Can Contractualism Be an Account of What We Owe to Each Other?

I. Introduction

Contractualism holds, very roughly, that an action x of type X is wrong, at least in one unified and distinct way, iff and because any principle permitting actions of type X could be reasonably rejected. It has been thought, however, that contractualism gives an account not only of when an action x is wrong, but moreover when x wrongs some person B. This paper articulates how exactly contractualism might give an account of when x wrongs some person B. This articulation in turn has several significant implications. Among these implications are that contractualism generates moral requirements whose violation wrong no one and that, because of this, contractualism may have a harder time giving an account of wronging than might otherwise have been thought.

According to contractualism, an action x of type X is wrong, at least in one unified and distinct way, iff and because—very roughly—any principle permitting actions of type X could be reasonably rejected. More precisely put: an action x of type X in circumstances C is wrong, at least in one unified and distinct way, iff and because any principle, P, permitting people to X in C could be reasonably rejected by an occupant of a (suitably generally characterized) standpoint, S, (e.g., that of “‘a person in need of help,’ ‘a person who is relying on an assurance that someone else has given’”¹) on the basis of generic, personal reasons, concerning how that occupant of S would be affected or treated by the general acceptance of P, even given the generic, personal reasons of occupants of other standpoints in favor of rejecting alternatives to P. Since this is a mouthful, let us henceforth use the following as shorthand: “when any principle, P, permitting people to X in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M.” “Affect or treat” is to be understood broadly, to include such things as being treated paternalistically, or being treated unfairly, or being exposed to risk of harm, even when this does not “set back one’s interests” on narrower construals of that latter phrase.

I take it to be uncontroversial that, at least sometimes when A is morally required not to x, A owes it to some B not to x. We can call the domain of moral requirements of which this is true, “relational morality.” To speak of “owing” is, moreover, to choose just one point of entry into a network of relational ideas. In saying:

A owes it to B not to x

one might just as well say any of the following:

A has a directed duty to B not to x,

B has a claim against A that A not x,

A wrongs B by x-ing out of disregard for B’s claim against A that A not x,

B has grounds to resent A for x-ing out of disregard for B’s claim against A that A not x.

Any of these statements, arguably, is necessary and sufficient for any other. At any rate, I will treat them as so.

According to what I will unimaginatively call the “Thesis,” contractualism, first, gives an account of relational morality, allowing us to specify when an agent, A, owes it to some patient, B, that A not x, and, second, does so by exploiting resources that are not available to other moral theories. The Thesis appears to be endorsed in the classic exposition of contractualism, T.M. Scanlon’s What We Owe to Each Other, as well in his latest major exposition of contractualism, “Contractualism and Justification.”² Certainly, Scanlon’s choice to label the domain of morality of which contractualism gives an account the morality of “what we owe to each other,” as well as his description of it as “having to do with our duties to other people,”³ suggests some sympathy with the Thesis. And such sympathy is also suggested by remarks in “Contractualism and Justification.” There he writes: “The kind of remorse made appropriate by wrongs of the kind I had in mind in developing my contractualist view is grounded in the sense that an individual who is affected by our action has a reason for objecting to it that cannot be answered satisfactorily.”⁴ He then adds in a footnote: “this makes it plausible to say this person has been wronged by such an action, not just that the action is wrong.”⁵ This suggests that when a moral requirement has a contractualist justification, it is owed to someone. He later writes, of a case in which we can save either a smaller or a larger group: “It thus seems that saving either group would be permissible according to contractualism.” He immediately adds, “But even if this is so, and no individual would be wronged by either course of action….”⁶ This suggests that when there is no contractualist justification of a moral requirement, there is nothing that is owed to anyone. Even clearer sympathy is expressed in “Wrongness and Reasons: A Re-examination”: “What I was concerned with in [What We Owe to Each Other]… is one particular way in which an action can be morally wrong, the way that involves wronging someone or, as I say there, violating ‘what we owe to others.’”⁷ Moreover, it is part of Scanlon’s case for contractualism, over alternatives such as consequentialism, that it is supposed to make sense, where alternatives cannot, of the distinctive, directed kind of “remorse” that we feel when we fail in what we owe to, say, a person in need.⁸ This remorse again focuses on the fact that they could object, on their own behalf, to our not helping them, as opposed to focusing on, say, the badness of a state of affairs in which a human need went unmet. And Scanlon is not alone in thinking that contractualism gives an account of relational morality. Significantly, R. Jay Wallace’s The Moral Nexus, the leading defense of the idea that relational morality is the whole of morality, also suggests it.⁹

I too am deeply attracted both to contractualism and the Thesis. While they might seem synonymous, however, work needs to be done to show that contractualism can be an account of what we owe to each other. This is a point that is often neglected in the vast literature on contractualism.¹⁰ When is it that A wrongs B by a particular action x? When is it that “an individual [B] who is affected by [A’s particular] action [x] has a reason for objecting to it that cannot be answered satisfactorily”? The answer is not given by the official contractualist formula for when x is wrong: namely, when any principle, P, permitting people to X in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M, and x-ing is an instance of X-ing in C. That formula says nothing about any particular person, B, to whom not x-ing is owed. So, more needs to be said, beyond the contractualist formula, for it to provide an account of what we owe to each other. We need a “bridge principle,” which carries us from the claim that action x of type X is wrong, because any principle permitting actions of type X could be reasonably rejected, to the claim that action x would wrong someone.

Sections II–IV of this paper aim to articulate what more needs to be said: how exactly contractualism gives an account of relational morality. This articulation takes the form of specifying the required bridge principle, which would bridge the gap between the claim that actions of type X could be reasonably rejected and the claim that some particular person A owes it to some particular person B that A not perform some particular action x. Very roughly, the bridge principle says that A owes it to B not to x iff and because actions of type X could be reasonably rejected, x is of type X, and x affects B in the way that justifies that reasonable rejection. As we will see, there are complications, but that is the rough idea.

The rest of the paper explores the implications of our bridge principle. First, as noted in Section V, our bridge principle implies that it is never the case that the fact that A has a contractualist duty not to x is explained by or just is the fact that A owes it to some B not to x. This is because the fact that A is under a contractualist duty not to x is sufficiently explained by the fact that actions of type X could be reasonably rejected and x is of type X, without need to appeal to the fact, even if it is a fact, that there is some B who would be affected by x in the way that justifies rejection. This means that contractualism cannot support the common, if not always explicitly articulated, thesis that in at least some cases when A is morally required not to x, the fact that A is morally required not to x is explained by or just is the fact that A owes it to some B not to x.

