Introduction

W

hat is a tort, and what is tort law for? On one leading scholarly account, torts are legal liability rules that seek to promote the welfare of society at large by disincentivizing socially suboptimal behavior and distributing the costs of accidents to those who can best bear them. Over the twentieth century, this instrumentalist view of tort law won powerful support in elite American legal culture. But it has never gained much traction in other common law jurisdictions, where judges and legal scholars standardly suggest that private law domains such as tort are “inextricably interwoven” with interpersonal morality. After all, the law of torts deals in concepts — negligence, deceit, assault — that are familiar from ordinary moral life. On the classical, moralist understanding of tort law, this moralized conceptual architecture reflects tort law’s basic normative substance. At its core, the common law of torts is an institutionalized application of “common morality and common sense” — a corpus of legal doctrine formed by courts “articulating inchoate lay ideas of right and wrong.”

Over the last several decades, this classical moralist view of tort law has enjoyed a revival of interest in the American legal academy, helping to revive interest in moralist approaches to the private law more generally. In the rest of the common law world, in both tort and other private law domains, moralism has long been academic and judicial orthodoxy. The intellectual lodestar of contemporary private law moralism, especially in tort, is the most famous opinion in the common law canon: Chief Judge Cardozo’s opinion in Palsgraf v. Long Island Railroad Co.

By now the facts of the case are familiar to us. A man attempted to jump onto a train as it was departing from a Long Island Railroad station. In the course of this attempt, the man was negligently jostled by a train attendant employed by the Long Island Railroad Company, causing the man to drop a package he was carrying underneath his arm. Quite improbably, the package contained some powerful explosives; the resulting explosion caused heavy scales at the other end of the train station to drop onto the head of Helen Palsgraf. Writing for the majority of a divided New York Court of Appeals, Cardozo refused to allow Mrs. Palsgraf to recover.

The lasting significance of Cardozo’s opinion rests less on its conclusion, however, than on its claims about the nature of the negligence tort and the law of torts more generally — claims that have proved highly influential in common law courts as well as in the legal academy. In virtually all legal systems, including civil law jurisdictions such as Germany and Italy, courts refuse to impose negligence liability for utterly unforeseeable injuries, such as the improbable injury suffered by Mrs. Palsgraf. But Cardozo’s opinion vigorously asserted an even more basic obstacle to her recovery than the unforeseeability of her injury. If a plaintiff is unforeseeable to a defendant, the opinion maintains, then the defendant cannot treat her wrongfully — the defendant cannot breach any duty owed to her, cannot commit a wrong against her as opposed to other people (or society at large). In tort law, Cardozo insisted, a plaintiff can only sue “for a wrong personal to her, and not as the vicarious beneficiary of a breach of duty to another.”

For tort scholars who join Cardozo in taking up the “Palsgraf perspective,” this basic doctrinal proposition — the “Palsgraf principle,” as it is sometimes called — reflects and reveals the moral heart of tort law. Tort law is about recognizing and redressing relational wrongs: breaches of duties owed by one private person to another. So unless a defendant has breached a duty to the plaintiff — unless she has been mistreated by his behavior — she cannot recover from him in tort.

That is because tort law, unlike the criminal law or administrative regulation, is not a set of legal liability rules that protect the public interest or incentivize socially desirable behavior. Instead, tort law is “irreducibly relational.” The law lays down relational norms of conduct — legal duties that we owe to other people. A tort such as negligence or battery is a sort of relational legal wrong, which consists in the breach of one such relational duty: “do not injure another person by treating her negligently” (the tort of negligence), “do not harmfully or offensively contact another person” (the tort of battery), “do not make a false and injurious communication about another person” (the tort of defamation), and so on. By laying down relational legal duties and wrongs that recognize relational moral duties and wrongs — duties and wrongs that we acknowledge in everyday moral life — tort law allows us to obtain redress from those who have treated us wrongfully. According to the Palsgraf perspective, that is tort law’s core aspiration and purpose: to recognize and redress relational wrongs.

This Article argues that the Palsgraf perspective is mistaken and proposes a different moral picture of tort law’s normative substance and doctrinal structure. Tort law may be centrally concerned with “common morality,” rather than promoting the welfare of society at large. But torts are not relational legal wrongs, and their purpose is not to recognize or redress relational moral wrongs. Torts are remedial pigeonholes: legal liability rules that identify the complex conditions under which a defendant is morally liable to provide a plaintiff with compensation or other forms of remedial relief. Rather than recognizing a species of relational moral wrongdoing or interpersonal mistreatment, a tort such as negligence is a coarse doctrinal device that identifies instances of “moral wrongdoing for which the offender must pay,” whether or not he has treated the plaintiff wrongfully. Other torts identify forms of behavior that render a defendant liable to pay compensation for resulting injuries although he has not behaved wrongfully at all. In both negligence and the other torts, a defendant may be liable to compensate a plaintiff even if he has not wronged or mistreated her. Contra the Palsgraf perspective, relational moral wrongdoing is not the basis of remedial liability, either in ordinary morality or the law of torts.

What is? Like the ordinary morality that it reflects, the common law contemplates many answers. Many of these answers, however, have a common core: the defendant’s moral responsibility for causing unjust damage to the plaintiff, in the sense of infringing some right against injury that the plaintiff holds. The idea of responsibility for unjust damage does not receive clean expression anywhere in the common law. It receives somewhat more transparent expression in many civil law jurisdictions, where the private law (including tort law) is largely codified in statute. So, for example, the most important tort law clause in the Italian Codice Civile, CC 2043, lays down that “[a]ny intentional or negligent act that causes unjust damage [danno ingiusto] to another person obliges the person who has committed the act to pay damages.” It is standardly supposed in Italian legal thought that a loss suffered by a plaintiff is compensable — that it counts as “damage” rather than mere “loss” — only if it infringes a right, or injures a protected interest, that she holds. There are similar clauses in other civil law codes. Thus, for example, section 823(1) of the Bürgerliches Gesetzbuch (BGB), the German Civil Code, lays down that “[a] person who, intentionally or negligently, unlawfully injures the life, limb, health, freedom, property or some other right of another person is liable to provide compensation to the other party for the damage arising therefrom.”

The civil law codes identify when a defendant is morally responsible for causing unjust damage, such that she is morally liable to compensate her victim, through generally worded legal liability rules such as CC 2043. The common law performs the same function in a more opaque and fragmentary fashion, through its “‘pigeonhole’ system of nominate torts,” a motley array of more specific liability rules such as battery, defamation, and negligence. Many of these causes of action (such as battery and defamation) take their names from species of relational moral wrongdoing, just like the pigeonhole causes of action found in the classical Roman law from which modern civil law codes largely derive. But it is a mistake to infer the character of the law’s moral substance from its outer doctrinal form — to insist, as defenders of the Palsgraf perspective often do, on “taking [tort law’s] structures and doctrines at face value.” Tort law’s formal structure should certainly be taken seriously. But that requires taking seriously the complex and unruly relationship between the law’s moral substance and its doctrinal form. Like the formulary actions in classical Roman law and the writs of action in medieval English common law, today’s common law torts do not wear their moral commitments on their sleeves.

Distinguishing between relational wrongdoing and moral responsibility for causing unjust damage (that is, infringing rights against injury) requires fixing some terms. In ordinary moral discourse, terms such as “duty,” “right,” and “wrong” are polysemous. The term “duty,” as it is most commonly used, refers to a sort of norm of conduct, or standard of behavior, which tells us to refrain from acting in a certain way. A moral duty, so understood, is largely or exclusively an evidence-relative phenomenon: It reflects those reasons that are accessible to an agent from her epistemic and deliberative position. To violate a duty owed to another person is thus to wrong or mistreat her by giving insufficient weight in one’s moral deliberation to moral reasons that reflect her status as a “self-originating source[] of claims.” Sometimes the term “right” is used as a synonym or conceptual flipside of “duty” (and thus the correlative of “wrong”): To violate a duty owed to another person just is to violate her right and to wrong her. This is the usage that Cardozo himself seems to have had in mind when he spoke of duties, rights, and wrongs in Palsgraf, and defenders of the Palsgraf perspective largely follow him in this regard.

There is, however, a different and equally important sense of “right,” one that is perhaps more familiar to moral philosophers and scholars in the civil law tradition than to common law tort scholars. A moral right, in this sense, is a reason for action — a reason that is grounded in facts about the person whose interests and standing the right protects, whether or not those facts are epistemically accessible to an agent. So, for example, it is possible for an agent to infringe another person’s property right innocently and unforeseeably, as by walking across land that does in fact belong to the plaintiff although the defendant could not reasonably have known it. Similarly, to borrow an example from Professor Judith Thomson, an agent might infringe another person’s moral right against bodily injury by flipping a light switch in her home that unforeseeably generates a freak current and electrocutes him.

Moral rights, in this sense, are fact-relative: Whether an agent infringes a fact-relative right, or right against injury, depends on what the agent actually does, not on what she reasonably believes she might do. Thus an agent who infringes another person’s right against injury does not breach any duty to him, or wrong him, except (perhaps) in the extremely thin sense that her action has in fact injured him. Put another way, an agent can infringe another person’s right against injury without in any way mistreating that person or behaving culpably or disrespectfully toward him.

Only in the property torts, such as trespass and conversion, does tort law directly encode and enforce moral rights against injury, and that is largely for contingent historical reasons involving the common law’s use of property torts to perform noncompensatory functions such as settling disputes about title. Most of the time, tort law — like the common sense morality that it reflects — holds a defendant liable to compensate a plaintiff only if the defendant’s infringement of the plaintiff’s right is attributable to and reflective of some morally significant feature of the defendant’s responsible agency (such as, paradigmatically, a culpable and wrongful choice). A great deal of tort doctrine consists in the law’s attempt to articulate, elaborate, and enforce “informed common sense” about “how far the responsibility of the defendant ought fairly to extend” for causing unjust damage to other people, that is, infringing their fact-relative rights against injury.

This Article argues that moral responsibility for causing unjust damage (that is, infringing fact-relative rights against injury) lies at the moral core of tort law — not relational wrongdoing (that is, the breach of relational duties). The following cases may help to provide an initial sense of how the two diverge.

Consider, first, a case in which a defendant is liable for negligently injuring an unforeseeable plaintiff. Suppose the defendant is conducting construction work near a building that appears to be owned solely by Alma, a local celebrity who is famously averse to litigation. To limit his legal risk, the defendant undertakes a laborious investigation of title — so laborious an investigation that it is fantastically improbable, once the investigation concludes, that anyone other than Alma owns the building. If against the odds there does exist another owner of the building, Bessie, and if the defendant foreseeably destroys the building through conducting his construction work carelessly, it seems clear that he is liable to compensate Bessie no less than Alma. That is true although Bessie is not a foreseeable victim, and the defendant has not (therefore) breached any duty to be careful toward her. Instead, the defendant’s liability to compensate Bessie rests on the fact that he has negligently and foreseeably risked causing a certain sort of unjust damage (that is, infringing a certain sort of right), this risk has materialized, and Bessie does in fact hold one of the rights thereby infringed.

The defender of the Palsgraf perspective might respond that Bessie is foreseeable under a suitably generic description (say, “an owner of that building, whomever he or she may be”). But this sense of victim foreseeability seems strained and artificial — and this response makes a hash, therefore, of other explanatory roles that victim foreseeability plays in legal and moral thought. Imagine, for example, that a construction mogul wishes to build a shopping center. To do so, she must engage in some blasting that will likely destroy a nearby (currently uninhabited) residence that reasonably appears to be owned solely by Abelard. Absent exigent circumstances, it is justifiable and nonculpable for the construction mogul to proceed if and only if she is reasonably certain that she has obtained the consent of Abelard. That is because Abelard is the only foreseeable owner of the building. Even if there exists some other, hidden, owner of the building (call him Baldwin), that fact is plainly irrelevant to whether it is justifiable and nonculpable for the construction mogul to risk destroying the residence. It is irrelevant precisely because Baldwin is not, in any natural or independently plausible sense of the term, a foreseeable victim (supposing his existence cannot be ascertained with reasonable diligence).

It is certainly possible to define an unorthodox sense of victim foreseeability under which hidden victims such as Bessie and Baldwin do count as foreseeable. That such a stipulative artifice is required to capture the intuitive moral phenomena, however, suggests that these phenomena are in truth explained by moral principles that do not hinge on victim foreseeability at all. In any event, even if the Palsgraf principle helps itself to this artificial sense of victim foreseeability, the principle cannot ultimately be saved. For the Palsgraf principle is not, at bottom, about victim foreseeability (or plaintiff foreseeability) at all. Rather, the principle maintains that a plaintiff must be wronged — must be treated carelessly or otherwise wrongfully by the defendant — in order to recover. Plaintiff unforeseeability is simply one sort of factor that precludes a defendant from treating the plaintiff carelessly, from breaching a duty to be careful toward her.

The Palsgraf principle is equally impugned, therefore, by negligence cases in which a defendant is liable to compensate a foreseeable plaintiff whom he has not treated carelessly. Imagine, for example, that a defendant runs a business that engages in a lucrative activity that spews toxic chemicals into the surrounding air — creating a significant risk of killing some nearby residents and damaging a neighboring piece of land, Blackacre. It is negligent for the defendant to direct his factory to engage in this activity, for the commercial advantages that he and his business’s shareholders will thereby gain do not justify imposing these risks of personal injury and property damage. Suppose the defendant has excellent reason to believe that the owner of Blackacre (whoever he or she may be) is a large shareholder in his factory — and that he or she has enthusiastically consented to the defendant’s continuation of his chemical-generating activity, notwithstanding the associated risk of damage to Blackacre, because he or she stands to reap a large financial gain (on balance) if this activity continues. It turns out, however, that the defendant’s evidence is misleading; in fact, the owner of Blackacre is not a shareholder in the defendant’s business and has not consented to the defendant’s pollution-generating activity.

If the defendant’s wrongful activity foreseeably damages Blackacre, he is surely liable to compensate its owner. But the defendant has not breached any duty to be careful toward her — has not treated her carelessly as opposed to other people. That is true whether she is identified under a specific description (such as her name) or under some abstract and generic description (such as “the owner of Blackacre” or “the owner of that parcel of land” or “the owner of the parcel of land that my action might foreseeably damage”). After all, the defendant had excellent reason to believe that the plaintiff, even if described in some such generic fashion, consented to his activity. No duty to be careful toward the plaintiff has been breached, and yet the plaintiff may plainly recover. That is because relational wrongdoing toward the plaintiff is not the true ground of recovery. Rather, the defendant is liable because he is morally responsible for infringing the plaintiff’s right against property damage.

In continental Europe, the non-relational normative principles operative beneath the relationally worded Romanist causes of action were distilled over many centuries by doctrinal scholars and moral philosophers, a process that culminated in the enactment of continental codes such as the Italian Civil Code and the German BGB. In England and America, with a few exceptions, such sweeping attempts at normative distillation and doctrinal reconfiguration have rarely occurred. Despite the formal doctrinal divergence between common law tort and civil law tort, this Article argues that the two categories’ moral substance is much the same. Like the general clauses of the civil law, the pigeonhole causes of action in the common law are centrally concerned with identifying and enforcing remedial moral liabilities that arise from a defendant’s moral responsibility for infringing a plaintiff’s rights against injury.

Call this the pigeonhole perspective. This Article argues for the pigeonhole perspective by examining a range of common law tort doctrines that utilize the language of relational wrongs to identify and enforce, often quite opaquely, remedial moral liabilities that are not grounded in relational wrongdoing.

Take the venerable doctrine of transferred intent in battery. Suppose that a defendant shoots at one person in an attempt to murder him. He thereby kills a different person who is unforeseeably on the scene. If this unforeseeable plaintiff’s estate brings suit, the defendant will be held liable to the plaintiff by means of the fiction that the defendant “intended” to harm her — although, in truth, the defendant did not intend harm to this plaintiff, or otherwise behave wrongfully toward her, any more than the train operator behaved wrongfully toward the unforeseeable Mrs. Palsgraf. For this reason, as Cardozo himself obliquely recognized in Palsgraf (but quickly brushed aside), transferred intent doctrine cannot be reconciled with the Palsgraf principle, which insists relational wrongdoing is the moral gravamen and sine qua non of tort liability.

