raise a presumption that defendant and its employees were negligent in each of the particulars alleged, and that every act or omission in plaintiffs specifications of negligence was the proximate cause of death; and it makes defendant liable unless it showed due care in respect of every matter alleged against it. 279 U.S. at 41 (emphasis supplied). 62. Id. at 640. 63. 165 U.S. 1 (1897). 64. Id. at 2. In holding the statute to be valid under the due process clause, the Court considered the competing interests of the railroad and property owners, and concluded: To require the utmost care and diligence of the railroad corporations in taking precautions against the escape of fire from their engines might not afford sufficient protection to the owners of property in the neighborhood of the railroads. When both parties are equally faultless, the legislature may properly consider it to be just that the duty of insuring private property against loss or injury caused by the use of dangerous instruments should rest upon the railroad company, which employs the instruments and creates the peril for its own profit .... The statutes not a penal one, imposing punishmentfor a violation of law; but it is purely remedial,making the party, doing a lawful act for its own profit, liable in damages to the innocent party injured thereby . ... Id. at 26-27 (emphasis supplied). 65. Ives v. South Buffalo Ry., 201 N.Y. 271, 94 N.E. 431. See note 3 supra. HeinOnline -- 18 Wm. & Mary L. Rev. 281 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 stood for the proposition that it is not unconstitutional to impose liability without regard to fault, though it may be unconstitutional, in the same circumstances, to impose liability on the basis of fault irrebuttably presumed to exist. The distinction is most plausibly explained in terms of stigma: an attribution of liability with no pretense of fault as its basis imputes no stigma to the party held liable, but an unwarranted imputation of fault reflects unfavorably and unfairly on the character of the accused. To summarize, two very different sorts of moral criticisms of strict liability have been considered, neither of which, I have argued, succeeds. First, critics contend that strict liability is an immoral standard of responsibility in torts because it imposes the costs of accidents on individuals who are not at fault in causing them. In other words, absence of injurer fault is not a defense to strict liability. However, because the type of strict liability considered here allows the defense of plaintiff misconduct, it imposes liability on faultless injurers only when neither they nor their victims are at fault. However, the fault system is committed to imposing liability on faultless parties in the very same circumstances; when neither litigant is at fault, the faultless victim bears the costs. This traditional argument against strict liability, focusing as it does on the absence of fault as a necessary defeating condition of liability in any just theory of tort liability, ignores the plain fact that the fault system involves the same sort of strict liability decision. As a second line of attack, critics attempt to define more narrowly the immorality of strict liability, arguing that there is at least a moral reason for preferring the no-fault decision in the fault system to the one in strict liability. To that end they advance considerations of administrative cost avoidance and injurer desert. But the desert claim is unfounded and the cost-avoidance claim is too weak. The most compelling argument advanced against strict liability at this level depends on the premise that only the strict liability nofault decision involves an act of compensation. According to this argument, the requisite act of compensation is inappropriate given the lack of defendant fault. But as we already have seen, compensation in torts only infrequently involves an expression of apology by the injurer to his victim. In brief, the argument rests on an inappropriate transposition of the ordinary concept of compensation to the legal context. HeinOnline -- 18 Wm. & Mary L. Rev. 282 1976-1977 1976] STRICT TORT LIABILITY IV, The question remains whether there is a moral reason for preferring the fault system to the absolute liability rule as a general matter. What, if anything, is unjust or immoral about absolute liability? There are two distinct kinds of absolute liability rules: absolute victim liability and absolute injurer liability. Absolute victim liability maintains not only that the costs of accidents ought to fall on victims initially, but also that they ought to remain there. Thus, absolute victim liability is nothing more than the initial liability rule of the fault system taken as an ultimate liability decision."8 Absolute injurer liability is simply strict liability without the defense of plaintiff fault or misconduct. Thus, it is the initial liability rule of strict liability, again taken as a final liability rule. The clearest example of a de facto absolute liability system is provided by workmen's compensation schemes that allow no plaintiff-related defenses to employer liability. In such statutory systems an employer may be required to bear the costs of his employees' on-the-job injuries even when these accidents