The remaining implications concern the possibility that at least some contractualist duties are not owed to anyone. Presumably, contractualism implies discretionary duties of aid. Yet it is sometimes suggested that discretionary duties of aid are not owed to anyone. One may be obligated to give something to the world’s needy, but it is not as though, or so it is suggested, one owes it to any particular needy person to do so. However, as I argue in Section VI, our bridge principle shows, on the contrary, that one owes it to every needy person, vindicating a view suggested by G.E.M. Anscombe and Wallace. This should be welcome to Scanlon, who, as we have seen, views duties of aid as paradigms of what we owe to each other. And it should be welcome to Wallace, who holds that all moral duties, and so a fortiori duties of aid, are part of what we owe to each other.

However, our bridge principle also implies that, with respect to other cases, contractualism does indeed generate moral prohibitions that are not owed to anyone. This is, at the highest level of abstraction, because it can be the case that actions of type X could be reasonably rejected and that action x is of type X, even though there is no B who would be affected by x in the way that justifies rejection. One set of examples, discussed in Section VII, are “conjunctive requirements,” such as the requirement to save both Mary and Jane, when you can save both.

Another set of examples, discussed in Section VIII, are “nonspecifically justified prohibitions.” These are prohibitions whose justification depends on the effects of the general acceptance of principles permitting violations of the prohibition, as opposed to the effects of particular violations of the prohibition. For example, it might be suggested that the reason why principles permitting spying could be reasonably rejected is not because of the effects of any particular episode of spying, but instead because of the widespread anxiety that would result from the general acceptance of those principles.

It is a surprising result, and worthy of note in itself, that some contractualist prohibitions are not owed to anyone. It means that contractualist morality is not quite co-extensive with the domain of what we owe to each other. There is a part of contractualist morality that lacks the special significance of relational contractualist prohibitions, which depends on there being “an individual who is affected by our action [and who] has a reason for objecting to it that cannot be answered satisfactorily.”

More importantly, if not all contractualist prohibitions are owed to others, then contractualism may have a harder time capturing the intuitive extension of relational morality, of what we owe to each other, than might otherwise be thought. Consider what seems intuitively part of relational morality: that we owe it to others not to spy on them. It might seem sufficient for contractualism to explain that fact—that we owe it to others not to spy on them—that contractualism generates a prohibition on spying. But this is to assume, falsely, that all contractualist prohibitions are owed to others. If this prohibition is, say, nonspecifically justified (as Scanlon suggests that it is), then it is not in fact owed to others. Thus, we have not explained why we owe it to others not to spy on them. Sections IX–XI explore this point and its consequences.

Section XII concludes that, even if the foregoing is mistaken, even if, say, the bridge principle has not been correctly specified, this paper still contributes something. This is to highlight that, while it might seem definitional that contractualism gives an account of what we owe to each other, there is a live question whether it does.

II. Articulating the Bridge Principle: First Steps

The Thesis holds that contractualism gives an account of relational morality, specifying when A owes it to some B that A not x. But what is this account? Wallace holds that “[B] can have a specific claim against [A] only if [B]’s interests would be affected negatively by [A]’s failure to honor the claim.”¹¹ This might suggest that A owes it to B not to x iff and because—very roughly—actions of type X could be reasonably rejected, x is of type X, and x affects B negatively. More fully, it suggests:

Bridge, First Pass   A owes it to B not to x (in the given circumstances) iff and because

• (i)  any principle, P, permitting people to X in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M,

• (ii)  (in the given circumstances) x is an instance of, or is a type of action every instance of which is an instance of, X-ing in C, and

• (iii)  (in the given circumstances)¹² x is an action that, or a type of action every instance of which, would affect B negatively.

Bridge, First Pass clearly won’t do. It implies that, if you place a bet that I will not torture Jones, then I owe it to you not to torture Jones. Surely any principle permitting people to torture others could be reasonably rejected, so clause (i) is satisfied. Torturing Jones is an instance of torturing someone, so clause (ii) is satisfied. Since losing the bet would affect you negatively, clause (iii) is satisfied. Hence, according to Bridge, First Pass, I owe it to you not to torture Jones. But I do not owe it to you not to torture Jones.¹³ I owe it to Jones.

The natural reply is that your losing the bet is not the right manner of being affected or treated, because it is not the manner of being affected or treated that justifies the prohibition on torture. Wallace goes on to write: “people are wronged by an action when it affects personal interests of theirs that would make it reasonable to reject principles permitting actions of the relevant kind.”¹⁴

He immediately goes on to offer an illustration:

we all have reasonable objections to principles for the general regulation of behavior that permit people to kill us for professional or private advantage, objections that are grounded in our basic personal interest in remaining alive. But in any particular case of impermissible killing, it is the individual whose interest in life is actually disregarded who will be wronged by the killer’s act: that is, the person who is killed.¹⁵

This suggests that A owes it to B not to x iff and because—very roughly—actions of type X could be reasonably rejected, x is of type X, and x affects B in the way that justifies that reasonable rejection. More fully, it suggests:

Bridge, Second Pass   A owes it to B not to x iff and because

• (i)  any principle, P, permitting people to X in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M,

• (ii)  x is an instance of, or a type of action every instance of which is an instance of, X-ing in C, and

• (iii)  x is an action that, or is a type of action every instance of which, would bring it about that B is affected or treated in manner M.¹⁶

For instance, and very roughly, any principle permitting people to torture others could be reasonably rejected by the people who might be tortured on the grounds that they would suffer torture. Torturing Jones is an instance of torturing someone. And my torturing Jones would bring it about that Jones would suffer torture—as opposed to being affected negatively in some other way, such as losing a bet. Hence, I owe it to Jones not to torture Jones. By contrast, while you will be negatively affected by losing the bet, this manner of being affected—as opposed to being tortured—is not the kind that justifies the reasonable rejection. Hence, I do not owe it to you not to torture Jones. More fully, I owe it to Jones not to torture Jones for the following reasons. Any principle, P, permitting people to torture others, could be rejected by an occupant of some standpoint, S, namely that of a person who might be subject to torture, on the grounds that the general acceptance of P would affect or treat that occupant by subjecting them to torture. This satisfies clause (i). Every instance of my torturing Jones is an instance of my torturing someone. This satisfies clause (ii). Every instance of my torturing Jones would bring it about that Jones is subjected to torture. This satisfies clause (iii), where Jones is B. Hence, I owe it to Jones not to torture Jones. By contrast, I do not owe it to you not to torture Jones. This is because not every instance of my torturing Jones would bring it about that you are subjected to torture. Clause (iii) is not satisfied, where you are B.¹⁷

