The Structure of Tort Law, Revisited: The Problem of Corporate Responsibility

1 Introduction

Legal theorists have offered many competing and complementary explanations of the normative principles that do and should underlie the law of torts, or civil wrongs. These include (1) deterrence of wrongful or costly behavior through incentives to take optimal levels of precaution against risks of injury, (2) insurance against concentrated losses, (3) fairness in the allocation of particular kinds of losses, and (4) responsibility, in the sense of treating individuals’ agential responsibility for injuring others as a reason to make the injurers bear the resulting costs, as a means of holding them accountable for their behavior. Since the advent of the law and economics movement, it has been extremely common for tort scholars to explain and justify tort law principally with reference to the goal of economically optimal deterrence—i.e., maximizing wealth, primarily by bringing about efficient levels of risk-taking. ^[1] As Jules Coleman has suggested, with evident unease: “we are [now] operating with a conception of tort law we have inherited from [law and economics scholar Guido] Calabresi, one in which tort law is a mere technology—a potential solution to a social problem and not a body of law rooted either in fundamental features of our agency or one that otherwise maps onto our moral lives or which reflects the normative structure of our relationships with one another.” ^[2] If Coleman is correct that the dominant approach to tort law today is to treat it as “a mere technology,” this is likely not just because of perceived successes of economic theories of tort law but also because of lingering doubts about how well tort law actually instantiates fairness and responsibility, and how morally meaningful are its conceptions of each. Economic approaches to tort law may seem especially powerful, and moralized accounts of tort law particularly impotent, in the many tort cases in which corporations are defendants held liable for defective products or vicariously liable for the torts of their employees. It is obvious that tort law may affect corporations’ incentives, but it is not self-evident that tort liability can be meaningfully understood as a form of moral accountability when it is imposed upon corporate rather than human persons.

In this article, I offer new defenses of the idea that tort law does embody distinctive interpersonal norms—in the first instance, norms of responsibility. First, I use recent work in the philosophy of responsibility to help clarify the role that morally significant forms of agential responsibility play in tort law. Second, I combine that analysis with recent work in the philosophy of group agency and responsibility in order to defend the idea that the basic rationale for applying tort law qua tort law—or tort law as distinct from other substitute institutions—extends to corporate persons as well as individual persons. Hence, the prevalence of corporate defendants in modern tort litigation should not incline us to adopt an economic approach to tort law that eschews a meaningful role for responsibility.

In Section 2 of the article, I begin with the observation, made by many tort theorists, that tort law has a distinctive structure: it unites in one procedure and one verdict the potentially distinct questions of what a plaintiff should receive in consequence of an injury and what a defendant should be made to bear for having caused it. Although this distinctive institutional structure severely limits the extent to which tort law can deliver deterrence, insurance, and fairness, it also builds upon and expresses ideas of responsibility that are plausibly understood as fundamental to how we understand ourselves and our lives. These ideas of responsibility are not obvious, however, because they differ in important ways from the kind of moral responsibility embodied in more obviously and heavily moralized practices, such as criminal law.

In Section 3, I introduce T. M. Scanlon’s distinction between that for which a person is attributively responsible (i.e., that for which a person may rightly be evaluated, morally speaking), and that for which a person is substantively responsible (i.e., what a person owes to other persons). Although tort liability is substantive responsibility, what makes tort law distinctive relative to other institutions that can deter, insure, and allocate losses with concern for fairness is that tort law grounds substantive responsibility for losses at least partly in attributive responsibility for causing those losses. Although attributive responsibility for acts and their outcomes is not a sufficient condition for tort liability, it does appear to be a necessary one. Tort liability gives a defendant the substantive responsibility for redressing an injury he is attributively responsible for causing another person. In Section 4, I explore the moral significance of the relation that individuals bear to the proximate consequences of their actions by revisiting Tony Honoré’s idea of “outcome responsibility” and the philosophical literature on moral luck, and by discussing Scanlon’s recent work on blame.

Although corporations may be eminently sensible targets of policies designed to deter unduly risky behavior and insure against or spread losses, it is not clear without argument that corporations can themselves be attributively responsible for acts and their harmful consequences in ways that would begin to justify imposing upon them substantive responsibility for those outcomes in tort law. The pervasiveness of corporate liability in tort law might therefore be taken to undermine the idea that tort law is, in the first instance, a morally meaningful, institutionalized practice of responsibility. Contrary to that view, I argue in Section 5 that with the aid of a novel conception of group agency and responsibility recently put forth by Christian List and Philip Pettit, we can see that tort law’s distinctive normativity remains plausibly applicable and important even in complex modern cases with corporate defendants.

2 The structure of tort law

The law of torts is a fundamental part of the curriculum of American law schools. Like other key substantive areas of law such as contracts and criminal law, tort law combines general principles with particular rules, forming a patchwork of more and less entrenched pieces. The rules and even the principles of tort law have varied somewhat across both time and place, and they will surely continue to change without ceasing to constitute the law of torts. Yet there is also enough of a stable core to tort law that it is possible to conceptualize tort law at a high level of generality.

