by John M. Adler

Copyright (c) 1991 Wisconsin Law Review.
University of Wisconsin

September, 1991 / October, 1991

1991 Wis. L. Rev. 867

LENGTH: 42227 words



* Assistant Professor, University of San Francisco School of Law. A.B., 1968, Harvard College; J.D., 1976, University of New Mexico School of Law. I wish to thank my research assistant, Jean Afterman, for her hard work and good spirits. I also wish to thank my colleagues, Professor Jeffrey Brand, Trina Grillo, Joshua Rosenberg and Stephanie Wildman not only for reading and commenting on earlier drafts, but also for their encouragement and support.

  ... The decision can be made adequately whether the defendant's behavior is characterized as misfeasance or nonfeasance, and whether or not the parties share a "special relationship," as long as a court explicitly analyzes the same kinds of factors that might in other situations prevent it from imposing a duty to behave reasonably. ... The babysitter's legal obligation stems from her special relationship with the child. ... PROBLEMS WITH THE SPECIAL RELATIONSHIP EXCEPTION ... There are two significant problems in relying upon special relationship analysis. First, in many situations it is difficult to determine what constitutes a special relationship. ... An example may help to illustrate the problems with the special relationship approach. ... The Tarasoff court found that the therapist had a duty to the victim, in part because of his special relationship to the assailant. ... Citing Tarasoff, the court stated that "[u]nder traditional tort law principles, one is ordinarily not liable for the actions of another and is under no duty to protect another from harm in the absence of a special relationship of custody or control." ... Plaintiff's argument that a special relationship existed between the police and the suspect was rejected. ... The court distinguished earlier cases in which a special relationship existed between the defendant and the plaintiff-victim. ... For example, the Michigan Supreme Court expands the notion of a special relationship to include social companions. ...  

In this Article, Professor Adler addresses a question that has confounded legal scholars for decades: Why is there no legal obligation to rescue another from impending peril when such protection could be rendered with minimal risk to, or effort by, the would-be rescuer? He begins by describing the traditional no-duty-to-rescue rules, and continues by explaining why they should be discarded. He observes that many courts have applied these rules in a cursory fashion, distorting them when necessary to reach decisions grounded in traditional tort law theory. He concludes by proposing that courts merely impose upon defendants an obligation to act reasonably under the circumstances -- the same duty prescribed in any negligence action. Consequently, his proposal is both radical, in that it suggests a significant restructuring of legal doctrine, and tradition-bound, in that it requires no new standards of care or procedural mechanisms.

 [*867]  For more than eighty years, commentators have argued about whether courts should require one to act affirmatively to protect a stranger in peril. n1 From the turn of the century to the present, legal scholars have continued to ask why a bystander who watches a toddler step off a sidewalk into the path of a careening truck has no legal obligation to protect the child, even if there would be no significant "cost" involved in reaching out to restrain the infant. n2

Most proposals for change in the law have urged that the bystander should have an obligation to protect the child from serious harm, at least when protection could be offered without serious inconvenience  [*868]  or risk. n3 But in spite of early and repeated calls for reform, no recorded case has expressly adopted the requirement -- which at first blush would appear to be relatively harmless -- that people have a responsibility to engage in even an "easy rescue." n4

Although examples like the one above dramatically pose issues about the relationship of law and morality, n5 and therefore are favorites of commentators and first year law professors, in reality, missed opportunities for easy rescue are rarely the stuff of reported opinions. n6 Most opinions considering failures to aid or protect individuals in peril involve public institutions, businesses, or professionals who have failed to protect an injured individual from the person who most "directly" caused the harm. n7 This Article suggests replacing the traditional rules with an alternative that will be applicable to the full range of cases.

 [*869]  Part I of the Article describes the traditional no-duty-to-rescue rule and the judicially created exceptions to it. n8 Most significantly, a "passive" defendant may be liable for a failure to rescue as a result of certain relationships between the defendant and either the victim or the person who has injured the victim. n9

Part II explains why the traditional no-duty rules are inadequate and should be abandoned. Courts that are intent on accurately applying these rules at times necessarily will fail to focus on, and will even frustrate, the aims of the tort system. n10 As a result, in spite of the seeming immutability of the no-duty-to-rescue rules, n11 the law has not been static in this area. Although they may pay lip service to these traditional common law rules, n12 courts often distort the rules (intentionally or otherwise) to reach decisions more directly, and honestly, justified by the same policies that inform their decisions in "ordinary" negligence cases. n13 Where courts do not distort the rules or abandon them outright, they risk nonsensical results. n14

 [*870]  Part III proposes that in lieu of the traditional approach, courts should impose upon a defendant nothing more nor less than an obligation to act reasonably under the circumstances unless, on balance, there are recognized policy concerns that militate against the imposition of that duty. The decision can be made adequately whether the defendant's behavior is characterized as misfeasance or nonfeasance, and whether or not the parties share a "special relationship," as long as a court explicitly analyzes the same kinds of factors that might in other situations prevent it from imposing a duty to behave reasonably. n15

The approach would greatly simplify this area of the law by sweeping away a web of complicated rules that depend upon unworkable factual distinctions and consistently fail to further any recognized aims of the tort system. The proposed approach would obviate the need for judicial hair-splitting over whether behavior is "misfeasance" or "nonfeasance," and, if it is characterized as nonfeasance, whether it falls within one of the exceptions to the traditional no-duty rule.

The proposed change is at once radical and tradition-bound. It is radical in the sense that it is not a proposal for a minor adjustment in the law. Nor is it a proposal that would significantly alter analysis only in a handful of cases. At the same time, it is both simple and traditional. It requires no new standards of care, no new procedural mechanisms. Rather, it urges only that courts considering failures to aid or protect use the standards and procedures long accepted in negligence cases involving active misconduct. Thus, failures to aid or protect would be analyzed in the same manner as cases involving "active" negligence. Unreasonable behavior -- however characterized -- could result in liability, unless such a result were precluded by a recognized public policy.

Such a change is made necessary by the same forces that led the California Supreme Court to abandon an outmoded approach to landholder obligations. The court in Rowland v. Christian n16 recognized that the traditional rules were not only confusing and complex, but they were no longer based on "proper considerations." n17 The court replaced an approach that depended upon overly technical status distinctions (trespasser, licensee, invitee) with an ordinary negligence approach. Now, nearly twenty five years after Rowland, it is time to consider making the same clean break from an established but flawed approach where affirmative duties are concerned.

 [*871]  Part III also considers anticipated objections to a policy-based approach. Interestingly, in response to calls for more limited reform than is proposed herein, a number of recent scholarly defenses of the no-duty-to-rescue rules have raised serious philosophical and practical questions about the wisdom of any approach that would legally require the protection of the infant in the example above. n18

Some argue that to impose a duty to aid or protect on an "innocent" bystander would be unfair and would violate fundamental notions of individual liberty. n19 Others are concerned that an expanded duty would be unmanageable and would result in a lack of certainty for courts and litigants alike. n20 The truth is that the traditional rules are no less vulnerable to these complaints than policy-based decisions would be. The difference is only that the traditional rules often fail to focus adequately upon the factors that are at the heart of the decisions they support.

The proposal made in this Article has an important secondary benefit. To the extent courts carefully articulate the considerations that actually underlie their decisions, the workings of personal injury law itself will be brought into focus. Both judges and those who study their opinions should be led to another level of inquiry about the knotty moral, philosophical, and legal problems that are so troubling in this area of the law. These inquiries may provide indications as to whether tort law can successfully fulfill the competing, and at times conflicting, aims which seem to be central to these cases. n21


As any first year law student can recite, liability in tort requires that the defendant have a duty to the plaintiff and that the plaintiff's harm result from a breach of that duty. n22 In cases involving a defendant's failure to protect another from a danger not of the defendant's making, courts traditionally have concluded that no legal duty exists to come to the aid of a stranger. n23

The common law's reluctance to require one to render aid to a stranger rests upon the distinction between misfeasance and nonfeasance. n24 Whatever practical difficulties a court may encounter in distinguishing nonfeasance from misfeasance, n25 once made, the distinction is more than academic. n26 If a court characterizes the defendant's behavior  [*873]  as nonfeasance, absent one of the exceptions discussed below, the defendant ordinarily will owe no duty to the plaintiff; the case may be dismissed. n27 As a result, unreasonable behavior may be immune from liability as long as a defendant can successfully characterize it as nonfeasance. n28

Thus, for example, where injury results from a water company's failure to provide water to a hydrant, a court may characterize the company's behavior as misfeasance, that is, improper maintenance of water pressure that created a risk of harm for the plaintiff. n29 In that instance, the defendant is unlikely to have the case dismissed before trial unless, for other reasons, the court concludes that no duty exists. n30 Thus, ordinarily, the jury will be asked to determine whether the behavior of the defendant was reasonable under the circumstances.

On the other hand, the court may take the case from a jury by characterizing the company's behavior as nonfeasance: a failure to provide a benefit. Because thelaw is loathe to require that one "go beyond the call of duty," the plaintiff's case will fail unless it can be shown that an exception to the no-duty rule should apply. n31 In that case, even behavior that is clearly unreasonable will not reach a jury.

There are limited exceptions to the general rule that nonfeasance results in no duty. In such cases, liability may be imposed even where the defendant's passive behavior would not have been otherwise actionable. n32 To  [*874]  illustrate, imagine several adults standing near a child about to step into the path of a truck. Each could easily protect the child, without inconvenience or significant risk. n33 One of the adults is the child's babysitter who has done everything reasonably possible prior to the appearance of the truck to keep the child from harm's way, but for whatever reason, fails to restrain the child as the child steps into the path of the truck. The second adult, who also fails to restrain the child, realizes she is partially responsible for the truck's lack of control, having hit the truck with her car, through no fault of her own, only an instant before. A third adult leaps in front of the child to wave off the truck. However, seeing that the truck is out of control, she jumps back to protect herself and fails to take reasonable steps to restrain the child. Finally, the fourt passive adult (the one in the best position to restrain the child) is a "pure" bystander.

The traditional approach would result in the dismissal of the case against the "pure" bystander. The others could be required to face a jury to defend the reasonableness of their actions. For each, the law may impose a duty to protect. n34 The babysitter's legal obligation stems from her special relationship with the child. n35 The second adult may  [*875]  have a duty if she recognizes that in some sense she created the risk. n36 The individual who waves the truck away may be legally obliged to continue to try to prevent the harm if the wave can be characterized as the taking charge of the helpless child's situation. n37

 [*876]  There is an additional exception to the general no-duty rule where a third person (rather than a dangerous condition) is posing the threat. In such a case a duty will exist even where the victim and the potential rescuer are total strangers, and even if the potential rescuer has neither created the risk nor undertaken a rescue in any way, as long as the potential rescuer shares a relationship with the individuals who pose the danger. Significantly, when the duty to act affirmatively is based upon a relationship between the potential aider and the dangerous person, the duty is expressed not as a duty to protect or aid the victim, but as a duty to control the dangerous individual. n38

For example, assume that several children stretch a wire across the road in order to injure an unsuspecting bicyclist in their neighborhood. A bystander (not acquainted with the children or their bicycling victim) who could prevent the serious injury simply by shouting a warning has no legal obligation to do so. n39 However, if one of the children who is devising the trap tells her therapist her plans to injure the victim, the therapist will have a duty to protect the bicyclist from the client, even though meeting the obligation may not be easy and may involve significant costs to the therapeutic relationship. n40

 [*877]  An increasing volume of cases turn on these rules in an ever-widening range of factual settings. n41 As a number of commentators have recognized, some courts have at times found a legal duty in situations where there would have been no duty found under the early common law. n42 But the exceptions have not, as some have suggested, swallowed the rule. n43 In general, courts have not eliminated limitations upon the duty to aid or protect; but the basis for the limitations is shifting. Simply put, the traditional rules fail to adequately account for the factors that guide many courts' decisions. Courts are struggling with difficult issues that are not adequately informed by the traditional rules.


The existing theoretical framework for handling behavior characterized as nonfeasance, involving as it does, the application of exceptions based upon "special relationships" or risk-creating acts or "undertakings to rescue" is impossible to apply consistently. Opportunities for mischief exist at every step in the analysis. Courts certainly could tolerate these line-drawing problems, however, if the existing rules generally furthered the aims of tort law. n44

The more fundamental problem with the no-duty-to-rescue framework, however, is substantive. All of the exceptions to the rule are made necessary by a distinction between misfeasance and nonfeasance that is itself fundamentally misguided. In many situations, it is difficult to see how the distinction is more than semantic play. But even assuming  [*878]  that a difference between misfeasance and nonfeasance can be defined (or at least that a difference will continue to be recognized), that difference is given far too much significance within the traditional legal framework. n45 As a result, the rules and exceptions based on the distinction often fail to focus properly on any pronounced or imaginable aims of tort law. Whether one views the goals of tort law to be the deterrence of unreasonably risky conduct, n46 the compensation of victims through loss distribution, n47 the imposition of values of fairness, n48 or some combination of these goals, the focus of the common law rule based upon the misfeasance-nonfeasance distinction will often undercut those aims.

A. Problems With the Underlying Principle: Misfeasance v. Nonfeasance

Two fundamental problems plague the distinction between misfeasance and nonfeasance: (1) in many cases it is impossible to distinguish the two; and, (2) in cases where intuitively there is a clear distinction, that distinction does not always coincide with generally accepted notions about whether liability should attach.

As noted, the traditional analytical structure is based upon a preliminary determination about whether the defendant's behavior constitutes  [*879]  misfeasance. As a general rule, it would be wrong, the traditional argument runs, to hold a defendant liable for acting negligently if she did not act at all. Despite the apparent appeal of this logic, it is flawed. Even where inaction and action can be distinguished, one defendant's "failure" to act may well be more reprehensible than another's affirmative misconduct.

In 1908, Professor Bohlen suggested that misfeasance differed from nonfeasance in two ways -- the behavior in question was different, and the effect on the plaintiff was different. n49 To analyze a case as nonfeasance, one would have to find both: (1) an omission rather than active misconduct; and, (2) behavior that did not worsen the position of the plaintiff. In 1928, Justice Cardozo provided an example of the already "time honored" distinction between misfeasance and nonfeasance in H.R. Moch Co. v. Rensselaer Water Co. n50 In Moch, the city's water supplier failed to provide enough water to fire hydrants. As a result, firefighters couldn't prevent plaintiff's property from burning. The court held that the water supplier owed no common law duty in tort to the property owner, explaining that the water company's failure was nonfeasance rather than misfeasance. n51

In Moch, Cardozo gave several examples of actionable omissions -- of what has more recently been called "pseudo-nonfeasance." n52 For example, a surgeon would be liable for harm resulting from the omission of failing to sterilize his instruments n53 and an auto manufacturer could be liable for the omission of failing to adequately inspect its product. n54

Cardozo suggested that the distinction between actionable and inactionable omissions rested not so much upon the nature of the omission, but upon the relationship that existed between the plaintiff and  [*880]  the party who failed to act. Presumably, a bystander who could see that the surgeon had failed to sterilize the instruments would have no duty to shout out a warning or otherwise protect the doctor's patient from possible infection. n55 Just as the surgeon fails to protect against infection at the moment she fails to sterilize the instruments, so does the person watching the procedure fail to protect when he fails to call out a warning. The omission of the doctor and the omission of the bystander are both, in traditional legal terms, "but for" causes of the infection. In Cardozo's terms, the bystander's omission would not be actionable because it was not enough to create a relation. The bystander's omission may have denied a benefit to the plaintiff, but, unlike the omission of the doctor, it did not actively injure the victim. n56

Recently, scholars have attempted to provide a more satisfying rationale for continuing to distinguish between misfeasance and nonfeasance. Professor Weinrib, whose article contributed enormously to a greater understanding of this area of tort law, has suggested that the difference between the omissions of the surgeon and the bystander lies in an analysis of the risk created by the omission. n57 Weinrib would agree with Cardozo that the bystander's behavior is truly nonfeasance. After all, Professor Weinrib would argue, the bystander's behavior did not create the risk of infection. Rather, by observing the incident passively, the bystander simply allows the physician-created risk to materialize. That risk existed (by virtue of the surgeon's failure to act) whether or not the bystander had been viewing the procedure. In contrast, the surgeon's omission in fact created the risk. n58

 [*881]  Ultimately, however, such a distinction is no more satisfying or productive than making a distinction between misfeasance and nonfeasance or between behavior that merely denies a benefit and that which actually harms. n59 However defined, the distinction invites decision-making based on some fragile factual characterizations. n60

Consider a situation where an employer requires an employee to work long hours and then drive himself many miles home. In an action against the employer for an injury to an innocent pedestrian caused by the employee falling asleep at the wheel, a court might find nonfeasance on the part of the employer and dismiss the case based on the common law rule that generally the employer-employee relationship does not create a duty to protect others from the employee's off-premises conduct. n61 On  [*882]  the other hand, a court might just as easily conclude that the employer's demands upon its employee actually created a foreseeable risk, n62 and analyze the case as a garden variety negligence action  [*883]  that would not depend on the finding of a special relationship between employer and employee.

Cases involving intoxicated drivers provide another excellent example of line drawing problems frequently faced by courts. If a tavern owner can, with impunity, let an inebriated patron drive off into a blizzard or hand a drunk patron his car keys, then should it make a difference that an employee of the tavern helps the patron to his car? n63 Or that a filling station attendant gives gas to an intoxicated customer? n64

The distinction between these acts is difficult to make on any principled basis. It is difficult to say why the person who helps the inebriated customer to his car has an affirmative duty to protect, if the person who has handed him his keys has not. Assuming that the act of comfort did not create or increase a risk by preventing others from aiding the victim, it is difficult to understand why the minimal act should result in a finding of a duty in one case and not the other.

But even assuming that these difficulties can be overcome and that a court could distinguish rationally between misfeasance and nonfeasance, the question remains, is that distinction standing alone significant enough to determine whether a case proceeds to trial? The limitations of an analysis founded upon the misfeasance-nonfeasance distinction can best be illustrated by an example borrowed from the world of medical ethics. n65 Assume that two doctors have similarly situated patients. Both are terminally ill, mentally competent and in considerable pain. Both wish to die. One doctor decides to administer an injection that quickly ends her patient's life. The other decides not to treat a minor respiratory problem, knowing that, although it will take some time, if the respiratory problem goes untreated, it will kill the patient.

 [*884]  One way to analyze the behavior of the two doctors is to consider whether their conduct was active or passive. Traditionally, the doctor who injected her patient would be viewed as unethical because her "active" behavior caused a death earlier than it would have otherwise come. The second physician would be viewed as having behaved ethically -- having done nothing more than let nature take its course (even though her decision not to treat did bring about death earlier than was necessary). n66

But to determine ethical conduct based on such a distinction avoids the underlying policy question -- should physicians speed the death of patients in this situation? If we believe that they should, then both have behaved ethically (even if one has been active and one s passive). Indeed, if we resolve the ethical dilemma in this way, the doctor who administered the injection has acted more humanely than the one who did not. She has complied with her patient's wishes and has helped her patient to die painlessly. The second doctor has also behaved ethically but has required her patient to suffer in the process. n67 Alternatively, if it is decided that it is not ethical for a doctor to accelerate the dying process, then irrespective of whether their conduct is characterized as active or passive, both doctors have acted unethically. In other words, when we focus on resolving the fundamental policy issue, the significance of the distinction between passive and active conduct recedes. n68

This limitation is easily illustrated in the civil litigation context as well. Absent affirmative conduct, some courts have limited the duty of landholders to protect one who comes on the land (even an invitee) from another dangerous individual on the property. n69 Where the landholder has tried to provide some sort of security, however, courts following the common law approach have at times based a duty on the existence of that affirmative act. n70

Assume that there are two competing restaurants in the same neighborhood. Both are aware of crime statistics in the neighborhood and can foresee that at some point there will be violence in their parking lots. One establishment chooses to ignore the threat; the other hires a security guard. Ultimately, customers are hurt and bring separate actions against the restaurants. The restaurant that ignored the threat successfully moves to have the case against it dismissed. On the other hand, the restaurant that hired a security guard is found to have a duty based upon its undertaking to protect its customers and faces a jury determination about the reasonableness of its conduct.

 [*885]  It is not only those who feel that tort law should be guided by considerations of fairness who should be disappointed with the law's willingness to countenance the most unreasonable behavior as long as it is described as the "mere denial of a benefit." So too, should those who believe safety considerations should guide the law, and in many situations, those who feel that the distribution of losses is a primary aim.

The results clearly operate as a disincentive to safety. Given the existing rules, new restaurants in the area would have no incentives to hire security guards. In fact, it is hard to imagine establishments rationally taking the safer course. n71 To the extent loss distribution is a viable goal of the tort system, neither is that goal adequately served by rules which result in a judgment against one restaurant and not the other. If the aim is to reduce the hardship to an individual plaintiff by distributing the costs of compensation to defendant's customers as a cost of doing business, the loss should be distributed in both situations. In either case, the restaurant is clearly in a better position to diffuse the burden than is its hapless customer. Either restaurant can pass on the cost of injury (and additional security measures) to the public. In that sense, the cases are indistinguishable, yet traditional rules may lead to widely different outcomes. And if notions of reasonable conduct n72 are considered primary, once again, the common law approach frustrates the goal by failing to focus sufficiently upon it. From what we know, the party most at fault may be protected by the judicial system while the party that acted more reasonably is punished for making the effort. n73

 [*886]  This rather shaky nonfeasance-misfeasance foundation supports several major exceptions to the no-duty rule, each with weaknesses of its own.

B. Problems With the Major Exceptions to the No-Duty Rule.

Because the misfeasance-nonfeasance distinction often fails to reflect social policy or to parallel any intuitive distinction between behavior that ought to be actionable and behavior that ought not, courts have generated substantial exceptions to the rule that nonfeasance means non-liability. This part of the Article explores the difficulties with these exceptions.


If a duty is imposed in spite of a finding of nonfeasance, most frequently it is because a special relationship has been identified. There are two significant problems in relying upon special relationship analysis. First, in many situations it is difficult to determine what constitutes a special relationship. Second, the presence or absence of a relationship does not always determine whether a duty will be imposed or not. For these reasons, special relationship analysis is of little predictive value and may not even accurately describe the concerns that determine the outcome of a significant number of cases. n74 In 1908, Bohlen urged that we examine these relationships "and the attributes which such relations possess in common, and so, if possible, discover their underlying basis." n75 But we may be further from a shared understanding about what constitutes a special relationship today than we were in Bohlen's time.

