Page 1 of 14 The doctrine of proximate cause is a complex phenomenon that responds to a number of unseen forces acting in the minds of judges. These unseen forces include the judges' intuitive attributions of causality and blame, as well as their conscious or unconscious biases regarding class, gender, and the role of industry and commerce. Economic efficiency may also play a limited role, but, it turns out, its influence is often trumped by other unconscious cognitive heuristics and biases. Law and Economics From an economic standpoint, it is possible to separate cases raising an issue of proximate cause into two categories: ones in which the tortuous act did not predictably increase the risk of the injury claimed, and ones in which the tortuous act did increase the risk of the injury, but only slightly or otherwise in ways that were somehow less than the risk increases caused by other factors. If the doctrine of proximate cause is responsive to economic influences, we should predict that the law will not impose liability on the former category, but may impose liability on the second category in cases where the risk is still substantial enough to be inefficient, or where it is difficult or impossible for tort law to influence the other, larger influential forces at work. For example, in cases where a person bumps into another on the street, causing them to miss their train and take another which later explodes, it is impossible to say that the bump actually increased the chances of the victim taking an exploding train. If we imagine train explosions as random and difficult to predict from the perspective of a passenger, the interpersonal collision was just as likely to make the victim miss a train that was going to explode, thus saving the victim's life, as it was to make the victim take a train that was going to explode, killing her. When our behavior has effects but we have no capacity to predict what Page 2 of 14 those facts should be, imposing liability for those the results of such behavior cannot give effective or clear incentives to actors and is therefore undesirable from an economic standpoint. However, some acts may clearly increase the chances of an injury occurring, even though the injury may be the result of intervening forces or be otherwise difficult to foresee. If we know an activity will, on a statistical level, make others more vulnerable to otherwise uncontrollable risks, efficiency dictates that we refrain from such activity when those risks to others outweigh the benefits to us. Leaving the keys in the ignition, for example, may not predictably lead to particular unauthorized and dangerous uses of the car, but it certainly can be said to increase, however marginally, the likelihood of such an event; moreover, while theft by a terrorist or dangerous driver may seem unlikely (although it wouldn't seem so unlikely in, say, the West Bank), the burden of taking the keys with you as you leave a car is so slight relative to even unlikely potential harms as to justify imposing liability on the careless owner. Thus, in McClenahan v. Cooley, a defendant who left the keys in the ignition of his car was held potentially liable when a thief stole the car, became involved in a police chase, and crashed into an innocent motorist, killing three people. The court left it to the jury to decide whether the defendant's acts were a substantial factor that foreseeably led to the harm. Similarly, in Vesely v. Sager, a lodge owner was held liable for selling alcohol to an intoxicated patron who he knew was driving home, and who then drunkenly crashed into another car; the economic benefits of selling alcohol were here outweighed by the rather high expected accident costs. However, courts do not reliably impose liability where efficiency dictates it should. In Ryan v. N.Y. Central R.R. Co., a court imposes only limited liability on a railroad that engages in behavior likely to set buildings on fire by limiting its liability only to the building that it actually set fire to, not to the buildings that predictably caught fire themselves as a result of that fire. Page 3 of 14 Here, a dangerous activity (negligently operating engines close to a woodshed) has a significant, measurable probability of causing harm for which the actor will be immune for liability even if found negligent. When actors not forced to internalize the costs of such activities, they run the risk of incurring more costs to the economy than the gains they produce. However, the Ryan court justifies its decision not to impose liability with the tacit assumption that the benefits of crowded buildings in a society that extensively uses flammable fuel outweigh the potential accident costs, although it argues at the same time that imposing liability for such accidents would be so costly as to destroy civilization as we know it. If it is in fact efficient to build houses close to one another, then the only wrongful behavior that the court can and should discourage is that of setting fires that may spread. Limiting liability for such behavior to exclude the consequences of crowded building arrangements fails to allocate the correct incentives to individuals who may start fires; when evaluating the worth of additional precautions, they will weigh the burden of extra care only against the expected damage to the first building that catches fire, and not against the expected damage of the fire's spread to nearby buildings. This calculation will result in inadequate levels of care. If, on the other hand, the prospect of paying for all the damage that one might do is truly impossible for the defendant to pay, then we are all better off if the defendant refrains entirely from potentially incendiary activities. In Pinder, like in Ryan, the defendant is held not liable for security lapses that