Tort law is a legal framework within which individuals are empowered to seek legal redress—typically monetary compensation—for injuries and losses they have suffered because of actions taken by others. Tort law has a distinctive “structure” that philosophers of tort law have called “bipolar”: tort law unites plaintiffs with individuals who have injured them, making a defendant’s liability the exact corollary of a plaintiff’s recovery. ^[3] This structure substantially limits tort law’s domain, restricting the extent to which it can promote such goals as deterring wrongs and unduly risky conduct, insuring individuals against concentrated injuries and losses, and bringing about fairness in the allocation of risks, injuries, and losses (even those specific ones that are the consequences of human agency). ^[4]

3 Attributive and substantive responsibility

Before we address the substance of questions about responsibility and its role in tort law, it will be helpful to distinguish between two related but distinct ideas of responsibility. In his influential work What We Owe to Each Other, T. M. Scanlon distinguishes between what he calls “responsibility as attributability,” and “substantive responsibility.” ^[15] What is attributable to a person means what can serve as a legitimate basis upon which to assess that person, morally speaking. Scanlon suggests that only expressions of “judgment-sensitive attitudes” can rightly be attributed to us in the sense of serving as bases for evaluating us morally. ^[16] For him, negative moral evaluation of a person means judging that person to have “faulty self-governance.” ^[17] By contrast, a person’s substantive responsibility is a matter of what that person owes to others or should endure for them. Scanlon contends that substantive responsibility is grounded in what he calls the “Value of Choice,” which “explains the role of choice in the justification of moral principles by appealing to the reasons... we have for wanting outcomes to depend on the way we respond when presented with alternatives.” ^[18]

As Scanlon notes: “It is tempting to say that the answer to the question of when a person is responsible in these two senses is the same: an action is attributable to a person just in case he or she performed it voluntarily, and a person is substantively responsible for a certain outcome just in case he or she chose, voluntarily, to bring it about.” ^[19] Scanlon is right, however, to conclude that “it is important to distinguish clearly between judgments of responsibility of these two kinds.” ^[20] One of the reasons the distinction is important for Scanlon is that it helps him defend—against fairness-based objections—the idea that people can be morally responsible and criticizable even if they are incapable of recognizing and acting on the moral reasons that apply to them, and even if it would be unfair to make them bear the costs of the actions for which they are morally responsible. ^[21] In his recent work on blame, Scanlon makes a similar point: “I do not think that blame is undermined by the fact that a person had no control over the factors that made him the kind of person that he is, or by the fact that, given the kind of person he is, he is incapable of understanding the reasons against acting the way he does.” ^[22] In holding these views about moral responsibility and blame, Scanlon differs from a number of other prominent moral philosophers, including Susan Wolf and R. Jay Wallace. ^[23] He does so partly because he is emphatic that attributive responsibility does not necessarily entail substantive responsibility. In particular, he does not view attributive moral responsibility as a kind of sanction or obligation, such that people cannot rightly be held morally responsible unless they have a reasonable or fair opportunity to avoid it. ^[24]

Scanlon aligns himself closely with Gary Watson, who earlier put forth a distinction between “two faces of responsibility” that is very similar to Scanlon’s. ^[25] Watson frames the relevant contrast as one between “attributability” (what reflects on a person), and “accountability” (what demands can appropriately be placed on a person). Like Scanlon, Watson argues that attributability is a morally fundamental notion. ^[26] Watson places particular emphasis on the relationship between attributability and personal identity, contending that “[t]he significant relation between behavior and the ‘real self’ is not (just) causal but executive and expressive.” ^[27] And, like Scanlon, Watson sees concerns about the fairness of responsibility as pertaining not so much to attributability and identity as to accountability and “the imposition of demands on people.” ^[28]

Yet another writer who draws a very similar distinction is David Miller. Miller argues that individuals who bear various kinds of responsibility for bringing about certain costs or harms through exercises of their agency do not necessarily incur “remedial responsibility” in the sense of being the parties who ought, all things considered, to bear those costs or redress those harms. ^[29] Although Miller believes that responsibility in the first sense of agential attribution is an important consideration that can ground remedial responsibility, he recognizes that there are other considerations that may outweigh it in determining where remedial responsibility most properly lies. ^[30]

Tort liability is what Scanlon would call “substantive” responsibility and Miller would call “remedial” responsibility. In principle, this kind of responsibility might be based on considerations other than attributive responsibility. For instance, we might conclude that society owes it to indigent persons to supply them with social insurance against such ills as hunger, sickness, and homelessness, even when they are the attributive responsibility not of society as a whole but only of the free, self-destructive choices of those who would suffer from them. However, if the only idea of responsibility operative in tort law were a substantive or remedial responsibility unconnected to attributive responsibility for creating the harms to be redressed, then the idea of responsibility would play only a trivial role in the assignment of tort liability—it would be another label for it, not a ground of it. So when we ask whether a conception of responsibility plays an important role in justifying tort liability, we are asking whether the substantive or remedial responsibility associated with tort liability is grounded partly in tortfeasors’ attributive responsibility.