Initially, the special relationship exceptions to the no-duty-to-rescue rule required one to protect another in limited situations where either the potential victim was sufficiently dependent upon the potential protector or the potential protector was sufficiently "in charge" of the individual posing the threat. n76 The difficulties with applying special relationship analysis are illustrated by those cases requiring one to control a threatening individual on behalf of a total stranger. n77

 [*887]  Initially, the cases simply held certain defendants liable to control people in their charge. Parents were responsible at times for protecting others from their children, employers were at times responsible for protecting others from their employees, property owners were at times responsible for protecting others from people who had permission to use the property, and those in charge of dangerous individuals were at times responsible for protecting others from their charges. n78

Some courts have expanded the sorts of relationships that would require one to control another for the benefit of a third person. Mental health professionals have been held responsible for failing to exercise reasonable care to protect others from dangerous patients. n79 Youth authorities and others in charge of dangerous persons have been required to take reasonable precautions to protect others, even when the dangerous person is not in custody. n80 Employers have been required to prevent employees from injuring others, even away from the workplace. n81 In spite of such developments, many cases have analyzed very similar relationships and have refused to find them sufficient. n82

In the 1930's, Harper and Kime found the ability to control another's conduct to be central to a number of situations where courts  [*888]  had imposed liability for nonfeasance that resulted in harm to strangers. n83 In fact, exceptions to the no-duty rule resulted in only a limited duty -- a duty to control. n84 Thus, for example, the possessor of a vehicle could be held liable for injuries to a stranger caused by another person allowed to drive in the car owner's presence. Liability was imposed not because of the negligence of the driver, but because of the owner's failure to exercise adequate control over the person to whom he had entrusted the auto. n85

However, the ability to control the person who causes the harm is not, in and of itself, sufficient to determine that a duty exists to act affirmatively to exercise that control. If it were enough, the duty would extend to any individual physically able to intervene effectively on behalf of another. n86 Instead, the Restatement and developing case law took the approach that a duty will be imposed where a person can exercise an ability to control and there is some additional ingredient to the relationship. The problem is that what the additional ingredient is remains unclear. Moreover, although some courts have used the absence of an actual ability to control to prevent the imposition of liability, n87 there are situations where a special relationship is found and  [*889]  a duty imposed even in the absence of any ability to control. n88 Thus the traditional rules fail to describe what is guiding the courts in these cases. n89

An example may help to illustrate the problems with the special relationship approach. n90 Assume that a family considers a suit against a church counselor, the campus police and a psychologist at a university's counselling service for failing to protect their child from an assailant. Assume further that the church counselor and the therapist both met with the assailant a number of times and could have warned the family of the threat. Both asked the campus police to detain the assailant because of his violent threats towards the plaintiff's child. The campus police detained and then released him.

Against whom may the plaintiffs proceed if, some time after his encounters with the potential defendants, the assailant does attack plaintiffs' child? What determines which, if any, potential defendant was legally obligated to warn the family? Arguably none has created the danger. Presumably, had none of them been involved, the assailant would have acted no differently. n91 Assuming that a special relationship with the assailant is sufficient to create an obligation to protect in spite of their failure to act, is there any basis for treating the defendants differently?

The example poses issues raised directly or indirectly by several cases decided by the California Supreme Court. n92 In Tarasoff v. Regents  [*890]  of University of Cal., n93 the court held that the plaintiffs could state a claim against a therapist employed by the University for failing to protect their daughter -- an "identifiable" stranger -- from an assailant who had threatened her life. At the same time, the court affirmed the trial court judgment sustaining the demurrer of the campus police without leave to amend. n94

The assailant had sought counseling at the student health facility at the University as a voluntary outpatient. In late August, 1969, the psychologist who saw him in outpatient psychotherapy determined that he was dangerous. As provided for by California law, the therapist notified campus police that the patient should be detained for evaluation. n95 The police interviewed the patient and released him. After his encounter with the police, the patient never returned to therapy. Several months later, when plaintiffs' daughter returned from travel outside the country, the patient killed her. n96

The Tarasoff court found that the therapist had a duty to the victim, in part n97 because of his special relationship to the assailant. Relying upon Restatement §§ 315-20, the court concluded that the therapeutic relationship with the assailant was sufficiently "special" to justify the creation of a new exception to the general no-duty rule in nonfeasance situations. n98 In support of its conclusion that the outpatient  [*891]  therapy created a relationship between the therapist and the assailant that would create a duty to the victim, the court skirted the issue of control; n99 the only authority cited for the proposition that a therapist owes a duty to protect a victim of a mentally ill person involved a hospitalized patient who was released periodically to work on a nearby farm. n100

The analogy is surprising. In general terms, a primary goal of outpatient psychotherapy is to encourage autonomous behavior by the patient and to discourage control by the therapist. n101 The facts of this particular case suggest that the therapist's actual ability to control the patient was in fact quite limited. The patient was not on a locked psychiatric ward. He was voluntarily obtaining counseling offered by the University to its students. Indeed, when he felt that the benefits of treatment were outweighed by the costs, he simply stopped visiting the center. The psychologist could not compel the patient's attendance or order his long term hospitalization. The therapist did have the ability to have the student detained for a brief observation. When he took that step, the patient discontinued treatment. At that point, the therapist lost whatever actual control he may have had. n102

 [*892]  Whether the decision was supportable based on established rules or not, n103 one would expect that if a sufficient relationship were found, a similar result should obtain for the church counselors in our example. However, recent case law casts doubt upon that conclusion. Twelve years after Tarasoff, parents whose son killed himself sued a church and church pastors for failure to counsel their son adequately prior to his death. n104 The son had sought counseling from the church which employed about fifty counselors, some of whom provided sessions resembling therapy. n105 Relying in part upon the Restatement approach, the California Supreme Court in Nally v. Grace Community Church of the Valley refused to recognize a duty on the part of the counselors. Citing Tarasoff, the court stated that "[u]nder traditional tort law principles, one is ordinarily not liable for the actions of another and is under no duty to protect another from harm in the absence of a special relationship of custody or control." n106

The court refused to "extend" the duty to "personal or religious counseling relationships in which one person provided nonprofessional guidance to another seeking advice and the counselor had no control over the environment of the individual being counseled." n107 Despite the distinctions drawn by the court, the supportive nature of the relationships in Nally and Tarasoff was virtually identical. n108 Nor does  [*893]  the ability of one party to control "the environment" of another seem significantly greater in Tarasoff than in Nally. n109 That a duty was found in Tarasoff and not in Nally cannot be explained by a difference in control available to the therapist or the counselor. n110

And what of the campus police in our example? The court in Tarasoff concluded that the campus police could have had no relationship with the patient or the victim sufficient to require that they protect the victim. The court did not choose to explain its conclusion, but a rationale can be gleaned from subsequent opinions.

Six years after the court's decision in Tarasoff, the victim of a violent assault brought a civil action alleging that defendant police had failed to exercise an ability to warn the victim of an impending attack. n111 The plaintiff, Yolanda Davidson, had been stabbed in a laundromat. The complaint stated that the police had the laundromat under surveillance. Allegedly, the police had concluded that a man they were watching was someone who had stabbed another woman at the same laundromat the previous evening. They watched the suspect go into and come out of the laundromat for about fifteen minutes without warning the plaintiff that she was in any danger. n112

 [*894]  Plaintiff's argument that a special relationship existed between the police and the suspect was rejected. n113 As in Tarasoff, the victim was not known to the defendants. n114 Although the police in Davidson had substantial certainty about the threat, could easily identify the potential victim, were physically present and capable of controlling the situation -- if not by making an arrest, n115 then certainly by warning the potential victim -- the court imposed no legal obligation upon them to do so. n116 The court pointed out that "a person's mere proximity to an assailant, even with a knowledge of his assaultive tendencies . . . does not establish a . . . [duty]." n117 Moreover, neither the fact that the defendants were peace officers nor the fact that they had had previous encounters with the assailant who had eluded them the previous evening after stabbing someone in the same locale were enough to create a special relationship. Once again, something other than the nature of the relationship would seem to be determinative. n118

 [*895]  One other post-Tarasoff opinion makes it absolutely clear that although the court faithfully recites special relationship rules, the decisions are actually based on other considerations. Suppose that the therapist in the example above had been employed not by the University, but at a county facility that released an assailant into the community and failed to protect a foreseeable victim. In Thompson v. County of Alameda, n119 plaintiffs sued the county for negligently failing to warn of the temporary release of a juvenile offender who made threats against young children in his neighborhood. n120 Within twenty-four hours of being released to the custody of his mother, the juvenile killed plaintiffs' five year-old son. The court held that the county's decision to release the juvenile into the community was immunized by a California statute, n121 but that no statute provided immunity for the county's failure to warn local police, the juvenile's mother, or neighborhood parents about the juvenile's expressed threat. Even absent an immunity, however, the county had no duty and consequently no liability. n122

In reaching its decision, the court looked only to the absence of a relationship between the county and the victim and failed even to consider the relationship between the county and the killer. The court distinguished earlier cases in which a special relationship existed between the defendant and the plaintiff-victim. n123 Without ever concluding that a special relationship between the county and the killer did not exist, the court distinguished Tarasoff which it read to require only that a warning be given to a "readily identifiable" intended victim. n124

 [*896]  But focusing on the adequacy of special relationship analysis in such cases fails to explain what the Thompson court itself recognized. Ignoring the possibility that a relationship between the county and the killer might be adequate to impose a duty upon the county, the court simply stated, "[a]s will appear, our conclusion is based in part upon policy considerations and in part upon an analysis of 'foreseeability' within the context of this case." n125 The court's reasoning is based upon a selective discussion of policy considerations rather than the common law special relationship formulation.

Although California courts are more explicit than most about considering a variety of factors in addition to the traditional special relationship rules, the results in the cases discussed above are not atypical in other jurisdictions. The Tarasoff case has been followed widely, and even expanded upon, with a number of courts imposing upon health care professionals duties to warn third parties, even where the victim, a member of a foreseeable class of plaintiffs, was not "readily identifiable." n126 The California cases finding no liability for police officers or public agencies also reflect trends in other jurisdictions. n127

Although a few decisions appear to have abandoned the no-duty rule and its special relationship exceptions altogether, n128 most courts  [*897]  continue to rely upon the Restatement rules. n129 However, when policy and rules appear to be at odds, courts sometimes have expanded liability beyond what the rules originally would have supported. In those situations, courts sometimes stretch the rules, but try to stay within them. For example, the Michigan Supreme Court expands the notion of a special relationship to include social companions. n130

Increasingly, however, cases (like the California cases described above) seem to be governed primarily by policy concerns. Some expand liability as a result. n131 Others have used policy considerations to prevent  [*898]  the imposition of liability when the traditional special relationship rules might well have led to a different result. n132

In a variety of cases, and irrespective of outcome, the significance of policies that are not related to special relationships is becoming more pronounced.


Other than the special relationship exception to the no-duty rule, the major common law exceptions -- those based upon undertakings to rescue and those based upon a defendant's initial creation of the risk -- are really variations on the nonfeasance-misfeasance distinction that makes the exceptions necessary in the first place. They involve the characterization of at least some part of the defendant's behavior as an affirmative act. Not surprisingly, these exceptions share the limitations that make the fundamental distinction unsatisfying. n133 Just as  [*899]  it is difficult to determine what relationship makes it appropriate for a court to find a duty to protect or aid, it is also impossible to generalize about what constitutes an undertaking to aid or protect that will remove a defendant's behavior from the protection of the no-duty rules.

Consider the treatment at common law of a hospital emergency room. Absent a pre-existing relationship between the hospital and the prospective patient, the common law would not require that treatment be offered unless the hospital somehow had undertaken to treat the patient. Presumably, if the emergency room doctor promised a patient that if she waited half an hour she would be seen by a doctor, most courts would find an undertaking. But suppose a nurse receptionist, after questioning the patient and consulting with the doctor in charge by telephone, turned the patient away? Would the receptionist's questions to the patient and consultations with the doctor be enough to create an undertaking? n134 If the receptionist's deliberations before turning away the patient are enough to constitute an undertaking, should the hospital be advised to avoid a duty by refusing to talk with, or unlock the doors for, a prospective patient? n135

At least one court has suggested that simply by holding itself out as having an emergency room, the hospital may have undertaken a duty to give assistance and must act reasonably to respond to the needs of those who appear at its doors. n136 Consistent with that approach is the finding that providing a 911 operator is enough of an undertaking to create an obligation to rescue. n137 Thus in some situations, little more  [*900]  than the defendant's existence is required to constitute an undertaking to protect a stranger. n138

However, such a broad reading is not universally accepted. If it were, doctors holding themselves out as specialists of one sort or another could be obliged to help those who sought their treatment. It is clear that not all courts will find an undertaking so easily and generally, at least where no special relationship exists, doctors are free to refuse to treat. n139

Because courts have been dissatisfied with the prospect of freedom from liability for all "nonfeasors," they have created exceptions to the no-duty rule that cover certain situations even where no pre-existing relationship exists between the parties. The problem is that the determination about whether the defendant has undertaken to rescue is a conclusion that results in the imposition or denial of liability and that is based on impossible factual distinctions that are not necessarily pertinent to the concerns driving the case. The fundamental issue is whether hospitals, for example, should be liable for turning away patients. If we decide that a hospital should be required to provide emergency treatment, then it makes little sense that one who locks its doors should be insulated from liability while one that answers questions before turning someone away would not. The problem is that in order to decide cases based on the broader policies of negligence law, courts must either manipulate and misapply the Restatement rules or must reject them outright.


A. The Proposal

The Discussion in Part II above has attempted to demonstrate that the no-duty-to-rescue-rule and its exceptions impede courts from focusing adequately on considerations of safety, loss distribution or fairness. The misfeasance-nonfeasance distinction is only serendipitously  [*901]  related to considerations that are really determinative. Therefore the no-duty rules must be manipulated by courts attempting to reach just results.

In this way the shortcomings of the rescue rules resemble the California Supreme Court's description of the failings of traditional rules in the landholder context:

Complexity can be borne and confusion remedied where the underlying principles governing liability are based upon proper considerations. Whatever may have been the historical justifications for the common law distinctions, it is clear that those distinctions are not justified in the light of our modern society and that the complexity and confusion which has arisen is not tied to difficulty in applying the original common law rule -- they are all too easy to apply in their original formulation -- but is due to the attempts to apply just rules in our modern society within the ancient terminology. n140

In spite of the shortcomings of the traditional no-duty rules, courts continue to rely upon them in varying degrees. The alternative proposed by this Article is very simple. The "terminology" of the Restatement approach should be put aside. n141 Society should require reasonable efforts to aid or protect others absent overriding competing societal interests. Courts should allow juries to consider the reasonableness of injury-causing behavior (even if the behavior could be characterized as nonfeasance) as long as there is no policy concern that outweighs the goals promoted by a finding of liability. n142

 [*902]  The Wisconsin Supreme Court has endorsed such an approach in considering conduct that ordinarily would be characterized as nonfeasance. In Schuster v. Altenberg, n143 the court reversed a judgment on the pleadings for defendants. Edith Schuster was killed when the car she was driving was involved in an accident. As a result of the accident, her daughter was injured. The daughter brought suit against her mother's psychiatrist for negligently treating her mother's psychosis and for failure to warn Edith Schuster or her family of the danger of her psychotic condition. n144

In discussing the failure-to-warn claim, the court stated:

[W]e need not engage in analytical gymnastics to arrive at our result by first noting that at common law, a person owes no duty to control the conduct of another person or warn of such conduct, and then finding exception to that general rule where the defendant stands in a special relationship to either the person whose conduct needs to be controlled or in a relationship to the foreseeable victim of the conduct. n145

Instead, the court concluded that if it were ultimately shown that the defendant's conduct was unreasonable under the circumstances, negligence would be established unless public policy considerations precluded liability. n146 The court concluded that although imposing a duty might be characterized as an "imposition of an affirmative duty to act" n147 no public policies would preclude liability in such cases generally and that therefore dismissal at the pleading stage was inappropriate. n148  [*903]  The court listed various policies that might prevent the imposition of liability for conduct found unreasonable.

(1) The injury is too remote from the negligence; or

(2) the injury is too wholly out of proportion to the culpability of the negligent tort-feasor; or

(3) in retrospect it appears too highly extraordinary that the negligence should have brought about the harm; or

(4) because allowance of recovery would place too unreasonable a burden on the negligent tort-feasor; or

(5) because allowance of recovery would be too likely to open the way for fraudulent claims; or

(6) allowance of recovery would enter a field that has no sensible or just stopping point. n149

Various jurisdictions, through their own common law development, have arrived at similar lists of policies to be considered before liability is imposed. n150 It is not the purpose of this Article to prescribe a particular list of factors to be adopted by all jurisdictions, but rather to encourage the adoption of an approach that may be adapted by courts  [*904]  working within the established framework of their own common law traditions.

An approach similar to Wisconsin's has been espoused by the California Supreme Court in a variety of cases. n151 The presumption in California is that a duty of ordinary care exists. n152 The presumptive duty may be overcome where a court considers one or more of an enumerated group of policy considerations sufficient to warrant shielding the defendant from liability. n153 Thus, even where conduct appears clearly unreasonable, any one of a number of policy concerns might lead a court to dismiss the action.

Such an approach, if exclusively employed in a nonfeasance case, n154 would turn the traditional rules on their head. The presumption  [*905]  created by the Restatement rules -- that no duty to aid exists absent an exception to the rule -- would be replaced by a presumption that reasonable conduct is required unless one or more policy concerns mandates a deviation from the norm. n155 Although the listed policies will conflict in some cases, conflict also exists where a court purports to be following the Restatement rules. What the suggested approach requires is that policy interests be made explicit and be discussed. n156

The suggested approach should give courts the flexibility to take into account interests that may compete with the interest in individualism. The approach not only represents a positive change from a rule-based approach, it has several advantages over proposals that suggest legally requiring only "easy rescues." Most significantly perhaps, the obligation of easy rescue, although academically interesting, is not particularly relevant in the large majority of cases which involve institutions that arguably have failed to take steps to reduce the risk of one individual hurting another. The proposal made in this Article offers another alternative, one which will apply to a large number of situations and should lead to judicial decisions that can adapt to changing conditions in an increasingly complex, dangerous society.

The approach should provide a court with an analytical framework not available under the traditional rules. For example, if a therapist has behaved unreasonably and harm has resulted, the question should not revolve around characterizing the conduct either as active mistreatment or a passive failure to warn, but rather should center upon whether there is any policy concern significant enough to shield the therapist from liability. Given otherwise identical facts would there be  [*906]  different concerns if the therapist were also a priest? What if the therapist worked not for a university, but for a county considering a weekend pass for the patient? n157 Adoption of the proposal would allow courts to analyze factual situations straightforwardly, giving full discussion to the fabric of the policy considerations that determine whether liability will be found. Such a process should ultimately lead to greater fairness in result and efficiency in process. n158 Defendants' obligations would not be expanded in every "rescue" situation were such an approach employed. For example, in considering the conduct of an electric utility that had caused a blackout, there would be no need to debate whether a failure to reduce electric load was nonfeasance or misfeasance; the analysis would be the same in either case. Still, the court might refuse to impose a duty if it determined that safety or loss distribution aims, for example, were outweighed in such a situation by concerns about the magnitude of liability. n159

Indeed there might be situations where even a very easy rescue effort would not be required because of countervailing policy concerns. For example, it might be held, because of privacy concerns, that one individual would not have a duty -- even where she could warn without personal cost to herself -- to warn another that a third individual was carrying a dangerous communicable disease. Similarly, where values of self-determination and privacy are significant, a doctor might have no duty to take even the most basic steps to rescue a patient who did not wish to be rescued. n160 Increasing technological developments may make the rescue of a dying patient easy from a medical perspective. However, where a competent, terminally-ill patient wants to refuse  [*907]  treatment, the societal costs of imposing a duty to rescue should be sufficient to defeat any obligation. n161

On the other hand, in some situations, the duty imposed under a policy-based approach might require affirmative efforts to aid or protect where they are not now required; and at times, the action required will be more than an easy rescue. A professional would be held to a professional standard, whether the conduct was characterized as misfeasance or nonfeasance. A landholder -- an innkeeper or a landlord -- might be required to make significant expenditures to provide a safer environment.

Most jurisdictions have not gone as far as Wisconsin has in abandoning the traditional approach, but several developments make likely continued movement toward the result that this article suggests -- the rejection of Restatement rules in a rescue context in favor of explicit consideration of policy factors for the analysis of duty issues. First, legislative action has laid the groundwork for a reconsideration of the traditional rules concerning affirmative duties. n162 More importantly, there is a strong, pre-Restatement common law tradition in California and other jurisdictions for analyzing duty questions without regard either to misfeasance-nonfeasance distinction or to the sorts of relationships considered by the Restatement's authors. n163 That tradition has continued to influence modern cases. And as already noted, courts sometimes have cited Restatement no-duty rules while appearing to decide cases on other grounds. n164

Nevertheless, although the groundwork has been laid, and in spite of a growing recognition of the value of requiring affirmative efforts to aid or protect in some situations, the move toward a rejection of the traditional rules has not yet been firmly established.

In other areas of tort law, courts have scrutinized common law rules that result in wholesale immunities from liability. n165 Where the  [*908]  immunity rules appear arbitrary or confusing, courts have been willing to sweep away the existing rules. n166 Perhaps the most dramatic example of such a development involved the traditional rules governing the duty of those who own or occupy land to those who come upon the land. In 1968, the California Supreme Court rejected the status-based classifications of trespasser, licensee and invitee in favor of a straight-forward negligence test. n167 The court looked to see whether any policy concerns should impinge upon the fundamental principle that one is liable for harm caused by a failure to exercise reasonable care under the circumstances; and it listed the major concerns to be balanced in determining whether a departure from the general rule was warranted. n168

The Rowland court reasoned that the application of the common law classifications was difficult, noting that the common law rules had been characterized as unrealistic, arbitrary and inelastic. n169 It noted that "exceedingly fine distinctions have been developed resulting in confusion and that many recent cases have in fact applied the general doctrine of negligence . . . rather than the rigid common law categories test." n170 Perhaps most importantly, the court concluded that the common law distinctions between trespassers, invitees and licensees, and the immunities based upon the distinctions did not always reflect the factors which should determine liability. n171

Rowland did not directly address the misfeasance-nonfeasance distinction. n172 Although subsequent California decisions have liberally cited the Rowland approach in cases involving affirmative duties to aid or protect, with rare exceptions, courts have not entirely abandoned either the nonfeasance-misfeasance distinction or the exceptions based upon it. Instead, the courts try to consider the Rowland factors and simultaneously rely upon the Restatement rules. It is as if, in the landholder context, the Rowland court had decided to rely upon policy  [*909]  factors and to continue to rely upon the distinctions between trespassers, licencees and invitees. While some decisions comfortably could rest on either analysis, some could not, and confusion would result. Cases involving questions of affirmative obligations to act for the most part continue to rely upon two different (and sometimes conflicting) approaches.

