Concept
When Causation Matters
Actual causation (but-for cause)
When Causation Matters
The but-for test of § 2.03(1)(a) measures conduct against "the result in question," which means it has work to do only when the offense has a result to measure against.1 Whether an offense has one depends on how it is defined. Some crimes are complete once the defendant performs the forbidden act with the required mental state, and nothing further has to happen. Perjury is committed by a material lie told on the witness stand with intent to mislead, even if the lie is so transparent that no one believes it and the outcome of the case is unaffected. Forgery is committed by making a false writing with intent to defraud, even if no one but the writer ever sees it, and reckless or careless driving is a crime although no one is harmed.2 These are conduct crimes, offenses whose definition never depends on an outcome. Because no harm must be produced for the offense to exist, there is no link between act and harm for the prosecution to prove, and causation drops out of the case.
Other crimes are written so that a specific outcome is part of the offense, and these are the result crimes. Intent-to-kill murder requires a death, intent-to-injure battery requires a bodily injury, malicious mischief requires an injury to property, arson requires a burning, and false pretenses requires that the victim actually pass title to property.3 The required mental state varies without changing the point. Manslaughter and battery can rest on reckless conduct that produces death or injury, depraved-heart murder on an extreme degree of recklessness that causes death, and some modern driving statutes on negligence that causes death or injury.4 A Florida appellate court stated the general rule in Velazquez v. State, observing that vehicular homicide resembles murder, manslaughter, aggravated battery, and arson in making a specified result caused by the defendant's conduct an essential element, so that there can be no liability unless the conduct was a cause in fact of that result.5 The wording of a statute often shows which kind of offense it is. Washington, for instance, defines promoting a suicide attempt as knowingly causing or aiding another person to attempt suicide, which writes a causal verb into the definition of the offense itself.6
A result crime requires more than culpable conduct followed by the forbidden result. Where A shoots at B intending to kill but misses, and at that moment B drops dead from a cause wholly unconnected to the shot, the two ingredients of murder, an intentional act and a death, exist at the same instant, yet A is not guilty of murder.7 What is missing is the causal connection between the conduct and the result, and that connection requires more than coincidence of time and place.8 The same point emerges from an illustration in which Jake, hoping to inherit from his wealthy Aunt Bea, makes a doll bearing her name and pushes a needle through its chest, and the next day Bea dies of a heart attack.9 Jake acted voluntarily, with intent to kill, against another person, and a death followed, so every element of murder seems present. The needle contributed nothing to the events that led to the heart attack, which means his conduct was not a cause in fact of her death. In both illustrations the culpable act and the result sit side by side in time, yet the act made no difference to when or how the result came about. That absence of difference is what the but-for test, taken up in the next section, is built to detect.
Civil courts reason the same way about proof of causation, and the criminal standard is higher still. In Kramer Service, Inc. v. Wilkins, a hotel guest was cut on the forehead by falling glass and two years later was found to have a skin cancer at the site of the wound. One physician put the chance that such a trauma would produce a cancer at one in a hundred, and the other denied any causal connection between trauma and cancer.10 The Mississippi Supreme Court rejected the idea that proof a past event possibly caused a result is enough to reach a jury, and it added that negligence and injury merely coexisting does not make the injury a product of the negligence.11 The inference the court condemned is the one known by its Latin name, post hoc ergo propter hoc, which treats an event as caused by an earlier one simply because it came afterward. A landowner whose water well goes dry after seismic blasting nearby generally cannot rest on the coincidence of the blasting and the damage either. Circumstantial proof has to show the conditions that existed beforehand and make it highly improbable that the injury would have occurred but for the operations.12
Two separate reasons explain why causation is its own element rather than something assumed whenever harm follows conduct. The first is fairness in attribution. Without a causal link, the harm may have happened even if the defendant had done nothing, and punishing the defendant for it would rest on something other than what the defendant did.13 The second explains why outcomes matter at all. Criminal law punishes culpable conduct that produces harm more severely than the same conduct that only risks it, and the principled case for that practice is contestable, since punishing only the risk-takers whose conduct happens to cause harm makes liability something of a lottery.14 The commentary to Model Penal Code § 2.03 offers a practical defense: jurors will not lightly return verdicts that lead to severe sentences without the resentment aroused by serious injury, and a prudent legislator cannot ignore that.15 Once the law lets severity turn on an outcome, it must say when an outcome belongs to a particular defendant. Harm itself is a visible fact, but responsibility for it depends on a judgment about which of the endless antecedent conditions should count as causes.16 Making causation an element that the state must prove forces that judgment to be justified in every result case.
Because causation is an element of a result crime, the prosecution must prove it beyond a reasonable doubt, as it must prove every other element.17 That proof has two steps, actual cause and legal cause. The first asks whether the conduct was a but-for cause of the result, and the second asks whether the result and the way it came about were close enough to what the defendant intended or risked to make responsibility fair.18 In the ordinary case the first step is easy. A, intending to kill B, a healthy person, aims at B's heart and fires, no one else is trying to kill B, and B dies within moments, so that but for A's conduct B would have lived and the death is exactly the one intended. The same is true of a driver who recklessly crashes into another car and kills its driver, because the death falls within the risk the recklessness created.19 The cases that need attention are the unusual ones, such as a second assailant who also shoots the victim or a victim whose preexisting illness turns a modest wound fatal. They are taken up in the sections ahead, beginning with the mechanics of the but-for test.