As we have seen, tort liability is not a free-floating form of substantive or remedial responsibility. Tort plaintiffs cannot recover from society at large or just anyone who happens to be well positioned to compensate them. Rather, only those who wrongfully injured them are in the pool of individuals who may potentially be held substantively responsible. Tort law determines whether someone is substantively liable by first asking whether she is attributively responsible for injuring the plaintiff. At the least, then, tort law implies that being attributively responsible for injuring another person makes one a better candidate to bear substantive responsibility for the resulting costs to the injured party. However, we have yet to clarify the moral significance of the kind of attributive responsibility presupposed by judgments of tort liability. To get a firmer grasp on this idea, it will be helpful to review and clarify Tony Honoré’s influential analysis of what he called “outcome responsibility”—a concept intended to capture the moral relationship that individuals bear to certain states of affairs they have brought about by their actions. ^[31]

4 Outcome responsibility: from attributive to substantive responsibility

For Honoré, “[o]utcome responsibility means being responsible for the good and harm we bring about by what we do,” and “[t]he main role of legal liability is to reinforce our basic outcome responsibility with formal sanctions such as compensation or punishment.” ^[32] Building on Honoré’s work, Stephen Perry has offered a detailed analysis of outcome responsibility and defended a particular conception of it, which he calls the “avoidability” conception. ^[33] According to that conception, an individual is responsible for a harmful outcome of her actions just in case she had the capacity and opportunity to foresee the potential harm and thereby act in such a way as to avoid creating it. ^[34]

Perry has contended that Honoré oscillates between “personhood” and “social” conceptions of outcome responsibility. ^[35] On the one hand, Honoré stresses how important and closely connected outcome responsibility is to our sense of ourselves as persons: “Outcome responsibility is, I contend, inescapable because it is the counterpart and at the same time a constituent of our personal identity and character. We could not dispense with outcome responsibility without ceasing to be persons.” ^[36] On the other hand, proceeding from the thought that “[c]hoosing is inescapably betting,” Honoré also conceptualizes outcome responsibility as a lottery and argues that in order for it to be justifiable, it must be “impartial, reciprocal and over a period, beneficial.” ^[37] And Honoré claims that, in fact, for those who meet the threshold level of capacity necessary to be outcome responsible in general, the system of outcome responsibility will generally bring more benefits than costs. ^[38] Justifiably, Perry replies that Honoré’s benefits/burdens story is either implausible or trivially true: “Unless minimum capacity is simply defined as not standing to profit ex ante from the system of outcome allocation—and Honoré comes close to doing this in a footnote—the claim that the system is beneficial to all seems to be, empirically, simply false.” ^[39]

Perry rightly observes that there are two distinct formulations of outcome responsibility here and it is not immediately clear whether or how they relate to one another. ^[40] I suggest we can better understand the two ideas and their relation by deploying Scanlon’s distinction between attributive and substantive responsibility. The “personhood” sense of outcome responsibility is akin to attributive responsibility in that it is a ground for evaluating individuals—albeit on the basis of their impact on the world, rather than simply their attitudes of concern for others. The “social” sense of outcome responsibility, by contrast, is best thought of as a form of substantive responsibility consisting of social and legal rules that create potential liability in moral practice and tort law for those who are attributively responsible for actions proximately causing costly outcomes.

Recall that Scanlon contends it is substantive responsibility rather than attributive responsibility that plausibly gives rise to fairness concerns about individuals’ capacity to avoid responsibility. Someone who negligently burns down his neighbor’s house is outcome responsible for destroying it in the attributive, “personhood” sense. It meaningfully attaches to him as an agent in the sense that it morally licenses certain judgments, attitudes, and actions toward him that might otherwise be inappropriate. But if the person who caused the fire acted with only minimally culpable ignorance of the risks—say, because he is a functioning adult with very low intelligence—it is not implausible to think that it is unfair to hold him substantively responsible by making him pay all the costs of rebuilding the home. ^[41] In this limited sense, he is the “winner” of a lottery in which nature gave him extra tickets. Honoré is right to be concerned about the fairness of outcome responsibility as substantive responsibility. But because he does not distinguish clearly between attributive and substantive responsibility, that concern leads him to a misguided attempt to show that outcome responsibility as such is, or should be, analogous to a fair lottery.

To take stock, I have suggested that what makes tort law distinct from other institutions that could be combined to fulfill many of its instrumental goals is the way in which it grounds substantive responsibility for an injury at least partly in attributive responsibility for being the agent of that injury. As scholars such as Honoré, Perry, and Miller have implied, using different language than Scanlon, something like attributive responsibility for outcomes is a coherent idea, and one that is plausibly of moral significance. Although attributive responsibility for an action is not in and of itself enough to ground substantive responsibility for its outcomes, ^[42] it does appear to be a necessary condition for tort liability, as well as the conception of responsibility with which the bipolar structure of tort law stands or falls. This is presumably what Perry means when he says that outcome responsibility is the “normative link between particular injurers and their victims.” ^[43]

This may be enough to show that a conception of responsibility lies beneath the structure of tort law. It is not yet enough, however, to show that this conception of responsibility actually does substantial work to justify tort law’s bipolar structure. When and why might attributive responsibility for an action even partially legitimate or justify making an individual substantively responsible for bearing the action’s specific costly consequences? It is helpful here to begin by distinguishing two different questions that are often run together. First, what affirmative aims are served by making a defendant who is attributively responsible for injuring a plaintiff bear substantive responsibility for the costs of that injury? Second, why is it permissible to pursue those aims by enabling the plaintiff to demand compensation from such a defendant? ^[44]