Occasionally, however, a court appears to abandon the misfeasance-nonfeasance distinction entirely. In Sprecher v. Adamson Co., n173 a natural landslide condition existed on defendant's 90-acres. n174 When a landslide damaged neighboring property, the landholder was sued. Defendant moved for summary judgment on the ground that it had no duty to abate a dangerous natural condition. The trial court ruled in defendant's favor and the Court of Appeal affirmed. The California Supreme Court reversed, however, agreeing with plaintiff that the common law rule of nonliability for a natural condition was "neither premised upon sound public policy nor in accord with modern principles of tort liability." n175

The court imposed a general duty to manage the property reasonably and remanded for a determination about whether such a duty had been breached. n176 In so doing, the court demonstrated a willingness to impose a duty where a defendant has done nothing to create or increase a pre-existing risk. n177 The rule of non-liability for natural conditions was merely a manifestation of the principle that one has no duty to undertake affirmative conduct to aid or protect others absent a special relationship. n178 The court made no suggestion that a special relationship existed between the neighbors, but stated, "[w]hatever the rule may once have been, it is now clear that a duty to exercise due care may arise out of possession alone." n179

 [*910]  Although the Sprecher decision focused on landowner liability rules and nowhere cited the Tarasoff decision, it is significant to the issues being considered here. Indeed, in applying ordinary duty criteria to a situation involving no active conduct on the part of the landowner, Sprecher laid the groundwork for a more dramatic rejection of the traditional rules than has come about. n180

One case in California did explicitly reject the Restatement approach in a rescue situation. In Soldano v. O'Daniels, n181 the California Court of Appeals held actionable a bartender's refusal to call police or to allow a good samaritan to use the telephone to call police to come to the aid of a man who was being attacked elsewhere. The court reached such a result even though the bar had no relationship either with the victim or with the attacker. Acknowledging that such a result was not supported by the traditional rules, the court employed a systematic analysis of the Rowland factors in concluding that a duty existed. n182

Even though Soldano represented a reasoned and cautious break from the traditional rules, and even though decisions are being determined by the policy considerations recognized in that case, courts have been slow to explicitly adopt such an approach. n183 It is time that they  [*911]  acknowledge the limitations of the traditional no-duty rules and adopt a workable alternative.

B. Discussion of the Proposal

In spite of the apparent need for change, many courts and a number of scholars have in recent years resisted any move to expand obligations to act affirmatively to protect another. It is therefore important to consider whether this recent scholarship suggests difficulties with the proposed approach that outweigh its benefits.

Opposition to proposals for expanding the duty to act affirmatively to aid or protect another is based upon both substantive and "process" concerns. The primary substantive arguments are that: (1) because a failure to act does not "cause" harm, no liability should attach without misfeasance; n184 and (2) positive duties would both interfere with individual liberties and would negatively affect society by requiring altruistic behavior. n185

The more telling arguments, however, are those that suggest that the judicial system would be strained by an expansion of the duty to rescue. Some of those arguments arise whenever a court chooses to rely upon standards rather than rules. n186 Other arguments specifically address difficulties anticipated if obligations to rescue were expanded. n187

 [*912]  The arguments, both substantive and process oriented, are generally offered against proposals for a duty of easy rescue, n188 but for the reasons set forth, they provide neither adequate theoretical support for the retention of the common law rules nor decisive criticisms of the approach proposed by this article.


The main causation argument used to support the common law distinction between misfeasance and nonfeasance is the long-recognized argument that even if an omission can physically cause another event, it would be unfair for our legal system not to distinguish legally between causing harm and merely allowing harm to occur. n189

It is argued that where there is a failure to rescue, the potential rescuer has in no way increased the risk of harm to the victim and therefore cannot be culpable. n190 However, in other contexts, the inability to prove physical causation will not always prevent the imposition of legal responsibility. In criminal law, for example, where a father fails to save an infant from drowning, "illegal inaction" may  [*913]  support criminal sanctions. n191 In tort law, courts have imposed liability in a wide variety of circumstances where policy concerns made it unnecessary for a plaintiff to prove that the defendant actually caused the plaintiff's harm. n192

Perhaps the best examples of tort rules that allow recovery for nonfeasance, however, are the common law exceptions to the no-duty-to-rescue rule themselves. Existing law allows actions to be brought where there is a failure to act, even where that failure does nothing to increase the risk of harm to the plaintiff, as long as a special relationship exists between the defendant and the plaintiff or between the defendant and the plaintiff's attacker. n193

Thus, in the Tarasoff case, for example, by concluding that plaintiffs could state a cause of action, the court implied that the action could proceed based on an allegation that a failure to act affirmatively "caused" Tatiana Tarasoff's death. n194 Where a special relationship has been recognized, the law has not hesitated to recognize, as "causes" of harm, omissions that did not independently increase the risk to the plaintiff.

Similarly, under existing rules, a rescuer who undertakes rescue efforts but is unsuccessful may be liable to a victim even though his actions have induced no reliance and have in no way increased the risk to the victim. Although the Restatement took the position that one who aids another is not ordinarily required to continue his assistance, n195 the Restatement also acknowledged the possibility of contrary results. n196

 [*914]  Regardless of how the causal dilemmas are characterized, standing on their own, clearly they do not always prevent liability under existing traditional rules. Therefore, standing alone, they are unpersuasive as arguments against a duty to act affirmatively based on a multi-factored policy approach. Unless one is a purist and refuses to recognize an obligation to rescue in any situation, it is disingenuous to insist that elimination of the no-duty-to-rescue rule in favor of a policy-based consideration of a duty to rescue would violate legal notions of causality; the existing rule frequently allows plaintiffs to recover even absent any showing that the defendant acted in any way to increase the risk of harm to the plaintiff.


a. Individual liberty

Another major philosophical argument against imposing additional rescue duties is rooted in a desire to protect individual liberty. n197 Proponents of continued no-duty-to-rescue rules argue that to require more would create unwarranted intrusions upon individual autonomy. n198

Those who base their opposition to the expansion of affirmative duties upon this limitation on liberty rely upon two premises. The first is that the imposition of a duty to rescue would be a more onerous burden on individuals than those imposed by legal prohibitions because prohibiting bad acts interferes with liberty less than requiring affirmative actions to protect others from harm. n199 The second is that even in our pluralistic society no legally recognizable interests are strong enough to counterbalance the intrusions on individual autonomy that an expanded duty rule might authorize. n200

The individualist might object that the government should never deprive him of the freedom to choose whether to engage in a rescue,  [*915]  distinguishing between rules that prevent people from doing acts and rules that require one person to help another. n201 The individualist would, however, accept rules that forbid harming others and thereby give up the ability to choose certain options, only because the rule provides him with a certain liberating protection. n202 Assuming that distinctions can be made between doing acts and failing to do acts, n203 do affirmative duties in general or a duty to aid, in particular, necessarily restrict personal liberty more than prohibitions? It seems that the relative burden of a rule depends more on the degree to which it limits choice than upon whether it is a prohibition or a requirement of positive action. The affirmative duty to pull an infant out of the path of a truck is certainly less onerous than a prohibition that allows no one with an infant to use the sidewalks without putting the child on a leash. n204

It is argued that the introduction of a general duty to aid others in certain situations would reduce the individual's ability to anticipate legal obligations; an emergency not of one's own making could arise at any time, imposing unexpected obligations. n205 It is true that a duty to aid a stranger in certain situations can be triggered fortuitously and that without a duty to rescue, the perfectly reasonable person could "mind her own business" and know what she might do within the law and what she might not. n206 But again it is not the nature of the obligation that determines how limiting the rule is. A requirement that one act reasonably to assist someone injured in an emergency that one witnesses is neither more nor less fortuitous, nor more or less a limitation  [*916]  on a person's ability to choose a course of action than the existing requirement that if involved in an emergency, she may not choose to do anything that a reasonable person would not do in the same situation. n207

One may always confine one's activities completely so as to try to avoid the risk of accidental occurrences. When individuals venture forth and encounter accidents, however, liability rules require that they behave reasonably. Thus, the degree to which the legal rule impinges upon individual freedom does not depend upon whether the rule prohibits unreasonable acts or requires reasonable ones. Instead, it depends upon the degree to which it takes choices from the affected individual. n208

Even assuming that the duty to rescue is more onerous than most prohibitions, few would suggest that a rescue obligation should never be imposed. Legislation is evidence of the fact that while individualistic values may be significant, they are not exclusively so. Indeed, they may be outweighed by other interests, notably the interest in public safety. Although statutory sanctions for failure to aid or protect another in danger have not been as numerous in this country as in Europe, n209 several states have enacted statutes that impose criminal penalties for a failure to take affirmative steps upon witnessing various types of assaults. n210 Other states have been more cautious, but nevertheless have indicated a willingness to limit freedom by imposing affirmative obligations under certain circumstances in order to promote safety. n211

In California, for example, the legislature has formally recognized the value of encouraging rescue n212 and has created limited affirmative  [*917]  duties to act on behalf of another in an emergency by making it a misdemeanor to fail to relinquish a party line that is needed for an emergency call to a police or fire department or ambulance or other rescue personnel. n213 The legislature has enacted a mandatory seat belt law that, among other things, imposes an affirmative duty to protect another in a non-emergency situation by imposing criminal penalties and the possibility of civil liability upon drivers who have failed to require their passengers to buckle up. n214

Many states require by statute that a driver involved in an accident, even if innocently, must offer assistance to accident victims. n215 Thus, even a faultless driver struck by one driving negligently must aid the negligent driver if the negligent driver is injured in the accident, or else face liability.

The question then, is not whether a duty to rescue can be imposed, but rather, given that the interest in preserving individual autonomy is not absolute, what are the interests that should be weighed in determining whether to impose a duty in spite of that interest? This proposal provides a more rational and explicit way of trying to make such determinations. n216

 [*918]  b. Societal values

A related argument supporting the continued use of the common law no-duty-to-rescue rules is that the law should not require altruism. n217 The argument runs that where altruism is required, it loses its moral value to society. n218 Epstein would argue that a person who protects a child about to step in front of a truck absent any legal obligation to do so is more virtuous than one who performs a similar rescue when it is required by law. Thus a rule requiring rescue deprives the rescuer, and indirectly the society itself, of a certain goodness. n219

To the extent that legislating conduct deprives the mandated action of moral value, it would justify the repeal of all criminal laws. It is doubtful, however, that requiring reasonable action to aid or protect would really reduce overall morality. First, it is not clear, even where the law requires that reasonable rescue actions be taken, that one who takes action and thereby saves a life will feel less virtuous because the act was legally required. Second, imposing a broader duty to rescue will not deprive society of all the moral worth that surrounds acts of real heroism. Actions that are beyond the call of duty will not be legally required in any case, so the law will not deprive society of heroes nor deprive individuals of all opportunities to perform heroically. n220 Thus, any reduction in morality that could be claimed as the result of increased requirements for reasonable behavior is minimal.

Moreover, any minimal reduction in morality that occurs is more than offset by the reduction of unreasonably dangerous conduct. By definition, the common law rules allow immunity for unreasonable failures to rescue (and sometimes unconscionable behavior). To the extent that the proposed rule would not condone such conduct, unconscionable behavior would be further reduced. n221 Thus, what society  [*919]  loses in the goodness of volunteerism, it will recoup in the reduction of blameworthy conduct. n222


Practical difficulties are often cited when courts consider abandoning any rule-based approach. n223 But the argument has been particularly telling in this area of tort law. Retention of the nonfeasance-misfeasance distinction often has been justified by reference to the "difficulties of setting standards of unselfish service to fellow men. . . ." n224 Professor Henderson has contributed the most thoughtful analysis of the systemic (as opposed to substantive) difficulties without abandoning the traditional approach to rescue situations. He argues that the common law rules should be maintained because they serve (better than would an expanded duty of rescue) to inform both primary behavior and adjudicative behavior. n225

 [*920]  To effectively guide primary behavior, Henderson argues, legal rules must inform people of liabilities that will be imposed for particular behavior. To do that they must be clear, they must refer to verifiable facts and they must not call for unachievable behavior. n226 Courts also need clear, verifiable rules that can be conformed to. In addition, the rules must be manageable; that is, they must not present the court with highly complex, polycentric problems. n227

To the extent that these suggestions argue against application of negligence principles in rescue cases, they argue with equal force against the application of negligence principles in all tort cases. It is the negligence standard itself that is vague, but it has proved adaptable and generally more durable than more rigid rules of conduct. n228 Henderson acknowledges that "the effects of negligence rules on primary behavior suggest that even superficially vague rules can guide behavior if they invoke widely understood and recognized moral principles." n229

The proposal in this Article has the virtue of having the comprehensibility of the negligence standard. People can understand a rule that says," generally, you are expected to conduct yourself reasonably" more easily than they can understand the common law rule with its various exceptions or, perhaps even a rule that requires one to rescue or protect, but only where to do so would be "easy." n230

The traditional no-duty-to-rescue rules are less comprehensible than the proposed standard because they are consistent with intuitively understood moral principles in certain instances and contrary to those principles in others. For example, it is not immediately apparent why a person need not make a reasonable effort to rescue an infant stepping in front of a truck; why a person who attempts a rescue generally risks liability more than one who turns away; why one who is innocently involved in an accident is obliged to rescue when the person who observed the incident is not. More significantly, businesses considering whether to provide security in a particular situation may be guided to  [*921]  safer conduct by focusing on whether it is reasonable to take steps to protect customers than by considering whether the jurisdiction would immunize inactivity by describing it as nonfeasance and finding no exception to the no-duty rule.

But even if "reasonableness" is an understandable concept, some would argue that the proposal made here would leave people in emergency situations without a way to determine whether action or inaction is reasonable. For example, a person witnessing a man and a woman struggle in the back seat of a car "may not be sure whether he is watching a rough and tumble courtship or an imminent rape. A call to the police may either be helpful or traumatic in its effect, depending on the circumstances." n231

But similar ambiguities exist in commonplace misfeasance cases. In determining what the most reasonable course of action is, people must consider a number of alternative courses of action, sometimes without all the facts at hand. A manufacturer considers alternative designs for a product and makes a decision about what is the most reasonable course of conduct. Among the alternatives that the manufacturer considers is leaving the design as it is. Thus, the consideration of the alternatives includes the option of doing nothing, just as it does in the example above. If anything makes the observation of the backseat activity more ambiguous, it is the emergency nature of it, not the fact that "non-action" is an alternative.

It is also instructive to consider the reasonableness of alternative courses of action by determining what the cost of error would be. Looked at in such a light, the decision does not seem so difficult. A mistaken call to police may result in embarrassment. On the other hand, if an observer does nothing because of a mistaken belief that no harm is being done, the costs to the victim may be devastating. n232

 [*922]  Once again the differences between nonfeasance and misfeasance situations may have been exaggerated. But even assuming that the ambiguities are greater in the nonfeasance context, it does not follow that a general duty is unacceptable on this score unless similar ambiguities are avoided under the common law rules. For example, if the car were parked in an area where it was observed by a security guard, the guard would be required to act in the face of identical ambiguities. n233 Similarly, a person who, upon seeing the situation, went to call for help and then changed her mind might also be liable under current law. n234 The person who innocently causes harm is also liable for a failure to rescue as the result of statute n235 or common law doctrines. n236 The ambiguity is in the situation, not in the difference between action and inaction. Under the existing law, the ambiguities still exist, but the person who ignores the situation is legally protected, while the person making a reasonable attempt to help by calling for help, flashing headlights, or calling for police on a car phone increases her risk of being sued. n237

 [*923]  Henderson also urges that the general duty to rescue would pose conformability problems. n238 That is, mentally incompetent persons or shy persons saddled with a duty to make an objectively reasonable rescue would be asked to deliver more than they are capable of. If, on the other hand, the standard of care were based on a subjective standard, Henderson argues, it would be difficult to verify that particular individuals had done what reasonably could have been expected. n239 The dilemma is not an easy one, and it has provided difficult issues for courts and commentators alike whenever the question of negligence arises. n240 Henderson reasons, however, that the situation is different in a rescue situation because in the rescue situation the mentally deficient defendant did not cause the harm whereas in the misfeasance situation he did. The innocence of the defendant n241 is not enough to exonerate him, but the innocence and inaction together are.

Thus, causation is really at the root of the argument once again. The argument that the distinction between action and inaction is not necessarily the difference between causation and the lack of it is made elsewhere. n242 Even if it were, as has been noted often, the characterization  [*924]  of behavior as an act or non-act is often susceptible to judicial manipulation. n243 And once again, to the extent Henderson's process criticism is valid, it also applies to the common law rules which often allow a defendant who is both "innocent" and inactive to be liable to another who suffers as the result of the defendant's inaction. n244 Henderson and others have made the related argument that rescue law should continue to immunize nonfeasance because of judicial and legislative line-drawing problems. n245 It would be impossible to decide when, if ever, not to impose such an obligation. n246 Professor Epstein put the argument graphically:

[If] an individual is required under some circumstances to act at his own cost for the exclusive benefit of another, then it is very hard to set out in a principled manner the limits of social interference with individual liberty . . . Even if the rule starts out with modest ambitions, it is difficult to confine it to those limits. Take a simple case first. X as a representative of a private charity asks you for $ 10 in order to save the life of some starving child in a country ravaged by war. There are other donors available but the number of needy children exceeds that number. The money means nothing to you. Are you under an obligation to give the $ 10? n247

But line-drawing problems exist in current negligence law without preventing its operation. Courts use policy considerations to place limits on liability that would not be limited by the application of rules of physical causation alone. For example, using either duty or proximate cause analysis, courts have made it very difficult for plaintiffs to recover for pure economic harm. n248 The proposal made in this Article does not  [*925]  bar courts from considering whether there are policies that would shield the behavior in Epstein's example from legal action. n249 In that sense, the example creates no real dilemma. The proposal made herein does nothing more than to require courts to articulate the real policies that they believe should prevent the imposition of liability if any such cases should arise. Although some might argue that a presumptive duty would increase the number of cases facing new line-drawing difficulties in situations like the Epstein example, it seems highly unlikely. Existing law, even in nonfeasance situations, faces line-drawing dilemmas of the type that Epstein anticipates without encountering significant difficulties. If Epstein's concerns were not exaggerated, under existing law courts might be presented with lawsuits every time a person sent a $ 10 check to a starving child and then neglected to continue to send checks to sustain the child. Arguably the person has undertaken the rescue and, under existing law, by so doing has a duty. n250 The absence of any such reported case is an indication that the fear is not substantial. The point is two-fold. Current law faces line-drawing problems identical to those feared by Epstein and others; and, the problems are not insurmountable.

The extreme cases for the most part will not get to court. n251 Those that do may be disposed of under existing principles. In Epstein's example, the lack of directness between the inaction and the harm would allow a court to refuse to find liability in such an instance. n252 In less extreme cases, the deliberation would be more difficult, but it already is. The current common law no-duty-to-rescue rules are riddled with line-drawing problems on tangential issues. n253 The adoption of a general policy-based duty analysis should, if anything, simplify the task of the court. n254

 [*926]  The second of the manageability problems that concerns those who have considered a broadened requirement of affirmative obligations is the difficulty recognized by the California Supreme Court in Tarasoff. "[The absence of a general duty to rescue] owes its survival to the difficulties of . . . making any workable rule to cover possible situations where fifty people might fail to rescue. . . ." n255 Some of the most notorious failures to rescue have involved situations where a large number of potential rescuers failed to do so much as make a telephone call to avert virtually certain serious harm to a helpless victim. In the well-known Kitty Genovese case, a young woman was attacked three separate times over a thirty-five minute period. Although thirty-eight people watched one or more of the attacks from their windows, the victim was dead before anyone even telephoned the police. n256 In a more recent and equally publicized event, people in a barroom watched while a woman was repeatedly raped on a pool table. n257

The position of some commentators is that the multiple defendant litigation resulting from such situations would pose difficult management problems for the court. n258 Plaintiffs might "join a number of defendants in order to increase their chances of recovery." n259 In addition, defendants would have an interest in joining parties that the plaintiff failed to join. Moreover, the court would have to sort out the various defendants in order to determine liability. But such problems are familiar in any multi-party litigation. n260

Even if the problems are identical, a general duty could be problematic if its adoption greatly increased the number of cases that had to deal with them. n261 Again, however, an argument that is interesting theoretically may be of little practical significance. First, although they receive tremendous publicity, the multiple party failures to rescue do not seem particularly common. A review of litigation where a general duty has been adopted does not indicate significant manageability problems  [*927]  based on the number of defendants. n262 Second, if one assumes that the law indeed can change behavior, n263 adoption of a general duty to rescue might actually reduce the number of occurrences of group failures to act. n264

Process concerns should not prevent a needed development in the law. Even where nonfeasance is an issue, courts can manage a legal approach that, absent overriding policy concerns, requires reasonable behavior. A number of courts already do it, if inconsistently. n265 Some have done it explicitly, refusing to be controlled by misfeasance-nonfeasance distinction and traditional relationship or undertaking exceptions. n266 Others have not explicitly abandoned the common law rules, but have increasingly relied upon discussions of policy. n267 A significant change in the law has taken place even without express rejection of the common law rules. It is now time to make explicit the changes that have begun.


The strengths of this proposal are the strengths of negligence law. So too, are the weaknesses. To the extent that criticisms of the approach are valid, they necessarily raise questions about negligence law itself.

The law in this area has resisted change not because such change would conflict with the aims of tort law, but because of fears that the litigation system would have difficulty handling a consistent application of negligence principles in cases involving failures to act affirmatively on behalf of another.