Rather, transferred intent doctrine indirectly recognizes a moral phenomenon that tort law elsewhere recognizes more explicitly: While the scope of a defendant’s moral responsibility is presumptively limited to the reasonably foreseeable outcomes of his action, sufficient culpability can expand the scope of a defendant’s moral responsibility (and the remedial moral liability it grounds) to outcomes that lie beyond the boundaries of reasonable foreseeability. The black letter of battery doctrine is relational — it demands a wrongful intention toward the plaintiff. Transferred intent doctrine remodels this relational formal architecture through an “arrant, bare-faced fiction of the kind dear to the heart of the medieval pleader,” in order to enforce a compensatory moral liability that does not arise from relational moral wrongdoing.

Taking tort law’s formal surface at moral face value distorts the nature of tort law’s underlying moral commitments, thus inhibiting the philosophical project of delineating their contours and assessing whether they can withstand reflective scrutiny. If torts are not relational moral wrongs, and relational wrongdoing is not required to hold a defendant liable to a plaintiff in tort, then the moral logic of central aspects of tort doctrine (such as the proximate cause element in negligence) cannot be understood along the lines that defenders of the Palsgraf perspective and other philosophically oriented tort theorists often suppose. Judges, of course, are not philosophers; their task is to faithfully elaborate legal principles in light of the received doctrine and its underlying normative commitments, not to overturn those commitments in the name of philosophical reflection. But the loss of analytical clarity incurred by the Palsgraf perspective distorts the judicial function as well. For there are cases, we will see, in which tort law’s underlying moral commitments imply that a plaintiff should recover more often than the relational formal structure of its causes of action allows. That is precisely why the law has been compelled to resort to doctrinal fictions such as transferred intent. Once we see how the law’s formal structure fails to fully implement its substantive commitments, we can formulate and entertain new possibilities for giving those commitments doctrinal and institutional expression. So, for example, we might devise new remedial structures, such as new derivative causes of action, that “piggyback” on the relational structure of existing torts in order to achieve non-relational remedial ends. To identify the divergence between the law’s formal structure and its moral substance is to liberate our imagination about the space of doctrinal possibilities — to appreciate that we can reconfigure the existing forms in service of the moral principles they are supposed to serve.

The Article proceeds as follows. Part I outlines the Palsgraf perspective and compares it to the pigeonhole perspective. It explains how, on the latter, the common law of torts is implementing a set of moral principles about a defendant’s responsibility for unjust damage (infringing the plaintiff’s rights against injury) and the resulting remedial moral liability, which are more directly implemented by the tort liability rules we find in civil law. Part II compares the Palsgraf perspective and the pigeonhole perspective across a wide range of issues in tort doctrine and theory, in each case arguing that the pigeonhole perspective yields a more plausible account of the law’s operation and its underlying normative commitments. The Article’s Conclusion briefly explores some methodological implications regarding the role of legal fiction and doctrinal opacity in interpretive private law theory, and some deeper philosophical questions about whether tort law’s underlying moral commitments can ultimately survive reflective philosophical scrutiny.

Before proceeding, a word about methodology may be warranted. In line with a venerable tradition of common law torts scholarship, including early scholarship on Palsgraf, this Article makes extensive use of hypothetical cases. Such cases can help to elicit the structure of the common moral sensibilities and convictions embedded in tort doctrine, and thereby to illuminate the law’s underlying moral commitments. The history of Palsgraf itself supplies a vivid demonstration of this point. Two days after Palsgraf was argued in the Appellate Division of the New York Supreme Court, a distinguished group of torts scholars and judges, convened by the American Law Institute (ALI), met to discuss a hypothetical case strikingly similar to the facts of Palsgraf. The case was offered for discussion (independently, it seems) by Professor Francis Bohlen, the reporter for the First Restatement of Torts. Among the luminaries present at this discussion was Cardozo. In the meeting’s transcript, we find Cardozo’s skeletal articulation of the moral account of tort liability that he would later defend in Palsgraf. We also find the equally distinguished jurist Learned Hand offering a skeletal articulation of something like the competing moral account of tort liability that I defend here. It is plausible to surmise that this discussion influenced how Palsgraf was decided, and thereby had a considerable effect on the course of the common law. But in that ALI meeting, it was Hand, not Cardozo, who better understood the moral character and doctrinal structure of tort law. That, at least, is what this Article immodestly attempts to show.

I.  Two Moral Pictures of Tort Law

A.  The Palsgraf Perspective: Torts as Relational Wrongs

Ordinary moral thought recognizes moral duties that are relational as well as duties that are impersonal. A relational moral duty is owed to another person, such that she is wronged by its breach. By contrast, an impersonal moral duty is owed to no one, or (perhaps) owed to the world at large. Suppose, for example, that A has reason to believe that detonating a piece of dynamite would destroy an ancient and beautiful cave that no one will ever see. Absent some good reason to do so, it would plausibly be morally wrong for A to detonate the dynamite, but doing so would not wrong anyone else — it would simply be wrong, full stop. If, by contrast, A has reason to believe that detonating the dynamite will kill another person, B, or destroy her property, then A’s duty not to detonate the dynamite is owed to B, such that he would wrong her by breaching this duty.

The Palsgraf perspective’s organizing interpretive commitment is that relational duties and wrongs — rather than impersonal duties and wrongs, or incentives for promoting societal interests or impersonal values — lie at the heart of tort law. Today, its scholarly defenders fall into two major camps. Pragmatic constructivists such as Dean John Goldberg and Professor Benjamin Zipursky believe that each tort — negligence, battery, defamation, and so on — defines a sort of relational legal wrong, which in turn encodes (though not precisely) a certain sort of relational moral wrong. Pragmatic constructivists tend to maintain a sort of studious philosophical quietism about the nature and content of these relational moral wrongs. Their focus is instead on elucidating the norms and practices of the courts and the legal profession, through which tort doctrine draws upon and roughly encodes relational moral wrongs recognized in ordinary moral life.

By contrast, philosophical formalists such as Professors Arthur Ripstein and Ernest Weinrib believe that tort law’s gallery of relational legal wrongs together encodes one overarching sort of relational moral wrong, the wrong of acting as another person’s master, by violating that person’s equal freedom to set their own ends — either by using that person’s means (that is, his body or property) without his consent, or by imposing on that person an abnormal risk of damage to his means. Thus, philosophical formalists are inclined to monism where pragmatic constructivists are inclined to pluralism. Unsurprisingly, philosophical formalists’ interpretive enterprise is far from philosophically quietist. Their aim is to elucidate the unified relational moral wrong of acting as another person’s master by reference to Kantian and Aristotelian philosophy, and thereby to defend it not only as an understanding of the law but also as a first-order normative view of how tort law ought to operate.

Thus pragmatic constructivists and philosophical formalists diverge in various important respects. Despite these differences, both camps agree that a tort just is a relational legal wrong that reflects a relational moral wrong — and that providing redress for relational wrongs (or empowering the victim of such a wrong to obtain redress at her option) is tort law’s central normative aspiration.

Consider how this understanding of tort law deals with the facts in Palsgraf itself. The freakish outcome that Mrs. Palsgraf suffered, as a result of the train attendant’s negligent treatment of the youth attempting to board his train, was not a foreseeable result of that negligent action. Thus many scholars and judges have supposed that Mrs. Palsgraf’s claim fails simply for want of proximate cause (which, to a rough approximation, requires that the plaintiff’s injury be the foreseeable materialization of the risk that made the defendant’s action wrongful). As defenders of the Palsgraf perspective note, however, treating Palsgraf simply as a proximate cause case does not comport with Cardozo’s reasoning in the opinion itself.

For Cardozo, Mrs. Palsgraf’s claim failed for a more fundamental reason than want of proximate cause: The train attendant’s negligence toward the boarding youth was not negligent toward Mrs. Palsgraf, since she was not a foreseeable victim. “Negligence . . . is . . . a term of relation”: To commit the tort of negligence is to (foreseeably) injure someone by breaching a duty to take reasonable care as to her. More generally, unless a defendant breaches a duty owed to the plaintiff — that is, commits a wrong against her — the Palsgraf perspective maintains that the plaintiff cannot recover for any resulting injuries, because redressing relational wrongs is tort law’s fundamental purpose.

Of course, as anyone who has taken introductory torts is aware, the defendant’s breach of a duty of care to the plaintiff is not enough to render him liable to her: Duty and breach are only two of the elements that together make up the tort of negligence, along with injury, cause in fact, and proximate cause. According to the Palsgraf perspective, all of these elements together define a complex relational duty: roughly, a duty not to foreseeably cause injury to another person by breaching one’s duty to take reasonable care against causing her such injury. So the tort of negligence — perhaps uniquely among the various torts — is associated with two relational legal duties: a duty of non-injuriousness (take reasonable care toward another person if your action might foreseeably injure her) and a qualified duty of non-injury (do not actually injure another person by breaching your duty to take reasonable care toward her). To violate this second duty — a qualified duty of non-injury — is to commit the tort of negligence. Or so the Palsgraf perspective says.

On its face, there is something strange about this way of understanding the tort of negligence — at least on the assumption that a tort such as negligence is a sort of legal wrong that tracks a sort of moral wrong. The negligence tort, so understood, does not appear to track any recognizable sort of moral wrong at all; it is rare to see anyone other than defenders of the Palsgraf perspective refer to such a thing as a qualified duty (or wrong) of non-injury. By itself this observation is hardly damning. Philosophical reflection can sometimes reveal that our ordinary moral practices are implicitly committed to the existence of some complex moral phenomenon that no concept in our ordinary moral vocabulary readily picks out. Nevertheless, the prima facie oddity of the idea of a qualified moral duty (or wrong) of non-injury invites the question of whether a more plausible moral understanding of what a tort is — and what tort law is doing — might be found.

B.  The Pigeonhole Perspective: Torts as Remedial Moral Liability Rules

As already suggested, the tort clauses of the civil codes point the way to such an account: A tort is a legal liability rule, which specifies a set of conditions under which a defendant has an enforceable legal obligation to compensate a plaintiff. Consider again the central tort clause of the Italian Codice Civile, CC 2043: “Any intentional or negligent act that causes [unjust damage] to another person obliges the person who has committed the act to pay damages.” That this clause lays down a compensatory legal liability rule is evident on its face. And nothing in such a clause appears to make any reference to the ideas that the Palsgraf perspective locates at the heart of tort law. Whether a defendant has acted wrongfully toward a plaintiff, mistreated her, affronted her personality, disrespected her, and so on — whether he has committed a wrong to her, not just a wrong to other people or society at large — is, on the face of these clauses, immaterial.

To be sure, Italian and German law would preclude recovery in a case such as Palsgraf no less than the common law would. But that is simply because civil law judges have read into such clauses something like a remoteness or proximate cause restriction, under which outcomes that are overly fortuitous or unrelated to the wrongful character of the defendant’s action are presumptively beyond the scope of his remedial liability. The civil law thus joins the common law in holding that, at least in normal circumstances, “a man must be considered to be responsible for the probable consequences of his act.” Put roughly, the injurious outcome or consequence must be foreseeable, but the foreseeability of the plaintiff is immaterial.

The plaintiff does figure in civil law analysis, but in a much more restricted fashion: The foreseeable consequence that the defendant causes, in a morally responsible manner, must infringe some right held by the plaintiff. Merely causing a loss to the plaintiff — even wrongfully and foreseeably — does not generate any compensatory liability toward her. Suppose, for example, that A negligently and foreseeably kills a rare and endangered sort of wild animal, thereby preempting B’s attempt to capture and exhibit this animal for profit in her zoo. If B had owned the animal, she would of course be entitled to compensation from A, for his killing of the animal would have infringed her property right. Because B has no such right, she is not entitled to compensation from A for the economic loss she suffers as a result of his action. Or suppose that A’s chattel falls onto the land of B, who innocently believes that the chattel belongs to her and plans to use it for her own gain. A is able to enter B’s land in a timely manner to regain the chattel, but doing so will (as a side effect) unavoidably destroy the property of another innocent party, C. If A recaptures the chattel nevertheless, A is liable to compensate C, but A is not liable to compensate B. That is for the simple reason that A’s action, however wrongful and culpable, did not infringe any property right (or right of another kind) held by B. Even if he deprives her of a gain she would have enjoyed without his wrongful action, she had no right to that gain, and thus, he is not liable to compensate her for it.

Civil law jurisdictions, like common law jurisdictions, generally recognize that wrongfully causing loss (without infringing any sort of right) does not ground compensatory liability. But the common law, unlike the civil law, insists (at least on its face) that a defendant must not only infringe the plaintiff’s right against injury; he must also breach a duty owed to her, that is, treat her wrongfully by violating a norm of conduct regarding how she should be treated. Thus, on its face, the common law denies recovery to an unforeseeable plaintiff whose property is foreseeably destroyed by the defendant’s negligence or a foreseeable plaintiff who is not treated carelessly by the negligent defendant. A civil law clause such as BGB section 823(1), by contrast, elegantly explains the propriety of recovery in such a case, and in many other cases to be discussed below.

This Article’s claim is that the common law’s formal veneer — its surface language of relational duties and wrongs — obscures its substantive moral commitments. In both common law and civil law, the core of tort liability is about tracking remedial moral liability — and remedial moral liability for inflicting injury does not ultimately rest on relational wrongdoing or mistreatment between the defendant and the plaintiff. Rather, it rests on the defendant’s moral responsibility for infringing a right against injury held by the plaintiff.

The idea of moral responsibility at work here is a familiar one — even if its nature and grounds are somewhat elusive on philosophical reflection. Ordinary moral thought recognizes that causing (or “doing”) harm can be highly morally significant. A person whose action is causally responsible for some harmful outcome is connected to that outcome in a way that other people, mere bystanders, are not — and this special connection may help to explain why certain responses to that outcome (regret, guilt, apology, compensation, and so on) are fitting or even obligatory for the person in question, while no such response would be fitting or obligatory for a mere bystander. But mere causal responsibility alone (sometimes called outcome responsibility) is inadequate to render onerous affective responses such as guilt fitting, or onerous practical responses such as compensation obligatory — for that, something more is required.

This is a general phenomenon that has nothing essentially to do with rights against injury or relational wrongs. To illustrate, consider a variation on Thomson’s example: Suppose that A innocently flips his kitchen switch and thereby unforeseeably causes the destruction of an (unowned) beautiful nature preserve, or some other environmental disaster. Although A is causally responsible for this outcome, there is an intuitively clear sense in which he is not morally responsible for it: The outcome is not connected, in the right way, to the quality of his responsible moral deliberation, judgment, and agency. And so, although mild forms of regret might be warranted, and limited forms of repair might be called for, it would be strange for A to be consumed by guilt or to regard himself as under a strong and onerous obligation to repair the harm he has caused.

Matters seem quite different if A has foreseeably and culpably risked such an outcome. In that case, moral attitudes such as guilt might be strongly warranted. And actions such as attempting to repair the harmful outcome in question, or (insofar as repair is impossible) make amends by preventing other such harmful outcomes, might well be morally obligatory, even at large personal cost. The same moral asymmetries between morally responsible and non-responsible causation would seem to exist when it is the infringement of another person’s right, rather than the destruction of some impersonally valuable object, that is at stake.

There may be various forms of moral responsibility — relationships between the quality of an agent’s responsible moral agency (on the one hand) and the outcome that her action causes (on the other) — that can ground remedial moral liability and liability in tort. But plausibly — as suggested by the central role of clauses like CC 2043 and BGB section 823(1) in the civil law, and the central role of the tort of negligence in the common law — the legally central form of liability-grounding moral responsibility is culpable wrongdoing. More specifically, I suggest, something like the following moral principle is implemented by vast swathes of tort law in both common law and civil law systems:

Liability for Culpable and Foreseeable Damage (CFD): An agent who culpably and foreseeably risks causing damage, that is infringing a moral right against being caused injury (to body, property, liberty, reputation, and so forth), is liable to compensate the person whose right he thereby infringes.

I will argue that this moral principle (as duly refined) furnishes the normative substance of torts such as negligence, battery, and defamation. To be sure, many torts (or central elements of them, such as the duty of care in negligence) have a facially relational structure. That this formal structure obscures the common law’s underlying substantive concerns is revealed by the common law’s handling of situations where relational wrongdoing and moral responsibility for rights infringement diverge: In such cases, the common law disregards the relational formal structure of its causes of action, deforms this structure through legal fiction, or utilizes nonrelational doctrinal appendages such as derivative causes of action.

