Concept
Palsgraf and Foreseeable Plaintiffs
Proximate Cause: Foreseeability and Scope of the Risk
Palsgraf and Foreseeable Plaintiffs
A careless act aimed at one person often reaches another, and the other person sues. The preceding sections asked what kind of harm falls within the dangers that made conduct careless. This unit asks whom those dangers reach. The risk principle answers both questions with one test: liability extends only to the consequences whose foreseeability made the conduct negligent, and only to persons within the foreseeable zone of danger. The foreseeability limits on proximate cause therefore match the limits that decide whether the conduct was negligent at all.1
The second limit rests on a premise about negligence: risk is always risk to someone. Judge Cardozo built Palsgraf on that premise. Negligence, he held, is a relational tort, so the defendant must have breached a duty to a particular victim or at least to a class of people that includes the plaintiff.2 A plaintiff outside the class cannot borrow a wrong done to someone else. Courts place the resulting limit in different doctrinal boxes: some speak of duty, some of proximate cause, and some of scope of liability.3 All of them ask who may claim the protection of a rule that makes certain conduct negligent.
The two opinions in Palsgraf disagree about where that limit belongs and about who decides whether a plaintiff meets it. Cardozo's majority opinion comes first, then Andrews's dissent, then the way courts apply the problem today.
Cardozo's Majority Opinion
In Palsgraf v. Long Island Railroad, a 4–3 majority of the New York Court of Appeals reversed a judgment for the plaintiff and dismissed her complaint.4 Judge Cardozo, writing for the majority, held that a defendant owes a duty of care only to plaintiffs within the range of reasonably foreseeable risk that its conduct creates. The range is fixed when the defendant acts. If a reasonable person at that moment would not have foreseen a risk of injury to someone situated like the plaintiff, no duty runs to her, and she cannot recover even though she was in fact hurt.5 Most states follow this approach, and a substantial minority follows the dissent taken up in the next subsection.6
Mrs. Palsgraf stood on a platform of the defendant’s station, having bought a ticket to Rockaway Beach, when a train bound elsewhere began to pull out and two men ran to catch it. One boarded without mishap. The other, carrying a small package wrapped in newspaper, jumped aboard but seemed about to fall. A guard on the car reached forward to help him, and a guard on the platform pushed him from behind. The package fell to the rails. It held fireworks, though nothing in its appearance said so, and they exploded. The shock threw down some scales at the other end of the platform, many feet away, and the scales struck Mrs. Palsgraf. By concession, nothing in the situation suggested to even the most cautious mind that a parcel wrapped in newspaper would spread wreckage through the station.7
Cardozo reasoned in three steps, each answering a different question.
- What did the guards’ act threaten? Pushing a man who is about to fall from a moving train endangers him and his package. The unreasonable risk the guards created was that the passenger or his package would be hurt, not that Mrs. Palsgraf would be, so the conduct, if wrongful at all, was no wrong in relation to her.8
- What could a careful observer have perceived? Cardozo tested the point by removing the inadvertence. A guard who threw the parcel down knowingly and willfully would not, so far as appearances could warn him, have threatened her safety, and liability cannot be greater when the act is merely careless.9
- What must the plaintiff prove? “Proof of negligence in the air, so to speak, will not do.” She must sue in her own right for a wrong personal to her, not as the vicarious beneficiary of a breach of duty to another.10
The demand for a wrong personal to the plaintiff follows from how negligence is measured. Conduct is careless because of the risks it creates, and a risk always threatens someone, which is why Cardozo called negligence, like risk, a term of relation.11 The probability and gravity of harm are weighed against the burden of precaution for the people the conduct endangers. A plaintiff outside that group never entered the weighing, so calling the defendant careless toward her applies the standard to a danger it did not measure. Cardozo’s image of the orbit makes the same point: “the orbit of the danger as disclosed to the eye of reasonable vigilance would be the orbit of the duty.” The risk reasonably to be perceived defines the duty to be obeyed, and risk imports relation, because it is risk to another or to others within the range of apprehension.12
Incentives point the same way. An actor can adjust his precautions only for risks he anticipates, so a rule that made him pay for unforeseeable harm would make the world no safer or more efficient, and the prospect of liability for risks no one could anticipate might even produce overdeterrence.13 The policy behind foreseeability is that a tortfeasor should have been able to recognize the danger and to adjust his conduct reasonably to avoid the harm.14 The test is also objective. It asks what a reasonable person in the defendant’s position should have realized about the consequences, not what the particular defendant actually contemplated.15
Cardozo placed this question ahead of causation. If no wrong was done to the plaintiff, there is no occasion to ask how far the consequences of a wrong extend, because the consequences to be followed must first be rooted in a wrong.16 That sequence is why the case is usually taught as a duty case. The labels vary, though. Some accounts say the guards owed Mrs. Palsgraf no duty, and others say the railroad did not breach the duty it owed her because the sequence of events was so extraordinary.17 Every account rests on one ground: she was an unforeseeable plaintiff.