Note that Bridge, Second Pass (as well as Bridge, Final to come) implies that there can be actions that are owed to everyone, such as not free riding on social practices to which everyone contributes, or not triggering a doomsday device that will kill everyone, since in those cases everyone happens to be affected or treated in the manner that grounds the reasonable rejection of principles that permit free riding or killing. But this is as it should be. While it is sometimes suggested that actions that are owed to some are not owed to others, this isn’t built into the very idea of relational morality. It is a substantive moral question whether there are actions that are owed to some people but not others. In principle, it might have turned out that every action that is owed to someone is owed to everyone. Of course, it seems overwhelmingly plausible, as a substantive matter, that there are actions, such as not torturing Jones, that are owed to Jones and not others.¹⁸

III. Articulating the Bridge Principle: Last Steps

However, Bridge, Second Pass is still not quite right. Suppose that:

F1: for whatever reason, if I torture Jones, I will cause someone else to torture Smith. And suppose that:

P1: any principle permitting people to torture others could be rejected by an occupant of some standpoint, S, namely that of a person who might be subject to torture, on the grounds that the general acceptance of P would affect or treat that occupant by subjecting them to torture.

Bridge, Second Pass, together with F1 and P1, then explain the fact that:

C1: I owe it to Smith not to torture Jones.

Clause (i) is satisfied directly by P1. Clause (ii) is satisfied because my torturing Jones is an instance of my torturing someone. And clause (iii) is satisfied because, given F1, my torturing Jones would bring it about that Smith is subjected to torture. However, Bridge, Second Pass should not, in combination with F1 and P1 alone, explain the fact that I owe it to Smith not to torture Jones. Surely, I do owe it to Smith not to torture Jones. But this is not because P1: that any principle permitting people to torture others themselves is rejectable. It is instead because:

P2: any principle permitting people to cause others to torture others could be rejected by an occupant of some standpoint, S, namely that of a person who might be subject to torture, on the grounds that the general acceptance of P would affect or treat that occupant by subjecting them to torture.

To put the point another way, suppose that, contrary to fact, while P1 was true—so that any principle permitting people to torture others themselves was rejectable—P2 was false—so that some principle permitting people to cause others to torture others was not rejectable. Then it would be false C1 that I owe it to Smith not to torture Jones. But Bridge, Second Pass, given F1 and P1, implies that C1 would still be true.¹⁹

What we need, I believe, is to add something to the contractualist formula, which is to understand the principles under consideration as being themselves “relational” in a very weak sense. This is that they are of the form: “agents are permitted to X patients.” In other words, the principles up for reasonable rejection articulate X in such a way as to identify a privileged role of patient. For instance, the principle might be: “agents are permitted to break promises to patients” or “agents are permitted to torture patients.” But the principle might also be “agents are permitted to torture the contemporaries of patients,” or “agents are permitted to torture people whose last names start with the same letter as patients’,” where the relationship between the action and the designated patient is less direct and more contrived. Not much is ruled out; the principle just has to anoint someone as the patient. Nevertheless, this is enough to allow us to distinguish principles that permit people to torture patients from principles that permit people to cause others to torture patients.

We can thus replace Bridge, Second Pass with:

Bridge, Final   A owes it to B not to x iff and because

• (i)  any principle, P, permitting people to X patients in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M,²⁰

• (ii)  x is an instance of, or a type of action of which every instance is an instance of, X-ing B in C, and

• (iii)  x is an action that, or is a type of action every instance of which, would bring it about that B is affected or treated in manner M.

The crucial difference is that Bridge, Final requires, for A to owe it to B not to x, not only that B be affected or treated in the way that justifies rejection of the relevant principle, but also that B occupy the role of patient specified in that principle. According to Bridge, Final, like Bridge, Second Pass, I owe it to Jones not to torture Jones, because P1—principles permitting people to torture patients are reasonably rejectable—and Jones is a patient who is directly tortured by my torturing Jones. According to Bridge, Final, like Bridge, Second Pass, I owe it to Smith not to torture Jones, because P2—principles permitting people to cause others to torture patients are reasonably rejectable—and, given F1, Smith is a patient who is caused to be tortured by someone else by my torturing Jones. However, according to Bridge, Final, unlike Bridge, Second Pass, I don’t owe it to Smith not to torture Jones because P1—principles permitting people to torture patients are reasonably rejectable. This is because Smith is not a patient who is directly tortured by my torturing Jones.

More fully spelled out: I owe it to Jones not to torture Jones. Why? Clause (i) is satisfied because any principle, P, permitting people to torture patients, could be rejected by an occupant of some standpoint, S, namely that of a person who might be tortured, on the grounds that the general acceptance of P would affect or treat that occupant by subjecting them to torture. Clause (ii) is satisfied because my torturing Jones is a type of action every instance of which is an instance of my torturing Jones. And clause (iii) is satisfied because every instance of my torturing Jones would bring it about that Jones is subjected to torture. I owe it to Smith not to torture Jones. Why? Clause (i) is satisfied because any principle, P, permitting people to cause others to torture patients, could be rejected by an occupant of some standpoint, S, namely that of a person who might be tortured, on the grounds that the general acceptance of P would affect or treat that occupant by subjecting them to torture. Clause (ii) is satisfied because every instance of my torturing Jones (in the imagined circumstances) is an instance of my causing others to torture Smith. And clause (iii) is satisfied because every instance of my torturing Jones would bring it about that Smith is subjected to torture. However, it is not the case, according to Bridge, Final, that I owe it to Smith not to torture Jones because any principle, P, permitting people to torture patients, could be rejected by an occupant of some standpoint, S, namely that of a person who might be tortured, on the grounds that the general acceptance of P would affect or treat that occupant by subjecting them to torture. It is not the case because it is not the case that every instance of my torturing Jones is an instance of my torturing Smith, although every instance of my torturing Jones will cause Smith to be tortured by someone else. Clause (ii) is not satisfied.

IV. Rejecting a Simpler Bridge Principle

Our revised clause (ii) requires that x be an instance of, or a type of action of which every instance is an instance of, X-ing B in C. However, given that there is already “directionality” in clause (ii)—that is, given that it invokes a B to whom the duty not to x might be owed—one might wonder whether we still need clause (iii) to explain “directionality.” But we do still need clause (iii). If we deleted clause (iii), we would get:

Bridge, Fourth Pass   A owes it to B not to x iff and because

• (i)  any principle, P, permitting people to X patients in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M, and

• (ii)  x is an instance of, or a type of action every instance of which is an instance of, X-ing B in C.

Now suppose that:

F2: Smith will be entirely unaffected by my torturing Jones.

