Concept
Manner and Extent of Harm
Proximate Cause: Foreseeability and Scope of the Risk
Manner and Extent of Harm
The preceding section asked whether the type of harm was one the defendant should have foreseen. Even when the answer is yes, two features of an accident can outrun anyone's prediction: how much damage resulted, and what chain of events carried the danger to the plaintiff. Section 435 of the Restatement (Second) of Torts speaks to both.1
Its two subsections pull in opposite directions. Subsection (1) favors the plaintiff: if the actor's conduct is a substantial factor in bringing about harm, the actor is liable even though the actor neither foresaw nor should have foreseen the extent of the harm or the manner in which it occurred.2 Subsection (2) is the backstop: the conduct may be held not to be a legal cause where, after the event and looking back from the harm to the negligent conduct, it appears to the court highly extraordinary that the conduct should have brought about the harm.3
The law keeps both rules because each guards against a different failure. Every accident follows its own sequence, and almost none could have been predicted, so a rule demanding foresight of the exact route would excuse nearly every defendant who created a recognized danger. A rule charging defendants with any consequence, however bizarre, would load losses far out of proportion to the fault onto defendants whose carelessness may have been slight. Section 435 resolves the tension by tolerating unforeseen detail as the general rule and reserving a narrow exception for the extraordinary case.
The choice of how broadly to describe the risk ties the two rules together, because the same accident can look typical or freakish depending on the level of generality at which it is told. The subsections below take up the general rule, the problem of description, and the backstop in turn.
Unforeseen Extent and Manner
A defendant may fail to foresee four different things about an accident: the plaintiff, the extent of the harm, the type of harm, or the manner in which it occurs. Plaintiff unforeseeability matters a great deal. Unforeseeability of extent or manner does not, and cases involving an unforeseeable type of harm are decided in ways that resist generalization.4 This subsection concerns the two categories the law forgives.
The reason lies in what the foreseeability limit measures. Most courts limit negligence liability for physical harm to foreseeable consequences within the scope of the risk and to foreseeable plaintiffs, so that liability reaches the physical harms that result from the risks that made the conduct tortious.5 A risk is a danger, and a danger can reach its victim by many routes. Fairness therefore asks only that the defendant should have foreseen, in loose terms, harm to the class of persons that includes the plaintiff and the general type of harm that occurred. Differences in the manner of occurrence are generally irrelevant, as is the fact that the harm was greater than expected.6
Marshall v. Nugent shows how far the tolerance runs. After a negligent traffic mix-up, “before the disturbed waters have become placid and normal again,” the events that follow may be “bizarre indeed,” yet the defendant may be liable, because no one in the defendant's position could predict in advance just how the negligent act would work out to another's injury. That unpredictability is no bar to recovery.7 The opinion also marks the boundary. Had the passengers' car been restored to normal travel, with the extra risks from the negligence over and the situation stabilized, a later injury would have come from the ordinary risks of being on the highway at all, and the earlier negligence would not be a proximate cause as a matter of law.8 An odd manner is forgiven while the danger the defendant created is still operating.
United Novelty Co. v. Daniels gives the clearest example of an odd path to a familiar harm. An employer let an employee clean machine parts with gasoline in a small room that held a heater with an open flame. A gasoline-soaked rat ran from the parts to the heater, caught fire, and ran back, and the explosion that followed killed the employee. The court conceded that the manner was unusual, but it held that an explosion was exactly the type of accident to be anticipated from using a volatile, flammable liquid in a small room with an open flame, and it held the employer liable.9 The reasoning compares the harm with the risk. What made the practice negligent was the chance that gasoline vapor would meet a flame and explode, hurting someone in the room, and that is what happened. The rat was only the means by which the flame reached the fuel.