Cardozo also said who decides. The range of reasonable apprehension is at times a question for the court and, if varying inferences are possible, at times a question for the jury. Because the facts conceded that nothing would have warned the most cautious mind, the question never reached the jury.18 The requirement bites rarely. In the run-of-the-mill negligence case, proof of negligence automatically satisfies foreseeability. The requirement matters only in the occasional case where injury to the plaintiff was not a harm within the risk that made the defendant’s conduct negligent, and Palsgraf is the model of that case.19
Three variations show where the edge of the orbit falls. The passenger the guards pushed stands inside it: shoving a man at the edge of a moving train risks that he will fall and be hurt, so he is a foreseeable plaintiff.20 Suppose instead that a traveler stood right beside him, close enough to be knocked down or hit by the falling parcel. The guards could foresee injury of that sort, but not an explosion, and most courts would find that her injury arose from a different risk than the one the railroad could anticipate and would deny recovery.21 A passenger who knew his parcel held fireworks is different again. He was negligent for bringing them into a crowded place precisely because an incident might detonate them, so most courts would hold him liable to those hurt in the explosion, even though the railroad escapes liability for the same injury.22 The injury is the same in each scene. What changes is whether a particular defendant’s conduct made that injury a foreseeable risk to that plaintiff.
Because the range of risk is fixed when the defendant acts, hindsight has no role. Ask what a reasonable person in the guard’s position, unable to see inside a newspaper-wrapped parcel, would have perceived at the instant of the push.23 Exam facts will usually reveal what the package held, and reasoning backward from the explosion makes the harm look foreseeable simply because it happened. Resist that pull. From the guard’s position, the explosion was a risk nothing in the situation disclosed.
Andrews's Dissent
Judge Andrews, writing for the three dissenters, would have affirmed Mrs. Palsgraf’s judgment.25 His analysis has two parts. Negligence, on his view, is the breach of a duty owed to everyone, so the plaintiff need only be someone the negligent act in fact injured. The limit on liability then comes from proximate cause, which he treated as a practical judgment about where the law stops tracing consequences. A substantial minority of jurisdictions follows this approach.26 Its emphasis on the directness of the connection between act and harm closely approximates the direct-causation test taken up in the next section.27
Andrews began by denying that negligence exists only toward a particular victim. “Every one owes to the world at large the duty of refraining from those acts that may unreasonably threaten the safety of others.” A driver racing down a crowded street is negligent whether or not anyone is hit, because on this view the act itself is the wrong, and it wrongs everyone within its reach. “Due care is a duty imposed on each one of us to protect society from unnecessary danger, not to protect A, B, or C alone.”28
He still required a duty owed to the one complaining, but not a duty owed to a particular individual because harm to him might be expected. He found the necessary relation between the actor and everyone the act in fact injures, a group wider than the people the actor might have expected to hurt. “Not only is he wronged to whom harm might reasonably be expected to result, but he also who is in fact injured, even if he be outside what would generally be thought the danger zone.”29 Because harm to someone was the natural result of the act, all those in fact injured may complain: “Unreasonable risk being taken, its consequences are not confined to those who might probably be hurt.”30
That reading does not make liability unlimited. “An overturned lantern may burn all Chicago,” and we may follow the fire from the shed to the last building.31 Every link in such a chain satisfies the but-for test, and the chain has no natural end. What the word proximate means, Andrews said, is that “because of convenience, of public policy, of a rough sense of justice, the law arbitrarily declines to trace a series of events beyond a certain point.” That, he added, “is not logic. It is practical politics.”32
His example was the fire rule then in force in his own state, under which a defendant whose negligence set one building afire answered only for the first building the fire reached. Lines of this kind move. Shortly after the early New York decision, the state extended recovery to the first adjoining landowner,33 and later to the first property the fire jumped to, even if not adjoining. Other courts rejected the “first building” test altogether and held defendants liable when fires spread a substantial distance.34 A line must fall somewhere, and the policy behind where it falls can change.
That placement of the limit separates the two opinions. Cardozo asked first whether the railroad owed this plaintiff a duty at all.35 Andrews, having found the act negligent toward the world, said that what remains is to “inquire only as to the relation between cause and effect”: “We deal in terms of proximate cause, not of negligence.”36
Andrews offered considerations in place of a formula, and the jury weighs them together:37
- Was there a natural and continuous sequence between the act and the harm?
- Was there a direct connection, without too many intervening causes?
- Was the result likely, and could prudent foresight have foreseen it?
- Was the result too remote from the act in time or in space?