In short, common law and civil law tort are both, in their heartland, concerned with moral responsibility for causing unjust damage — that is, infringing rights against injury — not with relational wrongdoing. The rest of this Article will attempt to provide a great deal of casuistical evidence for this claim. Before that, however, it is worth mentioning a more abstract sort of consideration: It would be surprising if the moral substance of the common law of torts turned out to differ starkly from the moral substance of the tort liability clauses in the civil codes. Prima facie, it seems more plausible that the common moral sensibilities embedded in the civil codes are largely continuous with the common moral sensibilities that have been embedded in the common law.

After all, the common law was at its inception largely derived from continental Europe — it was feudal law originally “developed on the Continent, from the days of Charlemagne onwards” and “imported into England by the Norman conquerors.” After that, the common law was substantially influenced by academic treatises on the civil law and regular interaction with ecclesiastical canon law, itself largely derived from the classical Roman law from which the modern civil law largely originates.

That no stark normative divide exists between the private law (including tort) in common and civil law systems was once, it appears, the common understanding of common lawyers. As one British judge put it, “the [c]ivil law is not of itself authority in an English [c]ourt, [but] it affords great assistance in investigating the principles on which the law is grounded.” In fact, some historians of the common law have argued that “the basic structure of the tort of negligence . . . was directly or indirectly derived from Roman law,” partly by way of continental moral philosophers, such as Grotius and Pufendorf, who were influential in shaping the eighteenth-century English legal consciousness as well as the civil law codes. Whether or not this strong claim is sound, there is ample evidence of robust influence and doctrinal borrowing. Some of the common law’s leading tort judgments, such as Lord Atkin’s famous speech in Donoghue v. Stevenson, conspicuously draw upon civil law concepts in order to frame or support their analyses of common law doctrine. To my knowledge, in none of these contexts did common lawyers ever suggest that the structural divergence between common law tort and civil law tort bespoke some significant normative discontinuity between them.

Reflection on the history of the civil law itself similarly provides some evidence for normative continuity. Modern civil law largely derives from classical Roman law. Like the common law today, classical Roman law contained “a somewhat haphazard assemblage of individual delicts,” which took their names from commonly recognized forms of relational moral wrongdoing: theft, violent robbery, loss inflicted through a wrong (the famous lex Aquilia, the most general form of action, with roughly the scope of the common law tort of negligence), offensive personal insult, and so on.

European legal systems largely inherited this doctrinal structure. Before the enactment of the BGB, for example, German tort law largely operated with such Romanist causes of action. Like the common law of torts, therefore, pre-BGB German tort law “refused to grant any claim outside the specific bases which were explicitly recognised.” Until the drafting of the BGB, “scholars never really considered unifying the list of torts into one general rule,” nor did the legal system. As with many other civil law tort codes, the intellectual pre-history and legal drafting of the BGB involved a process of abstracting and generalizing from the particular doctrinal details of the received Romanist causes of action, in order to arrive at general legal principles about fault, unjustifiability, rights, and remedial liability such as BGB section 823(1).

It is conceivable that remodeling the conceptual structure of German law in this way effected a profound transformation of its underlying normative substance. But it seems more plausible that this process involved the abductive extraction and refinement of normative principles immanent within the legal corpus, which ultimately received more transparent articulation in the form of the general legal principles laid down by the BGB. In my view, common law causes of action such as negligence and defamation ought to be understood in the same way. The pigeonholes of pre-BGB German tort law were coarse remedial devices whose underlying moral principles were never perspicuously disclosed by their formal structure. That these causes of action were named after various forms of relational wrongdoing did not indicate that relational wrongdoing was their basic moral concern. So, too, the common law torts are remedial legal pigeonholes — legal formulas for obtaining judicial relief — that, notwithstanding their surface form, are largely concerned with tracking a defendant’s moral responsibility for infringing a plaintiff’s rights and with enforcing the resulting moral liabilities to provide compensation or other forms of remedial relief.

Were the common law to be comprehensively remodeled in the form of a statutory code, this basic moral concern might receive cleaner expression in its formal structure. Most of the time, however, the common law’s pigeonholes do well enough in supplying recovery by the lights of their animating moral principles that any felt pressure to undertake the laborious process of abstract system-building, deeper normative explanation, and comprehensive doctrinal reform is defused.

These implicit moral principles can range across the boundaries of different torts. And it is because the common law’s relational formal structure — constituted by its motley assemblage of pigeonhole torts — does not clearly articulate these underlying moral principles that the common law must sometimes deform its formal structure through devices like legal fictions, such as the one we will now examine. To take the common law’s formal architecture at moral face value, as the Palsgraf perspective urges, is thus to misconstrue the nature of the commonsense moral principles that it houses.

II.  Comparing the Pictures

A.  The Fiction of Transferred Intent

Let us begin with an issue that Cardozo observed in Palsgraf but treated only in clipped and conclusory fashion: His claims about the nature of tort and tort liability are incompatible with the venerable common law doctrine of transferred intent.

The paradigmatic transferred intent case is from Professor William Prosser’s classic article on the subject:

Defendant shoots at A, intending to wound or kill him. His aim is bad, and he misses A. The bullet passes through a screen of bushes and hits B. Defendant is unaware of B’s presence, and he could not reasonably have been expected to be aware of it. There is no intent to hit B and no negligence with respect to B. The injury to B is an accident, pure and simple. Nevertheless, [D]efendant is liable to B for battery, an intentional tort.

If we wish to explain this sort of conviction, and the cases in which it finds expression, two options are available to us. The first is to say that, although B is unforeseeable to the defendant, the defendant nevertheless breaches a duty owed to B by shooting at A. To embrace this option is, of course, to abandon the idea that tort duties cannot be owed to unforeseeable plaintiffs. The second option is to say that B is entitled to redress against the defendant (and/or the defendant owes him an enforceable duty of compensation) despite the fact that the defendant has breached no duty owed to B. To embrace this option is to abandon the basic idea that liability must be founded on the breach of some duty owed to the plaintiff. Either way, transferred intent doctrine is incompatible with a fundamental assumption about tort law on which Cardozo’s reasoning in Palsgraf relies.

An obvious response to this tension is to dismiss the doctrine of transferred intent as the archaic relic of a bygone legal age. That response would have been consistent with Cardozo’s suggestion that allowing the victim of a harmful action to sue for the breach of a duty owed to another would be “to ignore the fundamental difference between tort and crime.” Transferred intent doctrine has its roots in the early writ of trespass, which sat somewhat indeterminately at the border of crime and tort (as we understand those legal categories today). The doctrine continues to play an important role in the criminal law, which few are prepared to abandon. But it is possible to maintain that the doctrine’s persistence in tort law is an archaism that ought to be abandoned.

That is not the strategy that Cardozo adopted, however, and it is not hard to see why. The intuitive moral conviction that, in cases such as Prosser’s hypothetical, the defendant owes his victim an enforceable duty to compensate for the damage he has inflicted on her — a remedial duty of precisely the kind that the tort system ought, in principle, to enforce — is a very powerful one. It finds expression in a longstanding body of case law on transferred intent; it is expressly affirmed by both the Second and Third Restatements; and it has deeply rooted doctrinal analogues in other legal systems, both common law and civil. And the intuition does not seem like it can be explained away as the creature of retributive instincts properly located in the criminal law (if anywhere). The law would not satisfy the intuitive requirements of justice to the victim simply by punishing the defendant in such a case, or by extracting from him a financial penalty to be deposited into a general fund for the victims of crime. It seems clear that the defendant is liable to his victim, unforeseeable though she may be — that he must compensate her, in preference to the victims of other wrongdoers.

It is understandable, then, that Cardozo did not reject transferred intent out of hand or admit the incompatibility of transferred intent’s normative logic with his opinion’s grand claims about the nature of tort liability. Doing so would have severely compromised the interpretive and moral plausibility of those claims and weakened the persuasive power of his judgment. Instead, he acknowledged the doctrine and intimated that its significance lies outside the domain of his discussion, in two passages. In the first passage, the reference to transferred intent is oblique:

Negligence is not a tort unless it results in the commission of a wrong, and the commission of a wrong imports the violation of a right, in this case, we are told, the right to be protected against interference with one’s bodily security. But bodily security is protected, not against all forms of interference or aggression, but only against some. One who seeks redress at law does not make out a cause of action by showing without more that there has been damage to his person. If the harm was not willful, he must show that the act as to him had possibilities of danger so many and apparent as to entitle him to be protected against the doing of it though the harm was unintended.

The implicit suggestion is that the logic of the italicized exception does not vitiate the logic of Cardozo’s general claim. But that suggestion plainly calls out for explanation and defense. If willfully harmful actions may render a tortfeasor liable for harm to an unforeseeable victim, one wishes to know, why not negligently harmful actions? If a tortfeasor cannot owe a duty against careless injury to an unforeseeable victim, how can a tortfeasor owe a duty against intentional injury to an unforeseeable victim? Absent some plausible explanation of this asymmetry, consistency would appear to require dispensing with the doctrine of transferred intent or renouncing the relationship between foreseeability and negligence liability that Cardozo posited. And nothing in the passage attempts to supply the needed explanation.

There is, however, a second passage in which Cardozo discussed transferred intent. In that passage, the germ of an explanation may be suggested:

Some acts, such as shooting[,] are so imminently dangerous to any one who may come within reach of the missile[,] however unexpectedly, as to impose a duty of prevision not far from that of an insurer. Even to-day, and much oftener in earlier stages of the law, one acts sometimes at one’s peril. Under this head, it may be, fall certain cases of what is known as transferred intent, an act willfully dangerous to A resulting by misadventure in injury to B. These cases aside, wrong is defined in terms of the natural or probable, at least when unintentional.

It is not entirely clear how to understand Cardozo’s suggestion, but there is a plausible interpretation that comports with views that some defenders of the Palsgraf perspective have offered on his behalf. In many circumstances, a defendant who shoots a gun at person A should be aware that doing so could result in harm to some other person B, even if no other person appears to be on the scene. For there is a nontrivial chance that another person is in fact present, invisibly to the defendant’s eye, and shooting a gun is so dangerous an activity that it poses a nontrivial risk of injuring any such person. Thus, B can recover for her injuries, in such a case, on ordinary principles of negligence: By shooting his gun at A, the defendant foreseeably risks causing harm to persons such as B, such that the agent breaches a duty of care to B by shooting the gun without adequate justification. Characterizing the situation in this way allows B to recover for injuries sustained through an act wrongful as to her, without invoking any conceit about transferred intent.

This maneuver does handle many of the cases that are standardly described as exhibiting the transferred intent fiction, but it distorts the moral gist of the doctrine. A simple modification to Prosser’s case will illustrate. Suppose a defendant is planning to murder A and that he makes laborious preparation to dispose of her body after the crime. The defendant ardently desires not to kill any bystanders, however, for he has not made any such preparations with respect to them, and inadvertently killing a bystander would threaten to expose his wrongdoing. Hence the defendant diligently examines the plot of land where he plans to kill A, so as to make sure that no one else is on the scene. In an exceptionally improbable turn of events, it turns out that some other person, B, has been buried alive, by a different villain, beneath that very plot of land. The police are on their way to dig out B (and thereby save him from impending asphyxiation). But before the police can arrive, the defendant shoots at A; the bullet misses A but goes into the land beneath him and pierces B’s heart, killing her. Here the risk that shooting at A might cause harm to anyone else, such as B, is so exceptionally small as to foreclose the conclusion that B was a foreseeable victim; the defendant cannot be plausibly accused of negligence toward anyone other than A. All the same, B’s estate appears to have a forceful moral claim to redress.

It is through the doctrine of transferred intent that the law would enforce these intuitive moral liabilities. The attempt to substitute negligence principles in the place of transferred intent doctrine would preclude the law from doing so. In practice, such clean test cases — in which the transferred intent plaintiff is clearly unforeseeable, rather than only arguably so — are unlikely to arise, and certainly they are unlikely to be litigated. Still, predicating the victim’s ability to recover in such cases on negligence principles distorts the character of the ordinary moral convictions that underlie the law’s provision for recovery. Intuitively, these convictions have little to do with any thought that the victims are in fact foreseeable to the wrongdoers who injure them. By itself, the doctrine of transferred intent does not elucidate or vindicate these underlying moral instincts, but it permits them greater expression than the Procrustean reliance on negligence principles would allow.

For that reason, excising transferred intent doctrine would compel the law to draw distinctions between plaintiffs that seem morally absurd. Suppose that a defendant attempts to murder his romantic rival A by shooting at a figure that looks like A, but the figure is in fact a mannequin; the bullet goes through the ground where it pierces the heart of another person who is unforeseeably buried alive beneath it (and would have been dug out in due course, had she not been shot). Even without a doctrine of transferred intent, the law would appear to hold the defendant liable in battery to the victim if the victim happens to be A, for in that case the defendant has killed A by acting on an intention to kill him (strange though the causal route may be). Without a doctrine of transferred intent, the law will hold the defendant liable to the victim only if the victim happens to be A, rather than some different person B. That seems absurd.

To credit such a distinction yields even greater absurdity in other cases. Imagine a version of Palsgraf in which the defendant train attendant is a bigot who intentionally jostles the boarding youth in order to murder her, because the attendant wishes to murder an Indian person with plausible deniability, believes that the youth is a famous stage actor, Sonia, and believes that Sonia is Indian. As it turns out, the train attendant is multiply mistaken: The youth is not Sonia, and Sonia is Italian, not Indian. The resulting explosion causes some scales to drop at the other end of the station, injuring three unforeseeable victims — Elise (French), Amna (Indian), and Sonia (who just happens to be at the station, waiting for a train). Without a doctrine of transferred intent, the law will of course deny Elise recovery, which seems perverse enough. As to Amna and Sonia, the law will need to determine how to characterize the content of the train attendant’s murderous intention (an intention to kill Sonia, or an intention to kill an Indian, or . . . ?). The exercise seems ludicrous, and there does not seem to be any morally or metaphysically principled way to carry it out.

That such an exercise is implied by the logic of the Palsgraf perspective should “enkindle doubt whether a flaw may not exist” in the logic that entails it. If the tort of battery is the legal recognition of a certain kind of relational moral wrong — the wrong of harming another person by acting on an intention to harm (or harmfully touch) her — and if providing redress for legally recognized relational moral wrongdoing is tort law’s basic normative aspiration, then tort law is, in such a case, committed to engaging in such an exercise as a matter of basic moral principle. That remodeling tort doctrine in line with the Palsgraf principle would in relevant cases commit the doctrine to conducting such an exercise is evidence that the Palsgraf principle distorts the moral sensibilities encoded by the law.

At this point, Cardozo’s followers might elect a more drastic maneuver. The idea here would be that, although in transferred intent cases the plaintiff is unforeseeable to the defendant, the defendant nevertheless breaches a legal duty owed to him: a legal duty not to injure him (unforeseeable though he may be) by acting on an intention to injure someone else. Once the requirement of victim foreseeability is jettisoned as to duties in battery, however, it becomes obscure why it should be retained as to duties in negligence. If (as the Palsgraf perspective’s defenders say) “the idea of owing [a] duty to someone who is unforeseeable” is incoherent — because a duty must be able to “guide [the defendant’s] conduct” (and a person deliberating about her conduct cannot take account of an unforeseeable victim) or else because the duty’s violation must express an “[a]ffront to [the plaintiff’s] personality” (and an unforeseeable victim’s personality cannot be affronted) — then such a duty is no less incoherent in battery than in negligence. If, by contrast, it is coherent and plausible to recognize a duty to unforeseeable victims in battery, there should be no obstacle to recognizing a duty to unforeseeable victims in negligence as well. But to recognize such a duty in negligence would, of course, vitiate the reasoning in Palsgraf entirely.

What is more, the envisaged maneuver would undercut the Palsgraf perspective’s basic supposition that the relational legal duties and wrongs we find in tort law track the relational moral duties and wrongs of ordinary moral life. The species of relational duty under consideration — a duty not to injure one person by acting on an intention to harm someone else entirely — seems like an unnatural creature, pieced together by dissecting familiar sorts of moral duties and conjoining them into a beast that does not roam the landscape of ordinary moral thought. In short, such a desperate maneuver would preserve the outermost legal form of the Palsgraf principle only by emptying it of any plausible or recognizable normative substance.