For at least three reasons, it is tempting to focus on the role that attributive responsibility likely plays (at least in conjunction with fault or some other condition) in making it morally permissible (rather than morally good or obligatory) that a defendant who is attributively outcome responsible for injuring a plaintiff should be liable for it. First, although the state (acting at various stages through police, prosecutors, judges, and juries) exercises considerable discretion in determining whether and how far to pursue criminal actions, political morality constrains the appropriate exercise of that discretion in ways that seem not to apply to private parties who bring tort suits. In tort law, the plaintiff with a winning tort case has the option to sue, but she also has wide moral discretion to forgo her claims if she wishes. ^[45] That a plaintiff has a winning case does not mean it would necessarily be morally good for her to stand on her rights and pursue it, or that it is morally obligatory or even good for the defendant to give her the compensation she could legally demand. Second, in part because the extent of actual damages suffered can vary significantly from the degree of wrongfulness of tortious conduct and the culpability of its perpetrator, it is doubtful that even those who wrongfully injure others necessarily, or even typically, deserve to bear the full costs of their conduct. It seems more plausible to suppose that whether or not they deserve to bear those costs, through their wrongful conduct they have diminished or lost their claim against having to bear them.

Third, there is a very plausible way of explaining why attributive responsibility for an injury, especially the wrongful infliction of injury, should have the moral effect of making it permissible (even if not yet morally good or obligatory) for a person to bear the costs of the injury. Recall that Perry defends an “avoidability” conception of outcome responsibility according to which a person is responsible for an outcome just in case it was one that she could have foreseen and avoided by acting on her foresight. We may better understand part of the case for grounding substantive tort liability in attributive outcome responsibility (conceived as applying to outcomes avoidable through foresight) if we recall Scanlon’s plausible principle that the substantive burdens a person can fairly be required to bear depend upon the choices she has with respect to those burdens. As Scanlon observes, having good choices or opportunities to avoid bearing a particular cost gives a person an important means of protecting herself against having to do so; hence, all else equal, a person has less of a complaint against having to bear a cost if she had a good opportunity to avoid bearing it than if she had little or no such opportunity. ^[46] Moreover, although it may be difficult or impossible to avoid setting back others’ interests over the course of one’s life (because people’s interests conflict and accidents happen), people have especially good opportunities to avoid wrongfully harming others, because doing so does not require them to refrain from anything other than acting wrongfully toward others.

To recap, the imagined argument runs as follows: attributive responsibility for causing a foreseeable type of injury implies an opportunity to avoid bringing about that injury, and therefore an opportunity to avoid having to bear the costs of it, which in turn diminishes a person’s complaint against having to bear those costs through tort liability. The argument is plausible on its own terms, but not as an explanation of tort law’s presupposition that attributive responsibility in particular helps ground substantive outcome responsibility. The problem is that conditioning substantive responsibility for bearing the costs of an injury on attributive responsibility for having caused that injury is not the only way to ensure that whoever bears those costs had a reasonable opportunity to avoid them. To see this, notice that were we to replace tort law with compensation for wrongful harms dispensed from an at-fault pool, the individuals who would have to contribute to the pool in proportion to their fault would also have a good opportunity to avoid having to contribute to the pool. In fact, their opportunity would be in an important respect better than the opportunity that people currently have to avoid tort liability, because a system of liability to an at-fault pool would, ex hypothesi, neutralize the effect of outcome luck, thereby giving people a means of avoiding liability more predictably tied to their choices than the ‘lottery’ of tort liability.

More generally, if the grounding relation between attributive and substantive outcome responsibility were merely that the former contributes to the permissibility of the latter, then we would have no reason to believe that the functions of tort law could not be handed over piecemeal, without moral loss, to morally permissible substitute institutions for dealing with the costs of interpersonal wrongs and harms—institutions without tort law’s bipolar structure. ^[47] To be sure, attributive outcome responsibility may well play an important role in the explanation of why it is permissible to impose tort liability upon defendants for certain harms they have proximately caused plaintiffs. However, to show that there is something normatively distinctive about tort law that gives us at least pro tanto reason to retain it rather than replace it with an amalgam of permissible alternative institutions, we need to answer the question of why attributive outcome responsibility might not only make it permissible but actually give us affirmative reason to single out an agent who has (perhaps wrongfully) injured another and make him liable to compensate that other person.

I will not attempt a systematic answer to that question here, but instead only gesture in the direction of where I believe one may ultimately be found. I will do so by briefly revisiting the idea of “moral luck” that Bernard Williams and Thomas Nagel influentially formulated, ^[48] which has often been mentioned in connection with tort law, and by discussing the new conception of blame that Scanlon has recently developed. ^[49]

Although a prominent strand in ordinary and academic moral theory seeks to cordon off luck from our moral assessments of one another, Williams, Nagel, and many others after them have resisted this impulse. In their view, moral judgments are inevitably infused with the influence of luck, so much so that morality would look far different were we to succeed in eradicating that influence. Nagel defines instances of moral luck as situations “[w]here a significant aspect of what someone does depends on factors beyond his control, yet we continue to treat him in that respect as an object of moral judgment.” ^[50] Moreover, one of the four basic types of moral luck he identifies is “luck in the way one’s actions and projects turn out.” ^[51] In other words, Nagel affirms precisely the idea that being attributively responsible for an action can connect a person to the consequences of that action in a morally meaningful way.