predictably created risks to third parties. Although the defendant's keeping the plaintiff's house key in an unlocked box made it possible for two men to steal the key and rape the plaintiff, the judge categorically denies recovery for a harm that was not “foreseeable,” refusing even to allow the question to go to a jury. Like the judge in Ryan, the decision in Pinder makes the implicit argument that the company's practice of keeping keys for the “convenience” of home buyers Page 4 of 14 should enjoy the presumption that it is, in fact, a desirable practice, despite the added danger of an assault or break-in. By focusing on the unforeseeability (and implied low probability) of the break-in and rape, the judge side-steps arguments the defendant could have easily added more security (and, in fact, he barely discusses this fact): in his analysis, no care need be taken against unforeseeable events. However, as in the McClenahan case, even an unlikely or difficult to foresee event should, from an efficiency perspective, give rise to liability if the defendant's burden of precaution is low and the defendant can foresee that some unspecified harm may result from inadequate care. It is important to note that proximate cause, as treated in the judge's analysis in Pinder, relies on a question of foreseeability and predictability that is on its face very difficult to distinguish from P (the probability of an injury), a crucial factor in a standard negligence analysis. In fact, when we separate out the attributional questions presented by proximate cause and focus solely on its effects in economic analysis, it stops having any special importance as distinguished from the question of negligence. There is no real economic reason to absolve actors from liability simply because some other “intervening” causal factors, natural or human, have significantly contributed to an adverse event. Such intervening forces are often quite possible to calculate, and, when it is in their interest to do so, people and businesses frequently do calculate such risks; they are perfectly capable of exercising care to protect against harm caused by third parties. Thus, using the tort system to provide people and companies with incentives to do such calculations would, in theory, lead to increased care and efficiency. Moreover, while courts tend to interpret the intentional acts of a third party as particularly damaging to establishment of proximate cause, the risk of someone stealing a car or a key is no more difficult to calculate than risks created by “blame-neutral” events, like an adverse drug Page 5 of 14 reaction or an economic downturn. It is no wonder, therefore, that the legal economist Calabresi referred to “proximate cause” as “weasel words.” Attribution theory When a legal concept is widely seen by scholars as inconsistently, obscurely, or capriciously applied, one potential explanation is that the inconsistency in its application is motivated by unseen biases or cognitive strategies. Proximate cause is one of those obscure concepts. Judicial intuitions about proximate causes, however, closely mirror cognitive strategies that social psychologists have identified in how humans attribute causation and blame in a variety of social contexts. As Glannon explains, it is difficult for the human animal to calculate “distant” causes and effects. For example, when presented with an auto accident, an observer's attention will most likely be limited to the behavior of the two drivers and possibly the weather and road conditions in operation at the time. Limitations on cognitive processing power prevent the observer from considering potential causes such as the reasons each driver got behind the wheel in the first place or events that may have influenced the emotional states of the drivers. Moreover, since tort doctrine, like other spheres of public life, is influenced by our tendency to see individuals as free choosers, it tends to assume that the actions of proximate actors are freely chosen and relatively impervious to situational factors. Thus, when the actions of one entity “set the scene,” as it were, for another individual to act in a harmful manner, the tort system assumes that its deterrence power is nonetheless best used on the second entity and not the first; that, while the first entity may have made the harm more possible or likely to occur, the second individual is still “in control” and thus deterrable through tort or criminal liability. However, this does not mean that those causal factors outside the immediate locus of Page 6 of 14 events were not important, or that it was impossible to tell how those more “remote” causes or situations would influence the probability of an accident. Consider, for instance, the example in Vesely v. Sager, in which a bartender supplied an intoxicated patron with alcohol knowing that the patron would then attempt to drive home late at night along an unusually dangerous road. Any ordinary person observing the patron's drinking habits could probably foresee that the patron might be later involved in an automobile accident. Thus, when we actually do look to actors outside the immediate locus of events, we may find people whose behavior contributed to a dangerous situation and who may be deterrable by tort law. Yet, as the Vesley court remarks, lodge owners who supply alcohol knowing that alcohol consumption will lead to dangerous behavior are generally shielded from liability because their actions occurred “behind the scenes.” Our failure to see situational influences is not just a matter of determining causation; it also influences our intuitive assignments of responsibility, control, and blame. As the judge in Vesely pointed out, although a bartender will not have too hard a time predicting that its patrons who drive away intoxicated may