“Is a tort a failure to do what one ought?,” asks the South African scholar Leo Boonzaier. In this book chapter, Boonzaier provides an insightful analysis of the question, which he frames as follows. A distinguishing feature of many non-instrumentalist theories of tort law is how they conceive of a tort—as a wrong, not merely as a cost or loss that the law has an interest in efficiently deterring through a compensatory remedy. But what does it mean to characterize negligent conduct or an intentional battery or a defamatory statement as a wrong?
Here is a very appealing answer: “The commission of a tort is a failure by the tortfeasor to do what, in the law’s view, he ought to do, all things considered.” (P. 169.)1 This answer helps explain why negligence is defined as unreasonable conduct, and why tort law sometimes awards injunctions to prevent the commission of torts and sometimes awards punitive damages. Note that the failure in question is the unjustifiability of the actor’s conduct, not the culpability of the actor. After all, tort law employs objective tests and does not recognize excuses: “one may blamelessly fail to do what one ought.” (P. 170.)
Yet this answer, Boonzaier points out, confronts significant counterexamples. In necessity cases such as Vincent v. Lake Erie Transportation Co.,2 the ship’s captain trespassed on the plaintiff’s property during a storm and was justified in doing so, yet he had to compensate the plaintiff. And in some nuisance cases, courts award compensatory damages yet refuse to enjoin the conduct. For example, in the British case Miller v. Jackson, a British court upheld nuisance liability based on numerous balls from a cricket field striking plaintiff’s adjoining property, but the court refused to award an injunction because of the substantial recreational value of cricket; only a damage remedy was warranted.3 It seems impossible to explain such examples if a tort is defined as conduct that the actor should not have engaged in. Although the trespasser and the creator of the nuisance do not satisfy that definition, tort law requires them to pay compensation.
How to explain these counterexamples? One solution proposed by some scholars is to characterize tort law as having two tracks.4 One track embraces the definition, but a second track explains the counterexamples on entirely different grounds. The first track includes most torts, while the second track offers a potpourri of quite different rationales for the small number of torts that the first track cannot explain. Thus, some would categorize Vincent as an instance of unjust enrichment, or the nuisance cases as reflecting property law rather than tort law principles.
Having nicely framed the problem, Boonzaier critiques the twin-track solution. The suggested explanations for the counterexamples are, he claims, ad hoc and inadequate. A theoretically satisfactory account of tort law should, he claims, find a unifying logic that embraces both the standard cases and the counterexamples. Boonzaier argues that another cricket case, Bolton v. Stone,5 reveals that logic. In Bolton, plaintiff was struck on the head by a cricket ball driven out of the cricket ground. The court rejected negligence liability due to the very low risk of injury, but one judge believed that compensation was morally required even in the absence of negligence, and many commentators criticized the no-liability result. Here is Boonzaier’s analysis:
[T]he prevailing doctrine, to which the Lords ultimately acceded, coupled tort liability with unreasonable conduct. But if the Lords had uncoupled it [and imposed liability], they would have made an equally viable choice, and quite possibly one that would have been less controversial.
(P. 186.) In Boonzaier’s view, it would be “rationally intelligible,” even if not the best rule, to impose liability in Bolton, especially in light of the imposition of nuisance liability in the first cricket case. Boonzaier then tentatively concludes that the “failure to do what one ought” standard is a sufficient ground of liability that can nevertheless be defeated by other values, especially the value of the defendant’s liberty. (P. 190.)
There is much to admire in Boonzaier’s well-written, carefully argued chapter, which offers a fresh perspective on well-trod issues. Let me now suggest some questions and additional thoughts about his analysis.
First, I commend Boonzaier for emphasizing the contrast between the liability result in the nuisance cricket case and the no liability result in the negligence cricket case, but more could be said. The contrast illustrates a doctrinal difference between nuisance and negligence, and between trespass and negligence, that is worthy of further study. When an actor knows or should know that her conduct is highly likely to cause repeated invasions of the property rights of a neighbor, she is potentially liable for nuisance; and when she knows that a single act will damage another’s property, she is potentially liable for trespass. But when she lacks such actual or constructive knowledge, and merely poses a small risk of property damage, at worst she is negligent. And in the former case but not the latter, a court is likely to require compensation even though her conduct was justifiable.6
Why the difference? Principles of fairness (e.g. based on imposing a nonreciprocal risk or obtaining a nonreciprocal benefit) arguably support liability in both cases. Perhaps the answer can be found in Boonzaier’s brief reference to the liberty of defendants. The lower-risk activities that negligence law governs are much more prevalent than the higher-risk activities that nuisance and trespass govern. Thus, expanding strict liability beyond the latter to the former might substantially burden the liberty of actors whose activities create risk.
