I am grateful to Benjamin Zipursky, Peter Vallentyne, Malcolm Thorburn, Hamish Stewart, Sophia Reibetanz Moreau, Martin O'Neill, John Norris, Christopher Morris, Dennis Klimchuk, Robert Gibbs, David Copp, Richard Bronaugh, and Donald Ainslie for helpful comments. Versions of this paper were presented at the University of North Carolina Legal Philosophy Conference, in October 2000, where Tom Christiano gave helpful comments and at Harvard University, New York University, and the University of Western Ontario. I am grateful to those present for their questions and comments.

1. Watson, Gary, “Two Faces of Responsibility” (1996) 24 Phil. Topics 227 Google Scholar. In adopting Watson’s distinction, I do not mean to commit myself to his application of it. In particular, Watson supposes that attribution is most at home in contexts in which a person is deciding how to live his or her life, while accountability is a matter of social practices of holding people responsible. I construe attributability more broadly, covering all of the senses in which an act might be thought to be “one’s own.” In this my use is similar to T.M. Scanlon’s distinction between what he calls “attributability” and what he calls “substantive responsibility”. Scanlon, T.M. draws a similar distinction between attributability and what he calls “ substantive responsibility” in What we Owe to Each Other (Cambridge, MA: Harvard University Press, 1998) at 248.Google Scholar

2. For example, Bernard Williams remarks that a responsible agent holds himself to a higher standard than others may rightly hold him to. See Williams, , “Voluntary Acts and Responsible Agents” in Making Sense of Humanity and other Philosophical Papers, 1982-1993 (Cambridge: Cambridge University Press, 1995) 22 at 32CrossRefGoogle Scholar. Again, Watson’s own use of the distinction suggest that attributability is important to individual identity, and so broader than moral accountability. But nothing either Watson or Williams says precludes the idea, prominent in the literature, that accountability is governed by some refinement of the concept of attribution.

3. The example is from Salmond, Sir John, The Law of Torts: A Treatise on the English Law of Liability for Civil Injuries, 6th ed. (London: Sweet & Maxwell, 1924) at 153 Google Scholar-55.

4. Nozick, Robert, Anarchy, State, and Utopia (New York: Basic Books, 1974) at 331.Google Scholar

5. See, for example, Kronman, Anthony, “Contract Law and Distributive Justice” (1980) 89 Yale L. J. 472 Google Scholar. Dworkin, Ronald, Sovereign Virtue: The Theory and Practice of Equality (Cambridge MA: Harvard University Press, 2000).Google Scholar

6. For example, Rawls acknowledges the role of “rules that govern the transactions and agreements between individuals and associations (the law of contract, and so on).” He goes on suggest that they be “framed to leave individuals and associations free to act effectively in pursuit of their ends and without excessive constraints.” Political Liberalism (New York: Columbia University Press, 1993) at 268. I discuss Rawls’s treatment of responsibility in “The Division of Responsibility and the Law of Tort” forthcoming in (April, 2004) Fordham Law Review. Dworkin subordinates private law to distributive justice in a different way, suggesting that “a regulatory constraint, or an article of tort law, is justified under the principle of correction only if there are good grounds for supposing that the corruption of the opportunity-cost test would be less with the constraint in place than it would be without it.” Dworkin, ibid at 157.

7. Ronald Dworkin’s influential work on equality and responsibility combines elements of both conceptions. In his seminal articles on equality of resources, reprinted in his Sovereign Virtue, Dworkin adopts a version of the reciprocity conception when he articulates the relation between individual responsibility and social equality in terms of the idea that people should bear the costs that their choices impose upon others, and again when he criticizes what he calls “starting gate” theories of justice. However, Dworkin’s arguments also include aspects of a version of the agency conception. That conception is at work, it seems to me, in Dworkin’s discussion of whether an agent’s preferences are genuine. He suggests that a preference is genuine if the agent whose preference it is identifies with it. A slightly different version of the agency conception seems to me to be at work when Dworkin introduces his distinction between what he calls “brute luck” and “option luck,” in terms of the idea that something is matter of option luck if it is the result of a choice that a person “might have declined.” Both the identification test and the idea of option luck, are monadic—that is, dependent solely on facts about the agent—in a way that reveals them to be specific versions of the agency conception. The idea that people should bear the costs that their choices impose upon others, by contrast is relational, in a way that suggests that it is a version of the reciprocity conception.