are the fault of the employee." Whatever reasons of deterrence there might be for adopting an absolute liability rule when the party held liable is the one best able to avoid the accident, it at least is arguable that there is a certain clement of injustice in the rule. But can we clearly define the injustice that may be present in the absolute liability rule? Schematically, the rule maintains that X, who is free from fault, must bear the costs of an accident sustained by Y and caused by Y's fault. Sometimes, X also is held liable for the costs of accidents that Y's fault causes Z to suffer. Absolute victim liability is simply a special instance of the latter formula in which X and Z are the same person. The absolute liability rule may require that a fault-free party bear the costs of an accident when a party whose fault caused the accident is available for liability. Is there anything unjust about such an arrangement of responsibili8 claims that there is a ties? Joel Feinberg, in his article Sua Culpa,1 principle of weak retributive justice that holds if a loss must fall on 66. As a general theory of loss allocation, absolute victim liability maintains that losses ought to be internalized by victims. In terms familiar to economists, absolute victim liability requires potential victims to be self insurers or to purchase first party insurance policies. Because of the availability of insurance, we really are deciding in many areas of tort law upon whom the burden of securing insurance ought to fall. 67. See note 43 supra. 68. FEINBERO, Sua Culpa, supra note 1. HeinOnline -- 18 Wm. & Mary L. Rev. 283 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 either of two parties, one of whom is at fault in causing it and the other of whom is faultless, the party at fault ought to bear the loss all other things being equal." Feinberg's principle of weak retributivism is grounded in the requirement that innocence or lack of fault ought to be protected, not in the view that those at fault ought to be penalized. The weak retributive principle admittedly has a limited scope and force. It does not purport to decide the incidence of a loss when both parties are at fault, nor does it determine liability when neither party is at fault. Moreover, the principle does not require that the party at fault bear the loss if his fault is inconsequential or if it does not contribute to the harm. Furthermore, the "all things being equal" escape clause means that even if the fault requirement is met by one party but not the other, conceivably in some situations it would not be unjust for the party free from fault to bear the loss. Apparently one such situation is when the party free from fault contractually agrees to bear the loss. One argument against the absolute liability rule could rely on Feinberg's principle. Thus, under this arguement, the absolute liability rule may require that someone who is free from fault bear the costs of injuries suffered by the party whose fault is responsible for them, but justice requires that the party at fault bear the costs. Therefore, absolute liability goes awry by violating this principle of weak retributive justice. Feinberg's principle requires only that when the only relevant moral difference between litigating parties is that one, but not the other, is at fault in causing the harm, liability ought to be imposed on the party at fault to protect the innocence of the faultless party. In contrast, the absolute liability rule, in some circumstances, would reach the opposite liability decision. What may be unjust about it, then, is that in the absence of a moral difference it could require the partyfree from fault to bear the costs of accidents caused by another's fault. I do not mean to suggest that one cannot envision instances in which the circumstances were such that it would not be unjust to require the faultless to bear the losses caused by another's fault. I mean to say only that in the absence of such circumstances, the absolute liability rule may be unjust. The same point could be put another way: what is wrong with absolute liability is that it requires an individual free from fault to 69. Id., at 217-21. HeinOnline -- 18 Wm. & Mary L. Rev. 284 1976-1977 1976] STRICT TORT LIABILITY 285 serve as the insurer of the faulty. Moreover, it imposes this obligation of insurer not by contractual agreement, but by liability rule. Although it may not be unjust for an individual to assume the role of an insurer by contract, it may be unjust to impose that role upon him through a liability rule in the absence of a standard liability trigger-mechanism, such as fault.0 The traditional view is that strict liability is an unjust theory of responsibility in torts because it does not allow the defense of freedom from fault to defeat an attribution of liability. I have argued that in both the fault and strict liability systems, an individual may be required to bear the costs of accidents that are not his fault. Moreover, if there is a moral shortcoming in strict liability, it is not an injustice in strict liability generally; instead, it is an injustice in the absolute liability rule only. It is the narrow injustice that one who is without fault is required, in the absence of a contractual agreement, to be the insurer