Other courts say the same thing in other settings. In Derdiarian v. Felix Contracting Corp., the New York Court of Appeals wrote that a defendant's inability to anticipate the precise manner of the accident or the exact extent of the injuries does not preclude liability as a matter of law where the general risk and character of the injuries are foreseeable.10 Products cases apply the rule the same way. In Wallace v. Owens-Illinois, Inc., a soda bottle exploded, and the plaintiff was hurt not by flying glass but by a fall as he cleaned up the spill. The court ruled that the maker of a defective bottle of liquid under pressure should foresee not only broken glass and a spill but also normal efforts to clean it up.11 A manufacturer need not divine the precise manner in which an accident will occur. It must foresee the general type of risk that caused the accident.12
The extent half of the rule has the same structure. When harm of a foreseeable type occurs, the defendant is not excused because it proved greater than expected. Suppose conductors carelessly knock a passenger's package to the platform. They could foresee that a package might fall if they were not careful and that its contents, whatever they were, might be damaged. If the package holds a valuable antique, the antique's value goes to the extent of the loss and leaves the nature of the risk unchanged, and courts usually hold defendants liable for the damage actually caused even when it exceeds what one would ordinarily expect.13 Section 461 of the Second Restatement states the personal-injury counterpart: the negligent actor is liable although a physical condition of the other, neither known nor knowable to the actor, makes the injury greater than the actor should have foreseen as a probable result of the conduct.14 A later section develops that eggshell-plaintiff rule.
Fire shows the principle at a larger scale. Andrews's Palsgraf dissent warned that once unreasonable risk is taken, its consequences are not confined to those who might probably be hurt: “An overturned lantern may burn all Chicago.”15 Courts have drawn the boundary in stages. An early New York rule limited recovery to the first building a fire reached. The state soon extended it to the first adjoining landowner and later to the first property the fire jumped to, and other jurisdictions rejected the first-building test and held defendants liable for fires that spread a substantial distance.16 Each step allowed recovery for a larger extent of loss.
A foreseeable result is recoverable by whatever route it arrives, while an unforeseeable result stays unrecoverable even when the defendant's act set the whole sequence going. Doughty v. Turner Manufacturing Co. falls on the second side of that line. A cover was negligently knocked into a vat of molten liquid. The fall risked splashing, but no splashing injury occurred. A short time later the liquid erupted from the vat and injured the plaintiff. His recovery at trial was reversed, because no one had reason to expect the eruption, which a chemical reaction between the liquid and the lid had caused.17 The plaintiff was hurt by a different risk from the one that made the act negligent.
Set beside Daniels, the case shows what separates a detail from a different risk. In Daniels the vapor, the flame, and the explosion were parts of the one danger that made the practice negligent, and the rat merely brought them together. In Doughty the danger of the falling cover was splashing, and the eruption came from a chemical process that nothing about the negligent act made foreseeable. The line is fine, because the same injury, hot liquid thrown on the plaintiff, can be described as a splash or as an eruption. The next subsection takes up how that description decides cases.
Step three does the real work, and answers go wrong there in two directions. Some deny recovery because the sequence was bizarre, which confuses an unusual manner with an unforeseeable type of harm. Others grant recovery because the defendant's act started the sequence, which forgets that a but-for cause of a different kind of harm lies outside the risk.
Defining the Risk
Section 435(1) excuses unforeseen extent and manner, but it leaves open a prior question: what counts as the same kind of harm? Any injury can be described at several levels of generality. A crushed foot is a falling plank, a dropped heavy object, a workplace accident, or simply a personal injury, and each description answers differently whether the harm lies within the risk that made the conduct negligent. Many scope-of-liability disputes are really disputes about description. The right level is the hazard that made the act unreasonable, such as explosion, fire, collision, or fall, and the analysis can fail in either direction.
The ordinary case never raises the problem. It is negligent to run a red light precisely because an intersection collision is foreseeable, so if the negligence caused the plaintiff's injury, proximate cause follows as a matter of law.19 Only the occasional freak case, in which injury to the plaintiff was not a harm within the risk that made the defendant's action negligent, calls on the harm-within-the-risk test to sharpen the foreseeability inquiry.20
Within the duty analysis, a foreseeable risk is a category of accidents, a “general field of danger,” that a reasonable person would consider in deciding whether to take a particular precaution. Once duty is established, the further question is whether the particular risk that factually caused the injury falls within that general category. If it does, the risk is necessarily foreseeable, but the issue is case specific, which makes it a question of proximate cause for the jury.21