get into an accident, observers of an accident involving a drunk driver will attribute all the blame to the drunk driver, the closest “blameworthy” actor. It is the driver, after all, who chose to drink the alcohol and then chose to drive while intoxicated. As previous opinions cited in Vesely argued, the seller merely provides the alcohol, but it is the buyer who chooses to consume it (this argument, of course, raises the question of exactly what else the alcohol seller expected the patron to do with the alcohol they bought). Notably, however, the common law had made some exceptions to this rule in cases where the buyer was known to have no control over their alcohol consumption, where the buyer was a minor or too intoxicated to make rational judgments about whether to continue drinking, or where the buyer was a slave. In these cases, the buyer is seen as an imperfect agent, one who is not capable of or entitled to Page 7 of 14 the exercise of control over his or her drinking habits. When observers are persuaded by such salient situational forces to see the intoxicated person as somehow less culpable than a “normal” intoxicated person would be, they will look to the next closest actor in their search for a good person to blame: in this case, the individual who knowingly provided an incompetent individual with alcohol. Attribution theory therefore suggests that courts are more likely to find liability despite “intervening forces” if we fail to find another suitable intervening party on whom to focus blame. In Vesely v. Sager, although the judge nominally based his opinion on the “unlawful” nature of selling alcohol after the regular closing time for bars, he included a discussion on, and most likely was highly influenced by, the facts that the patron was known to be “incapable of exercising the same degree of volitional control over his consumption of intoxicants as the average reasonable person” and that the lodge owner knew the patron intended to drive home along a dangerous road. In this case, the patron had started drinking at 10 P.M.; it is unlikely that the consumption of a beverage after 2 A.M. made a big difference in the driver's likelihood of being involved in an accident. Rather, it makes sense to infer that, finding the lodge owner blameworthy, the judge sought some legal reason to impose liability and found it in the prohibition on sale of alcohol after a certain time. Our attribution-of-blame heuristic also explains to some extent the holding in Doe v. Pinder Construction Company, in which a supervisor was not held liable for the criminal acts of an employee even though relatively elementary precautions on the supervisor's part would have prevented the injuries to the plaintiff, and the result in Watson v. Kentucky and Indiana Bridge & R.R. Co., in which the negligent acts of a railroad company subject them to liability when the harm is caused by the intervening acts of a negligent third party, but not the intervening acts of Page 8 of 14 an intentionally acting, criminally culpable third party, even though the railroad's negligence (and subsequent blameworthiness) is the same in either case. However, there remain some cases in which the criminal acts of an intervening party are not held to absolve the tortfeasor of liability, particularly in situations where criminal behavior is so expected that the actual criminal seems relatively anonymous and interchangeable, where we believe that if this particular person hadn't committed the crime, someone else surely would have. Thus, we find liability where a train drops off a passenger in a dangerous area and she is later assaulted, where a driver leaves a car unattended with keys in the ignition and a car thief steals it and injures someone in the predictably ensuing car chase with the police, and where a landlord fails to fix a lock on an apartment in a dangerous area. Critical Theory According to Professor Street, proximate cause is often determined based on “mixed considerations of logic, common sense, justice, policy, and precedent. . . .” As critical theorists have often pointed out, however, “common sense” and ideas of “justice” and “policy”are often influenced by our attitudes towards different social groups and our theories about which ones should enjoy power and protection. The manner in which a judge applies the concept of proximate cause will therefore often hinge on that judge's intuitive sense of which kinds of parties we need to protect, and, since the two issues so often go together, which kinds of harm merit compensation. These biases and intuitive theories of social justice have considerably shaped the application of the proximate cause doctrine. Its focus on “direct” and “proximate” causal factors currently makes it more difficult to hold people or corporations accountable for creating situational forces that lead to harms, shielding those with extensive power and knowledge of the Page 9 of 14 effects of situation from liability for the results of the situations they create. For example, gun manufacturers may use proximate cause doctrine to shield them from the effects of their attempts to make lethal weapons widely available and from advertisements encouraging aggressive people to buy guns, because any injuries due to assault with a weapon they sell involve the intervening actions of a wrongdoer. Moreover, by absolving an individual of responsibility for events that involved an intervening intentional act, proximate cause doctrine can preclude recovery to victims of rape, harassment, or assault, who are unable to obtain insurance recoveries against the direct aggressor (because the aggressor either acted intentionally or is judgment-proof) or against others who made that aggression possible. Ryan: When is proximate cause not proximate cause? Consider, for example, Ryan