Following Williams and Nagel, Scanlon has taken “the phenomenon of moral outcome luck” to be a datum that his account of blame has the putative virtue of explaining. ^[52] In Scanlon’s view, blameworthiness means the presence of “something about the agent’s attitudes toward others that impairs the relations that others can have with him or her.” ^[53] To actually blame a person means not just to appraise her as blameworthy but also to see reason to alter one’s relation to her accordingly. ^[54] Scanlon argues that in contrast to the “‘character assessment’ account of blame,” his relational impairment view makes good sense of the role luck plays in our blaming practices. ^[55] Chance variation in the outcome of a person’s action does not reflect on her character—understood as something like an amalgam of her beliefs, desires, values, and dispositions. ^[56] Yet Scanlon observes that it can affect blameworthiness and blame conceived on the relational impairment model. For according to that approach “blame is not a mere evaluation but a revised understanding of our relations with a person, given what he or she has done. Blame is therefore a function not only of the gravity of a person’s faults but also of their significance for the agent’s relations with the person who is doing the blaming.” ^[57]

Suppose that, of two equally negligent drivers, one kills a child and the other just misses doing so. As Scanlon observes, drawing the example from Nagel, the difference in outcomes indicates nothing about the character of the drivers. ^[58] Yet it is more than plausible to think that because the first driver has committed negligent homicide and the second has not, a host of people (family members, friends, community members) will have normative reasons to change some of their expectations about and dispositions toward the first driver in particular, and the first driver herself will have distinctive reasons for action that the second driver will not. ^[59] Williams uses the case of a non-negligent vehicular accident to emphasize both the salience and scope of what he calls “agent-regret”: a distinctively “first-personal” perspective upon one’s past behavior and its consequences. ^[60]

As Stephen Perry notes, even if a person’s outcome responsibility for an injury is not negligent or otherwise sufficient to create tort liability, it still may be a reason for her to do things such as “express [her] regret to the person harmed.” ^[61] Scanlon gives a helpful label to an idea he draws from Nagel when Scanlon calls such situations ones of “objective stigma” as distinct from truly moral luck. He contends that objective stigma resembles blame in involving “a modification of one’s relations with a person, a change in the meaning of one’s interactions.” ^[62] But objective stigma is distinct from blame in that unlike blame, objective stigma can arise from wholly faultless behavior.

The outcomes of our actions do not directly reflect our characters. Yet as Perry has noted, drawing on Honoré: “It is the outcomes of actions that give us a history as persons, and our history as persons contributes in very large measure to our identities as individuals.” ^[63] And as the distinction between moral outcome luck and objective stigma suggests, outcomes are likely to be particularly central to an agent’s identity when they are outcomes of actions that reflect poorly on the agent’s character. Some consequences will be too remote, of course, to ascribe to a person’s agency in any meaningful sense. That may help partially explain why in tort law individuals are generally only liable for types of injury that were reasonably foreseeable, and why Perry makes “avoidability” a necessary component of outcome responsibility. Yet what the moral luck literature and Scanlon’s conception of blame indicate is that outcomes of actions can reflect on agents in a way that changes how they have reason to comport themselves and how others have reasons to treat them. I suspect that much more can and should be said to explain and justify that claim. However, I also believe the discussions of Williams, Nagel, Honoré, Perry, and Scanlon, among others, are enough to give the idea substantial plausibility.

I have not attempted here fully to defend the idea that attributive responsibility for an injurious action can in fact give us affirmative reasons to make its agent substantively liable to bear its costs. What I hope to have shown is only that (a) tort law presupposes that people who are attributively responsible for foreseeably injuring others have special reasons to bear the costs of those injuries (at least or especially when they have caused them wrongfully), and (b) this implicit inference is at least plausibly justified and morally significant. In the following section, I will argue that the prominence of corporate defendants in tort law is perfectly compatible with—and therefore does not undercut the central role tort law accords to—the idea that a tort defendant must have been attributively responsible for a plaintiff’s injury in order to be a candidate for liability.

5 Corporate responsibility in tort law

Even if tort law in principle could embody special interpersonal norms of responsibility, the contemporary corporate form may routinely undermine, or at least substantially complicate, its ability to do so. The concern is that we may not stand in the same kinds of interpersonal relations with corporate persons as we do with human persons. It is not clear, in the absence of argument, whether corporations are or can be the kinds of moral agents to whom notions like attributive responsibility apply. Even if one accepts that a person’s being attributively responsible for injuring another activates affirmative reasons for her to bear substantive responsibility for redressing the injury, there remains the question of whether corporations qua corporations can be attributively responsible in a way that would begin to ground substantive responsibility. To draw that conclusion, we need a theory of how corporations can be responsible for actions and outcomes in a morally meaningful way, not just amenable to manipulation for social ends.