of one who is at fault. Thus, absolute liability may violate the principle of justice that holds that in the absence of such an agreement, or conditions of similar effect, the party at fault should bear the loss. V The plain but almost universally unappreciated fact is that in tort law we are dealing with losses-with activities and their accident costs. The question with which we began is the question with which we conclude: Who should bear these costs? My argument has been that the preoccupation with fault as an essential element in a just 70. Fault-free loss bearing is justified when it is agreed to by contractual arrangement. Certainly no one has seriously argued that it is unjust for insurance companies to offer indemnification policies. If there is a counterargument, it is not based on the view that it is unfair to the insurance company to impose the burden on it. After all, the company contractually agreed to bear it. Instead, it arguably undermines the restorative and characterbuilding elements of tort law to allow persons to indemnify themselves against the costs of injuries their fault causes others to suffer. More to the point, the "contract" justification for absolute tort liability is at the core of products liability holdings. Specifically, in products liability it often is hard to identify whether it is the tort theory of absolute liability or the contract doctrine of implied warranty that supports compensation. In fact, the development of the absolute liability rule in products liability has its roots in contract analysis. See, e.g., Henningsen v. Bloomfield Motors, Inc., 32 N.J. 358, 161 A.2d 69 (1960); Baxter v. Ford Motor Co., 168 Wash. 456, 12 P.2d 409 (1932); Winterbottom v. Wright, 10 M. & W. 109, 152 Eng. Rep. 402 (1842). In part because of contractual origins of the tort doctrine of absolute products liability and the tort character of the contract doctrine of promissory estoppel, some recent scholarship has focused on the viability of the toit-contract dichotomy. The most impressive of these works is Grant Gilmore's excellent book, G. GILMonE, Tim DF,ATH o CoNRmcr (1974). HeinOnline -- 18 Wm. & Mary L. Rev. 285 1976-1977 WILLIAM AND MARY LAW REVIEW [Vol. 18:259 theory of torts has obscured significant structural similarities between the criteria of fault and strict injurer liability. Foremost among these similarities is that we can conceive of the fault system itself as a form of strict victim liability with the defense of injurer fault.' Once we conceive of traditional fault and strict injurer liability as structural analogues, the importance of another issue comes to the fore because these standards of liability differ primarily only vith regard to who bears losses in the absence of fault. The fault principle, whether it is employed to justify transferring losses from victims to injurers (in the fault system) or from injurers to victims (in strict injurer liability), cannot provide a moral reason for preferring one standard of liability to the other. That issue can be resolved only by first attempting to fashion general principles for allocating losses in the absence of fault. My arguments in this Article have been that we cannot decide, without reference to particular kinds of categories of cases, which no-fault decision to make. Indeed, once we consider cases we are likely to discover that in certain instances the strict liability rule is morally superior to the criterion of fault. For example, in workmen's compensation cases and products liability cases considerations not of compensatory justice but of distributive justice could support imposing the burden of bearing accident costs on the employer or manufacturer rather than on the employee or consumer. But a thorough inspection of cases is left for another occasion. My point is simply that by focusing on the presence of fault in a just theory of liability, moral philosophers largely have ignored the question of who should bear a loss when no candidate for it is at fault. Consequently, philosophers have stopped contributing to the dialogue at that very point where they are needed most. 71. This argument holds whether one conceives of the concept of fault in tort law as primarily moral or primarily economic. If one conceives of the fault concept as a quasi-moral one, as I do, that an injurer is at fault may constitute a good reason for his bearing the costs of accidents caused by his fault. As I have argued earlier here, see notes 16-18 supra & accompanying text, and elsewhere, see Coleman, The Fault Sytem, supra note 1, at 488-90; Coleman, Justice, supra note 11, at 173-78, that an injurer is at fault is a good reason for conferring on the victims of his fault a right to recompense. By itself, however, his fault does not imply that he ought to discharge the corresponding obligation to provide recompense. On the other hand, if, as Posner and Calabresi contend, see note 13 supra, fault is infected by cost-avoidance implications, pursuit of economic efficiency could well justify having those at fault bear the costs of accidents for which they are responsible. Neither conception of the tort concept of fault succeeds in justifying any particular allocation of accident costs in the absence of fault. HeinOnline -- 18 Wm. & Mary L. Rev. 286 1976-1977