Then Bridge, Fourth Pass would incorrectly imply C1: that I owe it to Smith not to torture Jones. This is because, first, clause (i) is satisfied by the fact that:

P3: any principle permitting agents to torture the contemporaries of patients could be rejected by an occupant of some standpoint, S, namely that of a person who might be subject to torture, on the grounds that the general acceptance of those principles would affect that occupant by subjecting them to torture.

And second, cause (ii) is satisfied, since every instance of my torturing Jones is an instance of my torturing a contemporary of Smith. Given Bridge, Fourth Pass, it follows C1: that I owe it to Smith not to torture Jones. If we add clause (iii) back in, however, then we avoid this result: namely, that I owe it to Smith (who will be totally unaffected) not to torture Jones. This is because clause (iii) requires that, for me to owe it to Smith not to torture Jones, Smith must be affected or treated in way M—namely being tortured—by my torturing Jones, and by F2, Smith is not affected or treated in way M—namely being tortured—by my torturing Jones.

Even if Bridge, Fourth Pass were extensionally adequate, there would still be a problem with the general approach of which it is an instance: namely, to locate the “directionality” in the content of the principles alone, as opposed to the mechanism by which the principles are selected. A rule consequentialist proposing to give an account of relational morality might similarly say:

Rule-Consequentialist Bridge   A owes it to B not to x iff and because

• (i)  a principle, P, permitting agents to X patients in C, would be suboptimal, and

• (ii)  x is an instance of, or a type of action of which every instance is an instance of, X-ing B in C.

If we accept Bridge, Fourth Pass, in other words, then the contractualist cannot claim to give a better account of relational morality than can the rule consequentialist. This goes against the Thesis, which holds that there is something distinctive about the structure of contractualism that suits it to giving an account of relational morality and for that reason argues for it over rivals such as rule consequentialism. The problem will arise for any account that tries to place the “directionality” entirely in the content of principles, as opposed to the mechanism by which principles are selected: that is, for any account where the grounds for rejecting principles contribute nothing to “directionality.”²¹

V. Implications: Moral Requirements Neither Are Explained by nor Just Are Owings

Having argued for Bridge, Final, let us now turn to its implications. To begin with, it is often assumed, if not always explicitly articulated, that in at least some cases when A is morally required not to x, the fact that A is morally required not to x either is explained by or just is the fact that A owes it to some B not to x.²² This seems to follow, for example, from Wallace’s contention that in all cases when A is morally required not to x, the fact that A is morally required not to x just is the fact that A owes it to some B not to x.

If Bridge, Final is correct, however, then it is doubtful that it is ever the case that the fact that A has a contractualist duty not to x is explained by or just is the fact that A owes it to some B not to x. This is because contractualism presumably holds that clause (i)—that any principle, P, permitting people to X patients in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M—and clause (ii)—that x is an instance of, or a type of action every instance of which is an instance of, X-ing some patient, B, in C—together fully explain the fact that A is morally required not to x, without the addition of clause (iii)—that x is an action that, or is a type of action every instance of which, would bring it about that that same patient is affected or treated in manner M. It would be odd for a contractualist to say, “Yes, actions of type X are reasonably rejectable, and yes, action x is of type X, but all the same something more is needed to make it the case that x is wrong.” However, Bridge, Final implies that together clauses (i), (ii), and (iii) at least partly explain the fact that A owes it to some B not to x. If we maintain that the fact that A owes it to some B not to x at least partially explains the fact that A is morally required not to x, then we get the result that together clauses (i), (ii), and (iii) at least partially explain the fact that A is morally required not to x. But this contradicts the earlier finding that clauses (i) and (ii), without the addition of (iii), fully explain the fact that A is morally required not to x. And it is hard to see how the fact that A is morally required not to x can just be the fact that A owes it to some B not to x if what explains the former is not what explains the latter. So contractualism cannot support the common, if often implicit, idea that in at least some cases when A is morally required not to x, the fact that A is morally required not to x is explained by or just is the fact that A owes it to some B not to x.

VI. Implications: Discretionary Duties of Aid Owed to All in Need

The remaining implications of Bridge, Final concern the possibility that contractualism implies duties that are not owed to anyone. To begin with, consider discretionary duties of aid. It is natural to think both that there are discretionary duties of aid and that contractualism implies them. However, some suppose that discretionary duties of aid are not owed to anyone. One may be obligated to give something to the world’s needy, but it is not as though, or so some suppose, one owes it to any particular needy person to do so. If this supposition is right, then contractualism implies that there are duties that are owed to no one. But is it right?

Suppose, to simplify matters, that there are just two people in need, Mary and Jane, you have the resources to aid one, but only one, without excessive cost to yourself, and there are no other morally relevant features of the case (such as, say, that you have a contractual obligation to aid Mary but not Jane, that a bomb will detonate if you aid either Mary or Jane, etc.). Intuitively, and as one would expect contractualism to imply, you are morally required either to aid Mary or to aid Jane. If you don’t aid Mary, then you will have acted impermissibly unless you aid Jane. If you don’t aid Jane, then you will have acted impermissibly unless you aid Mary. Intuitively, though, and as one would expect contractualism to imply, you are not morally required to aid Mary. If you aid Jane instead, that’s OK. And likewise you are not morally required to aid Jane. If you aid Mary instead, that’s OK. So, assuming that you owe it to someone to do only what you are morally required to do, and as seems anyway plausible, you do not owe it to anyone to aid Mary, and you do not owe it to anyone to aid Jane. To whom, if anyone, do you owe it either to aid Mary or to aid Jane?

Wallace answers that you owe it to Mary and that you owe it to Jane.

If this is the right way of thinking about mutual aid, however, then it seems we have a way of identifying the parties to whom the duties are owed. They would include all of those with acute needs who are in the class of potential beneficiaries of a given agent’s beneficent efforts. These are the people who have objections on their own behalf to principles that would permit agents to do little or nothing. We might therefore wish to say that agents owe it to all of the individuals in this class that they should live up to the principles of mutual aid that it would be unreasonable for anyone to reject; by the same token, each of these individuals has a claim against agents that they should comply with the principles in question.²³

He grants that these moral claims have the “unusual” or “peculiar” feature that their satisfaction may affect claimholders in precisely the same way as their violation.