Assuming that a broader obligation to aid and protect others would be more consistent with recognized substantive tort aims -- safety, for example -- than are the Restatement rules, and that concerns about process problems are in fact what is preventing substantive change for the better, I would conclude that the law will be improved by the changes  [*928]  proposed herein. Another possible conclusion exists: if the tort litigation process cannot in fact consistently handle otherwise desirable substantive developments, it may be the process, not the aims, that ultimately give way.

To the extent courts increasingly attempt to articulate the underlying considerations that are actually guiding their decisions in this area of tort law, the decisions will be stronger for it. In addition, judges will be required to better articulate the reasons for their decisions. Whatever is then concluded about the relative strengths and weakness of the tort system and its alternatives, the development cannot help but be a healthy one.

n1. See, e.g., early articles in the debate: James B. Ames, Law and Morals, 22 HARV. L. REV. 97 (1908); Francis H. Bohlen, The Moral Duty to Aid Others As a Basis of Tort Liability, 56 U. PA. L. REV. 217 (1908).

n2. The generally accepted basis for the "no-duty-to-rescue" rule is set forth in RESTATEMENT (SECOND) OF TORTS 314 (1965) [hereinafter RESTATEMENT], which states in pertinent part: "The fact that the actor realizes or should realize that action on his part is necessary for another's aid or protection does not of itself impose upon him a duty to take such action." For a description of the rules most directly related to 314, see infra notes 32-38 and accompanying text.

n3. In his 1908 article, Ames, then Dean of the Harvard Law School, proposed that liability be imposed upon one who causes the death or serious injury of another by failing to undertake a rescue, but only where rescue was without "cost" to the rescuer. Ames, supra note 1, at 113. A number of commentators since have proposed expanded duties. See infra note 142.

n4. However, there have been changes in the way courts use the rules. Although courts continue to cite the Restatement rule and the exceptions to it, other considerations often determine whether a duty is imposed. See infra notes 74-139 and accompanying text.

n5. The presumptive immunity for one who turns away from someone in need of assistance has resulted in decisions "shocking in the extreme" and "revolting to any moral sense." W. PAGE KEETON ET AL., PROSSER AND KEETON ON THE LAW OF TORTS 56 (5th ed. 1984) [hereinafter PROSSER & KEETON]. For a defense of the consideration of moral values in decision-making, see generally Ernest J. Weinrib, The Case For A Duty To Rescue, 90 YALE L.J. 247 (1980). The expansion of a duty to affirmatively aid or protect has also been opposed as an inappropriate use of legal sanctions to enforce moral obligations. See Martin B. Rosenberg, The Case Against a Duty to Rescue, 19 COLUM. J.L. & SOC. PROBS. 1 (1985).

n6. Such paradigms were inspired by early cases like Yania v. Bigan, 155 A.2d 343 (Pa. 1959). The court in Yania concluded that a landholder who prodded a business invitee to jump into water had no duty to pull him out before he drowned. In 1959, Yania was the only reported state appellate decision to base a decision on the no-duty-to-rescue statement of 314 of the Restatement. Although cases involving one private individual's failure to engage in an effortless rescue do still arise, today they are relatively rare. Professor Rabin describes these simpler paradigms as "little more than an academic exercise" and speculates that expanded requirements for affirmative action have resulted in part because of the inability of law enforcement to cope with societal violence. Robert L. Rabin, Tort Law in Transition, 23 VAL. U. L. REV. 1 (1988).

n7. See, e.g., Kline v. 1500 Massachusetts Av. Apartment Corp., 439 F.2d 477 (D.C. Cir. 1970) (apartment buildings); Williams v. Saga Enterprises, 274 Cal. Rptr. 901 (1990) (taverns); Walker v. County of Los Angeles, 238 Cal. Rptr. 146 (1987) (municipal parks); Univ. of Denver v. Whitlock, 744 P.2d 54 (Colo. 1987) (universities); McIntosh v. Milano, 403 A.2d 500 (N.J. Super. Ct. Law Div. 1979) (mental health workers); Fuhrer v. Gearhart By The Sea, Inc., 760 P.2d 874 (Or. 1988) (hotels); Mostert v. CBL & Assoc., 741 P.2d 1090 (Wyo. 1987) (theaters). A significant number of cases involve failures to prevent another type of violence -- suicide. Nally v. Grace Community Church, 763 P.2d 948 (Cal. 1988) (church counselor-counselee), cert. denied, 490 U.S. 1007 (1989); Seiler v. Bethany, 746 P.2d 699 (Okla. Ct. App. 1987) (jail-prisoner).

These cases appear to be concerned not only with issues of safety, but with loss distribution as well. Where courts are willing to entertain such actions, litigation provides an avenue for an injured party to recover through the tort system when the most obvious defendant is not a likely source of recovery. Assailants are not always capable of paying a judgment, even when they can be apprehended. In this way, these cases resemble the triads so familiar in other types of cases, such as those in which a plaintiff who is unable to recover from its judgment-proof debtor resorts to seeking compensation from the accountants that vouched for the debtor's credit-worthiness. See, e.g., Int'l Mortgage Co. v. John P. Butler Accountancy Corp., 223 Cal. Rptr. 218 (1986); Credit Alliance Corp. v. Arthur Andersen & Co., 483 N.E.2d 110 (N.Y. 1985). See also Robert L. Rabin, Tort Recovery for Negligently Inflicted Economic Loss; A Reassessment, 37 STAN. L. REV. 1513 (1985). However, the interest in compensation is not absolute. Courts still regularly deny liability in rescue situations, using either traditional rules or other duty constraints to shield those who fail to aid or protect. See, e.g., infra notes 129, 132 and cases cited therein.

n8. See infra notes 22-43 and accompanying text.

n9. The relationship may be pre-existing or may be created when one party takes steps to rescue the other. See infra notes 34-40 and accompanying text.

n10. See infra notes 44-48 and accompanying text.

n11. The terms "no-duty rules," "no-duty-to-rescue rules," "no-duty-to-aid rules" and "no-duty-to-protect rules" will be used interchangeably, and broadly, to refer to the constellation of rules and exceptions set forth in Topic 7, RESTATEMENT, supra note 2, §§ 314-24. However, it should be noted that within Topic 7, some obligations are more circumscribed than others. See infra note 38 and accompanying text.

n12. Although the Restatement rules are referred to throughout as the "traditional" rules, a strong argument can be made that the common law provides a long-standing tradition (dating back to Heaven v. Pender, 11 Q.B.D. 503 (1883)) for deciding rescue situations without regard either to distinguishing acts and failure to act, or considering "special relationships" between parties as that term is used in the Restatement. See generally James P. Murphy, Evolution of The Duty of Care: Some Thoughts, 30 DEPAUL L. REV. 147 (1981) [hereinafter Murphy]. Professor Murphy identified the historical tension between a general duty of care and the no-duty-rules of the Restatement. Id. at 176-77.

n13. See infra notes 125-32 and accompanying text.

n14. See generally infra, Part II. Although it is true that a number of cases have greatly expanded the duty to rescue, a policy-based approach to the courts' duty determination does not automatically lead to that result. Such an approach may also lead courts to justify the protection of defendants. For example, in some instances courts have found no duty even where "special relationships" arguably would have required its imposition. See, e.g., Thompson v. Alameda County, 614 P.2d 728 (Cal. 1980) and discussion infra notes 119-24 and accompanying text. To the extent that courts are already relying more on policy and less on the Restatement rules, the proposal made by this Article is evolutionary, urging that we recognize what many courts are already doing in varying degrees.

n15. See infra notes 149-54 and accompanying text for an enumeration of the policies some courts have considered in determining whether to allow an action to proceed to trial.

n16. 443 P.2d 561 (Cal. 1968).

n17. Id. at 567-68.

n18. The failure of the pure bystander to undertake a cost-free rescue represents an extreme case. For that reason, it seems important to consider potential objections to the proposal made herein in that context. If the objections to the alternative proposed here are not convincing even in the hardest cases, they should give us no pause when we consider the more common situations. See supra note 7.

n19. See, e.g., Richard A. Epstein, A Theory of Strict Liability, 2 J. LEGAL STUD. 151, 200-01 (1973) [hereinafter Epstein]. Epstein's position has evolved since his theory was introduced. See, e.g., Richard A. Epstein, Causation -- In Context: An Afterword, 63 CHI.KENT L. REV. 653 (1987). Others suggest that to require warning or rescue might not maximize societal efficiency. The efficiency theorists are not considered in depth. Although they raise questions, they do not seem to come to a definite conclusion about the effects of an expansion of the duty. See, e.g., William M. Landes & Richard A. Posner, Salvors, Finders, Good Samaritans, and Other Rescuers: An Economic Study of Law and Altruism, 7 J. LEGAL STUD. 83 (1978).

n20. The concern is one that always accompanies a move away from rule-based decision-making. Such concerns play a major role in courts' decisions in a variety of contexts. For a recent example of a vigorous defense of a rule-based duty approach in negligent infliction of emotional distress cases, see Thing v. LaChusa, 771 P.2d 814 (Cal. 1989). However, the anxiety about the effect of eliminating traditional rules has also been significant in this particular area of tort law. The argument that practical constraints explain the resistance to change of rescue law has been best developed and articulated by Professor James A. Henderson. See generally James A. Henderson, Process Constraints in Tort, 67 CORNELL L. REV. 901 (1987), and discussion infra notes 224-64 and accompanying text.

n21. I have chosen in this Article to propose a change in the law that is readily obtainable. However, I remain interested in the larger changes that, while politically unobtainable in the near term, might ultimately be more satisfying. Although I am convinced that the proposal made herein would result in a significant, positive development in the law, the proposed changes ultimately share the limitations of a system that generally insists that compensation be limited to situations in which the plaintiff's injury is causally related to some unreasonable conduct on the defendant's part. For example, it is not clear why an injured plaintiff should be compensated only if lucky enough that either his attacker is not judgment proof or that the victims' lawyers are clever enough to identify parties that arguably could have prevented, or at least warned of, the danger. Of course a system that reimbursed the plaintiff without having to assign blame would eliminate the need for litigating failures to rescue that have been both philosophically and legally troublesome. For an excellent discussion of the weaknesses of tort litigation and a call for systematic reform, see Stephen D. Sugarman, DOING AWAY WITH PERSONAL INJURY LAW (1989).

n22. PROSSER & KEETON, supra note 5, 30, at 134-35.

n23. See, e.g., Osterlind v. Hill, 160 N.E. 301 (Mass. 1928); Buch v. Amory Mfg. Co., 44 A. 809 (N.H. 1898). Although courts most frequently analyze failures to rescue as duty problems for the courts themselves to resolve, issues of causation are also present. Causal links between a bystander's inaction and the plaintiff's injuries are said to be tenuous where the bystander did nothing to create the plaintiff's predicament initially. See infra notes 189-96 and accompanying text.

n24. The distinction may have historical roots that are both practical and philosophical. Practically speaking, early courts' concern with keeping the peace probably led to a focus on aggressive misconduct rather than failures to aid, however egregious the failure. See Bohlen, supra note 1, at 217. A more enduring interest may have been the more abstract aim of protecting individualistic values by refusing to legally require unselfishness. 2 FOWLER V. HARPER & FLEMING JAMES JR., THE LAW OF TORTS, 18.6, at 1049 (1956).

n25. For a discussion of some of the practical difficulties faced by courts trying to determine when to apply a "no-duty-to-rescue" analysis, see infra notes 49-64 and accompanying text.

n26. For an example of the significance of the distinction, see Weirum v. RKO General, Inc., 539 P.2d 36 (Cal. 1975). The court in Weirum recognized that the characterization of RKO's behavior determined which of the two distinct approaches to the duty question would be utilized. In Weirum, a radio station offered prizes for the first listener to locate one of its disc jockeys who was driving from one place to another. Two teenagers racing to find the disc jockey were involved in an accident and injured another person. The radio station argued that it had no duty to protect the victim because it shared no relationship either with the victim or with the teenagers. The California Supreme Court recognized that liability for nonfeasance -- failure to intervene to aid or protect -- "is largely limited to those circumstances in which some special relationship can be established." Id. at 41. However, the court obviated any need to discuss the existence of a special relationship by finding that the station's behavior amounted to misfeasance -- the creation of an unreasonable risk. Where the defendant's behavior was characterized as misfeasance rather than nonfeasance, "the question of duty is governed by the standards of ordinary care. . . ." Id. The proposal made by this Article eliminates the need for courts to rely upon the distinction.

n27. For a discussion of the commonly recognized exceptions to the rule, see infra notes 32-38 and accompanying text.

n28. Not surprisingly, the protection of such behavior has resulted in some appalling decisions. E.g., Osterlind v. Hill, 160 N.E. 301 (Mass. 1928) (no duty where owner rents canoe to intoxicated customer, then ignores repeated cries for help when customer falls into water, clings to canoe, and later drowns); Yania v. Bigan, 155 A.2d 343 (Pa. 1959).

n29. See Weinberg v. Dinger, 524 A.2d 366 (N.J. 1987) (private water company may be liable where negligent in maintaining adequate water pressure for firefighting); Defrancesco v. Western Pennsylvania Water Co., 478 A.2d 1295 (Pa. Super. Ct. 1984) (exculpatory clause in public utility tariff that prevented liability for losses due to insufficient water pressure void as against public policy).

n30. Even if the court finds misfeasance, other considerations may lead it to conclude that the defendant should not shoulder the burden of plaintiff's injury. See, e.g., Strauss v. Belle Realty Co., 482 N.E.2d 34 (N.Y. 1985) (no duty of care owed to tenant suffering injuries in the common area of the building where he lived even though the utility's conduct was grossly negligent, largely because of concerns with placing limits upon liability where huge numbers of potential plaintiffs). See also Robins Dry Dock & Repair Co. v. Flint, 275 U.S. 303 (1927) (placing limits on liability where misfeasance causes purely economic harm); Thing v. La Chusa, 771 P.2d 814 (Cal. 1989) (placing limits on liability when misfeasance causes pure emotional distress); Bovsun v. Sanperi, 461 N.E.2d 843 (N.Y. 1984).

n31. See infra notes 32-38 and accompanying text.

n32. Thus, an adult who observes an infant in danger has no obligation to act affirmatively to protect the child unless the adult either: (1) has a pre-existing relationship with the child; (2) has become involved with the child by virtue of having created the risk initially (even if non-negligently); or (3) has become involved by virtue of taking initial, gratuitous steps to help. See infra notes 35-37 and accompanying text.

n33. The example is simple and illustrative, but not typical. Most cases now involve institutions that arguably have failed to provide adequately safe conditions. In fact, the simple example involving a non-institutional defendant reacting to an emergency emphasizes a number of concerns that are not involved in the more typical case. Concerns for autonomy, the subjective state of mind of the defendant, and the danger faced by the bystander are not generally as significant where institutional safety conditions are involved. Fred C. Zacharias, The Politics of Torts, 95 YALE L.J. 698, 708 n.61 (1986).

n34. An indication of the fact that more than the Restatement rules are guiding the outcome is that a police officer at the intersection may be given more legal protection than the babysitter, the good samaritan, or the person who had hit the truck. Although the reasons for the outcome have been articulated in a variety of ways, generally, absent some articulated personal commitment, police officers have no duty to any particular citizen. This result obtains even where no governmental immunity protects the defendant. See infra note 118.

n35. The basic articulation of the no-duty principle is found in RESTATEMENT, supra note 2, 314. However, certain relationships traditionally have been sufficient to impose a duty upon an individual to act affirmatively to protect another from a dangerous situation, even where that individual did nothing to create the danger. Sections 314A and 314B identify five relationships that give rise to such an obligation. Section 314A states in pertinent part:

(1) A common carrier is under a duty to its passengers to take reasonable action

(a) to protect them against unreasonable risk of physical harm, and

(b) to give them first aid after it knows or has reason to know that they are ill or injured, and to care for them until they can be cared for by others.

(2) An innkeeper is under a similar duty to his guests.

(3) A possessor of land who holds it open to the public is under a similar duty to members of the public who enter in response to an invitation.

(4) One who is required by law to take or who voluntarily takes the custody of another under circumstances such as to deprive the other of his normal opportunities for protection is under a similar duty to the other.

See id. 314A.

Section 314B restates the Agency rule which imposed upon an employer a similar duty to "servants" acting within the scope of their employment. Cases based on the exceptions have been widely reported. See generally PROSSER & KEETON, supra note 5, 56.

The rules that apply in the situational crisis in this example also apply if a dangerous individual poses the threat, except that where a dangerous individual poses the threat, protection may also be required even if the potential rescuer shares a relationship only with the dangerous person. See infra note 38 and accompanying text.

The special relationship exception is the only "pure" exception to the rule that nonfeasance is not actionable. Where the special relationship exception is applied, the court finds no active conduct, but still imposes liability. In both other situations -- where an individual undertakes to aid or protect, or, where the individual created the risk -- a court may either: (1) conclude that the behavior amounts to misfeasance, obviating the need to rely upon the established exceptions; or (2) may characterize the behavior as nonfeasance and apply the appropriate exception. Either approach, however, depends upon characterizing the defendant's behavior as "active" conduct.

n36. The law recognizes that even if there is no other relationship between the parties, a risk-creator has a duty to provide reasonable aid or protection to one put in danger. RESTATEMENT, supra note 2, §§ 321, 322. For those who define nonfeasance (and the no-duty-to-rescue rules that accompany it) as the absence of risk-creating behavior, the risk-creating behavior may be seen as taking the potential rescuer's conduct out of the realm of nonfeasance altogether. Because the defendant created a risk, the court need not concern itself with no-duty arguments based on nonfeasance and may proceed to analyze the reasonableness of defendant's failure to rescue.

Alternatively, courts sometimes analyze this situation as one creating a relationship. See, e.g., Johnson v. State, 447 P.2d 352 (Cal. 1968). In Johnson, the court held actionable the state's failure to warn foster parents that a ward placed in their home displayed violent tendencies. The court found that because the State had placed the youth with the plaintiffs, the State's relationship with the plaintiffs was such that its duty extended to warning them of latent, dangerous qualities suggested by the youth's history or character. Id. at 355.

n37. RESTATEMENT, supra note 2, 324. Even where a defendant has done nothing to create a risk and has no special relationship with either a person requiring aid or an individual who is posing a threat to another, a duty may be found where the defendant has taken steps to assist the person in need. It is clear that an undertaking to aid or protect the person or property of another creates a duty to act reasonably either where a failure to do so increases the risk of harm or where harm results because the person in jeopardy relies upon the undertaking. However, the wording of 324 also suggests that a rescuer failing to use reasonable care in assisting one who is helpless may be liable even though the victim's situation did not worsen.

Although this example may appear to fall short of the "taking charge" referred to in 324, courts obviously have considerable latitude in determining what amounts to a rescue effort. See infra notes 133-39 and accompanying text.

The Restatement also explicitly leaves open the possibility that one who has thrown a rope to a drowning person, pulled him for a while and then unreasonably abandoned the effort may be liable even though the victim was not left in a worsened position as a result of the temporary efforts and clearly would have drowned had the potential rescuer never happened by. Id. 323, cmt. e. See also the caveats to both sections 323 and 324 (recognizing the possibility that liability may sometimes result even where discontinuance of the aid does not increase the risk of harm to the victim, as long as the abandonment leaves the victim exposed to an unreasonable risk). But see Miller v. Arnal Corp., 632 P.2d 987 (Ariz. Ct. App. 1981) (rejecting standard that would require those looking for avalanche victims not to abandon search).

Further, it would appear that if the rescuer had pulled the swimmer from the water, but then had left him in a weakened condition where he would be exposed to the elements, the rescuer might be liable for departing prematurely. Comment g to Section 324 states: "if the actor has succeeded in removing the other from a position of danger to one of safety, he cannot change his position for the worse by unreasonably putting him back into the same peril, or into a new one." See Parvi v. City of Kingston, 362 N.E. 2d 960 (N.Y. 1977) (duty of police officers who had removed drunks from danger not to leave them in another perilous situation).

n38. There is no duty to control the conduct of a third person as to prevent him from causing harm to another person unless

(a) a special relation exists between the actor and the third person which imposes a duty upon the actor to control the third person's conduct, or

(b) a special relation exists between the actor and the other which gives the other a right to protection.

RESTATEMENT supra note 2, 315 (emphasis added).

Relationships between a potential aider and a dangerous person that will support a duty to protect a stranger are set forth in §§ 316-19. Id. 315 cmt. c. The listed (and limited) obligations involve parents' responsibilities to control their children; masters' responsibilities to control their servants; possessors of land or chattels responsibilities to control the conduct of one who has been permitted to use the land or the chattels; and the responsibilities of those in charge of dangerous individuals.

Relationships between the potential aider and the person in need of protection may also support a duty. See id. 314A, 315 cmt. c., 320.

n39. See id. 315.

n40. This assumes that therapeutic relationships generally are sufficient to impose a duty. Finding a legal obligation here obviously represents an expansive reading of the duty to control set forth in 315. Although the duty is expressed as a duty to control, modern decisions have often imposed a broader obligation -- for example a duty to warn a potential victim -- where actual control over the dangerous person is not feasible. See infra notes 93-101 and accompanying text.

n41. For example, of the nearly two hundred reported state appellate decisions citing Restatement 314 since 1945, only three were decided before 1962, and roughly twenty cases each year have made reference to the section since the mid-1980's.

