+(,121/,1( Citation: 67 B.U. L. Rev. 213 1987 Content downloaded/printed from HeinOnline (http://heinonline.org) Wed May 19 14:57:55 2010 -- Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and Conditions of the license agreement available at http://heinonline.org/HOL/License -- The search text of this PDF is generated from uncorrected OCR text. -- To obtain permission to use this article beyond the scope of your HeinOnline license, please use: https://www.copyright.com/ccc/basicSearch.do? &operation=go&searchType=0 &lastSearch=simple&all=on&titleOrStdNo=0006-8047 VOLUME 67 NUMBER 2 MARCH 1987 BOSTON UNIVERSITY LAW REVIEW ASSUMPTION OF RISK AND CONSENT IN THE LAW OF TORTS: A THEORY OF FULL PREFERENCEt KENNETH W. SIMONS* TABLE OF CONTENTS I. A CRITIQUE OF THE MODERN CRITIQUE ............................... II. THE FULL PREFERENCE MODEL ....................................... A . THE BASIC M ODEL ............................................ 1. FULL RISK PREFERENCE ......................................... 2. WAIVER OF A TORT CLAIM DISTINGUISHED ..................... B. CLARIFICATION OF THE MODEL ..................................... D. E. III. 228 229 2. 3. REASONABLENESS IS LARGELY IRRELEVANT ...................... THE PREFERENCE CAN BE ATEMPORAL .......................... 233 235 4. MUST THE PREFERENCE BE "VOLUNTARY" 238 AND "KNOWING"? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . Is THE MODEL NECESSARY? VOLUNTARY ASSUMPTION OF A RELATIONSHIP DISTINGUISHED .... LIMITS OF THE MODEL ........................................ V. 239 241 244 247 COMPARING ASSUMPTION OF RISK WITH CONSENT TO AN INTENTIONAL T ORT ............................................................ IV. 224 1. THE NATURE OF THE FULL PREFERENCE ........................ 5. UNAVOIDABLY DANGEROUS ACTIVITY ........................... C. 215 218 218 218 THE MODEL IN PERSPECTIVE: TORT CASES AND SCHOLARSHIP ........ 248 258 SOME APPLICATIONS OF THE MODEL ................................... 269 A. B. COMPARATIVE FAULT ............................................... 270 INJURIES IN SPORTING EVENTS ..................................... 271 C . STRICT LIABILITY ............................................. V I. CONCLUSION ..................................................... APPENDIX: THE DEBATE OVER THE RESTATEMENT (SECOND) OF TORTS PROVISIONS CONCERNING ASSUMPTION OF RISK ................ 274 279 281 According to most commentators and a growing number of courts, the tort doctrine of assumption of risk should be abolished.' The dramatic emert © Copyright 1987 by Kenneth W. Simons. * Associate Professor of Law, Boston University. I thank Jane Cohen, Clayton Gillette, James Henderson, John Leubsdorf, and Paul Wallace for helpful criticism, and Thomas Gunning for valuable research assistance. See, e.g., 4 F. HARPER, F. JAMES & 0. GRAY, LAW OF TORTS§ 21.8, at 259 (2d 213 HeinOnline -- 67 B.U. L. Rev. 213 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 gence of comparative negligence in recent years requires courts to listen carefully to the modernists' plea for abolition. If, as the modernists claim, assumption of risk is either a limitation upon duty or a form of contributory negligence, then it bars recovery in the former guise but not in the latter. If, as the traditionalists claim, it is a viable independent defense, it might always bar recovery.2 The recent expansion of strict products liability also provokes close attention to the abolitionist cause, since only assumption of risk, and not contributory negligence, is a defense to strict liability, according to the traditional view. 3 The confident demands for abolition are tempting. After all, the modern trend simplifies existing doctrine and dovetails neatly with the larger move4 ment toward liberalizing recovery in tort actions. In this article, I suggest a cautious resistance to temptation. I propose that we take seriously the traditional notion that one who consents to a risk may not obtain damages when the risk materializes in harm. At the same time, I believe that traditional doctrine fails to isolate the proper type of choice between alternatives that should constitute assumption of risk-in brief, a plaintiff's true and full preference for the risky alternative that he chose over the alternative that defendant tortiously failed to offer. From this view, traditional doctrine deserves sharp revision, but not necessarily abolition. Some puzzles in existing tort doctrine suggest the need for a revisionistshall I say postmodern?-view of assumption of risk. Why has assumption of risk often served as a defense to strict liability claims, when contributory negligence has not? Consent is an accepted defense to an intentional tort; why then are we uncertain about the status of assumption of risk, which is ed. 1986) ("[T]he concept of assuming the risk is purely duplicative of other more widely understood concepts, such as scope of duty or contributory negligence."); C. MORRIS & C.R. MORRIS, JR., MORRIS ON TORTS ch. 8, § 5 (2d ed. 1980) (observing that when the defense applies even a defendant who is "at fault" may escape liability); V. SCHWARTZ, COMPARATIVE NEGLIGENCE 167-73 (2d ed. 1986) (discuss- ing merger of implied assumption of risk into contributory negligence); Gaetanos, Essay-Assumption of Risk: Casuistry in the Law of Negligence, 83 W. VA. L. REV. 471 (1981) (decrying the confusion assumption of risk has engendered); James, Assumption of Risk, 61 YALE L.J. 141, 169 (1952) ("Except for express assumption of risk . ..the term and the concept should be abolished."); Kionka, Implied Assumption of Risk: Does it Survive Comparative Fault?, 1982 S. ILL. U.L.J. 371, 376, 400-01; see also id. at 388 n.68 (listing cases demonstrating judicial trend toward abolition). Contra RESTATEMENT (SECOND) OF TORTS §§ 496A-496G (1965); W. P. KEETON, D. DOBBS, R. KEETON & D. OWEN, PROSSER & KEETON ON TORTS § 68 (5th ed. 1984) [hereinafter W. PROSSER & W. KEETON]. THE LAW OF See infra text accompanying notes 201-06. See infra text accompanying notes 221-29. 4 Abolition is ordinarily viewed as a victory for plaintiffs because reasonable assumption of a negligently created risk no longer bars recovery. See infra text accompanying notes 6-17. Plaintiffs are no better off, however, insofar as reasonable assumption of risk cases are reclassified as "no duty" cases. 2 3 HeinOnline -- 67 B.U. L. Rev. 214 1987 ASSUMPTION OF RISK 1987] often equated with consent? Why has the doctrine of assumption of risk persisted in many states 5 despite overwhelming criticism and the ascension of comparative fault? These questions are partially answered by my revisionist view of assumption of risk as full preference. Before delineating that view, and how it helps explain these puzzles, I will set forth the modern criticism of assumption of risk, and suggest how that criticism is deficient. 1. A CRITIQUE OF THE MODERN CRITIQUE Professor Fleming James has ably articulated the modernists' criticism of the traditional doctrine of implied assumption of risk 6-- the doctrine that a plaintiff's negligence claim should be barred if he has voluntarily encountered a known risk.7 In the following passage, he explains why he believes implied assumption of risk should be abolished: A plaintiff's reasonable assumption of risk would not bar him unless the risk was one which defendant had a legal right to put up to plaintiff; and in such a case defendant breached no relevant duty. A plaintiff's unreasonable assumption of risk would constitute contributory negligence on his part; and this would be a defense without the need to invoke any separate doctrine.8 The leading case of Meistrich v. Casino Arena Attractions, Inc. 9 adopts essentially the same analysis. In cases of "primary" assumption of risk, defendant is not negligent-either he owed no duty or he did not breach the duty owed. "Secondary" assumption of risk is an affirmative defense to an established breach of duty, and exists only if plaintiff was contributorily negligent.10 In Meistrich, defendant arena was negligent in the preparation of See, e.g., Riley v. Davison Constr. Co., 381 Mass. 432, 441, 409 N.E.2d 1279, 1283 (1980); Sandberg v. Hoogensen, 201 Neb. 190, 199, 266 N.W.2d 745, 750 (1978); Kennedy v. Providence Hockey Club Inc., 119 R.I. 70, 75, 376 A.2d 329, 332 (1977). 6 See James, supra note 1. Although he is an abolitionist, James acknowledges that express assumption of risk by contract remains, subject to possible contract limitations upon enforceability. Id. at 162-166. 7 Traditionally, assumption of risk is also a defense to claims of both recklessness and strict liability. See RESTATEMENT (SECOND) OF TORTS §§ 402A comment n, 496A-496G, 515, 523, 524 (1965); see also infra Part VC. James, Assumption of Risk: Unhappy Reincarnation, 78 YALE L.J. 185, 185 (1968) (footnotes omitted, emphasis in original) (restating Bohlen's views). 9 31 N.J. 44, 155 A.2d 90 (1959). 10Id. at 49, 155 A.2d at 93. Meistrich places in a separate category "an express contract not to sue for injury or loss ... and also situations in which actual consent exists, as, for example, participation in a contact sport." Id. at 48, 155 A.2d at 93. see also Blackburn v. Dorta, 348 So. 2d 287, 290 (Fla. 1977) (describing both situations as "express assumption of risk"). The separate treatment of "consent" is a serious mistake, as I will attempt to show. HeinOnline -- 67 B.U. L. Rev. 215 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 an ice rink for its patrons. After noticing the unusually slippery conditions, plaintiff continued to skate and was injured. Under the court's analysis, plaintiff would be barred from recovery only if defendant breached no duty to plaintiff," or if plaintiff was contributorily negligent. This modern form of analysis helps to explain cases that traditional analysis justifies weakly, if at all. Consider the famous case of Eckert v. Long Island R.R. 11 A jury had found that the railroad's engineer was operating the engine negligently, endangering a young child sitting on the track. Plaintiff's deceased valiantly attempted to rescue the child, but was killed in the effort. At trial, the court refused to instruct that if the deceased voluntarily placed himself in peril, he could not recover. The majority of the New York Court of Appeals upheld the refusal to grant this instruction, with one dissent. Remarkably enough-and to the perverse delight of some teachers of confused first-year torts students-the majority and dissenting opinions simply ignore each other. The dissent emphasizes that the deceased had assumed the risk, by voluntarily and knowingly encountering the risk. 13 The majority opinion virtually ignores assumption of risk, and is satisfied that the decedent's estate should recover because the deceased was not contributorily 14 negligent in voluntarily encountering the risk. The dissent in Eckert is correct: the rescuer did "assume the risk" in that he voluntarily and knowingly encountered it. Attempts to avoid this conclusion by characterizing his conduct as "involuntary" 15 are desperate fictions. And yet the majority is surely correct that the rescuer's meritorious conduct should not prevent or reduce his recovery. How can this tension be resolved? The modern solution, abolishing assumption of risk as a separate defense, works well here. If, as in Eckert, the plaintiff was not contributorily negligent then, by definition, his action was reasonable and his "secondary" assumption of risk will not prevent recovery. Likewise, if defendant owed and breached a duty to the plaintiff-a condition satisfied in Eckert 16 -then primary assumption of risk is also inapplicable, and plaintiff may recover. But this solution works poorly in other situations. Suppose I decide to engage in hang gliding. I join an outing organized by the defendant, who warns me of the serious but unavoidable risks. (I can avoid the risks only by " Perhaps plaintiff was an unusually talented skater who could skate easily under these conditions. 12 43 N.Y. 502 (1871). 13Id. at 506. 14 Id. at 505-06. 15 See RESTATEMENT (SECOND) OF TORTS § 496E illustration 3 (1965) (bystander's non-negligent rescue of blind man is not assumption of risk). 16 For simplicity, I refer to the rescuer in Eckert as "plaintiff." In fact, the rescuer was plaintiff's decedent. Furthermore, I do not analyze the court's assumption that the defendant owed a duty to the rescuers. HeinOnline -- 67 B.U. L. Rev. 216 1987 19871 ASSUMPTION OF RISK declining to engage in hang gliding altogether.) I am injured in a regrettably predictable way. I cannot recover damages. Why? Because I was negligent in encountering the risk? But then defendant was presumably negligent in offering me the choice to encounter it. The dubious result is that I might recover some portion of the damages under comparative negligence. Alternatively, perhaps I cannot recover because my conduct in encountering the risk was indeed reasonable-and so was defendant's conduct in offering me the choice to encounter it. But again, we have a difficulty. Would a reasonable person engage in such risky conduct? No doubt, one could interpret the "reasonable person" standard to approve of hang gliding. But it is more useful to focus on the distinctive aspect of this case. I truly prefer a risky alternative; I would rather have the opportunity to hang glide, with the necessary incidental risks, than be restricted to the opportunity of flapping my arms as I run along the ground. In Eckert, by contrast, presumably the plaintiff's decedent would rather not have been put in the position of having to rescue at all. The core of the traditional doctrine of assumption of risk, I believe, is the plaintiff's expressed preference for a risky activity as compared to a less risky alternative that defendant could and (the plaintiff now claims) should have offered. In this sense, assumption of risk should prevent the hang glider's claim, but not the rescuer's claim in Eckert. Consider another example, from the law of intentional torts. In athletic contests, participants are not infrequently injured. When can an injured participant successfully bring a damages action against the injurer? Doctrinally, the question is whether the plaintiff has actually or apparently consented to the harm. 1 7 But the notion of consent is notoriously elusive. Participants in a game know that injuries often occur. And they choose to participate in the game with that knowledge. Do they legally consent to the risks of all injuries that they can foresee? I seriously doubt that they do. If the most egregious examples of misconduct in a sport lead to successful liability suits-and they do-then many participants will become aware that the misconduct might occur. In a recent, well-publicized case, a professional football player was intentionally punched in the back of the head during a game by an angry opponent."8 Does every football player aware of the case now consent to such a contact? The problem with applying the notion of consent here is that, under the circumstances, the consent is quite limited. I would rather play professional football with the risk of such a brutal blow than not play at all. But I would also rather play and not run that risk. By the same token, the rescuer in 17 Although voluntarily encountering the risk of an intentional harm is traditionally classified as consent, a more apt characterization may be assumption of risk. See infra text accompanying notes 113-19. " See Hackbart v. Cincinnati Bengals, 601 F.2d 516 (10th Cir.), cert. denied, 444 U.S. 931 (1979). HeinOnline -- 67 B.U. L. Rev. 217 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 Eckert preferred rescuing to inaction; but undoubtedly he would have preferred that defendant had not negligently created the difficult predicament at all. On the other hand, much as I would like to play football without ever being hit by another player in violation of the rules, I know that enforcing such a policy strictly will drastically alter the game-to a degree that I may find unacceptable.. Thus, I may have consented to those lesser, more typical contacts. 19 The doctrine of assumption of risk is partially valid, then, if interpreted as a type of consent. But the concept of consent itself is ill-defined and requires further articulation. 20 The task of the remainder of this article is to propound a defensible theory of full preference that underlies both consent and assumption of risk. II. THE FULL PREFERENCE MODEL The Basic Model A. I have informally sketched the outline of a model of consent and assumption of risk. It is time to present the model more formally. 1. Full Risk Preference What justifies barring relief because of plaintiff's consensual behavior? Consensual behavior sometimes constitutes contributory negligence, of course, which may bar or reduce recovery. But the classic doctrines of consent and assumption of risk also have independent significance. When plaintiff asserts that defendant acted negligently or otherwise tortiously because defendant provided plaintiff with choice A rather than choice B, yet plaintiff actually prefers A to B, ordinarily there is no good reason to grant plaintiff recovery. This intuitively powerful proposition can be further justified in various ways: plaintiff should not obtain the benefits of a choice without incurring the expected risks; 21 sometimes defendant has not acted ' For a further discussion of consent to intentional torts in athletic events, see infra text accompanying notes 207-20. 21 See Rice, The Automobile Guest and the Rationale of Assumption of Risk, 27 MINN. L. REV. 323, 460-61 (1943) (suggesting that "consent" is no more helpful than "assumption of risk" in explaining the cases). 1 See Mansfield, Informed Choice in the Law of Torts, 22 LA. L. REV. 17, 73 (1961), which states the argument as follows: The plaintiff was satisfied enough with the defendant's conduct when he hoped to gain by it. Only when things have gone badly does he complain that the defendant has violated a duty owing him and demand compensation for the consequences. The plaintiff wanted something for nothing, but when a cost was exacted seeks to cast this upon the defendant. Mansfield, however, does not specifically approve of the argument. Id. at 74-75: see also infra text accompanying notes 175-79. HeinOnline -- 67 B.U. L. Rev. 218 1987 1987] ASSUMPTION OF RISK tortiously by offering such a choice;2 2 sometimes defendant relies upon plaintiff's choice; 23 plaintiff is the "co-author of his own harm; ' 24 individual choice will be undermined if it is not enforced against the later wishes of the 2 6 chooser;25 and recognizing autonomous choice increases social wealth. Often, however, defendant has acted tortiously. See infra Part IIC. But often he does not. Consider Eckert, for example, and other cases in which defendant's negligence precedes plaintiff's assumption of risk. See RESTATEMENT (SECOND) OF TORTS § 496C comment f (third category of assumption of risk) (1977) ("[Diefendant has already violated his duty to the plaintiff, by creating a dangerous condition or situation, which continues after the defendant's conduct has terminated."); R. Keeton, Assumption of Risk in Products Liability Cases, 22 LA. L. REV. 122, 163-65 (1961); Mansfield, supra note 21, passim. 24 R. Keeton, supra note 23, at 151-52. I doubt, however, that plaintiff's causal contribution to his own harm is a sufficient justification for true assumption of risk. After all, a plaintiff's inaction may express his choice, see infra text accompanying note 28, and a plaintiff often acts affirmatively without either being negligent or fully preferring the risk. Consider the cases in which plaintiffs injure themselves by their understandable frightened responses to defendant's tortious conduct. Consider also the cases in which plaintiff expresses only a limited preference for risk. 25 The argument might be that plaintiffs as a class are better off if their choices are irrevocable, at least when defendant has relied upon them. See Mansfield, supra note 21, at 25; see also Simons, Rescinding a Waiver of a Constitutional Right, 68 GEO. L.J. 919, 950 (1980) (discussing and rejecting claim that waiver of rights should be irrevocable to ensure that party carefully considers whether to waive). This argument is problematic. At the most obvious level, the individual plaintiff believes she is better off if she has the power to revoke a prior choice. One needs to explain why her individual welfare must be sacrificed to the welfare of the class. The argument might also be that the defendant's freedom not to act would unfairly be compromised. See Mansfield, supra note 21, at 25, 30, 41. This argument, however, may be best analyzed as an instance of voluntary assumption of a relationship. See infra Part IID. 26 See generally R. POSNER, THE ECONOMICS OF JUSTICE ch. 3 (1981). The argument that free choice has moral value in promoting autonomy and well-being is seriously flawed. In a recent article, Professor West argues powerfully against Judge Posner's view that consensual transactions maximize well-being. West, Authority, Autonomy, and Choice: The Role of Consent in the Moral and Political Visions of Franz Kafka and Richard Posner, 99 HARV. L. REV. 384 (1985). West emphasizes that we often lack the motivation to improve our well-being through choice; indeed, we might choose as we do in order to please authority, id. at 400, or to relinquish control over our life or well-being. Id. at 412. As she argues: Many ...consensual changes leave both individual and community not just worse off, but miserable. It is not obvious why we should assume that all of these consensual changes in the world are moral changes on the ground that they promote autonomy. It does not follow from the fact that coerced states are immoral by virtue of their coercive element that voluntary world states are of positive moral value by virtue of their voluntariness. Id. at 427. For Posner's reply and West's rejoinder, see Posner, The Ethical Significance of Free Choice: A Reply to Professor West, 99 HARV. L. REV. 1431 (1986); 22 23 HeinOnline -- 67 B.U. L. Rev. 219 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 Still, the proposition itself may be stronger and more persuasive than any potential justification for it.27 In the typical consent or assumption of risk situation, the plaintiff has three actual or hypothetical choices: (1) Not engaging in an activity (and also obtaining certain benefits (B1 )); (2) Engaging in the activity and encountering a tortiously created risk (and also obtaining certain benefits (B2)); or (3) Engaging in the activity and not encountering that risk (and also 28 obtaining certain benefits (B3 )). The argument in the typical assumption of risk case goes as follows. Plaintiff argues that defendant was negligent for presenting choice (2) and not (3).29 Thus, the skater in Meistrich asserts that defendant should have West, Submission, Choice, and Ethics: A Rejoinder to Judge Posner, 99 HARV. L. REV. 1449 (1986); see also Kelman, Choice and Utility, 1979 Wis. L. REV. 769, 782-95 (critiquing the assumption of neo-classical economics that consumers will be better off if they can choose goods freely, and noting that people often make dissatisfying choices, including those resulting from habit or duress). I agree with many of West's criticisms. My own thesis, however, is much more limited. I do not claim that all instances of full preference have positive moral value. I do not dispute the moral claim that plaintiffs should often be compensated by government (and in some cases by defendants) even when they have fully consented to a risk. Rather, I believe that a plaintiff's tort claim against a defendant should depend upon whether plaintiff fully preferred the risk in the narrow sense that I describe. 27 Cf. L. WITTGENSTEIN, ON CERTAINTY 17e (1969): "I know that I have never been on the moon." That sounds quite different in the circumstances which actually hold, to the way it would sound if a good many men had been on the moon, and some perhaps without knowing it. In this case one could give grounds for this knowledge .... I want to say: my not having been on the moon is as sure a thing for me as any grounds I could give for it. 28 The matrix is oversimplified. First, in choice (1), "not engaging" in an activity may well carry its own risks. In the typical case, however, the risks involved in abstention will be less than the risks of participation as per choice (2). Moreover, the model can be modified to describe only the particular risk that came to fruition as plaintiff's injury, so long as it is one of the aggregate risks within defendant's scope of duty owed to the plaintiff. Second, choice (3) may present other background risks beyond those created by defendant's tort. But this fact is unlikely to alter plaintiff's preference, since defendant's tort ordinarily adds to, rather than displaces, background risks. 29 Note that the rescuer in Eckert seems to be asserting that the defendant was negligent, not for offering choice (2) instead of (3), but for offering choice (2) instead of (1). In this case, option (3) apparently does not exist. Nevertheless, if the point of "engaging in the activity" is solely to save another's life, then the model may apply. The third option is the option to save another's life without risking your own. This is the option that the rescuer in Eckert really preferred-not option (1), the option to avoid risk to himself yet not save the child. HeinOnline -- 67 B.U. L. Rev. 220 1987 19871 ASSUMPTION OF RISK offered a properly prepared rink, not an improperly prepared one. A worker asserts that the employer should have supplied nondefective rather than defective machinery. A passenger asserts that the automobile manufacturer should have designed the seat brackets not to break off in a low-speed collision. A jaywalking pedestrian asserts that the driver should have kept a closer look-out. The defendant then asserts that, even if he was negligent, plaintiff assumed the risk by choosing (2) over (1). After all,. the skater could have asked for his money back. The worker could have stopped working. The passenger, if duly warned of the defect, could have decided not to ride in the car. The jaywalker could have waited for the "walk" signal. Note what the defendant does not, and often cannot, assert: that plaintiff preferred (2) over (3) as well as over (1). This is the nub of the problem. The defendant's minimal assertion is too weak to justify an assumption of risk defense. Plaintiff may have preferred option (2) to option (1), but in the above examples, he would have preferred (3) overall. Plaintiff's limited "consent" to or assumption of the risk does not justify non-liability. To be sure, sometimes choosing (2) over (1) will be contributory negligence. But sometimes it will not. In any event, the classic consent and assumption of risk doctrines, if understood to reflect true full preference, simply do not support the defendant here. Reconsider, however, the hang gliding example above. Suppose this sport could be made safer by placing an iron cage around the glider. This would also diminish the participant's enjoyment, it is fair to say. If the plaintiff would prefer the sport as it is (2) over this safer alternative (3), then he has truly assumed the risk. Such cases, though rare, do exist. Perhaps the skater in Meistrich actually prefers skating on "improperly" prepared ice, compared to "properly" prepared ice, because of the greater thrill. Perhaps the pedestrian enjoys dodging cars, especially those negligently driven. (The analogous possibilities for the worker or the passenger are even more implausible.) The traditional view gives assumption of risk too broad a scope, barring a plaintiff who has made only a limited choice. Plaintiff may have chosen (2) over (3) only because it is the best of the currently available options. I'd rather skate on improperly prepared ice than not skate; I'd rather work with defective machinery than not work at all; I'd rather ride in a car with defective seat brackets than walk home; I'd rather not wait for the "walk" signal. The real question, however, is whether plaintiff prefers that option in a full sense over the option that the defendant has tortiously failed to offer him, i.e., whether plaintiff prefers (2) over both (1) and (3).30 The traditional 30 Plaintiff's full preference refers to the option that defendant should have offered. It is not enough that plaintiff prefers the risk inherent in another mode of conduct, if defendant did not create that risk. For example, some of those who fail to use seat belts prefer taking the risk of injury to the inconvenience and the psychic and symbolic costs of buckling up. But I doubt HeinOnline -- 67 B.U. L. Rev. 221 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 broad view of assumption of risk carries a rhetorical danger, since it places plaintiff on the defensive even though he has "chosen" to accept a risk only in a limited sense. Moreover, it causes courts (such as the New York Court of Appeals in Eckert) to avoid, or commentators to distort, assumption of risk doctrine. At the same time, the modem view of assumption of risk is too narrow, for it fails to account for those cases, admittedly few, in which the plaintiff truly prefers the option that he has chosen over the option he claims the defendant 31 should have offered. The argument in the typical case of consent to an intentional tort is even more straightforward. If I agree to a friendly wrestling match with you, I had the choice of not wrestling [option (1)] or wrestling with a known risk of some injury [option (2)]. In the usual case, there is no third option that I can insist you should have offered. 32 Thus, my preference for (2) over (1) will be full in the relevant sense. There are complications, of course. How does one determine which intentional invasions and which risks I agreed to? Can a defendant sometimes be faulted for not offering a third option? Suppose you post a warning that a deadly spring gun protects your garden, and I decide, notwithstanding the warning, to make a gallant rescue of a pea-hen in the garden.33 My consent is that they prefer to confront negligent rather than nonnegligent drivers. Therefore, they do not assume the risk of injury under my model. They might, of course, be negligent. 