Causation in tort law is the legal connection a plaintiff must prove between a defendant’s wrongful conduct and the specific harm suffered, and it comes in two parts: actual cause (cause in fact) and proximate cause. Both have to be established. A defendant who was clearly careless still walks free if the plaintiff cannot show that the carelessness actually produced the injury and that the injury was a foreseeable consequence of what the defendant did. Most tort cases turn on these two questions rather than on whether the defendant behaved badly in the first place.
Cause in Fact: The But-For Test
The first question is factual. Did the defendant’s conduct actually produce the plaintiff’s injury? Courts answer this with the but-for test: if the injury would not have happened but for what the defendant did, the defendant is a cause in fact. If the injury would have happened anyway, factual causation fails and the claim ends there.
A grocery store employee mops a floor and skips the warning sign. A customer slips on the wet tile and fractures her wrist. Had the floor been dry, she would have walked through without incident. That is a clean but-for link, and the failure to warn is a factual cause of the broken wrist.
Compare a driver going fifteen over the limit who rear-ends another car at a stoplight. If reconstruction shows the collision would have happened with the same force at the legal speed, factual causation gets shaky. The speeding was careless, but it did not change the outcome. Courts care about whether the specific negligent act triggered the specific harm, not whether the defendant was generally behaving badly.
Proximate Cause and Foreseeability
Even when a factual link exists, the law draws a line around how far liability extends. That line is proximate cause, and it is built on foreseeability. A defendant answers only for consequences that a reasonable person in the same situation could have anticipated. Without this limit, a single careless act could generate liability for an endless chain of ever more remote outcomes.
The landmark case is Palsgraf v. Long Island Railroad Co., decided by the New York Court of Appeals in 1928. Railroad guards helped a passenger board a moving train and knocked loose a small newspaper-wrapped package he was carrying. The package contained fireworks. It fell to the tracks and exploded. The blast knocked over heavy scales at the far end of the platform, and the scales struck Helen Palsgraf. Judge Cardozo held that the guards owed no duty to Palsgraf because no one could have foreseen that helping a passenger aboard would injure someone standing far away through that sequence. “The risk reasonably to be perceived defines the duty to be obeyed.”1New York State Courts. Palsgraf v Long Is. R.R. Co.
Foreseeability does not require the defendant to have predicted the exact sequence of events. It requires that the general type of harm to the general class of people fell within the range of foreseeable risk. Leave a stack of heavy boxes on a high shelf and it is foreseeable that a box could fall and injure a shopper. You do not need to foresee which shopper, or whether the box would land on a foot or a head.
The Eggshell Skull Rule
Foreseeability has one important exception on the damages side. Under the eggshell skull rule, sometimes called the thin skull rule, a defendant who causes a foreseeable type of injury is liable for the full extent of the harm, even if the victim’s injuries are far worse than anyone would have predicted. The defendant takes the victim as they find them, pre-existing conditions and all.
A delivery driver rear-ends another car at low speed. A healthy person might walk away with a sore neck. The actual driver has a degenerative spinal condition, and the same impact herniates a disc and requires surgery. The at-fault driver is liable for the full surgical costs, not just what a healthy plaintiff would have incurred. The initial collision was foreseeable. The severity was not, and under this rule that does not matter. Many courts extend the same principle to pre-existing psychological conditions.
Intervening and Superseding Causes
Sometimes a new event enters the picture between the defendant’s negligent act and the final injury. These are intervening causes, and they force courts to decide whether the original defendant still bears responsibility for what happened after.
Not every intervening event breaks the chain. If the event was itself a foreseeable consequence of the defendant’s negligence, the original defendant remains liable. A driver who causes a highway collision is responsible for injuries that worsen because the ambulance is stuck in rush-hour traffic for forty minutes. Delayed care after a crash is a predictable risk of causing the crash.
A superseding cause is different. When an intervening event is so unforeseeable and independent that it fundamentally breaks the connection between the defendant’s conduct and the final harm, it relieves the original defendant of liability for what follows. A contractor leaves a trench uncovered on a sidewalk. A criminal walks up and throws a pedestrian into the trench. The trench created a risk that someone would fall in, not a risk of assault. Courts would likely treat the criminal act as a superseding cause.
Criminal acts by third parties are not automatically superseding, though. When a business operates in an area with a known history of violent crime, courts have found that attacks on customers are foreseeable, and the business can be liable for inadequate security. The question is always whether the intervening event fell within the scope of risk the defendant’s negligence created. A poorly lit parking garage in a high-crime neighborhood where a customer is mugged is not producing a cosmic surprise; it is producing the exact risk the negligence made more likely.
When Two Causes Would Each Be Enough
The but-for test has a well-known blind spot. When two or more independent forces would each have caused the same harm on their own, neither one is technically a but-for cause, because the injury would have happened without it. Read literally, the test would let both defendants escape. Courts refused to accept that result.
The classic illustration involves two fires. In Kingston v. Chicago & Northwestern Railway Co., one fire was started by the railroad’s negligence and another of uncertain origin merged with it, and the combined blaze destroyed the plaintiff’s property. The court held that when two negligent acts concur to produce a single injury and either alone would have been sufficient, each wrongdoer is responsible for the entire damage. Separating the contributions is impossible, and letting both point fingers at each other would leave the victim with nothing.