Foreseeability sits in that list as one consideration among several, and it is never determinative.38 Remoteness likewise counts without ending the inquiry, because distance is not the same as unforeseeability. Fireworks negligently manufactured in New York and shipped 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.39 Courts likewise have no trouble imposing liability for toxic substances to which people are exposed decades after the substances were produced.40
The viewpoint for judging foresight also differs from Cardozo’s. Cardozo fixed the range of risk at the instant of the act, from the standpoint of an observer who could not see inside the package.41 Andrews looked back from the catastrophe and asked what might be expected to follow once the explosion occurred.42 That starting point makes the result far easier to foresee, since injury to someone on a crowded platform is a predictable consequence of an explosion even when no one could predict the explosion.
Andrews’s own hypothetical shows how the factors work at different distances. A chauffeur negligently collides with a car that, unknown to him, is filled with dynamite. The explosion kills a pedestrian on the sidewalk, cuts a person sitting in a window across the street, and cuts another sitting in a window a block away. Ten blocks off, it startles a nursemaid into dropping a baby.43 Andrews doubted that the baby would recover, treated the person across the street as a question for court or jury, and said that denying recovery to the person a block away would rest on a finding that the injury was not the proximate result of the negligence. Distance matters because the greater it grows, the more surely other causes intervene to shape the result.
Mrs. Palsgraf stands at the other end of that range. Andrews held that knocking an apparently harmless package onto the platform was negligent, and that the railroad answered for the proximate consequences: to the owner if the contents broke, to a passenger if the package crushed a foot, and to anyone in the immediate vicinity if it exploded.44 She stood perhaps 25 or 30 feet away. No remote cause intervened. The explosion directly caused her injury, the only intervening event being that the concussion smashed a weighing machine that then fell on her, and little time and space separated act from harm. Given an explosion of that force, it took no great foresight to predict injury to someone on the platform. Andrews could not say as a matter of law that her injury was not the proximate result of the negligence, so he would have affirmed.
The opinions also disagree about who decides. Cardozo’s majority kept the case from the jury, ruling that the plaintiff was unforeseeable as a matter of law. Andrews would let the jury set the limits.45 One commentator reads the dissent as aimed less at that outcome than at the majority’s rule-based formulation, which can never capture everything that goes into deciding that there is no proximate cause as a matter of law. On that reading, Andrews’s objection to withholding the case from the jury is almost secondary.46 Under Andrews’s approach, a plaintiff who is unforeseeable may still recover, because foreseeability is one factor and not a gate.
Students often misdescribe the dissent in two ways. It does not reject foreseeability, which it counts as one factor, and it does not impose liability for every harm an act causes, because it still cuts off the chain when the connection is too attenuated or remote.47
Applying Palsgraf Today
Most states follow Cardozo, so the working question in them is whether the plaintiff stood within the zone of risk that made the defendant’s conduct negligent. A substantial minority follows Andrews and leaves the limit to proximate cause.49 Either way, the inquiry begins as Cardozo began it: identify what made the conduct dangerous, then ask whether this plaintiff stood within the danger.
Courts disagree about what to call the inquiry. On similar facts, some analyze the issue as duty or “proximate cause,” while others use scope-of-liability rules.50 In Thompson v. Kaczinski, for example, the Iowa Supreme Court applied the Restatement (Third)’s scope-of-liability rule, under which the factfinder decides whether the type of harm that occurred was among the risks that made the conduct negligent. The label can change who decides: the Palsgraf majority took the question from the jury, while Andrews would have given it to the jury. An exam answer should therefore use the jurisdiction’s own vocabulary when the facts supply it, and otherwise ask the question every version shares, whether the plaintiff fell within the risk that made the conduct negligent.
Duty rules can also settle a whole category of cases in advance, as a matter of policy, before any proximate-cause analysis begins. Most courts hold that a social host owes no duty to third parties to refrain from serving a guest too much alcohol, but that a bartender in a commercial bar does owe that duty.51 Only when the duty question is answered yes, and the plaintiff proves breach, damages, and cause in fact, does the proximate cause issue arise.52 A later section returns to such no-duty rules.
Modern litigation applies the foreseeable-plaintiff question to facts far from a railroad platform. After the September 11 attacks, families of victims sued airlines, airport security companies, and airport operators for negligently failing to prevent the hijackings. The defendants conceded a duty to their passengers but denied any duty to people on the ground. In re September 11 Litigation, the federal court, applying New York law, held that those present in the Twin Towers were foreseeable plaintiffs, because the screening was performed for the benefit of passengers and of people on the ground.53 The explanation lies in what screening is for. Checkpoints exist to keep weapons off aircraft, so the risk that makes lax screening careless is that a plane will be seized and turned on a target, and a worker in a tall office tower stands inside the orbit of that danger. The harm realized the very hazard that made the omission careless.