3' The analysis can also be applied when defendant fully prefers that plaintiff act negligently. By analogous reasoning, defendant's consent should preclude reduction of the plaintiff's recovery. For example, an amusement park operator should not be permitted to assert the negligence of customers who smash bumper cars against each other in a manner that the operator advertised and invited. Some of the early worker injury cases denying recovery are troubling not only because of their broad view of assumption of risk, but also because, insofar as the employee's conduct may have been careless, the employer may have invited or accepted that carelessness to benefit his business. Cf. Micallef v. Miehle Co., 39 N.Y.2d 376, 348 N.E.2d 571 (1976) (employee's customary but dangerous practice of "chasing hickies" in a printing press benefitted the employer (though not defendant manufacturer) by avoiding shutting down the printing process). 32 Indeed, Prosser suggests that consent "is not, strictly speaking, a privilege, or even a defense, but goes to negative the existence of any tort in the first instance." W. PROSSER & W. KEETON, supra note 1, at 112. I demur. There is an intentional invasion of plaintiff's interests, whether or not he consents, just as defendant may negligently create a risk, whether or not plaintiff assumes it. In each case, the legal conclusion that plaintiff cannot recover could be expressed in terms of "no tort" or "no duty." But it is clearer to identify consent or assumption of risk as the specific reason for that conclusion. See in fra text accompanying notes 89-94. 33 Cf. Bird v. Holbrook, 4 Bing. 628, 130 Eng. Rep. 911 (1828) (where no warning was given). HeinOnline -- 67 B.U. L. Rev. 222 1987 1987] ASSUMPTION OF RISK limited here, in the relevant senses: I prefer that you had not, and can insist that you should not have, placed me in this predicament. The example differs from typical instances of consent to an intentional tort, such as the wrestling example, because ordinarily I cannot insist on the third option. But these distinctions between negligence and intentional torts are breaking down to some extent, as discussion of patients' informed consent to bat36 35 tery3 4 and athletes' consent to athletic injuries will suggest. One sign of the inadequacy of the modern framework for discussion of assumption of risk is the controversy over the proper treatment of "reasonable" assumption of risk. This is the key to the modern debate, according to many commentators.3 7 The traditionalists think that "reasonable" assumption of risk should be a bar, the reformers that it should not. Neither view is correct. What is often called "reasonable" assumption of risk is a preference between the two currently available choices (choices (1) and (2) in the model). This limited preference is not tantamount to a full preference for the chosen option over all alternatives, including the alternative of engaging in the activity with less or no risk (choice (3)). When "reasonable" assumption of risk is only a limited preference, then I agree with the reformers that it should not be a bar. But when it reflects full preference, then it ordinarily should be a bar. The traditionalists' examples of "reasonable" assumption of risk typically reflect only limited preference. The Restatement (Second) of Torts, which recognizes "reasonable" assumption of risk as a bar, gives the example of an employer who furnishes an independent contractor with a machine in a dangerous condition, which the employer fails to repair. 3 8 The contractor may not be negligent in continuing to work with the machine, because the risk may be relatively slight in comparison with the utility of his own conduct, but, according to the Restatement, he should still be denied recovery. 39 I disagree. The contractor's decision only reflects limited preference. He would have preferred to work with nondangerous machinery. Thus, his 40 consent to take the risk should be insufficient to bar recovery. See in fra text accompanying notes 57-60. 35 See infra Part VD. 36 When defendant acts recklessly, the doctrine of assumption of risk has traditionally applied. See RESTATEMENT (SECOND) OF TORTS §§ 496A-496B (1965). My narrower full preference theory applies as well. 37 RESTATEMENT (SECOND) OF TORTS § 496A comment d; id. at § 496C comments e, f, g (1965); W. PROSSER & W. KEETON,supra note 1, at 493-95; James, supra note 34 8, at 188-92; Twerski, Old Wine in a New Flask-Restructuring Assumption of Risk in the Products Liability Era, 60 IOWA L. REV. 1, 1 (1974). 38 RESTATEMENT (SECOND) OF TORTS § 496A comment c (1965). 39 Id. 40 Consider a second example from the Restatement, which may or may not be intended as an example of "reasonable" assumption of risk: HeinOnline -- 67 B.U. L. Rev. 223 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 Similarly, Judge Robert Keeton posits the following case: [T]he plaintiffs Black and Blue are persons to whom the purchaser of a defectively designed motorcycle lends it, after discovering the defect and with full warning to Black but not to Blue, the plaintiffs having need of a vehicle and reasonably choosing to take this, the only vehicle available. If... both plaintiffs could overcome any arguments that the defendant manufacturer's duty was one of warning only. . ., Blue, who did not know of the defect, would recover. As in the case of Blue, Black was not contributorily negligent, but, having known of the defect and 41 having fully appreciated the danger, he consented to the risk. In this case, too, Black's consent was limited. Again, the full preference theory would not bar his recovery. 2. Waiver of a Tort Claim Distinguished Consent, in the sense of full preference, is the most viable concept underlying assumption of risk and intentional tort consent doctrine. But the term consent can be misleading here. Sometimes a plaintiff specifically agrees not to hold defendant liable for conduct that would otherwise be tortious. Such waivers will often be enforced, especially when they are in contractual form. But neither consent to an intentional tort nor assumption of the risk of negligent and other tortious behavior usually reflects this form of agreement-even if "agreement" encompasses implicit agreements or unilateral expressions of intent. The suggestions of many commentators notwithstanding, 42 people typically give little or no thought to the liability A, the owner of a baseball park, is under a duty to the entering public to provide a reasonably sufficient number of screened seats to protect those who desire it against the risk of being hit by batted balls. A fails to do so. B, a customer entering the park, is unable to find a screened seat, and although fully aware of the risk, sits in an unscreened seat. B is struck and injured by a batted ball. Although A has violated his duty to B, B may be barred from recovery by his assumption of the risk. RESTATEMENT (SECOND) OF TORTS § 496C illustration 4. I see no firm ground for barring recovery here if B would have preferred a screened seat, and if B was not negligent in remaining in his chosen seat. However, if B actually preferred an unscreened seat to a screened seat, aware of the concomitant risks, then his preference should ordinarily bar recovery. 41 R. Keeton, supra note 23, at 157-58. 42 See RESTATEMENT (SECOND) OF TORTS§ 496C comment b (1965) (when plaintiff "impliedly" assumes the risk, he "manifests his willingness to accept it, and to take his chances as to harm"); W. PROSSER & W. KEETON, supra note 1,at 484-86; see also Kionka, supra note 1, at 381-86 (supporting Prosser's argument and viewing true assumption of risk as a consent to waive liability). Most British commentators appear to suggest that consent to waive a tort claim is the proper scope of assumption of risk, and they criticize courts for not fully adopting this narrow theory. See G. WILLIAMS, JOINT TORTS AND CONTRIBUTORY NEGLI- HeinOnline -- 67 B.U. L. Rev. 224 1987 1987] ASSUMPTION OF RISK consequences of behavior that courts characterize as consent or assumption of risk. Rather, courts characterize the behavior as consensual because they conclude that such a plaintiff is not entitled to relief, without regard to 43 whether he intends to waive his tort claim. Prosser, in a well-known passage, demonstrates this confusion: It is not true that in any case where the plaintiff voluntarily encounters a known danger he necessarily consents to any future negligence of the defendant. A pedestrian who walks across the street in the middle of a block, through a stream of traffic travelling at excessive speed, cannot by any stretch of the imagination be found to consent that the drivers shall not use care to watch for him and avoid running him down. On the GENCE 307-08 (1951) (noting that the defense should be very rare because it reflects consent to bearing the legal, in addition to the physical, risks); Ingram, A History of the Defence of Volenti Non Fit Injuria, 26 JURID. REV. i, 27 (1981); Jaffey, Volenti Non Fit Injuria, 44 CAMBRIDGE L.J. 87, 88-91, 100, 109 (1985); Morrison, The "Volens" Principle, I SYDNEY L. REV. 77 (1953); see also Nettleship v. Weston, [1971] 3 W.L.R. 370, noted in Goodhart, Notes, 87 LAW Q. REV. 444, 444 (1971) (holding that driving instructor did not assume risk of injury from inexperienced driver because he had not agreed to waive any claim for negligence). True to the waiver of liability theory, Professor Jaffey is forced to conclude that a plaintiff is much less likely to have assumed the risk if he knows that defendant carries insurance. Jaffey, supra, at 102, 104. Certainly no American court would find insurance coverage relevant to assumption of risk. Fleming properly criticizes the very narrow "agreement to waive a legal right" test as "both psychologically unrealistic and incompatible with a course of decisions which have found plaintiffs volentes on the basis of far less positive conduct." J. FLEMING, THE LAW OF TORTS 269 (6th ed. 1983). Unfortunately, he fails to offer a clear alternative. His suggestion that the judicial test is consent "not merely to the risk of injury, but also to the lack of reasonable care which may produce that risk," id. (footnote omitted), hits closer to the mark. 43 In the criminal procedure literature, the same distinction is drawn between losing a right by a knowing and voluntary waiver, and by a forfeiture that need not express the individual's actual consent to the loss. See Westen, Away From Waiver: A Rationale for the Forfeiture of Constitutional Rights in Criminal Procedure, 75 MICH. L. REV. 1214, 1214-15, 1254-61 (1977); see also Simons, supra note 25, at 920. The confusion with true agreement is understandable, since plaintiff does voluntarily and knowingly accept a risk. But this is still short of voluntarily and knowingly waiving a tort claim. See R. EPSTEIN, A THEORY OF STRICT LIABILITY 99 (1980); James, supra note 1,at 162; R. Keeton, Assumption of Risk and the Landowner, 22 LA. L. REV. 108, 110-1I, 117 (1961); R. Keeton, supra note 23, at 146-49 (distinguishing "consent to exculpation" from "consent to risk" and from consent to intentional conduct). In a recent article, Professor Barnett defends a consent theory of contract based on the parties' manifestation of an intention to alienate rights. Barnett, A Consent Theory ofContract, 86 COLUM. L. REV. 269, 300-09 (1986). This view is quite similar to the narrow "waiver" conception of consent and assumption of risk in tort law. Indeed, perhaps that conception is not really a tort conception at all. HeinOnline -- 67 B.U. L. Rev. 225 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 contrary, he is insisting that they shall. This is contributory negligence pure and simple; it is not assumption of risk. And if A leaves an automobile stopped at night on the travelled portion of the highway, and his passenger remains sitting in it, it can readily be found that there is consent to the prior negligence of A, whose control over the risk has terminated, but not to the subsequent negligence of B, who thereafter runs into the car from the rear. In both cases, the plaintiff has exposed himself to the risk of future harm, but he has not consented to relieve the defendant of any future duty to act with reasonable care. This is a distinction which has baffled a great many law students, some judges, 44 and unhappily a few very learned legal writers. The distinction is baffling indeed. So is Prosser's confidence in its merit. Darting into a street with knowledge of the risks is a classic instance of assumption of risk, as conventionally understood. 45 The "consent" that Prosser insists is critical appears to cover only a very narrow categorywaiver of a tort claim that plaintiff would otherwise have against the defendant. 46 Considering the vehemence with which Prosser argued for the reten- 44 W. PROSSER & W. KEETON, supra note 1, at 485 (footnotes omitted, emphasis added); see also RESTATEMENT (SECOND) OF TORTS § 496C comment h (1965); W. PROSSER, THE LAW OF TORTS 445 (4th ed. 1971) (Prosser, the Reporter for the Restatement, makes the same point in the earlier edition). 45 Indeed, does Prosser's distinction undermine the express waiver category, as well? Why can it "readily be found that there is consent to the prior negligence of A"? The passenger has decided to stay in the car, but it does not follow that he has agreed not to sue A, or that he approves of A's negligence. See also G. WILLIAMS, supra note 42, at 309, whose examples show the difficulty of deciding when to infer "agreement" from the conduct of encountering a known risk. Williams, like Prosser, seems to sidestep the difficulty by constructing an artificial concept of -'agreement." Perhaps Prosser means to distinguish past negligence, about which the defendant can now do nothing, and future negligence, which the defendant as well as the plaintiff can avert. If Prosser meant to endorse some such variation on last clear chance, he should have explained why. I remain baffled. Finally, perhaps Prosser is arguing indirectly for the full preference theory. In the jaywalking example, the pedestrian prefers risking injury to waiting for the light, but no doubt he also would prefer that the drivers not proceed negligently. (Needless to say, I agree with Prosser's conclusion that the jaywalker should not be barred by assumption of risk.) However, Prosser's example of the passengers in the car assuming a risk is inconsistent with the full preference theory. 46 Concededly, Prosser is ambiguous about this. At times he speaks of relieving defendant of responsibility and taking one's chances, W. PROSSER & W. KEETON, supra note I, at 480-81, which sounds like a waiver of plaintiff's tort claim against defendant, at other times he speaks of not insisting that defendant exercise due care, id. at 485, 490, which might differ from a waiver. While both interpretations are much narrower than traditional assumption of risk, the second is closer to a full preference. This ambiguity has muddled discussions of assumption of risk from the very inception of the doctrine. In what was apparently the first sustained academic HeinOnline -- 67 B.U. L. Rev. 226 1987 1987] ASSUMPTION OF RISK tion of assumption of risk in the Restatement (Second) 47 it is remarkable 48 that he fought for so little. Similarly, the Restatement (Second), which retained assumption of risk after strenuous debate, 49 gives the justification that a plaintiff who voluntarily encounters a known risk thereby agrees to take her chances. ° But the nature of this "agreement" is woefully ambiguous. Does it mean that plaintiff has waived a tort claim? This interpretation is often highly artificial. Does it mean that plaintiff prefers taking her chances to withdrawing from the risk? Such a preference is only a limited consent, i.e., for (2) over (1) in my model. It does not, in my view, warrant barring plaintiff's claim. Does "agreeing to take your chances" describe a full preference, then? Does it suggest that you prefer encountering the risk (2) to engaging in your activity with the tortious risk removed (3)? Sometimes it may. Only in these cases 51 does the Restatement (Second) coincide with the proposed model. Even consent to intentional torts is typically less than a knowing waiver of an otherwise valid tort claim. A patient who consents to an operation, thereby relieving the doctor of liability for battery, does not usually consciously agree to refrain from suing for battery. No doubt she would ordinarily be quite surprised if she did retain a claim for battery. But consent and 52 waiver of a tort claim remain distinct concepts. discussion of the doctrine, Professor Warren adopted the traditional definition, voluntary exposure to a known risk. Warren, Volenti Non Fit Injuria in Actions of Negligence, 8 HARV. L. REV. 457, 459 (1895). Warren justified the doctrine on this ground: "You have hence shown that you agree to take your chances of the danger." Id. at 460; see also id. at 461 (the question is whether plaintiff "had consented... to run the chances ...of [the] danger"). Again, the justification is ambiguous: it could reflect either waiver or full consent. 47 See the Appendix for a synopsis of these debates. 48 See also Morison, supra note 42, at 83 ("There is no occasion to abolish the general rule [of assumption of risk], for its interpretation [as consent to relieve defendant of liability] has rendered it innocuous."). 19See the Appendix. 50See RESTATEMENT (SECOND) OF TORTS § 496A comment c, proposition 2; comment d; id. at § 496C comments b, h; id. at § 496E comment a. 51 But it is doubtful that the Restatement is meant to apply only in these cases. Indeed, many of the examples are inconsistent with the full preference model. See § 496A illustration 2 (B assumes risk When he knowingly accepts ride with drunk driver): id. at illustration 4 (B assumes risk when he accepts ride with incompetent driver): id. at § 496C illustration 3 (B assumes risk of obviously slippery floor and open trap door), illustration 4 (B assumes risk of unscreened seat, even though he would apparently prefer a screened seat). 52 An express waiver of a tort claim is itself a type of consent, of course-namely, consent not to hold the other party liable. Such a waiver may or may not coexist with consent to an intentional contact or to an unintentional injury. A patient might sign a form relieving a doctor of all potential liability, yet might not consent to an operation, nor to risks of injury from a doctor's negligence. She might object to the operation, HeinOnline -- 67 B.U. L. Rev. 227 1987 BOSTON UNIVERSITY LAW REVIEW B. [Vol. 67: 213 Clarification of the Model Several additional points about the model deserve mention. At the start, I should clarify the nomenclature. The model of full risk preference does not assume that the plaintiff in question is a "risk-preferrer," as that phrase is used in economic and public choice theory. A "risk-preferrer," unlike a risk-neutral person, would choose an uncertain detriment over a certain detriment even when both were of equal expected value.53 Instead, the model posits only that the agent would choose54 the tortious over the but agree not to sue for battery. Or she might consent to the operation, yet actually believe that she is entitled to sue for any negative consequences (say, on a strict liability or contract theory). Furthermore, a waiver of a tort claim need not require much knowledge of the particular risks waived, as long as it satisfies the usual contract rules for valid consensual agreements. See R. EPSTEIN, supra note 43, at 97-98. But see G. WILLIAMS, supra note 42, at 311. (Williams also believes that the waiver need not satisfy the usual rules of contract. Id. at 312. He does not explain why not.) m For example, suppose you are offered a choice-either you lose 500 outright or you lose $1.00 if you call the toss of a coin incorrectly; or, either you gain 50g outright or you gain $1.00 if you call the toss correctly. In each case, a risk-neutral person would be indifferent, while the risk-preferrer would choose the second and more uncertain option. See generally K. ARROW, ESSAYS IN THE THEORY OF RISK-BEARING (1974); see also C. GOETZ, LAW AND ECONOMICS: CASES AND MATERIALS 75-82 (1984); A. POLINSKY, AN INTRODUCTION POSNER, ECONOMIC ANALYSIS OF LAW To LAW AND ECONOMICS 51-53 (1983); R. 75-77 (2d ed. 1977). 5 In referring to a plaintiff's "preferences," I am sensitive to Professor Sen's admonition: [T]here are several related but different statements about a person's interests, actions, etc., that need to be distinguished, even though they are often identified in the literature: (1) the person gets more satisfaction in state x than in state y (statement about satisfaction or pleasure); (2) the person thinks that he or she is better off with x than with y (statement about introspective welfare); (3) the person is better off with x than with y (statement about individual welfare that may or may not be introspective); (4) the person prefers that x rather than y occurs (statement about the mental condition of preference, or desire, regarding states); (5) the person would like to so choose that x rather than y occurs (statement about desired choice); (6) the person believes that it would be right to so choose that x rather than y occurs (statement about normative judgment regarding choice); (7) the person believes that it would be better if x were to occur rather than y (statement about normative judgment regarding states of affairs); (8) the person so chooses that x rather than y occurs (statement about actual choice). None of these statements logically entails any of the others, and it is a matter for substantive empirical or normative analysis to check how in particular cases any two of these statements link with each other. HeinOnline -- 67 B.U. L. Rev. 228 1987 1987] ASSUMPTION OF RISK nontortious alternative, even though others might not. Technical "riskpreference" need not explain the choice. 1. The Nature of the Full Preference The critical question, in my model, for both assumption of risk and the intentional tort doctrine of consent is the fullness or comprehensiveness of plaintiff's preference of a course of conduct. In a later section, I will examine intentional tort consent doctrine more carefully. 55 Here I will try to clarify what counts as a full preference. An initial problem is determining what plaintiff has consented to. Often plaintiff agrees to the particular conduct of a defendant, or to a particular result. Does plaintiff also necessarily agree to all further consequences of that conduct or result, or to similar conduct or a similar result? Such questions arise in the simplest of cases. You agree to resolve your argument with defendant by fighting it out. What kind of weapons and what degree of force have you agreed to? And have you agreed to fight, but not to be injured? In part these questions are factual: What were your actual prefer56 ences? What did defendant reasonably interpret your preferences to be? (I will discuss apparent consent later.) But these questions are not simply an inquiry into a person's actual or apparent state of mind. For in a literal sense, plaintiff's preference will usually be to obtain the benefits of the contact or activity without having to suffer any risk. I would prefer not to be injured in the fight. I would prefer not to suffer a harmful side effect from a beneficial operation. I'd prefer to hang glide without injury. How can consent be reconciled with such preferences? As a start, consent must be given a more practical interpretation than the abstract set of preferences that plaintiff would like to have satisfied in an ideal world. In this world, difficult choices must be made. If I want the beneficial operation, SEN, CHOICE, WELFARE AND MEASUREMENT 8-9 (1982) (footnote omitted). The concept of preference most relevant here is close to statement (4). However, I am not certain that the preference must be a state of mind or "mental condition" in the narrow sense. You may truly prefer that defendant had never put you in the predicament of having to undertake a rescue, yet you may never have given the option a moment's thought. A preference is what you do, or would do, given a choice. Although it is an expression of desire, it is not necessarily an expression of a conscious thought process. Moreover, the preference need not be a strong desire; willingness or acceptance captures the mood as well. See Mansfield, supra note 21, at 34. In sum, many different states of mind-conscious or not, strong desire or merely a shade above indifference-are consistent with "preferring" option A to option B. 55 See infra Part Ill. 56 See RESTATEMENT (SECOND) OF TORTS § 892A comments c-i (1965) (discussing the factual scope of consent); see also R. Keeton, supra note 23, at 128 (distinguishing consent to conduct or condition involving an obvious risk from consent to the incidental risks of that conduct or condition). A. HeinOnline -- 67 B.U. L. Rev. 229 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 some risks of harm are inevitable. If I have been adequately informed of those risks, my decision to have the operation may be deemed a consent to all of them. My choice is a package; if I prefer this necessary combination of benefits and risks over no operation, then my preference is sufficiently full to be a binding consent. But this practical interpretation only takes us so far. Sometimes the plaintiff can have his cake and eat it too, i.e., obtain the benefit and incur a lesser risk. The plaintiff's consent need not be a bar in this situation if defendant is negligent or otherwise tortious for not creating the lesser risk. Suppose you are rushed to the hospital for an emergency operation. The only doctor available, although employed to render such emergency treatment, may be incompetent to perform the particular operation. The doctor accurately informs you that if he performs the operation, it will pose a 5% risk of a serious side-effect. Suppose a fully competent doctor could perform the same operation with only a 1%risk. In this case, your limited consent to the 5% risk should not bar your recovery because the hospital has failed to offer you a choice of alternatives consistent with its duty to provide qualified 57 emergency service. Suppose, however, you could obtain the benefit and incur a lesser risk, but defendant is not negligent or otherwise tortious for not providing the choice of a lesser risk. Suppose there are a handful of expert surgeons who could have performed the operation with only a 0.1% risk, but the cost would be so astronomical that the hospital is not negligent for failing to offer that choice. If you prefer that option, but the doctor fails to inform you of it, can you 5 successfully sue because your consent was not informed? 57 If instead you have a choice between the incompetent and the competent doctor, and you prefer the first because his fee is lower, then your choice is relevantly full. Thus, the model by itself would not permit recovery. However, other principles not directly addressed by the model-including a requirement that the consent be sufficiently knowing and voluntary, and principles that override consent in some circumstances, see infra text accompanying notes 68-74, 132-36, would undoubtedly permit recovery. See, e.g., Kozan v. Comstock, 270 F.2d 839, 845 (5th Cir. 1959) (holding that physician cannot contract to free himself from liability for negligence). But see Epstein, Medical Malpractice: The Case for Contract, 1 AM. B. FOUND. RES. J. 87 (1976) (arguing that private contract law permits a physician to limit her liability). 51 There is some question whether informed consent cases are properly treated as negligence or battery. The prevailing view treats them as negligence, see W. PROSSER & W. KEETON, supra note 1, at 189-90; Shultz, From Informed Consent to Patient Choice: A New ProtectedInterest, 95 YALE L.J. 219, 224-27 (1985), except "in those relatively unusual situations where a medical procedure has been carried out without any consent." Id. at .226. The analysis in the text helps.explain this view. Intentional tort doctrine protects plaintiff's freedom to decline any intentional invasion more strictly than does negligence law. But once plaintiff agrees to accept some invasion, the strict protection relaxes into the more flexible protection of negligence law. See infra Part II1. HeinOnline -- 67 B.U. L. Rev. 230 1987 19871 ASSUMPTION OF RISK The answer here is no. But the answer depends on the scope of defendant's duty as well as the scope of plaintiff's consent. The full preference model alone cannot account for this result, 59 since plaintiff's limited preference for an operation performed with a 0.1% risk is not a full preference. Rather, the best explanation is that defendant's duty to plaintiff, even in the intentional tort context, is limited. Although plaintiff can insist that defendant not perform an operation that plaintiff prefers to forego, plaintiff cannot insist that, if defendant performs the operation, he perform it with more than the requisite degree of care. This is a result not merely of plaintiff's consent but also of defendant's duty: the plaintiff simply cannot 60 ask so much of such a defendant. 