The lesson, I believe, is that the Palsgraf principle is unsound: One need not act wrongfully toward another — need not mistreat her, as opposed to other people — to incur a liability in morality or law to compensate her for the injuries one has inflicted on her. If that lesson is correct, however, what should we make of the black-letter proposition, which is embraced by most common law jurisdictions, that the duty of care in negligence is owed only to foreseeable victims, such that only foreseeable victims can in ordinary cases recover? In my view, this black-letter rule does not reflect any principled commitment to the significance of relational wrongdoing in grounding negligence liability (a commitment that would sit in stark tension with the underlying moral commitments of transferred-intent doctrine). There is, rather, a more complex moral principle operating beneath the surface of both doctrinal areas — a principle that explains why unforeseeable transferred intent victims can recover, and why unforeseeable victims of mere negligence (as in Palsgraf) typically cannot.

This principle is simply the principle that we have already located at the core of civilian tort clauses such as CC 2043, more accurately formulated:

Liability for Culpable and Foreseeable Damage* (CFD*): An agent who culpably and foreseeably risks causing damage, that is, infringing a right against injury, is liable to compensate the person whose right he thereby infringes, where the degree of foreseeability required depends on the degree of the agent’s culpability.

Such a principle finds considerable support in the modern civil law and the classical Roman law from which it derives. The same principle is recognized in the common law as well — occasionally with great clarity, as in a leading British judgment that forthrightly observes that “[t]he more culpable the defendant the wider the area of loss for which he can fairly be held responsible.” Most of the time, however, the common law recognizes this principle more coarsely and obliquely, through a crude categorical distinction between intentional torts and the tort of negligence.

Thus, it is a familiar observation in the case law, doctrinal commentary, and restatements, that intentional torts such as battery will subject a tortious wrongdoer to a far wider scope of liability than the tort of negligence: “The rule of legal (proximate) cause . . . for intentional torts sweeps very broadly, almost to the full reach of factual causation.” The law seems to draw the boundaries of proximate cause more widely in battery than in negligence because the law takes intention to harm as a rough but serviceable proxy for agential culpability, which (as the law recognizes) expands the range of risks for which a wrongdoer bears remedial moral responsibility. As an early twentieth-century American case put it: “For an intended injury the law is astute to discover even very remote causation . . . because it has been felt to be just and reasonable that liability should extend to results further removed when certain elements of fault [are] present.”

The same general principle operates within the boundaries of the negligence tort alone. Consider a defendant who has imposed an unreasonable risk upon a foreseeable plaintiff, such that he has plainly breached a duty of care owed to her. In determining whether the resulting injury was “foreseeable,” such that the proximate cause element of the tort of negligence is satisfied, the law will treat a tortfeasor more harshly the more culpable he is. Thus, a grossly reckless tortfeasor will be held liable for the materialization of a wider range of risks than a barely negligent one. Put another way, the proximate cause inquiry in the tort of negligence is culpability-sensitive.

My proposal is that all of these common law phenomena — the fiction of transferred intent, the broader scope of liability for intentional wrongdoing as opposed to negligent wrongdoing, and the culpability-sensitivity of proximate cause within the tort of negligence — are simply particular doctrinal expressions of the same underlying moral principle, namely, something like CFD*.

Consider, in more detail, how this proposal makes sense of the kinds of transferred-intent cases discussed above. It explains why the defendant who maliciously shoots at A ought to be liable to compensate another person, B, whom he thereby kills, even if B’s presence on the scene is utterly improbable. Similarly it explains why, if the defendant attempts to kill A by shooting at a mannequin that mistakenly appears to be A, and the defendant thereby kills another person unforeseeably on the scene, the defendant is liable to compensate that person regardless of whether she happens to be A or some other person B.

There are certain respects in which the proposal does not track all of the fine-grained structural contours of the transferred intent fiction — but that is a virtue of the proposal, not a vice. Suppose that the defendant negligently pets A’s dog, which he mistakenly believes to be his own. By doing so, he triggers a hidden motion detector that detonates a bomb, blowing up a cat that belongs to another person, B. By its terms the doctrine of transferred intent seems to apply to the tort of trespass to chattels, so it yields the result that the defendant is liable in trespass to B for the injury he has unforeseeably inflicted on her cat. But this seems entirely at odds with common moral understanding; the defendant seems no more morally liable than the train attendant in Palsgraf.

In this sort of case, therefore, the scope of a legal rule plainly diverges from the common moral sensibilities that underlie and rationalize it. By its terms, the transferred intent fiction prescribes a recovery here that “does not seem consonant with current ideas of justice or morality.” Common sense morality does not suppose that “for an act of negligence, however slight or venial, which results in some trivial foreseeable damage the actor should be liable for all consequences however unforeseeable and however grave.” The scarcity of transferred intent cases, in life and litigation, prevents this divergence between legal form and moral substance from working much injustice in practice.

The idea that the operative instincts behind transferred intent doctrine sound in agential culpability is far from a novel one. This idea is suggested by Prosser and Keeton on the Law of Torts and equally authoritative English authority, as well the Third Restatement, which accordingly prescribes that the doctrine should be interpreted and applied in light of the tortfeasor’s degree of culpability. There is no way of interpreting the doctrine, however, that will entirely cure its mismatch with the common moral convictions that underlie and justify it. For there are situations, such as the hypothetical just discussed, in which the terms of the transferred intent fiction plainly call for recovery, but this underlying justification supplies no basis for it. Conversely there are cases in which this underlying justification is fully implicated but the terms of the transferred intent fiction are silent.

Imagine, for example, a variation of Prosser’s case in which the defendant is grossly reckless with respect to his possible commission of some deadly bodily intrusion against A: Without any good reason, the defendant releases a cat in A’s direction, knowing that there is a one-in-two chance that the cat will come into contact with A and kill her, since A is deathly allergic to cats. As it happens, the cat does not come into contact with A but does come into contact with B, a different person who is unforeseeably on the scene and who is also deathly allergic to cats. Since the defendant did not possess any “substantial certainty” as to the harm he might cause A, he cannot be deemed even constructively to have intended harm to B. Thus, the doctrine of transferred intent will be unable to secure any recovery to B. And since, as Palsgraf lays down, there is no doctrine of transferred negligence — not even in the case of gross negligence or gross recklessness — B will be left without any doctrinal peg on which to hang a claim for recovery. Yet the same common moral instincts that counsel in favor of permitting B recovery in Prosser’s original case seem to speak powerfully in favor of affording B recovery here. If such a case should actually arise, it would not be at all surprising to see the court defy the formal strictures of doctrine in order to send it to the jury, or to see the jury grant the plaintiff recovery.

Such cases suggest that the transferred intent fiction is most plausibly understood as an extremely coarse fictional device by which the law roughly tracks one dimension of the more general moral phenomenon articulated by CFD*: The scope of an actor’s remedial moral responsibility for the materialization of the risks of injury posed by his action is sensitive to his degree of culpability for imposing the risks in question. Is it implausible to suppose that the law should implement its underlying moral principles with so coarse a device? No; we will see several more such coarse devices below. Before that, however, it is worth dwelling a bit more on one of the doctrinal phenomena we have just encountered, and others in its vicinity.

B.  Proximate Cause, Punitive Damages, and Holistic Culpability

As just discussed, the proximate cause determination in negligence is sensitive to a defendant’s level of culpability in performing the action that injured the plaintiff. Similarly, and even more obviously, the defendant’s level of culpability determines whether the plaintiff whom he has injured can exact punitive damages from him. In both ways, the law takes cognizance of the defendant’s holistic culpability, rather than his culpability as toward the plaintiff. It is difficult to square these doctrinal phenomena, therefore, with the Palsgraf perspective’s insistence that a defendant’s liability to a plaintiff must turn entirely on his mistreatment of her.

Let us begin with punitive damages. Consider again the classic transferred intent hypothetical in which the defendant attempts to shoot and kill one person, A, and ends up, improbably, shooting and killing some different person, B, instead. Suppose a variation of the case in which B is a foreseeable, albeit improbable, victim. In such a case, it is indeed possible to explain the defendant’s liability to B (or her estate) in conformity with the Palsgraf principle, for he has acted wrongfully toward her as well as toward A.

Now let us imagine, however, that B’s estate presses not only for compensatory damages but for an award of punitive damages as well. It would be entirely unremarkable, indeed prosaic, for a jury to grant such an award, and no judge would dream of reversing it. Notice, however, that the propriety of such a punitive damages award cannot be grounded in the magnitude of the defendant’s mistreatment of B, for the defendant was not very negligent as toward B at all. The defendant’s vulnerability to punitive damages, in such a case, depends upon his culpability tout court, not his culpability vis-à-vis the plaintiff. There is no strict inconsistency in supposing that, while the defendant’s liability to compensate B depends on whether he has mistreated her, his liability to pay punitive damages to her may depend on how gravely he has mistreated others. But the tension between these propositions is glaring, and it is unlikely that both of these propositions will survive any plausible attempt to work out the normative basis of liability to pay punitive damages.

Take proximate cause next. Suppose that in attempting to shoot and kill A, the defendant triggers a nest of hornets, unforeseeably hidden on the scene, that lethally sting both A and B. The defendant will be liable to A in battery, and liable to B in battery on the fictional basis that the defendant intended B’s death alongside A’s. But suppose that, in line with the maneuver described above, we dispense with that fiction and instead attempt to ground the defendant’s liability to B on the basis of ordinary negligence principles. Can we do so? Only if we acknowledge that the defendant’s enormous culpability as to A can expand the range of harms to B for which he is liable, beyond the boundaries of reasonable foreseeability which would normally delimit this range. If (for example) the defendant were merely careless in playing with a gun, notwithstanding the risk that by doing so he might shoot and kill A and B, it is unlikely that any court would hold him liable for bringing about B’s death by shooting and triggering an unforeseeable nest of hornets. The scope of the defendant’s liability for causing unforeseeable harm to B in the preceding case, just like his liability for causing unforeseeable harm to A, reflects his holistic culpability for his action. That is true whether the defendant’s liability to B is doctrinally secured by the fiction of transferred intent or instead by the proposition that the defendant’s heightened culpability can expand the ambit of harms for which he is liable in negligence, beyond those harms that he can reasonably foresee.

That proximate cause determinations are sensitive to a defendant’s holistic culpability is readily intelligible if the law is implementing a moral principle such as CFD*. Predicating liability for punitive damages on holistic culpability coheres naturally with this position as well (although any sustained treatment of punitive damages is beyond the scope of this Article). These phenomena are much harder to square with the Palsgraf perspective. If a court must determine whether a defendant is liable to a plaintiff at all by asking only whether he has mistreated her (as opposed to other people), why should it be proper for the court to determine the scope and magnitude of the defendant’s liability by looking at his mistreatment of the plaintiff among other people?

Indeed, the idea that culpability in any form should bear upon proximate cause is mysterious if we adopt the Palsgraf perspective. Recall its central interpretive commitment: that the elements of each tort together define a qualified relational legal wrong of injury, which consists in the breach of a qualified relational legal duty of non-injury, and that each such legal duty and wrong encode a moral duty and moral wrong recognized by the law. On this view, the proximate cause element of the negligence tort must be located within the content of the qualified relational duty of non-injury that the negligence tort defines. For this reason, accounting for the culpability-sensitivity of proximate cause determinations requires characterizing the content of this relational duty in a baroque way: It must be characterized as a duty not to foreseeably injure another person by breaching one’s duty to take care not to injure him, where one’s level of culpability in breaching that duty of care determines whether the resulting injury counts as sufficiently foreseeable that one has breached the qualified duty of non-injury in which the duty of care is embedded.

So even if we are prepared to recognize qualified relational moral duties of non-injury, it seems especially difficult to construe the tort of negligence as encoding any plausible instance of this sort of duty (or correlative wrongs) once the holistic culpability-sensitivity of the proximate cause element is brought clearly into view. It is more natural to regard considerations such as a tortfeasor’s holistic culpability for an action and the fortuitous or foreseeable character of the resulting injury, as helping to determine his responsibility for occasioning that injury.

C.  Form and Substance in the Tort of Battery

Even if it is unnatural to understand the negligence tort as a relational wrong, there are other torts that — at least on their face — are more natural to understand in this way. Consider battery. We have seen that the law sometimes deforms the relational structure of battery, by means of the transferred intent fiction, in order to recognize and enforce a sort of compensatory moral liability that does not rest on relational moral wrongdoing between defendant and plaintiff. But at least in its nonfictional core, one might think, the tort of battery plainly articulates a sort of relational wrong. A defendant batters a plaintiff by harmfully or offensively contacting him, through intending such a contact, without adequate privilege or excuse.

This understanding gains support from the fact that at least certain defenses to battery — which prevent a prima facie case of battery from counting as an actionable wrong, all things considered — appear to have relational structure. Consider the apparent consent privilege. If a defendant reasonably but mistakenly believes that a plaintiff has consented to some harmful contact, then he will escape liability in battery although the plaintiff did not in fact consent. To understand this privilege in relational terms is, on its face, quite natural: It is because A does not treat B wrongfully that she cannot recover from him in these circumstances, although she did not in fact waive her right against being so contacted.

Consider, however, the following case: Suppose that, as part of a puerile reality show, the defendant sprays a stinky gas that could potentially prove lethally injurious at two people, A and B. It is clear that A does not consent, but the defendant simply does not care; by contrast, the defendant has strong reason to believe (in part on the basis of a reasonable interpretation of her gestures) that B does consent. Suppose it turns out that B does not in fact consent (and, for good measure, suppose that the prospect that her gestures might be interpreted as consent, or that she might otherwise be taken to consent, was not reasonably foreseeable to B). If both A and B are killed by the gas, it seems clear that the defendant is liable to compensate both of their estates. But the apparent consent privilege, if taken at face value, denies B’s estate such recovery.

That the apparent consent privilege misfires, in such a case and others like it, indicates that the privilege is only a rough proxy for the common moral judgments that it registers. If a plaintiff reasonably appears to consent (but does not in fact consent) to a defendant’s injurious action, it will often be the case that the defendant cannot be accused of any fault. It may also be true that the plaintiff has foreseeably induced in the defendant a mistaken impression that she is aggressing against him, in which case it may be plausible that she waives her right against the resulting injury via some moral mechanism akin to estoppel. Where neither is true, the defendant is morally liable to compensate the plaintiff for culpably and foreseeably causing an infringement of her right against being harmfully touched (a right that she has not, in fact, waived or otherwise lost). Presented with facts of this kind, it is difficult to believe that a court would not find a way to reach the same conclusion. And if such facts were to arise with some regularity, it would not be surprising to see the common law devise fictional maneuvers to provide for such recovery as a matter of course.

The self-defense privilege can misfire in a similar way. The privilege immunizes an actor from liability for using lethal defensive force against an apparent aggressor if the actor reasonably believes that “the [apparent aggressor] is about to inflict upon him an intentional contact or other bodily harm” and “he is thereby put in peril of death or serious bodily harm or ravishment, which can safely be prevented only by the immediate use of such force.” Thus it is natural to interpret the self-defense privilege, too, as embodying a concern for relational wrongdoing: Although the merely apparent aggressor has in fact done nothing to forfeit or diminish her right against being killed, the defendant who uses lethal force against her cannot be said to mistreat her, given his reasonable belief to the contrary.

But now suppose that the defendant is walking along the ridge of a high cliff when he notices A, proceeding in his direction, with a hunting rifle. A is a perfect doppelganger for a hit woman that has been trying to kill the defendant. The defendant reasonably but mistakenly believes that A is the hit woman and is imminently about to murder him, such that he can save his life only by preemptively killing her. Thus the defendant grabs an innocent and obviously unthreatening bystander who happens to be on the scene, B, and throws B at A, sending both of them off the cliff and to their demises. In fact, A was also an entirely innocent and unthreatening person, returning to her home after using her hunting rifle to cull an overgrown herd of wild animals (and she could not reasonably have known, we may suppose, that she might be mistaken for an aggressor on her way back home). The defendant is straightforwardly liable, both in morality and the law of battery, to compensate B’s estate for lethally using her body without her consent. Intuitively, however, the defendant is morally liable to compensate A’s estate for culpably killing her, too, and one strongly suspects that a common law court would find a way to enforce this liability. But the self-defense privilege, as articulated on its face, blocks such a recovery.

Again the most plausible diagnosis is that the doctrinal contours of this privilege track the contours of the underlying moral terrain only roughly. When a defendant uses evidently necessary deadly force against a plaintiff who reasonably appears to be threatening his life, it will very often be the case that the defendant’s action is not culpable. Sometimes, it may also be the case that the plaintiff has foreseeably and responsibly induced in the defendant this reasonable appearance, such that the plaintiff should be understood to have waived her right against this action in an estoppel-like fashion. In some cases, however, neither will be true — and in such cases, the privilege will immunize from liability a defendant who ought in justice to compensate his victim, because he has culpably and foreseeably infringed her right against being killed or injured.