The caveat to section 314A recognized the possibility of such a development ("The Institute expresses no opinion as to whether there may not be other relations which impose a similar duty.").

n42. In some situations, for example, courts have expanded liability by stretching the exceptions to no-duty-to-rescue rules beyond their historical boundaries. See, e.g., Farwell v. Keaton, 240 N.W.2d 217 (Mich. 1976). The Michigan Supreme Court found that a teenager had a duty to aid a friend who had been beaten while the boys were on a social venture. Id. at 222. (The result was also supported by the court's finding that the teenager had undertaken to assist his cohort by placing ice on his head and driving him in a car.)

n43. See, e.g., Weinrib, supra note 5, at 248 n.9 (citing Caldwell v. Bechtel, 631 F.2d 989, 1000 (D.D.C. 1980)) (court reads recent cases to suggest that expanding exceptions based on special relationship "eroding" the general rule). The suggestion that the exceptions to the no-duty-to-rescue rules may swallow the rule ignores the fact that many courts have continued to refuse to find a duty to rescue even where judicially recognized exceptions to the no-duty-to-rescue rules would have allowed a jury to consider the imposition of liability.

n44. Although the alternative proposed here will not eliminate line-drawing problems for courts, it will be able to consider the individualistic values of the society supposedly protected by the traditional rule. See HARPER & JAMES, supra note 24, 18.6. Unlike the traditional rule, however, the proposed approach will be able to give play to other societal interests as well.

n45. See infra notes 65-74 and accompanying text.

n46. See, e.g., GUIDO CALABRESI, THE COSTS OF ACCIDENTS (1970); WILLIAM M. LANDES & RICHARD A. POSNER, THE ECONOMIC STRUCTURE OF TORT LAW (1987). A recent, highly publicized incident illustrates how the existing legal approach may operate as a disincentive to safety. A good samaritan stopped after an accident to pull some people from an overturned car that was spilling gasoline. When individuals involved in the crash were struck by another car, he was subsequently sued for a failure to set out flares. Stephen Braun, Some Hero's Welcome: Rescuers Sued, L.A. TIMES, Nov. 2, 1990, at B1.

n47. The idea that tort law should relieve plaintiffs of burdens by spreading losses through defendants' use of price increases or liability insurance has gained a hold in American courts. For a critical discussion of the trend, see George L. Priest, Modern Tort Law and Its Reform, 22 VAL. U. L. REV. 1 (1987).

n48. There may be little agreement upon when it is "fair" or morally right to impose liability, even among those who see justice as a primary goal of tort law. For example, Oliver Wendell Holmes urged that defendants should pay when it is morally right to do so -- when they have created an unreasonable risk. See generally OLIVER W. HOLMES, THE COMMON LAW (1881). Professor Fletcher propounded a theory of corrective justice basing liability upon "nonreciprocal risk taking." George P. Fletcher, Fairness and Utility in Tort Theory, 85 HARV. L. REV. 537 (1972). Professor Epstein's notion that causation is at the root of questions of when it is fair to impose liability is discussed infra notes 189-96 and accompanying text.

Notions of corrective justice continue to evolve. For example, arguing that tort law cannot be justified by its regulatory or compensatory functions, Professor Catherine Pierce Wells has concluded that an alternative to fault-based liability theory may be advanced by replacing the question, "Is the defendant at fault?" with the query "Is it fair, all things considered, to make this defendant financially responsible for the plaintiff's loss?" Catherine P. Wells, Tort Law as Corrective Justice, A Pragmatic Justification for Jury Adjudication, 88 MICH. L. REV. 2348, 2363 (1988).

n49. Bohlen, supra note 1, at 220.

n50. 159 N.E. 896, 898 (N.Y. 1926). Cardozo's opinion in Moch provides both a good starting point for trying to understand the distinction between nonfeasance and misfeasance and a good example of why such distinctions are difficult to make.

n51. Id.

n52. Harold F. McNiece & John V. Thornton, Affirmative Duties in Tort, 58 YALE L.J. 1272, 1273-82 (1949); See also Weinrib, supra note 5, at 251-58.

n53. Moch, 159 N.E. at 898.

n54. Id. (citing his own opinion in MacPherson v. Buick Motors Co. 111 N.E. 1050 (N.Y. 1916)) Cardozo failed to explain why the auto manufacturer's omission was misfeasance and the water company's omission was not. For a discussion of the tension between Cardozo's approaches in Moch and in Macpherson, see Murphy, supra note 12, at 152-56. Professor Murphy views the cases as representing two strains of duty analysis that have survived to the present -- the Moch strain focussing upon special relationships and the MacPherson strain focussing upon the legal requirement that one must avoid causing foreseeable harm -- irrespective of the existence of special relationships or the distinction between misfeasance and nonfeasance. Professor Murphy's article is significant in that it demonstrates that there would be substantial legal authority for an abandonment of the misfeasance-nonfeasance dichotomy and the relatively recently developed no-duty-to-rescue rules that rest upon it.

n55. Recently, in a closer case, a patient unsuccessfully sought to recover from a physician who proctored the patient's surgery when the proctoring physician failed to intervene during the operation when errors were made. The appellate court found no special relationship exception to the no-duty-to-aid rule. Clarke v. Hoek, 219 Cal. Rptr. 845 (Ct. App. 1985).

n56. What we need to know is not so much the conduct to be avoided when the relation and its attendant duty are established as existing. What we need to know is the conduct that engenders the relation. It is here that the formula, however incomplete, has its value and significance. If conduct has gone forward to such a stage that inaction would commonly result, not negatively merely in withholding a benefit, but positively or actively in working an injury, there exists relation out of which arises a duty to go forward.

Moch, 159 N.E. at 898. On the face of it, the water company's failure to take adequate precautions to insure a sufficient flow of water to hydrants would appear to be more analogous to the surgeon who fails to take adequate precautions to avoid foreseeable infection than to an otherwise uninvolved bystander. Cardozo sheds no light on the rationale for characterizing the company's behavior as a mere withholding of a benefit rather than an omission amounting to injurious conduct.

n57. Weinrib, supra note 5, at 253-54. Professor Weinrib might impose a duty of easy rescue on the bystander in spite of the nonfeasance, however. Id. at 257-58, 283.

n58. Id. Although he would agree with Cardozo that the bystander's behavior is nonfeasance, Professor Weinrib's emphasis in making that determination is slightly different. For example, where Cardozo could conclude that the water company in Moch had not gone far enough in relation to the plaintiff to injure him (rather than deny him a benefit), Weinrib's analysis might yield a different result. Weinrib would ask whether the water company's behavior created the risk that there would be insufficient water at the hydrant at the moment of emergency. Certainly for Weinrib, the water company's conduct must be more analogous to that of the surgeon in our example. However, even Weinrib's distinction may not be as helpful as it first appears. For a discussion of the problems involved in trying to make consistent distinctions between misfeasance and nonfeasance, see infra note 60.

n59. As Harper and Kime noted in discussing the hazards of trying to distinguish between misfeasance and nonfeasance, the analytical technique is simple: the outcome depends upon what point in time is focused upon. Fowler V. Harper & Posey M. Kime, The Duty to Control the Conduct of Another, 43 YALE L.J. 886 (1934). The technique works as nicely in looking at risk-creation. Weinrib focuses on events prior to the defendant's omission. In that sense, an emergency room that failed to admit a patient did not create the need for emergency treatment. On the other hand, at the instant the hospital refused the patient, the potential patient's risk was increased substantially. Modern doctrines of proximate causation and notions of comparative fault recognize that there may be multiple causes for a particular injury and that there is no need to attempt to distinguish one which creates a risk from one that increases a risk.

n60. Such a concern is not new. In Rochester Tel. Corp. v. United States, Justice Frankfurter stated that attempts to distinguish misfeasance from nonfeasance were "mischiefmaking" and, even in 1939, "outmoded." 307 U.S. 125, 142 (1939). The point has been made again recently by Justice Brennan, dissenting in DeShaney v. Winnebago County Dep't of Social Serv., 489 U.S. 189 (1989). In declining to find a constitutional violation by the state, the majority in DeShaney concluded that the state might have been aware of a danger that a child's father posed to the child but argued that the state's failure to protect placed the child in no worse position than he would have been otherwise. Id. at 201. Justice Brennan was critical of the determinative division between action and inaction. "In a constitutional setting that distinguishes sharply between action and inaction, one's characterization of the misconduct alleged under 1983 may effectively decide the case." Id. at 204 (Brennan, J., dissenting).

A classic illustration of the silliness of the distinction is provided in Pridgen v. Boston Hous. Auth., 308 N.E.2d 467, 476-77 (Mass. 1974). The court held that defendant owed a trespasser trapped in an elevator shaft a duty of reasonable care, including reasonable affirmative action.

[I]t is unthinkable to have a rule which would . . . [impose liability if the defendant] pushed the "go" button on the elevator . . . [knowing that plaintiff] was trapped in the elevator shaft, but would not hold it liable if, being reasonably able to do so, the employee knowingly failed or refused to turn off the switch to the electrical power to the same elevator.


n61. See RESTATEMENT, supra note 2, 317. See also Pilgrim v. Fortune Drilling Co., 653 F.2d 982 (5th Cir. 1981) (Employer required employee who lived three hours from work site to work a twelve-hour shift. No duty found to protect employee from off-premises injury.).

n62. See Robertson v. LeMaster, 301 S.E.2d 563 (W. Va. 1983). Under common law rules, the employer had no duty in this situation to control the behavior of the employee acting outside the scope of his employment, but nevertheless the employer's affirmative conduct prior to the accident -- requiring such long hours and providing no transportation home -- created a foreseeable risk of harm.

A recent unreported trial court opinion suggest once again the significance of the court's characterization of the defendant's behavior (and the "mischief-making" nature of the distinction). In Kargul v. Sandpiper Dunes Ltd., WL 28051 (Conn. Super. Ct. Jan. 29, 1991), the court imposed a duty upon a female tenant who was in a relationship with an assaultive male to protect another tenant in the building from being attacked by him. Defendant shared her apartment with a man whom she knew to be a danger to women. She had counseled him while he was incarcerated for sexual offenses and later had become romantically involved with him. Immediately prior to his attack on her co-tenant, she had reason to know of the risk he presented.

Interestingly, the court did not find a special relationship between the defendant and the victim or the defendant and the assailant, but reached the same result by concluding that tenants have a duty to one another not to create an unsafe condition for others on the premises. The court found that the defendant's "act of allowing the [assailant] to reside with her in the same apartment building in which the plaintiff resided, increased the risk of harm occurring to the plaintiff." Id. at 11. See also Pamela L. v. Farmer, 169 Cal. Rptr. 282 (Ct. App. 1980). In Pamela L., the no-duty rule was found inapplicable where defendant had not only failed to prevent harm but allegedly had created a foreseeable risk that her husband would sexually abuse plaintiffs by leaving them alone with him. The court cited RESTATEMENT 302B: "An act or an omission may be negligent if the actor realizes or should realize that it involves an unreasonable risk of harm to another through the conduct of the other or a third person which is intended to cause harm." Id. at 284.

Courts may reach the same result by finding in an affirmative act evidence of an undertaking to protect or of an effort to take charge or exercise control. See Otis Engineering Corp. v. Clark, 668 S.W.2d 307 (Tex. 1983). A divided Texas Supreme Court found an affirmative act in the employer's ordering the employee to leave the work site, followed by accompanying him to his car and trying to ascertain that he was capable of driving. The court concluded that the employer had a duty to exercise control over an incapacitated employee and to prevent him from driving home from work intoxicated.

A number of courts have declined to expand the duty of employers since the Otis decision. See D'Amico v. Christie, 518 N.E.2d 896 (N.Y. 1987) (distinguishing the employer's affirmative act evidencing exercise of control in Otis where employer sent employee home after checking his condition, with the employer's lack of control where employer summarily fired employee for drinking and employee was involved in an accident after leaving the premises). See also Bruce v. Chas. Roberts Air Conditioning, 801 P.2d 456 (Ariz. Ct. App. 1990); Meyers v. Grubaugh, 750 P.2d 1031 (Kan. 1988); Poplaski v. Lamphere, 565 A.2d 1326 (Vt. 1989).

The fine-line distinction between Otis and D'Amico illustrates that the traditional rules may lead a court to focus on minutiae. Should liability really hinge on whether an employer walks an employee to his car before sending him away? Should an on-the-spot dismissal of an employee be what protects the employer from liability? Should the tort law encourage that conduct on an employer's part? In any of these situations, the real question is whether there are strong reasons not to allow a jury to consider the reasonableness of the employer's procedures.

n63. Lather v. Berg, 519 N.E.2d 755 (Ind. Ct. App. 1988); McGee v. Chalfant, 806 P.2d 980, 984-85 (Kan. 1991) (citing Stepnes v. Adams, 452 N.W.2d 256, 259 (Minn. Ct. App. 1990)). But see Leppke v. Segura, 632 P.2d 1057 (Colo. Ct. App. 1981) (tavern owner refuses drinks to intoxicated patron, but liability imposed because employee jump-starts patron's car).

n64. See Holloway v. Martin Oil Serv., Inc., 262 N.W.2d 858 (Mich. Ct. App. 1978) (not foreseeable that gasoline would be used for arson; however, gas station would have a duty to people injured by intoxicated driver if selling gasoline had resulted in use of the auto). Such distinctions lead to some anomalous results. One who provides liquor to an obviously inebriated driver may be immunized from liability while one who provides an automobile to the inebriated driver may be liable under a negligent entrustment theory. See RESTATEMENT, supra note 2, 390: "One who supplies . . . a chattel for the use of another whom the supplier . . . has reason to know to be likely because of his youth, inexperience, or otherwise, to use it in a manner involving unreasonable risk of physical harm . . . is subject to liability. . . ."

n65. The following discussion is based upon the ideas set forth in an article by James Rachels, Active and Passive Euthanasia in ETHICAL ISSUES IN DEATH AND DYING (Tom L. Beauchamp & Seymour Perlin, eds. 1978). Rachels' conclusion that the distinction between active and passive is morally irrelevant is challenged by Tom L. Beauchamp, A Reply to Rachels on Active and Passive Euthanasia, supra at 246. (Beauchamp concludes that the distinction is not always irrelevant, but ultimately agrees that the distinction should probably be dispensed with in the euthanasia context. Id. at 257.)

n66. Rachels, supra note 65, at 241, 242.

n67. Id. at 242.

n68. Id.

n69. See, e.g., Toups v. Hawkins, 518 So. 2d 1077 (La. Ct. App. 1987) (where no security provided, nightclub had no duty to protect patron from being shot).

n70. See, e.g., Harris v. Pizza Hut, Inc., 455 So. 2d 1364 (La. 1984) (duty exists to provide patrons with reasonable protection, once security is provided).

n71. In Craig v. A.A.R. Realty Corp., 576 A.2d 688 (Del. Super. Ct. 1989), the court imposed a duty upon a mall owner to protect the invitee of tenant from abduction from mall and physical harm. The court relied upon the following analysis:

[O]f all the involved parties, the cost of crime reduction is cheapest to the landowner. For the criminal, imposing civil liability on him in addition to existing criminal sanctions does not deter him from committing the crime. Imposing duty on the patron, so that he must protect and compensate himself, may result in crime reduction, but only at the expensive cost of the patron staying home. While the patron can prevent crime by not going out at night, the price of staying home is high not only for him but also for society in general. As opposed to the transient patron, who has little information about the crime problem on the landowner's premises and little ability to directly influence it, the landowner can be much more effective in dealing with the problem. While the patron holds just one expensive option, staying home, the landowner holds many options, ranging from installation of better lighting, fences, or guard service, to even varying hours of operation. All of these options should be less expensive and much more effective in deterring crime than the patron's sole choice of staying home.

Id. at 693 (quoting Michael J. Bazyler, The Duty to Provide Adequate Protection: Landowners' Liability for Failure To Protect Patrons from Criminal Attack, 21 ARIZ. L. REV. 727, 747 (1979)).

n72. See HOLMES, supra note 48.

n73. Even if corrective justice notions depend not upon conduct objectively measured as unreasonable, but upon causation, for example, or even upon broader notions about when it is appropriate to impose liability, it is difficult to imagine persuasive arguments that the restaurant that hired security guards should be responsible for the loss while the other restaurant is not. Id.

n74. See infra notes 90-132 and accompanying text.

n75. Bohlen, supra note 1, at 221.

n76. RESTATEMENT, supra note 2, 351.

n77. The line-drawing problems are also endemic to relationships built upon dependency. See RESTATEMENT, supra note 2, 314A, 320 (concerning relationships between a person in need of protection and a person who might provide such protection).

n78. See supra note 38. Parents continue to be responsible for their children's negligent or intentional misconduct where the parent has reason to know or knows of her ability to control the child and of the need for that control. See, e.g., Basler v. Webb, 544 N.E.2d 60 (Ill. App. Ct. 1989) (six-year-old on bike with no brakes injures another child); Huston v. Konieczny, 556 N.E.2d 505 (Ohio 1990) (affirming denial of defendants' summary judgment motion where passenger of teenage driver suffered injuries after driver allegedly was given alcohol by children of defendants at an authorized unsupervised teenage party at their home).

Employers continue to be responsible for protecting others from their employees in certain situations even while the employees are acting outside the scope of employment. See, e.g., Destefano v. Grabrian, 763 P.2d 275 (Colo. 1989) (diocese may be liable for failing to protect couple in marriage counseling where clergyman providing counseling allegedly induced wife of couple to have sexual relations with him); Kigin v. Woodmen of the World Ins. Co., 541 N.E.2d 735 (Ill. App. Ct. 1989) (camp counselor's employer may be liable for failure to prevent counselor's sexual molestation of camper, even where outside scope of employment). Notice that in Kigin and Destefano, defendants arguably also had relationships with the victims that might require their protection on independent grounds.

Landholders continue to have a similar duty to protect people from individuals who have been permitted on the property. See, e.g., Craig v. A.A.R. Realty Corp., 576 A.2d 688 (Del. Super. Ct. 1989); Moody v. Cawdrey & Assoc., Inc., 721 P.2d 708 (Haw. App. 1986) (court upheld denial of summary judgment motion by condo owner whose guests were assaulted).

n79. E.g., Tarasoff v. Regents of Univ. of California, 551 P.2d 334 (Cal. 1976).

n80. E.g., Johnson v. State, 447 P.2d 352 (Cal. 1968).

n81. See Otis Eng'g Corp. v. Clark, 668 S.W.2d 307 (Tex. 1983) (off-premises accident caused by intoxicated employee returning home from work).

n82. See, e.g., Hasnei v. United States, 541 F.Supp 999 (D. Md. 1982) (psychiatrist had no duty to protect victim from dangerous outpatient); Rayfield v. South Carolina Dept. of Corrections, 374 S.E.2d 910 (S.C. Ct. App. 1988) (duty ends when custody ends; murder only hours after release of prisoner); Sharpe v. South Carolina Dept. of Mental Health, 354 S.E.2d 778 (S.C. Ct. App. 1987) (mental health hospital had no duty to warn of release of murderer). For cases limiting employer liability, see supra note 62.

n83. See generally Harper & Kime, supra note 59.

n84. The Restatement noted that "the relations between the actor and a third person which require the actor to control the third person's conduct are stated in §§ 316-319." RESTATEMENT, supra note 2, 315 cmt. c (emphasis added).

n85. Harper & Kime, supra note 59, at 890. Harper & Kime find actual control, as opposed to respondeat superior or joint enterprise liability, as the theory that explains the imposition of liability in auto cases. See, e.g., Samson v. Aitchison (1912) A.C. 844, 850 quoted at 890-91:

Collins, when he took the wheel came under the control of the defendant. . . If Collins had been going too quickly, and Sampson had told him to go slow, and Collins persisted in going too quickly, Sampson would have had the right to say to him: "If you wish to continue to drive my car you must drive it as I direct, and if you will not do so you must cease to drive it." In such circumstances Collins would have had to obey or stop driving.

n86. Harper & Kime contrasted the actual ability to control with the "fictitious" control of respondeat superior or joint enterprise analysis. Harper & Kime, supra note 59, at 891. However, at times they appeared to be describing a "right to control" rather than a physical ability. Thus, presumably an owner of a car would be held responsible for a burly driver's negligence. On the other hand, a guest would not be held legally responsible for the owner's faulty driving, even if actually able -- either physically or through persuasion -- to eliminate the risk.

Harper and Kime suggest that such efforts would not be required of the guest because the guest "naturally feels reticent about offering suggestions and criticisms of [the owner's] driving. Such back-seat operation of a motor vehicle is not only commonly regarded as discourteous but is invariably resented by the owner of the car." Id. Thus, it appears that at a minimum some social relationship separate from the actual ability to control is necessary to the imposition of liability. The mention of reticence, discourtesy, and resentment raises the possibility that the real defining principle may be nothing more than that a duty to control will be found where it is reasonable to expect such a control to be exercised.

n87. See, e.g., Campbell v. Haiges, 504 N.E.2d 200 (Ill. App. Ct. 1987). In affirming the trial court's summary judgment for the parent defendants, the appellate court in Campbell noted that "the simplest answer to plaintiff's complaint is that the [defendant's parents] did not have the ability to control their child because they were both at work at the time of the alleged attack." Id. at 203.

n88. See, e.g., infra note 99 and accompanying text.

n89. Increasingly, and usually without expressly rejecting the Restatement position, courts have imposed obligations that are based on reasonable care and, therefore, may or may not require control of another. For example, a therapist with no ability to control an outpatient's behavior between sessions might be called upon to warn a potential victim.

Some modern courts have based the duty more broadly upon a right to control. See, e.g., Hasenei v. United States, 541 F. Supp. 999 (D. Md. 1982). (In spite of finding that a special relationship might be based on the right to control another's conduct, the relationship between a psychiatrist and a voluntary patient lacked sufficient elements of control to give rise to a legal duty to protect third parties from violent acts by the patient.) Id. at 1009-10.

n90. The example is based upon several cases decided by the California Supreme Court. The opinions of the court are illustrative, but not unique. Although the California court has given more explicit recognition to policy concerns in making duty determinations than have many courts, each of the California opinions purports also to follow the Restatement rules in making its determination and reaches results that are consistent with decisions in other jurisdictions. See infra notes 126-27.

n91. On the other hand, the conduct of each may be considered a but-for cause of the injury suffered, and artful pleading might characterize defendants' conduct as creating a risk. If the assailant struck on campus, the college itself might be primarily negligent, if not for a failure to protect them from something as mundane as providing cover for assailants by failing to adequately trim foliage. See Peterson v. San Francisco Community College Dist., 685 P.2d 1193 (Cal. 1984) (immunity prevented suit against public institution for failure to provide police protection but cause of action recognized for failure to warn and failure to trim foliage). Id. at 1202.

n92. The advantage to selecting decisions from one jurisdiction is that the differences in result cannot be attributed to differing jurisdictional approaches. The disadvantage is that the selection may appear vulnerable to the criticism that that jurisdiction, particularly California, is simply not representative. However, what is illustrated by these cases is that the traditional special relationship exception to the no-duty rule actually determines the outcome of these cases less and less frequently. And that development is one that clearly is significant elsewhere. See infra notes 126-32 and accompanying text.

n93. 551 P.2d 334 (Cal. 1976).

n94. Id. at 353.

n95. The Lanterman-Petris-Short Act provides that:

[W]hen any person as the result of mental disorder, is a danger to others . . . a peace officer . . . may upon probable cause take, or cause to be taken, the person into custody and place him or her in a facility . . . for 72-hour treatment and evaluation.

