Concept
Elements of the Doctrine
Alternative liability when the responsible defendant cannot be identified
Elements of the Doctrine
Section 433B(3) of the Restatement (Second) of Torts states alternative liability in a single sentence: where the conduct of two or more actors is tortious, and it is proved that harm has been caused to the plaintiff by only one of them, but there is uncertainty as to which one has caused it, the burden is upon each such actor to prove that he has not caused the harm.1 Courts have unpacked that sentence into conditions the plaintiff must meet before the burden moves. The court in Conn. Interlock Risk Mgt. Agency v. Jackson lists three: all of the defendants acted negligently and harm resulted, all possible tortfeasors have been named as defendants, and the tortfeasors' negligent conduct was substantially simultaneous in time and of the same character so as to create the same risk of harm.2 Another formulation lists four elements: the plaintiff cannot identify which defendant caused the injury, all potentially liable parties are joined, each defendant breached a duty of care owed to the plaintiff, and the defendants are in a far better position than the plaintiff to offer evidence of causation.3
Courts guard these conditions closely because the shift is expensive. It asks each defendant to disprove a fact the plaintiff would ordinarily have to prove, and a defendant who fails risks paying for a harm he did not cause. Joint and several liability, moreover, is inherent in the application of the doctrine.4 In Michigan, defendants unable to exculpate themselves were held jointly and severally liable, and after the legislature restricted joint and several liability a federal court concluded that the change precluded alternative liability in a case against pesticide manufacturers.5 Each condition therefore blocks a different route to an unfair result. The first confines the shift to wrongdoers, the second limits the missing proof to a question of identity, the third keeps the true cause inside the courtroom, and the fourth keeps the odds among the defendants close to even. The hunting accident satisfied all four, which is why it remains the measure for later cases.
Tortious Conduct by Every Defendant
The burden moves only to defendants the plaintiff has already proved to be wrongdoers. Section 433B(3) applies where the conduct of two or more actors is tortious,6 and in a negligence case that means each defendant failed to use reasonable care toward the plaintiff. The plaintiff must establish that failure by a preponderance of the evidence before the burden of proof on causation shifts.7
The reason lies in what the exception is for. It rests on the unfairness of letting several tortfeasors who acted at the same time escape liability merely because their conduct and the resulting harm make it hard for the plaintiff to show which of them caused the harm.8 Summers made the same point in its own terms. The hunters were both wrongdoers, both negligent toward the plaintiff, and they had put the plaintiff in the unfair position of having to point to which of them caused the harm, so it should rest with each to absolve himself if he can.9 That argument works only for defendants who were careless. A defendant who may have used due care did nothing to create the plaintiff's difficulty, and making him disprove causation would charge him with an injury that no wrongdoing of his may have produced. The California Supreme Court later singled out this feature of the hunting accident: each hunter was a wrongdoer who had negligently fired toward the plaintiff at about the same time, and no fact distinguished the nature or extent of their negligence.10
Courts have refused the doctrine when proof of tortious conduct was missing. One court declined to apply the rule because a plaintiff injured in an automobile accident never contended that both defendant drivers were negligent. In Goldman v. Johns-Manville Sales Corp., the Ohio Supreme Court explained that the theory relaxes only the requirement that the plaintiff show which defendant caused the injury, and that the relaxation is warranted only when the plaintiff shows that all defendants acted tortiously.11
Tort law has no rule that injured plaintiffs must always win. A plaintiff recovers only by proving every element of the claim, and, Summers or no Summers, the uncertainty of making that proof generally falls on the plaintiff.12 The doctrine relaxes one element, the identity of the actor who caused the harm, and it does so only after the plaintiff has proved that each defendant acted tortiously.
A second failure arises when the plaintiff can show wrongdoing by at most one defendant. In McLaughlin, the plaintiff sued two distributors and invoked alternative liability, but the court granted the distributors summary judgment. It reasoned that the Restatement's requirements were not met: because the case alleged a manufacturing defect, one distributor acted tortiously and not both, so there was no justification to join both.13 The reasoning tracks the nature of the claim. A manufacturing defect is a flaw in a particular product, so at most the distributor who sold that product was a tortfeasor, and the plaintiff's doubt ran between a wrongdoer and a stranger. In Summers the doubt ran between two men who had each committed the same dangerous act.
Both failures confuse two burdens. Proving that each defendant acted tortiously is the plaintiff's burden under the ordinary elements of the claim. Proving which defendant's conduct caused the harm is the burden that moves once the first has been carried. A plaintiff who runs the two together treats an unexplained injury as if it proved breach, when it shows only that the identity of the cause is unknown. A torts-text problem makes the point from the other side: the principle of Summers could not apply until the plaintiff proved that both defendants were negligent.14
One Cause, Unknown Source
The second element asks the plaintiff to prove that exactly one defendant caused the harm. Section 433B(3) applies where it is proved that the harm was caused by only one of the actors, but there is uncertainty as to which one,16 and courts treat the first half of that clause as a presumption the plaintiff must satisfy: the plaintiff must prove that all defendants acted tortiously and that the harm resulted from the conduct of one of them.17 The second half names the gap that remains. The cause of the harm is certain, and only its source is unknown, so the problem is one of identity.
