Causation in law is the required connection between what a defendant did and the harm a plaintiff or prosecutor is complaining about. Courts test it in two steps. Factual causation asks whether the defendant’s conduct actually produced the injury. Legal causation asks whether that kind of harm was a reasonably foreseeable consequence of the conduct. If either link fails, there is no liability, no matter how careless the defendant was.
Factual Causation and the But-For Test
Factual causation is the physical link. Courts ask a simple question: would the injury have happened if the defendant hadn’t acted the way they did? This is the “but-for” test, and it applies in both tort and criminal cases.1Legal Information Institute. But-For Test
A driver who runs a stop sign and strikes a cyclist satisfies the test cleanly: but for the failure to stop, the cyclist would not have been hit at that moment. The chain gets harder to prove when the connection depends on what might have happened. If the cyclist would have been struck by another car seconds later anyway, the defendant’s conduct starts to look dispensable rather than indispensable.
Lawyers build the chain with police reports, medical records, surveillance footage, and expert reconstruction testimony. The goal is to map a direct physical progression from act to injury. Cases often fail here, before anyone reaches the harder legal questions, because the link rests on speculation rather than evidence.
Legal Causation and Foreseeability
A physical link alone isn’t enough. Legal causation, also called proximate cause, asks whether the harm was a reasonably foreseeable consequence of the defendant’s conduct.2Legal Information Institute. Proximate Cause A single careless act can set off an improbable chain of events, and the law draws a line somewhere along it. If a reasonable person in the defendant’s position couldn’t have anticipated the kind of injury that occurred, the defendant generally isn’t liable for it.
The classic case is Palsgraf v. Long Island Railroad Co., where a railroad employee helping a passenger board a moving train knocked loose an unmarked package of fireworks. The explosion caused a scale to fall on a bystander far down the platform. The court held the railroad owed no duty to someone so far outside the foreseeable zone of risk.3New York State Reporter. Palsgraf v Long Is. R.R. Co. The core question survives in modern practice: was this the kind of harm the defendant should have seen coming?
The Eggshell Skull Rule
Foreseeability has an important limit. Under the eggshell skull rule, a defendant who causes an injury is liable for its full extent, even when the victim’s condition makes the harm far worse than anyone would have expected.4Legal Information Institute. Eggshell Skull Rule You take the victim as you find them. Rear-end a driver with a pre-existing spinal condition, and you pay for the serious injury that follows, not the minor one most people would have suffered.
Courts reconcile this with proximate cause by separating the type of harm from its severity. The defendant has to be able to foresee the general kind of harm, not its unusual magnitude. The same rule applies in criminal law: a defendant who throws a punch can face homicide charges if the victim had an undiagnosed condition that made the blow fatal.
Intervening and Superseding Causes
Things sometimes happen between the defendant’s act and the final injury. An intervening cause is any independent event that enters the chain after the defendant’s conduct but before the harm is complete.5Legal Information Institute. Intervening Cause Whether it breaks the chain comes down to foreseeability.
A foreseeable intervening event does not release the original defendant. If a car accident sends someone to the emergency room and a doctor commits a routine treatment error, the driver typically remains liable. Medical complications are a predictable consequence of causing an injury that needs treatment.
The chain breaks when an intervening event is so extraordinary that it qualifies as a superseding cause. A random act of intentional violence by a stranger or a sudden natural disaster can qualify. Context matters, though. A landlord who neglects security in a high-crime building cannot credibly claim a break-in was unforeseeable. The more predictable the event, the weaker the argument that it supersedes the original negligence.
When Multiple Parties Cause the Same Harm
The but-for test works cleanly for one defendant and one injury. It struggles when several independent acts combine to produce a single, indivisible harm, and courts have built doctrines to handle those situations.
Substantial Factor
When two or more causes merge into one injury, courts ask whether each defendant’s conduct was a “substantial factor” in producing the harm.6Legal Information Institute. Substantial Factor Test The contribution doesn’t have to be the sole or primary cause, but it must be more than trivial.