CAL. WELF. & INST. CODE §§ 5000, 5150 (West 1976).

n96. The assailant was originally convicted of second degree murder. The conviction was reversed for failure to give proper diminished capacity instructions. See People v. Poddar, 518 P.2d 342 (Cal. 1974).

n97. The court also based its decision upon a determination that society's interest in confidentiality must be subordinated to an interest in public safety. Tarasoff, 551 P.2d at 437. ("The protective privilege ends where the public peril begins.") Although the majority opinion did not systematically analyze societal interests, it quoted Dean Prosser for the proportion that "[Duty] is not sacrosanct in itself, but only an expression of the sum total of those considerations of policy which lead the law to say that the particular plaintiff is entitled to protection." Id. at 342 (citations omitted). The court went on to say that ordinary negligence principles would only be departed from upon the balancing of various policy considerations. Id. See generally infra discussion at Part III. If the court perceived any conflict between the Restatement's rule-based approach and the "policy consideration" approach, it did not acknowledge it. And California courts have continued to rely simultaneously on both approaches. See, e.g., cases discussed infra notes 104-25 and accompanying text.

n98. Id. at 343.

n99. The court did not focus on the therapist's ability to control the patient, nor did it frame the duty that it imposed in terms of control. Instead, the court recognized a duty to "protect" the victim, by warning if necessary. Id. at 343-44. The duty imposed is broader than a duty to control, and it does not seem dependent on the ability to control which, on these facts, did not exist.

n100. Id. at 344, citing Merchants Nat. Bank & Trust Co. v. United States, 272 F. Supp. 409 (D.N.D. 1967). The court's use of authority has been roundly criticized. See, e.g., Alan A. Stone, The Tarasoff Decisions: Suing Psychotherapists to Safeguard Society, 90 HARV. L. REV. 358 (1976); Note, The Dangerous Violet Lee Patient Exception and the Duty To Warn: Creation of a Dangerous Precedent?, 9 U.C. DAVIS L. REV. 549, 551 (1976). The decision has also been critically received by some jurists. See Alegria v. Payonk, 619 P.2d 135, 140 (Idaho 1980) (Shepard, J., dissenting) (bar owner liability).

n101. In all of those Restatement relationships, the actor has either the right or the ability to control the third person's conduct. Thus, in the absence of a relationship involving such control, the exception to the general rule, that there is no duty to control the conduct of a third person for the protection of others, should not be applicable. . . [T]he typical relationship existing between a psychiatrist and a voluntary outpatient would seem to lack sufficient elements of control necessary to bring such relationship within the rule of 315. Indeed, lack of control by the therapist and maximum freedom for the patient is oft times the end sought by both the psychiatric profession and the law.

Hasenei v. United States, 541 F. Supp. 999, 1009-10 (D. Md. 1982) (footnotes omitted).

n102. If the ability to have the student held for seventy-two hours represents control sufficient to establish a special relationship, it is reasonable to ask why the police who actually picked up the student and had physical custody of him did not also have a special relationship with the student that would require of them a duty to warn. Although the passage of time between the defendants' contact with the assailant and his attack was not discussed in Tarasoff, it raises an interesting question about the duration of obligations to warn. See Tresemer v. Barke, 150 Cal. Rptr. 384 (Ct. App. 1978) (Doctor provided patient with I.U.D. Two years later, new information about the risks of I.U.D.s was available. Patient had not seen the doctor since her initial visit, but doctor's failure to warn of subsequent developments was actionable.).

n103. Commentators have suggested various justifications in support of the duty found in Tarasoff. See, e.g., Thomas J. Murphy, Affirmative Duties in Tort Following Tarasoff, 58 ST. JOHN'S L. REV. 492, 527 (1983) (therapist's ability to control a patient results from the fiduciary nature of the relationship with the patient); Note, Professional Obligation and the Duty to Rescue: When Must a Psychiatrist Protect His Patient's Intended Victim, 91 YALE L.J. 1430 (1982) (therapist's obligation arises by virtue of selecting an occupation that is likely to lead to discovery of helpless or endangered persons).

n104. Nally v. Grace Community Church, 763 P.2d 948 (Cal. 1988). The facts of Nally are not those of the example, but give an indication of the court's reluctance to further expand Tarasoff.

n105. Id. at 964 (Kaufman, J., concurring) (Justice Kaufman agreed that nonsuit was properly granted. He argued that defendants did owe a duty of reasonable care to plaintiffs, but that evidence of a breach was lacking.). Id.

n106. Id. at 956. Characteristically, the court also acknowledged that part of the duty determination involved consideration of policy concerns. That the latter is controlling, and is further discussed infra note 107.

n107. Id. at 959. The court observed "[n]either the Legislature nor the courts have ever imposed a legal obligation on persons to take affirmative steps to prevent the suicide of one who is not under the care of a physician in a hospital." Id. (citing Katona v. County of Los Angeles, 218 Cal. Rptr. 19 (1985)); Searcy v. Hemet Unified Sch. Dist., 2 223 Cal. Rptr. 206 (Ct. App. 1986) (school district owes no duty to protect student once student has left for home). In refusing to find a special relationship, the court distinguished (as it had not in Tarasoff) cases where the relationships existed with patients on locked hospital wards. Id. (distinguishing Meier v. Ross General Hosp., 445 P.2d 519 (Cal. 1968); Vistica v. Presbyterian Hosp. & Medical Ctr., 432 P.2d 193 (Cal. 1967)). However, as was noted by Justice Kaufman, the stand was not unequivocal. The court would not "foreclose imposing liability on non-therapist counselors, who hold themselves out as professionals, for injuries related to their counseling activities." Id. at 961 n.8 (cited in Id. at 964 (Kaufman, J., concurring)).

n108. The pastors arguably held themselves out as capable of dealing with suicidal depression, and developed a close counseling relationship with Nally, recognizing his problem. Id. at 964 (Kaufman, J., concurring).

n109. There is also an additional basis for finding a special relationship in this case. Here, the victim and the person to be controlled are one and the same. As a potential victim, Nally had a direct and dependent relationship with the therapist that the victim in Tarasoff, a total stranger, did not.

n110. It would have been difficult to predict the result in Nally based upon a special relationship analysis. In spite of the Nally court's stated reliance on a special relationship analysis, the status of the defendant, the nature of the harm suffered, and the societal burden of imposing liability played a more significant role than did the nature of the relationship that had been established.

Indeed, in Nally far more than in Tarasoff, the court entered into an extended discussion of the policies that led to its conclusion. First, the court expressed concern that if liability were imposed upon unlicensed church counselors, line-drawing would become a problem and other non-professional helpers would be deterred. Second, safety concerns that were present in Tarasoff arguably were absent in a situation involving a suicide. "'[Tarasoff] recognized the importance of the confidential relationship which ordinarily obtains between a therapist and his patient, holding that "the therapist's obligations to his patient require that he not disclose a confidence unless such disclosure is necessary to avert danger to others. . . .'" Id. at 958, (emphasis added) (quoting Bellah v. Greenson, 146 Cal. Rptr. 535 (1983)). In addition, the court was concerned that if liability were imposed upon unlicensed church counselors, there would be a threat to religious liberty. Id.

For an exploration of issues that arise in the area of clergy counselling where the clergy's activity is neither completely secular nor completely religious, see Michael J. Fiorillo, Clergy Malpractice: Should Pennsylvania Recognize a Cause of Action for Improper Counseling by a Clergyman? 92 DICK. L. REV. 223 (1987) (examining both constitutional and "practical" issues in considering liability for clergy counselors). See also Robert J. Basil, Clergy Malpractice: Taking Spiritual Counseling Conflicts Beyond Intentional Tort Analysis, 19 RUTGERS L.J. 419 (1988).

n111. Davidson v. City of Westminster, 649 P.2d 894 (Cal. 1982).

n112. Id. at 895.

n113. Plaintiffs also argued that Ms. Davidson had a special relationship with the police. The court found no special relationship, determining that neither her dependance nor the fact that she was a foreseeable victim were sufficient to create such a relationship. Id. at 900.

n114. Although the court considered the relationship between the victim and the defendants in the Davidson case, the victim in Tarasoff was so far removed from the defendant that the relationship was not even considered.

n115. The police were provided statutory immunity under CAL. GOV'T. CODE 845 (West 1980) for deciding whether or not to make the arrest. However, the immunity did not prevent the court from considering whether the defendants were obligated to warn the victim of the danger.

n116. There is also a question about whether the behavior of the police could be considered misfeasance. If it were found that the police had somehow contributed to the risk that Ms. Davidson encountered, then the need for a discussion of special relationships would be obviated, and the conduct, at least to the extent it was not protected by governmental immunity, would be evaluated by a jury as reasonable or not. See supra note 26.

n117. Davidson, 649 P.2d at 898.

n118. In Davidson, as in Tarasoff and Nally, the court acknowledged that it was considering other factors. As the court stated in Davidson, to recognize an obligation to warn potential victims "would raise difficult problems of causation and public policy." Id. at 900. Had the killer in a case like Tarasoff told only police officers that he was intending to kill a particular young woman but not mentioned his inclination to his therapist, the plaintiffs probably would have recovered nothing. Even where an attack is highly foreseeable, courts tend to protect police departments and other governmental agencies, even where their actions would not be protected by statutory immunities. Some refuse to find liability under special relationship theory, explaining that because the police have a special relationship with all, they cannot have a special relationship with any one individual, and therefore the traditional no-duty-to-rescue rule is in force. See, e.g., Riss v. City of New York, 240 N.E.2d 860, 862 (N.Y. 1968). Even when the victim is identified and has asked for protection, no legal duty exists absent an undertaking. See Schuster v. City of New York, 154 N.E.2d 534 (N.Y. 1958) (holding that a duty existed to protect police informant).

Some courts have been critical of the approach as a form of sovereign immunity that is "outmoded and artificial." See City of Kotzebue v. McClean, 702 P.2d 1309, 1312 (Alaska 1985). See also Gerald P. Krause, Municipal Liability: The Failure to Provide Adequate Police Protection -- The Special Duty Doctrine Should be Discarded, 1984 WIS. L. REV. 499.

Policy concerns are rarely far beneath the surface even in cases that rely upon the Restatement rules. See, e.g., Casey v. Geiger, 499 A.2d 606 (Pa. Super. Ct. 1985). (In spite of foreseeability of an assault on young girl using a city pool, the Casey court found no special relationship to support her recovery). The court's concern was that:

It is questionable how long any municipality could maintain its parks, playgrounds and swimming pools. Due to the cost of increased insurance premiums and added police protection, municipalities will lack the necessary funds to provide recreational services, The end result is that the welfare of the public at large will suffer.

Id. at 614.

n119. 614 P.2d 728 (Cal. 1980).

n120. Id. at 730.

n121. CAL. GOV'T CODE 820.2 (West 1980) states: "Except as otherwise provided by statute, a public employee is not liable for an injury resulting from his act or omission where the act or omission was the result of the exercise of the discretion vested in him, whether or not such discretion be abused."

n122. Thompson, 614 P.2d at 738.

n123. Id. at 734.

n124. In fact, Tarasoff supports the existence of a much broader duty to a potential victim when a relationship exists between one who can protect and the person causing the threat. Tarasoff is based on the existence of a duty of due care to all those foreseeably endangered. Tarasoff, 551 P.2d at 342. Where the victim is identifiable, a warning to that person might discharge the defendant's duty. However, where a foreseeable victim cannot be identified with specificity, other action may be required. See Thompson, 614 P.2d at 739 (Tobriner, J., dissenting). See also Florence L. Di Benedetto, Thompson v. County of Alameda: The Demise of the Special Relationship Doctrine in California, 15 U.S.F. L. REV 563 (1981).

n125. Thompson, 614 P.2d at 734 (citing Rowland v. Christian, 443 P.2d 561 (Cal. 1968)). The court was unwilling to allow a jury to consider the county's conduct in part because to have done so would have so burdened the state's rehabilitation efforts. "Obviously aware of the risk of failure of probation and parole programs, the Legislature has nonetheless as a matter of public policy elected to continue those programs even though such risks must be borne by the public." Id. at 735 (citation omitted).

n126. For example, the Washington Supreme Court read Tarasoff more broadly and consistently with general foreseeability principles. In Petersen v. State, 671 P.2d 230 (Wash. 1983), the court found Tarasoff to support a duty to take reasonable precautions to protect anyone who might foreseeable be endangered by the hospital's patient. Id. at 237. In Peterson, there was no way to warn the victim who was hit by the patient's car after plaintiff had been released from the hospital. See also Semler v. Psychiatric Inst., 538 F.2d 121 (4th Cir. 1975), cert. denied sub nom., Folliard v. Semler, 429 U.S. 827 (1976) (holding that a psychiatric institute had duty to protect public by retaining custody over mental patient ordered confined at institute; institute's release of patient without approval from the court resulted in a killing for which institute and probation officer were held liable); Lipari v. Sears, Roebuck & Co., 497 F. Supp. 185, 192 (D. Neb. 1980); McIntosh v. Milano, 403 A.2d 500 (N.J. Super. Ct. Law Div. 1979).

n127. See supra note 118.

n128. For example, in Lombardo v. Hoag, the New Jersey Superior Court considered the duty of a passenger who relinquished the keys of a vehicle to the vehicle's intoxicated owner. Although the court could have concluded that absent a special relationship the passenger had no obligation to protect the driver or other passengers by denying the keys to him the court ignored the traditional approach. Instead, relying upon a policy against drunk driving, the court required nothing less than reasonable conduct.

An enlightened society should no longer excuse the immoral and outrageous conduct of a person who allows another to drown, simply because he doesn't wish to get his feet wet. Society demands more than that of its citizens. It demands that a person exercise a duty of care towards another person in order to insure that the other person remains free from harm, if he can do so without peril to himself. And it demands an atmosphere in which all persons will expect that others will conduct themselves in such a manner. Defendant Niemeyer had an obligation in the law to do what he could to see that Hoag did not drive his vehicle while intoxicated. And, it is of no particular moment whether we express that obligation in terms of duty, or in terms of proximate cause or foreseeable risk. . . .

The bottom line is that if defendant Niemeyer knew or should have known that defendant Hoag was intoxicated and unable to drive, then he should have done whatever a reasonable person would have done under the circumstances to see that Hoag did not drive his vehicle. It was not difficult conduct that was expected of him, only reasonable conduct. . . . Instead, he drove to his house, arrived safely, and delivered the auto over to Hoag, without any consideration at all of the consequences of his conduct. This behavior, if true, is simply immoral and demands legal sanctions. Those of a higher authority take much too long.

It will be up to a jury to determine whether defendant Hoag was intoxicated and unable to drive, whether Niemeyer knew or should have known of this fact, and whether Niemeyer's conduct in delivering the vehicle over to Hoag was reasonable under the circumstances . . .

566 A.2d 1185, 1189-90 (N.J. Super. Ct. Law Div. 1989).

n129. Indeed, in most situations, state courts are not reading the exceptions to the no-duty rule broadly to find liability, but are reading the exceptions narrowly, so as to preclude it. See, e.g., Kaminski v. Town of Fairfield, 578 A.2d 1048 (Conn. 1990); Heigert v. Riedel, 565 N.E.2d 60 (Ill. App. Ct. 1990); Poplaski v. Lamphere, 565 A.2d 1326 (Vt. 1989); Meyers v. Grubaugh, 750 P.2d 1031 (Kan. 1988); D'Amico v. Christie, 518 N.E. 2d 896 (N.Y. 1987); Lane v. Messer, 689 P.2d 1333 (Utah 1984); McGee v. Chalfant, 806 P.2d 980 (Kan. 1991). However, even when traditional rules lead to predictable results, it is often clear that policy concerns that have little or nothing to do with the relationship of the parties are central to the outcome. See supra note 118.

n130. Farwell v. Keaton, 240 N.W.2d 217 (Mich. 1976).

n131. See, e.g., Kelly v. Gwinnell, 476 A.2d 1219 (N.J. 1984). The New Jersey Supreme Court held that social hosts may be liable for making liquor available to a guest, knowing that the guest is intoxicated and will be driving, when the guest injures someone as a result. Id. at 1224. The court relied heavily upon social policy in reaching its decision.

While we recognize the concern that our ruling will interfere with accepted standards of social behavior; will intrude on and somewhat diminish the enjoyment, relaxation, and camaraderie that accompany social gatherings . . . we believe that the added assurance of just compensation to the victims of drunken driving as well as the added deterrent effect of the rule on such driving outweigh the importance of those other values.

Id. In Kline v. 1500 Massachusetts Ave. Apartment Corp., 439 F.2d 477 (D.C. Cir. 1970) (holding that a duty may be placed upon a landlord to protect tenants in certain situations from foreseeable assaults), the court overcame a host of concerns in departing from the traditional rule in order to encourage "the only one in the position to take the necessary measures" to do so, even if it meant that the costs of additional security would be passed on to tenants. Id. at 488.

See also Mostert v. CBL & Assoc., 741 P.2d 1090 (Wyo. 1987) (Considering a number of factors including the burden to the defendant and safety considerations, the Wyoming Supreme Court imposed a duty not traditionally supported by the Restatement, upon a movie theater to protect invitees from off-premises risk. The theater had failed to inform patrons of storm warnings that had been posted while patrons were watching a film. As a result, one of the patrons drowned after leaving the theater when a flood hit the vehicle she was in.).

n132. In Wright v. Webb, 362 S.E.2d 919 (Va. 1987), for example, the Virginia Supreme Court declined to follow a Restatement special relationship approach which might have led to liability. The use of a policy-based alternative led to a contraction rather than an expansion of liability. In deciding that no duty existed, the court considered "'the likelihood of injury, the magnitude of the burden of guarding against it, and the consequences of placing that burden on the defendant must be taken into account.'" Id. at 921. (citations omitted.) Accordingly, in spite of earlier violent incidents, the court found that a landholder had no duty to protect an invitee from assault by criminals in a parking lot, because both the invitor and the invitee were innocent, and because the burden of providing security would be prohibitive. Id. at 922.

In Lopez v. McDonald's, 238 Cal. Rptr. 436 (Ct. App. 1987), survivors and family members of victims of a mass murder at a McDonald's restaurant claimed that the restaurant's failure to provide adequate security resulted in death and injury. The court recognized the well-established rule in California that the relationship between a business establishment and its customers as a matter of law places an affirmative "duty" on the proprietor to take reasonable precautions to protect patrons from foreseeable criminal acts. Id. at 440. (citing Isaacs v. Huntington Memorial Hosp., 695 P.2d 653, 657 (Cal. 1985); Taylor v. Centennial Bowl, Inc., 416 P.2d 793, 797 (Cal. 1966)). Nonetheless, basing its decision on a policy-oriented approach, the court concluded that the defendant could not be liable as a matter of law, instead of leaving the issue of foreseeability to the jury. Id. at 443 n.6.

Thus, the primary issue here is not whether a fast-food proprietor has a duty to protect plaintiffs from the potential criminal attacks perpetrated by unknown third parties, but rather to determine whether the boundaries of McDonald's general 'duty' encompasses the burden to protect against once-in-a-lifetime massacres. We do this by evaluating this specific event's reasonable foreseeability, or likelihood of occurrence under the circumstances, with applicable policy considerations in resolving whether liability should be restricted.

Id. at 441. (citing Ballard v. Uribe, 715 P.2d 624, 628 n.6 (Cal. 1986)).

n133. See supra notes 60-73 and accompanying text.

n134. See O'Neill v. Montefiore Hosp., 202 N.Y.S.2d 436 (App. Div. 1960) (finding a duty created).

n135. This particular situation serves as a graphic example, but with increasing state and federal regulation of hospitals, such issues may be less likely to be litigated in common law terms. For a discussion of statutory and common law remedies in such cases, see Andrew J. McClurg, Your Money or Your Life: Interpreting the Federal Act Against Patient Dumping, 24 WAKE FOREST L. REV. 173 (1989). The author argues that the common law now imposes duties when there is an undertaking, or when the person seeking treatment attains status of "patient" or has relied upon a custom to furnish emergency services. The author further concludes that because of the hospital's status as a protector, it may have a legal obligation to accept a desperate patient even absent any additional factors.

n136. Chandler v. Hosp. Auth., 548 So. 2d 1384 (Ala. 1989). Summary judgment was denied a hospital where it refused admission of infant with spinal meningitis because the mother was without insurance or the $ 54 emergency room fee. The court found an issue of fact remained where the hospital's published admissions policy presented a 'scintilla of evidence' from which jury could reasonably determine that the hospital had assumed a duty to provide emergency care to indigent patients in emergency situations. The concurring opinion conclude that a duty existed whether it had been "assumed" or not, taking note of congressional disapproval of the no-duty rule manifested in federal "anti-dumping" legislation. Id. at 1387 (Kennedy, J., concurring).

n137. See Delong v. County of Erie, 457 N.E.2d 717 (N.Y. 1983) (duty existed where community had been encouraged to call 911 and victim was assured help would be forthcoming). But see Zepeda v. City of Los Angeles, 272 Cal. Rptr. 635 (Ct. App. 1990) (no common law duty for paramedics to give assistance to shooting victim).

n138. Professor Murphy suggests that the special relationships enumerated in Restatement §§ 316-319 are all undertakings. Murphy, supra note 12, at 171 n.147. In other words, simply by being a parent, for example, one undertakes to control a child in certain situations. Similarly, being a hospital or a 911 service, may very nearly be enough in itself to allow courts to impose obligations to "strangers."

n139. See Sullenger v. Setco Northwest, Inc., 702 P.2d 1139 (Or. Ct. App. 1985) (relationship as consulting physician to osteopath (primary care-giver in this case) did not give rise to a duty to treat child). The fact that physicians do not have such duties has led legislatures to enact "Good Samaritan" legislation in all states, beginning in the late 1950s. See Samuel J. Hessel, Good Samaritan Laws: Bad Legislation, J. LEGAL MED. 40 (May-Jun. 1974); Eric A. Brandt, Good Samaritan Laws -- The Legal Placebo: A Current Analysis, 17 AKRON L. REV. 303 (1983).

n140. Rowland v. Christian, 443 P.2d 561, 567-68 (Cal. 1968).

n141. Although some might raise stare decisis questions about such a development, the discussion in Part I indicates that the law has been moving incrementally toward such a position and, in fact, that much of the recent caselaw development can be better explained by traditional negligence principles than by the common law rules concerning rescue.

n142. In some senses this is a more far reaching proposal than many. Most who have urged change have suggested imposing obligations that are more limited in the affirmative acts they would require. See, e.g., JEREMY BENTHAM, AN INTRODUCTION TO THE PRINCIPLES OF MORALS & LEGISLATION 292-93 (J. Burns & H. Hart eds. 1970); Antony M. Honore, Law, Morals and Rescue, in THE GOOD SAMARITAN AND THE LAW 225 (James M. Ratcliffe ed. 1966); MARSHALL S. SHAPO, THE DUTY TO ACT: TORT LAW, POWER & PUBLIC POLICY xii, 64-68 (1977); Anthony D'Amato, The "Bad Samaritan" Paradigm, 70 NW. U.L. REV. 798 (1975); Ames, supra note 1, at 113; Marc A. Franklin, Vermont Requires Rescue: A Comment, 25 STAN. L. REV. 51 (1972); Robert J. Lipkin, Beyond Good Samaritans and Moral Monsters: An Individualistic Justification of the General Legal Duty to Rescue, 31 U.C.L.A. REV. (1983) [hereinafter Beyond Good Samaritans]; Wallace M. Rudolph, The Duty to Act: A Proposed Rule, 44 NEB. L. REV. 499, 509 (1965); Weinrib, supra note 5.