The first error describes the risk too narrowly, building the odd details of the accident into the hazard. Conductors who carelessly handle a passenger's package cannot know it holds an antique, and if the risk is stated as damage to antiques, no one could have anticipated it. Most courts define the risk more generally. The risk that makes the conductor's act negligent is that the passenger or his belongings would fall and be injured, and when one of those risks causes the injury, the defendant is liable. The antique's value goes to the extent of the risk, not its nature.22
Products cases show the same instinct. A teenager was burned when cologne she was using ignited and flared. She argued that the product was defective for failing to warn that cologne is highly flammable, and the defendant answered that the product had been misused and that the damage was unforeseeable. The court reasoned that the alcohol in perfume creates a risk of fire, since a lit cigarette might touch perfumed skin, so a jury could find proximate cause.23
The opposite error stretches the risk until it covers almost anything. If the defendant need foresee only personal injury of some sort, the plaintiff who tripped over a box lid left on the sidewalk, missed her train, and was hurt when a later train crashed could recover, because the lid created a risk of personal injury and she suffered personal injury.24 The court must focus instead on the particular types of harm the defendant's act made likely, while avoiding the idiosyncratic details of the accident.25 The bystander on a railroad platform shows the same discipline. Conductors who knock a package loose can foresee injury from falling or from being hit by the package, but not an explosion, and most courts would conclude that the injury arose from a different risk than the one the railroad could anticipate and deny recovery.26
Shih v. Starbucks Corp. applies the same discipline to a hot drink. The plaintiff walked to a table, put down her cup, removed the lid, bent over the table, pushed out her chair, lost her balance, grabbed the table, and knocked her drink off it. The court held that Starbucks was not liable for the resulting burns, which she attributed to a defect in the cup. Although a customer could foreseeably lose her balance at a table, that event was not “within the scope of the risk . . . created by a restaurant's decision to serve a hot beverage that is filled to the brim or that does not have a sleeve.”27 A description pitched at burns would have covered her. The court asked instead which risks made serving the drink in that condition wrongful, and it recognized that an event can be foreseeable in the sense of imaginable and still fall outside the set of risks the defendant should have weighed in deciding how to act.
Description also controls when the harm falls on a different interest from the one the conduct threatened, and courts divide there. Entrusting a pistol to a child is negligent because of the danger of accidental discharge. If the child instead drops it on someone's foot, a risk for which the entrustment alone was not unreasonable, many courts will say the defendant owed no duty to protect against the harm that actually resulted, and others reach the same result through proximate cause or scope of liability. Not all courts deny liability, however, because the nature and scope of the risk can be defined broadly enough to encompass the harm that occurred, and duty is no less elastic.28 The labels vary as well. Courts sometimes say that the defendant's conduct was not the proximate cause of the harm when they really mean that the defendant was not negligent at all, or that the plaintiff's own conduct was a superseding cause.29
Statutes fix the risk through their purpose. In Gorris v. Scott, sheep were washed overboard from a ship that violated a regulation requiring that they be kept in pens, but the regulation was meant to protect the animals from contagious disease. The harm that occurred fell outside the risk the rule addressed, and recovery failed. The inquiry has the same shape as the common-law one, with the legislature's aim identifying the risk.
Decided cases do not always reconcile. In Wallace, the court held that a fall while cleaning up after an exploding bottle lay within what the manufacturer should foresee.30 In Crankshaw v. Piedmont Driving Club, Inc., a woman vomited on a restaurant rest-room floor, and her companion, who rushed in less than a minute later to help, slipped and broke a hip. The court held that the companion's suit could not go to the jury.31 Both plaintiffs were hurt while dealing with the aftermath of an earlier harm, and the results differ. One explanation lies in the description of the risk. An exploding bottle creates a mess of glass and liquid that someone must clean up, and that mess contains the fall, while an illness creates a risk to the person who is ill, which does not obviously contain a fall by a helper.
A plaintiff can also try to redescribe the case, and the mechanism of harm is the test. Suppose a plaintiff in a Palsgraf-style case shifts attention from the package to the scale that fell on her and argues that the railroad should have anticipated the scale's toppling. The court would still ask whether the general mechanism that caused the harm was foreseeable. It would probably hold that the mechanism here was a fireworks explosion, which a reasonable person weighing the risks of placing the scale on the platform need not anticipate, even though she could anticipate other ways the scale might topple.32
Because every consequence can be redescribed, a chosen description earns its place only if it traces back to a danger that made the conduct careless. A description that cannot be traced is a conclusion presented as a premise.
Steps three and four do the real work, and exam answers usually perform only one of them. A student who tests only for narrowness will accept any description that rescues the plaintiff, and one who tests only for breadth will reject every unusual sequence. A correctly described risk still leaves a backstop, which the next subsection takes up.