Not only are they not claims to any specific kind of performance on the part of the agent against whom they are held, since agents have wide discretion to determine for themselves which kinds of contribution they are going to make to alleviating the distress of those who are in dire need. They are not even claims that agents should do anything to help the claimholder in particular, since they could fully be discharged through discretionary efforts that end up assisting completely different members of the class of potential beneficiaries.²⁴

However, he insists, despite this, that:

moral claims, and the corresponding directed duties, may nevertheless intelligibly be ascribed to the parties in these cases, and that it can be illuminating to think in these terms. Affluent agents owe it to each of the individuals in the class of potential beneficiaries to do their fair share to provide needed assistance. And each of those individuals in turn has claims against individual affluent agents that they should so contribute. We are perhaps accustomed to thinking of moral claims as demands, compliance with which would redound to the benefit of the claimholder in particular, but I don’t see this as something that is built into the very meaning of a claim.²⁵

The last quoted sentence, however, is in tension with Wallace’s earlier assertion that: “[B] can have a specific claim against [A] only if [B]’s interests would be affected negatively by [A]’s failure to honor the claim.” At least this is so if we interpret Wallace’s earlier assertion as saying that A owes it to B not to x only if some interest of B’s would be necessarily better served by A’s not x-ing than by A’s x-ing—that compliance with the claim would necessarily “redound to the benefit of the claimholder in particular.” That is what Wallace, in the last quoted sentence, says is not “built into the very meaning of a claim.”²⁶

So, if Wallace wants to say that you owe it to Mary and to Jane to aid either Mary or Jane, then he may need to retract his earlier assertion. Nevertheless, Bridge, Final vindicates Wallace’s claim that you owe it to Mary and owe it to Jane that you aid either Mary or Jane. This is so at least if we accept the following claim:

P4: any principle permitting an agent to fail to aid a sufficiently needy patient in circumstances in which they could do so without excessive cost, in which they are not using the necessary resources to save someone else at least as needy, and in which no other exception obtains,²⁷ could be reasonably rejected by an occupant of some standpoint, S, namely that of a person in need of aid, on the grounds that the general acceptance of those principles would affect that occupant by leaving them without aid.

Now recall:

Bridge, Final   A owes it to B not to x iff and because

• (i)  any principle, P, permitting people to X patients in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M,

• (ii)  x is an instance of, or a type of action of which every instance is an instance of, X-ing B in C, and

• (iii)  x is an action that, or is a type of action every instance of which, would bring it about that B is affected or treated in manner M.

According to Bridge, Final, you owe it to Mary not to aid neither Mary nor Jane—that is, you owe it to Mary to aid either Mary or Jane. This is because, first, clause (i) is satisfied by P4: any principle permitting an agent to fail to aid a sufficiently needy patient in circumstances in which they could do so without excessive cost, in which they are not using the necessary resources to save someone else at least as needy, and in which no other exception obtains, could be reasonably rejected by an occupant of some standpoint, S, namely that of a person in need of aid, on the grounds that the general acceptance of those principles would affect that occupant by leaving them without aid. After all, the principle would say, in effect, even if you are not going to help anyone else, you don’t have to help this person. Surely, that person could reject such a principle on the grounds that, in a situation in which you aren’t going to help anyone else, your not helping them would leave them without aid, indeed aid that they would otherwise have enjoyed.²⁸ Second, clause (ii) is satisfied. Take any instance of aiding neither Mary nor Jane (i.e., not aiding either Mary or Jane). It is, in the imagined circumstances, an instance of failing to aid Mary when you could do so without excessive cost to yourself. It is an instance of doing so when you are not using the necessary resources to save someone else at least as needy; by hypothesis, the only such person is Jane, and you are aiding neither Mary nor Jane. And it is an instance of doing so when no other exception obtains; by hypothesis, there are no other morally relevant features of the case. And third, clause (iii) is satisfied, since every instance of aiding neither Mary nor Jane brings it about that Mary goes without aid. The same goes for Jane.

Now recall that we want to avoid the result that you owe it to Mary to aid Mary. After all, you are permitted to aid Jane instead. Does Bridge, Final, even given P4, avoid the result that you owe it to Mary to aid Mary? Yes, because, in the circumstances, clause (ii) is not satisfied. Not every instance of failing to aid Mary an instance of failing to aid Mary in circumstances in which you are not using the necessary resources to save someone else at least as needy. In particular, if you aid Jane, then you are using the necessary resources to save someone else at least as needy.

It thus might help to make explicit something left implicit in Bridge, Final, to give us:

Bridge, Final, Explicit   A owes it to B not to x iff and because:

• (i)  any principle, P, permitting people to X patients in C, could be rejected by the occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M,

• (ii)  x is an instance of, or a type of action every instance of which is an instance of, X-ing B in C, whatever other actions A may perform and

• (iii)  x is an action that, or a type of action every instance of which, would bring it about that B is affected or treated in manner M, whatever other actions A may perform.

The qualifications are strictly speaking redundant. However, they do helpfully make explicit something left implicit in Bridge, Final. Regarding clause (ii), it is not the case that you owe it to Mary to aid Mary, since it is not the case that every instance of not aiding Mary is an instance of not aiding Mary in circumstances in which you are not aiding someone at least as needy, whatever other actions you may perform. If you aid Jane, then you are aiding someone at least as needy.

This discussion helps us, I think, to understand a somewhat difficult passage in Anscombe’s classic paper “Who is Wronged?”²⁹ There Anscombe is considering a case of saving either five people or a sixth different person. She begins by saying: “Suppose I am the doctor, and I don’t use the drug at all,” saving no one. “Whom do I wrong? None of them can say: ‘you owed it to me.’… [I]f one can say that, all can; but if I used it, I let one at least go without and he can’t say I owed it to him.”³⁰ That seems to suggest that Anscombe takes the position that we have been questioning in this section: namely, that you owe it neither to Mary nor to Jane to save either Mary or Jane. After all, changing the numbers from five versus one to one versus one, Anscombe might have written: “Neither Mary nor Jane can say: ‘you owed it to me.’ If one can say that, both can; but if I used it, I let one go without and she can’t say I owed it to her.”

Puzzlingly, however, Anscombe then goes on to suggest the position that we have been defending in this section: namely, that you owe it to Mary to save either Mary or Jane and that you owe it to Jane to save either Mary or Jane.

Yet all can reproach me if I gave it to none. It was there, ready to supply human need, and human need was not supplied. So any one of them can say: you ought to have used it to help us who needed it; and so all are wronged. But if it was used for someone, as much as he needed it to keep him alive, no one has any ground for accusing me of having wronged himself. Why, just because he was one of the five who could have been saved, is he wronged in not being saved, if someone is supplied with it who needed it? What is his claim, except the claim that what was needed go to him rather than be wasted. But it was not wasted. So he was not wronged.³¹

The most natural way to solve the interpretive puzzle is to distinguish between owing it to Mary to save Mary and owing it to Mary to save Mary unless you save someone at least as needy. What Anscombe is asserting in writing, “None of them can say ‘you owed it to me,’” is that none can say that you owed it to them to save them, period. But this is consistent with asserting that each of them can say, “you owed it to me to save me unless you saved someone at least as needy.” After all, that is the claim that, according to Anscombe, each had: namely, “the claim that what was needed go to him rather than be wasted.” And so, when none is aided, “all are wronged.” Properly interpreted, Anscombe’s remarks anticipate the point that, as we have been arguing in this section, our bridge principle implies: that discretionary duties of aid are owed to every person who stands to be aided.