In recent work, Christian List and Philip Pettit have argued that corporations themselves, and not just the individuals who collectively constitute them, can be agents—not just in a loose sense but also in a very rigorous sense with important implications for how we should understand and treat them. ^[64] They have done so by building a theory of group agency that attempts to show how a group can be more than a straightforward sum of its individual constituent decisionmakers and thus in an important sense an “agent in its own right.” ^[65] On his own and in his work with List, Pettit has extended the theory of group and corporate agency in an effort to show that an appropriately structured group or corporation can also be responsible qua corporation, not just in a causal sense, but in a moral sense as well. ^[66] In this section, I argue that the kind of moral responsibility that Pettit shows is possible for corporations is attributive responsibility. Hence, corporations may be attributively responsible in a way that partially grounds tort liability—not just “responsibilizable” entities fit to be held substantively responsible as though they were attributively responsible. ^[67]

The account of group agency that List and Pettit offer begins with their analysis of what they call the discursive dilemma. The discursive dilemma is a problem that arises when a group of people aggregates the individual judgments of its members about interdependent issues. List and Pettit credit the seed of their idea to Lewis Kornhauser and Lawrence Sager, who identified an instance of the phenomenon in the context of judicial decisionmaking. ^[68] As Kornhauser and Sager observed, there are many legal cases in which judgments about several subsidiary issues combine to yield an overall judgment in favor of or against a particular party. In tort law, for instance, to prove that the defendant committed the tort of negligence, a plaintiff must show that the defendant owed a duty of care to the plaintiff, breached that duty, and in breaching that duty proximately caused harm to the plaintiff. ^[69] Imagine a panel of three judges considering those four issues—(1) duty, (2) breach, and (3) proximate causation of (4) harm. Suppose that each element were found to be met by two out of three judges yet no single judge were to find all four elements met and therefore no single judge were to find liability. If the judges were to apply a majority criterion to each individual element of negligence, then look to see whether the group as a whole affirmed the conjunction of the four elements, they would come to a different result than if they were to apply the majority criterion only to each judge’s overall conclusion on negligence.

The Kornhauser and Sager observation, which they termed the “doctrinal paradox,” was valuable in its own right, illuminating a problem for judicial decisionmaking. What List and Pettit have shown, however, is a more general result: such potential inconsistency in judgment aggregation is not unique to legal or majoritarian decision procedures but extends across a broad range of functions that satisfy important criteria for judgment aggregation. ^[70] We can set aside many of the technical details regarding the range and nature of the conflict. For our purposes, it should suffice to describe the dilemma in general terms—as between what Pettit has helpfully characterized as “individual responsiveness” and “collective rationality.” ^[71] The potential conflict between these values that group decisions can respect or violate can occur both synchronically and diachronically, but it is easier to appreciate if we consider the more salient diachronic case. ^[72] Part of what it means to be able to co-reason with a person or group is to be able to draw out the logical entailments of the person’s or group’s prior commitments. ^[73] The diachronic aspect of the conflict between individual responsiveness and collective rationality consists in the following problem: if a group is to be responsive to aggregations of individual judgments on new issues, there always exists the possibility that such responsiveness will conflict with the logical entailments of the group’s prior judgments about logically related issues.

List and Pettit contend that groups with a collective purpose typically will, and should, prioritize collective rationality by adopting self-conscious aggregation procedures that would prevent, or force a group to revise, any group judgments in conflict with collective rationality—even when doing so would sacrifice individual responsiveness by making the group’s judgments conflict with the group members’ judgments about some specific issues. ^[74] For instance, giving up the demand for individual responsiveness makes it possible to adopt “[a]n aggregation function such as a premise-based or sequential priority procedure [which] enables a group not only to form rational intentional attitudes but also to do so in a way that collectivizes reason.” ^[75] List and Pettit argue that collectivizing reason “establish[es] a certain kind of autonomy for group agents,” ^[76] and “groups that collectivize reason deserve ontological recognition as intentional and personal subjects.” ^[77] Groups that collectivize reason become not mere “aggregates” but “integrates,” to use Pettit’s language. ^[78] This gives them a distinctive intentional capacity to form “beliefs” and “desires” that are their own, in the sense that they cannot be read off straightforwardly from simple issue-by-issue aggregation of the group members’ beliefs and desires. ^[79]

It is not just theoretically possible for groups to collectivize reason. As Pettit observes, “a purposive collectivity” “will be under enormous pressure to collectivize reason” when, as is bound to happen, discursive dilemmas arise. ^[80] Like individuals, purposive groups attempt to co-reason with other individuals and groups. But as noted already, persons and entities can only co-reason with one another insofar as they are able to draw out the logical implications of one another’s past commitments. As Pettit puts it: “any such purposive collectivity must avoid automatic recourse to the revision of past commitments; it must show that those commitments are sufficiently robust for us to be able to expect that the group will frequently be guided by them in its future judgments.” ^[81]

Imagine an integrated group that to a substantial extent holds itself to its prior beliefs, desires, intentions, and actions—and anticipates others will hold it to them as well. Such an entity already applies to itself something like a conception of responsibility. So it should not be surprising that Pettit has used the forgoing account of integrated groups to argue that it is possible for a group agent to incur moral responsibility of its own—moral responsibility that is distinct from the moral responsibility individual members of the group bear for creating the group, being part of the group, and carrying out its actions. ^[82]