8. This puts my view squarely in the camp that G.A. Cohen has recently criticized. See If You’re an Egalitarian, How Come You’re So Rich? (Cambridge, MA: Harvard University Press, 2000) I don’t disagree with Cohen’s claim that the justice of a society is in part a matter of its ethos as well as its coercive structure. But I think he is wrong to conclude that that important observation undermines the idea that the coercive structure itself requires a special kind of justification. In a society with an egalitarian ethos, different coercive structures might be needed than in one that lacked one. But that does not change the fact that the coercive structure poses special problems about justification.

9. For a more general elaboration of this theme, see Dan-Cohen, Meir, “Responsibility and the Boundaries of The Self” (1992) 105 Harv. L. Rev. 959.Google Scholar

10. Kant makes this point in the opening sections of the Doctrine of Right, the first part of the The Metaphysics of Morals, in Gregor, Mary, trans. and ed., Immanuel Kant: Practical Philosophy (Cambridge: Cambridge University Press, 1996) 353 at 386ff.Google Scholar

11. I develop these Kantian themes in more detail in “Authority and Coercion” (2004) 32 Phil. & Pub. Aff. 2.

12. Is the reciprocity conception really Kant’s view of responsibility? In several places Kant appears to endorse something close to a different view, which I discuss below, since he says “The good results of a meritorious action, like the bad results of a wrongful action, can be imputed to the subject. “ (Metaphysics of Morals, supra note 10 at 382). But the quoted passage comes in the introduction to the Metaphysics of Morals, and is appealed to explicitly in the doctrine of virtue, for example, when Kant says that a servant who lies on behalf of his master will have the master’s wrongful deeds imputed to him “in conscience.” It is not clear that it is meant to apply to issues of justice. Moreover, I think the rest of Kant’s legal philosophy makes it clear that what he regards as the “bad” consequences of a wrong are to be understood in terms of that wrong, rather than simply in terms of natural causation and its effects on persons. This is particularly clear, I think, given his tripartite division of the ways in which something can be “ones own.” One can have a right to a thing, as against all other persons. One can have a right to another’s performance, as against that person only, and one can have a right to another person that extends to unspecified acts, and gives one a claim to that person has against all other persons. The differences between the first and the second ways in which a thing can be “one’s own” illustrate the role of the reciprocity conception. If two people make a contract, each of them has a right to the other’s performance of the promised thing. But their exchange of rights does not change the rights of third parties. If I have a contractual right to a thing, I can complain against my contracting partner if I do not receive it has specified, but I have no complaint against third parties who might act in ways the effect of which is to deprive me of it. This is so, even if the third party deprives me of it as a side effect of some wrong. I cannot complain because my contracting partner cannot give me a right against anyone else that I did not already have. As a result although the wrongdoer occasioned my loss through his wrong, he is not responsible for that loss. In the notorious essay “On a Supposed Right to Lie from Philanthropy,” Kant appears to go further, but, for reasons I allude to in a later footnote, I do not believe that this follows from the rest of his account. I am grateful to Dennis Klimchuk and Arnulf Zweig for discussions of this issue.

13. I may gain a prescriptive right with the passage of time, in a way that complicates but does not undermine the general point here.

14. I put to one side complications concerning tortuous interference with contract. I hope to explore these issues elsewhere.

15. Perhaps some less attractive conception of justice carries with it a different understanding of responsibility that nonetheless incorporates norms in a similar way, so that, for example, a serf is responsible for the consequences if he violates norms of fealty. I do not explore that possibility here, because my concern is with the relation between justice and responsibility, not that between injustice and responsibility.

