Concept
Direct Consequences Versus Foreseeability
Proximate Cause: Foreseeability and Scope of the Risk
Direct Consequences Versus Foreseeability
A negligent defendant need not have foreseen how badly the plaintiff would be hurt or how the injury would come about. Section 435(1) of the Restatement (Second) of Torts says so: if the actor's conduct is a substantial factor in bringing about harm, the actor's failure to foresee the extent of the harm or the manner in which it occurred does not prevent liability.1 The rule forgives some surprise, and this unit asks how much. When the plaintiff was a foreseeable victim but suffered a different kind of harm from the one a careful person would have feared, must that kind of harm itself have been foreseeable?
Two tests answer differently. The direct-consequences test looks back from the injury. It asks whether the harm followed from the negligent act in an unbroken natural sequence, and if it did, the defendant answers for it whether or not anyone could have predicted it.2 The foreseeability test looks forward from the moment the defendant acted. It asks what a reasonable person in the defendant's position should have realized about the consequences of the act, not what this defendant actually contemplated.3 Most courts and the Restatement take the forward-looking view: they limit negligence liability for physical harm to foreseeable consequences within the scope of the risk and to foreseeable plaintiffs.4 The backward-looking view is held by fewer courts and writers.5
The split repeats the one in Palsgraf. Judge Cardozo asked who fell within the range of risk reasonably foreseeable when the defendant acted.6 Judge Andrews treated foreseeability as only one factor and would have let the jury set the limits, so his view closely approximates a direct-causation test.7 Polemis and the Wagon Mound cases carry that disagreement from the question of which plaintiffs may sue to the question of which kinds of harm a defendant must pay for. Courts sort unforeseeability into four problems: unforeseeable plaintiffs, unforeseeable extent of harm, unforeseeable types of harm, and unforeseeable manner of harm. Plaintiff unforeseeability matters a great deal, unforeseeable extent and manner do not, and cases involving an unforeseeable type of harm are decided in ways that resist generalization.8 This unit takes up the type of harm. The next takes up extent and manner.
Polemis and Direct Consequences
The direct-consequences test takes its name from re Polemis and Furness, Withy & Co., a 1921 English decision that arose under a time charterparty by which the respondents chartered their vessel to the appellants.9 The court held the defendant liable for all the damages directly caused by the dropping of a plank, even though the result was unforeseeable. It added that defendants are not liable for remote causes.10 The two halves of the rule work together. A consequence is direct when it follows in an unbroken natural sequence from the effect of the defendant's act on conditions already existing and forces already in operation, without the intervention of external forces that were not then in active operation, and direct consequences are always proximate.11 A harm reached only through new outside forces is remote, and the defendant does not answer for it. Once negligence is shown, the inquiry runs along the path from act to injury, and what anyone could have predicted drops out.
The hypothetical isolates the point on which the two tests part company most visibly: harm that follows directly from the negligence but falls outside the risks that made the conduct careless.
The weakness of the direct-consequences test is that “direct” describes a result without explaining it. A restaurant owner who leaves rat poison near a stove appears to be the direct cause of the explosion that follows, since the placement of the poison led it to heat up and blow up. Yet many courts would be uncomfortable imposing liability for that unexpected consequence, and it is hard to escape the conclusion that “direct” is just a word rather than a method of analysis.12 Remoteness is no better as a stand-alone limit. 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.13 Courts likewise have no trouble imposing liability for toxic substances to which people are exposed decades after the substances were produced.14 Neither “direct” nor “remote,” by itself, tells a judge where to stop.
The directness idea has not vanished. Judge Andrews's Palsgraf dissent closely approximates it,15 and some courts still define proximate cause as a cause that, in natural and continuous sequence, unbroken by any efficient intervening cause, produces the plaintiff's harm.16 In products cases, the extent to which a defect caused the plaintiff's damages encompasses both direct harms and indirect, foreseeable harms, a rule that corresponds to the directness test. The eggshell-skull rule also makes a defendant pay for unforeseeably high damages directly caused by the tortious contact. Any unfairness is offset because compensatory damages equal the amount of harm in question, which can be substantially below the foreseeable harms.17 A later section develops that rule.
On an exam, apply the rule the facts name. When they name none, say that most courts use foreseeability, show how that test resolves the stem, and note whether the direct-consequences test would change the result.