As such cases suggest, it is facially plausible but implausible upon reflection to construe the self-defense privilege as encoding a principled moral commitment to predicating the defendant’s liability to the plaintiff on whether he treated her wrongfully. In such cases, a relational understanding of the privilege’s moral logic would implausibly divorce this logic from common moral intuition. The more plausible view is that relational wrongdoing between defendant and plaintiff, in mistaken self-defense cases, is not the ground of the former’s liability to the latter even if it mostly tracks this ground well enough. Like the apparent consent privilege, the self-defense privilege’s moral function is not to identify and redress a relational moral wrong between defendant and plaintiff; it is to roughly track features of the defendant’s conduct (in particular, his culpability) and the plaintiff’s conduct (in particular, her responsibility for inducing a relevant mistake about whether she has consented or is aggressing) that help to determine whether the former is liable to the latter.

An obvious objection to the preceding arguments is that they infer too much about the moral substance of battery law from its doctrinal periphery. Most paradigmatic battery cases, of the sort that figure prominently in casebooks and treatises, do indeed involve relational wrongdoing or mistreatment: A defendant intentionally touches another person, in a manner not unlikely to be at least somewhat harmful or offensive, without any adequate reason for doing so. And so, one might suggest, the tort of battery is, at least in its heartland, the law’s recognition of a sort of relational moral wrong — a moral wrong that is distinctive in kind from, and generally graver than, the moral wrong recognized by the tort of negligence, for which reason the law separately recognizes it and attaches distinctive consequences to its commission.

But it proves very difficult to articulate any plausible account of what this distinctive sort of relational moral wrong is supposed to be. Attempts to do so, including by defenders of the Palsgraf perspective, tend quickly to lose sight of battery doctrine — not only its periphery, but also its heartland. Kantian philosophical formalists, for example, suggest that battery and trespass encode the relational wrong of using another person’s body, which (along with her property) belongs to her, without her consent. The suggestion is attractive because harmfully instrumentalizing other people’s bodies is widely regarded as an especially serious form of interpersonal mistreatment: To kill someone in order to eliminate him as a romantic rival is seriously wrongful, but to enslave him, or kill him in order to appropriate his organs, is even more so.

In many paradigmatic cases of battery, however, the defendant does not in any ordinary sense of the term use or instrumentalize the plaintiff. To shoot another person in order to eliminate him as a romantic rival is classic battery — but it cannot be understood as “using” him except by defining a sense of the term that is so capacious that it is essentially stipulative and conclusory. Even waiving this objection and granting that intentionally making contact with an object to eliminate it can count as “using” it, the proposal fails to track well-settled law. It is well-established that if a defendant intentionally makes contact with an object that she mistakenly regards as an animal or inanimate object, and the object is in fact the plaintiff’s body, the plaintiff’s action will sound in negligence, not battery. Nor does the Kantian view make room for the fact that a defendant can batter a plaintiff entirely as a side effect of pursuing her own goals, by unintentionally imposing on the plaintiff a “substantial certainty” of harm. To be sure, the law equates such “substantial certainty” to intent, for the purpose of imposing liability in battery and the other intentional torts. But this equivalence is plainly a fiction.

The existence of such a fiction, moreover, tells against another tempting account of the distinctive moral wrong of battery, on which the wrong consists in deliberately harming or offending (or deliberately touching in a foreseeably harmful or conventionally offensive way) another person. Suppose that the defendant is a hit man seeking to murder A with a sniper rifle. A is in his line of sight, but then B, whom the defendant has no desire to kill, moves in front of A. As the defendant knows, therefore, if he pulls the trigger of his rifle, the bullet will unavoidably kill B on its path to killing A. When he pulls the trigger and kills both A and B, the defendant seriously wrongs both of them. And it does not seem that the defendant’s wrong to A is significantly graver than, or in some important respect morally distinct from, his wrong to B: These victims are more or less equally mistreated. Certainly these culpable and wrongful injurings do not seem morally distinct in any way that bears upon the defendant’s remedial moral liabilities: The defendant is liable to B’s estate no less than A’s estate, and it would not be morally permissible or proper for the defendant, if rendered insolvent, to give any significant priority to satisfying one liability over the other.

The common law’s fictional contrivance to treat “substantial certainty” of harm as equivalent to an intention of harm, for purposes of liability in battery and the other intentional torts, is in part an implicit recognition that substantially certain injurers are often no less culpable than intentional injurers and thus properly exposed to a similar scope of liability. To be sure, this fictional contrivance serves multiple ends; so, for example, substantial certainty can function as an evidential proxy for an actual intention of harm, thus obviating the difficulties that might attend a plaintiff’s attempt to establish that the defendant actually intended her harm. But such a function could also be served by treating substantial certainty as raising an extremely strong but defeasible inference of actual intent and imposing upon the defendant the burden of undercutting this inference. That no such modification of the current regime has been entertained is some further indication that the fiction also serves other ends.

In particular, the fictional assimilation of substantial certainty to intention enables the law to enforce ordinary moral convictions that its formal structure would otherwise leave unenforced. Thus, for example, it allows the law to deny the defendant, as an intentional tortfeasor, the ability to plead the plaintiff’s own fault in order to eliminate or diminish her recovery; precludes the defendant from discharging the liability in bankruptcy; and allows the law to readily widen the scope of the tortfeasor’s liability beyond the bounds of reasonable foreseeability. These results seem proper when the batterer is highly culpable, as will often be true of the substantially certain injurer (just as it will often be true of the intentional injurer). It is widely recognized by thoughtful legal commentators that these results are more difficult to defend when the batterer is not highly culpable. That fact helps to confirm what reflection on the battery tort otherwise suggests: The distinction between intentional and nonintentional harm set up by the law’s distinction between battery and negligence is a rough proxy for culpability, rather than the law’s reflection of distinct classes of relational moral wrongs or mistreatments.

It is an open question whether using intention (and substantial certainty) as a proxy for culpability, in these various contexts, is ultimately defensible. Especially in light of developments such as the shift from contributory negligence to comparative fault, it is arguable that the law ought to determine the availability of the comparative fault defense, bankruptcy discharge, and similar matters through a direct inquiry into the defendant’s level of culpability. For interpretive purposes, the point is that the moral reasons against retaining the proxy are considerations widely recognized by participants in legal practice as integrally connected to the purpose (to register defendant culpability) the proxy is thought to serve.

In short, while the concepts that lie on the formal surface of the battery tort — intention, apparent consent, and so on — suggest that the tort’s function is to identify and redress relational wrongs or mistreatments, there is good reason not to take this surface language at face value. It is because the battery tort is largely performing the same normative function as the tort of negligence — identifying and enforcing remedial moral liabilities that arise from the culpable and foreseeable infringement of rights against bodily injury — that battery doctrine must make use of fictional devices such as assimilating substantial certainty to intent and transferring intent between victims. The civil law codes have little need for such fictions, for clauses such as BGB section 823(1) and CC 2043 seek to perform this function on their face.

The vast majority of nonprivileged batterers are at least somewhat at fault, even if the fault is slight. There is, however, a small residuum of battery cases that do not involve fault at all. In some of these, the defendant justifiably imposes a heightened risk of harm, so as to satisfy the “substantial certainty” prong of battery’s intent element. Such liability can be grounded in a non-fault-based form of moral responsibility, the same sort of moral responsibility that underlies strict liability for abnormally dangerous activities and the doctrine in Vincent v. Lake Erie Transportation Co.

Other cases may require a different account, however. In particular, battery law typically holds liable a mentally insane defendant who acts on an intention to make contact with a plaintiff for the resulting harm or offense, even if the defendant is incapable of appreciating that this contact might harm or offend. This fact might be adduced as evidence that battery law truly is concerned with a type of relational wrongdoing or mistreatment — intentionally touching another person, in a harmful or conventionally offensive manner, without that person’s consent — whether or not the defendant can be regarded as at fault or otherwise responsible for engaging in that conduct.

The argument is ultimately unconvincing, in my view. Battery law’s insensitivity to insanity is entirely of a piece with tort law’s general insensitivity to cognitive disability, including in the heartland of negligence. In the case law, this general insensitivity is most often explained by administrative, evidentiary, and practical considerations, as well as suggestions that the cognitively disabled should in fairness be strictly liable for the heightened risks they impose on others by going through the world in their condition. Both lines of explanation may be dubious, on reflection. It is arguable that, like most of the civil law codes, the common law should decline to impose tort liability on defendants that truly lack mental capacity except where their lack of capacity has been self-induced through responsible actions they previously performed. The point for present purposes is that, by its own lights, the common law does not appear to impose tort liability on defendants lacking mental capacity because it believes they have genuinely mistreated or morally wronged the plaintiffs they have injured. Whether or not justifiable, the common law’s harshness toward mentally incapable defendants is a departure from its general policy — operative across torts including battery and negligence — of holding defendants liable for infringing others’ rights in a morally responsible manner.

D.  Wrongful Death and Loss of Consortium Actions

The divergence between formal structure and moral substance is even more glaringly apparent in the domain of the derivative actions for wrongful death and loss of consortium, an area with far greater significance than the infrequently litigated law of battery. Suppose the defendant negligently drives into a pedestrian, injuring or killing her, and leaving the pedestrian’s wife bereaved. The pedestrian’s wife may recover against the driver on behalf of the pedestrian’s estate. But she may also recover on her own behalf, for losing the social, financial, and inherent benefits of her relationship with her wife, and (in many states, at least) for her resulting emotional distress. So, too, if she loses such benefits because her wife is grievously injured rather than killed.

Such actions for wrongful death and loss of consortium are derivative, rather than freestanding: They allow the kin of a primary victim to recover but only upon demonstration that the primary victim was tortiously injured. These causes of action are not torts, but doctrinal appendages to the various torts. Thus it might be suggested they are best understood as an inaccurately taxonomized variant of survival actions (which are brought on behalf of a decedent’s estate rather than in the name of her kin, even if the estate’s assets are then predictably distributed to the decedent’s kin, and even if a kinsman might serve as the estate’s representative). In the large majority of states, however, this suggestion does not comport with the remedial legal consequences of wrongful death suits. Wrongful death damages are not swept up into a decedent’s estate; the decedent’s creditors cannot lay claim to them. Similarly, the tortfeasor cannot set off debts owed to him by the decedent against the wrongful death damages that he owes to the decedent’s kin, as would be possible if these damages formed part of the decedent’s estate.

These remedial legal implications of wrongful death suits comport with the plausible moral judgment that the decedent’s kin have claims against the tortfeasor in their own right, although he has not breached any legal duty of care owed to them. The same goes, mutatis mutandis, for loss of consortium actions. The availability of these kinds of actions is readily explicable if the common law is implementing a moral principle such as CFD*. When the defendant negligently kills or injures his primary victim, he does not only infringe the primary victim’s moral rights; he infringes moral rights held by the victim’s kin. These parties have moral rights that the defendant not cause a sort of damage (the death or injury of the primary victim) that will destroy or injure the morally valuable intimate relationship they share with her.

That is why kin are legally empowered to recover for the defendant’s foreseeable injuring of the primary victim, even if they are unforeseeable victims. Suppose that Amelia is a cop, who has been deep undercover in a drug syndicate for many decades. Ron, a member of the syndicate, spends virtually all of his time together with Amelia — and it is thus exceptionally improbable, on Ron’s evidence, that Amelia has any kin. If one day Ron negligently and foreseeably injures Amelia, he will still be liable to compensate her kin, both in morality and in tort. The common law recognizes and enforces their moral rights only indirectly — not in the form of any primary legal right held by the kin, but only by means of the derivative actions for wrongful death and loss of consortium — but it recognizes and enforces these rights all the same.

The derivative actions for wrongful death and loss of consortium are thus strong and pervasive counterexamples to the Palsgraf perspective, for they involve plaintiffs suing for injuries sustained through defendants acting wrongfully toward other people. It is surprising, then, that Cardozo does not deign to discuss wrongful death suits or loss of consortium actions at all in his opinion. Some defenders of the Palsgraf perspective follow suit, simply ignoring the threat to their view that such cases pose. But others, such as Goldberg and Zipursky, have heroically sought to ward off this threat. Their main argument is that — as infamously laid down by Lord Ellenborough in the early nineteenth-century English case Baker v. Bolton — the common law did not provide for wrongful death suits until legislative intervention forced it to. The ingenious claim is that reflection on wrongful death suits thus supports the Palsgraf perspective: Because predicating liability on a relational wrong to the plaintiff is so central a pillar of the common law’s architecture, a statutory wrecking ball was required to (partially) demolish it.

One problem with this argument is that the history is much less tidy than Goldberg and Zipursky suggest. While English courts remained hostile to wrongful death suits until Parliament intervened, the courts of many American states allowed husbands and fathers to claim against the killers of their wives and children even before the passage of wrongful death statutes by state legislatures. In fact, there are plausible arguments — as the eminent Judge Baron Bramwell suggested sixty-five years after Baker, and as suggested afterward by leading historians of English law — that Baker v. Bolton was itself “a careless overstatement of the law.” It is not obvious, therefore, that Baker fairly reflected the broader corpus of English common law, even if it was binding precedent for a time. To take the case as divining a deep truth about the normative character of the common law may be suspect, even in England.

Leaving these objections aside, Goldberg and Zipursky’s argument relies on a strong interpretive distinction between the judicially created common law of torts and the body of tort law that is produced by statutory enactment. That distinction seems dubious; tort law, like the private law more generally, has been a creature of statute as well as judicial decision for many centuries. More importantly, at least for our purposes, the argument seems to run roughshod over the intuitive character of the moral instinct behind the wrongful death statutes. The instinct was not (only) that wrongful death actions might better deter wrongdoing, or punish wrongdoers, or secure humanitarian assistance for the innocent victims of misfortune. It was (also) that interpersonal fairness and justice supported recognizing the decedent’s kin’s claims to redress as against the tortfeasor.

So, for example, the seminal New York wrongful death statute maintains that “in every such action the jury may give such damages as they shall deem fair and just, with reference to the pecuniary injury resulting from such death to the wife and next of kin of such deceased person.” The language of such an enactment precludes construing it as an exercise in engineering social incentives or a humanitarian gambit for distributing the costs of misfortune. Rather, it reflects the widespread moral sentiment that we often have a claim, in justice, to redress for the injuries we suffer through the killing or injuring of our loved ones. It is simply that the formal structure of the common law, as (perhaps infelicitously) described and cemented by Baker v. Bolton, did not permit this prosaic sentiment even rough expression.

Now the common law does permit this, of course, with the benefit of statutory intervention; but even now the derivative character of the actions for wrongful death and loss of consortium refracts the sentiment only obliquely. By contrast, many of the civil codes explicitly permit secondary victims to recover in their own right for such injuries. By taking the common law’s doctrinal architecture “at face value,” the Palsgraf perspective obscures how it is founded in ordinary moral sensibilities that do not differ as between common law countries and the European continent.

Understanding the normative logic of such derivative actions opens up intellectual and conceptual space for doing practical justice elsewhere in tort law. Suppose that a therapist negligently treats his patient, a child, and thereby induces in her false memories of abuse at the hands of her parent. It is plain that the child may recover from the therapist, but may the parent? Courts that have addressed such cases are split. Some recognize a duty of care to the patient’s parent; others do not. All of these courts agree that whether the parent may recover turns on whether the therapist owes a duty of care to her, on top of the duty of care inarguably owed to her child.

These courts tend to frame the duty analysis in nakedly instrumental terms, as a matter of “whether the social benefits of imposing a duty outweigh the social costs of imposing a duty.” Defenders of the Palsgraf perspective are not amenable to this mode of analysis; after all, much of the early impetus for the Palsgraf perspective’s development was its aversion to collapsing duty analysis into a “blunderbuss” judicial weighing of social costs and benefits. Rather than reverse engineering conclusions about the duty of care by way of instrumental calculations about liability, defenders of the Palsgraf perspective believe that judges should closely examine the existing legal materials and the customary moral norms that they embody in order to discern whether those materials and norms are best understood to recognize that the defendant owes the plaintiff a duty. Conclusions about duty, so reached, should drive conclusions about liability rather than vice versa.