59 Sometimes, to be sure, defendant can rely upon apparent consent. See infra text accompanying notes 125-33. A defendant might reasonably believe that a plaintiff fully consents if she has been adequately warned about only those benefit or risk alternatives that it would be negligent for defendant not to offer. Yet defendant cannot always rely on such apparent consent; for example, he may know that plaintiff is not paying her own bill, and that plaintiff would prefer a socially unjustifiable degree of care. 60 If defendant has reason to know that plaintiff would consider extraordinarily remote risks important to her decision whether or not to undergo treatment, defendant may sometimes be required to warn plaintiff of these risks. The informed consent doctrine defines the scope of the duty. But that is different from plaintiff relying on the "full consent" doctrine to insist that defendant actually perform an operation that it is unreasonable to ask defendant to perform. Informed consent, like other duty to warn doctrines, places unusual emphasis on plaintiff's actual preferences. The ultimate question is not simply what a reasonable defendant would do, but whether plaintiff has agreed (however foolishly) to the treatment. Informed coisent does examine the reasonableness of defendant's belief that plaintiff has knowledgeably consented, but that differs from ignoring plaintiff's consent entirely and asking only whether defendant acted reasonably in deciding what information to disclose. See infra note 129. Insofar as we examine what a reasonable plaintiff would find relevant to the decision, see Canterbury v. Spence, 464 F.2d 772, 787 (D.C. Cir.), cert, denied, 409 U.S. 1064 (1972), we may still be inquiring about defendant's reasonable beliefs. In short, apparent consent is the usual framework. Many courts award damages in an informed consent case only if a "reasonable patient" would have declined the treatment if properly informed of the risks. See W. PROSSER & W. KEETON, supra note 1, at t91. Under this approach, a plaintiff with unusual preferences who would not have consented to the operation might not be able to recover simply because a "reasonable" plaintiff would have consented. This result is difficult to justify if the autonomy and idiosyncratic preferences of the individual plaintiff are the critical values that informed consent doctrine serves. See Shultz, supra note 58, at 288. However, the proof problems that a more subjective test of causation would create do help explain this objective test. See, e.g., Canterbury, 464 F.2d at 790-791. The same problem arises in cases involving a duty to warn of product defects. See Reyes v. Wyeth Laboratories, 498 F.2d 1264, 1281-82 (5th HeinOnline -- 67 B.U. L. Rev. 231 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 When plaintiff asserts that defendant has acted negligently, and intentional tort liability is not an issue, it may be even clearer that a plaintiff's full preference for certain non-negligent alternatives is irrelevant, because those alternatives are beyond defendant's duty to offer. I, as a pedestrian, may truly prefer that all automobile drivers limit their speed to five miles per hour, but my preference is relatively full if I choose to walk alongside traffic 6 that is proceeding at a non-tortious (non-negligent) speed. 1 Finally, suppose a plaintiff fully prefers a risk, but also asserts that. he prefers to be able to recover damages if the risk materializes. (As a factual matter, the latter preference is more likely than its converse, which is a true consent to waive a claim against defendant.) Why not add this as a general fourth option in the model? A plaintiff who prefers (2) over (1) and (3) will often fully prefer the following option: (4) Engaging in an activity, encountering a tortiously created risk, obtaining certain benefits (B,), and receiving compensation for the risk if it 6 materializes. Cir. 1974) (rebuttably presuming that plaintiff would have declined vaccination if properly warned). 61 A full preference for certain alternatives may also be irrelevant in strict liability cases-though here, there is a complication. See infra Part VC. 62 Often, of course, a plaintiff who fully prefers (3) to (2) will also prefer (3) over (4)-that is, he would prefer to avoid defendant's tortious act entirely rather than incur an injury for which he is compensated. In one sense, this last preference merely suggests that the usual tort measure of compensation is inadequate. In a deeper sense, however, it suggests that we view most tort rights seriously-as personal rights of the plaintiff that should not be violated. We do not believe that a defendant, by his unreasonable conduct, should be permitted to "condemn" or "take" a plaintiff's interest in bodily integrity or in freedom from physical or emotional harm. Cf. Ingraham v. Wright, 430 U.S. 651, 696-97 (1977) (White, J., with whom Brennan, Marshall, and Stevens, JJ., join, dissenting) (objecting that majority's emphasis on availability of lawsuit for unjustified paddling in school would impermissibly authorize the state to deliberately violate constitutional rights and then remit citizen to a subsequent damage action). When the tort theory is strict liability, however, and when plaintiff cannot demonstrate that defendant should have acted differently, then in effect defendant is permitted to "condemn" plaintiff's interests. See Boomer v. Atlantic Cement Co., 26 N.Y.2d 219, 228-30, 257 N.E.2d 870, 876-77, 309 N.Y.S.2d 312, 319-22 (1970) (Jasen, J., dissenting) (objecting that award of permanent damages in lieu of enjoining nuisance permits polluting cement company to "take" plaintiff's interest in clean air). In this sense, Boomer is a strict liability case, for the majority determined that the company was not required to act differently. Professor Cooter's analysis of prices and sanctions is consistent with this approach. Cooter defines a sanction as a detriment imposed for doing what is forbidden, and a price as money extracted for doing what is permitted. Officials should create HeinOnline -- 67 B.U. L. Rev. 232 1987 1987] ASSUMPTION OF RISK The short answer is that this type of preference for a tort recovery is irrelevant. If the full preference theory is persuasive, it justifies denying recovery to a plaintiff because her full preference among options (1), (2), 63 and (3) shows that she has suffered no injury that the law should protect. In summary, only after we have determined the scope of defendant's tort duty to plaintiff, can we decide whether plaintiff has any basis in tort law for complaining about defendant's failure to offer her an alternative that she fully preferred. To paraphrase Cardozo, a full preference "in the air, so to 64 speak, will not do." 2. Reasonableness is Largely Irrelevant The second observation, stated above but worth reiterating, is that the full preference model does not examine the reasonableness of the plaintiff's conduct. We do not ordinarily ask whether a plaintiff's contractual waiver of prices to compel decisionmakers to take into account the external costs of their acts, whereas officials should impose sanctions to deter people from doing what is wrong. Cooter, Prices and Sanctions, 84 COLUM. L. REV. 1523, 1523 (1984). Under this model, strict liability imposes a price, and negligence imposes a sanction. Id. at 1524, 1538-43. However, Cooter employs his model for the exclusive purpose of evaluating which legal approach more effectively changes behavior to approximate the socially efficient level. The approach I sketched above does not presuppose such a purpose, but is consistent with corrective justice, rights-based, and other noneconomic theories of tort law. 63 However, some strict liability doctrines may be an exception. See supra note 62, and infra note 236. Of course, a defendant might agree to remain liable notwithstanding plaintiff's full preference. Such an agreement to waive nonliability, like a plaintiff's agreement to waive defendant's liability, will often be enforceable. However, it raises issues distinct from assumption of risk. 6 Palsgraf v. Long Island R.R., 248 N.Y. 339, 341, 162 N.E. 99, 99 (1928) (quoting F. POLLOCK, THE LAW OF TORTS 455 (1 Ith ed. 1920). In other areas of law, I believe "consent" should receive a similar interpretation, tailored, of course, to the relevant policies. For example, the rape defense of "consent" to sexual intercourse shares the general features of consent that I have described. A women who "agrees" to intercourse when threatened by a gun has certainly not consented. She would have strongly preferred, and has a right to insist, that the defendant had not placed her in that predicament. On the other hand, if an employer threatens a woman with the loss of her job, or a high school boy threatens a girl with social scorn for being behind the times, her limited consent might be sufficient, even though the victim would prefer that defendant had not placed her in that predicament. The question is whether the rape statute gives the woman the right to insist that that degree of pressure not be imposed. If it does not, then her limited consent may be "full" enough. Of course. lesser criminal penalties, or civil damages, may be appropriate in cases where the type or degree of pressure is insufficient for a rape conviction. HeinOnline -- 67 B.U. L. Rev. 233 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 a tort claim against defendant was negligent, or whether plaintiff's consent to a battery was unreasonable. Why then ask whether her preference for risk was negligent or unreasonable? If we are willing to characterize plaintiff's choice of a risk as negligent, then we have concluded that the conduct is faulty, inefficient, culpable, or deficient in some other manner, according to the theory of contributory negligence that we hold. 65 At a minimum, we believe that plaintiff should have acted differently. But a judgment that plaintiff truly and fully prefers a risky alternative does not depend on such a conclusion or belief. Often, it is simply impossible to say whether or not risk-preferring conduct is negligent. Would the reasonably prudent person ever try the experimental sport of hang gliding? Although it may sometimes be possible to characterize riskpreferring conduct as negligent-e.g., playing Russian roulette-the reasons for denying or limiting recovery by the plaintiff are nevertheless different under the two theories. Assumption of risk emphasizes that plaintiff fully preferred to take the risk, and has no basis for complaint when it materializes. Contributory negligence emphasizes the unreasonableness of the 66 conduct. '5 See generally RESTATEMENT (SECOND) OF TORTS §§ 463, 464 (1965) (defining contributory negligence as conduct that falls below that of a reasonable man under like circumstances and that is a contributing cause of plaintiff's harm)- R. POSNER, supra note 53, at § 6.3: W. PROSSER & W. KEETON, supra note 1, at § 65 (discussing the elements and justifications of traditional contributory negligence); Schwartz, Contributory and Comparative Negligence: A Reappraisal, 87 YALE L.J. 697 (1978). 66 Professor Gary Schwartz, in his illuminating discussion of contributory negligence, notes the following partial explanation of why a person might voluntarily encounter a known risk: [O1n the basis of the values that the individual holds, he may conclude that his conduct is reasonable. If the law deems his behavior negligent, it does so because the community (viz., the jury) adheres to a different set of values. Consider the victim who walks into a dark, unknown room because he is unusually curious and places an unusually high value on the satisfaction of his curiosity, or the "risk-preferring" victim who is willing to jaywalk partly because he does not really mind (or even enjoys) dodging cars. Schwartz, supra note 65 at 715-16 (footnote omitted). Schwartz then questions whether contributory negligence is properly applied here: [ln the case of the victim with atypical values, it is hardly clear that his is conduct that society should really want to prevent. Certainly, the economist would not automatically argue in favor of prevention. Economics is keen on leaving to each individual the right to define and determine his goals; the economic assumption of rationality relates to choices that the person makes in pursuing his goals. Id. at 716 (footnotes omitted). I agree that society should not want to prevent all choices flowing from idiosyncratic values, though I do believe society may appropriately seek to prevent some choices (such as playing Russian roulette). When one cannot conclude that plaintiff should have acted otherwise, then by definition the atypical value choice cannot be HeinOnline -- 67 B.U. L. Rev. 234 1987 1987] ASSUMPTION OF RISK True assumption of risk cases should not be analyzed in reasonableness terms at all. If risk-preferring conduct is also negligent, its character as risk-preferring should ordinarily be decisive against recovery--even in a 67 comparative negligence regime, as we shall see. 3. The Preference Can Be Atemporal A third point about the full preference model is its atemporal character. Analysis of consent or assumption of risk should not refer only to existing actual choices available to the plaintiff. An employee who chooses to continue to work with defective machinery, rather than insist that the assembly line be shut down, has not assumed the risk under the model, because he would have preferred that the manufacturer or employer had supplied nondefective machinery in the first place. That choice is not now available to him. But since the critical question is plaintiff's actual full preferences, he should ordinarily be barred by consent or assumption of risk if he would not in any event have chosen the unavailable option. Conversely, he should not be barred on this. ground if he would have chosen that option. (To be sure, another ground exists: a plaintiff sometimes will be contributorily negligent for the choice that he does now make, among the available options.) The atemporal character of full preference creates obvious evidentiary problems. How do we know what plaintiff would have done if presented with the nontortious alternative? 68 Common sense will be a guide in significant condemned as negligent. Even when that conclusion is possible, however, the value choice is sometimes better analyzed as full preference. 67 See infra Part VA. 68 Professor Leubsdorf suggested this example to me: an employer claims that if he had bought the safe machine, its cost-and the greater desirability of safe jobswould have caused him to cut wages by 15%. Which would the employee prefer? Mansfield gives another example that vividly illustrates the problem. "A man may expose himself to the risks from a fireworks display in a public street in order to continue along the street to his office, and yet enjoy the fireworks as he goes." Mansfield, supra note 21, at 62. Mansfield worries that we might have to investigate the reason or motive for the man's conduct. Id. at 61-62. But I think the inquiry into preference is a little more manageable. To decide whether he prefers to enjoy the fireworks, we should ask whether he would go out of his way to see them, or whether he would have taken the same route even if they had been absent. Note, however, the possible causal overdetermination with the last option: perhaps either the brevity of the route or the presence of the display would have sufficed to cause him to take that route, and neither was a but-for cause. Causal overdetermination is a problem whenever mental states are legally dispositive, since persons act from a variety of motives or reasons. See, e.g., Eaton, Causation in Constitutional Torts, 67 IOWA L. REV. 443, 452-61 (1982); see generally Brodin, The Standard of Causation in the Mixed-Motive Title VII Action: A Social Policy Perspective, 82 COLUM. L. REV. 292 (1982). HeinOnline -- 67 B.U. L. Rev. 235 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 categories of cases, such as injuries to spectators and participants in sporting events, to customers of amusement parks, or to employees working with dangerous machinery. 69 For example, an employee who complains about an unsafe condition without avail clearly does not fully prefer the risk.70 Indeed, in general, employees will rarely fully prefer the risk of their employment, since they will only rarely obtain distinct personal benefits from the risk. By contrast, it will be difficult to credit the claim of a roller coaster patron that 71 he preferred a slower ride, since speed is part of the usual enjoyment. Nevertheless, the problem of proof should not be underestimated. Indeed, it plagues the informed consent and duty to warn cases, in which courts struggle to construct causal tests that avoid relying on a plaintiff's self72 serving recollection about his hypothetical past preference. A related problem is whether the hypothetical alternative should be determined prospectively or retrospectively. If we ask whether, in hindsight, the plaintiff fully preferred the road not taken when the road he took led to the realization of a risk, obviously plaintiff will usually, and honestly, recite that he now wishes that he had done otherwise. But is it obvious that we must therefore determine full preference prospectively, according to what 73 plaintiff would have chosen before he knew how matters would turn out? Why not allow a plaintiff to recover based on his full preference, in hindsight, not to choose the road that he chose? This question brings us back to the essential justification for the full preference theory. Consent should be a distinctive bar if, but only if, it demonstrates that plaintiff "prefers" the alternative he actually chose over the other alternatives that tort law requires defendant to offer. It is possible, consistent with this theory, to acknowledge plaintiff's regret and measure 69 "Common sense" jury determinations may, in practice, blur the clear conceptual line between contributory negligence and assumption of risk. Clarity can only be achieved through the usual devices for controlling jury discretion, such as instructions, directed verdicts, and appellate review. 70 See Lamson v. American Axe & Tool Co., 177 Mass. 144, 58 N.E. 585 (1900). See in fra note 151. "2 See supra note 60. 71 73 One possible answer is that in duty to disclose cases, a defendant's breach of duty only causes harm if the plaintiff's antecedent choice would have been different. But assumption of risk is an issue in cases other than disclosure cases. Nor is it a sufficient answer that defendant would have relied on plaintiff's antecedent expression of choice, because sometimes defendant would not in fact have relied. (Compare a doctor's subsequent reliance on a patient's consent to the doctor's disclosure with a manufacturer's inability so to rely on a consumer's consent to disclaimers, due to the manufacturer's loss of control over the product.) For discussions of the ambiguous content of consent because of different interpretations of the relevant time frame, see Kelman, Interpretive Construction in the Substantive Criminal Law, 33 STAN. L. REV. 591, 614-16 (1981) (consent to criminal acts); Kelman, supra note 26, at 778-87 (consent in economic theory). HeinOnline -- 67 B.U. L. Rev. 236 1987 1987] ASSUMPTION OF RISK his preference by his current desires. However, the purposes of tort liability, at least apart from strict liability, militate in favor of the foresight measure. Defendant is to be faulted for not doing otherwise, for breaching a duty, for falling below a standard. It is then appropriate to ask: if defendant had acted as he should have, would he have satisfied plaintiff's full preferences? And plaintiff's preferences, in this hypothetical world, are important because they disclose what plaintiff would have then chosen to do. Thus, hypothetical foresight is usually the proper test for measuring plaintiff's full preference. 74 The atemporal character of full preference does provoke the following objections. Perhaps the law should be primarily concerned with incentives for plaintiff's future behavior, and should ignore hypothetical alternatives that no longer exist. Or if deterrence is not the guiding policy, perhaps fairness demands holding plaintiff responsible for actual present choices. The first objection is at least overstated; we need not artificially confine the time frame for deterrence. Although it may be beneficial to confront plaintiffs with appropriate liability incentives to make the right present choices among limited options, those incentives can conflict with incentives for proper long-term care. The first set of incentives is not necessarily more important. 75 As to the fairness objection, perhaps plaintiffs do bear some responsibility for their limited present choices, and this responsibility is 74 But strict liability is another matter. See infra text accompanying note 237. In a different context, Professor Dworkin makes the arresting statement: -[A] counterfactual consent is not some pale form of consent. It is no consent at all." R. DWORKIN, A MATTER OF PRINCIPLE 278 (1985). Dworkin is responding to the common argument of law-and-economics scholars that legal rules express what the parties would have consented to under certain hypothetical conditions, such as absence of transaction costs. See, e.g., Gordon, Fair Use as Market Failure: A Structural and Economic Analysis of the Betamax Case and its Predecessors, 82 COLUM. L. REV. 1600, 1641-45 (1982); Posner, The Ethical and PoliticalBasis of the Efficiency Norm in Common Law Adjudication, 8 HOFSTRA L. REV. 487, 492 (1980); Posner, Epstein's Tort Theory: A Critique, 8 J.LEGAL STUD. 457, 460 (1979). In the context at hand, Dworkin's objection is not telling. A plaintiff has made a choice. We are not simply reconstructing a hypothetical choice out of none; we are trying to decide the significance of a limited choice. Still, I do believe it is appropriate to give dispositive weight to a counterfactual preference. Perhaps I do not differ dramatically from Dworkin on this matter, for his principal objection is not to any use of counterfactual consent, but to a failure to justify such use within a broader theory. R. DWORKIN, supra, at 275-80. I have tried to give such a justification, based not on an idealized market, but on what a plaintiff would actually do if confronted with the nontortious alternative. 5 See, e.g., Shavell, Torts in Which Victim and Injurer Act Sequentially, 26 J.L. & EcON. 589 (1983) (analyzing the last clear chance doctrine as a question of conflicting incentives over a broad time frame); see generally Wittman, Optimal Pricing of Sequential Inputs: Last Clear Chance, Mitigation of Damages, and Related Doc- trines in the Law, 10 J. LEGAL STUD. 65 (1981). HeinOnline -- 67 B.U. L. Rev. 237 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 reflected in contributory negligence doctrine. But it does not seem fair to preclude a plaintiff's claim because of her autonomous choice when the defendant has narrowed the temporal contours of choice. 4. Must the Preference be "Voluntary" and "Knowing"? A fourth observation is that the model may not fully address an actor's voluntary or knowing assumption of risk. Early assumption of risk cases often scanted this requirement, 7 . thus deservedly incurring the criticism of modern abolitionists. 77 The central tenet of the proposed model, however, addresses this issue. A plaintiff's claim should not be barred unless her chosen course of action was based on a full and true preference, i.e., made with knowledge of all the alternatives that defendant had a duty to offer, including that alternative which plaintiff claims defendant tortiously failed to offer. Moreover, the degree ,3f knowledge required should at least extend to 78 matters that would affect plaintiff's full preference. Voluntariness is a more problematic requirement. What types of constraints on a plaintiff's choice undermine its validity as a true expression of preference? Here, the model provides only a partial answer, for it is only concerned with those constraints that the defendant tortiously imposes. 79 Consider the well-known consent case, O'Brien v. Cunard Steamship Co. See Meistrich v. Casino Arena Attractions, Inc., 31 N.J. 44, 49-55, 155 A.2d 90, 93-95 (1959); see also Rabin, The Historical Development of the Fault Principle:A 76 Reinterpretation, 15 GA. L. REv. 925, 941 (1981). 77 Another problem is fixing the time at which plaintiff's knowledge should be measured. What if plaintiff understands the risk at one time, but forgets about the risk later at the actual point of choice? Twerski argues that such plaintiffs should not be barred from recovery, at least if their forgetfulness is not negligent. Twerski, supra note 37, at 19-22. I tend to agree, but for a different reason: the forgetful plaintiff's actual choice is a much less reliable guide to her full preference if she does not have the full range of alternatives in mind at the moment of choice. 78 The knowledge requirement may be broader if the point is to protect plaintiff's right to decide in an informed manner, even if plaintiff would not have decided differently with the information. See Shultz, supra note 58, at 225-26 n.27; Twerski, supra note 37, at 47 ("The issue in battery is not whether but for the failure to warn the plaintiff would not have taken the drug; but rather, whether plaintiff was subjected to an unconsented touching.") (footnote omitted). Perhaps such a broader right to decide should also be recognized in negligence cases. A distinct issue is whether defendant has an affirmative duty to ensure that plaintiff has given his informed consent. Such a duty, imposed in connection with medical treatment and (in a sense) product sales, typically requires the defendant to disclose risks. But consent and assumption of risk can be defenses apart from such a duty: they indicate that an otherwise established breach of duty is not a legally cognizable harm from plaintiff's particular perspective. 79 154 Mass. 272. 28 N.E. 266 (1891). HeinOnline -- 67 B.U. L. Rev. 238 1987 ASSUMPTION OF RISK 1987] The plaintiff, a passenger on a steamship, may have held out her arm to be vaccinated only out of terror and fear of quarantine. The court held that her apparent consent barred her battery claim.80 However, let us suppose defendant was aware of her unexpressed fears. Would that undermine the voluntariness of plaintiff's consent? Not under the proposed model, because it was the government's vaccine regulations, not the defendant, that restricted plaintiff's choice. Traditional assumption of risk doctrine apparently takes the same position."' But voluntariness in this modest sense may be an unacceptably weak requirement. The considerations that prompt courts or legislatures to relieve consumers from some types of contractual "choices" (including waivers of liability) certainly apply here. 2 Such relief may often be wise, whether it is viewed as paternalistic, as recognizing social and political influences that prevent individuals from expressing their true choices, or as expressing a 83 judgment that the very concept of "true choice" has uncertain meaning. For example, if fear of quarantine motivated the passenger in O'Brien to accept the government-required vaccination, and if she otherwise would not have consented, then it may be sensible to permit her to recover in tort. In this article, I make no attempt to describe or defend such policies. I note only that the voluntariness question is significant, difficult, and partially independent of the model that I present. 5. Unavoidably Dangerous Activity The final point of clarification concerns unavoidably dangerous activity. Ordinarily, an activity can be made safer at some cost. Numerous alternatives will often exist, each offering a distinct mix of reduction in certain risks 80 Id. at 275, 28 N.E. at 266. See RESTATEMENT (SECOND) OF TORTS § 496E comment b (1965); see also Mansfield, supra note 21, at 48 n.48 (analyzing O'Brien in this manner). But see Twerski, supra note 37, at 40 (suggesting that it is appropriate to deny the defense in "extreme" duress situations even if defendant is not responsible for the situation). 82 Cf. RESTATEMENT (SECOND) OF TORTS§ 496B commentj (express agreement to assume risk will generally not be enforced if there is "such disparity of bargaining power between the parties that the agreement does not represent a free choice on the part of the plaintiff"); id. at § 496C commentj (1965) (implied assumption of risk may be invalid defense against certain classes of plaintiffs who have been unreasonably taken advantage of). Assumption of risk is no longer a defense in worker's compensation systems. See W. PROSSER & W. KEETON, supra note I, at 573. 83 On the indeterminacy of the concept of choice, see generally Kelman, supra note 26. On paternalism, see Mansfield, supra note 21, at 42. On the question when, and why, government should decide that private preferences are an improper basis for social choice, see Sunstein, Legal Interference with Private Preferences, 53 U. 8I CHI. L. REV. 1129 (1986). HeinOnline -- 67 B.U. L. Rev. 239 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 and increase in certain costs. The model then asks whether the plaintiff would prefer the safer alternative, given the additional cost. But sometimes an activity is unavoidably dangerous: it cannot be made safer. In that event, the defendant cannot offer choice (3), i.e., she cannot offer the activity's benefit without its costs. If the plaintiff has chosen to engage in such an activity, with knowledge of the risks, then he has consented or assumed the risk, in the full and not the limited sense. Perhaps he wishes that there were a safer way to engage in the activity, but there is not. His consent, therefore, should ordinarily be effective. Defendant has not deprived him of any feasible options. Indeed, in offering choice (2) (the benefits with the attendant risks), the defendant may not even have owed or breached a duty. If this is the case, we need not reach the consent issue. Activities such as lending a defective chattel that one has no power to fix, or distributing an unavoidably dangerous but highly beneficial drug, may fall within this category. However, some activities, such as skydiving, are more difficult to classify as negligent or not negligent. Thus, when a plaintiff is injured in such an activity, it is sometimes clearer that he has assumed the risk than that defendant has tortiously created it. Nevertheless, sometimes the unavoidable dangers are so extreme that it is negligent for the defendant even to offer choice (2). In these situations, plaintiff's knowing acceptance of or engagement in the patently dangerous activity would ordinarily be a full preference. For example, defendant might market an unavoidably dangerous drug with very slight benefits. A plaintiff using this drug should ordinarily be barred by assumption of risk. The plaintiff actually prefers the negligent alternative that defendant has marketed to defendant's alternative of not marketing at all. 84 In addition, here, the plaintiff will be contributorily negligent as well.8 5 In jurisdictions that treat the two doctrines differently, it matters whether plaintiff has assumed 86 the risk as well as acted negligently. Note, finally, that many strict liability cases involve unavoidably danger- 84 Plaintiff might insist that he was deprived of a real alternative choice by defendant's marketing of the product: perhaps plaintiff could have resisted the temptation to use the product if defendant had not marketed it! Still, plaintiff's preference was to use the product. The temptation argument is better considered as part of defendant's duty to protect plaintiff against his own voluntary choices. See supra text accompanying notes 82-83. 