Courts historically used a “substantial factor” test in these situations, asking whether the defendant’s conduct was a significant contributor rather than a trivial one. The Restatement (Third) of Torts moved away from that language because courts had started applying it too loosely, but the principle survived. When multiple acts each would have been sufficient to cause the harm independently, each act counts as a factual cause. This comes up constantly in toxic exposure cases where a worker develops disease after years of exposure at multiple job sites and no single employer’s contribution can be isolated.
When the Plaintiff Cannot Identify the Defendant
A separate problem arises when the plaintiff clearly suffered harm from one of several defendants but genuinely cannot tell which. Standard causation rules would block recovery because the plaintiff cannot prove that any specific defendant was the but-for cause. Two doctrines developed to handle this.
Alternative Liability
Alternative liability comes from Summers v. Tice, a 1948 California Supreme Court case. Two hunters fired shotguns simultaneously in the plaintiff’s direction, and one shot hit him in the eye. He could not prove which hunter’s pellet caused the injury. The court held that when all defendants acted negligently and the plaintiff cannot identify which one caused the harm through no fault of the plaintiff’s own, the burden shifts to each defendant to prove they were not the cause. A defendant who cannot make that showing is liable.2Stanford University. Summers v. Tice – 33 Cal.2d 80 It is fairer to make the wrongdoers sort out responsibility than to leave an innocent victim without recourse because two people were negligent instead of one.
Market Share Liability
Market share liability extends this idea to mass-produced products. The leading case is Sindell v. Abbott Laboratories, where women developed cancer from DES, a drug their mothers had taken during pregnancy decades earlier. Every manufacturer made the drug from an identical formula, and the cancer did not appear for ten or more years, so identifying the specific manufacturer was impossible. The California Supreme Court held that when all defendants produced an identical product and the plaintiff cannot identify which manufacturer supplied the one that caused the harm, each defendant is liable in proportion to its share of the market.3Justia. Sindell v. Abbott Laboratories A company that sold 15% of the DES on the market bore 15% of the liability. A defendant that could prove its product could not have been the one that injured the plaintiff escaped entirely.
Market share liability has not been adopted everywhere. Courts that recognize it require specific conditions: the product must be identical across manufacturers, the plaintiff’s inability to identify the source must not be the plaintiff’s fault, and the plaintiff must join enough defendants to represent a substantial portion of the market. These theories remain the exception.
Proving Causation
Knowing the doctrine is one thing. Meeting the burden of proof is another, and this is where most causation disputes actually play out.
The Standard of Proof
In civil tort cases the plaintiff must prove causation by a preponderance of the evidence, meaning it is more likely than not that the defendant’s conduct caused the harm.4United States District Court for the District of Vermont. Burden of Proof – Preponderance of Evidence That is a lower bar than the criminal standard of beyond a reasonable doubt, but it is not nothing. A plaintiff who shows only a possibility of causation, rather than a probability, loses.
Expert Testimony
In straightforward cases, causation is obvious. Someone runs a red light and hits your car; no scientist needs to explain the connection. In medical malpractice, toxic exposure, pharmaceutical injury, and other technically complex cases, expert testimony is essential. A plaintiff claiming a surgical error caused nerve damage needs a medical expert to explain why the symptoms are consistent with the alleged mistake and not with some other condition.
Courts do not accept just any expert opinion. Under the Daubert standard, followed in federal courts and most state courts, judges act as gatekeepers who evaluate whether an expert’s methodology is scientifically reliable before letting the testimony reach a jury. Judges look at whether the methods have been tested and peer-reviewed, the known error rate, and whether the methods have gained acceptance in the relevant scientific community. An expert who cannot explain the methodology behind a causation opinion, or who relies on speculation rather than data, will be excluded before the jury hears anything.
Loss of Chance
A particular problem arises in medical malpractice when a doctor’s negligence did not clearly cause the patient’s death or worsened condition but reduced the patient’s chances of a better outcome. Under traditional rules, if a cancer patient already had only a 40% chance of survival before a delayed diagnosis dropped it to 15%, the patient cannot prove by a preponderance that the delay caused the death. The patient would have been more likely to die regardless.
Some states address this through the loss of chance doctrine, treating the lost chance itself as a compensable harm. Rather than asking whether the patient would have survived, the court asks what the reduction in survival probability was worth, and damages are proportional to the percentage of chance lost. Other states, including California and Texas, have rejected the doctrine and hold plaintiffs to the traditional preponderance threshold. Where the doctrine is recognized, expert testimony is critical to quantify both the baseline survival odds and the reduction caused by the negligence.
Where Tort Cases Are Actually Won and Lost
In most tort cases, whether the defendant was negligent is the easy part. Causation is where the real fight happens. Defendants rarely argue that speeding through a school zone was reasonable. They argue that the speeding did not cause the accident, or that the injuries the plaintiff claims are not actually connected to the incident. Causation is the element where the plaintiff carries the heaviest practical burden and where factual ambiguity is hardest to resolve.
Plaintiffs lose causation arguments most often in two situations: when they cannot connect a specific injury to a specific act, and when an intervening event gives the defense an argument that the chain was broken. Evidence gathered early matters enormously. Medical records created right after an injury carry far more weight than records generated months later. Accident reconstruction, toxicology reports, and contemporaneous documentation build the foundation these arguments stand or fall on. The doctrines above give courts a framework for analyzing causation, but the outcome almost always comes down to the strength of the evidence behind the claim.