Pettit argues for group moral responsibility by setting out three conditions that must be true of an agent and then arguing that those three conditions can be simultaneously satisfied by groups that have integrated or collectivized reason in the manner just discussed. ^[83] In his view, moral responsibility applies if (1) there is a morally significant choice, and the agent making it has the capacity (2) to recognize and (3) to act on, the sorts of normative reasons that apply to the choice. ^[84]

As entities of collectivized reasoning, corporations can face morally significant choices just like individuals and they can at least in principle recognize normative reasons in their integrated capacity. The more difficult question is how a group can act in its own right, much less on particular normative reasons. As Pettit observes, a common objection to the idea of groups acting in their own right proceeds by observing that groups necessarily act only through their members and then asserting that “[o]ne and the same action cannot be subject both to the reason-sensitive control of the group agent and to the reason-sensitive control of one or more individuals.” ^[85]

To respond to this objection, Pettit appeals to a more general theory of tiers of explanation that he has developed with Frank Jackson. The basic idea, which they have called “program explanation,” is that relatively coarse-grained causal explanations “program” for an outcome they explain, in the sense that they ensure or make probable the presence of one or another set of fine-grained factors that will be sufficient to bring about the relevant result. ^[86] Although it is individual members of a corporation who carry out its actions, they do so within the confines of corporate rules, procedures, and decisions designed to ensure that such members will act in particular ways or within particular bounds. What corporations qua corporations are responsible for, then, is programming for particular actions of their members. ^[87] Although the corporation through its rules, procedures, and collective decisions does not guarantee that any person will act in any particular way, it can and does regulate what individuals can do, and are likely to do, in the corporation’s name. A corporation can often all but ensure that even if not employee X specifically, someone will do the corporation’s bidding.

In formulating his view of corporate responsibility, Pettit has distinguished “thinking” someone responsible from “holding” someone responsible, and both of these from holding someone “accountable.” ^[88] Although he does not employ Scanlon’s terminology, it is evident that Pettit understands “thinking” a person responsible and “holding” a person responsible to entail or express judgments of attributive responsibility. On the other hand, he views holding a person accountable in the way Miller views assigning “remedial responsibility”—as not requiring grounding in attributive responsibility. Pettit believes there are reasons for “holding an agent accountable” or “identifying that person as the one who carries the can, the one who sits at the desk where the buck stops” quite apart from whether she is attributively responsible. ^[89] One such reason is the conditioning or “responsibilization” of not yet fully responsible agents. ^[90]

It if were not already clear from the foregoing, Pettit goes even further to make it clear that his account of corporate responsibility is, in the first instance, only an account of corporate responsibility in the attributive sense. He does so by explicitly noting that he offers no assessment of “the practical considerations that are relevant to determining the best sorts of sanctions to impose in criminal law or indeed in the law of tort, the law of contract, or any other branch of jurisprudence.” ^[91]

However, as I have argued in the previous section, tort law’s distinctive structure presupposes that if an individual is attributively responsible for an act that injures another, then that attributive responsibility is an important consideration that helps license or justify holding her substantively responsible by shifting those losses to her. In that sense, at least by tort law’s internal standards, Pettit’s defense of corporate attributive responsibility is highly relevant to the justifiability of corporate tort liability. In fact, it is precisely the kind of account of corporate responsibility that is necessary to show that corporate tort liability is consistent with the underlying normativity of the bipolar structure of tort law.

Suppose that a vertically integrated chemical manufacturing company has mistakenly manufactured and distributed a cleaning fluid with the presence of dangerous, unlabeled toxins. Further assume that the company has done so not because specific employees acted maliciously or otherwise outside of the scope of their employment but rather because they acted within the bounds of organizational structures surrounding the coordination of different parts of the company’s chemical production and waste removal operations. Individual employees of the company collectively put a defective product on the market. And specific individuals bear attributive responsibility for putting in place the organizational structure of the company, perhaps years ago. But there is also a meaningful sense in which the corporation itself has “programmed” for the distribution of a defective product. By “programming” its employees to carry out specific tasks within its organizational structures, the corporation qua corporation pursued the aim of profit maximization in a way that led its individual employees collectively to bring a defective product to the market on its behalf. The company’s organizational structures and their effect on employees’ actions made the company, as such, an abstract entity engaged in practical reasoning not straightforwardly reducible to the practical reasoning of its individual employees. And the result of its pursuit of its aims was the distribution of a defective product.

There is a question regarding which if any specific tortious actions by employees should be attributable to the corporation in such a way as to bring about vicarious liability in the corporation. (Though I will not develop the thought here, my sense is that the idea that a corporation “programs” for actions of its employees corresponds nicely to ideas of corporate control underlying the theory of vicarious liability). There is also a question of products liability law regarding the standard of care to which we should hold the corporate agent, qua programmer of employees’ actions that collectively bring about an injury, as through the manufacture and distribution of a defective product. I have described the case of the chemical manufacturer in such a way as to leave open whether the corporation’s organizational structures programmed so unreasonably for the production and distribution of a defective product that we should say the corporation qua corporation was negligent. Depending upon the standard of care to which we want to hold the corporation, we might make it liable only if it can be shown that there was negligence by its employees or in its organizational structures, or, instead, we might make it liable for all the defective products it manufacturers and distributes. The point I hope to illustrate is simply the general one that a corporation can be attributively outcome responsible for a plaintiff’s injury. Hence, corporate liability in tort law is consistent with the idea that substantive tort liability is grounded in part in a defendant’s attributive responsibility for causing a plaintiff’s injury.