In the sort of case at hand, defenders of the Palsgraf perspective have suggested, liability to the parent should not be imposed. That is because recognizing a duty of care to the parent would be incompatible with the therapist’s fiduciary obligations to his patient, which require that his attention in treating her be focused exclusively on her interests rather than the interests of third parties. Not only may the therapist not take care as to the interests of others, he must not take care as to the interests of others. So there is no duty of care to others, such as a parent, on which the parent might found a claim to redress if she is injured by the therapist’s mistreatment of her child.

The pigeonhole perspective suggests a different way of approaching these cases and duty analysis more generally. It is probably true, in such a case, that the therapist ought to be more or less exclusively attentive to his patient’s interests: In making decisions about how to treat his patient, he ought not to act on any concern for the interests of third parties. Probably he should attempt to screen such concern from his deliberative radar entirely. In this sense, it is accurate to describe him as lacking any moral duty of care to third parties. Yet it seems perverse to infer from this observation that, if the therapist carelessly induces in his patient false memories of severe abuse at her parent’s hands, her parent ought to be precluded from recovery, as a matter of moral principle.

That conclusion does not follow if a defendant’s remedial legal and moral liability toward a plaintiff is ultimately grounded in his wrongfully and culpably acting in a manner that foreseeably causes damage that infringes her rights. That is plausibly the case here. After all, the parent’s interests in avoiding a serious injury to her relationship with her child, and the resulting reputational and emotional injuries, would seem amply sufficient to ground a strong moral right against the therapist’s action.

That the therapist should not direct his attention or care toward satisfaction of this right does not imply otherwise. For it is a mistake, in general, to connect duties of care too closely with rights against injury or remedial liability for their infringement. Suppose a therapist has undertaken contractual and promissory obligations that he cannot discharge compatibly with discharging his fiduciary obligations to his patient. Absent exceptional circumstances, he should exclude this fact from his deliberation about how to treat his patient — his care, in treating her, should concern her interests alone. From this it does not follow that, if somehow he breaches his promises in the course of treating his patient, he is not morally liable to compensate the promisees (or ought not to be legally liable). That the therapist should be liable to compensate the patient’s parent is similarly plausible.

It is then a further question how this sort of remedial moral liability ought to be recognized and enforced in the doctrinal architecture of the negligence tort. The standard strategy would be to recognize a legal duty of care running between therapist and parent, as some courts have done. But this strategy risks the perverse effects that have led other courts (along with defenders of the Palsgraf perspective) to decline to recognize any such duty of care: It risks perverting the proper shape of the therapist’s deliberation, by conveying to him that he ought in treating his patient to incorporate a substantial concern for the interests of another party. And it risks generating legal incentives for him to act on such a concern, including in circumstances where it would conflict with a concern for the interests of his patient.

The more elegant and plausible solution is simply to allow the parent in such a case to bring a claim for her injury and join it to her child’s action — in the manner of a wrongful death or loss of consortium claim — and allow the parent to recover on this derivative claim if but only if the therapist is found to have breached his duty of care to the child. Nothing about the parent’s interests is folded into the content of the defendant’s duty or the determination of breach. These interests come in only at the remedial stage, once breach has been established on independent grounds. Thus this structure does not impose upon the therapist deliberative duties incompatible with his fiduciary role, convey to him that he must entertain a concern for third-party interests in the course of treating his patient, or risk incentivizing him to attend to such interests in cases where they conflict with the interests of his patient. In this way, the law can recognize and enforce the plaintiff’s underlying moral rights against injury in a more subtle way: It need not, and should not, predicate the plaintiff’s recovery on the defendant’s breach of any duty of care or other legal duty to her, as the Palsgraf perspective would insist.

The example helps to illustrate why the divergence between tort law’s moral substance and conceptual structure might, despite its propensity to inhibit perspicuous analysis and understanding, nevertheless sometimes be desirable. The relational structure of the duty of care in negligence allows us to entertain remedial structures such as the above, in which the injured plaintiff gets to piggyback on the defendant’s breach of duty to another. If the law had to supply such a recovery through an omnilateral duty of care, of the kind urged by Judge Andrews and the Third Restatement, it might risk inducing the sort of problematic moral deliberation just discussed.

E.  The Borderland Between Negligence and the Intentional Torts

The torts of negligence and battery figure heavily in defenses of the Palsgraf principle, but other torts sometimes make an appearance as well. Suppose a defendant builds a hideous fence on A’s land, without her consent, and thereby causes the value of neighboring land owned by B to drop; or that the defendant defrauds A of her life savings and thereby prevents A from financially supporting the indigent B; or that the defendant makes a defamatory statement regarding A and thereby causes A’s business partner, B, to lose out on lucrative business opportunities. In each case A has an action against the defendant — in trespass, deceit, and defamation, respectively — but B does not. According to scholars who defend the Palsgraf perspective, it is the Palsgraf principle that explains these kinds of asymmetries: A tort defendant must behave wrongfully toward the plaintiff, by mistreating her in the manner articulated by the relevant tort, in order for her to recover.

But these asymmetries are equally explicable on the view that the fundamental normative basis of compensatory tort liability is the defendant’s moral responsibility for infringing some right against injury held by the plaintiff. In each of the preceding cases, the defendant may be morally responsible for causing B to suffer a loss, for such loss is a foreseeable consequence of the defendant’s wrongful behavior. But in none of these cases does B have a right against the loss in question; in general we do not have a right against being deprived of future profit or otherwise caused to suffer economic loss. Thus, each of these results is as ably explained by supposing that common law torts such as defamation are remedial pigeonholes — which obliquely implement normative principles about moral responsibility for rights infringement — as they are by supposing that torts are legally recognized relational moral wrongs. It is because the civil law’s general clauses are (somewhat more transparently) implementing the same principles that in each of these cases the civil law reaches the same result as the common law.

In other cases, however, the formal structure of common law intentional torts does appear to produce different implications from the structure of the general clauses in the civil codes. Take the tort of defamation. Suppose that a defendant engages in the following behavior, and thereby negligently causes reputational injury to the plaintiff: The defendant makes a certain true communication regarding the plaintiff’s wife to X, which foreseeably leads X to form a belief in a different proposition and communicate that proposition to Y, on which basis Y foreseeably forms the belief that the plaintiff is a serial killer and communicates this belief to Z. Because the tort of defamation affords redress only where a defendant’s communication is “of and concerning” the plaintiff, the plaintiff will have no redress for this “reputational injury.” Without such a communication, injuring the plaintiff’s reputation by initiating a chain of foreseeable inferences does not engage the terms of the defamation tort. Similarly, suppose that a defendant negligently and foreseeably causes another person to intentionally confine the plaintiff. Since a defendant can commit the tort of false imprisonment only by intentionally confining the plaintiff, here the plaintiff will have no false imprisonment claim against the defendant for negligently and foreseeably depriving him of his liberty (or any consequential economic loss).

How should we understand the common law’s restrictions on recovery for negligent injury to reputation and liberty? By the lights of the Palsgraf perspective, such restrictions reflect the content of the relational wrongs that defamation and false imprisonment aim to recognize and redress. Thus, defamation is a wrong that essentially consists in making a defamatory communication about the plaintiff. To suppose that negligently inflicted reputational injury might properly be recoverable is simply to misconstrue the nature of the wrong. Similarly, the wrong of false imprisonment “consists in asserting control over” a person, thus treating that person “as subject to [one’s] choice.” To negligently cause a false imprisonment is not to morally mistreat another person in the manner recognized by the tort of false imprisonment.

But these moral understandings of defamation and false imprisonment are implausible, it seems to me, and cannot make sense of the law. Take them in turn. It is certainly morally intelligible for the law to narrowly draw the boundaries of negligence liability for communicative activity in order to refrain from unduly impinging on the free flow of information and expression. Thus it is intelligible for the law to restrict fault-based liability for communicatively injuring another’s reputation — whether under the aegis of defamation or negligence — to statements “of and concerning” the plaintiff, just as the law tightly draws the boundaries of fault-based liability for causing physical injury by communicative means.

By contrast, it would not be morally intelligible for the law to suppose that negligent communicative activity cannot in principle ground a plaintiff’s claim to compensation unless some communication has been made “of and concerning” him. Any such view would fetishize the “of and concerning” requirement in a manner that can claim little support from ordinary moral intuition. We care about false and derogatory statements “of and concerning” us simply because such statements are particularly likely to injure something — our reputation — to which we are entitled.

Communicative activity of other kinds will generally be much less likely to falsely injure reputation. Thus, it may be unwise to expose all such activity to the specter of litigation. But that is not because the law limits its principled moral concern for the negligent infliction of reputational injury to statements of and concerning injured plaintiffs. After all, in cases involving noncommunicative activity that negligently injures reputation, the law is much less wary of affording redress, which it generally affords under the auspices of the negligence tort rather than defamation. So, for example, courts have allowed physicians to recover in negligence for reputational injury foreseeably inflicted by the negligent servicing or design of medical products.

Whether and when the tort of negligence ought to allow recovery for reputational injury is a difficult issue. It implicates technical doctrinal questions about the analytical relationship between the elements of negligence and defamation, predictive questions about the possibility that negligence plaintiffs might make an end run around important values encoded in restrictions on defamation recovery, and philosophical questions about how to weigh the importance of protecting free communication against the importance of enforcing remedial moral liabilities that arise from the careless infringements of rights against reputational injury. It is not obvious how these trade-offs should be struck.

The important point for our purposes is that, once we observe that infringements of rights against reputational injury will often generate powerful remedial entitlements to redress in principle, that will change how we understand those elements of the defamation tort that regularly operate to preclude such relief. So, to return to the example above, when someone foreseeably suffers grievous reputational and economic injury because her intimate has been defamed, we will not interpret the law’s bar on her recovery as founded upon the deep normative nature of the defamation tort. Rather, we will treat it — far more plausibly — as a concession to institutional limitations, a recognition that a widened defamation tort might unduly suppress communication, or an unwillingness to allow the secondary victims of defamation to drag its primary victims into potentially costly and embarrassing court battles that they would rather avoid.

Recognizing that the normative considerations that support the restriction are less deeply rooted in the nature of the tort than the Palsgraf perspective supposes should, in turn, open up room to relax or remodel those restrictions in appropriate cases. Take the live issue of whether the developers of large language models may be held liable to persons injured by their models’ “defamatory” outputs. It may be dubious to characterize these outputs as communications or publications made by the developers, and the models themselves are not (for now, at least) legally cognizable as agents who can commit torts and subject developers to vicarious liability. Thus, the tort of defamation arguably supplies no redress for the victims of such defamatory outputs. Contra the logic of the Palsgraf perspective, however, nothing in the normative nature of the defamation tort precludes extending either the tort of defamation or negligence (or negligence-like forms of products liability) in order to afford such redress.

So, too, with false imprisonment. It is understandable that the law should be wary of recognizing claims for negligently imposed false imprisonment, given how readily liability for minor and relatively harmless detentions might then attach. But it would be a mistake to suppose that a company that negligently designs or deploys an artificial intelligence model in order to surveil suspected shoplifters, and thereby foreseeably causes innocent customers to be imprisoned without cause, is insulated from claims to redress by the deep normative logic of tort law. The structure of civil law clauses such as BGB section 823(1) straightforwardly accommodates claims for negligently and unjustifiably infringing rights against liberty-depriving detention. Common law development might prudently do the same — whether by recognizing such detentions as a new form of actionable damage in negligence or by recognizing a new and delimited tort, such as a tort of recklessly causing false imprisonment.

The point is not just that such developments would be normatively attractive. These developments would be the sort of “modest but principled extension[s]” of existing doctrine that are recognizably within the legitimate boundaries of the judicial role. That is because these doctrinal developments would give expression to a principle of common morality — that a defendant who negligently, foreseeably, and unjustifiably infringes a plaintiff’s right against deprivation of reputation or liberty is liable to compensate her — that is already recognized and implemented, even if indirectly and imperfectly, by the common law of torts.

F.  Torts Without Wrongs and Intra-Doctrinal Moral Inconsistency

We have discussed many cases in which a plaintiff injured by a defendant’s culpable and unjustifiable action can recover from him, notwithstanding that he has not mistreated her. The propriety of relief in such cases undercuts the Palsgraf perspective. But so, too, does the vast amount of tort doctrine in which the law empowers a plaintiff to obtain relief from a defendant who has not acted culpably or unjustifiably at all.

The situations I have in mind are familiar ones. If a defendant uses a plaintiff’s dock to save his ship in a storm, the defendant will be liable to compensate her for the resulting damage, although the trespass was entirely reasonable. If a defendant reasonably decides to conduct some ultrahazardous blasting in order to demolish an abandoned shack and build a new house on her land, and foreseeably ends up injuring her neighbor as a result, the neighbor will be able to recover. If a defendant runs a small business, and hires and trains a new employee who negligently causes a customer to suffer harm a few weeks into the job, the defendant will be liable to the customer even if she cannot be accused of any fault in hiring, training, or supervising the employee. The doctrines that secure recovery in these cases are familiar features of tort law.

The Palsgraf perspective struggles with all of these doctrines because none of them involve the defendant engaging in wrongful behavior toward anyone, let alone the plaintiff. In response, defenders of the Palsgraf perspective reject the legitimacy of these doctrines (and the powerful moral instincts embedded in them); or relegate them to the interpretive margins of tort law; or rely on metaphysical claims, such as the claim that an employer acts through the employee for whom it is vicariously liable, that are legal fictions (and thus call for a substantive normative rationale that the defenders of the Palsgraf perspective do not then supply); or appeal to the claim that (some of) these defendants are committing legal wrongs, which is a formal observation that does not address the substantive objection that these putative legal wrongs are not tracking moral wrongs.

These are all existing criticisms, and I am not trying to add to the stock of criticism here. My point is that the Palsgraf perspective is only impelled to engage in such strained interpretive maneuvers because of its fundamental assumption that tort liability is essentially about redressing wrongs. All of these doctrines make more sense if tort liability seeks to enforce remedial moral liability, and remedial moral liability is grounded in a defendant’s moral responsibility for infringing the plaintiff’s rights against injury. In the preceding cases, of course, the defendant does not exhibit the sort of fault-based moral responsibility that is tracked by the tort of negligence. But an agent’s exercise of her responsible agency can sometimes be morally significant for reasons other than its wrongfulness and culpability: because (say) the agent has responsibly decided to pose a substantial and nonreciprocal risk of infringing others’ rights against injury or because she has innocently engaged in a certain sort of collective enterprise with someone who has culpably and wrongfully infringed the rights of another person.

Arguably, ordinary moral thought recognizes non-fault-based species of remedial moral responsibility less confidently and robustly than it does fault-based remedial responsibility. There is, after all, a recurring impulse in moral philosophy to reject non-fault-based species of moral responsibility on considered reflection. The same impulse recurs in the history of normative reflection on tort law, sometimes with profound effects on the shape of doctrine. In Germany, for example, this impulse has led to a significant curtailment of the scope of strict liability and the reconceptualization of old strict liability doctrines in terms of fault. In America, this impulse may have helped to curtail the twentieth-century expansion of enterprise liability and to reinstate the historical dominance of the tort of negligence.

The impulse reflects the appeal of explanatory unity and simplicity in legal doctrine and moral theory. But it also derives its force from the fact that our moral judgments about fault-based and non-fault-based (“strict”) species of remedial moral responsibility may well prove to be inconsistent with one another — generating incoherence in legal doctrine and ordinary moral thought and reflective pressure to tame this incoherence by pruning the species of remedial moral responsibility that we ultimately recognize. Consider an agent who saves her own life from a credible threat by turning her car in the direction of a pedestrian, knowing that it is extremely likely to collide with him and inflict a moderately serious injury on him (but not intending to do so). The agent will be held liable in battery to her victim, on the basis that she was “substantially certain” her action would injure him. Neither the law’s articulation of the battery tort, nor any of the battery tort’s defenses, will allow the agent to escape liability to her victim on the basis that her action was permissible and faultless. Yet if the agent saves her own life by imposing (say) only a five percent chance of inflicting such an injury, her action will be treated under the tort of negligence, and she will easily evade liability, without even needing to plead her absence of fault as an affirmative defense: The law articulates the content of the tort of negligence such that performing a permissible and faultless action, however injurious, does not even count as committing the tort.

There does not obviously exist any coherent, consistent, and reflectively plausible set of moral principles that implies that the agent should be liable in the first scenario but not in the second. And, if we attempt to iron out the inconsistency so as to produce consistent moral and legal principles, we may well produce a fearfully broad account of the conditions of strict remedial liability (under which we are strictly liable to compensate for the vast majority of the injurious consequences of our actions, far more often than almost all of us assume and far more often than the law currently holds us liable) or else an implausibly anemic account, according to which we are hardly ever strictly liable to compensate for such consequences. Some tort theorists have adopted such revisionary normative accounts in an attempt to follow principled consistency where it leads. But ordinary moral intuition and tort doctrine both balk at following in their steps.