85 But not always. Plaintiff's idiosyncratic interests might justify his encountering the risk even though defendant remains negligent. Suppose defendant manufactures an unavoidably dangerous drug creating enormous risks and slight benefits to almost all potential users. The benefit to an individual plaintiff, however, may be much greater and outweigh the risks. 86 See infra text accompanying notes 201-02. HeinOnline -- 67 B.U. L. Rev. 240 1987 ASSUMPTION OF RISK 1987] ous activities.8 7 Thus, a plaintiff who prefers such a risk also prefers it fully. 88 C. Is the Model Necessary? Is the full preference model necessary? Cannot the modem framework accommodate the concept of full preference? Cannot cases of full preference be analyzed as cases of either no duty, no breach of duty, or contributory negligence? I have already reviewed and rejected the last possibility. 89 But what about no duty, or no breach? If plaintiff actually prefers the option that defendant has placed before her, can we not conclude that defendant owes that plaintiff no duty to protect against the harm? Robert Keeton has criticized a similar argument treating reasonable assumption of risk as "no duty." 90 In his view, the argument can be misleading and can create bizarre effects. For example, when defendant creates an unreasonable risk to two persons, and only one appreciates and voluntarily encounters the risk, the result is that no duty is owed to him, while a duty is owed to the other.9 1 I, too, have my doubts that "no duty" is the appropriate category.9 2 More 87 Consider two strict liability categories: abnormally dangerous activities under the RESTATEMENT (SECOND) OF TORTS §§ 519-520 (1965), and product defects involving manufacturing flaws and unconscious design defects. See J. HENDERSON & R. PEARSON, THE TORTS PROCESS 698-700 (2d ed. 1981). In many of these cases, the risks are unavoidable, or nearly so. 88 See infra Part VC. 89 See supra text accompanying notes 65-67. 90 R. Keeton, Assumption of Products Risks, 19 Sw. L.J. 61 (1965). I' at 68. As Keeton explains: Id. The only difference is not in the quality of the defendant's conduct but rather in the quality of the plaintiffs' participation in the event. Talking about this one difference as a duty or no-duty proposition is confusing. Duty and no duty connote something about the defendant's conduct or the relation between the plaintiff and the defendant rather than something about the plaintiff's participation. Assumption of risk, on the other hand, connotes something about the plaintiff's participation, and since this is the point of distinction, use of this term seems the preferable way of saying it. Id. For James's response, see James, supra note 8, at 194. 12 A related concern is whether plaintiff or defendant should have the burden of persuasion. A "no duty" characterization allocates the burden to plaintiff, while a "defense" characterization allocates it to defendant. Moreover, the first characterization would likely give the judge a larger role than the second. See RESTATEMENT (SECOND) OF TORTS § 496C comment e (1965); W. PROSSER & W. KEETON, supra note I, at 493-95; Twerski, supra note 37, at 9 & n.35, 10 & n.43. Although these questions are significant, I believe they can be answered either way without undermining the independence of the concept of full preference. (Note, for example, that jurisdictions disagree whether defendant has the burden of proving consent or plaintiff has the burden of proving nonconsent to an intentional tort. See the cases cited in W. PROSSER & W. KEETON, supra note 1, at 112 nn.2-3). HeinOnline -- 67 B.U. L. Rev. 241 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 to the point, however, I believe it unwise to submerge a distinct and significant doctrine within the vast sea of "no duty" rules. In Keeton's words, "we might as readily say that the defendant has no duty to one who is contributorily negligent." ' 93 The concept of duty is unruly enough. We would do well to remember its separate doctrines-such as no duty to rescue, or limited duty not to cause emotional harm negligently-and their separate rationales. Even if assumption of risk is sensibly placed in the no duty category, that category is such a mishmash that the placement does not further analysis.9 I do agree, however, that for some purposes, such as comparative negligence, the "no duty" characterization of full preference accurately describes the appropriate legal import of an assumption of risk finding. 95 Of course, when the duty of a defendant consists only in supplying information to the plaintiff-as by warning the user of a product of significant dangers-then a plaintiff who obtains that information cannot recover from the defendant. 96 Under the full preference model, since the defendant's duty is only to supply information, any decision that plaintiff subsequently makes with that information is an expression of preference for which the defendant cannot be held responsible. Put another way, the defendant's limited duty represents a social judgment that he need not act differently, but need only 97 facilitate the plaintiff's informed individual choice. R. Keeton, supra note 23, at 160 (footnote omitted); see id. at 161-64. California's "formula" for duty, see Rowland v. Christian, 69 Cal. 2d. 108, 112-13, 443 P.2d 561, 564, 70 Cal. Rptr. 97, 100 (1968), is a hodgepodge of factors. Trying to factor in plaintiff's choice to take the risk of defendant's conduct would only add to the confusion. But see Gaetanos, supra note 1, at 478 ("[A]ssumption of risk should be subsumed into the reasonable man formula."); see also Twerski, supra note 37, at 48 (emphasis in original): To be sure one can broadly state that defendant owes no duty to a plaintiff who perceives a risk and encounters it, but the realities of the litigated cases are that an individual plaintiff has been barred because of his peculiar relationship to a risk. The duty of the defendant to act or not to act and the scope of his responsibility to others remains essentially untouched. 93 94 95 See infra Part VA. 96 This is true even if the defendant failed to supply the information, and plaintiff obtained it independently. See RESTATEMENT (SECOND) OF TORTS § 496C comment e (1965). Although one might say that defendant owed no duty to one who would otherwise acquire the information, or that plaintiff assumed the risk in the looser Restatement sense, see id., I would employ a simpler analysis: defendant's negligence did not cause plaintiff any harm. See R. Keeton, supra note 23, at 145-46. But see Twerski, supra note 37, at 5 n.23. 9 See RESTATEMENT (SECOND) OF TORTS § 496G comment b (1965) (explaining that rules establishing no duty beyond disclosure reflect "elements similar to those of assumption of risk"). Information about an activity or a product can be valuable to a plaintiff in two basic ways-first, by describing risks that plaintiff can avoid while engaging in the activity or using the product, and second, by warning of risks that cannot be avoided unless HeinOnline -- 67 B.U. L. Rev. 242 1987 1987] ASSUMPTION OF RISK Professor James has taken this duty analysis to its next logical step. Consider a legal duty requiring a defendant to do more than warn, and to take precautions on plaintiff's behalf. If assumption of risk is a valid defense in such a case, James insisted, the legal duty is self-defeating. 98 1 agree that assumption of risk should not ordinarily be a defense in this situation. But under the full preference model, it usually would not be. A manufacturer of a product with obviously dangerous exposed moving parts may have a duty to install a safety guard, and a plaintiff who uses the product aware of its danger may well be entitled to recover. But such a plaintiff undoubtedly consents to the risk of injury only in a limited sense; he would rather use the product than not use it. Only rarely will he actually prefer that the product not have a safety guard, and thus fall within my narrower conception of assumption of risk. There is, however, an aspect of duty analysis that overlaps substantially with the full preference theory. Insofar as a defendant's duty depends specifically upon the full preferences of the class of potential plaintiffs, "no duty" and "full consent to risk" involve the same analysis. For example, modern product liability doctrines often consider the preferences of consumers, among other factors, in determining whether a product's design is defective. 99 When a defendant's duty is limited to warning of dangers, the justification for the limitation is sometimes that most potential plaintiffs would fully prefer that the defendant not also take an additional precaution. Nevertheless, full preferences are not always considered in determining the scope of duty, and are usually not the only factor relevant to that determination. Moreover, the full preference theory is a more individualistic doctrine; even if a defendant owes a duty in part because most potential plaintiffs would not prefer the risk, a particular plaintiff who does fully prefer the risk would be barred from recovery. plaintiff discontinues the activity or use. See M. AND ALTERNATIVES FRANKLIN & R. RABIN, TORT LAW 608-14 (3d ed. 1983). The full preference theory could apply in either category. In the first, once defendant discharges his duty, ordinarily plaintiff will suffer harm only if he has been contributorily negligent. However, sometimes such a plaintiff may fully assume the risk, and act negligently as well. If defendant fails to warn that its lighter fluid should not be squirted onto an open flame, but plaintiff concludes that the product's ability to work under these circumstances outweighs the risks, then plaintiff prefers the risky alternative. If the state's comparative fault law treats contributory negligence and true assumption of risk differently, as it sensibly might, see infra Part VA, then the additional assumption of risk finding can be significant. 98 James, supra note 8, at 192. 99 See, e.g. Barker v. Lull Eng'g Co., 20 Cal. 3d 413, 432, 573 P.2d 443, 455, 143 Cal. Rptr. 225, 237 (1978) (a product may be found defectively designed if it fails to perform as safely as an ordinary consumer would expect); Cepeda v. Cumberland Eng'g Co., 76 N.J. 152, 174, 386 A.2d 816, 826 (1978) (examining, inter alia, "the usefulness and desirability of the product," citing Wade, On the Nature of Strict Tort Liability.for Products, 44 Miss. L.J. 825, 837 (1973)). HeinOnline -- 67 B.U. L. Rev. 243 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 Voluntary Assumption of a Relationship Distinguished D. Is there a plausible argument for an assumption of risk doctrine broader than full preference? Should a non-negligent, limited preference, as in Keeton's last example, be a bar? At the very least, the justification for this will be much weaker than that for full preference as a bar. 1° ° One possible justification for the broader form of assumption of risk links the doctrine to the general absence of a duty to act in tort law. When the defendant offers a product, condition, or alternative to the plaintiff that defendant has no initial duty to offer, and which plaintiff is free to take or leave, why can't defendant offer it on limited terms, terms which might amount to negligence if defendant did owe a duty? James articulates this justification with numerous examples.10 1 If I lend you a car that I tell you has bad brakes and bald tires, I am not liable if the defects cause injury. Indeed, this rationale may have been a major justification of assumption of risk in its principal early application, to employment: an employee has no right to a job, is free to reject it, and therefore cannot complain if the employer offers it on limited terms or conditions, conditions that would 10 2 amount to negligence if defendant did owe a duty to provide a job. I agree that this type of justification sometimes has merit. Its merit should not be overstated, however. The claim that "no duty" implies a privilege to extend only a limited duty is one of many instances in the law of the more general argument that "the greater includes the lesser"--the greater power to deny a legal benefit includes the lesser power to grant it upon any condition. The argument is not universally valid.10 3 An employer may have For example, Robert Keeton suggests that the plaintiff's "coauthorship" of the harm along with the defendant justifies the relatively broad consensual form of assumption of risk. R. Keeton, supra note 23, at 151-52, 163-64. But Keeton gives little explanation of why causal contribution should be so critical. Causation alone cannot be the justification. See infra note 165. By the term "coauthorship," Keeton must be relying on some unexplained concept of agency or choice. 101 James, supra note 1, at 142-43. 102 This rationale seems implicit, for example, in Lamson v. American Axe & Tool Co., 177 Mass. 144, 144-45, 58 N.E. 585, 585 (1900). For judicial and scholarly reviews of the development of assumption of the risk doctrine, see Tiller v. Atlantic Coast Line R.R. Co., 318 U.S. 54, 58-67 (1943); G. WHITE, TORT LAW IN AMERICA: AN INTELLECTUAL HISTORY 41-45 (1980); Bohlen, Voluntary Assumption of Risk, 20 HARV. L. REV. 14, 20-22 (1906); Rabin, supra note 76, at 939-44 ; Schwartz, Tort Law and the Economy in Nineteenth-Century America: A Reinterpretation,90 YALE L.J. 1717, 1768-71 (1981); Warren, supra note 46. 103 See Kreimer, Allocational Sanctions: The Problem of Negative Rights in a Positive State, 132 U. PA. L. REV. 1293, 1310-14 (1984); Simons, Equality as a ComparativeRight, 65 B.U.L. REV. 387, 447-48 (1985); Westen, The Rueful Rhetoric of "Rights," 33 UCLA L. REV. 977, 993-95, 1010-18 (1986); Westen, Incredible Dilemmas: Conditioning One Constitutional Right on the Forfeiture of Another, 66 IOWA L. REV. 741, 747 n.20 (1981). 100 HeinOnline -- 67 B.U. L. Rev. 244 1987 1987] ASSUMPTION OF RISK no duty to hire me, but it does not follow that he can hire me and then torture 10 4 me for misbehavior-even if he has provided full disclosure of his policy. Some risks, however, he may fairly require me to accept. As a member of a construction crew, I cannot demand that the builder of a skyscraper reduce risks of physical harm to the level of a desk job, even if this would be feasible at some cost. And the individual who lends me a car and informs me of its bad brakes is fairly held to only a limited duty. On these occasions, when a defendant who has no duty to the plaintiff in the first instance is permitted to limit that duty, the limitation should not be analyzed as assumption of risk. A better category would be voluntary assumption ot"a relationship. This separate category is necessary because such cases do not necessarily involve full preference. When someone lends me a car and warns me about its bald tires, I do not fully prefer what he's offered me; most likely, I would prefer that he had obtained new tires. Nevertheless, because I voluntarily assumed this relationship, he is not liable. 105 But when is defendant's duty properly limited due to plaintiff's voluntary assumption of a relationship? How much can defendant limit his duty? Doesn't the plaintiff's full preference .have some bearing here? These difficult questions are relevant to many areas of tort liability-including the liability of land owners to licensees and invitees;10 6 the duties of landlords, therapists, and others whom courts find in a "special relationship" either to third parties or to potential victims; 1 0 7 and the remaining vestiges of the privity doctrine. 10 104 Conversely, does the "greater" duty of a defendant to offer a benefit imply the "lesser" duty not to condition the benefit on plaintiff's willingness to be exposed to a tortious risk? Mansfield gives the example of a common carrier, obliged to serve the public, attempting to limit its liability agreement. Mansfield, supra note 21, at 52. Again, as Mansfield suggests, the greater does not logically entail the lesser, though strong arguments may support the lesser duty. Id. 105 For discussions of voluntary assumption of a relationship or association, see Hoar v. Sherburne Corp., 327 F. Supp. 570, 575-78 (D. Vt. 1971) (Oakes, J.) (containing a thoughtful critique of the concept of "voluntarily" assuming a relationship); RESTATEMENT (SECOND) OF TORTS § 496C comment d (1965) (second type of assumption of risk); RESTATEMENT (FIRST) OF TORTS § 466 comment d (1934); id. at § 893 comment b (second type of assumption of risk); James, supra note 1, at 142-46 ("primary" assumption of risk); R. Keeton, supra note 23, at 128 ("associational" assumption of risk). 106 See J. HENDERSON & R. PEARSON,supra note 87, at 459-61; W. PROSSER& W. KEETON, supra note 1, ch. 10. 107 See RESTATEMENT (SECOND) OF TORTS §§ 314-320 (1965); R. EPSTEIN, C. GREGORY, & H. KALVEN, CASES & MATERIALS ON TORTS 399-413 (4th ed. 1984) [hereinafter MATERIALS ON TORTS ]; W. PROSSER & W. KEETON, supra note I, at 383-85. 108 See, e.g., Ultramares Corp. v. Touche, 255 N.Y. 170, 174 N.E. 441 (1931) (limiting recovery for economic harm from misrepresentation); Moch Co. v. HeinOnline -- 67 B.U. L. Rev. 245 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 Although these questions are complex, voluntary assumption of a relationship and voluntary assumption of a fully preferred risk deserve separate analysis. Voluntarily assuming a relationship at least implies that plaintiff prefers the relationship, with its attendant risks, to not entering it at all, but does not imply that the plaintiff made a full preference. Factors other than voluntariness and knowledge are relevant to whether defendant's duty in the relationship is limited. Such factors, including foreseeability, the defendant's ability to reduce the harm, and the other considerations in the typical tort calculus, are not independently relevant to full preference.10 9 Moreover, the full preference theory can apply whether or not plaintiff has voluntarily assumed a limited relationship. A business visitor to whom a duty of care is owed fully assumes the risk of danger from a gaping hole in defendant's building if the visitor proves his athletic prowess by trying to leap the hole in a single bound. In brief, full preference emphasizes plaintiff's full consent, while assumption of a limited relationship considers defendant's ordinary lack of duty along with general tort principles. Assumption of risk doctrine is most frequently abused when plaintiff has voluntarily assumed a relationship. It is tempting to apply the doctrine here, since voluntariness and choice are at least a part of the relevant analysis. But to indulge that temptation is a dangerous mistake. If plaintiff does not fully prefer the risk, then the simple fact that she assumed a relationship does not show that she should be barred from recovery. Courts should tend to a more complete analysis of the fullness of plaintiff's preferences, and not be deceived by the disarming simplicity of holding the plaintiff to her "choice." Rensselaer Water Co., 247 N.Y. 160, 159 N.E. 896 (1928) (property owner has no cause of action against water works company for failure to maintain water pressure at fire hydrants based on breach of contract between water company and municipality); MATERIALS ON TORTS, supra note 107, at 429-35, W. PROSSER & W. KEETON, supra note 1, at 740-49. 109 In Rowland v. Christian, 69 Cal. 2d 108, 443 P.2d 561, 70 Cal. Rptr. 97 (1968), for example, the California Supreme Court announced that the following considerations should be balanced in deciding whether a person's duty of care should be limited: [Tihe foreseeability of harm to the plaintiff, the degree of certainty that the plaintiff suffered injury, the closeness of the connection between the defendant's conduct and the injury suffered, the moral blame attached to the defendant's conduct, the policy of preventing future harm, the extent of the burden to the defendant and consequences to the community of imposing a duty to exercise care with resulting liability for breach, and the availability, cost, and prevalence of insurance for the risk involved. Id. at 113, 443 P.2d at 569, 70 Cal. Rptr. at 100; see also R. Keeton, supra note 23, at 129-30 (explaining how voluntary assumption of a relationship relies on different policies, including a different conception of voluntariness, than voluntary assumption of risk). HeinOnline -- 67 B.U. L. Rev. 246 1987 1987] E. ASSUMPTION OF RISK Limits of the Model Can the model be criticized from the opposite perspective-as too strong, rather than too weak? Should a full preference always preclude the plaintiff's recovery? I do not think it should. Defendant sometimes has a duty to protect plaintiff from the consequences of his own voluntary choice, even when the choice amounts to a full preference. 110 Perhaps the marketer of a highly dangerous drug with slight benefits should be forbidden from marketing it; some patients may fully consent to the risks, yet society may be prepared to require protection from the manufacturer's harmful and irresponsible choice. Moreover, as I have suggested,"' the model only gives a partial answer to what is a sufficiently voluntary or knowing consent to a risk. And there is a deeper problem. Although the categories of limited preference and voluntary assumption of a relationship are conceptually distinct, they shade into one another in practice, as follows. Recall that sometimes a plaintiff is properly barred on the ground that she has voluntarily entered into a relationship with the defendant which the defendant is free to decline, even though the plaintiff may fully prefer a slightly different relationship. For example, in an emergency you hail a passing car and agree to a ride despite the bad brakes; you would fully prefer, however, that the Good Samaritan had exercised more care in maintenance. The full preference model cannot 2 explain nonrecovery here."1 Unfortunately, many situations involving choice do not reveal full preference. Much therefore depends on the scope of the category of voluntary assumption of a relationship. And much depends on whether a relatively full preference can justifiably bar recovery. Consider the facts in Meistrich again. Suppose the owner advises the patrons that the ice is quite unsafe, but they are free to skate if they like. In twenty minutes, moreover, he will have properly prepared a sizeable portion of the rink. You decide to skate, not because you like to skate on unusually slippery ice, but because it is inconvenient to wait twenty minutes. Isyour preference full? No; you would have been happiest if he had properly prepared the ice in the first place. Yet you are unusually willing to tolerate his negligence; a slight inconvenience prompts you to choose the risky course. In such a case, the line between full preference and voluntary assumption of a relationship blurs. You are almost as willing to accept defendant's tortious conduct as a fully risk-preferring person would be. Your decision should bar recovery for your injuries, I believe. But we can increase the inconvenience to the point where this conclusion is much less certain. 110 For an illuminating discussion of such a duty, with reference to the traditional broad conception of voluntary choice, see Twerski, supra note 37, at 25-34, 43-48, 50-51. "ISee supra text accompanying notes 78-83. 112 See supra Part liD. HeinOnline -- 67 B.U. L. Rev. 247 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 The cases of full preference may be few. Nevertheless, when the preference is less than full, I believe that the arguments underlying the full preference model simply apply with less force, while the arguments underlying voluntary assumption of a relationship (and other "no duty" rules) correspondingly become more prominent. III. COMPARING ASSUMPTION OF RISK WITH CONSENT TO AN INTENTIONAL TORT Assumption of risk is often casually described as the doctrine of consent applied to nonintentional torts, and especially to negligence. But few have examined this relationship between consent and assumption of risk carefully. One of the most fruitful analyses is Professor Mansfield's: Consent is the right term to use when the plaintiff was willing that a certain event occur, probably some conduct on the part of the defendant, because he desired an invasion of a normally protected interest. Ordinarily he will believe that the invasion is substantially certain to follow the event, and the term consent focuses on his belief in the certainty of the invasion. Assumption of risk is the right term to use when the plaintiff was willing that a certain event occur, but he neither desired an invasion of a normally protected interest nor did he suppose that such an invasion was substantially certain to result. The focus is on the uncertainty of the result from the plaintiff's point of view ... As so defined, the terms consent and assumption of risk do not parallel the distinction between intentional torts and negligence. The defendant may intend an invasion and the plaintiff even realize that he intends it, and yet it is a case of assumption of risk if the plaintiff does not desire the invasion or think it substantially certain to occur. Or the defendant may be merely negligent and not intend an invasion, but if the plaintiff thinks an invasion substantially certain to occur, it is a case of consent. 113 I largely agree. Assumption of risk should refer to acceptance of a risk of harm (or of some other invasion of a legal interest) that plaintiff does not believe is substantially certain to materialize. Consent should imply relative certainty that the risk will materialize.11 4 However, conventional doctrine appears to draw a different line, describing consent as the appropriate 113 Mansfield, supra note 21, at 31-32 (footnote omitted). 114 However, I do not agree with Mansfield that plaintiff's desire for an invasion of a legally protected interest means that she has consented and not assumed the risk. For example, if plaintiff desires to risk harm, she might only assume that risk. I would distinguish between assumption of risk and consent according to the certainty of plaintiff's belief that the risk will materialize, and not according to desire. After all, the full preference model does not even require a positive desire for the risk or the invasion. It requires only a preference, and only a preference for an alternative Ipackage" that includes that risk along with distinctive benefits. See supra note 54. HeinOnline -- 67 B.U. L. Rev. 248 1987 1987] ASSUMPTION OF RISK concept for intentional torts and assumption of risk as the appropriate 115 concept for torts of negligence, recklessness, and strict liability. Let me provide some examples of Mansfield's position. A person who consents to an operation agrees to a physical touching that would otherwise be a battery. He agrees to a fairly certain (or highly probable) intrusion, and the defendant intends the intrusion. A passenger who assumes the risk of the defective brakes of the car in which he is riding agrees to accept an uncertain risk (they may not fail), and a risk that was at best negligently created. These are the classic types of consent and assumption of risk. However, if an athlete is deemed to assume the risk of some intentional contacts during a 11 6 sporting event, he has agreed to an uncertain risk of an intentional tort. Thus, there are at least two senses in which plaintiff can be said to have agreed to a "risk of harm": she may have agreed to a risk of a tortious invasion, intentional or otherwise; or she may have agreed to a negligently created risk, certain or otherwise. Apart from this small and fairly unimportant distinction in the strength of plaintiff's belief, Mansfield views consent and assumption of risk as the same.1 7 That appears to be the received view,118 though as noted above, it is 115 See RESTATEMENT (SECOND) OF TORTS ch. 17A, Scope Note; id. at § 892 comment a (1965). 116 Less frequently, a plaintiff might consent to a relatively certain risk that defendant negligently and unintentionally created. For example, a rescuer might risk almost-certain injury from defendant's prior negligence, though from defendant's ex ante perspective, the risk may have been much less certain. 117 Mansfield, supra note 21, at 33. 118 See, e.g., G. WILLIAMS, supra note 42, at 295-96; R. Keeton, supra note 23, at 152-53, 166 ("the doctrine of assumption of risk continues to commend itself if ...it is limited to cases in which either objectively consensual assumption of risk or an even more consensual form can be proved"); see also Kionka, supra note I, at 373-76 (arguing that only express assumption of risk, in the sense of a release or waiver of liability, truly exemplifies consent.) Professors Schwartz and Kionka, however, attempt to distinguish consent to an intentional tort from assumption of risk. Consent to an intentional tort, they argue, reflects plaintiff's agreement to hold defendant harmless, to release the defendant from liability. But voluntary assumption of a known risk only rarely reflects such an agreement or release. V. SCHWARTZ, supra note 1, § 9.5, at 173-74; Kionka, supra note I, at 384-85. The argument is unpersuasive. Consent to an intentional tort does not necessarily reflect a waiver of liability. In any event, waiver is an unduly narrow conception of assumption of risk. See supra Part IIA(2). Dean W. Page Keeton's attempt to characterize consent to an intentional tort as radically different from assumption of risk is also unpersuasive. See W. Keeton, supra note 43, at 117-18. Keeton simply stipulates that consenting to a known risk properly bars recovery, unlike consenting to a less certain risk. He also proposes a symmetry argument: if defendant is liable only for creating an unreasonable risk, plaintiff should only be barred for taking an unreasonable chance. This argument HeinOnline -- 67 B.U. L. Rev. 249 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 the nature of the tort that defines the distinction, under conventional doctrine. Thus, the Restatement (Second) comments: "Consent to conduct that is merely negligent, creating an unreasonable risk of harm, is commonly called 'assumption of risk.' "119 The relevant "fullness" of the preference is an important question for both consent and assumption of risk, as I have suggested. 