16. Anscombe, G.E.M., “Modern Moral Philosophy” in The Collected Philosophical Papers of G.E.M. Anscombe, Vol. 3: Ethics, Religion, and Politics (Oxford: B. Blackwell, 1981) Minneapolis, MN: University of Minnesota Press, 1981.) 26 at 35–36.Google Scholar

17. Hart, H.L.A., Punishment and Responsibility (Oxford: Oxford University Press, 1970) at 152.Google Scholar

18. For another example of an account that is normative in this sense, see Wallace, R. Jay, Responsibility and the Moral Sentiments (Cambridge, MA: Harvard University Press, 1994.Google Scholar) In political philosophy, the account of responsibility implicit in Locke’s discussions of consent property, and slavery, (discussed below) can be read as incorporating norms determining whether someone had a fair chance. More recently, G.A. Cohen’s view of responsibility, discussed below, can be read as an application of Hart’s approach, although Cohen himself sees the question of responsibility as leading political philosophy into the question of free will. See Cohen, , “On The Currency of Egalitarian Justice” (1989) 99 Ethics 906 at 934.Google Scholar

19. The capacities relevant to the reciprocity conception all concern the ability to bring one’s conduct into conformity with the norms of conduct that the reciprocity conception incorporates. A person is not responsible for what she does while sleepwalking, and the reciprocity conception carves out narrow exemptions for those who, because of age or disability, cannot be guided by particular norms of conduct. These exemptions do not represent the intrusion of considerations about agency into the reciprocity conception. Instead, they reflect the reciprocity conception’s focus on norms governing the ways people treat each other.

20. Thin versions of the agency conception can accommodate responsibility for such consequences, but only at the cost of supposing that a person is accountable for all of the consequences of his actions.

21. I explain the role of institutions in making the demands of justice determinate in “Authority and Coercion” supra note 11.

22. See Aristotle, , Nicomachean Ethics, trans by. Crisp, Roger (Cambridge: Cambridge University Press, 2000)CrossRefGoogle Scholar; Stone, Martin, “The Significance of Doing and Suffering” in Postema, Gerald ed., Philosophy and the Law of Torts (Cambridge: Cambridge University Press, 2001) 131.CrossRefGoogle Scholar

23. Some defenders of distributive justice have doubted that corrective justice makes sense, because it leaves so much to luck. See, for example, Waldron, Jeremy, “Moments of Carelessness and Massive Loss” in Owen, David, ed., Philosophical Foundations of Tort Law (Oxford: Oxford University Press, 1995) 387.Google Scholar I respond to Waldron’s and similar objections in Equality, Responsibility and the Law (Cambridge: Cambridge University Press, 1999) ch. 3.

24. Here as in other legal contexts, one can have a duty that one lacks the capacity to fulfil. If I owe you a hundred dollars, I owe you the money, even if I have difficulty paying. The reciprocity conception explains the foreseeability requirement in terms of the idea that there cannot be a duty that nobody could fulfill, because nothing would count as fulfilling it.

25. There are two apparent exceptions to the foreseeability requirement. One is the so called “thin skull rule”, which holds an injurer liable if what would have been a minor injury for most people turns out to be a severe one because of some preexisting condition of the plaintiff. It is only an apparent exception, however, because an injurer is always liable for the full extent of an injury provided that injuries of that type were foreseeable. If the injurer had a duty to avoid that type of injury; its unforeseen extent just is the injury he was supposed to avoid. The other apparent exception is fraud. Kant says that the person who tells a lie is “responsible for its consequences … however unforeseen they may have been.” See “On a Supposed Right to Lie From Philanthropy” in Practical Philosophy, supra note 10 at 613. That most unKantian British judge, Lord|Denning, says that “it does not lie in the mouth” of the fraud to complain that he could not have foreseen what the dupe would do in response to his lie. Doyle v. Olby (Ironmongers) Ltd. and others [1969] 2 Q.B. 158 at 167 (C.A.). Here too the exception is only apparent. While he cannot foresee what the dupe will do in response, the one thing the fraud can foresee is that his dupe will act on the belief he has induced in him. As a result, he cannot complain that the actions taken were unforeseeable, because those actions just are the injury he had a duty to avoid.

26. I assume here that storms are not seasonal—if they are, the initial injurer could be liable for delaying a ship so that it set out in the bad season. But if the season is predictably dangerous, the master of the injured ship might be judged responsible for not staying longer in harbour, since any such delay would be at the injurer’s expense.

27. Weinrib, Ernest develops a similar example in “Right and Advantage in Private Law” (1989) 10 Cardozo L. Rev. 1283.Google Scholar

28. Epstein, Richard, “A Theory of Strict Liability” (1973) 2 J. Leg. Stud. 151 Google Scholar. It is essential to Epstein’s project that liability is not conditioned on norms of conduct. Indeed, that is the root of his account’s difficulties.