v. N.Y. Central R.R. Co. As discussed earlier, the court denies recovery for damage to houses based on a particularly confusing proximate cause argument. Setting fire to a house through negligence, Judge Hunt argues, cannot be the proximate cause of damage to other houses that catch fire as a result of the original fire, even though intentionally setting fire to a house may be the proximate, albeit unintentional, cause of damage to other houses that may catch fire. While the modern reader of this decision is still trying to figure out why an individual's motives behind setting a fire have anything to do with whether the act of doing so “proximately caused” other fires (one imagines that the spread of fire caused by arson is no more “natural and expected” than the spread of a fire caused by negligence), she is hit with yet another hard-to-explain line of reasoning: that the defendant could not control the spread of the fire to the other buildings, and therefore cannot be held liable. Since the defendant could control the setting of the original fire (as is evidenced by the fact that they were found negligent), the spread of a fire in a “populous city” is not difficult to anticipate, Page 10 of 14 and since the other buildings probably would not have caught fire if the first one had not, it is again difficult to see exactly why the judge insists on attributing “no control” to the defendants for spread of a fire they caused. The judge goes on to apparently distinguish the present case from Guille v. Swan, in which a balloonist was found to be a proximate cause of a crowd's trampling of a garden because his descent into the garden “substantially requested the presence of the crowd there.” It is difficult, however, to grasp exactly how a descending balloonist could be the proximate cause of damage done by a number of thinking human beings if a house fire is not the proximate cause of a spark jumping onto the house next door and setting it alight. Finally, the judge argues that, if defendants were held liable for fire damage caused to houses other than its own, they would be “subject to a liability against which no prudence could guard, and to meet which no private fortune would be adequate,” liability that would result in no less than “the destruction of all civilized society”: strong words to use in a case where a large railway company is alleged to have negligently caused damage to a private dwelling. This liability, argues Hunt, would be made all the more burdensome by the fact that it is impossible even for a vigilant person to guard against “accidental or negligent fires” and that one “cannot insure his neighbor's building or furniture, for the reason that he has no interest in them.” Taken to its full logical extent, however, this argument implies that no actions for negligent property damage should recover, including cases where proximate cause is not an issue. If, for example, the operator of the railroad had set of sparks near the plaintiff's house instead of near its own woodshed, thus directly causing the plaintiff's house to burn down, then the railroad would not be liable such direct damage because its negligence was inevitable and it could not insure the plaintiff's belongings. The non sequiturs and contradictions in this opinion imply that pure logic and precedent Page 11 of 14 are not truly moving this decision. Moreover, as discussed earlier, it would be very difficult to say that economics that is moving this case. Something else must, therefore, be influencing the judge's opinion. One explanation for this opinion is articulated in Morton Horwitz's theory of tort-law subsidization of industry. In the 19th century in particular, tort liability shrank in order to protect the interests of the large industries that had started expanding as a result of the industrial revolution. Courts protected these industries from liability in the belief that industry was an instrument of societal progress that needed to be encouraged, not discouraged, from carrying out its [often dangerous] regular activities. Liability was holding America back. However, this subsidization hid the costs of industry and made it difficult to make certain that its benefits really were outweighing the costs of the injuries it caused, so courts had to put some level of faith in industry advocates that they knew what was best. This leap of faith was most likely made easier by the fact that such advocates tended to make up the most economically powerful sectors of society, and that their victims were often the most socially disadvantaged. Not only did these powerful members of society have more resources to influence/capture the opinions of the courts and the rest of the American public, but also judges, like other humans, were likely to see successful individuals and corporations as inherently good and worthy of praise and assistance, while simultaneously blaming the disadvantaged for their distress. Thus, blaming the victim also probably had some influence on the court's decision in Ryan. Social psychology has taught us that, when we see others who have undergone some misfortune and cannot point directly to some “bad actor” who has caused it, we feel discomfort and try to reconcile this observation with our sense that the current world is just by blaming the victim of the misfortune. Hunt, who has labeled fires such as that in this case as inevitable Page 12 of 14 consequences of civilized society and who is generally in favor of civilized society, must therefore find some way to justify the fact that society sometimes burns down people's homes by blaming those whose homes burn down. He does so by arguing that it was the plaintiff's responsibility to insure his own home, and that by “neglect[ing] such precaution, and. . . call[ing] upon his neighbor. . . to indemnify him