If the common law of torts were ever to be comprehensively remodeled in the manner of a civil law code, the various tensions and inconsistencies between its underlying ordinary moral sensibilities about fault-based and strict moral liability — which seem to be reflected by the law’s gallery of torts and the doctrinal tensions between them, as we have seen — might face greater pressure to be ironed out. But to date, at least, the pigeonhole structure of the law has largely defused any such pressure toward general and systemic normative consistency and its potentially revisionary effects on the doctrine. There is generally little appetite in common law decisionmaking or scholarship “to put the entire law into systematic and exhaustive order,” as there was in European legal culture in the decades preceding and during the drafting of the civil codes.

As an interpretive matter, the most plausible position may be that the moral principles regarding remedial moral liability that underlie the various torts are sometimes genuinely inconsistent with one another. Some of these tensions are much more than surface deep. It is an interesting and important normative question about the design of tort doctrine whether and how such recalcitrant normative tensions between the principles about remedial moral liability recognized in different parts of the law should be ironed out. An interpretive theory of tort law need not (and probably should not) resolve these first-order normative questions about the revision of doctrine, but it should at least frame them perspicuously.

Defenders of the Palsgraf perspective have, by and large, skirted any serious engagement with such issues about inconsistency between the moral principles underlying different torts. Pragmatic constructivists such as Goldberg and Zipursky do so by adopting a sort of philosophical quietism, according to which they need not offer any systematic account of the moral principles that underlie the various torts: “[W]e do not need a substantive theory of wrongs if we have a positive account of what the wrongs of tort law are and a normative theory of adjudication appropriate to this domain of law.” That is because common law adjudication is “not a matter of finding the objective truth as to which acts really are wrongs, but of constructively carrying forth ideas, principles, and norms that are already in the law.”

But this claim relies on a false dichotomy. The intellectual pressure to articulate the relationship between tort law’s legal structures and its underlying moral principles is not simply an alien demand imposed upon pragmatic common law judges by legal theorists with a taste for overly nice philosophical abstraction. It is a pressure that arises from within the enterprise of principled and consistent common law adjudication (and academic reflection on it), for such adjudication unavoidably encounters the evident normative tensions between different planks of tort doctrine — such as the tension between the rule in Palsgraf and the doctrine of transferred intent, or the tension between the conditions under which negligence and battery (respectively) enable plaintiffs to recover for harm that is justifiably inflicted.

The former tension proves only superficial once we see that foreseeability limitations on duty and proximate cause in negligence and the doctrine of transferred intent in battery together implement, if only roughly, a coherent set of complex moral principles about fault-based remedial responsibility. The latter tension is more recalcitrant, as we have just seen. The suspicion thus arises that an important portion of negligence doctrine implements a conception of remedial moral responsibility that cannot be reconciled with the conception of remedial responsibility underlying the relevant battery doctrine. In both cases, the relationship between the torts of negligence and battery and the underlying moral principles they implement is far less tidy than pragmatic constructivists assume. That is true whether the underlying principles are internally coherent or incoherent, jointly consistent or irremediably in conflict.

By contrast to pragmatic constructivists, who attempt to skirt these issues through philosophical quietism, philosophical formalists often handle these issues by subtly misconstruing the doctrine. Thus, Ripstein, following a similar argument by Weinrib, argues that both the tort of negligence and the strict liability torts can be understood as prohibiting the wrong of harming another person by subjecting her to an “excessive risk” (that is, a risk greater than the background risks that attend ordinary social life). The argument is intolerably strained. Someone who imposes an unusually large risk on another person does not seem to treat her wrongfully (in any recognizable and nonfictive sense) if that risk is justifiably imposed. The argument is also unfaithful to a vast swath of settled doctrine. Notwithstanding a couple of famous old English cases, such as Bolton v. Stone (on which Ripstein and Weinrib heavily rely), it is black-letter law that a plaintiff will not recover in negligence against a defendant who has injured him by reasonably and carefully imposing a risk upon him, unusually large though that risk may be.

What may produce the distortion is the insistence that all causes of action in tort law be understood as encoding wrongs. That insistence requires treating strict liability causes of action such as strict liability for abnormally dangerous activities as encoding a sort of wrong, strange as the putative wrong may be — injury through the imposition of an excessive risk, reasonable or not. This implausible characterization of strict liability for abnormally dangerous activities in turn requires mischaracterizing negligence doctrine, since the philosophical formalist must then maintain that negligence doctrine does not permit the reasonable imposition of unusually large risks on others, which plainly it does.

Again the lesson, I believe, is that the core moral function of torts is not to encode and redress wrongs. Rather, a cause of action such as strict liability for abnormally dangerous activities implements common moral judgments according to which justifiably and nonculpably risking infringement of others’ rights against injury can generate moral liability to compensate them. This cause of action is, like negligence, a remedial pigeonhole that roughly tracks and enforces widely recognized remedial moral liabilities, grounded in a certain species of moral responsibility for rights infringement — one that is very different from the fault-based species of moral responsibility recognized by civil law clauses such as CC 2043 and BGB section 823(1) and the common law tort of negligence.

That said, there are certain instances of strict liability in tort that cannot be entirely understood as reflecting a defendant’s moral responsibility for rights infringement, whether fault-based or non-fault-based. The property torts, in particular, resist explanation along these lines. These torts are famously prepared to impose strict liability on a defendant who unforeseeably causes harm, such as by crossing land or converting a chattel that he could not reasonably have predicted might belong to anyone other than him. To be sure, it is possible to culpably commit trespass or convert a chattel. Similarly it is possible, as Vincent shows, to commit a property tort in a manner that imposes a substantial and nonreciprocal risk of infringing another person’s rights against property damage. In large part, therefore, the property torts can be seen as identifying and enforcing remedial liabilities resting on both fault-based and non-fault-based forms of moral responsibility for rights infringement. As the case of the unforeseeable and innocent trespass shows, however, that is not all they do.

One response to this observation is to conclude that the Palsgraf perspective is right: The unforeseeable trespasser does commit a moral wrong, and the tort of trespass aims to recognize and redress this wrong. But it is difficult to identify any plausible candidate for what such a wrong would be. There is an exceptionally thin sense (the fact-relative sense) in which infringing another person’s right, however unforeseeably, is a moral wrong. But tort law nowhere else holds a defendant liable simply for committing a moral wrong in this exceptionally thin sense.

Philosophical formalists have suggested that the wrong is a more specific one: It is the wrong of using another person’s means, what belongs to them, without their consent. But neither ordinary moral thought nor the rest of tort law appears to suppose that unforeseeably using another person’s means exposes a defendant to compensatory liability. Suppose a defendant is camping in a forest, and he burns some large logs, lying on the ground, for a campfire. Unforeseeably, the logs are persons whose bodies have been masterfully disguised by a villain. Neither battery, nor negligence, nor any other tort would provide a cause of action against the defendant; the battery claim would fail because the defendant has not intended to contact the person of another. The absence of any such action comports with the common moral intuition that such utterly unforeseeable harm lies beyond the scope of “how far the responsibility of [a] defendant ought fairly to extend.”

The interpretive task is to explain why, unlike battery or negligence, property torts such as trespass and conversion do impose liability for harm without any inquiry into fault or foreseeability. As judges have recognized, dispensing with any such inquiry can offend against “notions of ‘elemental justice.’” For example, if in the preceding hypothetical the defendant had burned a log of wood that unforeseeably belonged to the plaintiff, rather than an apparent log of wood that unforeseeably constituted the plaintiff’s body, the law would hold the defendant liable. Rather than conjuring up a strained sort of moral wrong to supply the explanation of this starkly counterintuitive position, the better view is that the explanation does not sound in wrongdoing or responsibility at all.

The most plausible explanation instead rests on the way that the common law has called upon tort law to protect property. Property torts such as trespass and conversion serve remedial purposes that have little to do with any compensatory moral liability incurred by the defendant — purposes such as settling disputes to title and allowing for the recovery of property that has fallen into a defendant’s hands (however innocent). These noncompensatory remedial purposes are facilitated by defining the scope of the actions for trespass and conversion entirely in terms of the plaintiff’s property right, rather than anything about the defendant. “You have encroached on land that is in fact mine” is a pleading more suited to settling title than “you have encroached on land that you should reasonably have foreseen might belong to another person, such as me.”

In modern civil law systems, as in classical Roman law, these other remedial functions are served by distinct “proprietary” actions directly designed to serve them. Unsurprisingly, then, the civil law does not impose liability upon an innocent defendant who unforeseeably trespasses on, or converts, the property of a plaintiff. That the common law does impose compensatory legal liability on such a defendant — liability to make good any losses the plaintiff has sustained, not just an obligation to return the property in question or refrain from further interference — is a side effect of the common law’s use of liability rules, such as trespass and conversion, for these noncompensatory purposes, not an indication that it takes a fundamentally different position on the elemental justice of compensatory liability for innocent and unforeseeable interference with property than it does with respect to innocent and unforeseeable interference with person.

If the material circumstances of human life were different, causes of action such as negligence and battery might be structured along similar lines as trespass and conversion. If, for example, disputes about the ownership of bodies arose as often as disputes about ownership of land and chattels, the intent element in battery might look more like the intent element in trespass — requiring only an intent to touch an object that is in fact a person’s body (rather than an intent to touch an object that is believed to be a person’s body), just as trespass to land or chattels requires only an intent to touch an object that is in fact another person’s property (even if the defendant reasonably believes otherwise). Given the actual circumstances of human life, there has been little need for torts to person to serve remedial functions other than holding a defendant legally liable to pay a plaintiff when, because she is morally responsible for infringing his bodily rights, she is morally liable to pay him.

G.  Responsibility and Relationality in the Law of Negligence

We have ranged broadly, but we can now return to the heartland of tort law, where we began: liability for the negligent infliction of injury to person and property. The Introduction of this Article offered two cases in which a defendant is intuitively liable to a plaintiff although he has not behaved wrongfully toward the plaintiff, even under some abstract and generic description (such as “the owner of the parcel of land, Blackacre, that might be damaged by this negligent activity”). In such cases, the defendant’s liability must in truth be founded on the fact that he has culpably and foreseeably caused a certain rights-infringing outcome, and the right infringed in fact belonged to the plaintiff.

Reflection on plaintiff-focused defenses supports the point. Defenders of the Palsgraf perspective have said that plaintiff-focused defenses, such as comparative negligence, indirectly reflect the relational character of the tort of negligence: If the defendant is required to exercise due care toward the plaintiff, it follows that the plaintiff in fairness must be required to exercise similar care toward herself. But casuistry casts doubt on this relational understanding of plaintiff-focused defenses, and thus casts doubt on the relational conception of the negligence tort that generates them.

Suppose Jacopo is aware that if he races to his office in his Hummer he is very likely to destroy a car parked by the side of the road, and that one of the many cars parked by the road belongs to him. Jacopo chooses to race in order to close a lucrative business deal. The odds that he might hit a car are high, but the odds that the car will belong to him are low, such that Jacopo’s expected benefit from wrongfully speeding down the road outweighs the expected cost to him of doing so. Jacopo does, unluckily, hit the car that belongs to him. It is Jacopo’s negligence toward all relevant parties that will properly diminish his recovery against whoever negligently parked his car by the side of the road. Since the action is perfectly in his rational self-interest, Jacopo cannot plausibly be said to behave negligently toward himself.

So, too, a plaintiff can be comparatively negligent although she is an unforeseeable victim of her own negligence. Suppose that a defendant, the custodian of a rare car collection owned by Monty Moneybags, negligently parks one of Moneybags’s cars, a rare BMW, by the side of the road. The defendant thereby puts the car at risk of being hit and damaged by a speeding driver. Moneybags dies and in probate his car collection passes to his long-lost third cousin, Clara. As it happens, Clara is at that moment speeding down the road, and negligently crashes into the BMW, damaging it. The custodian should and will be able to plead Clara’s negligence in order to reduce his liability to her. But it can hardly be said that by risking damage to cars by the side of the road, Clara has been negligent toward herself. The odds that by speeding she might crash into and damage an expensive car that she owns, because she has acquired title to the car through a long-lost relative (or any similarly fantastic mechanism), are so low as to render the possibility unforeseeable. Clara herself, in such a case, is an unforeseeable victim of her own negligence, but her negligence will still diminish her recovery.

These cases suggest that, although in most cases of comparative negligence a plaintiff has been negligent toward herself (among others), the defense reflects a more general moral phenomenon: the plaintiff’s responsibility for foreseeably causing the wrongful rights-infringing outcome of which she now complains. Indeed, this is an area in which the non-relational character of the underlying moral dynamics lies close to the formal doctrinal surface of the law: Jury instructions on comparative fault often ask whether the plaintiff’s negligence was causally responsible for the injury of which she complains without any suggestion that the plaintiff need have been negligent toward herself.

In other cases, the law of negligence must engage in doctrinal gymnastics to enforce the ordinary moral principles operative beneath its formal structure. Consider the facts in Shell UK Ltd. v. Total UK Ltd., in which the defendant negligently damaged a piece of property held on trust for the plaintiff, preventing the plaintiff from accessing and exploiting this property, thereby causing it to suffer economic loss. The beneficiary of trust property is not its legal owner: It is the property’s trustee who holds title to it, and thus the beneficiary has a legal right against third parties that they not use or interfere with the property against the trustee’s will. Thus, on black-letter principle, the defendant owed no duty of care to the plaintiff, and the economic losses suffered by the plaintiff were pure economic losses that are unrecoverable (rather than economic losses “consequential” upon the defendant’s infringement of a duty owed to it, which are standardly recoverable).

Faced with this evidently unjust result, the English Court of Appeal simply refused to follow the operative black-letter rule, on the basis that doing so would be “legalistic” — the sort of “triumph of form over substance” that would inhibit “the impulse to do practical justice.” Similarly, it would surely be “legalistic” to deny a retiree recovery for economic losses suffered as the result of the destruction of his retirement fund on the basis that, in truth, it is the investment trustee who owns the fund, and the retiree is only its beneficiary (such that the retiree’s resulting economic losses are pure rather than consequential). As a matter of common social understanding, it is the trust beneficiary in such cases who is its owner, morally speaking — it is the trust beneficiary who has a moral right against the asset’s destruction, not the investment manager or legal entity that legally owns the asset and holds it on trust. Faced with the exceptional case in which the formal structure of the legal duty of care was inadequate to provide a trust beneficiary with just recovery for losses arising from the negligent and foreseeable damaging of an object that was obviously its property — in morality, as recognized in social custom and indirectly in law — the Court of Appeal was readily willing to disregard the relational formal structure of the duty of care in order to enforce the negligence tort’s animating moral instincts.

The case is recent, but the general phenomenon is not a novel one. Take the rule announced in Wagner v. International Railway Co., another famous Cardozo case decided seven years before Palsgraf. In Wagner, Cardozo held that a defendant who had negligently endangered another person could be liable to that person’s companion, for injuries sustained in the course of attempting a rescue. It might seem that such injured rescuer must sue as the “vicarious beneficiary” (in Palsgraf’s phrase) of the negligent defendant’s breach of his duty of care to the primary victim endangered. Not so, Cardozo maintained: “The wrong that imperils life is a wrong to the imperiled victim; it is a wrong also to his rescuer. . . . The risk of rescue, if only it be not wanton, is born of the occasion.” The most natural reconstruction of Cardozo’s thought, as the Palsgraf perspective’s defenders have recognized, sounds in foreseeability: “[T]he prospect of a rescuer who might be injured [is] within the scope of the hazards the negligent defendant [can] be expected to foresee.” Because the rescuer is a foreseeable victim of the defendant’s negligent action, the defendant breaches a duty of care owed to him, not just a duty of care owed to the directly imperiled party. “Danger invites rescue,” as Cardozo memorably put it.