There is one lingering problem to address. Pettit’s account of corporate responsibility does not necessarily extend to any particular corporations, such as American business corporations. As already noted, various moral philosophers, including Pettit, believe that in order to be morally responsible, an agent must be capable of recognizing and adhering to the moral reasons that apply to its choices. Yet arguably many American business corporations are systematically structured so as not to recognize and act on specifically moral reasons. That is to say, arguably they are structured in such a way as to ensure only that their employees reliably act on their behalf in profit-maximizing ways, and that their employees act in morally responsible ways only to the extent that doing so coincides with the profit maximization imperative.

As Pettit suggests, one reason we might hold corporations that are not (fully) morally responsible to a standard that does not strictly speaking apply is that by doing so, we will inculcate in them the practice of responsible agency and give them an incentive to restructure themselves in such a way that they will be capable of adhering to the moral reasons that apply to their decisions. ^[92] However, this “developmental rationale,” as Pettit puts it, ^[93] is an argument for substantive responsibility grounded in considerations other than attributive responsibility. What we have been searching for is an account of how corporations can incur attributive responsibility of the kind that substantive liability in tort law seems to presuppose.

Fortunately, the developmental rationale is not our only recourse here, for two reasons. First, we must be careful to distinguish between corporations’ capacity to appreciate and act on moral reasons, and their disposition to do so. Like individuals, corporations may be poorly disposed to follow the normative reasons that apply to them, but this does not necessarily mean they lack the capacity to do. Although the profit maximization imperative often disposes corporations not to act on moral reasons, corporations also often have broad charters as well as procedures by which to restructure themselves. So we should not be too quick to assume that real-world business corporations lack the relevant moral capacities necessary for moral responsibility.

Second, even when corporations do lack the capacity to recognize and act on specifically moral reasons, it is not clear that this disqualifies them from the kind of attributive responsibility evidently presupposed by tort liability. Recall that as against some philosophers, Scanlon contends that if an agent is through no fault of his own “unable to see the force of morally relevant reasons,” this is not alone enough to think that he cannot be attributively responsible in a morally significant sense. ^[94] As Scanlon puts it: “If he commits [terrible] crimes because he does not place any value on other people’s lives or interests, what clearer grounds could one have for saying that he is a bad person and behaves wrongly?” ^[95] In Scanlon’s view, this immoral person is “the real agent” and his inability to avoid being that agent raises fairness concerns only where substantive responsibility is concerned—not where mere ascriptions of attributive responsibility are at stake. ^[96]

Even if a corporation is not responsible for having structured itself so as not to recognize moral reasons, its failure to act on those reasons still reflects on its character, and its actions still have consequences in the world that can be rightly ascribed to it. Likewise, whether it is unfair to hold a corporate person substantively responsible in tort law is not the issue here—that is a separate concern that applies equally to human persons. The issue for our immediate purposes is simply whether a corporation can be attributively responsible for its actions and thereby connected in a morally meaningful way to their consequences. As to that question, arguably the answer is yes and fairness is irrelevant. There may, of course, be good reasons of fairness to be concerned about actual corporate tort liability. For instance, corporations can and sometimes do pass on their liabilities to their constituent parts in ways that undermine norms of fairness. But as we already observed, we can simply add these fairness problems to a long list of plausible fairness objections to tort law in general.

Like persons, groups and corporations have identities and histories as agents. The consequences of their actions sometimes give us reasons to evaluative them differently, bolstering or impairing the kinds of relations that it is reasonable for us to have with them, and appropriately changing our expectations and dispositions toward them. Corporate liability is therefore consistent with the distinctive normative structure of responsibility that tort law reflects.

6 Conclusion

Tort law is a bipolar framework that brings together plaintiffs and defendants, liability for injuring and compensation for injury. While that structure limits its ability to pursue a variety of goals, it reflects—and enables tort law to give legal expression to—what is arguably a morally significant relationship that individuals bear to certain consequences of their actions. The bipolar structure of tort law implies that being responsible for injuring another reflects on an individual in a way that can sometimes give us at least pro tanto reasons to make her liable to bear the costs of the injury.

In this article, I have argued that although the conception of responsibility in tort law is distinct from the kind of attitudinal, will-focused notions of responsibility that infuse our thinking about criminal law, it is arguably morally significant in its own right. Moreover, I have argued that the prevalence of corporate defendants in modern tort practice does not undermine the role of responsibility in tort law, because the conception of responsibility upon which the structure of tort law rests can apply to corporations qua corporations, not just individual persons. What I have not done is fully defend the morality of the bipolar structure of tort law—i.e., the way tort law treats a person’s attributive responsibility for having injured another as sometimes supplying a good reason for her, rather than others, to bear the costs of the injury.