Necessary But Not Sufficient

Necessary But Not Sufficient

The but-for test shows that the defendant's conduct was one of the conditions the result needed, and the earlier sections showed how to run it and when it clears a defendant. Satisfying it still leaves causation unfinished. Section 2.03(1)(a) treats conduct as a cause when the result would not have occurred without it, and that clause states only the first requirement.¹ A second filter, usually called legal cause or proximate cause, asks whether the result that actually occurred resembles the result the defendant intended or risked closely enough, in kind and in manner, that holding the defendant responsible is fair.² The Code frames the same question as whether the actual result is too remote or accidental in its occurrence to have a just bearing on the actor's liability.³ The law needs this second step because the first has no natural stopping point.

Every result has an unbounded chain of antecedent conditions, running back through events that seemed unimportant when they happened. The standard illustration is a robbery in which Lewis kills Joe, a liquor store clerk. Lewis was conceived at a particular moment, and had his parents conceived a child earlier, later, or not at all, Lewis would probably never have existed to kill Joe, so the conception is a condition without which the death would not have occurred.⁴ By the test as stated, the conception is a but-for cause. Yet no one would describe the conception of Lewis as the killing of Joe, and common sense refuses to count every necessary condition as a cause.⁵ The test is therefore over-inclusive: it admits conditions the law has no wish to treat as causes, and it offers no way of deciding which ones to exclude.

Over-inclusion also appears in realistic cases. In 1994 in Buffalo, 71-year-old Eleanor Szablicki suffered a fatal heart attack after discovering that burglars had ransacked the home she had lived in for 48 years, and the medical examiner ruled the death a homicide.⁶ In Oklahoma City in 1984, 67-year-old Mildred Pruner suffered a fatal heart attack in court as she described being robbed at gunpoint, and Oklahoma's chief medical examiner likewise ruled her death a homicide.⁷ The but-for test is met in both. Because the result tested is the death as and when it actually occurred, each victim would not have died at that moment without the crime.⁸ Whether the burglar or the robber 'killed' the victim is a different question, and the sources that report these deaths leave it open, so the safe statement is that the but-for link exists and that legal cause must decide the rest.

The two burglaries differ in one fact: whether the burglary played any part in bringing on the fatal attack. In the Buffalo case it did, so deleting it changes the result and the analysis moves on to legal cause. In the invented version it did not, so the analysis stops at the threshold, which is where the earlier rule on results that would occur anyway operates. This section is about what happens to the defendants who pass that threshold.

Because the test cannot choose among necessary conditions, something else has to. Human decisions affect the time and circumstances of most deaths, yet the law singles out only a small number as homicides, meaning deaths caused by human action, and that selection is primarily a normative judgment about which necessary conditions deserve the name of cause.⁹ Courts do not ask what a philosopher would call the cause, since a philosopher may treat the cause as the sum of every antecedent. They ask which antecedent the law should treat as the one for which the defendant answers, and considerations of fairness bear on the answer.¹⁰

Two features keep that judgment manageable. The first is procedural: the inquiry starts with the defendant who is charged, so the offender's mother does not enter the case, much less more distant ancestors.¹¹ The second is substantive: a defendant who was a but-for cause of the harm is not responsible unless the conduct was also the imputable, or legal, cause.¹² Each limit has its own reason. Cause in fact is demanded first because punishing a result that would have happened even had the defendant done nothing is arbitrary, since the punishment would not rest on anything the defendant did.¹³ Legal cause is demanded second because, without it, a defendant would answer for any result that his conduct happened to precede, however coincidental the link.

The last step is the one most often skipped, and a real decision shows it doing independent work. In King v. Commonwealth, the defendant and a copilot were carrying 500 pounds of marijuana in a light plane when they became lost in fog and crashed into a mountainside, killing the copilot, and the defendant was convicted of felony murder. The Virginia Court of Appeals reversed: the crash would not have occurred but for the drug felony, so the but-for requirement was met, but the crash was not a foreseeable result of the felony because the contraband cargo did not make it more likely, and proximate cause was missing.¹⁶ The court added that proximate cause might have been found had the crash resulted from flying at low altitude to avoid detection.¹⁷ Both King and the heart-attack deaths pass the first filter, and what differs is the second. In the burglary deaths the victim reacted to the very wrong committed, so whether to attribute the death is a live question. In King the wrong did nothing to raise the odds of the crash. Students lose this issue when they treat a but-for cause as a legal cause, because King shows that the first can be conceded and the second still defeat liability.

The same shape appears outside homicide, in a field with different purposes. Under the Fourth Amendment's exclusionary rule, a defendant seeking to suppress evidence must show that the authorities would not have obtained it but for the illegality.¹⁸ In Hudson v. Michigan the Court said that but-for causality is "only a necessary, not a sufficient, condition for suppression."¹⁹ The reason given is that suppression must also serve the interests the violated rule protects, so a link that is causal in the logical sense can still be too attenuated to justify exclusion.²⁰ In United States v. Bailey a defendant struggled with an officer after an arguably unlawful arrest, and the officer would not have seen the resistance but for the arrest, yet evidence of the resistance was not excluded.²¹ These are rules of criminal procedure with their own justifications, and they appear here only because the structure matches: a but-for link opens the question, and a second inquiry closes it.

Over-inclusion is only half of the difficulty with the but-for test. The next section takes up the opposite failure. When two independent forces are each enough to bring about the result, deleting either one leaves the result in place, so a literal reading of the test would say that neither caused it, and the law makes an exception for simultaneous sufficient conditions.²²