29. Hart, H.L.A. & Honoré, Tony, Causation in the Law, 2nd ed. (Oxford, 1985) at 268 CrossRefGoogle Scholar, quoting McLaughlin, James Angell, “Proximate Cause” (1925) 39 Harv. L. Rev. 149 at 164.Google Scholar

30. These include the absence of liability for pure economic loss, and for negligent invasion of contractual interests.

31. As Perry, Stephen has done. See particularly “The Moral Foundations of Tort Law” (1992) 77 Iowa L. Rev. 449 especially at 496–514 Google Scholar, and “Responsibility for Outcomes, Risk, and the Law of Torts” in Philosophy and the Law of Torts, supra note 22 at 72. Perry’s account is a normative agency account, as evidenced by his invocation of Hart’s capacity/opportunity principle.

32. The agency conception cannot avoid this problem, a because agency is subject to what Joel Feinberg once called “the accordion effect.” Just as an accordion can be pulled out or com Pressed to varying degrees, so to can our sense of what a person has done. As I flip the switch, I also turn on the light and alert the burglar to the fact that someone is home. There is no particular place where it is more or less natural to draw the line between what I have done and its more distant consequences. Of course, that doesn’t mean that a line cannot be drawn. But it does mean that the concept of agency itself will be of no assistance in drawing it. See Feinberg, Joel, “Action and Responsibility” in Doing and Deserving: Essays on the Theory of Responsibility (Princeton, NJ: Princeton University Press, 1970) 119 at 134.Google Scholar

33. The agency conception could be narrowed in one other way that would seem to make limits to responsibility inherent in it, by limiting responsibility to the known or intended consequences of a deed. On such an understanding, I might be answerable if my bonfire burns down your fence but not your house, if I realised that the former could happen, but didn’t think of the latter. Much of the literature on free will seems to work with just such a conception—a person is responsible for those things that he set out to do.

34. See, for example, Nozick, , Anarchy, State, and Utopia, supra note 4 at 283 Google Scholar. So far as I know, Randy Barnett is unique even among libertarians in advocating debt slavery for those who impose costs on others through wrongdoing. Although perhaps not to his moral credit, it is at least to his intellectual credit that he pursues this line of thought consistently. See The Structure of Liberty: Justice and the Rule of Law (New York: Oxford University Press, 1998)Google Scholar

35. Locke, John, Two Treatises of Government, Laslett, Peter, ed., (Cambridge: Cambridge University Press, 1988) at 285.CrossRefGoogle Scholar

36. Hegel, F., Elements of the Philosophy of Right, ed. by Wood, Allen, trans. Nisbet, H.B. (Cambridge: Cambridge University Press, 1991) at 96.Google Scholar

37. Wood, Allen Hegel’s Ethical Thought (Cambridge: Cambridge University Press, 1990) at 99.CrossRefGoogle Scholar

38. Cohen, G.A., Self-Ownership, Freedom, and Equality (Cambridge, Cambridge University Press 1995) at 47.CrossRefGoogle Scholar

39. I take this to be Cohen’s considered view of responsibility, since it is the only view compatible with both the limits he places on it in “On The Currency of Egalitarian Justice”(supra note 18) and his claim in If You ‘re an Egalitarian, How Come You ‘re So Rich? (supra note 8) that talented people who charge whatever the market will bear for their services are responsible for so doing. The talented are presumably responsible because they have other acceptable options. I am grateful to Susan Hurley for drawing my attention to this tension in Cohen’s several treatments of responsibility.

40. This is not Cohen’s only objection to slave contracts. He says, in passing, (Self-Ownership, supra note 38 at 21) that we might want to prohibit slavery simply because it is wrong. He declines to pursue that line of thought, whatever its merits, because those sympathetic with Nozick will not be “moved by it.” This argument depends less on the contingent effects of slavery on third parties, but it, too, is an extrinsic limit on an agency conception. The agency conception itself has nothing to say against slave contracts.

41. The category of mechanic’s tools is broad, and includes, among other things, a barber’s mirror and chairs. See, for example, Terry v. McDaniel 103 Tenn. 415 (1899).