But foreseeability does not seem to supply the sole or even principal normative basis of the rule. That is why, as an earlier generation of doctrinally oriented torts scholars observed, negligence law often supplies recovery in cases where the assumption that a rescuer is foreseeable seems “artificial and fanciful.” We can construct cases that exhibit the point sharply. Imagine two seamen aboard a ship that contains no other crew or passengers, or so the seamen have every reason to believe. One night the first seaman drinks to excess and, in his drunken stupor, crashes his body into the second seaman, sending him careening into the sea. In an extremely improbable turn of events, another person has been stowed away in the bowels of the ship, in a room with a window onto the sea. Upon seeing the second seaman careen past her window, she jumps out of her window and into the sea in a bid to save him, and suffers injury in the course of the rescue.

The first seaman is liable to compensate the stowaway, under the rule in Wagner, even if the prospect of her presence on the scene was so improbable as to be unforeseeable in any natural sense of the term. That is because the rule in Wagner bestows a categorical claim to redress on rescuers, so long as their rescue attempts are not grossly irresponsible. No matter how utterly improbable the presence of a rescuer, or the prospect that she might attempt a rescue, the rescuer may avail herself of the rule. If the rule were otherwise, moreover, certain absurd implications would follow: A prospective tortfeasor would be able to conclusively extricate himself from the prospect of negligence liability to rescuers by ensuring that it is sufficiently unlikely that any potential rescuer is on the scene.

To be sure, courts sometimes stipulate that “rescuers always should be regarded as foreseeable plaintiffs,” such that imperiling one person is always a breach of duty not only toward him but toward his rescuer as well. But the perfectly conclusive and categorical character of this stipulation suggests that it is a semifictive legal construction, by which the law supplies a recovery that is often motivated on other grounds. What those grounds might be is a question that lies outside the scope of this Article. The point is that here, again, the law fictionally deforms the relational structure of the legal duty of care in order to enforce a species of compensatory moral liability that does not actually rest on the defendant’s breach of any duty to be careful toward the plaintiff. The same is true, I have argued, in the heartland of negligence and battery: The law uses a relational formal structure (including a relational duty of care) to enforce non-relational moral principles, such as CFD*, about responsibility for rights infringement.

None of this is to say that the idea of a relational duty of care is incoherent or an illusion. The law of negligence does announce a legal duty to take care against foreseeably injuring others, a duty that is most often formulated as owed to those whom one might foreseeably injure, and the expressive effect of this announcement is to recognize such a moral duty as well. It is no illusion to suppose that there can exist such a relational moral duty, or that it may play various roles in ordinary moral thought. Rather, the normative illusion persistently engendered by the place of the relational legal duty of care in the formal structure of the negligence tort is that the breach of any such relational duty is the ground of liability to compensate for negligently inflicted injury (by the lights of either ordinary moral thought or the law of negligence). This illusion is dispelled only by comparative legal reflection, and by closely attending to cases in which relational wrongdoing (breach of relational duty) extensionally diverges from the overlapping but distinct moral phenomenon of fault-based responsibility for rights infringement.

If this is correct, an obvious question to ask is why the legal duty of care in negligence should have assumed a relational form in the first place. The answer is no doubt complex, and ultimately requires sustained historical analysis of a kind that cannot be performed here. Still, the history of the common law, as it is commonly told, supplies the basis of a plausible answer.

There is a sort of negligence case — in which a defendant does not cause a plaintiff any damage, that is, does not infringe any right against injury that she holds — in which the defendant’s liability to the plaintiff truly does rest upon his breach of a duty to be careful toward her, a duty grounded in some morally significant undertaking or other special relationship that he bears toward her and not others. Imagine, for example, that the defendant is a firefighter who undertakes to A that he will use reasonable care to safeguard A’s lands if it should ever appear that any of A’s lands are under threat of fire damage. If the defendant carelessly fails to prevent a fire from burning down a piece of land that belongs to both A and B, the defendant is liable to A and not to B.

The early common law of negligence handled situations of this kind — situations in which liability genuinely rests upon the breach of a relational duty of care, one that is grounded in an undertaking or other special relationship, to the plaintiff — alongside situations involving negligently inflicted damage to a plaintiff. Without much intellectual self-consciousness, judges used the language of duty to identify when cases of both kinds were actionable. Thus it was natural enough, once judges and doctrinal writers started to distill the doctrinal structure of the modern tort of negligence from this catalog of cases, to subsume actionable cases of both kinds under the overarching doctrinal conceit of a relational duty of care.

On at least one plausible and common understanding of the common law’s history, therefore, that history may explain why we now have a duty of care in negligence with two faces. One face perspicuously discloses the substantively relational, special relationship–based ground of a certain sort of compensatory moral liability. The other face is a mask — a formal device in the nature of a fiction, invoked to enforce a sort of compensatory moral liability that is, in truth, grounded in moral principles of a different kind. And it is this second sort of liability that is at work in the heartland of tort, in Palsgraf and other cases of negligently inflicted damage to a plaintiff.

That the common law of tort has not worked itself pure of such fiction — by explicitly putting tort liability in the second sort of case onto a foundation of general principles regarding the defendant’s responsibility for causing damage to the plaintiff — is perhaps partly explained by the utility of a foreseeability-inclusive duty of care in controlling jury discretion and enabling courts to dismiss less colorable claims at an early stage of litigation. Probably, though, it is also explained by the insistently practical orientation of the common law’s juristic culture. The civil law’s general tort principles were extracted from the coarse formal structure of classical Roman law, with its common law–like pigeonhole causes of action, by means of extensive conceptual and philosophical reflection conducted over several centuries by doctrinalists, philosophers, and theologians on the European continent. Occasionally, tort scholars propose that the judiciary or the legislature should restructure large swaths of the common law in broadly similar form. But common law jurisdictions have never had much appetite for such an enterprise, and they probably never will. If it is asked why the common law of torts has not worked itself pure, by distilling its normative substance into general legal principles not straitjacketed by its received pigeonhole causes of action, “[w]e . . . have got on well enough without” any such purifying exercise is an answer with great force.

Conclusion: Form, Fiction, Substance, and the Moral Foundations of Tort Law

Legal historians have often observed that legal fictions, and other doctrinal mechanisms that create daylight between the common law’s formal apparatus and its substantive normative concerns, have played a significant role in the common law’s development. Medieval actions, such as ejectment and trover, protected property in chattels and land through fictions involving nonexistent lessors and imaginary findings of “lost” property. Until quite recently, the common law ordered reversals of mistaken payments by fictionally stipulating an implied promise by the recipient to reverse any such payment. Creative pleadings asserting injury in breach of the king’s peace — pleadings sometimes strained, sometimes outright false — progressively expanded the range of tort actions cognizable in the royal courts. And, concealed beneath the facially strict character of trespass (the central action in medieval tort law), it is likely that a substantive concern for fault-based moral responsibility was opaquely at work. In all of these ways and others, “the common law has [engaged] in the abuse of its elementary ideas” — manipulating, obscuring, and sometimes ignoring the doctrinal strictures ostensibly imposed by its basic forms — in order to “reach[] reasonable results behind institutions and procedures of quite unreasonable artificiality.”

It is tempting to believe that, now that the common law has escaped from the old forms of action, it “has outgrown the need for legal fictions,” and its formal doctrinal structure can thus “be taken more or less at face value.” That is exactly what the Palsgraf perspective, and much of contemporary private law theory, seeks to do. On its face, the common law of torts is largely concerned with relational wrongdoing or mistreatment — forms of behavior, such as intentionally and unjustifiably contacting another person in a harmful or offensive manner, that violate ordinary moral norms regarding how we must treat one another. If we take this formal structure at moral face value, the Palsgraf principle naturally follows: Only if the defendant has mistreated the plaintiff, treated her wrongfully by breaching a legally recognized moral norm regarding how she may be treated, can she recover from him.

But the Palsgraf principle fails to track a good deal of tort doctrine and its underlying moral sensibilities. That is because the common law’s conceptual architecture — which largely sounds in the language of relational duties and wrongs — does not neatly map onto its substantive moral concerns. Some tort liabilities truly are grounded in the breach of relational duties, the commission of relational wrongs. But these liabilities sit at some distance from the heartland of tort law. In its heartland, tort law makes use of a wrongs-based conceptual architecture in order to implement a complex set of moral principles regarding the conditions under which a defendant is morally responsible for infringing a plaintiff’s rights against injury, such that he is liable to compensate her — whether or not she was foreseeable to him or he has treated her wrongfully. A wrongs-based conceptual architecture can house these substantive principles without undue strain because there is a great deal of overlap between relational wrongdoing or mistreatment (on the one hand) and moral responsibility for rights infringement (on the other). Nevertheless, there is enough divergence between substance and form that the contours of the formal architecture must sometimes be bent by legal fiction in order to implement these underlying principles.

Just as in medieval tort law, then, modern tort law’s organizing conceptual structures are sometimes “twisted and tortured to inappropriate uses” in order to meet the felt requirements of remedial justice — to enforce ordinary moral commitments about remedial moral liability that are widely recognized, often implicitly and somewhat inarticulately, by judges and other participants in the practice. The substantial overlap between relational moral wrongdoing and moral responsibility for rights infringement helps to occlude the artificiality of the legal fictions and doctrinal appendages that the law uses to enforce principles about the latter through a conceptual vocabulary that largely speaks of the former.

Unlike legal historians, contemporary private law theorists tend to pay scant attention to legal fictions, and when they do they often criticize the fictions under scrutiny as regrettable deviations from principled legal decisionmaking. In my view, however, the law’s use of fictional and quasi-fictional forms — devices that facilitate “legal inferences founded on imaginary facts, but intended to forward the substantial ends of justice” — tells us something quite important about its nature and functions. That the law must make use of such devices reflects the disconnect between the principles of justice that it recognizes and the conceptual structures that it uses to enforce them.

It is only when we examine the doctrine in granular casuistical detail that these underlying moral principles become apparent. Once we do, the Palsgraf perspective’s insistence that tort law’s moral core is robustly relational becomes untenable. True, there is a thin sense in which a defendant’s moral liability to compensate a plaintiff (and liability in tort) is “irreducibly relational”: The defendant must infringe some fact-relative moral right, a moral right against actually being injured, held by the plaintiff. But this is a very thin sense of relationality. It need not involve delivering any “[a]ffront to [the plaintiff’s] personality” or subjecting her to an excessive risk of damaging her body or property. The plaintiff might be unforeseeable to the defendant, or the defendant might treat her with all due care — because, say, she has to all reasonable appearances given consent to his action — and yet the defendant may be liable to compensate her, because he has responsibly risked infringing a certain sort of moral right, he did in fact infringe such a right, and the right in question did in fact belong to her.

All of this invites the question, of course, of why moral responsibility for rights infringement should ground remedial moral liability. As an interpretive matter, it is not obvious that any such explanation is required. After all, it is quite possible that the characteristic commitments of ordinary moral thought are in fact mistaken — that the true moral theory is some theory that is radically discontinuous from commonsense morality, such as act utilitarianism. Even if that were true, it is unlikely that tort law — a thoroughly human institution — would encode any such moral theory. If this Article’s arguments are sound, the law’s underlying moral commitments are more than a mishmash of particular convictions; they have a deeper structure that can be elicited through casuistry and cast in the form of general principles. But the principles are ones that are recognized, if only implicitly, in ordinary moral life.

Ultimately, however, ordinary moral principles are worth close attention — and the manner in which tort law encodes them is worthy of close examination — largely because these principles might track or at least substantially illuminate important moral truths. And, when we ask why it should be true that remedial moral liability is grounded in moral responsibility for rights infringement, the answer proves to be elusive.

An initially tempting answer appeals to the simple observation that moral rights are moral reasons. As a general matter, when an agent acts against a moral reason and thereby damages or sets back the value in which it is grounded, the agent can acquire a derivative moral reason to undo the damage or make an offsetting contribution to promoting the value in question. Thus an agent who pollutes a beautiful lake, even unforeseeably, will have a reason to clean it up (if she is able to) or perhaps to clean up other lakes and other polluted parts of the environment (if she is not). So, too, it is plausible to suppose, an agent who infringes another person’s right against bodily injury, even if she has done so unforeseeably, will acquire a remedial reason to help undo the injury (such as by financing remedial surgery) or otherwise promote the person’s welfare.

But this explanatory strategy is obviously incomplete, for such remedial reasons appear to be comparatively weak. An agent who unforeseeably burns down her neighbor’s home by flipping a kitchen switch may have a reason to undo the damage, but this reason is nowhere near as strong as the reason that the agent would have if she had burned down her neighbor’s home culpably and foreseeably. Culpably and foreseeably infringing a right (or otherwise causing an outcome that damages something of impersonal moral value) can ground an extensive and onerous remedial moral duty — the sort of remedial moral duty that the law can and should enforce with highly burdensome remedial measures, such as make-whole compensation.

The question is why such an agent’s remedial moral reasons are so much stronger than if she is not morally responsible, in any such way, for causing the outcome in question. It is not enough merely to suggest that the law should in fairness decline to enforce such a moral duty unless the agent has acquired the duty through some exercise of his responsible agency. Not only does that suggestion fail to explain the robust and basic conviction at stake — that no such onerous remedial duty exists even in morality — but it is also obscure why, if there did exist such a strong and enforceable remedial moral duty, it would be unfair for the law to enforce it. Similarly, it does not gain us much explanatory ground to suggest that the strength of an agent’s remedial reasons (and whether they ground a strong and enforceable remedial moral duty) reflect the extent to which the agent is the agent or author of the outcome in question. Without more, that suggestion simply restates the phenomenon that needs to be explained.

Perhaps the most promising explanation sounds in a line of thought once voiced by Justice Holmes, whose hostility to moralist accounts of tort law (and the commitments of ordinary moral thought) was decidedly inconstant. Holmes suggested that an agent cannot properly be held responsible for an outcome, in morality or in law, unless the agent could have avoided it by choosing differently, for otherwise the agent would lack control over incurring such responsibility. It is an attraction of this explanation that it might make sense of the intuitive moral phenomenon embedded in doctrines such as transferred intent and the culpability-sensitivity of proximate cause: Culpability can widen the scope of a wrongdoer’s liability beyond the boundaries of reasonable foreseeability. The complaint that one could not have avoided causing an outcome rings more hollow if one has caused it through an egregiously culpable and wrongful action, for such actions are easier to avoid than mildly or moderately culpable ones.

But ultimately even this line of thought comes perilously close to restating what it purports to explain. To see why, suppose that an agent culpably imposes a moderate risk of death on another person through some prosaic means, such as playing around with explosive fireworks. If the agent causes harm to the other person in a predictable way (such as, blowing him up), then she will of course be morally responsible for this outcome and liable to compensate him for it. If, by contrast, she causes him equivalent harm through some freakish and unpredictable way (such as awakening a slumbering and murderous bear), she will not. But there is an obvious sense in which the agent is equally able to avoid the injurious and rights-infringing outcome in both cases: She can simply avoid playing with the fireworks, something that she is anyway obligated to do.

From a certain perspective it seems clear, moreover, that this is the only sort of avoidability or control worth caring about. If a parent knows that her child is about to risk killing himself through some imprudent action, she might of course pay good money to reduce the magnitude of this risk. But it would be senseless for the parent to pay any money to ensure that, if her child kills himself through his carelessness, the lethal outcome will be of a foreseeable kind rather than an unforeseeable one. Perhaps a parent might sensibly pay good money to ensure that, if her child’s culpable wrongdoing is about to kill another person, the lethal outcome will be of an unforeseeable rather than a foreseeable kind. If that is so, however, it would seem to be because the unforeseeability of an outcome undercuts an agent’s moral responsibility for it, and it is in the child’s interest that he not be morally responsible for killing another person (in part, perhaps, because it is in his interest that he should not be morally or legally liable to compensate). And, of course, the connection between outcome foreseeability and moral responsibility is the very phenomenon that this Holmesian appeal to avoidability and control is supposed to ground. The explanatory power of this Holmesian account may, therefore, be illusory: Such an account may rely on defining notions such as avoidability and control in a covertly stipulative sense that bakes in the very moral distinctions that these notions are being invoked to explain.

Thus, it may prove difficult to explain and justify ordinary moral sentiments according to which the scope of moral responsibility for causing an outcome is presumptively determined and delimited by the boundaries of reasonable foreseeability. But the intuitive moral phenomenon here is a perfectly general one. It has nothing in particular to do with relational wrongdoing (or the infringement of rights against injury, for that matter). After all, outcome foreseeability seems equally to limit moral responsibility for damaging objects of impersonal value — in precisely the same, reflectively mysterious way. Here, as elsewhere, some of the most interesting and difficult philosophical questions about tort law come into clear view only if we look beneath the formal surface of the doctrine and its ostensible concern for relational duties and wrongs.