The Wagon Mound Cases
In Overseas Tankship (U.K.) Ltd. v. Morts Dock & Engineering Co., known as Wagon Mound No. 1, the Privy Council rejected the direct-consequences test of Polemis and held that a defendant is liable only for damage of a foreseeable type.19 The defendant's ship discharged fuel oil into Sydney Harbor. The owner of a wharf went ahead with its welding after “experts” assured it that the floating oil would not burn. Molten metal set afire cotton waste or rags floating in the oil, and the flames destroyed the dock. The dock owner was a foreseeable plaintiff in Cardozo's sense, because some harm to the dock, such as fouling it with oil, was foreseeable.20 The fire was not. The plaintiff could therefore recover for fouling the slips of its dock, a foreseeable consequence of releasing the oil, but not for the unforeseeable fire that destroyed the dock itself.21
The Privy Council's objection to Polemis concerned the vantage point from which responsibility is judged. Looking back after the fire, anyone can see that the spill led to the loss of the dock. “After the event even a fool is wise. But it is not the hindsight of a fool; it is the foresight of the reasonable man which alone can determine responsibility.” The Polemis rule, by “substituting ‘direct’ for ‘reasonably foreseeable’ consequence leads to a conclusion equally illogical and unjust.”22
The line the Privy Council drew follows the type of harm. A plaintiff inside the zone of danger recovers for the kinds of harm a reasonable person would have foreseen, here the fouling, and not for a different kind of harm, here the fire, even though the fire came from the same negligent discharge. Unforeseen detail is a different matter. Under section 435(1), the actor's failure to foresee the extent of the harm or the manner in which it occurred does not prevent liability once the conduct is a substantial factor,23 and courts give no weight to unforeseeable extent or manner.24 The difference makes sense. The test asks whether the kind of danger was among those that made the conduct careless. Once it was, the defendant created that danger and cannot insist that the loss stay within the size or the route the defendant predicted.
Wagon Mound No. 1 therefore does not say that fire can never be foreseeable from a spill of oil. The owners of ships damaged in the same fire sued in Overseas Tankship (U.K.) Ltd. v. Miller Steamship Co., known as Wagon Mound No. 2, and they won.25 The plaintiff shipowner was not bound by the earlier decision, and unlike the dock owner it was not hampered by possible charges of contributory negligence or assumption of risk. It introduced evidence that some risk of fire was reasonably foreseeable by the defendant's engineer, and Lord Reid, speaking for the Privy Council, distinguished Wagon Mound No. 1 and affirmed a judgment for the plaintiff.26 The defendants should have been aware of some slight risk of fire, and because their conduct had no utility, they were not justified in ignoring even that slight risk.27
Lord Reid noted that the plaintiff lost in Bolton v. Stone because the risk of harm was small and the activity that caused it was lawful.28 In Wagon Mound No. 2, nothing of value stood on the other side of the scale, so even a slight foreseeable risk could not be brushed aside. The two decisions thus apply one test to two records. In the first, the unforeseeability of fire was established, and in the second, the evidence showed a slight but real risk that a reasonable engineer would have taken into account. Foreseeability does not demand probability, and the difference between the cases lies in the findings of fact.
How a court describes the risk often decides how the test comes out. Foreseeability is an accordion concept, depending on the detail and precision with which foresight is required.29 A court can go wrong in either direction. If the defendant need only foresee “personal injury” of some sort, a plaintiff who tripped over a box lid left on a 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. The court must instead focus on the particular types of harm to be anticipated, while avoiding the idiosyncratic details of the accident.30 In one case the accident happened in an unusual way, but the court concluded 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 defendant liable.31
The opposite result appears in Doughty v. Turner Manufacturing Co. A cover was knocked into a vat of molten liquid. No injury of the expected kind followed the fall, but a short time later the liquid erupted from the vat and injured the plaintiff. The trial court's recovery was reversed, because there was no reason for anyone to expect the eruption, which was caused by a chemical reaction between the liquid and the lid.32 In the explosion case, an unusual manner did not defeat recovery because the harm was the kind that made the conduct careless. In Doughty, nothing made the eruption a risk anyone should have anticipated.
The Restatement (Third) of Torts states the same limit as the risk principle. An actor's liability is limited to the physical harms that result from the risks that made the actor's conduct tortious, and the Restatement adopts that view for most courts.33 Applied to the oil spill, the question is whether fire was among the risks that made the discharge negligent. The answer was no in the first case, and on the evidence in the second it was yes.
Two cautions apply when you use these cases on a fact pattern. First, do not assume that every court applies Wagon Mound mechanically. Outcomes in unforeseeable-type cases are harder to generalize,34 and there is still much room for flexibility and special rules even within the risk principle.35 Second, a conclusion that harm was unlikely is not a conclusion that it was unforeseeable. The question is whether a reasonable person would have taken the risk into account, and a slight risk of serious harm from conduct that has no value can count.