42. Earlier, I described Nozick’s endorsement of slave contracts “passing” because he says that a free society would allow them, not that such a society would enforce them. His failure to attend to this distinction reflects his acceptance of the idea that the state can have no powers of enforcement that individuals lack. As we have seen, the fact that I can agree to do your bidding does not, on its own, entail anything about anyone’s right to enforce such an agreement.

43. Although it may, through education and the provision of adequate opportunities make servility a less appealing option. See Hill, Jr., Thomas, “Servility and Self-Respect” in his Autonomy and Self-respect (Cambridge: Cambridge University Press, 1991).CrossRefGoogle Scholar

44. Kant makes essentially the same when he claims that people cannot give up their equality by any rightful deed. See “On the Common Saying: That may be correct in Theory but it is of no use in practice” in Kant, , Practical Philosophy, supra note 10 at 293-93.Google Scholar

45. De Francesco v. Barnham, (1890) 45 Ch.D. 430 at 438 per Fry, L.J.

46. Again, Kant’s development of the reciprocity conception makes this contrast clear. See Kant, , Practical Philosophy, supra note 10 at 471-72.Google Scholar

47. Do the slave gamblers in Cohen’s example bind each other equally because they enter their gamble on fair terms? They do not because how things turn out is crucial on the reciprocity conception; it matters that people interact as equals, not that they have once so interacted. Once one becomes master and the other slave, they never again interact as equals, because the slave is entirely at the master’s mercy, since the master has rights against the slave which the slave does not have against him.

48. See, for example, Dworkin, , Sovereign Virtue, supra note 5Google Scholar; Cohen, , “On the Currency of Egalitarian Justice,” supra note 18Google Scholar; Roemer, John E., Equality of Opportunity (Cambridge, MA: Harvard University Press, 1997)Google Scholar; Arneson, Richard “Liberalism, Distributive Subjectivism, and Equal Opportunity for Welfare” (1990) 19 Phil. & Pub. Aff 158 Google Scholar

49. Indeed, the two developments of the reciprocity conception that I have appealed to—Kant and the common law—certainly do not suppose that the conception of justice underlying it mandates material equality. But if the reciprocity conception does not mandate material equality, it provides an interpretation of the impetus behind the call for material equality. Rather than supposing that material equality is somehow valuable apart from its effects, the basic idea of justice underlying the reciprocity conception supposes lets us see the value of material equality in the way that it enables people to interact as equals. On this understanding, material inequality is objectionable for the same reason that slave contracts are: it renders one person utterly dependent on the whims of others. Here, as in the slave contract example, the difficulty is not that a person is left without the means or opportunities to set and pursue his or her own ends. Instead, the difficulty is that where inequality is of wealth is severe, some are entirely subject to the generosity or bargaining power of others.

50. Roemer, John E., “Egalitarianism and Responsibility” (1995) 20 Boston Rev.. FIRST PAGE? Google Scholar

51. Dworkin, , Sovereign Virtue, supra note 5 at 344.Google Scholar

52. The same concern for material equality as a condition of fair interaction underwrites a variety of other institutional arrangements. I mention only one example here. A commitment to fair equality of opportunity requires that new parents be eligible for parental leave on terms that do not compromise their careers. To make new parents bear the cost of such programs by making leaves unpaid would thus be unacceptable. But to place the burdens on employers is also unacceptable, because it compromises their ability to function in a competitive setting. Instead, the burden should be held in common by the society. The reason is not that well cared-for children are a benefit to all—thought that is doubtless true—but that equality requires it.

53. Of course, that argument rests on the idea that people should interact as equals; that idea is not derived from an idea of responsibility.

54. This is not the place to consider the appropriate mechanism for redistribution, in particular, whether it should take the form of transfers or minimum wages, and whether taxation should be based on income or expenditures.

55. So, for example, Dworkin criticises Rawls’s “difference principle” for being insufficiently sensitive to considerations of responsibility, because it focuses on the condition of the worst-off without considering how they got to be badly-off.

56. I said earlier that the reciprocity conception can explain why something close to an agency conception would be appropriate to the criminal law, and these examples provide the outline of just such an explanation: the person who sets out to exploit norms of justice commits a wrong against those norms in addition to the wrong he commits against particular persons. Such wrongs are important, but are not the appropriate starting point for thinking about responsibility.

57. Rawls, John, A Theory of Justice (Oxford: Oxford University Press, 1972) at page number? Google Scholar