VII. Implications: Conjunctive Duties Owed to No One

Even if Bridge, Final does not imply that discretionary duties of aid are owed to no one, nevertheless it implies that other contractualist duties—that is, duties with a contractualist justification—are owed to no one. One class of such contractualist, but not relational, duties are “conjunctive duties,” which I discuss in this section. Another class of such contractualist nonrelational duties are “nonspecifically justified duties,” which I introduce in the following section.

Suppose, to use a colorless example, that you are confronted with three buttons. Pressing button A will save Mary, but leave Jane to perish. Pressing button B will save Jane, but leave Mary to perish. Pressing button C will save both Mary and Jane. Clearly, you are morally required to press button C. Moreover, contractualism implies as much, at least if we accept:

P5: any principle, P, permitting one to fail to aid both a patient in need and someone else in need (when one could do so without excessive cost to oneself, when no further exceptions obtain), could be rejected by an occupant of some standpoint, S, namely that of a person in need, on the grounds that the general acceptance of P would affect or treat that occupant by leaving them without aid.

I stress that the conjunctive moral requirement to aid both Mary and Jane—to press button C—is directly implied, in light of P5, by contractualism. It is not the result of applying an agglomeration principle, which might be questioned on independent grounds, to a contractualist duty to aid Mary and a contractualist duty to aid Jane.

However, on reflection, it is not the case that you owe it to Mary to press button C. After all, if you instead press button A, which also saves Mary, you don’t fail to do what you owe it to Mary to do. And, indeed, Bridge, Final implies that you do not owe it to Mary to press button C: to save both Mary and Jane. Granted, clause (i) would be satisfied, because of P5. Granted, clause (ii) would be satisfied, since every instance of not pressing button C (in the imagined circumstances) is an instance of not aiding both Mary and someone else in need (when one could do so without excessive cost to oneself), no matter what other actions you perform. However, clause (iii) would not be satisfied. It is not the case that every instance of not pressing C brings it about that Mary is left without aid whatever other actions you perform. If you press button A, then Mary also enjoys the aid. So, it is not the case that you owe it to Mary to press button C. What you owe Mary is either to press button A or to press button C. Likewise you do not owe it to Jane to press button C. Instead, you owe it to Jane either to press button B or to press button C. If you press button B, then you do not fail to do what you owe it to Jane to do. And if you don’t owe it to Mary or Jane to press button C, then to whom do you owe it? The answer would seem to be: no one.

VIII. Implications: Nonspecifically Justified Duties Owed to No One

So, you are morally required to press button C, to save both Mary and Jane, but you don’t owe it to anyone to do so. Perhaps this problem of conjunctive duties is not so troubling, though, since it remains the case that, if you fail to press button C, then you fail to do what you owe it to someone to do. By failing to press button C—that is, by either pressing A or pressing B instead of C—you either fail to do what you owe it to Mary to do—which is either to press A or to press C—or you fail to do what you owe it to Jane to do—which is either to press B or to press C. There is, however, another class of contractualist but nonrelational duties of which the same cannot be said: nonspecifically justified duties.

These duties arise from the fact that contractualism tests whether a particular action is wrong by considering what is entailed by the general acceptance of a principle permitting actions of that kind. Contractualism thus shares a structure with rule consequentialism and other “two-level” views. This makes possible a nonspecifically justified prohibition, which:

• 1.  prohibits people from performing actions of type X,

• 2.  because any principle permitting people to perform actions of type X could be reasonably rejected, on the grounds of what the general acceptance of the principle would entail,

• 3.  but not solely on the grounds of the specific effects of an action x of type X: that is, not solely on the grounds of how x, taken in isolation, affects or treats people.

In other words, the effects of the general acceptance of a principle, over and above specific effects, must be taken into account in order to nonspecifically justify it.

Here are some illustrations of possible nonspecifically justified prohibitions. First, consider lying. It might well be that my lie, taken in isolation, would not set back your interest in being able to rely on the testimony of others or mislead you in any harmful way. Nevertheless, it might be said that the general acceptance of a principle permitting lies could be reasonably rejected, on the grounds that general acceptance of a principle permitting lies would affect people by making them unable to rely on the testimony of others. Here we would be making use of Scanlon’s observation that the “widespread performance of acts of a given kind can have very different effects from isolated individual instances.”³²

Second, consider privacy. Suppose that I can snoop on you in a way that you will never find out about and that won’t otherwise adversely affect you. Nevertheless, it might be said that the general acceptance of a rule permitting snooping would lead to anxiety and defensive burdens, such as inhibiting what we say on phone calls or put in personal files. As Scanlon writes, “Our need for privacy, for example, is not met simply because, as a matter of fact, other people do not listen in on our phone calls and go through our personal files. In order to have the benefits of privacy we need to have assurance that this will not happen, and this is something that general acceptance of a principle can provide.”³³ Moreover, Scanlon suggests, since the principles that we generally accept “constrain the reasons we may, or must, take into account, they can affect our relations with others and our view of ourselves.”³⁴ This is the case, he urges, with respect to privacy. “The fact that others recognize reasons to restrain themselves so that I may be free from observation and inquiry when I wish to be is important in defining my standing as an independent person who can enter into relations with others as an equal. If the principles we all accepted did not recognize these reasons, this would crucially alter my relations with other people, and even my view of myself.”³⁵

Third, consider bodily integrity. Suppose that you need a kidney, and I have one to spare. Suppose you could take it from me safely and painlessly, without anyone’s ever finding out. Isn’t your interest in the kidney—your reason for rejecting principles prohibiting taking it—more significant than my interest in it—my reason for rejecting principles permitting taking it? The situation might be said to be analogous to privacy. It might be said that the general acceptance of a rule permitting the theft of organs would lead to anxiety and defensive burdens. And it might also be said that the general acceptance of a rule permitting stealing organs would damage our conception of ourselves and our relations to others. As Scanlon writes, “Principles defining my distinctive rights over my own body—rights to say who can even touch it, let alone claim its parts for other purposes—are an even clearer example” than that of privacy, of how the acceptance of principles that we accept affect our relations with others and our view of ourselves.³⁶

Finally, consider property. If you have much more and I have much less, then why am I not permitted to take your property? It might be replied that the general acceptance of a rule permitting theft could be reasonably rejected on the grounds that David Hume famously cites in his discussion of the “artificial” virtue of justice.³⁷

Now recall:

Bridge, Final, Explicit   A owes it to B not to x iff and because:

• (i)  any principle, P, permitting people to X patients in C, could be rejected by an occupant of some standpoint, S, on the grounds that the general acceptance of P would affect or treat that occupant in manner M,

• (ii)  x is an instance of, or a type of action every instance of which is an instance of, X-ing B in C, whatever other actions A may perform and

• (iii)  x is an action that, or a type of action every instance of which, would bring it about that B is affected or treated in manner M, whatever other actions A may perform.