The hunting accident shows the structure. The one shot that entered the plaintiff's eye was the major factor in assessing damages, and it could not have come from the guns of both defendants; it came from one or the other only.18 The California Supreme Court, recounting the case, drew out features that go to this either/or structure. The injury was instantaneous and indivisible rather than latent and progressively deteriorating. No contributing or concurrent causation existed, because one hunter was the cause in fact of the entire injury. And the plaintiff had no evidentiary means whatsoever of showing from which hunter's shotgun the pellet had been fired.19
The first two questions guarantee that the dispute really is either/or. A harm that develops over time, like a disease that follows repeated exposures, invites the possibility that several sources each added something, and the question becomes how much each contributed and not which one acted. A single indivisible injury leaves no room for that. In Summers one hunter caused all of the harm from a single pellet, and no factual basis existed on which to apportion fault.20 The structure also removes the shift's worst risk. When exactly one defendant caused the harm and every possible actor is before the court (a condition taken up below), the group necessarily contains the true cause, so the shift reaches a wrongdoer and never a stranger to the harm.
The third question explains why the plaintiff receives help at all. The California court described a shift in the burden of proof on causation as a fundamental departure from ordinary principles, justified only by a showing of necessity.21 A plaintiff struck by a single pellet from a gun he never saw fired has no evidence to find. Without the shift both tortfeasors would have escaped, and the plaintiff would have had no legal means to seek redress for a negligently inflicted injury.22 Summers put the point more bluntly: if one defendant can escape, the other may also, and the plaintiff is remediless.23
The structure matches Summers, with produce in place of birdshot. The cause of the illness is known, and only its source is unknown. One tomato cannot have come from both farmers, just as one pellet could not have come from both guns, and the customer cannot trace a tomato already eaten. The farmers are no better placed. Since neither can show whose tomato caused the harm, both will be held liable and will split the loss under contribution principles if both are solvent.24 The result may lack intellectual elegance, but it supplies the rough justice that courts must often settle for in real disputes.25
Summers gave one further reason for placing the burden on the defendants: ordinarily they are in a far better position than the plaintiff to offer evidence on which of them caused the injury.26 One formulation therefore lists superior access to causation evidence as a fourth element.27 The tomato farmers strain that formulation, since neither knows more than the customer about the tomato he ate. In Sindell v. Abbott Laboratories, a DES case, the court concluded that the defendants' lack of greater access to information identifying the manufacturer does not per se prevent application of the Summers rule.28 The shift can survive without superior access because its central rationale is fairness. Where one of several negligent defendants must have caused the harm, the risk that no one can say which should fall on the negligent defendants and not on the plaintiff.29 Superior access is best treated as the usual explanation for the shift, and not as a requirement that decides cases by itself.
Two kinds of case look like Summers but fall outside it, and each fails the either/or test in a different way.
On the surface the scene resembles Summers, since both defendants were negligent and the plaintiff cannot say which one caused the injury. The causal structure differs. In Summers one hunter caused all of the harm and the other none, whereas here each collision probably caused part of the disk injury. That is an apportionment dilemma and not an identity problem, and most courts handle it differently. Once the plaintiff proves both defendants negligent and an injury difficult to apportion, the burden shifts to the defendants to show what part each caused, and those who cannot separate the damages are held jointly liable for all of it or severally liable in proportion to their fault.30
The second case is one in which the evidence lets the jury decide in the ordinary way.
The evidence in the plum case does the sorting that the shift would otherwise do, so no identification gap remains. Courts invoke the shift only on a showing of necessity,31 and in more typical multi-defendant cases the jury has some basis for a more-probable-than-not finding of who caused the harm, so a court is unlikely to invoke it.32
Joinder of All Actors
Section 433B(3) places the burden on each actor to prove that he did not cause the harm, but it says nothing about who must be sued.34 The joinder condition comes from the commentary on the rule. The commentary observes that every case applying it had involved actors who were all joined as defendants, and it warns that modification may be needed where one of the actors is not or cannot be joined, or where the actors' conduct or the risks they created differ substantially.35 Courts have turned the observation into a requirement. The Connecticut court lists as its second condition that all possible tortfeasors be named as defendants, and the opinion in Smith v. Cutter Biological, Inc. states joinder of all responsible parties as the second of two presumptions that follow from the theory.36
The hunting accident met the condition on its facts. The single pellet could not have come from both guns, so it came from one hunter's gun or the other's, and both hunters were before the court.37 The California Supreme Court listed this first among the features that made the shift fair: all the tortfeasors were named as defendants, one of them certainly caused the eye injury, and there were no other potential tortfeasors.38 The plaintiff did not have to wonder whether a stranger to the lawsuit was the shooter.