Highly Extraordinary Results
Section 435(2) sets the outer limit on the tolerance that § 435(1) extends. The conduct may be held not to be a legal cause where, after the event and looking back from the harm to the negligent conduct, it appears to the court highly extraordinary that the conduct should have brought about the harm.34 The vantage point is the distinctive feature. Negligence and the scope of the risk are judged at the moment of acting, by a reasonable person who cannot see the future, and this test runs the other way. The court knows what happened, traces the path backward from the injury to the careless act, and asks whether the connection is so remarkable that charging the defendant with the result would offend common sense.
A hindsight test follows from what § 435(1) gave up. Because that subsection forgives unforeseen detail, a defendant could otherwise be charged with coincidences that attached themselves to a chain of events the defendant set going. The backward-looking inquiry catches those extremes without asking anyone to have predicted the particulars. The wording keeps it narrow. The conduct may be held not to be a legal cause, so exclusion is a conclusion the court is permitted to reach and not an automatic result of strange facts, and the word highly means that a result that is merely unusual, improbable, or unprecedented does not qualify.
The doctrine is not confined to civil cases. In People v. Acosta, a California murder appellant argued that there was insufficient evidence his conduct was the proximate cause of three deaths.35 Because the standard is a screen and not a formula, it is best learned from cases on either side of it.
In Jensen v. Schooley's Mountain Inn, a tavern allegedly served alcohol to a visibly intoxicated patron. He drove away, parked, and began climbing a tree for no known reason. A branch broke, he fell, and he drowned. The court assumed the service was negligent but affirmed summary judgment for the tavern, treating the events as past the line beyond which liability should not be extended as a matter of fairness or policy. Looking back from the drowning, the chain runs through a drive and a climb that nothing in the tavern's conduct pointed toward.
The result went the other way in Hairston v. Alexander Tank & Equipment Co. A dealer negligently failed to tighten the lug bolts on a customer's car, and a wheel fell off. The customer brought the car safely to a stop on a bridge, and a van stopped behind to offer help. A truck then rammed the van into the customer, who was killed. The court held the dealer liable even though the truck driver's conduct was inexcusable, because the collision was not so highly improbable and extraordinary as to bear no reasonable connection to the harm the dealer's negligence threatened. The later driver's negligence cannot be what separates the two decisions, since the dealer answered despite it. What separates them is how the last event relates to the danger the defendant created. A car that loses a wheel and stops on a bridge is disabled where traffic moves, and being struck by other traffic is the danger a disabled car presents, so the second driver's carelessness was the channel through which the dealer's original risk reached the owner.
Remoteness alone does not trigger § 435(2). A defendant can foresee that fireworks it negligently manufactures in New York and ships across the country in January may cause harm in California on July 4th. That harm is remote in time and space, and indirect, but it is foreseeable nonetheless, and courts have no trouble imposing liability for toxic substances that injure people decades after they were produced.36 Courts also let juries decide proximate cause in some quite extraordinary situations. In one case the plaintiff hit a horse that suddenly appeared in front of his car at 45 to 50 mph. The horse was thrown into the air and fell on the roof, collapsing the roof rail on the passenger side and killing his wife instantly. Rejecting the car maker's argument that the accident was “freak and bizarre,” the court upheld a verdict for the plaintiff.37 An odd route does not by itself trigger § 435(2). The backstop applies only when the oddity severs the harm from the danger that made the conduct wrongful, and courts differ on when that has happened.
The line between unusual and highly extraordinary is hard to draw, and Yun v. Ford Motor Co. shows judges drawing it differently on a single record. A spare tire assembly fell off a Ford van on the Garden State Parkway. The driver stopped safely on the berm, and a passenger got out, crossed two lanes of traffic to retrieve the tire, and was struck and killed on the way back. The Appellate Division majority affirmed summary judgment for the defendants, relying on § 435(2) and Jensen: because the van had come safely to rest, the passenger's trip across the highway was highly extraordinary, and reasonable people could not differ. Judge Baime, dissenting in part, saw a jury question. A jury could find it foreseeable that the tire would come loose while the van was moving and that someone might be hurt trying to retrieve it, and what seems strange to judges may seem ordinary to others. Any carelessness in the attempt would be weighed as comparative fault. The New Jersey Supreme Court issued a per curiam opinion in 1996 reversing for the reasons in the dissent. The doctrine is therefore real, since several judges applied it, and contestable, which is why courts use it sparingly and why it leads to the question of who decides foreseeability.
Step four does the real work, and the usual exam error is to treat step two as decisive. A strange route to a foreseeable kind of harm is forgiven by § 435(1), and strangeness matters here only to the extent that it cuts the injury off from the risk that made the defendant negligent.