The difficulty is simply that clause (iii) requires that A’s x-ing have a specific effect in order A to owe it to B to x. Every instance of A’s x-ing must bring it about that B is affected or treated in manner M. Yet, when it comes to nonspecifically justified prohibitions, it is not my token action x that affects or treats people in manner M, but instead the general acceptance of a principle permitting actions of the type to which x belongs. So Bridge, Final implies that, where nonspecifically justified prohibitions of x-ing are concerned, I owe it to no one that I not x.

We can illustrate this with the nonspecifically justified prohibition on spying. Again, the adverse effects of the general acceptance of principle permitting spying are being subjected to anxiety, defensive burdens, and compromised conceptions of oneself and one’s relations to others. Suppose, as I know, I could read Jones’s diary in secret, which would not lead anyone to feel anxiety, or impose on anyone defensive burdens, or compromise their conception of themselves and their relationships to others. Then, while it is wrong to read Jones’s diary, I do not owe it to Jones or anyone else not to read Jones’s diary, because it is not the case that my reading Jones’s diary would affect anyone by causing them anxiety, imposing on them defensive burdens, or compromising their self-conceptions and relations with others.³⁸

In the case of the conjunctive duty to save Mary and Jane, the contractualist could say that, while you did not owe it to anyone to save Mary and Jane, nevertheless if you fail to save Mary and Jane, then you fail to do what you owe it to someone to do. But the contractualist cannot say the same about the nonspecifically justified prohibition on reading Jones’s diary. It is not the case that if I violate that prohibition, then I fail to do what I owe it to someone to do.

It is, or so it seems to me, a surprising and noteworthy result that not all contractualist prohibitions are relational. It means that contractualist morality is not quite, despite its billing, the morality of what we owe to each other. Part of contractualist morality lacks the special force of relational contractualist prohibitions, which depends on there being “an individual who is affected by our action [and who] has a reason for objecting to it that cannot be answered satisfactorily.” One might say that, in a sense, contractualist morality is more than the morality of what we owe to each other.³⁹

IX. Does Contractualism Capture the Intuitive Extension of Relational Morality?

This also leads to a further point, which will occupy us for the rest of the paper. There is also a sense in which contractualist morality may be less than the morality of what we owe to each other. What I mean by this is that if not all contractualist prohibitions are owed to others, then contractualism may have a harder time capturing the intuitive extension of relational morality, of what we owe to each other, than we might otherwise have thought. Again, intuitively, I owe it to Jones not to read his diary. If we assume, falsely, that all contractualist prohibitions are owed to others, then it will appear to us to be enough for contractualism to explain that fact—that I owe it to Jones not to read his diary—that contractualism generates a prohibition on my reading Jones’s diary. But if nonspecifically justified prohibitions are owed to no one, and if the prohibition on reading Jones’s diary is nonspecifically justified, then we have not explained why I owe it to Jones not to spy on him.

One might put the point, as concerns nonspecifically justified prohibitions, in terms of a tension. On the one hand, contractualism’s two-level structure, by allowing for nonspecifically justified prohibitions, appears to extend contractualism’s extensional reach. It promises to explain the wrongness of actions whose wrongness would otherwise remain difficult or impossible to account for. On the other hand, explaining the nonrelational wrongness of those actions is not enough. We also want to explain why they relationally wrong others. And nonspecifically justified prohibitions do not in fact explain that.

X. An Illustration: Promissory Obligation

This tension makes itself felt in Scanlon’s discussion of promises.⁴⁰ There he puts forward a principle of promissory obligation, “Principle F,” which has as a necessary condition that the promisor assure the promisee that they will x. Yet one might worry that requiring prior assurance for promissory obligation prevents us from promising just when we need it most. What if there are no grounds, independent of promissory obligation itself, for you to believe that I will x? Then it seems that there is no way for me to undertake a promissory obligation. And yet we might both benefit from my undertaking a promissory obligation. Suppose we are Hume’s farmers.⁴¹ We would both benefit if you could help me with my harvest this week in return for my helping you next week. But how can I assure you that I will help you then if you help me now? It would be nice if I could create such assurance by promising you that I will help you then if you help me now. But according to Principle F, I need to be able to create such assurance independently of promissory obligation, since I do not incur a promissory obligation until I have created such assurance. So, unless I can appeal to another kind of motivation for helping you then if you help me now, we both suffer from an inability to cooperate.

To be sure, as Wallace and I emphasized in a comment on Scanlon on promising, a contractualist can grant that there are promissory obligations that can be triggered without prior assurance, at least in the presence of a convention.⁴² Such convention-based promissory obligations derive from the principle of fairness, which roughly requires us to do our fair share in conventions from which we benefit, on the assumptions that we benefit from the convention of promising and that keeping our promises counts as doing our fair share.

However, once the two-level structure of contractualism is squarely in view, a third possibility presents itself: namely, a kind of promissory obligation that would require neither assurance nor conventions. After all, we would benefit from the general acceptance of an “ex nihilo” promissory principle, which would require us to keep promises, whether or not the promisee was assured of performance, and whether or not breaking such promises would unfairly exploit an existing convention. The general acceptance of this ex nihilo principle would see us past the impasse facing Hume’s farmers, even where there was no convention. If you knew that I was party to the general acceptance of the ex nihilo principle, then, when I promised to help you then if you help me now, you would be assured that I would help you. Our cooperative endeavor could go forward. Can’t a principle permitting the breach of ex nihilo promises be reasonably rejected, simply on the grounds that its general acceptance would deprive us of these benefits?