The condition serves two purposes. The first concerns what the shift decides. It answers a single question, which defendant in the group caused the harm, and it assumes that the answer lies somewhere in the group. The Connecticut court explained that naming all possible tortfeasors eliminates from the jury's consideration the theory that some other cause, besides a joined defendant's conduct, caused the injury. Without that step it would not be fair and equitable to relieve the plaintiff of proving which tortfeasor caused the harm.39 Drop the condition and a defendant who cannot disprove causation might pay for an injury that an absent actor inflicted.
The second purpose concerns the odds. The rule asks each defendant to prove a negative and accepts that one may pay for a harm he did not cause, so it is defensible only if each defendant has a realistic chance of being the one who did it. Every unnamed actor lowers that chance for every named defendant. The defendants in Sindell made the point with numbers. In Summers there was a 50 percent chance that one of the two defendants was responsible for the plaintiff's injuries, whereas any one of 200 companies that manufactured DES might have made the product that harmed the plaintiff.40
Sindell shows the condition failing. The court held that the plaintiff could not use the Summers rationale to fix the whole liability for her injuries upon the defendants, at least as those principles had previously been applied. In Summers all the parties who were or could have been responsible for the harm had been joined as defendants, whereas about 200 drug companies made DES, any of which might have made the injury-producing drug.41 The theory courts developed to answer that difference, market share liability, is taken up below.
The commentary leaves room for a court to modify the rule where an actor cannot be joined.42 A court that did so would be departing from the rule as it is usually stated, and it would have to explain why the shift remains fair after the odds have fallen and an outside cause has become possible.
The settings in which courts have found the doctrine useful are those where the group of possible actors is small and visible: multiple-car collisions, pollution by several defendants, and injuries during an operation on a sedated plaintiff.43 The participants in each were present when the harm occurred, so naming all of them is realistic. A plaintiff hurt in a pileup who sues every driver in it satisfies the condition. A plaintiff who sues a handful of the many makers of a mass-marketed drug has the Sindell problem.
Similar, Simultaneous Conduct
Section 433B(3) says nothing about when the actors must have acted or what they must have done, yet courts add a condition on both points. The Connecticut court lists as its third requirement that the tortfeasors' negligent conduct be substantially simultaneous in time and of the same character so as to create the same risk of harm, a requirement it traced to comments f and g of § 433B.45
The requirement protects the odds. The court explained that it would be unreasonable to require defendants to absolve themselves from liability unless “the likelihood that any one of them injured the plaintiff is relatively high.”46 Similar conduct keeps the chances level, because when defendants did the same thing nothing in the evidence ranks one above another. Simultaneity works alongside it. Defendants who acted at different times usually leave different evidence, and different evidence lets a jury make an ordinary more-probable-than-not finding about who caused the harm.
The hunting accident in Summers supplies the measure for both halves. The two hunters negligently fired their shotguns toward the plaintiff at about the same time, and no facts distinguished the nature or extent of either hunter's negligent conduct; they were coequals from the standpoint of fault.47 Because the defendants did exactly the same thing, the chances were equal that either caused the harm, and there was no logical way to decide which one had.48
Acting alike at the same moment is not the same as acting together. The concert of action theory imposes collective liability when, under a common plan or scheme, a party assists, participates with, or actively encourages one or more wrongdoers who tortiously harm the plaintiff. The classic cases involve drag races, in which the driver of the car that did not hit the plaintiff is also liable for helping create the risk.49 Alternative liability needs no agreement or encouragement among the defendants. It turns on whether the defendants stand on equal footing as to fault and timing, so that no evidence tells their conduct apart.
The same hour does not equalize the odds when one actor created a much larger risk, which is why the conditions join timing to character. The commentary reflects the same concern when it says the rule may need modification because of “substantial differences in the character of the conduct of the actors or the risks which they have created.”50 A student who finds negligence by every defendant and an unknown source must therefore still compare how dangerous each defendant's conduct was.
Courts disagree over how strictly to insist on simultaneity. The opinion in Smith reports that the first presumption has been interpreted to mean that the tortious actions must occur simultaneously, and it cites a federal district court considering an asbestos-related injury. It adds that another court, applying the theory in a Factor VIII case, disagreed.51 Some formulations leave simultaneity out of their list of elements and treat timing and similarity under the commentary's reference to modification.52 The Connecticut standard asks for conduct that is substantially simultaneous, and the hunters fired at about the same time, so closeness and not the identical instant is the usual measure.53 Because the cases leave the point open, the dependable approach is to treat both halves as measures of one question, the odds. The closer the defendants' acts are in kind and in time, the stronger the case for shifting.