Second, the doctrinal schizophrenia that Boonzaier identifies is arguably more widespread than he suggests. Boonzaier decides not to categorize traditional strict liability categories (such as the Rylands v. Fletcher rule) as counterexamples to the “failure to do what one ought” conception of tort wrongs. (P. 176.) But this choice depends on a controversial conception of “ought.” Invoking Derek Parfit, Boonzaier relies on the “fact-relative” conception—that is, whether the actor’s conduct was justifiable given the facts about what actually transpired. But it is problematic to employ this sense throughout tort law. A determination that an actor is negligent, for example, relies on an ex ante perspective, not an “actual facts” perspective.7 Suppose I flip a switch in a room that, as it turns out, has poor wiring, resulting in harm to a neighbor.8 If I do not know and should not reasonably know about the wiring, I am unjustified in flipping the switch in the fact-relative sense, but I am not negligent. On an ex-ante understanding of “ought,” traditional strict liability categories such as abnormally dangerous activities, liability for dangerous animals, and product manufacturing flaws are indeed counterexamples to the “ought” conception of tort law.
To be sure, there is a difference between these traditional strict liability categories and liability under the Vincent necessity rule. In the former cases, but not the latter, the actor’s conduct on the particular occasion that caused harm is unjustifiable and regrettable ex post. If defendant knew that the dynamite blast would kill the victim, or that the soda bottle with the flaw would explode, defendant should not have set that blast or sold that bottle. But in Vincent, there is nothing unjustifiable, even ex post, in damaging a dock in order to save a ship. If the same fact pattern were to arise again, it would again be justifiable to cause the damage. Whether this distinction requires that different principles be invoked to justify strict liability is a fertile question for inquiry.
Consider also a third, related issue: the problem of reasonable mistakes. Sometimes, even if a mistake about an element of a tort is reasonable, tort law ignores the mistake and imposes liability. For example, when a property owner makes a reasonable mistake in determining the boundary between their property and their neighbor’s, the owner is still liable for trespass. But sometimes a reasonable mistake precludes liability, as when X using defensive force makes a reasonable mistake in concluding that the plaintiff was threatening force, or when Y makes a reasonable mistake in concluding that the plaintiff consented to a physical touching. Yet Boonzaier, like many scholars, treats reasonable mistakes as excuses, not justifications. On this understanding, it would seem that X and Y should be liable, for they ought to have acted differently. By contrast, some scholars treat (many) reasonable mistakes as justified, not merely excused, thus supporting a no-liability outcome.9 Once again, the fact-relevant conception of “ought” is more controversial than the chapter suggests.
I hope that Boonzaier explores some of these questions in future work. In the meantime, this illuminating chapter is an impressive exploration of fundamental questions about the nature and scope of tort liability.
- Boonzaier’s qualification “in the law’s view” helpfully differentiates this sense of “ought” from a purely moral sense.
- 124 NW 221 (Minn. 1910).
- [1977] QB 966. The American counterpart is the famous case Boomer v Atlantic Cement Co., 257 N.E.2d 870 (N.Y. 1971).
- See John Goldberg & Benjamin Zipursky, Recognizing Wrongs 189-198 (2020) (treating abnormally dangerous activities as an anomalous licensing-based category of tort liability that departs from the norm that torts are wrongs); Gregory Keating, Strict Responsibilities, in Reasonableness and Risk (2022) (treating strict liability as one of two types of wrong, the conditional wrong of harming-without-repairing and the wrong of violating a person’s right or sovereignty, that are distinct from fault-based liability).
- [1951] AC 850.
- Suppose, in Vincent, that the ship owner only created a small risk of damage to the dock and nonetheless damaged it. The owner would probably not be liable either for trespass (because of lack of intent) or for negligence (because of the overriding necessity).
- In Parfit’s terminology, the “ought” applicable to negligence is evidence-relative—that is, wrongful in light of the evidence available to the actor—rather than fact-relative (in light of the actual facts) or belief-relative (in light of the actor’s beliefs). Derek Parfit, On What Matters (2011).
- This is a famous example offered by philosopher Judith Thomson, in The Realm of Rights 229 (1990). In later work, she rejected her earlier view that the actor “ought” not to flip the switch. Judith Thomson, Normativity 198 (2008).
- See, e.g., Kenneth W. Simons, Self-defense, Necessity and the Duty to Compensate, in Law and Morality, 55 San Diego L. Rev. 357, 374-377 (2018); R. A. Duff, Rethinking Justifications, 39 Tulsa L. Rev. 829 (2004).