Some commentators have limited the duty by requiring only that individuals report certain incidents that they witness. See, e.g., Kate E. Bloch, The Role of Law in Suicide Prevention: Beyond Civil Commitment -- A Bystander Duty to Report Suicide Threats, 39 STAN. L. REV. 929 (1987); Mark K. Osbeck, Bad Samaritanism and the Duty to Render Aid: A Proposal, 19 Mich. J.L. Ref. 315 (1985) (proposing a duty to notify police for witnesses of violent crimes) [hereinafter Bad Samaritanism]; Jack Wenik, Forcing the Bystander to Get Involved: A Case for a Statute Requiring Witnesses to Report Crime, 94 YALE L.J. 1787 (1985).

Several commentators have urged an analysis based not upon the misfeasance-nonfeasance distinction, but upon the sort of policy-based analysis suggested by this Article. Murphy, supra note 12. David P. Leonard, The Good Samaritan Rule as a Procedural Control Device: Is it Worth Saving? 19 U.C. DAVIS L. REV. 807 (1986). Professor Leonard rejects the "per se" Good Samaritan Rule in favor of a multi-factor determination of duty based upon the California court's approach in Rowland. Id. at 862-68. However, Professor Leonard undoes much of his proposal by urging that the plaintiff in nonfeasance cases be required to meet a "clear and convincing" burden in proving the causal link between the defendant's conduct and the plaintiff's harm. Id. at 872. In so doing, he urges an approach that might allow additional cases to go to trial but would then require judges to make critical procedural decisions based on the elusive misfeasance-nonfeasance distinction -- an inappropriate focus for decision-making in most cases. See supra notes 49-73 and accompanying text.

n143. 424 N.W.2d 159 (Wis. 1988).

n144. Id. at 160-61. The court categorized the allegations of the complaint as: negligent diagnosis and treatment; failure to warn the patient's family of her condition and its dangerous implications; and failure to commit the patient. Id.

n145. Id. at 165 n.3.

n146. Id. at 166-67.

n147. Id. at 165.

n148. The court's analysis really involves two questions. First, does the behavior of the defendant create an unreasonable risk of foreseeable harm? Second, if it does, is there a policy that will militate against the imposition of a liability? The court has at times even eschewed discussion of "duty." In Walker v. Bignell, 301 N.W.2d 447 (Wis. 1981), the court concluded: "We find it unnecessary to ground our decision relative to the defendants' ultimate liability upon the somewhat nebulous concept of duty even though the parties to this review, the courts below, and many decisions from other jurisdictions have used the term." Id. at 453. See also Murphy, supra note 12, at 166-67. For a similar approach to duty issues where economic harm results from accounting error, see generally Howard B. Wiener, Common Law Liability of the Certified Public Accountant for Negligent Misrepresentation, 20 SAN DIEGO L. REV. 233 (1983). For a study concluding that a policy-based approach can be effectively employed in cases involving pure economic harm, see David A. Fisher, Tort Law: Expanding the Scope of Recovery Without Loss of Jury Control, 11 HOFSTRA L. REV. 937, 961-65 (1983).

n149. Schuster, 424 N.W.2d at 167. (quoting Garrett v. City of New Berlin, 362 N.W.2d 137, 143 (Wis. 1985)). The Schuster court considered societal interests in safety and confidentiality, but did not specifically discuss the factors it identified and particularly (within factors (4) and (6)) the larger concerns that have been expressed about the adoption of such an approach -- the manageability of the duty analysis and the philosophical concerns about the imposition of affirmative obligations to act. See Leonard, supra note 142, at 853. Professor Leonard reviewed the functions of traditional no-duty rules, concluding that in addition to reflecting deeply ingrained values of individualism, they provided predictability and administrative efficiency and urging that any alternative approach must continue to adequately protect the judicial process.

n150. For example, the New Jersey Supreme Court has concluded that the policies underlying the concept of duty include "the economic good of the group, practical administration of the law, justice as between the parties and other considerations relative to the environment out of which the case arose." Suter v. San Angelo Foundry & Mach. Co., 406 A.2d 140, 151 (N.J. 1979) (quoting Leon Green, Duties, Risks, Causation Doctrines, 41 TEX. L. REV. 42, 45 (1962)).

n151. See infra note 154 and accompanying text. The use of such factors raises questions about the relationship between judge and jury. Most notably, both California and Wisconsin include foreseeability as essential to a court's duty determination. At the same time, it is also clearly important to questions more appropriately left to the jury. Recognizing that foreseeability issues permeate questions of both duty (an issue for the court) as well as determinations of "reasonableness" and proximate cause (issues for the jury), and acknowledging some confusion about the role that foreseeability should play in those determinations, the California Supreme Court has attempted to provide some guidance to courts. The court has suggested that a judge's duty determination should focus on "whether the category of negligent conduct . . . is likely to result in the kind of harm experienced." (emphasis added). Ballard v. Uribe, 715 P.2d 624, 628 n.6 (Cal. 1986). On the other hand, in considering fault and proximate cause issues, the jury may consider whether the "particular" defendant's behavior made foreseeable a "particular" plaintiff's harm. Id.

n152. Ballard, 715 P.2d at 628, citing Rowland v. Christian, 443 P.2d 561 (1968) and CAL. CIV. CODE 1714 (West 1980).

n153. Id.

n154. Id. at 628 n.6 (citations omitted). Rowland listed as major policy factors: "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." Rowland, 443 P.2d at 561.

In fact, California courts have adjusted the list of policy factors to be considered, depending upon the circumstances. For example, where a duty to prevent pure economic harm has been the issue, the California Supreme Court has referred to slightly f different factors than it has in cases where landholder obligations are in question. See Biakanja v. Irving, 320 P.2d 16, 19 (Cal. 1958) ("The determination whether in a specific case the defendant will be held liable to a third person not in privity is a matter of policy and involves the balance of various factors, among which are the extent to which the transaction was intended to affect the plaintiff, the foreseeability of the harm to him, the degree of certainty that the plaintiff suffered injury, the closeness of the connection between the defendant's conduct and the policy of preventing future harm.").

When the defendant has been a governmental agency, the court has acknowledged that additional policies will be addressed. See, e.g., Raymond v. Paradise Unified Sch. Dist., 31 Cal. Rptr. 847, 852 (1963) (where public agency is a defendant, additional elements include: the extent of agency powers, legal limits upon the role of the agency, and budgetary constraints).

Some courts have interpreted Rowland's "consequences to the community" factor to include the "administrative factor" -- the difficulties of fashioning a rule and applying it -- addressed by professor Green. E.g., Soldano v. O'Daniels, 190 Cal. Rptr. 310, 316 (Ct. App. 1983), citing Leon Green, The Duty Problem in Negligence Cases, 28 COLUMB. L. REV. 1014, 1035-45 (1929) reprinted in LEON GREEN, THE LITIGATION PROFESSOR IN TORT LAW: NO PLACE TO STOP IN THE DEVELOPMENT OF TORT LAW 174-75 (2d ed. 1977). Green found policy to be determined by the "administrative factor," i.e., institutional concerns of the courts, and the "ethical or moral," "economic," "prophylactic or preventive" and "justice" factors. Green, supra, at 1034. See also Leon Green, Tort Law Public Law in Disguise, 38 TEX. L. REV. 1 (1959).

Several commentators have proposed variations upon the California approach. See generally Weiner, supra note 148 and Leonard, supra note 142. Ultimately, however, courts should be able to adapt their existing approaches to the rescue context. Rigorously applied, the existing approaches should develop sufficiently to meet the concerns expressed by supporters of the traditional rule-based approach. See generally infra discussion at part B. See also infra note 239.

n155. See Weiner, supra note 148, at 254 n.5 (citing Rowland, 443 P.2d at 564).

n156. For example, while California has been willing to consider the availability of insurance as a factor, other jurisdictions have been critical. Oregon, for example, has refused to consider the availability of life insurance in determining liability. "[A] person's liability in our law still remains the same whether or not he has liability insurance; properly, the provision and cost of such insurance varies with potential liability under the law, not the law with the cost of insurance." Winn v. Gilroy, 681 P.2d 776, 784 n.9 (Or. 1984). Such issues cannot be discussed until they are made explicit.

n157. In each case, a special relationship exists. Yet the results in the three cases would not necessarily be the same. See supra notes 90-125 and accompanying text.

n158. That is not to say that there will not continue to be contradictions and uncertainties as the law evolves. In fact, at times the court has failed to engage in any in-depth consideration of policy factors, even while acknowledging their importance. See generally Tarasoff v. Regents of Univ. of Cal., 551 P.2d 334 (Cal. 1976); Sprecher v. Adamson Co., 636 P.2d 1121 (Cal. 1981). But for more recent decisions, attempting a more systematic discussion of the Rowland policy factors, see, e.g., Nally v. Grace Community Church, 763 P.2d 948 (Cal. 1988); Soldano v. O'Daniels, 190 Cal. Rptr. 310 (Ct. App. 1983). To the extent that courts rigorously apply the policy-based approach, they will be required to develop law that is based upon the concerns that actually determine their cases' outcomes rather than upon fine-line distinctions about whether the behavior is nonfeasance or not.

n159. For an example of the considerations that might be argued even where conduct could legitimately be considered nonfeasance, see majority and dissenting opinions in Strauss v. Belle Realty Co., 482 N.E.2d 34 (N.Y. 1985) (holding that defendant Consolidated Edison's conduct was misfeasance).

n160. The ethical-legal issues involved cannot be adequately addressed within the vocabulary of the traditional rules. For a discussion of the competing concerns involved, see Bruce A. McDonald, Ethical Problems for Physicians Raised by Aids and HIV Infection: Conflicting Legal Obligations of Confidentiality and Disclosure, 22 U.C. DAVIS L. REV. 557 (1989).

n161. See, e.g., In re Bouvia v. Superior Court, 209 Cal. Rptr. 297 (Ct. App. 1986); In re Farrell, 529 A.2d 404 (N.J. 1987) (recognizing the competent patient's self-determination interests in refusing life sustaining treatment). See A. Samuel Oddi, The Tort of Interference with the Right to Die: The Wrongful Living Cause of Action, 75 GEO. L.J. 625 (1986) (the author argues against the use of good samaritan rules where a patient has not consented to treatment).

n162. See infra notes 209-15.

n163. A line of authority for a general duty to aid or protect that predates the development of the Restatement rules and dates back to Lord Escher's words in Heaven v. Pender:

[W]henever one person is by circumstances placed in such a position with regard to another that every one of ordinary sense who did think would at once recognize that if he did not use ordinary care and skill in his own conduct with regard to those circumstances he would cause danger or injury to the person or the property of the other, a duty arises to use ordinary care and skill to avoid such danger.

11 Q.B.D. 503, 509 (1883). See generally Murphy, supra note 12.

n164. See, e.g., cases discussed supra notes 131-32 and accompanying text.

n165. See generally Neil M. Levy and Edmund Ursin, Tort Law in California; At the Crossroads, 67 CAL. L. REV. 497 (1979) (reviewing developments in California Law during the 1960s and 70s, and concluding that the court's interest in loss distribution, especially, led to dramatic changes in the court's approach to tort liability).

n166. Id. at 505-15 and cases cited therein.

n167. Rowland v. Christian, 443 P.2d 561, 566 (Cal. 1968) (citing Fernandez v. Consolidated Fisheries, Inc. 219 P.2d 73 (1950)).

n168. See supra note 154.

n169. Id. See also Kermarec v. Campagnie Generale Transatlantique, 358 U.S. 625, 630-31 (1959) (Refusing to adopt the common law classifications for the law of admiralty. "Through this semantic morass [confusing application of the rules] the common law has moved, unevenly and with hesitation, towards "imposing on owners and occupiers a single duty of reasonable care in all circumstances.") (footnotes omitted).

n170. Rowland, 443 P.2d at 567.

n171. Id.

n172. Ironically, although Rowland addressed only landholder obligations, the case could have been analyzed as a failure to aid or protect, involving as it did, a tenant's failure to warn of danger in her apartment. Murphy, supra note 12.

n173. 636 P.2d 1121 (Cal. 1981).

n174. All parties agreed that human activity did not contribute to the movement of the land. Id. at 1121-22.

n175. Id. at 1122-23. Although most courts had not taken such bold steps, a number had been willing to extend traditional rules concerning a possessor's liability for harm to an adjoining landowner resulting from natural conditions when those rules conflicted with general principles of negligence law. See, e.g., Dudley v. Meadowbrook, Inc., 166 A.2d 743, 744 (D.C. 1960) (possessor of land has a "duty of common prudence in maintaining his property, including trees thereon, in such a way as to prevent injury to his neighbor's property"). See also Mahurin v. Lockhart, 390 N.E.2d 523 (Ill. App. Ct. 1979); Kurtigian v. City of Worcester, 203 N.E.2d 692 (Mass. 1965); Barker v. Brown, 340 A.2d 566 (Pa. Super. Ct. 1975).

n176. Sprecher, 636 P.2d at 1130.

n177. Id. at 1127.

n178. Id. at 1125-26 citing Fleming James Jr., Scope of Duty in Negligence Cases, 47 N.W.U.L. REV. 778-809 (1953); Dix W. Noel, Nuisances from Land in its Natural Condition, 56 HARV. L. REV. 772, 796-97 (1943); Glenn A. McCleary, The Possessors' Responsibilities as to Trees, 29 MO. L. REV. 159 (1964); HARPER & JAMES, supra note 24, 27.19, at 1521-22 (1956).

n179. Sprecher, 636 P.2d at 788. In a footnote, the court did suggest that the act of possession could be equated to an act preventing an act of rescue because possession forecloses possession by another who might abate the condition. The court noted that while there may be no liability for failing to throw a rope to a drowning victim, liability does exist for hiding a rope so as to prevent another from rescuing. Id. at 789 n.8. For a discussion criticizing the court's suggestion that ownership may be equated with the act of preventing a rescue, see David W. Burcham, Sprecher v. Adamson Companies: Nonfeasance Immunity Slides By the California Supreme Court, 16 LOY. LA. L. REV. 625, 635 (1983). The California Supreme Court has recently extended liability even further, holding that in the case of commercial landlords possession alone may in some situations result in the imposition of strict liability principles. See Becker v. IRM Corp., 698 P.2d 116 (Cal. 1985).

n180. Notably, such an opinion goes far beyond the easy rescue proposals so frequently made. In Sprecher, the problem was a long term condition rather than an emergency situation and abatement of the condition might have been enormously costly to the defendant.

n181. 190 Cal. Rptr. 310 (Ct. App. 1983).

n182. Applying the Rowland factors, the court found: the harm in the case was not only foreseeable, but imminent; the injury was certain; there was arguably a connection between the bartender's refusal to help and the death; there was a callous disregard for life; the burden of a phone call was minimal; and imposition of a duty would promote safer conduct by requiring citizens not to hamper good samaritans. The court limited its holding out of a concern for a workable rule -- explicitly not requiring "a citizen to open up his or her house to a stranger." Id. at 315-16.

The Soldano court's break from the common law was tempered somewhat by the court's conclusion that the facts very nearly came within RESTATEMENT 327 which states: "[o]ne who knows or has reason to know that a third person is giving or is ready to give another aid necessary to prevent physical harm to him, and negligently prevents or disables the third person from giving such aid, is subject to liability for physical harm caused to the other by the absence of the aid which he has prevented the third person from givin."

n183. On the contrary, some courts have been critical of Soldano. See, e.g., Clarke v. Hoek, 219 Cal. Rptr. 845 (Ct. App. 1985) ("Appellant attacks the rule that a person may not be liable in tort for nonfeasance absent the existence of a special relationship, urging that this rule has been discarded in the recent case of Soldano v. O'Daniels. Appellant's attempt to declare the rule dead and buried is premature, at best."). Id. (citations omitted). And shortly after Soldano, in the rescue area, the California Supreme Court continued to apply both a policy-based duty test and a traditional rule-based approach simultaneously. See, e.g., Williams v. State, 664 P.2d 137 (Cal. 1983) ("[One] who has not created a peril is not liable in tort merely for failure to take affirmative action to assist or protect another unless there is some relationship between them which gives rise to a duty to act"). Yet the Soldano case itself illustrates again the frailty of even indirect reliance upon the traditional rules. Should the result really be dismissal of the case if the bartender refuses when the samaritan asks him to dial 911, and potential liability if he refuses when the samaritan asks to dial his own call?

n184. See, e.g., Bad Samaritanism, supra note 142, at 329, (citing Robert Nozick, Anarchy, State and Utopia ix; Hospers, What Libertarianism Is, in THE LIBERTARIAN ALTERNATIVE at 12-13 (Machan ed. 1974)).

n185. See, e.g., Epstein, supra note 19, at 198-201.

n186. The importance of manageability has been underscored recently by the California Supreme Court in Thing v. LaChusa, 771 P.2d 814 (Cal. 1989). The court established bright line rules to limit the number of plaintiffs who could recover for the negligent infliction of emotional distress.

Reversing the court of appeal, the Supreme Court denied recovery to a mother who had not actually witnessed the accident that injured her child. The court weighed:

the impact of arbitrary lines which deny recovery to some victims whose injury is very real against that of imposing liability out of proportion to culpability for negligent acts. We also weigh in the balance the importance to the administration of justice of clear guidelines under which litigants and trial courts may resolve disputes.

Id. at 827. See also Elden v. Sheldon, 758 P.2d 582 (Cal. 1988) (limiting recovery for negligently caused emotional distress so as to exclude non-married cohabitants, and expressing similar administrative concerns).

n187. Although the proposal made by this Article would expand the obligation to rescue in some situations, expanded obligations would not result in all instances. As has been noted, a determination that the defendant has acted unreasonably does not automatically result in liability. "Liability does not necessarily follow even when negligence and negligence as a cause-in-fact of injury are present. Public policy considerations may preclude liability." Schuster v. Altenberg, 424 N.W.2d 159, 166 (Wis. 1988) (citations omitted).