The same question produced the opposite answer in Stutts v. De Dietrich Group. Plaintiffs exposed to chemical weapons stored in Iraq sued the banks that supplied letters of credit and the suppliers of the raw materials, and the court dismissed, relying on Cardozo’s statement that “the risk reasonably to be perceived defines the duty to be obeyed.”54 The plaintiffs’ theory required the banks to have foreseen a long chain of independent events: that the chemicals would be sold to Iraq, made into weapons, stockpiled, later released, and inhaled by the plaintiffs.55 Nothing would have suggested to the most cautious mind that financing the sale would cause that harm. What separates the two decisions is how close the harm lay to the center of the risk and how many independent links separated act from injury. Screening exists to stop the attack that injured the tower occupants. A letter of credit sat at the head of a long chain of independent choices.
The defendant need not have foreseen this plaintiff by name. The duty runs to a particular victim or at least to a class of people among whom the plaintiff is a member, so the plaintiff need only belong to the endangered class.56 How the risk is described decides who belongs to that class, and a court can err in either direction. A court that defines the risk too specifically gets bogged down in the idiosyncratic details of the accident. A court that defines it too generally, as “personal injury” of some sort, would let a plaintiff who tripped over a box lid on the sidewalk, missed her train, and was hurt when a later train crashed recover, because the lid created a risk of personal injury and she suffered personal injury. The court must focus on the particular types of harm the defendant’s act made foreseeable.57
Passing the plaintiff test does not guarantee recovery, because the harm must also be of a kind the conduct made foreseeable. Where molten liquid erupted from a vat a short time after a cover had been knocked into it, a recovery was reversed because no one had reason to expect the eruption, which came from a chemical reaction between the liquid and the lid.58
A recurring fact pattern shows how much room the first two steps leave for disagreement. A vehicle negligently strikes a person, and that person’s body then strikes the plaintiff, who had been standing in apparent safety. Several courts have allowed recovery, among them the Minnesota court in Kommerstad v. Great Northern Railway, the Alabama court in Alabama Great Southern Railroad v. Chapman, and the Massachusetts court in Wolfe v. Checker Taxi Co. The Pennsylvania Supreme Court, by contrast, has twice declared that the pattern presents an unforeseeable plaintiff, once in Wood v. Pennsylvania Railroad, which involved a train, and once in Dahlstrom v. Shrum, which involved a bus.59 The split is a disagreement about how widely to draw the orbit of danger. A court can describe the risk of a collision as injury to the person struck, which excludes everyone else, or as injury to people near the point of impact, which includes the plaintiff, and either description fits the same facts. In a modern variation, a train struck a commuter who dashed across the tracks, and the commuter’s body then struck a plaintiff standing a hundred feet away. An Illinois appellate court reversed a summary judgment for the commuter’s estate that rested on a no-duty argument and sent the case to a jury to consider breach and proximate cause.60
The foreseeability inquiry also decides whether a later wrongful act ends liability. Kentucky, for example, no longer treats criminal conduct as automatically interrupting the causal link. The spark that ignited trash negligently left beside a building may have been struck accidentally, negligently, intentionally, or even criminally, and it can still be foreseeable and so leave liability in place.61 The earlier section on foreseeable acts of others develops that point.
Rescuers are the one class that courts treat as foreseeable as a matter of policy. In Wagner v. International Railway Co., the defendant negligently left a car door open and a passenger was thrown out as the car rounded a curve. The passenger’s cousin was injured trying to rescue him. Cardozo held it error to instruct the jury that the defendant’s negligence was not a proximate cause of the rescuer’s harm unless the rescuer had been asked to help, because “danger invites rescue.” The reasoning is that negligently imperiling anyone makes an attempt at rescue predictable, so harm to the rescuer falls among the risks that made the conduct negligent. Many jurisdictions now apply the rule as a matter of law, even where a rescue probably was not foreseeable, which reflects a policy favoring voluntary rescue.
Belonging to a foreseeable class does not prove that the defendant caused the particular injury. The Washington Supreme Court in McCoy, a product liability suit against Suzuki, rejected the argument that the rescue doctrine excuses a rescuer from proving that the defendant proximately caused the injury. The court relied on Maltman v. Sauer, where a helicopter summoned to an automobile accident crashed on the way and killed its crew. The court held as a matter of law that the car accident was not the proximate cause of the crash, which was too remote a result of the original accident. In McCoy itself, a rescuer struck by an approaching car at the scene of a vehicle rollover presented a foreseeability question close enough for the jury. A rescuer’s injury from the dangers of the rescue stays within the scope of liability, while an injury from an unrelated or remote cause must meet the ordinary proximate-cause requirement.