While this ex nihilo principle would seem to be directly implied by contractualism’s two-level structure, however, the promissory obligations that it implies would not be owed to anyone. Consider a case in which I promise to x, but in the absence of a convention or any expectation on your part that I will x. Secretly breaking my promise, while violating the ex nihilo principle, does not in fact affect or treat you, or anyone else, in a way that would ground rejection of principle permitting what the ex nihilo principle forbids. So I do not wrong you, or anyone else, by breaking it. By contrast, insofar as the promissory obligations that derive from Principle F require prior assurance, it is guaranteed that breaching such an obligation, by undermining that assurance, affects or treats the promisee in a way that would ground the rejection of a principle permitting such breaches. So the promisor wrongs the promisee by breaching such a promissory obligation. And insofar as the promissory obligations that derive from the principle of fairness require that others have contributed to an existing convention of promising, it is guaranteed that breaching such an obligation, by unfairly exploiting those other contributors, affects or treats them in a way that would ground the rejection of a principle permitting such breaches. So the promisor wrongs, if not the promisee, then at least those other contributors by breaching such a promissory obligation.

Put another way, if the principle of promissory obligation is, like the ex nihilo principle, a nonspecifically justified prohibition, then we cannot explain how it is owed to the promisee or anyone else. If, however, the principle of promissory obligation is, as when it is derived from Principle F or the principle of fairness, a specifically justified prohibition, then we can explain how it is owed to someone. This is one instance of the broader tension that we highlighted in previous section. On the one hand, contractualism’s two-level structure, by allowing for nonspecifically justified prohibitions, extends the range of actions that can be classified as wrong. On the other hand, violating these nonspecifically justified prohibitions wrongs no one.

XI. Are Nonspecific Justifications Otiose?

This suggests that a contractualist prohibition can be part of what we owe to each other only if the prohibition has a specific justification. But if there is a specific justification for every prohibition, then aren’t nonspecific justifications of prohibitions otiose?

If nonspecific justifications of prohibitions were otiose, one might then conclude there remains no reason to frame contractualism in terms of the general acceptance of principles at all. There remains no rationale for its two-level structure. But this would be too quick. For even if there are no nonspecifically justified prohibitions, there might nevertheless be nonspecifically justified permissions. For example, the general acceptance of principles requiring us to treat others with perfect impartiality might prevent us from having the patterns of attitudes that are constitutive of love and friendship.⁴³ So while accepting that nonspecific justifications of prohibitions were otiose would vitiate much of the rationale for contractualism’s two-level structure, it would not vitiate all of that rationale.

Nevertheless, Scanlon himself clearly does not believe that nonspecific justifications of prohibitions are otiose, since he offers such justifications when discussing privacy and bodily integrity. But what might nonspecific justifications contribute, if there are already specific justifications?

One possibility, perhaps, is that the nonspecific justification of a given prohibition might help the necessary specific justification to be sufficient for reasonably rejecting alternatives. Suppose that we have some specific justification for prohibiting, say, spying. Nevertheless, it may be insufficient, on its own, in competition with the reasons that others have in favor of a permission to spy. If, however, others themselves have nonspecific grounds not to want the general acceptance of a principle permitting spying—such as the anxiety and defensive burdens that they would suffer—then their reasons, overall, in favor of a permission to spy are weaker. Perhaps they are then so weak as to make the specific justification for prohibiting spying sufficient. We might call this sort of specific justification of a prohibition—which is too weak on its own, but sufficient when supplemented with a nonspecific justification—a “supplementable” specific justification. If nonspecific justifications of prohibitions are not otiose, it would appear, it is because there are supplementable specific justifications.

There remain of course the questions of whether there is a specific justification for every intuitive prohibition, and if so, whether these specific justifications are supplementable. In the case of organ theft, for example, perhaps there is just a basic interest in control over others’ accessing one’s body. If so, however, then that interest might be strong enough on its own to justify a prohibition on organ theft. That is, it seems as though it might be a specific justification, but not a supplementable specific justification, which leaves something for a nonspecific justification to contribute. I leave open whether this is so. I also leave open whether specific justifications are to be found with respect to other prohibitions, such as those on lying, spying, and theft, and if so, whether these specific justifications are supplementable, so that they leave something for a nonspecific justification to contribute. What we can say conclusively is that, if contractualism is to be an account of what we owe to each other, then it cannot avail itself of the full explanatory resources that might at first seem to be made available by its two-level character: by its testing principles on the basis of what their general acceptance would entail. Nonspecific justifications of prohibitions can, perhaps, supplement specific justifications, but they cannot extend the extensional reach of contractualism to cases where there are no specific justifications to supplement.

XII. Conclusion

The overarching question of this paper has been whether contractualism provides an account of what we owe to each other. We labored to articulate a bridge principle, which would bridge the gap between the claim that actions of type X could be reasonably rejected and the claim that some particular person, A, owes it to some particular person, B, that A not perform some particular action, x. We then explored the implications of this bridge principle. To begin with, the bridge principle implied that contractualism cannot support the idea that the fact that A has a duty not to x is explained by or just is the fact that A owes it to some B not to x. It also implied that contractualism justifies duties that are not relational, that are owed to no one. While this might not be true of discretionary duties of aid (which, in fact, are owed to everyone who stands to be aided), it is true of conjunctive duties and, more importantly, nonspecifically justified duties. Consequently, contractualism may have a harder time accounting for the intuitive extension of what we owe to each other than might otherwise be thought. If we are seeking to explain what we owe to each other, it is no help to appeal to prohibitions that are not owed to anyone.

Now, perhaps this is mistaken. In particular, perhaps I have overlooked ways in which contractualism can account for what we owe to each other, while retaining the full explanatory resources of its two-level structure. And perhaps this is because I have not articulated the correct bridge principle. I would welcome this, since I find both contractualism and the Thesis highly attractive. All the same, as I hope to have shown, more must be said to establish the Thesis. And if this much is so, then this paper contributes something. Its most fundamental point is that, while it might seem a matter of definition that contractualism is an account of what we owe to each other, there is a real question whether it can be.

Acknowledgements

This paper grew out of notes for a graduate seminar on the structure of wronging at the University of California, Berkeley in the spring of 2025. I am grateful to participants in that seminar for discussion. I am also indebted to comments on an earlier version of this paper from R. Jay Wallace and Erik Zhang, as well as to discussion of this paper in a workshop at the Australian National University, where Nic Southwood gave prepared comments. I thank Selim Berker, a participant at that workshop, for alerting me to the papers by Salomon and Webber that I cite. I am greatly indebted for the extraordinary patience and constructive comments of two anonymous reviewers for Free and Equal, which saved me from a number of serious mistakes and vastly improved my understanding of the underlying issues. Some last-minute comments from Wallace saved me from further errors, as did the excellent copyediting of Atticus Carnell, Isabel Fulda Graue and Amanda Lopatin.

Competing interests

The author has no competing interests to declare.