120 And for both doctrines, the answer to that question is similar: to decide whether the consent is sufficiently full, we must determine whether the defendant has a duty to provide the option that plaintiff would fully prefer. If not, then plaintiff's more limited preference is full enough to bar recovery. Consider another example. Plaintiff and defendant agree to resolve a verbal argument by a physical fight. Suppose plaintiff agrees to fight, but says, "I'm not agreeing to be injured." Then, if defendant decides to fight, he may be liable for plaintiff's injuries (and may well lower his level of physical aggression). More likely, however, defendant would reply: "Forget it. Do you want to fight or not?" If plaintiff then decided to fight, defendant could not reasonably believe that plaintiff fully consented. Clearly plaintiff has expressed only a limited preference for fighting (and incurring risk of injury) over not fighting. But just as clearly, plaintiff should not recover for injuries suffered in the fight. Plaintiff has no right to insist that defendant offer one type of combat rather than another, so long as plaintiff is free to refrain. 121 The result may seem in tension with the usual approach to intentional tort law: a defendant has a strict duty not to intentionally invade plaintiff's legal does not discredit the full preference theory, under which plaintiff is only barred if he was in fact willing that defendant act tortiously. (Symmetrically, defendant cannot rely on plaintiff's contributory fault if indeed defendant fully preferred that plaintiff act that way! See supra note 3 1). Consider again Meistrich v. Casino Arena Attractions Inc., 31 N.J. 44, 48, 155 A.2d 90, 93 (1959), distinguishing, without explanation, consent by participation in contact sport from assumption of risk. One case took Meistrich's approach to its logical extreme, recognizing express assumption of risk where plaintiff voluntarily consents to take certain chances of negligently-inflicted injury in a sporting event. Kuehner v. Green, 436 So. 2d 78, 80 (Fla. 1983). Though logical, this analysis threatens to resurrect the doctrine that Meistrich promised to inter. 119 RESTATEMENT (SECOND) OF TORTS § 892 comment a (1965). 110 See supra Part IIA(l). Indeed, in this and some other cases, the plaintiff may have actually entered an informal contract with the defendant, agreeing to a certain degree of force or level of 2I care. In contract law, it is irrelevant that you only had a limited preference for what the other party was willing to offer you. You have voluntarily assumed a contractual relationship, usually for consideration: that is enough. But even if the course of dealings do not amount to a contract, the tort duty is insufficient for recovery here. In some jurisdictions, of course, consent to an illegal act is ineffective to bar recovery. HeinOnline -- 67 B.U. L. Rev. 250 1987 1987] ASSUMPTION OF RISK interests, absent consent or some other privilege. 122 We do not ordinarily think of the duty as qualified in scope. But its scope is qualified, as I have just illustrated. 2 3 Other qualifications unrelated to consent also exist, though they are tangential to this paper. 124 Consent and assumption of risk doctrine are directly analogous in significant ways. Both traditionally operate as complete defenses. Both essentially rely on the plaintiff's subjective preferences or desires.125 But there are some subtle contrasts, as well. Let us examine whether the body of doctrine surrounding one concept plausibly applies to the other. I will begin by applying several features of consent doctrine to assumption of risk. First, does the doctrine of apparent consent apply? Is plaintiff barred 122 See generally RESTATEMENT (SECOND) OF TORTS chs. 2-11 (1965) (covering the range of intentional torts and defenses thereto). 123 Reconsider the earlier discussion of informed consent and the incompetent doctor, supra text preceding note 57. With respect to battery, (a) a defendant must give plaintiff the informed choice not to suffer any intentional invasion of his bodily integrity. But (b) plaintiff has no right to.insist that defendant propose an intentional invasion that creates a lesser risk of harm than a reasonable defendant would permit. Yet (c) plaintiff does have a right to insist that any proposed invasion creates no more risk of harm than a reasonable defendant would permit. Both intentional tort and negligence doctrine are involved here. Battery doctrine requires (a), negligence doctrine requires (c), and the two doctrines together require (b). Note that the "greater includes the lesser" fallacy reappears here. See supra text accompanying notes 103-04. Although plaintiff has the right not to suffer any intentional invasion, (a), it does not follow that she can insist that defendant's intentional invasion satisfy all of her preferences-including an extraordinary standard of care, (b). For the defendant has discharged his duty when he has satisfied (a) and (c), unless she specifically contracts for a higher standard .ofcare (b). In sum, intentional tort doctrine gives her no right to her preferred mode of treatment, but only the limited right to reject a form of treatment that has been inadequately explained to her. This narrower right protects the interest in bodily integrity without extending that interest to constrain defendant's freedom if he does act. In a sense, consent or preference protects that interest as a shield, but not as a sword. 124 Most qualifications of defendant's intentional tort duty enter into the defenses rather than the prima facie case, and thus are not qualifications of "duty" in a technical sense. Self-defense, necessity, and similar defenses create an important reasonableness component to the intentional tort. See, e.g., RESTATEMENT (SECOND) OF TORTS ch. 4 (Defenses of Person, Land, and Chattels-Recaption); id. at §§ 196-197 (1965) (public and private necessity). Many intentional tort cases differ from negligence cases mainly in burden of proof, and perhaps causation and damages, not in the ultimate standard of liability. See id. at § 435A (more liberal rules of causation for intentional torts); id. at § 908 (punitive damages); see also id. at § 826 ("intentional" nuisance divided into socially negligent harms, subsection (a), and harms that in effect create strict liability, subsection (b)). 125There are other similarities. For example, it seems that consent is no more likely than assumption of risk to be a bar when defendant has violated a criminal statute. Compare RESTATEMENT (SECOND) OF TORTS § 892C (1965) (consent to a crime) with id. at § 496F (assumption of risk and violation of statute). HeinOnline -- 67 B.U. L. Rev. 251 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 if a reasonable defendant would believe that plaintiff assumed the risk, even if plaintiff did not actually consent? For example, suppose the skater in Meistrich continued to skate on the improperly prepared ice, and the owner of the rink reasonably inferred that she fully preferred that condition when in fact she did not. Or suppose a doctor offers you the choice of two medicines, and you choose the more risky but more potent one; he reasonably believes that you understand his warnings, but you don't. In such circumstances, under traditional assumption of risk doctrine, it is not clear that plaintiff is barred. 126 The Restatement (First) specifically rejects apparent assumption of risk; 27 the Restatement (Second) does not discuss it, but emphasizes the requirements of subjective proof,12 an emphasis that may be inconsistent with an "apparent assumption of risk" defense. To be sure, subjectivity in this context might only draw attention to plaintiff's individual preference or state of mind, actual or apparent. Even apparent consent is "subjective" in examining whether a reasonable person in defendant's shoes would believe that plaintiff had (subjectively) consented, not whether a reasonable person in plaintiff's shoes would have consented. But traditional assumption of risk doctrine seems to read "sub29 jectivity" even more narrowly. 126 I have uncovered no cases specifically addressing apparent assumption of risk. Some commentators, however, recognize the concept. See R. Keeton, supra note 23, at 127-28, 152 ("objectively manifested consent to risk" should ordinarily be treated the same as subjective consent to risk); Mansfield, supra note 21 (discussed in.fra at note 130). 127 See RESTATEMENT (FIRST) OF TORTS § 893 comment a (1934): The rule stated in this Section differs from that stated in § 892 (Consent) in that the existence of the defense is dependent upon the plaintiff's state of mind. Thus, unlike the defense of consent, voluntary assumption of risk is a defense although there has been no manifestation of assent by the plaintiff to the defendant. Again, while consent is a defense if the defendant [sic: plaintiff] has caused the plaintiff [sic: defendant] reasonably to believe that assent exists, the defense of assumption of risk is not dependent upon the defendant's knowledge or belief as to the plaintiff's state of mind but depends upon the plaintiff's awareness of the situation and his realization of the risks to which he exposes himself. The Restatement (First) consent provision differs from the Restatement (Second) in requiring a manifestation of consent. Compare RESTATEMENT (FIRST) OF TORTS § 892 (1934) (requiring manifestation) with RESTATEMENT (SECOND) OF TORTS § 892 comment b (1965) ("Even without a manifestation, consent may be proved by any competent evidence to exist in fact."). 128 RESTATEMENT (SECOND) OF TORTS § 496B comments c-d (1985). 129 The reasonableness test incorporated in apparent consent doctrine might seem to submerge all of consent doctrine into negligence, but it does not, for two reasons. First, as suggested above, the apparent consent inquiry is still ultimately subjective: what would a reasonable defendant believe was the individual plaintiff's preference? The point is of great significance in informed consent law; recently courts have disagreed with the view, supposedly articulated in Canterbury v. Spence, 464 F.2d HeinOnline -- 67 B.U. L. Rev. 252 1987 1987] ASSUMPTION OF RISK Under my proposed full preference model, apparent consent and apparent assumption of risk are equally sound doctrines. If full preference is reason enough to bar recovery, then it suffices for both relatively certain and relatively uncertain risks. And for both types of risk, there are good reasons for more broadly permitting a defendant to escape liability when he reasonably believes that a plaintiff fully consents, and then acts upon that belief. Defendant's reliance is critical to his apparent consent or apparent assumption of risk claim. As Mansfield has noted, there is no good reason to protect a defendant's reasonable, but possibly mistaken, beliefs unless his beliefs affected his conduct. 13 0 Consider the Meistrich hypothetical above. If the owner reasonably believed that a particular skater preferred the risk, 772, 787 (D.C. Cir.), cert. denied, 409 U.S. 1064 (1972), that the informed consent duty to disclose should be examined from a reasonable plaintiff's perspective. See Scott v. Bradford, 606 P.2d 554, 558-59 (Okla. 1979). Second, different types of reasonableness judgments are required to determine whether defendant should know that plaintiff has consented and whether defendant was negligent. The first inquiry is much more limited in scope. And if the consequences of negligent misperception are severe-e.g., when, absent consent, defendant will have committed an intentional tort-then one might view negligence as a more serious shortcoming. Conversely, when plaintiff more understandably misperceives, as when she is estimating her right to use self-defense, we might view a negligent mistake as a less serious shortcoming. In theory, the negligence test itself is flexible enough to accommodate these differences: it will require greater care of the actor in the consent case in light of the seriousness of a mistake, and less care in the self-defense case in light of the emergency circumstances. But factoring the autonomy value or the emergency into a negligence equation may be difficult, and in practice, factfinders might inappropriately equate the two negligence inquiries. For another analysis of the differences between these reasonableness judgments, see Mansfield, supra note 21, at 38 n.29a, 66-67. 130 When the defendant acted on the basis of a belief in the plaintiff's willingness, then the question of actual willingness is in a sense ... subordinated to the manifestations of willingness that preceded or accompanied the defendant's action. We are concerned to give the defendant a practical guide to conduct. When the defendant did not act on the basis of a belief in the plaintiff's willingness, then our concern is less with manifestations than with the plaintiff's actual state of mind itself. If manifestations are important, it is not as a guide for the defendant's conduct, but as a basis for the court's or jury's own inference about the plaintiff's state of mind. For this purpose the plaintiff's words and actions after the event are as relevant as those that preceded it. The plaintiff's willingness is a circumstance to be inferred from his total conduct. Mansfield, supra note 21, at 33-34; see also id. at 57-59. Mansfield also agrees that defendant's belief in plaintiff's consent must be a reasonable belief. Id. at 26; see also RESTATEMENT (SECOND) OF TORTS § 892 comment c (1965) (describing apparent consent in part as "consent that will justify the other in acting in reliance upon [the manifestations of consent]"); W. PROSSER & W. KEETON, supra note 1, at 113 ("In our society we must perforce rely upon the overt works and acts of others, rather than upon their undisclosed minds."). HeinOnline -- 67 B.U. L. Rev. 253 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 when she did not, his belief is irrelevant to liability unless the owner relied upon it. If the owner refrained from warning plaintiff on this basis, perhaps the owner would not be liable. But if the owner provided a warning, thus treating the plaintiff the same as all other skaters who did not appear to prefer the risk, then the relevant question should be plaintiff's actual full preference. The reliance requirement helps explain why apparent consent is a more clearly established doctrine than apparent assumption of risk. Ordinarily, consent to an intentional tort precedes the intentional contact and is directly communicated to the defendant, who acts in reliance on plaintiff's expressions of consent. Apparent consent is therefore frequently invoked. By contrast, consent to, or assumption of, a risk of unintentional injury more often does not precede the conduct creating the risk, and more often is not directly communicated to the defendant. Thus, apparent assumption of risk, at least in the most obvious form, arises much less frequently. There are, however, less obvious forms of apparent assumption of risk. Consider the duty to warn in negligence and products liability cases. It is tempting to analyze all duty to warn cases under an apparent assumption of risk doctrine-tempting but perhaps misleading. When a manufacturer places adequate warnings on a product, those who use the product after reading the warning may have assumed the mentioned risks. Wouldn't a reasonable manufacturer believe that all users would read the warning? Not necessarily. Unlike the typical apparent consent to an intentional tort, in which defendant has specific knowledge about the individual plaintiff to justify his belief that plaintiff consents, a large number of plaintiffs may assume a risk created by defendant. Often the defendant will reasonably believe that most of them consent, but not that, each consents. Such a reasonable belief may well discharge defendant's duty, but it may not constitute apparent consent as to each plaintiff. Indeed, defendant may be virtually certain that some foreseeable plaintiffs, children and bystanders for example, do not consent. Yet perhaps duty to warn and apparent consent doctrine are not so different, after all. The doctor who reasonably but erroneously believes that an individual patient has understood and assumed a risk is relying on probable evidence. His usual method of ascertaining that the patient consents may involve a small risk of error. If he advises enough patients, he can be virtually certain that one of them will misunderstand. Yet his method may be reasonable in light of the costs of greater care in giving advice. The doctor's method is not substantively different from a drug manufacturer's placement of warnings on a product, such warnings being reasonably adequate for most users.131 Thus, duty to warn doctrine effectively incorporates apparent as131 To be sure, the manufacturer relies on what he expects will be the consent of the users, while the doctor sometimes relies on past expressions of consent or on more individual information. But again, the difference does not seem significant, at HeinOnline -- 67 B.U. L. Rev. 254 1987 1987] ASSUMPTION OF RISK sumption of risk. Instead of reasonably believing that an individual user has consented to a risk, when it is most likely that he did, the defendant reasonably believes that most users will consent to the risk, when it is most 132 likely that each, considered individually, will. A second comparison of consent with assumption of risk concerns the fullness of the preference. Consent doctrine as applied to intentional torts less often creates the troubling problem: has the plaintiff given full or only limited consent? If defendant suggests to plaintiff that they engage in an arm wrestling match, and plaintiff agrees, it is unlikely that plaintiff's consent is limited. Perhaps plaintiff would have preferred to engage in the match in a safer way-for example, by calling a draw after one minute. But such an undisclosed preference is unlikely to have existed, because it is so easy for plaintiff to disclose it. Moreover, defendant may also be able to rely on plaintiff's apparent consent. When the intentional tort consists, as it typically will, in a direct individual encounter between a plaintiff and defendant, and when plaintiff has the opportunity both to object that his consent is limited and to obtain the alternative he fully prefers, then consent will 133 usually be either actually or apparently full. Similar opportunities less frequently exist when defendant has only created a risk of harm to plaintiff. Plaintiff's only present opportunity may be to avoid the risky activity, not to secure a less risky alternative. When you notice a dangerous condition on your host's premises, it may often be impossible or infeasible to correct the problem while you are present. Nevertheless, there are cases in which a defendant, rather than creating a risk and then dropping out of the picture, responds to your manifestation of consent to the risk. For example, the organizer of a skydiving expedition may take you along only after you have consented to the risks. In such cases, assumption of risk very closely resembles classic cases of consent to intentional torts. 134 least under the flexible "'reasonableness" test for defendant's belief in plaintiff's consent. 132 See R. Keeton, supra note 23, at 143 (noting some overlap between "'objective" assumption of risk and the doctrine that plaintiff cannot recover for "'obvious" risks). If defendant manufacturer can identify a theoretical subclass within the class of users who would not apparently consent to the risk, such as those with an unusual allergic reaction, and if it would be prohibitively expensive to warn those users effectively, then apparent assumption of risk does not explain their nonrecovery, while no duty or no breach doctrine does. 133 But suppose defendant insists on an unlimited arm wrestle or none at all? Then if you accept, your recovery for ensuing damages would be barred not because you fully preferred the risk, but because you voluntarily assumed the relationship, which defendant was free not to offer. See supra Part lID. 134 See Mansfield, supra note 21, at 22 (describing as the strongest cases for a consensual bar "those cases in which the defendant acted only because he believed HeinOnline -- 67 B.U. L. Rev. 255 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 Third, according to the Restatement (Second), consent to an intentional tort need only be to conduct intended to invade plaintiff's interests, and need not extend to the invasion or resulting harm. 135 Curiously enough, such consent seems to require less knowledge of the consequences than assumption of risk. Consider the following illustration from the Restatement (Second): In a friendly test of strength, A permits B to punch him in the chest as hard as he can. B does so. Unknown to either A or B, A has a defective heart and as a result of the blow136he drops dead. A's consent is effective to bar recovery for his death. Suppose A had consented only to a risk of harm from a physical blow. For example, he sits very close to ringside at a boxing match, and the defendant organizer negligently permits the boxers to swing into the crowd. One boxer strikes A, who has a heart attack. It is unlikely that assumption of risk would be a bar, since most courts employing the doctrine require very specific knowledge of the risks assumed. 137 Moreover, the substantive test of negligence requires creation of an unreasonable risk of physical harm, not simply a physical touching. Thus, the content of one's consent should refer to resulting injury, not simply to invasive contact. But in substance, the two approaches may not be very different. Under neither approach is it necessary that plaintiff knowingly waive defendant's potential tort liability to him. Consent to intentional conduct with unforeseen results probably does express a preference for the dangerous conduct, even if the unforeseen risk were factored in. After all, the foreseen intended conduct itself may pose serious risks. By contrast, when plaintiff has assumed the risks of negligent conduct, the unforeseen results of that conduct may more frequently affect plaintiff's preference. Although one might examine, case by case, the relevance of unforeseen risks to a plaintiff's decision, that the plaintiff was willing to be exposed to the risk created by his conduct"). Compare RESTATEMENT (SECOND) OF TORTS § 496C comment f (third category of assumption of risk) (1965) ("[D]efendant has already violated his duty to the plaintiff, by creating a dangerous condition or situation, which continues after the defendant's conduct has terminated."), with id. § 496C comment c (first category of assumption of risk) ("[D]efendant is doing a dangerous act at a place where the plaintiff is privileged to be, or to remain .... and the plaintiff enters or remains within the area of danger, under circumstances which manifest willingness that the activity shall begin or continue."). 135 RESTATEMENT (SECOND) OF TORTS § 892A(l) comment e (1965); see I F. 3.10. illustration 5 (1965). HARPER, F. JAMES & 0. GRAY, supra note 1, § 136 RESTATEMENT (SECOND) OF TORTS § 892A 137See, e.g., Carr v. Pacific Telephone Co., 26 Cal. App. 3d 537, 542-43, 103 Cal. Rptr. 120, 124-25 (1971) (assumption of risk not a bar because plaintiff's decedent unaware of risk that tree would spring free from telephone lines while decedent was removing lines from tree). HeinOnline -- 67 B.U. L. Rev. 256 1987 1987] ASSUMPTION OF RISK judicial economy and concern about fraud may explain the categorical judgments. A fourth point of comparison between consent and assumption of risk concerns voluntariness. "Consent is not effective if it is given under duress," according to the Restatement (Second). 3 8 Although the Restatement does not define duress, it notes that case law has recognized only "those forms of duress that are quite drastic in their nature and that clearly and immediately amount to an overpowering of the will." 13 9 Again, assumption of risk doctrine differs. A much broader set of circumstances, including deprivation of any "right or privilege of which the defendant has no right to deprive [the plaintiff]," 140 renders an assumption of risk involuntary, as we shall see. In theory, the voluntariness requirements of consent and assumption of risk should be the same. To anticipate later discussion, 14' much of the rather stringent involuntariness doctrine for assumption of risk is more relevant to contributory negligence. Or it may reflect full preference indirectly, by elaborating the notion of unfair alternatives; if so, consent should be similarly analyzed. Perhaps the different Restatement (Second) requirements do have some factual basis, however. Duress is more likely to apply in paradigmatic consent situations, where plaintiff and defendant are communicating face-to-face, than in many assumption of risk situations, where plaintiff often confronts the risk outside of the physical presence of the defendant. 4 2 Finally, the different voluntariness requirements of consent and assumption of risk may be more apparent than real: the wording of the 43 consent requirement may be narrower through sheer inadvertence. Does assumption of risk doctrine help inform consent doctrine? Perhaps assumption of risk doctrine contributes by prompting close attention to whether the consent was reasonable. On first impression, it seems odd that traditional consent doctrine is not subject to the controversy that has beset assumption of risk doctrine-whether "reasonable" assumption of risk does or should exist. Does "reasonable" consent exist? Of course it does. Reasonable people frequently consent to operations, friendly physical contacts, and other intentional invasions that would be tortious but for their consent. 1-8 RESTATEMENT (SECOND) OF TORTS 139 Id. § 892B comment j: see also W. PROSSER & W. § 892B(3) (1965). KEETON, supra note 1, at 121. (SECOND) OF TORTS § 496E(2)(b) (1965). See inIra text accompanying note 186-97. 142 Not surprisingly, the Restatement (Second) gives no examples of an assumption of risk that is involuntary because of duress in the narrow sense of physical threats. See RESTATEMEMT (SECOND) OF TORTS § 496E (1965). 140 RESTATEMENT 141 143 The commentary to the two Restatement sections does not suggest a deliberate decision to employ different voluntariness requirements. See Id. comments a-d, § 892B comment j (1965). HeinOnline -- 67 B.U. L. Rev. 257 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 The more difficult question may be: does "unreasonable" consent exist? Here the answer has two parts. First, unreasonable conduct by a plaintiff is generally not a defense to an intentional tort. 144 Thus, even if "unreasonable" consent exists, its unreasonableness as such would not ordinarily bar recovery. Reasonableness is generally considered less important to intentional torts than to non-intentional torts. However, the supposed irrelevance 4 of plaintiff's unreasonable conduct is exaggerated. . Second, and more significantly, the concept of "unreasonable" consent is as unstable as the concept of "unreasonable" assumption of risk. In the sense of full preference, consent is neither reasonable nor unreasonable. It simply expresses what plaintiff wants or prefers. And that is often good enough reason to bar recovery. Assumption of risk controversy has also developed over whether the concept is better considered as part of defendant's duty, or as a defense based on plaintiff's conduct. 146 As applied to consent, I again doubt the significance of the controversy. With respect to intentional torts, even if one eschews the language of duty, one could say either that no intentional tort is committed when the plaintiff consents to the contact, 147 or that plaintiff's consent is a good reason for denying recovery for an intentional invasion of her interests. In either case, consent is a distinct reason for denying recovery, and it is more important to analyze the strength and scope of the reason than to place consent in a "no tort" or "no recovery for a tort" pigeonhole. 148 IV. THE MODEL IN PERSPECTIVE: TORT CASES AND SCHOLARSHIP The full preference model did not originate by spontaneous generation. The perspectives of case law and commentary help explain the formation of the model, as well as its place in tort theory. Case law relying upon the plaintiff's "venturesomeness,"' 149 and emRESTATEMENT (SECOND) OF TORTS § 481 (1965). For example, a plaintiff who could reasonably avoid the consequences of defendant's intentional tort may often be barred from recovery, under traditional law. See Id. § 918. Under comparative fault, some jurisdictions do reduce plaintiff's recovery against an intentional tortfeasor when plaintiff was negligent, a position that many commentators applaud. See Sobelsohn, Comparing Fault, 60 IND. L.J. 413, 441-43 (1985). 144 145 146 See supra notes 90-91 and accompanying text. See RESTATEMENT (SECOND) OF TORTs § 892A comment a (1965) ("no wrong is done to one who consents"). 148 Of course, burdens of production and persuasion may turn on whether consent 147 is an affirmative defense, or whether its absence is part of the prima facie case. Conventionally, consent is viewed as a defense. 149 See, e.g., Power v. Brodie, 460 N.E.2d 1241, 1243 (Ind. App. 1984) (plaintiff did not assume the risk of being struck by another vehicle); King Markets, Inc. v. HeinOnline -- 67 B.U. L. Rev. 258 1987 1987] ASSUMPTION OF RISK phasizing the independence of assumption of risk and contributory negligence, 150 lends some support to the model. So do decisions emphasizing that the risk was an essential part of the pleasure that plaintiff derived from his choice. 