The textbook example is two fires started by different people that merge and burn down a house. Each fire alone would have been enough, so the but-for test paradoxically lets both defendants off. The substantial factor test fixes that. The same logic appears in toxic exposure cases, where a worker encountered hazardous materials from many manufacturers over years and no one can pinpoint which exposure caused the illness. When multiple defendants each qualify as a substantial factor, courts often impose joint and several liability, which lets the plaintiff collect the full judgment from any one of them.7Legal Information Institute. Joint and Several Liability
Alternative and Market Share Liability
A related problem comes up when several defendants acted negligently but only one actually caused the injury and the plaintiff can’t tell which. Under alternative liability, the burden shifts to the defendants: each must prove it was not the cause, and if none can, all are jointly liable.
In rare cases involving identical products from multiple manufacturers, some courts apply market share liability. The plaintiff who cannot identify which company made the specific product that caused the injury recovers from each manufacturer in proportion to its share of the market.8Legal Information Institute. Market Share Liability Only a handful of states recognize this theory, and it applies almost exclusively to interchangeable defective products.
When the Plaintiff Is Partly at Fault
Causation analysis also looks at the plaintiff. If the plaintiff’s own conduct contributed to the injury, most states apply comparative negligence, reducing recovery in proportion to the plaintiff’s share of fault.9Legal Information Institute. Comparative Negligence A jury that finds you 30% responsible for a $100,000 injury awards you $70,000.
The specifics vary. A 50% bar rule denies any recovery to a plaintiff who is 50% or more at fault. The 51% bar sets the cutoff at 51%. Pure comparative negligence states let a plaintiff recover even when they bear most of the blame. A small number of states still follow contributory negligence, which bars recovery entirely if the plaintiff was even 1% at fault.9Legal Information Institute. Comparative Negligence
Causation in Criminal Cases
Criminal law uses the same two-step framework but with a much higher burden of proof. A civil plaintiff proves causation by a preponderance of the evidence, meaning more likely than not.10Legal Information Institute. Preponderance of the Evidence A prosecutor must prove it beyond a reasonable doubt.11Legal Information Institute. Beyond a Reasonable Doubt
The prosecution does not have to show the defendant’s act was the only cause. If two people independently attack a victim and either attack alone would have killed them, both can be convicted of homicide, because each act was a sufficient cause. The but-for test is modified in these concurrent-cause situations so defendants cannot use each other’s conduct as a shield.
Intervening events can break the chain in criminal cases too, but the bar is high. The event must be truly unforeseeable and must be the sole direct cause of the final harm. A victim’s pre-existing vulnerability does not break the chain, which is why defendants are routinely convicted of homicide when the victim died only because an underlying condition made an otherwise non-fatal assault lethal.
Proving Causation at Trial
Establishing causation in court often comes down to expert testimony, especially in medical injury, toxic exposure, and complex accident cases. Courts act as gatekeepers to keep speculation off the stand.
The Daubert Standard
Federal courts and most state courts evaluate expert testimony under Daubert v. Merrell Dow Pharmaceuticals, Inc. The trial judge assesses whether the expert’s reasoning is scientifically reliable before the jury hears it. The Supreme Court identified several factors: whether the theory has been tested, whether it has been peer-reviewed, its known error rate, and whether it has gained acceptance in the relevant scientific community.12Justia. Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579
Federal Rule of Evidence 702 codifies this gatekeeping role. The party offering an expert must show that it is more likely than not that the expert’s knowledge will help the jury, that the testimony rests on sufficient facts, and that the expert has reliably applied sound methods.13Legal Information Institute. Federal Rules of Evidence Rule 702 – Testimony by Expert Witnesses
The Frye Standard
A minority of states still follow the older Frye standard, which focuses on a single question: is the expert’s methodology “generally accepted” in the relevant scientific field? Frye is narrower than Daubert and cares about scientific consensus rather than a broader reliability inquiry. In Frye jurisdictions, novel or untested techniques face a higher bar because they haven’t yet achieved widespread acceptance.
Differential Diagnosis
In medical malpractice and toxic tort cases, experts often establish specific causation through differential diagnosis. The expert identifies all plausible causes of the plaintiff’s condition, then rules out alternatives until the most probable cause remains. Courts scrutinize whether that process follows the same rigor expected in clinical practice. An expert who works backward from a litigation theory rather than forward from the evidence is likely to be excluded. Causation opinions built on speculation fail under both Daubert and Frye, and losing a causation expert usually means losing the case.