However, even a court that understands that such an approach does not abandon all limits upon liability, still must be convinced that its approach is practical. In Soldano, the California Court of Appeal used the Rowland approach to impose a duty where the traditional rules would have barred the action. Before it could do so, however, it acknowledged the "pragmatic concern of fashioning a workable rule and the impact of such a rule on the judicial community," Soldano, 190 Cal. Rptr. at 316, and was satisfied that its action "would not involve difficulties in proof, overburden the courts or unduly hamper self-determination or enterprise." Id.

n188. The arguments concerning causation and liberty have been made in various forms by advocates of a limited duty to rescue and are treated summarily here. The survival of the traditional rules cannot be explained on liberty or causation grounds. However, the process concerns are very real to courts. See supra note 20.

n189. Professor Bohlen described the distinction thusly, "In the case of active misfeasance the victim is positively worse off as a result of the wrongful act. In cases of passive inaction plaintiff is in reality denied protection which, had it been afforded him, would have benefited him." Bohlen, supra note 1, at 220; see also Ames, supra note 1.

n190. A related argument is that an event must be a positive occurrence to cause another. Because allowing harm to occur is a failure to act rather than a positive occurrence, it is by definition "negative causation" and cannot therefore be the basis of liability. See, e.g., Epstein, supra note 19, at 194-95; Eric Mack, Bad Samaritanism and the Causation of Harm, 9 PHIL. AND PUB. AFF. 230 (1980). For arguments that omissions can be considered the cause of harm, see, e.g., JOEL FEINBERG, HARM TO OTHERS 165, 185 (1984), HEBERT L. A. HART & ANTHONY HONORE, CAUSATION IN THE LAW 24-47, 58-64 (1959). Feinberg points out that if it is not always unfair to impose liability without causation, the argument that nonfeasance cannot be a "positive occurrence," is moot. For a more in-depth discussion of these two versions of causation arguments against liability and the relationship between them, see FEINBERG, supra, at 165-81.

n191. J. HALL, GENERAL PRINCIPLES OF CRIMINAL LAW 195-96 (1960). See also Arthur Leavens, A Causation Approach to Criminal Omissions, 76 CAL. L. REV. 547 (1988) (rejecting, in the criminal context, the distinction between acts and omissions in favor of inquiring simply whether the act or omission caused the proscribed harm).

n192. See, e.g., Hicks v. United States, 368 F.2d 626 (4th Cir. 1966) (finding defendant in medical malpractice case answerable where traditional cause in fact could not be shown if defendant had destroyed even a substantial possibility of patient's survival); Sindell v. Abbott Laboratories, 607 P.2d 924 (Cal. 1980) (holding that case against drug manufacturers survives plaintiff's inability to prove cause in fact); Haft v. Lone Palm Hotel, 478 P.2d 465 (Cal. 1970) (finding that safety concerns require a shift in burden of proof where defendant's failure to act deprives plaintiff of witness necessary to a proof of cause in fact). Several commentators have recognized the role that policy plays in deliberations about cause in fact. See, e.g., E. Wayne Thode, The Indefensible Use of the Hypothetical Case to Determine Cause in Fact, 46 TEX. L. REV. 423 (1968).

n193. Weinrib, supra note 5, at 260; Bad Samaritanism, supra note 142, at 329; Beyond Good Samaritans, supra note 142, at 267-69.

n194. The court overruled the trial court's sustaining of the therapist's demurrer without leave to amend and remanded. Tarasoff, 551 P.2d at 353.

n195. See RESTATEMENT, supra note 2, 323 cmt c.

n196. RESTATEMENT 323 comment e suggests that there may be cases "in which one who has entered on performance of his undertaking, and cannot withdraw from it without leaving an unreasonable risk of serious harm to another, may be subject to liability even though his conduct has induced no reliance and he has in no way increased the risk."

n197. The argument has led proponents of an expanded duty to limit the scope of their proposals. See, e.g., Weinrib, supra note 5, at 268. (duty to rescue limited to emergency situations where no inconvenience would be caused by imposition of the legal obligation).

n198. See McNiece & Thornton, supra note 52, at 1287-88. Epstein, supra note 19, at 198. The view may be rooted in the view of Americans as independent individualists. See, e.g., HARPER & JAMES, supra note 24, 18.6. See also Rosenberg, supra note 5, at 12-13. Professor Rosenberg argues that even a required easy rescue violates liberty interests, but acknowledges that other interests may outweigh "values of privacy and personal liberty." Id. at 13.

n199. Perhaps an even more pertinent question is whether the proposed change in the law would be more burdensome than are the common law no-duty-rescue rules that now exist.

n200. The value of liberty which underlies the traditional rules need not be ignored in the approach proposed by this Article, but it cannot be considered without giving consideration to competing interests.

n201. "[W]hen a government requires a person to act, it is necessarily interfering more seriously with his liberty than when it places limits on his freedom to act -- to make man serve another is to make him a slave, while to forbid him to commit affirmative wrongs is to leave him still essentially a free man." Robert L. Hale, Prima Facie Torts, Combination, and Non-Feasance, 46 COLUM. L. REV. 196, 214 (1946) (emphasis included) (describing individualism).

n202. See JOEL FEINBERG, SOCIAL PHILOSOPHY 26-31 (1973). This "harm principle" might not be accepted by a "pure" individualist, but a "reflective" individualist will recognize that he is responsible primarily to himself but that, because of his interrelationships, he has some duty to others as well. Id. Of course it may be argued that benefits accrue to each of us if others are taking additional affirmative steps to provide safe environments. The argument has been made, therefore, that even individualists might, consistent with their principles, support an expanded duty to rescue. Beyond Good Samaritanism, supra note 142, at 252.

n203. As previously noted, even if there are situations where most people would agree that behavior was nonfeasance (for example, a bystander comes upon an accident and fails to help), the overwhelming majority of cases cannot be so easily characterized.

n204. FEINBERG, supra note 190, at 231-34; Bad Samaritanism, supra note 142, at 336.

n205. Feinberg finds persuasive the argument that affirmative duties make planning difficult. FEINBERG, supra note 190, at 212. See also Bad Samaritanism, supra note 142, at 335.

n206. FEINBERG, supra note 190, at 163-65.

n207. See Ferrer v. Harris, 434 N.E.2d 231 (N.Y. 1982) (holding that failure to instruct jury on emergency doctrine was error where child suddenly stepped in front of defendant's vehicle).

n208. See FEINBERG, supra note 190, at 165; Beyond Good Samaritans, supra note 142, at 277.

n209. Most European countries impose liability for a failure to rescue. See generally Aleksander W. Rudzinski, The Duty to Rescue: A Comparative Analysis, in THE GOOD SAMARITAN AND THE LAW 91 (James M. Ratcliffe, ed. 1981); Kristin A. DeKuiper, Stalking the Good Samaritan: Communists, Capitalists and the Duty to Rescue, 1976 UTAH L. REV. 529.

n210. See, e.g., MASS. GEN. LAWS. ANN. ch. 268, 40 (West 1990); MINN. STAT. ANN. 604.05(1) (West 1983); R.I. GEN. LAWS 11-37-3 (1981); VT. STAT. ANN. tit. 12, 519(a) (1973).

n211. For example, all states now have statutes that require reporting child abuse. Some impose a legal obligation upon "any person" who suspects abuse. Mary H. Mitchell, Must Clergy Tell? Child Abuse Reporting Requirements Versus the Clergy Privilege and Free Exercise of Religion, 71 MINN. L. REV. 723, 729 (1987) (citing IRVING J. SLOAN, CHILD ABUSE: GOVERNING LAWS AND LEGISLATION at 24-25 (Legal Almanac Series No. 79, (1983))).

However, the existence of such statutes does not always result in civil liability for a failure to report abuse. See Fischer v. Metcalf, 543 So. 2d 785 (Fla. Dist. Ct. App. 1989) (finding no civil cause of action for failure of a psychiatrist to report abuse of children by the psychiatrist's children where statute made such conduct a misdemeanor).

n212. "[R]escuing a person in immediate danger of injury or death as the result of fire, drowning, or other catastrophe also benefits the entire public." CAL. GOVT. CODE 53100 (West 1983 & Supp. 1984).

n213. CAL. PENAL CODE 384 (West 1980 & Supp. 1984). For a review of statutes that suggest that the law is moving toward expanded affirmative obligations, see Teresita Rodriguez, I Am My Brother's Keeper: A Trend Towards Imposing a General Duty Upon a Bystander to Assist a Person In Danger, 26 B.C. L. REV. 497, 507-14 (1985). For a discussion of California's legislative efforts on behalf of victims, see James E. Culhane, California Enacts Legislation to Aid Victims of Criminal Violence, 18 STAN. L. REV. 266 (1965). See also Soldano v. O'Daniels, 190 Cal. Rptr. 310 (Ct. App. 1983). Soldano cited a number of statutes to support its expansion of duty to rescue. Id. at 315.

n214. "No person shall operate a private passenger motor vehicle on a highway unless that person and all passengers four years of age or over are restrained by a safety belt. . . ." CAL. VEH. CODE 27315(d) (West 1990).

n215. See id. at 20001 (West 1971).

n216. Some who would concede that the interest in individual autonomy may at times give way to safety, or other, interests, nevertheless may be concerned that the proposal will involve courts in policy decisions that are better left to legislatures. Although the concern about the relationship cases between the judicial and legislative branch is a serious one, it is a concern that is common to all cases involving affirmative obligations -- whether a case is resolved using traditional no-duty rules or the proposed negligence approach. A court that articulates the policies that are guiding its decision is not necessarily more "legislative" than a court which is guided by the same concerns but leaves them unarticulated. Further, it is not unprecedented in this area of the law for the legislature to overrule a court's decision. For example, in Coulter v. Superior Ct, 577 P.2d 669 (Cal. 1978), the California Supreme Court imposed a duty on social hosts. Shortly thereafter, however, the legislature immunized both social hosts and commercial enterprises (in most situations) from liability for harm suffered as the result of serving alcohol. CAL BUS. & PROF. CODE 25602 (West 1985).

It should be clear that neither the common law rules nor the more recent cases which have dealt with rescue issues have ever considered libertarian principles to be solely determinative. The proposed approach would allow courts to consider libertarian values that are represented by these arguments, but would not necessarily treat them as controlling in every situation.

n217. Epstein, supra note 19, at 200-01; laws already require individuals to contribute to charity quite directly, authorizing money raised through taxation to be spent for food stamp benefits to low-income individuals and emergency relief aid to famine victims. Bad Samaritanism, supra note 142, at 335-36.

n218. Epstein, supra note 19, at 200-01. For a discussion of Epstein's more significant point -- that a general duty would make it impossible to limit societal interference with individual freedoms and would result in forced exchanges -- see infra note 240 and accompanying text.

n219. Epstein, supra note 19, at 200-01.

n220. Weinrib shares a concern that any attempt to require more than an easy rescue would obscure the distinction between "the praiseworthy and the required." Weinrib, supra note 5, at 281. However, the choice is not between an absolute duty to behave altruistically and an easy duty to rescue. Rather, under the proposal made herein, a duty to behave reasonably will be imposed unless there are interests that would prompt the court to reject liability in spite of the unreasonable conduct. Thus, the court will not be without limiting power. The line will have to be drawn between the reasonable and the "praiseworthy." See supra notes 140-64 and accompanying text.

n221. Although it is difficult to measure how much changes in the law affect action, there is some suggestion that social opprobrium for failure to undertake a reasonable effort to aid another would increase were the law to change. See Jay Silver, The Duty to Rescue: A Reexamination and Proposal, 26 WM. & MARY L. REV. 423, 429 (1985), citing D'Amato, supra note 142, at 809 (taking the position that law enforcement shapes moral attitudes), and Harry Kaufmann, Legality and Harmfulness of a Bystander's Failure to Intervene as Determinants of Moral Judgment, in ALTRUISM AND HELPING BEHAVIOR: SOCIAL PSYCHOLOGICAL STUDIES OF SOME ANTECEDENTS AND CONSEQUENCES 77-81 (Jacqueline R. Macauley & Leonard Berkowitz, eds. 1970) (study suggesting that a legal obligation to rescue would reinforce the notion that it would be immoral to passively fail to attempt certain rescues).

n222. Some commentators have suggested that a broader duty to aid or protect would not have the desired safety effect. Landes and Posner tentatively suggest that increased imposition of liability for non-rescue might possibly prompt potential rescuers to avoid activities that might require rescues. Landes & Posner, supra note 19, at 120. Others have suggested that expanded duties could increase risk not by prompting avoidance but by prompting involvement (thus encouraging officious intermeddlers). See, e.g., Weinrib, supra note 5, at 281. In addition to the dangers that might accompany "intermeddler" problems, some have posited that criminals would take advantage of a legally mandated duty to "lure" good samaritans into traps by pretending to be injured or in danger. See D'Amato, supra note 142, at 811 n.47. But the obligation to act reasonably under the circumstances is not an unlimited obligation to rescue and should not require or encourage intervention that increases risk to the victim, the intervening party or others -- something that should be easily and intuitively grasped.

n223. See supra note 20.

n224. PROSSER & KEATON, supra note 5, at 56. Henderson, supra note 20.

n225. Henderson argues that courts' refusal to adopt a general duty to rescue cannot be explained on substantive grounds. He characterizes the theorists who have opposed duties to rescue as comprising three groups. The first group argues that fairness concerns are or should be primary. See, e.g., Epstein, supra note 19, at 151; George P. Fletcher, Fairness and Utility in Tort Theory, 85 HARV. L. REV. 537 (1972). The second group argues that efficiency concerns are or should be determinative. See, e.g., Richard A. Posner, The Ethical and Political Basis of the Efficiency Norm in Common Law Adjudication, 8 HOFSTRA L. REV. 487 (1980). The third believes that some combination of efficiency and fairness concerns guide the development of the tort law. See, e.g., GUIDO CALABRESI, THE COSTS OF ACCIDENTS (1970). Instead, Henderson concludes, constraints of process "are what have made the existing rules so durable." Henderson, supra note 20, at 901.

n226. Henderson, supra note 20, at 904-05, 911-15.

n227. See Id. at 905-18. Henderson concedes that the process constraints are not absolute. Nor will they all necessarily support a particular rule choice. Rather "they must balance against each other and against the substantive objectives they are intended to complement." Id. at 911. Henderson's process analysis fails to explain the movement away from the traditional rules. Henderson concedes there is such a movement, taking the position that the no-duty-to-rescue rule has nearly been swallowed by exceptions. Id. at 928.

n228. See, e.g., Pokora v. Wabash Ry. Co., 292 U.S. 98 (1934). In Pokora the Supreme Court returned to reasonable person standard only seven years after supplanting it with a judicial rule of conduct for persons crossing railroad tracks in Baltimore & Ohio R.R. Co. v. Goodman, 275 U.S. 66 (1927).

n229. Henderson, supra note 20, at 920.

n230. Even assuming that Henderson's process objections are valid, they may not outweigh certain positive process results that would accompany abandonment of the Restatement rules together with the substantive benefits that would be derived from the change.

n231. Id. at 934 n.163. Of course, cases have faced similar problems under existing law. As a result, a therapist must not act to protect a possible victim of the therapist's client unless the therapist knew or under objective standards should have known of the danger. Presumably, the same limitation could be effective in all rescue situations. Unless the potential rescuer should reasonably have known of the danger, no liability would attach. See Tarasoff v. Regents of Univ. Of Cal., 551 P.2d 334, 345 (Cal. 1976). Indeed in some senses, the person observing the couple in the back seat has a simple task. If she is able to determine there is a danger, it is immediate. There is a foreseeable victim and a foreseeable type of harm that may not be present in many misfeasance cases.

n232. See Bad Samaritanism, supra note 142, at 340 (discussion of Henderson's example in context of proposal for a duty to notify authorities). It should be noted that courts and commentators have begun to recognize that men and women may see such a situation differently. See Ellison v. Brady, 942 F.2d 872 (9th Cir. 1991) (holding that reasonable woman standard is employed in sex discrimination case brought under Title VII of the 1964 Civil Rights Act). See also Nancy S. Ehrenreich, Pluralist Myths and Powerless Men: The Ideology of Reasonableness in Sexual Harassment Law, 99 YALE L.J. 1177, 1207-08 (1990); Stephanie Wildman, The Rise of the Reasonable Woman, S. F. BANNER D.J., Mar. 11, 1991, at 4.

n233. Henderson argues that duties imposed under the special relationship exceptions to the no-duty rule are reasonably focused enough to avoid process difficulties. Henderson, supra note 20, at 941 n.193. But the Tarasoff case is a good example of a fact situation which belies the contention. The Tarasoff duty is very broad -- a duty to protect the victim of the ex-patient -- in any way that is reasonable.

n234. Henderson argues that where there has been an undertaking to rescue, the duty of the potential rescuer is limited to a duty to carry through with the specific course of action that he has undertaken and by the fact that the residue was achievable. However, under common law principles, a person who does undertake a rescue might be found negligent for failing to do more under the circumstances. In fact, an initial undertaking is not always a clearer indication of the scope of the duty than a general reasonableness standard would be. Whether in fact a general standard would require any additional effort or change the result in such a case seems unlikely.

n235. See generally Annotation, Violation of Statute Requiring One Involved in an Accident to Stop and Render Aid as Affecting Civil Liability, 80 A.L.R.2d 299 (1961).

n236. Henderson acknowledges that the common law rule that requires an individual to aid another whose injury he innocently has caused appears to have the process difficulties that are of concern to him since there is nothing to guide the behavior of the potential defendant in that situation. He argues that courts relying upon the rule may really be deciding the cases on alternate grounds -- for example, a theory that highway travelers share a special relationship. See, e.g., Montgomery v. National Convoy & Trucking Co., 195 S.E. 247 (S.C. 1938). But if a special relationship explains such cases, why not acknowledge the special relationship between all persons? Surely it is not ownership of a vehicle that produces a special bond. Certainly a special relationship rationale that includes those on the highways, but not those on trails or on horseback cannot be touted as a model of comprehensibility or verifiability.

n237. Although it is unlikely that a court would find it reasonable to risk one's own safety, once a helper has been identified, she is more likely to encounter arguments that it would have been reasonable to do "something more than was done." For a recent example, see supra note 46.

To the extent that a duty to act reasonably puts more helpers at legal risk, good samaritan legislative protection could be expanded. To right the safety incentives in such situations, it must be clear first that an unreasonable failure to aid is actionable, and that, absent gross misconduct, the helping effort will not be. Although this Article has not addressed statutory developments, the proposal made here and suggestions for further protection of, or encouragement for, good samaritans are not mutually exclusive. For a discussion of legislative encouragement of aid to individual in distress, see Rosenberg, supra note 5.

n238. Henderson, supra note 20, at 935.

n239. Id. at 932.

n240. Henderson, supra note 20, at 920. See, e.g., Vaughan v. Menlove, 132 Eng. Rep. 490, 493 (C.P. 1837) (opinion of Tindale, C.J., ["Defendant contends] . . . that the question ought to have been whether the Defendant had acted honestly and bona fide to the best of his own judgement. That, however, would leave so vague a line as to afford no rule at all, the degree of judgement belonging to each individual being infinitely various. . . .") See generally HOLMES, supra note 48; PROSSER & KEETON, supra note 5, at 173; Henry W. Edgerton, Negligence, Inadvertence and Indifference: The Relation of Mental States to Negligence, 39 HARV. L. REV. 849 (1926); James W. Ellis, Tort Responsibility of Mentally Disabled Persons, 1981 AM. B. FOUND. RES. J. 1079; Warren A. Seavey, Negligence -- Subjective or Objective?, 41 HARV. L. REV. 1 (1927); Jacobus Tenbroek & Floyd W. Matson, The Disabled and the Law of Welfare, 54 CAL. L. REV. 809 (1966).

n241. Even where a defendant is delusional, she may be liable for negligent acts. See, e.g., Breunig v. American Family Ins. Co., 173 N.W.2d 619 (Wis. 1970) (defendant believed God was controlling the steering wheel of her car). However, the law has established a lesser standard of care for children; See e.g., Ellis v. D'Angelo, 253 P.2d 675 (Cal. Ct. App. 1953) (four year old exonerated because of limited mental development of children that age); and has allowed consideration of physical limitations. See, e.g., Hammontree v. Jenner, 97 Cal. Rptr. 739 (Ct. App. 1971) (defendant stricken unforeseeably by epileptic seizure before car accident). Perhaps most importantly, courts have made concessions where defendants are involved in emergency situations. See, e.g., Ferrer v. Harris, 434 N.E.2d 231 (N.Y. 1982):

"[I]t is common sense and good law that, when one is confronted with a sudden and unexpected event or combination of events which leave little or no time for reflection or deliberate judgment, this itself may be a significant circumstance which, realistically as well as conceptually, should enter into the determination of the reasonableness of the choice of action pursued. . . ."


n242. See supra notes 188-95 and accompanying text.

n243. See supra notes 49-64 and accompanying text.

n244. For example, a mentally deficient person whose failure to act resulted in harm to another would be no more likely to be exonerated under common law rules than under an expanded duty rule if the person were involved in a special relationship with the victim, had innocently been involved in an accident with the victim, or, in some jurisdictions, had started a rescue attempt and then shyly withdrawn. See supra notes 34-37 and accompanying text. It is unlikely that there would be a significant number of cases raising the issue, but to the extent that there are such cases, the difficulty is shared by existing m common law rules.

n245. Henderson argues that not only potential litigants, but courts too, must be able to enjoy rules that are comprehensible and verifiable, and which require no more of people than they can deliver. However, his major concern is that the rule would simply not be judicially manageable. Henderson, supra note 20, at 936.

n246. See, e.g., Epstein, supra note 19, at 198-99; Hale, supra note 201, at 215. But see, Carl S. Hawkins, Premises Liability After Repudiation of the Status Categories: Allocation of Judge and Jury Functions, 1981 UTAH L. REV. 15. Professor Hawkins studied cases in jurisdictions that had eliminated status categories and concluded that judges had been able to maintain constraints on the judicial process. Id. at 61. Nothing suggests a different result in situations involving affirmative duties to act.

n247. Epstein, supra note 19, at 198-99.

n248. See, e.g., In re Kinsman Transit Co., 388 F.2d 821 (2d Cir. 1968); Credit Alliance Corp. v. Arthur Andersen & Co., 438 N.E.2d 110 (N.Y. 1985).

n249. See, e.g., Walker v. Bignell, 301 N.W.2d 447 (Wis. 1981) (employing a policy-based approach but declining to find liability where accident allegedly resulted from community's failure to clear away growth and weeds near roadway).

n250. See RESTATEMENT, supra note 2, 324.

n251. Good attorneys are likely to advise their clients not to bring such actions. See Beyond Good Samaritans, supra note 142, at 275 n.1.

n252. In addition to the practical problems that would make such a suit unlikely, current duty or proximate cause doctrines would prevent floods of successful cases. Under the approach proposed by this Article, a court might dismiss on the grounds that no duty exists if, for example, it determined that requiring charitable contributions were a matter of tax policy not appropriate for adjudication. Note that such an approach could be consistently applied whether one had simply refused to make a contribution or had made contributions regularly and then stopped. Using the California policy criteria for determining duty, a court might simply find that the connection between the defendant's conduct and plaintiff's harm to be too tenuous to allow the suit to proceed. See supra note 154.

n253. The common law rules are distracting because they require line-drawing on a number of issues of dubious relevance. See supra notes 65-73 and accompanying text. That they help to define and narrow the duties owed seems unlikely.

n254. The situation is parallel to the rejection of the common law immunities for landholders. There too, the rules were complex and not always focused on the relevant considerations. See supra note 140.

n255. Tarasoff, 551 P.2d at 343 (quoting PROSSER, TORTS (4th ed. 1971) 56, p. 341).


n257. Diane Kiesel, Who Saw This Happen?, 69 A.B.A.J. 1208 (1983).

n258. See, e.g., Henderson, supra note 20, at 937.

n259. Id.

n260. In the criminal context, conspiracy trials may present similar difficulties. Even where a number of manufacturers are sued, they may all present complex "sorting out" problems, which ultimately are simply considerations of the reasonableness of their behavior. To the extent that the existence of joint and several liability reduces the court's difficulty in sorting out defendants (a point which Henderson accepts) it may serve the same purpose in nonfeasance situations. To the extent that such rules present fairness problems, they present them also in "ordinary" negligence claims. See Weinrib, supra note 5, at 262 (such difficulties should not be determinative).

n261. This may ultimately be Henderson's major point. See Henderson, supra note 20, at 938 n.179.

n262. Wisconsin, for example, has abandoned the Restatement rules in favor of a policy-based approach. In 1981, the Wisconsin Supreme Court decided Walker v. Bignell, a case involving the failure of a municipality to keep brush cut at highway intersections without reference to the misfeasance-nonfeasance distinction or special relationship analysis. 301 N.W.2d 447 (Wis. 1981). The approach does not seem to have triggered a flood of multi-party rescue issues on appeal during the last decade.

n263. See supra note 221.

n264. Such a result would of course simplify matters for the courts, and not incidentally, might save some lives. See Silver, supra note 221 (author proposes statutory remedies for failure to rescue, concluding lives saved would justify any new incursions on individualism).

n265. See generally supra Part II.

n266. See supra note 128 and accompanying text.

n267. See supra notes 131-32 and accompanying text. In varying degrees courts are already engaging in analysis that makes the Restatement rules less significant.

Prepared: February 8, 2002 - 5:02:29 PM
Edited and Updated, February 12, 2002

Back to
Philosophy of Law
Home Page