151 Of course, a larger number of cases endorse the broader and vaguer concept of "voluntarily encountering a known risk," without regard to whether plaintiff fully preferred the risky course; and "modernist" courts eschew assumption of risk altogether. One of the most explicit discussions of full preference in academic commentary is Judge Posner's economic analysis of assumption of risk. According to Posner, the ordinary negligence calculus assumes risk neutrality. The assumption of risk doctrine can increase economic efficiency as follows: People who are risk preferrers may not be willing to pay for what to the risk neutral would be a cost-justified precaution ....The economist is no more disposed to look behind the taste for risk than he is the taste for pickled herring .... In a world without risk preferrers, an assumption of risk defense would have no applications. The function of the defense is to permit those who like risk to trade on their taste. The defense of contributory negligence is unrelated. A plaintiff's willingness to waive cost-justified precautions by the defendant does not imply that the plaintiff could have prevented the accident by cost-justified precautions of his own. 52 Posner usefully focuses on the importance of a taste for risk. He is correct that the reasonable person standard differs from a properly conceived assumption of risk doctrine. However, his explicitly economic interpretation, though largely consistent with the proposed full preference model, 53 is not Yeats, 226 Va. 174, 181-82, 307 S.E.2d 249, 254 (1983) (plaintiff did not assume the risk of slipping in parking lot): Hunn v. Windsor Hotel Co., 119 W. Va. 215, 219, 193 S.E. 57, 59 (1937) (plaintiff assumed the risk when walking on unstable steps). 15oSee, e.g., Kennedy v. Providence Hockey Club Inc., 119 R.I. 70, 76, 376 A.2d 329, 332-33 (1977) (defining assumption of risk as "knowingly encountering the danger," and contributory negligence as "negligently encountering a risk"). 151 See, e.g., Murphy v. Steeplechase Amusement Co., 250 N.Y. 479, 481, 166 N.E. 173, 174 (1929) (Cardozo, C.J.), holding that plaintiff assumed the risk of injury from an amusement park attraction called "The Flopper": A fall was foreseen as one of the risks of the adventure. There would have been no point to the whole thing, no adventure about it, if the risk had not been there ....The tumbling bodies and the screams and laughter supplied the merriment and fun ....Whether the movement of the belt was uniform or irregular, the risk at greatest was a fall. This was the very hazard that was invited and foreseen. 152 R. POSNER, supra note 53, at 127-28. see also R. POSNER, ECONOMIC ANALYSIS OF LAW 158 (3d ed. 1986). Posner refers to "willingness to waive cost-justified precautions." R. POSNER, supra note 53, at 128. Full preference need not amount to an actual agreement to waive liability, as noted above. See supra text accompanying notes 42-52. Con153 versely, however, if a waiver of liability occurs in a market setting expressing full risk preference, it is more likely to be enforced. See R. Keeton, supra note 90, at 65 HeinOnline -- 67 B.U. L. Rev. 259 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 the only viable interpretation of the model. Suppose a market provides two versions of a product, one with a safety precaution and one without, and the plaintiff pays less for the less safe version.1 54 Plaintiff may well have assumed the risk. But the explanation need not be the principle that permitting risk preferrers to market their tastes has economic value. Rather, it may be the principle that you get what you've paid for (and fully prefer). Moreover, in many circumstances in which the proposed model applies, the economic interpretation does not. As one commentator has noted, courts have applied assumption of risk doctrine to persons who cannot trade over risk.1 55 Generally, consumers are more likely to be risk-averse than riskpreferring; buyers often do not discover risks until after the purchase; and buyers do not receive enough information about risk to calculate the appropriate risk-dollar tradeoff.15 6 Although these applications of assumption of risk are inconsistent with the economic interpretation, in some cases they would be consistent with my model. Consider, for example, a non-purchasing user of a product who fully prefers the less safe version because it is easier to use. Part of the explanation for the difference is that the model is not limited to persons who are risk-preferrers in the technical sense. It applies also to those with unusual or idiosyncratic tastes, without assuming that they generally prefer risk, or that they prefer taking a specific uncertain risk to its present expected value. 157 1 might choose the risks of skydiving even if in all ("[L]awyers who are employed to draft exculpatory clauses would do well to consider proposing a two-price option-one price with exculpation, and a higher price without.") Of course, full preference is only one factor relevant to whether an exculpatory clause should be enforced. 154 The plaintiff might prefer risk to such a degree that he willingly pays more to encounter risk than to avoid it; he does not simply assign less value to avoiding risk than does a risk-neutral person. That is, a risk-neutral person would pay a certain amount to avoid a risk with an equal expected value; a risk-preferrer might not only be content to pay less to avoid the risk, he might "pay" a negative amount for avoidance, i.e., pay to encounter it. For example, some hang gliders, mountain climbers, tightrope walkers, and indeed some ordinary patrons of amusement parks, prefer some risks in this stronger sense. Professor Gillette has mentioned to me a vivid example of the strong form of risk preference. A San Francisco museum recently presented an exhibit with a mechanical device connected to a rifle. The device was set to trigger the rifle exactly once in a 100 year period. Museum-goers were invited to sit in a chair in front of the rifle. Despite-or because of-full warning that the device might trigger at any time, patrons lined up for the opportunity to be subjected to the risk. 155 See Note, Assumption of Risk and Strict Products Liability, 95 HARV. L. REV. 872 (1982). 156 Id. at 880. This economic interpretation of assumption of risk has important distributional consequences. In theory, the same allocative effect would be achieved by allowing the injurer to "buy out" from liability. See Calabresi & Hirschoff, Toward a Test for Strict Liability in Torts, 81 YALE L.J. 1055, 1073 n.66 (1972). 157 See supra text accompanying note 53. HeinOnline -- 67 B.U. L. Rev. 260 1987 1987] ASSUMPTION OF RISK other respects I am unusually timorous, and without regard to the present value of those risks. For all of these reasons, neither traditional assumption of risk doctrine, nor the narrower full preference model, is restricted to situations in which the market model applies. The strict liability proposal of Professor Calabresi and Mr. Hirschoff' a overlaps the full preference model to some extent, but it is fundamentally different. They suggest that a party should be strictly liable in tort if he is in the best position to make the cost-benefit analysis between the cost of accidents and the cost of avoiding them.1 59 If that party is the plaintiff victim, then he is "strictly liable" in the sense that he cannot recover for his injuries, without regard to the possible "reasonableness" of his decision. In their proposal the question is not whether plaintiff should or should not have acted as he did, but who, plaintiff or defendant, was in the best position to decide that question. Traditional assumption of risk "is, and always has been, a kind of plaintiff's strict liability,"' 160 because it examines factors 1 relevant to whether plaintiff was the cheapest cost-avoider in this sense. 61 This analysis parallels my model in distinguishing assumption of risk from negligence or unreasonable conduct. But the explicitly economic or costbenefit interpretation has much narrower application and is probably a less appropriate rationalization of traditional doctrine. Calabresi and Hirschoff Calabresi & Hirschoff, supra note 156, at 1055. Id. at 1060-61. They also state that distributional goals might properly affect liability. Id. at 1076-84. 160 Id. at 1065. Calabresi and Hirschoff's argument deserves a fuller quotation: The doctrine of assumption of risk-though grossly misapplied by courts which have not looked realistically to whether the plaintiff in practice had the requisite knowledge and possibility of choice the doctrine implied-is essential 158 159 to an understanding of a non-fault world. It is, and always has been, a kind of plaintiff's strict liability-the other side of the coin of defendant's strict liability. It may even go to negate defendant's negligence, by expressing a judgment that although the defendant's conduct was not worth its costs (i.e., was negligent), the plaintiff was in a better position than the defendant to evaluate the costs and benefits involved (i.e., the plaintiff assumed the risk). Just as the employer may be in the better position to evaluate the costs and benefits of a piece of equipment given the likelihood of occasional employee negligence (defendant's strict liability), so a spectator at a baseball game may be best suited to evaluate the desirability of sitting in an unscreened bleacher given the likelihood of occasional negligent wild throws by the players during the game which may result in the spectator's being hit on the head (plaintiff's strict liability, or assumption of risk). In both these situations, the conclusion as to whether an accident cost should be shifted depends not on whether a party was negligent, but rather on a judgment as to which party was in a better position to make the cost-benefit analysis irrespective of the other's negligence. In each situation, strict liability (whether defendant's or plaintiff's) is imposed regardless of whether the other party "ought" to have done what he did. Id. (footnotes omitted). 161 Id. at 1062-66, 1073-74; see also Calabresi, Optimal Deterrence and Accidents, 84YALE L.J. 656, 664 (1975). HeinOnline -- 67 B.U. L. Rev. 261 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 presuppose that plaintiff and defendant, victim and injurer, are making the same ultimate inquiry: is the accident worth avoiding, given the costs and benefits? I do not so suppose. Indeed, one of the distinctive values of both traditional assumption of risk and my model is their refusal to inquire into reasonableness, or cost-benefit balancing, while relying instead on plaintiff's idiosyncratic desires and preferences. To be sure, one can capture some idiosyncracies in a cost-benefit equation. But it is one thing to say that a plaintiff possesses unique information relevant to such an equation. (The consumer, but not the manufacturer, knows that she has an allergy to the product.) It is quite another to recognize that plaintiff has unique desires and values. (The consumer has an unusual preference for the hallucinogenic side effects of the drug.) The full preference model better captures the latter possibility. 162 Neither of the two recent tort theories conventionally described as moral or corrective justice theories adequately explains assumption of risk. Professor Epstein's causal strict liability theory 163 has some features in common, for he emphasizes that assumption of risk does not evaluate the reasonableness of plaintiff's conduct. 16 4 However, the causal metaphysic of his theory See Schwartz, supra note 65, at 715-16; see also R. Keeton, supra note 23, at 154 (footnote omitted): But there is a difference between causing risk to oneself and causing risk to others. Consider, for example, the objective of amusement. It deserves attention in the weighing of utility against risk to determine whether conduct is unreasonably risky. Weighing the actor's interest in amusement against risk to others is quite a different matter from weighing the actor's interest in amusement against risk to himself. It may be reasonable to have fun at one's own expense or risk, when not so at the expense or risk of others. I agree that the plaintiff might not act "unreasonably" when he has fun at his own risk, but I would not give a causal explanation. Rather, such a plaintiff has made an individualistic decision not aptly described either as reasonable or as unreasonable. 163 R. EPSTEIN, supra note 43. Epstein extends and constitutionalizes the analysis of consent and assumption of risk in R. EPSTEIN, TAKINGS: PRIVATE PROPERTY AND THE POWER OF EMINENT DOMAIN ch. 11 (1985) [hereinafter R. EPSTEIN, TAKINGS]. Curiously, in one passage Epstein appears to adopt the very narrow "waiver" interpretation: "Assumption of risk [at common law] required some waiver, express or implied, of the rights of recovery that an injured party prima facie had against the defendant." Id. at 153 (emphasis in original). 164 The question is not whether the conduct of the plaintiff should be changed, but whether his assumption of the risk provides a reason within a system of corrective justice to shift back to the plaintiff the loss now presumptively 162 fastened to the defendant on causal grounds .... The prudence of the plaintiff's decision to assume the risk is his own affair, not that of the courts. R. EPSTEIN, supra note 43, at 100-01; see also id. at 106-07; Epstein, Products Liability: The Search for the Middle Ground, 56 N.C.L. REV. 643, 655 (1978) ("[Aissumption of risk applies when the plaintiff has knowledge of a dangerous situation and decides for his own benefit to encounter that danger."). But Lf. R. EPSTEIN, supra note 43, at 109 (noting that at a later stage of pleading, defendant can HeinOnline -- 67 B.U. L. Rev. 262 1987 1987] ASSUMPTION OF RISK requires that plaintiff "bring about" the harm to himself in one of the narrow paradigmatic ways central to Epstein's theory. 16 5 Under my model, only a decision to encounter a risk is required, even if this consists of a simple failure to leave a place of danger. Of course, the decision must also reflect 166 full preference-a condition that Epstein does not impose. Professor Fletcher's theory of reciprocal and nonreciprocal risks 16 7 bears less resemblance to the proposed model, and does not capture much of traditional assumption of risk theory. Under his theory, if a victim has decided to engage in an activity whose risks are reciprocal to the injurer's activity, the victim cannot recover. 168 Fletcher emphasizes the external risks characteristic of the victim's activity, not the victim's knowing and voluntary encounter of a risk created by the defendant.16 9 However, Fletcher's rejoin that "the plaintiff did not choose the course of action that allowed him to achieve his private objective ... with the greatest safety to himself and at the least possible cost to the defendant"). Notwithstanding Epstein's apology, this analysis does appeal in part to "notions of balancing or reasonableness." Id. at 109-10. 165 "Whenever the plaintiff assumes the risk of harm inflicted by the defendant, the case presumptively should be treated as though the plaintiff inflicted harm upon himself." R. EPSTEIN, supra note 43, at 97. For the exposition of the four causal paradigms-force, fright, compulsion,.and dangerous conditions-see id. at 22-46; see also Gordon, Wrong Turns in the Volens Cases, 61 LAW Q. REV. 140 (1945) (claiming that assumption of risk is inapposite when plaintiff has injured himself, because plaintiff should be barred on simple causal grounds). The causal argument is hardly so simple, however. See G. WILLIAMS, supra note 42, at 306 n. 1. Under modern proximate cause tests, plaintiff's self-injury will rarely be a superseding cause relieving defendant of responsibility. See also Twerski, supra note 37, at 37-39. 166 Thus, Epstein analyzes Eckert as a case of assumption of risk "because the plaintiff both knew of the particular hazard that he encountered and chose, for whatever reasons, to subject himself to it." R. EPSTEIN, supra note 43, at 106. Epstein does not consider whether the choice was sufficiently full. 167 Fletcher, Fairness and Utility in Tort Theory, 85 HARV. L. REV. 537 (1972). at 543-51. 168 Id. 169Fletcher briefly discusses contributory negligence as involving, in some cases at least, the creation of a risk "that unduly exceeds the reciprocal norm." id. at 548. But he also notes that in many cases of contributory negligence the plaintiff does not impose a nonreciprocal risk on defendant. Id. at 549 n.44. Fletcher does not, however, specifically discuss assumption of risk doctrine. In more recent writing, Fletcher does explicitly discuss assumption of risk. Fletcher, The Search.for Synthesis in Tort Theory, 2 LAW & PHIL. 63, 82-85 (1983). He asserts that plaintiff's bringing himself within the arena of defendant's risk properly reduces defendant's liability from strict to negligence. Id. at 83-84. He also notes that plaintiffs sometimes "not only expose themselves to an expectable risk, but... accept the risk of injury as the price of their activity." Id. at 84. Fletcher gives as examples informed consent and athletic injuries. Id. Fletcher nevertheless accepts the modern view that contributory negligence is the proper framework for analyzing such cases. Id. at 84-85. HeinOnline -- 67 B.U. L. Rev. 263 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 theory could easily be adapted to encompass full preference. The principle that defendants should not impose nonreciprocal risks may have little moral weight when a plaintiff is willing to accept a risk that would otherwise be nonreciprocal. 71 0 In 1961, Professor Mansfield wrote an extremely thoughtful essay 171 about assumption of risk that contains some observations relevant to the full preference model. His elaborate analysis isolates two significant categories of cases in which a plaintiff might be properly barred because of assumption of risk. 7 2 The first consists of cases in which defendant relies upon plaintiff's manifestation of consent, 173 a category I discussed earlier as apparent consent or apparent assumption of risk. 74 The second consists of cases in which the plaintiff seeks an advantage arising out of the defendant's conduct or property. 175 In the second category, Mansfield is equivocal about whether plaintiffs' claims should be barred, though he describes plaintiffs as seeking "something for nothing." 1 7 6 Rather, he approves only a narrower bar when plaintiff receives the very thing she desires. 177 His somewhat sexist example describes a coy young woman who wants to be kissed but indicates an aversion to it, and then brings a battery action. 178 "All that can be said is that 179 [she] has changed [her] mind about the value of this result." The full preference model only partially reflects Mansfield's views because Mansfield is unnecessarily reluctant to recognize assumption of risk when an undesired consequence occurs. The critical question is whether plaintiff fully preferred the alternative encompassing the undesired conse170 The following passage suggests that the broad principles underlying Fletcher's thesis are consistent with my model: The paradigm of reciprocity.., is based on a strategy of waiver. It takes as its starting point the personal rights of individuals in society to enjoy roughly the same degree of security, and appeals to the conduct of the victims themselves to determine the scope of the right to equal security. By interpreting the riskcreating activities of the defendant and of the victim as reciprocal and thus offsetting, courts may tie the denial of liability to the victim to his own waiver of a degree of security in favor of the pursuit of an activity of higher risk. Fletcher, supra note 167, at 569. It is not clear why plaintiff's "waiver of a degree of security" must occur only by his imposing risks on others, and not by his accepting risks to himself. But see supra note 169 and accompanying text. 171 Mansfield, supra note 21. 172 1 say "might be" because even in these two categories, Mansfield believes that defendant should sometimes be liable. Id. at 41-52, 68-69. 173 Id. at 22-30, 33-35. 174 See supra text accompanying notes 125-32. 175 Mansfield, supra note 21, at 59-62, 65, 71-76. Examples include viewing unlicensed fireworks, id. at 60; using a negligently placed steel platform as a personal canopy, id.; and participating in a guinea hen race, id. at 65-66. 176 Id. at 73. 177 Id. at 74-76. 178 Id. at 74. Id. 179 HeinOnline -- 67 B.U. L. Rev. 264 1987 1987] ASSUMPTION OF RISK quence over other alternatives that defendant should have offered. I do not desire to be injured by a firecracker, but I might fully prefer the package of benefits and risks that comprises watching the display. Mansfield's reluctance may stem from a sense that plaintiff's immediate choice between encountering and avoiding the risk is insufficient to work a bar. Agreed-but 180 the requirement of a full preference addresses the problem. The full preference model also derives some limited support from both Restatements of Torts. Under the Restatement (First), one form of assumption of risk is "where the defendant's conduct would be wrongful to the plaintiff but for the plaintiff's assent to its continuance and where the plaintiff is willing or desirous of having the conduct continue." 18 Under the 180 Professor Twerski would recognize a narrow and "pure" form of assumption of risk when defendant, although negligent or the manufacturer of an unreasonably dangerous product, has legitimately left plaintiff with a real voluntary choice. Twerski, supra note 37, at 39-43. The reasonableness of plaintiff's choice is irrelevant. Id. at 43. Twerski's approach bears some similarity to mine. However, he defines voluntariness, somewhat vaguely, as the absence of extreme duress from any source. Id. at 40. He does not specifically address the question whether the choice is limited or relevantly full, although the "legitimacy" of the choice left to plaintiff by defendant might bear on the question. British scholars who would restrict assumption of risk to waiver of liability have recognized exceptions that seem to confirm the full risk preference theory. Glanville Williams would permit a distinct assumption of risk defense "where the plaintiff deliberately brings the damage upon himself," or "where the plaintiff desires the consequences." G. WILLIAMS, supra note 42, at 313-14. Williams's only example is one who attempts suicide by throwing himself before a negligently driven bus. Williams does not attempt to justify this exception. However, it is rather clear that such a plaintiff fully prefers the risk-he is glad that defendant has provided the opportunity. Jaffey also acknowledges that an exception to his restrictive view exists. He posits the case of a woman deciding, after an unsuccessful sterilization operation, that she wants after all to have another child. In this case, he notes, unlike Williams's example, the plaintiff's conduct is reasonable, yet he would recognize a defense. Jaffey, supra note 42, at 98. Jaffey explains that such a woman-wants the child for its own sake, and deliberately suffers the loss, rather than merely taking a risk of loss. Id. More persuasive, I think, is the full preference model's explanation: she prefers, after all, that defendant had been negligent. 181 RESTATEMENT (FIRST) OF TORTS § 893 comment b (1934). As Twerski comments: [T]here is apparent good sense in a rule which denies a plaintiff who sought out the defendant and exposed himself to the danger from recovering. Perhaps, the no-duty advocates would argue that the relationship between the parties negates the duty when defendant does not pose the choice to the plaintiff as to whether he may continue in the reasonable exercise of his rights but, rather the plaintiff imposes himself on the defendant and seeks out the danger for the very purpose of benefiting from the dangerous condition. Twerski, supra note 37, at 6 n.28. HeinOnline -- 67 B.U. L. Rev. 265 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 Restatement (Second), a plaintiff who assumes the risk agrees to "take his own chances," 8 2 and "manifest[s] his willingness to accept" a risk.183 "The basis of assumption of risk is consent to accept the risk." 18 4 Of course, neither Restatement requires full consent in the sense that I have described.' 8l5 The Restatement (Second)'s treatment of the necessity of a voluntary assumption of risk lends indirect support to the full preference model. 8 6 Section 496E(2) provides: The plaintiff's acceptance of a risk is not voluntary if the defendant's tortious conduct has left him no reasonable alternative course of conduct in order to (a) avert harm to himself or another, or (b) exercise or protect a right or privilege of which the defendant has 187 no right to deprive him. If the broad terms "reasonable" alternative course, "harm," and "right or privilege" are given a wide scope, this language could essentially convert assumption of risk into contributory negligence. If they are given a narrow 182 RESTATEMENT (SECOND) OF TORTS § 496A comments c-d (1965). 183 Id. 84 § 496C. Id. § 496C comment h. Some of the Restatement examples, however, are of full preference. Consider the fireworks example from the RESTATEMENT (FIRST): 1. A is illegally and dangerously setting off fireworks on his land adjacent to a public highway, thereby endangering people on the highway. B, a neighbor, fully conscious of the risk of so doing, approaches on the highway for the sole purpose of seeing the fireworks and without A's knowledge stops to watch A. B is hurt by an apparently sound but defective rocket. A is not liable to B. RESTATEMENT (FIRST) OF TORTS § 893 comment b, illustration 1 (1934); see also RESTATEMENT (SECOND) OF TORTS § 496C illustration 1 (1965) (virtually identical example). This may not be a clear case of full preference, however, since B may have fully preferred that A not use defective rockets. A stronger case is another fireworks example: 1I. A is setting off large firecrackers in a public place with reckless indifference to possible harm to persons in the neighborhood. B and C approach the place where A is acting. B, fully conscious of the risks, approaches A for the purpose of enjoying the explosion. C is not desirous of hearing the explosion but carelessly continues in order to attend to business at a place beyond that where the firecrackers are being set off, although he realizes that some danger is involved. B and C are both hurt by the explosion of a large firecracker. A is not liable to B, but is liable to C. RESTATEMENT (FIRST) OF TORTS § 893 comment d, illustration 11 (1934); see also RESTATEMENT (SECOND) OF TORTS § 496C comment 1, illustration 7 (1965) (plaintiff assumes known risk of injury from a thrilling roller-coaster ride when she encourages defendant driver's high speed). 186 See RESTATEMENT (SECOND) OF TORTS § 496E (1965). 187Id. at § 496E(2). For an early discussion of this approach, see Bohlen, supra note 102, at 20-22. 185 HeinOnline -- 67 B.U. L. Rev. 266 1987 ASSUMPTION OF RISK 1987] scope, however, then the involuntariness criterion begins to approximate limited consent under the full preference model, and the remaining "voluntary" assumptions of risk begin to resemble full preferences. Let me explain. The Restatement comments on this section indicate that the rescuer's actions in Eckert should be considered "involuntary," since the rescuer had "no reasonable alternative" by which to "avert harm to... another." 188 If the rescuer acts unreasonably in choosing to rescue, however, then both contributory negligence and assumption of risk are a bar. 189 This analysis not only gives "involuntary" a highly peculiar meaning, it comes perilously close to equating assumption of risk with contributory negligence, in practical effect. Thus, in situations covered by § 496E(2), if plaintiff acts reasonably, she has acted "involuntarily," and is barred by neither contributory negligence nor assumption of risk. Conversely, if she acts unreasonably, she is barred by both. 90 But how broad is § 496E(2)? If it applies in every case,.then assumption of risk is a gratuitous doctrine. The language appears to be all-encompassing: "harm" and "right or privilege" might include every interest that would also 188 See (1965). 189 See RESTATEMENT (SECOND) OF TORTS § 496E comment c, illustration 3 id. at § 496E comment d ("As to either defense, the factors to be consid- ered in determining the existence of a reasonable and adequate alternative are much the same."). 190 Every example in § 496E where defendant leaves plaintiff with a "reasonable" choice is also an example of plaintiff's contributory negligence. See id. at illustrations 7-9. The fourth illustration, of a negligent rescue that does not amount to an assumption of risk, must refer to negligence in the manner of rescuing, not in the choice whether to rescue or not. Similarly, comment a to § 496E states that plaintiff's reliance on defendant's assurance of safety or promise of protection precludes assumption of risk unless the reliance is not "reasonable." Again, why the departure from a subjective standard? It is possible, but not likely, that the comment means to describe apparent assumption of risk. See supra text accompanying notes 125-32. The original version of the Restatement (First) defines voluntary assumption of risk as a type of contributory negligence, and then notes, in language strikingly similar to the "involuntariness" language in § 496E, some of the circumstances in which plaintiff's conduct is not unreasonable: In many cases the plaintiff is confronted with the necessity of either foregoing the exercise of a right or privilege or encountering a danger which he realizes or has reason to realize is involved in the condition created by the defendant.... RESTATEMENT (FIRST) OF TORTS § 466 comment c (1934); see also Jaffey, supra note 42, at 97 (claiming that under British law, a voluntary assumption of risk will almost always be an unreasonable assumption of risk). Professor Wade, among others, has also noted how misleading it is to "use the term 'voluntary' to pass on the reasonableness of a plaintiff's conduct." Wade, The Place of Assumption of the Risk in the Law of Negligence, 22 LA. L. REv. 5, 13 1961; see Rice, supra note 20, at 369-72. HeinOnline -- 67 B.U. L. Rev. 267 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 be relevant to whether plaintiff was contributorily negligent. Consider again Robert Keeton's "Black and Blue" example. 