instead,” the plaintiff is actually in the wrong. Pinder: Thinking like a feminist In Doe v. Pinder Construction Co., the court categorically denies that a construction company cannot be liable for allowing an employee to steal a homeowners' house key and raping her, because the criminal acts of taking the keys and raping the plaintiff were “unforeseen, intervening causes of the plaintiff's injuries.” The construction company is not its employees' keeper; any unforeseen criminal activities that the employees engage in, regardless of how easy the construction company made it for those activities to be carried out, is the sole responsibility of the employees themselves. As discussed in the law and economics section above, the holding in Pinder is economically inefficient. Assume, for example, that, as the jury decided in Wassell v. Adams, a rape by an intruder does about $850,000 in damage to the victim, and that a reasonably sturdy padlock costing about $10 would have prevented the crime. Assume also that an extra $10 worth of precaution on Doe's part would not have prevented the rape; after all, it is difficult to see what she should have done beyond locking her door, and the judge in this case nowhere implies that she was negligent. If, therefore, the odds were only a little over one in a hundred thousand that someone with access to the demo house would steal a key and rape someone, the additional cost of a padlock would be efficient. With those odds, there is simply no need for the construction company to have any reason to suspect of its workers in order to be reasonably interested in Page 13 of 14 investing the added money for a padlock. Nor does the court in Pinder have substantial precedential grounds for ruling as it does. In its opinion, it quite inexplicably decides that, although the construction company had the keys to Doe's home, it should not be held to the same duty as an inkeeper or a landlord, because the latter relationships include a “contractual obligation regarding security or maintenance and a control over the plaintiff's premises. . . .” This reasoning ignores the fact that, in fact, Doe and the construction company did have a contractual obligation regarding security and maintenance: Doe had actually entrusted the construction company with the keys to her own home, under a contract that required them to enter her home and perform repairs. Moreover, while the company did not have full control over the premises, their possession of her key and their ability to enter the premises at will certainly gave them a considerable amount of control. The judicial opinion also overlooks the fact that, as explained by Glannon, an atypical “manner” of harm causation is not seen as an unforeseeable intervening cause if the defendant's negligence had already set up a situation where that kind of harm was likely to occur. For example, when a worker was told to use gasoline in a small room near an open flame, the fact that a rat caught fire and ran under the machine the worker was cleaning, causing an explosion, does not constitute a superseding intervening event because an explosion is exactly what could be expected from the situation. However, in Pinder, the judge considers the defendants' lack of care securing the keys irrelevant because the rapists in this case gained access to the keys by stealing keys to the model home and entering it at night, and because the defendant could not have anticipated that the rapist would steal the keys to the model home. However, given the defendant's lack of security, there were a number of ways for ill-intentioned people to obtain keys to people's homes, and the uniqueness of this particular sequence of events should not under Page 14 of 14 traditional doctrine have absolved the defendant from liability. The judicial opinion is particularly problematic when we approach the case from the point of view of the victim. It is difficult to imagine that Doe herself did not think they were assuming some responsibility to guard her house keys against theft. Moreover, if she had been given the option of choosing what level of security to keep the keys in, it is reasonable to believe that Doe would have wanted her keys locked in a cabinet and inaccessible to every single person who entered the model house, regardless of the fact that, as the judge asserts, they were all authorized to be there and “not strangers.” Also problematic here is the fact that the responsibility to guard against the acts of third parties is here apportioned unequally among the parties. When it comes to looking after their own safety, women are assigned a rather strict responsibility to guard against intentional harms by third parties, including those with no known criminal history (and certainly those who they know, as one of the defendants here did, have recently served jail time). If Doe had purposefully or negligently allowed two construction workers into her house and they had raped her, she, like Wassell, would likely have been found seriously negligent. However, the “common-sense” awareness that all men, especially lower-class men such as construction workers and above all men who are often drunk and/or have served jail time, is not attributed to the defendants in this action. Instead, the judge attributes knowledge of the rapists' potential dangerousness to mere “hindsight,” despite the fact that one of the defendants knew that one of the rapists had served time in jail, but apparently hadn't bothered to inquire further about what he had served jail time for. Thus, the holding in Pinder absolves non-women of any responsibility for protecting women from rape or even awareness of situations that would be dangerous to women, illustrating that the “reasonable man” standard and the “reasonable woman” standard are fundamentally different.