191 Keeton states that Black cannot recover for knowingly using the defective motorcycle, even though he may have reasonably chosen to use it. a2 But is Black's use "voluntary" under § 496E? Suppose Black were injured in an accident, and obtained a ride to get to the hospital. Hasn't the manufacturer left plaintiff "no reasonable alternative course of conduct in order to ... avert harm to himself "?' 93 Suppose Black's decision to use the motorcycle was "reasonable," as Keeton posits, for some other reason-say, because he needed it for important business and could not "reasonably" wait for a nondefective machine. Isn't his use of the product for this, or any, "reasonable" purpose the "exercise of a right or privilege of which the defendant has no right to deprive him"? Perhaps not. The comments are not terribly helpful in explaining this phrase, but the illustrations are of "illegally" blocking or rendering danger- ous a public sidewalk; 194 "illegally" carrying on blasting operations;19 5 barring a person's access to his stable; 9 6 and a city's failure to maintain a public sidewalk.197 Apparently, then, "right or privilege" refers to a statutory, property, or other independent legal right apart from those undifferentiated interests that one considers in determining whether plaintiff or defendant was negligent. 198 In addition, subsection (a)'s reference to "harm" may 191 See supra text accompanying note 41. 192 See R. Keeton, supra note 23, at 157. 193 See RESTATEMENT (SECOND) OF TORTS § 496E comment b, illustration 1 (1965) (suggesting a negative answer, at least where defendant is the driver rather than the manufacturer). 194 Id. § 496E comment c, illustration 5, comment d, illustration 7. 195 Id. § 496E comment d, illustration 8. 196 Id. § 496E comment c, illustration 6. 197 Id. § 496E comment d, illustration 9. 198 See id. § 523 comment f, illustration 1 (plaintiff does not assume risk of abnormally dangerous activity when he is exercising right to use highway); W. PROSSER & W. KEETON, supra note 1, at 491 (discussing "a valuable legal right"). Epstein adopts the narrower interpretation, though this is to be expected in light of his theory's explicit protection of property rights against utilitarian balancing. See R. EPSTEIN, TAKINGS, supra note 163, at 153-54; R. EPSTEIN; supra note 43, at 108-10. Dworkin also adopts a rights-based, non-utilitarian analysis in his recent work, urging that egalitarian principles govern conflicts of rights, but not conflicts of personal interests. See R. DWORKIN, LAW'S EMPIRE 293-95, 300-09 (1986). However, examples in W. PROSSER & W. KEETON, supra note 1, at 491 nn.7-8, do not seem to be restricted to a narrow class of rights. Since Prosser was the Reporter for the Restatement (Second), these examples support a very broad interpretation of the phrase. (The relevant text is unchanged from the fourth edition. W. PROSSER, THE LAW OF TORTS 451 nn. 16-17 (4th ed. 1971)); see also Rutter v. Northeastern Beaver County School Dist., 496 Pa. 590, 601, 437 A.2d 1198, 1205 (1981) (losing "privilege" HeinOnline -- 67 B.U. L. Rev. 268 1987 ASSUMPTION OF RISK 1987] include only physical harm. A narrower view of § 496E offers some support to a full preference theory. The drafters justify § 496E(2) as follows: The plaintiff's acceptance of the risk is not to be regarded as voluntary where the defendant's tortious conduct has forced upon him a choice of courses of conduct, which leaves him no reasonable alternative to taking his chances. A defendant who, by his own wrong, has compelled the plaintiff to choose between two evils cannot be permitted to say that the plaintiff is barred from recovery because he has made the choice.1 99 I would put the matter a little differently. When the defendant's tortious conduct has forced a choice upon a plaintiff, and when plaintiff is thereby foreclosed from engaging in the activity free of the risks created by defendant, then plaintiff's choice, though perhaps "voluntary," is not full or complete enough to bar his claim for damages. In many of the cases addressed by § 496E-rescues and risky activity to protect other valuable property rights-the plaintiff clearly has made only a limited choice. He prefers to encounter a risk rather than lose an independent legal right or suffer (or permit another to suffer) other harm. Because of the importance of what he wishes to avoid, his choice is probably not a full preference for risk, but rather a limited preference for the only available means for protecting his rights or preventing harm. The rescuer in Eckert much preferred saving the child to seeing the child be killed; but very likely he would have much 20 0 preferred avoiding the predicament altogether. V. SOME APPLICATIONS OF THE MODEL At the outset, I asserted that the full preference model helps answer certain puzzles within existing doctrine. Now I hope to vindicate that assertion by reviewing comparative fault, injuries in sporting events, and the traditional preservation of assumption of risk as a strict liability defense. to play on school football team raises "voluntariness" question under RESTATEMENT (SECOND) OF TORTS § 496E(2)(b) (1965)). The A.L.I. debates shed little light on the question. Dean W. Page Keeton suggested that the "involuntariness" provisions of § 496E(2) were a narrow exception, but others believed that they expressed a broad exception for lack of any reasonable alternatives. W. Keeton, Annual Meeting, 40 A.L.I. PROC. 405-06 (1963); Boskey, Breitel, id. at 459; R. Keeton, id. at 462-63 . 199 RESTATEMENT (SECOND) OF TORTS § 496E comment c (1965). 200 This likely relation is not a necessary one, to be sure. If you much prefer encountering a tortious risk, (2), to losing a right, (1), it may be because you fully prefer the risk in (2) to the lower risk in alternative (3), if defendant were not negligent. You love thrilling rescues. Or if your strong preference for (2) over (1) is based on dislike for (1), undoubtedly you will prefer (3) to (1); but it is still possible that you will prefer (2) to (3). HeinOnline -- 67 B.U. L. Rev. 269 1987 BOSTON UNIVERSITY LAW REVIEW A. [Vol. 67: 213. Comparative Fault The full preference theory helps resolve whether comparative fault should absorb assumption of risk. The modern view, that assumption of risk cases involve either contributory negligence or no negligence by defendant, answers the question in simple fashion: contributory negligence allows partial recovery; no negligence allows no recovery.2 0 1 And the modernists can point to an apparent dilemma for stubborn defenders of assumption of risk. Since unreasonable assumption of risk is a form of contributory negligence, the modernists argue, it should logically be compared to defendant's negligence. How, then, do we analyze reasonable assumption of risk? Should it be compared to defendant's negligence? But how can "reasonable" conduct be compared with fault? Should it be a complete defense? Then a reasonable plaintiff recovers less than an unreasonable one! This quandary may have 0 2 accelerated the trend toward abolition. But the dilemma is a false one. It does make considerable sense to treat similarly plaintiffs' reasonable and unreasonable assumptions of risk. 20 The appropriate equal treatment, however, is not to compare each type of assumption of risk to defendant's negligence, but to compare neither. If the assumption of risk reflects full preference, then it should ordinarily be a complete bar, 204 however one might characterize it. 205 If the same conduct Nga Li v. Yellow Cab Co., 13 Cal. 3d 804, 824-25, 532 P.2d 1226, 1240, 119 Cal. Rptr. 858, 872 (1975). See cases cited in Annot., 16 A.L.R. 4th 700, 711-15 (1982). 202 See, e.g., Blackburn v. Dorta, 348 So. 2d 287, 289, 293 (Fla. 1977) (refusing to allow reasonable implied assumption of risk to bar plaintiff's suit when partial recovery may be had for unreasonable and contributorily negligent conduct under the rubric of comparative negligence); J. FLEMING, supra note 42, at 264 (noting the dilemma); W. PROSSER & W. KEETON, supra note i, at 497-98 (noting same dilemma); Kionka, supra note 1, at 372 & n.9 (noting that advent of comparative negligence has prompted many states to abolish assumption of risk); Van Eman, Ohio's Assumption of Risk: the Deafening Silence, II CAP. U.L. REV. 661, 680 (1982) (preserving "consent" doctrine, but apparently only in contractual form). 203 Nevertheless, I have questioned our ability even to draw the distinction in many cases. See supra text accompanying notes 65-67. 204 See Kennedy v. Providence Hockey Club Inc., 119 R.1. 71, 76-77, 376 A.2d 329, 333 (1977) ("[Clontributory negligence and assumption of the risk do not over201 lap; the key difference is . . .the exercise of one's free will in encountering the risk."). However, I believe that the Kennedy court errs in treating any voluntary encountering of a known risk as "absolv[ing] the defendant of creating the risk." Id.; see also Riley v. Davison Constr. Co., 381 Mass. 432, 436-38, 409 N.E.2d 1279, 1282-83 (1980) (taking the same approach as Kennedy). Courts have had little difficulty concluding that express consent to waive a tort claim is not to be compared with defendant's negligence. See, e.g., Keegan v. Anchor Inns, Inc., 606 F.2d 35, 40 (3d Cir. 1979) (separating consent from defendant's negligence, but failing to address conduct that is both contributorily negligent and evidence of waiver or consent); Salinas v. Vierstra, 107 Idaho 984, 990, 695 P.2d HeinOnline -- 67 B.U. L. Rev. 270 1987 1987] ASSUMPTION OF RISK can be characterized as both contributory negligence and true assumption of risk, and if the first would allow partial recovery while the second would not, 20 6 then recovery should be barred. Consider the facts of Meistrich. If the plaintiff's decision to continue skating is characterized as unreasonable, it can be compared to defendant's negligence. If, in light of the peculiar benefits to plaintiff, her decision is characterized as reasonable assumption of risk, it cannot easily be compared. But these characterizations should be irrelevant if the facts demonstrate a true assumption of risk. If the plaintiff really prefers skating on "improperly" prepared ice to skating on "properly" prepared ice, then her full preference should ordinarily be a complete bar to recovery. B. Injuries in Sporting Events The full preference model helps explain why courts tend to allow recovery for some, but not all, risks knowingly and voluntarily encountered in athletic events. When one player injures another intentionally, and the victim sues either the player or another responsible defendant for battery, the defendant may raise as a defense consent or (in Mansfield's sense) assumption of risk. 20 7 The plaintiff is unlikely to recover for risks that "are almost inevita369, 375 (1985) (differentiating between contributory negligence and express assumption of the risk); Arbegest v. Board of Educ. of S. New Berlin Central School, 65 N.Y.2d 161, 169-70, 480 N.E.2d 365, 371, 490 N.Y.S.2d 751, 757 (1985) (express contractual assumption of risks of playing donkey basketball precludes liability even under broad comparative fault statute that abolishes assumption of risk, except as public policy proscribes such an agreement); Anderson v. Ceccardi, 6 Ohio St. 3d 110, 113-14, 451 N.E.2d 780, 783 (1983) (express assumption of risk separated from contributory negligence); see also J.FLEMING, supra 42, at 269 (under English law, contributory negligence permits apportionment, while assumption of risk remains a complete defense). 205 If there is a good argument for treating it as only a partial bar, for exacting some damages from the defendant notwithstanding the usual reasons for barring the plaintiff, then all cases of true assumption of risk should be so treated. See R. Keeton, supra note 23, at 159-60 (suggesting possible equal proration because plaintiff and defendant are each coauthors of the harm). But see supra note 162. 206 Recovery should be barred unless there is an affirmative reason for requiring defendants in such cases to pay damages, a reason that does not equally apply to "reasonable" assumption of risk cases. I am not aware of any convincing argument to this effect. But see RESTATEMENT (SECOND) OF TORTS § 496A comment d (1965) (suggesting that in a comparative negligence jurisdiction, all forms of contributory negligence should only reduce damages, without explaining why "reasonable" assumption of risk should be a more complete bar than "unreasonable" assumption of risk): see also Keegan v. Anchor Inns, Inc., 606 F.2d 35, 40-41 (3d Cir. 1979) (following the same reasoning). 207 Actual or apparent consent may be involved. For general discussions of consent and assumption of risk in connection with athletic injuries, see J. WEISTART & C. LOWELL, THE LAW OF SPORTS § 8.02 (1979); Lambert, Tort Law and Participant HeinOnline -- 67 B.U. L. Rev. 271 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 ble from the conduct on the part of the other contestants to which he... gives his assent." 20 8 Even violation of the rules of the sport might not lead to recovery. 20 9 However, a plaintiff may recover for injuries suffered from 210 certain acts, including but not limited to flagrant violations of the rules. Courts and commentators have found it difficult both to draw these lines and to justify their choices once made. The Restatement (Second) simply asserts: Taking part in a game manifests a willingness to submit to such bodily contacts or restrictions of liberty as are permitted by its rules or usages. Participating in a game does not manifest consent to contacts which are prohibited by rules or usages of the game if such rules or usages are designed to protect the participants and not merely to secure the better playing of the game as a test of skill. This is true although the player knows that those with or against whom he is playing are habitual violators of such rules.211 We receive no explanation of why plaintiff is not barred when he voluntarily accepts a known risk that defendant might injure him while merely violating Sports: The Line Between Vigor and Violence, 4 J. CONTEMP. L. 211 (1978); Note, Tort Liability in ProfessionalSports, 44 ALB. L. REV. 696, 697-99 (1980); Note, Torts in Sports-DeterringViolence in Professional Athletics, 48 FORDHAM L. REV. 764, 790-93 (1980); Note, Compensating Injured ProfessionalAthletes: The Mystique of Sport Versus Traditional Tort Principles, 55 N.Y.U. L. REV. 971, 976-81, 988-90, 994-96 (1980) [hereinafter Note, Compensating Athletes]; Comment, Torts in Sports-"I'll See You in Court!" 16 AKRON L. REV. 537, 546-49 (1983) [hereinafter Comment, Torts in Sports]; Comment, Assumption of Risk and VicariousLiability in Personal Injury Actions Brought by ProfessionalAthletes, 1980 DUKE L.J. 742, 745-58 [hereinafter Comment, Assumption ofRisk]; see also G. SCHUBERT, R. SMITH & J. TRENTADUE, SPORTS LAW 230 (1986) (broadly and incorrectly claiming that "[p]articipants do not consent to or assume the risk of injury from another's intentional or willful acts"). 208 1 F. HARPER, F. JAMES & 0. GRAY, supra note I at § 3.10; see also 4 id. § 21.5, at 238 n.17. 209 See, e.g., Kabella v. Bouschelle, 100 N.M. 461, 672 P.2d 290 (1983) (negligent violation of informal rules insufficient to state tort claim); see also RESTATEMENT (SECOND) OF TORTS § 50 comment b, illustration 5 (1965) (participants consent to violations of rules that are "merely to secure the better playing of the game as a test of skill," such as the offside rule in football); J. HENDERSON & R. PEARSON, supra note 87, at 71. But see Hackbart v. Cincinnati Bengals, Inc., 601 F.2d 516, 520 (10th Cir.), cert. denied, 444 U.S. 931 (1979) ("it is highly questionable whether a professional football player consents or submits to injuries caused by conduct not within the rules"). 210 See Hackbart, at 520-21; Nabozny v. Barnhill, 31 111. App. 3d 212, 215, 334 N.E.2d 258, 261 (1975) (liability for reckless violation of safety rule); Ross v. Clouser, 637 S.W.2d 11 (Mo. 1982) (same). 211 RESTATEMENT (SECOND) OF TORTS § 50 comment b (1965). HeinOnline -- 67 B.U. L. Rev. 272 1987 1987] ASSUMPTION OF RISK a "skill" rule.21 2 And this standard permits recovery whenever defendant violates a safety rule, even if the violation is quite routine--e.g., roughing the kicker or passer in football.2 13 Others have suggested that a player consents only to reasonably foreseeable contacts.2 14 But this is an odd limitation, since assumption of risk ordinarily requires subjective awareness of a risk in any event. Moreover, this suggestion is excessively broad, for some notoriously malicious and reckless acts are now all too foreseeable. 215 Some propose as a test whether the risk is inherent in, or "part of," the game. 216 Although this test has promise, it is regrettably vague. 217 And some courts have suggested or implied, without much explanation, that a player does not assume the risk of another's reckless conduct.2 1 8 Most puzzling of all is the artificiality of these tests of "consent." Why not simply apply the 219 usual test of consent to the facts of any sports injury? See id. illustration 5 ("offside" tackle in football does not create liability if it is no more violent than football rules would permit if player were "onside"). 213 Moreover, the distinction between rules of safety and rules of skill is not easy to 212 draw. Many safety rules could readily be given a "skill" justification. The penalty for roughing a kicker or passer may be designed not only to prevent physical harm but also to encourage the kicker or passer to play the game with more confidence and skill, by reducing his fear of injury. It is possible that customary "usages" would permit certain contacts that the rules themselves forbid. But this interpretation of RESTATEMENT (SECOND) § 50 comment b is problematical. See Note, Compensating Athletes, supra note 207, at 980 n.73. 214 See Bourque v. Duplechin, 331 So. 2d 40, 42 (La. App. 1976); Comment, Torts in Sports, supra note 207, at 547 nn.93-94; see also Woolridge v. Sumner, [1962] 2 Q.B. 43, 57 (Sellers, L.J.) (apparently combining reasonable foresight with reckless disregard for safety standard). 21' See supra text accompanying notes 17-19. 2'6 See J. WEISTART & C. LOWELL, supra note 207, at 936, 942; Comment, Assumption of Risk, supra note 207, at 758. 217 See Comment, Assumption of Risk, supra note 207, at 758 (test is whether risk is "normally associated with the game"). 2'8 See Note, Compensating Athletes, supra note 207, at 995-96. Courts adopting this analysis must be using the term "reckless" in a special sense. In the ordinary sense, it refers to a risk of harm less certain than an intentional tort requires; also, it may refer to defendant's subjective awareness of that risk. See RESTATEMENT (SEcoND) OF TORTS § 500 (1965). But in the sports injury cases, defendants often do have sufficient intent for a battery. Perhaps some courts have adopted recklessness as the liability standard because the term can also refer to an attitude of indifference to the welfare of others, a personal callousness that goes well beyond mere violations of rules against certain types of contacts. But that brings me to issues beyond the scope of this article. See also Lambert, supra note 207, at 216 (noting that recklessness requirement in jurisdictions that have abolished assumption of risk serves to limit duty and preclude liability for customary contacts, functions that assumption of risk formerly served). 219 Some courts have applied the usual consent or assumption of risk test to sports HeinOnline -- 67 B.U. L. Rev. 273 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 The full preference model may provide some insight. A player will ordinarily expect to incur some risk of injury from an athletic contest, and he obviously prefers taking that risk to not playing. But what is his full preference? I doubt that a professional hockey player prefers a game with no checking, or even a game with such stringent rules or officiating that no illegal checking would occur. This (reasonably apparent) full preference helps determine the proper scope of liability. But full preference theory does not entirely explain the proper scope of liability for sports injuries. Sometimes, although a class of plaintiffs clearly does not fully prefer a risk, the class cannot recover damages. For example, the class of football quarterbacks may well prefer more stringent protection against roughing the quarterback than the present rules afford. If they cannot recover for certain predictable injuries, even those resulting from violations of existing football rules, the reason must be something other than true consent. Perhaps it is because the quarterbacks' preference here cannot be satisfied without directly impairing the preferences of other players, spectators, owners, advertisers, and the radio and television audience. Or the legal standard barring recovery may simply ignore or override the quarterbacks' preferences to serve another goal, such as private profit or football tradition. Or the standard may be thought to promote aggregate social wealth. Whatever the precise justification, doctrinally we would say that neither the owners nor the other players have breached a duty to the objecting quarterbacks. As I have noted, 20 one salient no-duty category is plaintiff's voluntary assumption of a relationship. Many athletic injury cases fit this category, requiring a more complex method of analysis than the full preference model of assumption of risk. We cannot simply ask whether plaintiff agreed to the limited choice to play or not. Rather, we must ask whether defendant can reasonably demand that plaintiff play on these terms if he is to play at all. The different contemporary tort standards for consent to injuries in sporting events are unnecessarily confusing and incomplete because they often conceal these underlying problems within the plain fabric of "consent." C. Strict Liability According to the Restatement (Second), contributory negligence is not a 2 general defense to a strict liability tort, while assumption of risk is. 21 injuries. See Kuehner v. Green, 436 So. 2d 78 (Fla. 1983); Ross v. Clouser, 637 S.W.2d II (Mo. 1982). Other courts, however, have held that assumption of risk is inapposite. See Carey v. Toles, 7 Mich. App. 195, 151 N.W.2d 396 (1967) (adopting abolitionist approach). For such courts, the question arises why (or whether) consent to a more certain or more intentional sports contact should continue to bar recovery. 220 See supra Part liD. 221 See RESTATEMENT (SECOND) OF TORTS § 402A comment n (1965) (strict products liability); id. § 515 (wild or abnormally dangerous domestic animals); id. §§ 523- HeinOnline -- 67 B.U. L. Rev. 274 1987 ASSUMPTION OF RISK 1987] Curiously enough, the Restatement (Second) allows an assumption of risk defense in products liability only if the assumption was unreasonable, i.e., contributorily negligent. 222 The Restatement (Second) gives little explanation of why contributory negligence is not a defense to strict liability, and no explanation of why assumption of risk is.223 Modern commentary analyzing 224 strict liability and plaintiffs' conduct is considerably more sophisticated. 524 (abnormally dangerous activities): see also Franken v. City of Sioux Center, 272 N.W.2d 422 (Iowa 1978) (following the reasoning in RESTATEMENT (SECOND) OF TORTS § 515 (1965)). Although the Restatement (Second) recognizes one form of contributory negligence, it is simply unreasonable assumption of risk. See §§ 515(2), 524. 122 RESTATEMENT (SECOND) OF TORTS § 402A comment n (1965). Also curious is the position that only the "assumed risk" form of contributory negligence works a bar. "Ironically, the Restatement seems to reward plaintiffs who do not inspect the products they use." Gomberg, Alvis in Wonderland-Assumption of Risk No Longer a Complete Bar in Strict Liability Actions?, 32 DE PAUL L. REV. 299, 308 (1983). And it seems to encourage manufacturers to make defects more obvious. See Daly v. General Motors Corp., 20 Cal. 3d 725, 737-38, 575 P.2d. 1162, 1169, 144 Cal. Rptr. 380, 387 (1978). 223 The Restatement (Second) explains that contributory negligence is not a defense to strict liability for an abnormally hazardous activity because of the policy to place "full responsibility for preventing the harm resulting from [such an activity] upon the person who has subjected others to the abnormal risk." RESTATEMENT (SECOND) OF TORTS § 524 comment a (1965). The drafters give no specific explanation for preserving assumption of risk: presumably the usual consent rationale governs. Strict liability for animals and defective products is analyzed similarly. Id. at §§ 402A comment n, 515 comment b. For a rather unhelpful discussion, see W. PROSSER & W. KEETON, supra note I, at 565-67. Perhaps there is some sense to the Restatement (Second) position recognizing only unreasonable assumption of risk as a strict products liability defense. Some view the assumption of risk form of contributory negligence as more culpable, or more important to deter, than unknowing contributory negligence. See Williams v. Brown Mfg. Co., 45 111. 2d 418, 425-46, 261 N.E.2d 305, 309-10 (1970): see also Gomberg, supra note 222, at 315 (assumption of risk is simply "a more serious degree of plaintiff misconduct" than other forms of contributory negligence): Kionka, supra note 1,at 392-94 (defending same view, but without much argument): Note, supra note 155, at 890-91 ("Principles of fairness, while not sufficiently strong to support a classical assumption of risk defense, may be responsible for the current restriction of the contributory negligence defense to cases in which the plaintiff proceeded in the face of a known risk."). However, the full preference model does not justify this result. 224 See, e.g., R. POSNER, supra note 53, at 138-41; Calabresi & Hirschoff, supra note 156 passim (whether an accident cost should be shifted depends not on whether a party was negligent, but rather on which party was in a better position to make the cost-benefit analysis): Epstein, supra note 164, at 654-57: Noel, Defective Products: Abnormal Use, Contributory Negligence, and Assumption of Risk, 25 REV. 93, 107-19 (1972). HeinOnline -- 67 B.U. L. Rev. 275 1987 VAND. L. BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 The case law contains a variety of approaches. Although there is some support for recognizing contributory negligence as a strict products liability defense,22 5 and general support for recognizing unforeseeable misuse as a negation of duty or breach, 226 there is little support for recognizing traditional assumption of risk as a defense, at least where the assumption is not also contributory negligence. 22 7 For example, courts now disfavor the rule that a plaintiff cannot recover for open and obvious product defects. 22 8 Some conduct that is courts accept the Restatement (Second) position that 2only 2 9 both negligent and an assumption of risk is a defense. Whatever the appropriate treatment of contributory negligence, there is good reason to treat assumption of risk (in the full preference sense) as a bar. The full preference theory justifies such a bar whatever the nature of defendant's tort, and whether or not plaintiff was also negligent. Moreover, in those strict liability cases in which defendant cannot make the product or condition safer, it may be especially clear that plaintiff fully preferred the risk.2 30 "Thus a passenger who chooses to travel by air in an abnormally dangerous jet plane, still of an experimental character, at supersonic speed, will assume the risk inseparable from that type of transportation." ' 23 ' And 225 See Daly v. General Motors Corp., 20 Cal. 3d 725, 575 P.2d 1162, 144 Cal. Rptr. 380 (1978); see also Noel, supra note 224, at 114-15; Schwartz, Strict Liability and Comparative Negligence, 42 TENN. L. REV. 171, 178 (1974). 226 For a discussion of the case law, see Kionka, supra note 1, at 397-98; Noel, supra note 224, at 95-105; Schwartz, supra note 225, at 172-73. 227 See Daly v. General Motors Corp., 20 Cal. 3d 725, 742, 575 P.2d 1162, 1172, 144 Cal. Rptr. 380, 390 (1978) (assumption of risk abolished insofar as it is a form of contributory negligence); Note, supra note 155, at 873-74 (describing limited support for assumption of risk defense in the case law); see also Kionka, supra note i, at 392-96, 400-01 (advocating recognition of unreasonable assumption of risk as a strict liability defense that may reduce recovery under comparative fault). Glanville Williams claims that assumption of risk, in the narrow sense of waiver, is more likely to exist when liability is strict, because of the "presumption that people do not intend to excuse others from their legal duty of [reasonable] care." G. WILLIAMS, supra note 42, at 313-14. But it is not clear why we should not also presume that people do not intend to excuse others from their strict liability duties. 228 See, e.g., Micallef v. Miehle Co., 39 N.Y.2d 376, 348 N.E.2d 571, 384 N.Y.S.2d 115 (1976). 229 See, e.g., Simpson v. General Motors Corp., 108 Il. 2d 146, 483 N.E.2d 1 (1985); Austin v. Raybestos-Manhattan, Inc., 471 A.2d 280, 287 (Me. 1984); Duncan v. Cessna Aircraft Co., 665 S.W.2d 414, 432 (Tex. 1984); see also Kionka, supra note 1, at 372. 230 In determining whether an activity is abnormally hazardous, for example, the RESTATEMENT (SECOND) OF TORTS § 523 (1965) considers "inability to eliminate the risk by the exercise of reasonable care" as one important factor; see also id. comment d (participant assumes "necessary" risks). 231 Id. § 523 comment g. I omitted the final phrase, "even though the plane is a common carrier." Id. The Restatement drafters are troubled that one ordinarily has HeinOnline -- 67 B.U. L. Rev. 276 1987 1987] ASSUMPTION OF RISK "one employed as a lion tamer in a circus may be barred from recovery by 232 his assumption of the risk when he is clawed by a lion.However, the full preference theory needs to be applied with special care in strict liability cases. Often, a strictly liable defendant is required to 233 compensate for harm not because defendant should have acted otherwise, 2 34 In such but because of other social goals, such as loss-spreading. "pure" strict liability cases, since plaintiff cannot legitimately complain that defendant should have acted otherwise, she cannot complain that she would have fully preferred a third option. For example, a consumer injured by a manufacturing flaw cannot necessarily insist that defendant should have adopted a better manufacturing process. Defendant may have adopted the the "right" to use a common carrier. Under their peculiar analysis, the assumption of such a risk would be involuntary. To distinguish this case as voluntary, they must rely on the weak arguments that the risk is "necessary" and that "free alternatives" exist. Id.; see supra text accompanying notes 186-98. 232 RESTATEMENT (SECOND) OF TORTS § 515 comment e (1965); see also id. at § 523 comment d (1965) ("Thus a plaintiff who accepts employment driving a tank truck full of nitroglycerin, with knowledge of the danger must be taken to assume the risk when he is injured by an explosion."). Such cases give partial support to the "fireman's rule," under which a professional rescuer, such as a police officer or firefighter, cannot sue for injuries inherent in her hazardous job. See, e.g., Walters v. Sloan, 20 Cal. 3d 199, 571 P.2d 609, 142 Cal. Rptr. 152 (1977). When the tort claim is based upon strict liability for an abnormally dangerous activity, the fireman's rule may illustrate relevantly full risk-preference, since the plaintiff cannot insist that the defendant should have acted otherwise. For some examples of the fireman's rule applied to abnormally dangerous activity, see Lipson v. Superior Court of Orange County, 31 Cal. 3d 362, 182 Cal. Rptr. 629, 644 P.2d 822 (1982) (fireman's rule held inapplicable to acts of misconduct unrelated to fireman's duties); Armstrong v. Mailand, 284 N.W.2d 343, 352 (Minn. 1979) (reasoning that the fireman's rule reflects "primary assumption of risk" or no duty, even in a strict liability case); see also Peneschi v. National Steel Corp., 295 S.E.2d I (W. Va. 1982) (applying assumption of risk to employee of independent contractor hired to work with abnormally dangerous instrumentality). However, when the basis of liability is negligence, the fireman's rule typically does not illustrate full preference, since the firefighter prefers that the defendant had not negligently caused the fire. 233 Alternatively, he should have acted otherwise only in the somewhat circular sense that he should not have caused the harm without paying for it. See R. Keeton, Conditional Fault in the Law of Torts, 72 HARV. L. REV. 401, 418-23 (1959). 234 To be sure, sometimes strict liability is a kind of strict medicine to induce non-negligent behavior. See Escola v. Coca-Cola Bottling Co., 24 Cal. 2d 453, 461-64, 150 P.2d 436, 440-41 (1944) (Traynor, J., concurring). Indeed, "strict liability" in the products context often is a type of negligence. A defective product is sometimes defined as one that reasonably should have had a better warning, or should have been better designed in light of the risks and utility of the existing and proposed designs. See W. PROSSER & W. KEETON, supra note I, at 697, 699. HeinOnline -- 67 B.U. L. Rev. 277 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 best possible process (or at least a non-negligent one). If the consumer is aware of a small risk of a defect and uses the product, has she assumed the risk? She might fully prefer the impossible, or the possible but extraordinarily expensive, i.e., a safer manufacturing process. Because this alternative is beyond defendant's duty to offer, it is irrelevant to whether she has assumed the risk. Her decision to use the product indicates that she prefers the risk in a relevantly full sense. Does every duly informed plaintiff in a pure strict liability case assume the risk? That is a possible conclusion. But strict liability rules have justifications different from other tort rules, justifications which could support the opposite conclusion, that assumption of risk should never be a defense to strict liability. If loss-spreading is a principal goal, 235 for example, why should it matter that plaintiff prefers using the product to not using it? Her use as such is unlikely to be negligent. After all, we may have denominated the product defect as one of strict liability, not negligence, precisely because we have concluded that the defendant was not negligent for offering it to the public; a typical user cannot then be negligent simply for deciding to use it. Moreover, the defendant's duty may be to compensate an informed plaintiff despite her predictable full preference. Consent is insufficient; the duty requires more.2 36 Finally, it is also possible to measure plaintiff's full preference in a pure strict liability case through hindsight by considering plaintiff's regret at how matters turned out.2 37 In practice, this will invariably result in a conclusion that consent is not a bar. .,5 See Escola v. Coca-Cola Bottling Co., 24 Cal. 2d 453, 462, 150 P.2d 436, 441 (1944) (Traynor, J., concurring); G. CALABRESI, THE COSTS OF ACCIDENTs 39-67 (1970) (recounting justifications for loss spreading). 236 See Twerski, supra note 37, at 4 (emphasis in original): [Wie shall ask of each product and each situation: is it the desire of the law to impose a duty upon defendants to preclude plaintffs from choice-making? ... Only after squarely facing this major policy question can assumption of risk as a doctrine of the law designed to evaluate plaintiff's behavior come into play. See also R. Keeton, supra note 23, at 166 (assumption of risk would be relevant, but no longer dispositive, if strict liability were imposed for all risks "typical" of an enterprise). Here is another possible analysis: the relevant additional option that plaintiff fully prefers is using the product, with a risk of harm, and receiving compensation for that harm. As noted above, this option is ordinarily beyond defendant's duty to offer. See supra text accompanying notes 62-63. But some strict liability doctrines may amount to a duty of the defendant to compensate, despite plaintiff's otherwise full preference for the risk. Plaintiff will certainly "prefer" the exercise of that duty in the fullest sense, for this simply amounts to a refusal to waive compensation. Whether such a duty exists is the important and difficult question. 237 In cases of pure strict liability, when liability is not premised on the judgment that defendant should have acted otherwise, a hindsight test is sometimes applied to determine defendant's liability. See, e.g., Beshada v. Johns-Manville Prods. Corp., 90 N.J. 191, 204-07, 447 A.2d 539, 546-48 (1982); see generally Henderson, HeinOnline -- 67 B.U. L. Rev. 278 1987 1987] ASSUMPTION OF RISK VI. CONCLUSION In a small number of cases, assumption of risk can continue to play a valuable role in tort law. If we isolate the important common element in assumption of risk and consent doctrine, a narrow concept of full preference emerges. We should not simply ask whether plaintiff voluntarily and knowingly encountered the risk that defendant created, when she could have avoided that risk. Rather, we should ask whether plaintiff fully preferred to take the risk, i.e., whether she preferred the risky alternative that she chose to the alternative that defendant tortiously failed to offer. Consider a classic assumption of risk problem, the liability of the owner of a baseball park to a spectator who is hit by a ball while sitting in an unscreened seat. Suppose the owner owes and breaches a duty to provide an adequate number of screened seats in certain locations.2 38 But suppose that the spectator, aware of the risk of injury, nevertheless prefers an unscreened to a screened seat, because it provides a slightly better view.2 39 Then she fully prefers to take the risk, and should ordinarily be barred from recovery. This analysis improves upon both traditional assumption of risk doctrine and the modem view that assumption of risk should be assimilated into other doctrines, especially contributory negligence. The traditional doctrine must treat many cases (such as rescues) in which plaintiff reasonably and knowingly encounters a risk as "involuntary," a tortured interpretation. (We Coping with the Time Dimension in Products Liability, 69 CALIF. L. REV. 919 (1981) (questioning whether courts should rely upon hindsight in judging design and marketing decisions). An intriguing question is whether it is then appropriate to apply a hindsight test to plaintiff's risk preference. Ordinarily, I have suggested, plaintiff's hypothetical preference should be determined by a foresight test: what would plaintiff have fully preferred, without his present knowledge of how matters actually turned out'? However, if the point of pure strict liability is not that defendant should have acted otherwise, but only that he should pay for the harm he has caused, then there is much less reason to ask what plaintiff would have chosen prior to the injury. Consider a manufacturer's liability for an exploding bottle of a carbonated beverage. If the theory of liability is negligence, then we want to know whether the plaintiff gets a thrill out of the greater danger of explosion, or is willing to pay less for a more dangerously manufactured product. And we want to know how plaintiff would have answered prior to receiving the injury, to see whether her preference is inconsistent with recovery. But if the theory is pure strict liability, then why not ask whether she now regrets the preference she once had? Acknowledging her regret as a measure of her preference is not obviously inconsistent with demanding that defendant compensate for the harm he does. 238 See, e.g., Brown v. San Francisco Ball Club, Inc., 99 Cal. App. 2d 484, 487-88, 222 P.2d 19, 20-21 (1950); 4 F. HARPER, F. JAMES & 0. GRAY, supra note I, § 21.5. 239 Her preference would be clearest if she actually had the opportunity to sit behind the screen, but she declined. HeinOnline -- 67 B.U. L. Rev. 279 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 might even describe the baseball spectator's plight in my example as "involuntary" under the Restatement, since she had no reasonable alternative way to protect her privilege to see the baseball game. 2 40 ) Some, possibly including Prosser, apply the traditional doctrine only when the plaintiff has impliedly agreed to waive his tort claim against the defendant. This application is so narrow that the value of preserving the defense under the title of assumption of risk is questionable. (Under this interpretation, the baseball spectator probably would not have assumed the risk.) On the other hand, the modern assimilationists cannot easily account for cases of true full preference. One who prefers a risky alternative is not necessarily contributorily negligent (consider the spectator, again), and the defendant who presents her with the risky choice may have breached a duty to her. The modern view also ignores the link between assumption of risk and the traditional doctrine of consent, a doctrine which is not restricted to plaintiffs who consent "unreasonably." However, I do not wish to overstate the importance of assumption of risk as full preference. Sometimes a plaintiff should be permitted to recover despite such a preference. And the modern assimilationists are correct that most traditional assumption of risk cases are best analyzed differentlyeither the plaintiff's recovery may be barred or reduced because of plaintiff's contributory negligence, or it may be barred because defendant breached no duty to plaintiff, or perhaps it should not be reduced at all. In particular, a defendant who has no duty to engage in a relationship with plaintiff may often have only a limited duty when plaintiff voluntarily assumes that relationship. And some duties of care depend quite directly on the preferences of those who will interact with defendant, defendant's activity, or defendant's product. These and other "no duty" or "limited duty" concepts do rely in significant part on policies of respecting individual informed choice and protecting a defendant's reliance on that choice. Nevertheless, assumption of risk, like traditional consent, relies so exclusively on such policies that it deserves recognition as a separate tort category. And what deserves recognition is not the traditional notion that any voluntary encountering of a known risk bars the plaintiff, but the germ of truth within that notion: a plaintiff "consents" in a dispositive way only if he fully prefers the course he chooses over all alternatives that the defendant tortiously failed to offer.. Thus, those cases traditionally analyzed as "voluntary assumption of a known risk," and in which plaintiff's recovery is properly barred or reduced, should be analyzed as falling within one of three categories: contributory negligence; no duty or no breach (including voluntary assumption of a relationship); or full preference. If I am correct, it might be well to discard the nomenclature of "assumption of risk," with its multiple and mainly unacceptable meanings. In cases of true, full preference, courts 240 See supra text accompanying notes 186-90. HeinOnline -- 67 B.U. L. Rev. 280 1987 1987] ASSUMPTION OF RISK might say that plaintiff "fully consented to," "fully agreed to," or "fully preferred" the risky alternative. In cases of voluntary assumption of a relationship, courts might say that plaintiff "chose" or "agreed to" the relationship with the defendant. Such rhetorical changes only have value, however, if accompanied by greater insight into the underlying concepts. Only in this way can we successfully resist seduction by an unrestricted doctrine of assumption of risk. APPENDIX: THE DEBATE OVER THE Restatement (Second) of Torts PROVISIONS CONCERNING ASSUMPTION OF RISK The debate over the assumption of risk provisions in the proposed Restatement (Second) of Torts is quite revealing. Dean Prosser, the Reporter, indicated that these provisions created the first major disagreement among the Restatement Advisors.2 41 Despite the vehemence of their disagreement, however, the argument seemed technical and narrow when the issue reached the floor of the American Law Institute. The main reason, I believe, is that many of the principals had only a vague conception of the meaning of "consent" to a risk. In the account that follows, I emphasize the portions of the debate that address the meaning of "consenting" to, "acquiescing" in, or being "willing to accept" a risk. Prosser's proposed draft, as presented to the A.L.I. membership, contained two critical sections. "Express" assumption of risk bars a plaintiff "who by contract or otherwise expressly agrees to accept a risk of harm arising from the defendant's negligent or reckless conduct. ' 242 "Implied" assumption of risk bars a plaintiff who understands a risk of harm "and who nevertheless manifests his acquiesence by voluntarily" encountering the risk.243 When the A.L.I. members debated the proposed assumption of risk chapter, they had before them documents detailing the broad disagreement among Restatement Advisers. Laurence Eldredge's memorandum to the A.L.I. Council recognizes a situation of "true consent," in which the plaintiff suffers no legal injury and defendant is not a wrongdoer. Consent is determined by plaintiff's state of mind and "willingness" to suffer a risk of bodily invasion. 244 A plaintiff soliciting a ride from a drunk is an example of such consent, according to Eldredge. Apart from consent, assumption of RESTATEMENT (SECOND) OF TORTS § 893, at 70 (Tent. Draft No. 9, 1963). Id. § 496B, at 92; see also RESTATEMENT (SECOND) OF TORTS § 496B (1965) (using identical language in the adopted version). 243 RESTATEMENT (SECOND) OF TORTS § 496C, at 96 (Tent. Draft No. 9, 1963); iee also RESTATEMENT (SECOND) OF TORTS § 496C (1965) (the adopted language is similar; plaintiff is barred if he voluntarily encounters a known risk "under circumstances that manifest his willingness to accept it"). 244 RESTATEMENT (SECOND) OF TORTS § 496C, at 72-74 (Tent. Draft No. 9, 1963). 241 242 HeinOnline -- 67 B.U. L. Rev. 281 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 risk should be abolished. Professor Malone's memorandum expresses sub245 stantial agreement with Eldredge. In his reply, Prosser proclaims this position "little short of astonishing. ' '1 4 Why, asks Prosser, do the anti-riskers-whom Prosser dubbed "the Confederate Army"--accept assumption of risk when it is express, in words, but reject it when it is tacit and by conduct? 47 They also err in assimilating too many assumption of risk cases into the "no duty" category. "Duty is a matter of what the defendant should reasonably expect, while assumption of risk is a matter of what the plaintiff in fact knows and is willing to accept." 248 Prosser also notes certain doctrinal differences between assumption of risk, on the one hand, and no duty or contributory negligence, 2 49 on the other. Professor Wade's "replication" takes issue with Prosser on the meaning of consent. Consent can be by conduct, Wade agrees, but Prosser errs in giving it a fictional meaning, assuming that the concept applies whenever plaintiff encounters a known risk. Rather, consent is a subjective concept that requires "actual agreement, whether express or implied. ' 250 For example, if a buyer knowingly purchases a defective chattel at a reduced price, "a ' 251 true assumption of risk would properly be implied. With these memoranda, together with Prosser's proposed draft and comments, in hand, the A.L.I. takes up the debate. After Prosser explains the history of the debate, Dean Page Keeton sets forth the "Confederate" position. Assumption of risk should be narrowly restricted to what I have called the "waiver" conception of consent-where plaintiff has "agreed not to hold the defendant responsible." 252 This intent may be express or tacit, Keeton explains, in response to Prosser's charge that the Confederates arbitrarily accepted only "express" consent. "So our distinction is not between express and implied assumption of risk, but rather between a case where he agrees not to hold the defendant responsible and a case where 2 53 there is no such agreement." Eldredge's subsequent statement takes up the same theme. His interpretation of consent, however, like his earlier written statement, seems broader than the "waiver" interpretation. "Is the plaintiff willing to have this thing happen to him, or is he willing to take the risk that it will happen to him? And if he is, then the defendant who does this invasion just is not a tortfeasor at 246 Id. at 76-77. Id. at 78. 247 Id. 145 Id. at 79. Id. at 79-83. 250 Id. at 84. 251 Id. at 85. 248 249 252 W. Keeton, Annual Meeting, 40 A.L.I. PROC. 371 (1963). 253 Id. at 372. HeinOnline -- 67 B.U. L. Rev. 282 1987 1987] ASSUMPTION OF RISK all.' '254 For example, if plaintiff walks into a room with rattlesnakes and chooses to stay, he assumes the risk.255 Prosser asks the Confederates just what it is they wish to strike in the chapter. 25 6 Dean Keeton explains that § 496B, covering "express" agreements, may actually be too narrow, but is acceptable if extended to any agreement to relieve defendant of responsibility. 257 However, he strongly objects to the critical section, § 496C, which declares that plaintiff is barred in certain cases by his voluntarily encountering a known risk. "So our primary difficulty is with Section 496C, if I understand 496C correctly to mean that it will apply even though there is no tacit agreement on the part of the plaintiff not to hold the defendant responsible. ' 258 Professor Wade underscores this point of difference. He notes that unlike earlier drafts, the current draft of § 496C is closer to the "Confederate" position, since it now requires that plaintiff "manifests his acquiescence" to a risk, not that he simply encounter it voluntarily and knowingly. Still, there is a gap: "You are saying that the person who deliberately decides to stay does indicate his acquiescence, and we are trying to insist that there must be an actual agreement, actual consent, on the part of the plaintiff in the case." 219 In response to this and other arguments, Prosser shifts to a different point: whether a given situation is analyzed as "no duty" or "assumption of risk" significantly affects the burden of proof. 26 0 Dean Keeton clarifies that the major disagreement is not over this point, but rather over the breadth of 261 § 496C. Judge Breitel expresses the belief, and perhaps the hope, that the differences between Prosser's draft and the Confederates' position are narrow- ing.2 62 In Breitel's view, there really is an "agreement" whenever "there is full knowledge of ...the risk involved and one has a completely free choice whether to engage in the conduct or not." 263 Unfortunately, Breitel neglects to explain what he believes such a plaintiff has agreed to. Again, he suggests that " 'voluntary' means more than mere acquiescence in an appreciated risk,''264 yet he does not explain what else it means. In a brief comment, Prosser amplifies the confusion by insisting that he Eldredge, id. at 374. Id. at 375. 256 Prosser, id. at 396-97. 257 W. Keeton, id. at 397. 258 Id. 259Wade, id. at 398. (It is not clear, however, whether Wade endorses Dean Keeton's narrow "waiver" interpretation.) 260 Prosser, id. at 404. 261 W. Keeton, id. at 405. 262 Breitel, id. at 407. 254 255 263 Id. 264 Id. at 408-09. HeinOnline -- 67 B.U. L. Rev. 283 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 had not substantially changed the message of Restatement (First) § 893, which Professor Warren Seavey had belatedly drafted to give some recognition to assumption of risk.2 65 Prosser's claim of fealty is dubious. Section 893 spoke only of voluntarily encountering a risk; it did not also require an "acquiescence" in or "willingness to accept" a risk. Professor Robert Keeton then criticizes the Confederates for treating assumption of risk as a no duty rule. 266 Eldredge's response repeats the point that the valid form of assumption of risk is consent, a willingness to encounter the risk. 2 7 And he insists that his conception is narrower than Prosser's, for it does not include certain reasonable choices by a plaintiff to encounter known risks, such as a plaintiff's decision in an emergency to take a ride with a drunk. 26 But it is not clear why we should not categorize this case as plaintiff's "willingness to encounter a risk," on Eldredge's own theory. Not surprisingly, one participant at this point declares, "I seem to be in very near agreement with both sides. ' 269 Eldredge does not advance matters by conceding that Restatement (First) § 893 is perfectly acceptable. 270 That concession is perplexing inasmuch as § 893 is a classic statement of the broad form of voluntary assumption of risk. Presumably both Prosser and Eldredge claim to be perpetuating that section in order to demonstrate continuity with the past and consistency with the views of Seavey, the respected drafter. In a subsequent exchange, Eldredge and Prosser restate the difference between them as whether assumption of risk ever applies when defendant has breached a duty to plaintiff, but plaintiff is not negligent. 271 They do not seem to realize, however, that the scope of assumption of risk as consent is critical to this issue. Eldredge claims that consent indicates no breach of duty, no wrongdoing by defendant. Prosser claims that assumption of risk is a defense based on consent by conduct. If they employ the same concept of consent, then the difference is at best about burden of proof-consent as no breach of duty, or as an affirmative defense. After a vote is taken demonstrating general support for Prosser's approach,72 a discussion of the "express" assumption of risk provision (§ 496B) begins. Dean Page Keeton repeats his point that in terms of the form of the agreement, the section should be broadened to include conduct as well as words, but in terms of the content of the agreement, it should 273 apply only when plaintiff intends to relieve defendant of all responsibility. 265 Prosser, id. at 410. R. Keeton, id. at 415. Prosser, id. at 415-16. 2681 Id. at 417-18. 269 Shewmaker, id. at 422. 270 Eldredge, id. at 425. 271 Prosser, Eldredge, id. at 425-26. 272 Id. at 426. 273 W. Keeton, id. at 432-33, 436-37. 266 267 HeinOnline -- 67 B.U. L. Rev. 284 1987 1987] ASSUMPTION OF RISK (Justice Braucher expresses what seems to be a different understanding: the "express" § 496B refers to mutual agreements, while the "implied" § 496C refers to unilateral agreements.2 74 However, Keeton says that he accepts Braucher's formulation. 2 75 Perhaps Keeton assumes that Braucher shares his narrow view of consent as waiver.) In reply, Prosser unhelpfully asserts that the difference is "between that which is express and that which is implied.12 76 But he later explains that § 496B is consent by words, usually given in advance, while § 496C is consent by conduct, usually expressed after the defendant has acted.2 77 Why then, asks Eldredge, not combine the sections, since each is just a way of manifesting willingness?2 78 Because, replies Prosser, words are for ajudge to 279 interpret, while conduct is for a jury. The debate turns to the central issues-the meaning of "manifesting acquiescence" in § 496C and its relation to voluntarily encountering a known risk. Prosser emphasizes that voluntarily encountering the risk does not suffice, if plaintiff does not "manifest his acquiescence in the risk and the situation." 2 0 Dashing into speeding traffic is not an example of assumption of risk, but accepting a ride with a drunk is. 281 Prosser gives little explanation why these examples are not treated the same. Indeed, he later suggests that if plaintiff could wait a minute or two before crossing, he assumes the risk by proceeding! 282 Prosser soon catches himself, 283 but it is troubling that he is so uncertain about his own example. The problem, I believe, is the vagueness of "acquiescence" and the uncertainty about what one must accept or acquiescence in. Prosser addresses the first part of the problem by accepting Eldredge's suggestion to change "manifests his acquiescence" to "manifests his willingness to accept" the risk.28 4 But what does "willing" mean in this context? Although Prosser insists that not every voluntary encountering of a known risk expresses "willingness," he utterly fails to explain how we are to decide whether it does. In later discussion, one A.L.I. member points out that the language of § 496C will probably be interpreted as ordinarily allowing an inference of "willingness" from "voluntary encountering.- 2 8 5 Prosser seems amenable 274 275 276 277 278 279 280 211 282 283 284 285 Braucher, id. at 434. W. Keeton, id. Prosser, id. at 435. Id. at 435, 438. Eldredge, id. at 439. Id. Prosser, id. at 445, 448. Id. at 445-46. Id. at 461-62. Id. at 462. Id. at 446. Shewmaker, id. at 449. HeinOnline -- 67 B.U. L. Rev. 285 1987 BOSTON UNIVERSITY LAW REVIEW [Vol. 67: 213 to a change explicitly providing that voluntarily encountering a risk "is usually evidence of such willingness.''26 Dean Lockhart then suggests a ' ' 28 7 redrafting to give less evidentiary weight to "voluntarily encountering. Although there is no discussion, in fact Prosser's final draft incorporates Lockhart's phrase, "under circumstances that manifest his willingness to accept" the risk. Prosser's ambivalent, or indifferent, attitude bespeaks an inability either to define "willingness" clearly or to clarify its relationship to "voluntary encountering." Significantly, no one responds to Judge Dooling's objection that the two concepts are completely independent and have no evidentiary bearing on each other.288 The most plausible explanation for the silence, apart from the participants' fatigue, is that "willingness" is so obscure a concept that Dooling's objection cannot be fairly evaluated. Prosser is clear about one point. He disagrees with Dean Page Keeton's narrow interpretation of consent as waiver of a tort claim against defendant. When Keeton points out an official comment suggesting that plaintiff's conduct must "indicate that the plaintiff is willing to take his chances, and relieve the defendant of responsibility," 28 9 Prosser explains that he means only that willingness indicates consent, and "does in fact thereby relieve defendant of responsibility. ' 29 0 In the final version of comment h, the portion of the quoted phrase after the comma is omitted. Several lessons can be drawn from this review of the debate. First, some of the principal antagonists are not so antagonistic, after all. They share both a desire to relieve defendants of responsibility when plaintiffs "consent" and an inability to clarify that critical term. To be sure, Page Keeton consistently takes the extreme Confederate position that only consent to free defendant from liability suffices. But Eldredge and Wade, also Confederates, seem to have a much looser interpretation. At the same time, although Prosser apparently begins his efforts by endorsing the traditional broad view that a plaintiff is barred whenever he voluntarily encounters a known risk, ultimately he accepts the requirement that plaintiff "acquiesce" in or "willingly" accept the risk. Second, insofar as the antagonists do differ, and they certainly believe that they do, the difference largely concerns the proper evidentiary weight of plaintiff's voluntary encounter of a known risk. Yet it is impossible for the two sides to justify the difference in weight, since each purports to be proving the same type of consent. Third-and this is most speculative-perhaps the full preference theory helps explain the different positions. Aside from Page Keeton, the other Confederates seem, from most of their examples and their opposition to 2816Bunn, id. 287 Lockhart, 218 at 455. id. Dooling, id. at 455-56. W. Keeton, id. at 448 (noting comment h to § 496C). Prosser, id. at 448-49; see also id. at 456-57 (colloquy between Eldredge and Prosser). 219 290 HeinOnline -- 67 B.U. L. Rev. 286 1987 1987] ASSUMPTION OF RISK 287 Prosser's draft, to believe that a plaintiff should be barred only when he fully prefers the risk. (Recall the examples of the rattlesnakes or the reducedprice defective chattel.) Prosser's view is less consistent with my theory. In any event, both sides seem to be groping toward an acceptable definition of consent to a risk. The full preference theory does provide a relatively clear intermediate definition between waiver and the traditional broad view. HeinOnline -- 67 B.U. L. Rev. 287 1987 HeinOnline -- 67 B.U. L. Rev. 288 1987