IN ETHICS AND THE LIMITS OF PHILOSOPHY, Bernard Williams writes, “There is an everyday notion of obligation, as one consideration among others, and it is ethically useful.” He also writes, “It is a mistake of morality to try to make everything into obligations.”1 These remarks bear on the relational interpretation of morality that I have been developing in this book. The obligations in the everyday sense to which Williams refers include a range of apparent duties, but the most familiar of them are the obligations created by promises.2 These are obligations that are owed to another person, which we register in reflection as presumptive constraints on our ongoing planning about what to do. The relational interpretation does indeed try “to make everything into obligations” in this core sense. As noted in section 2.3 above, it understands modern morality to consist of a set of directed obligations that derive from our membership in a notional community of moral equals, and that are owed specifically to the individual members of this community. Actions are morally right, on the resulting approach, just in case they are ones that others have a claim against us to perform, and the set of potential claim-holders in this sense is maximally inclusive.
Williams had reasons of his own for thinking that it is a mistake for morality to make everything into obligations in the “everyday” sense. I do not propose to go into those reasons here, at least not directly. But I think there are some important concerns that are at least related to Williams’s skepticism about whether it is really plausible to extend the original notion of directed obligation to the entirety of the moral domain, as the relational interpretation proposes to do. I develop these concerns in the first two sections of this chapter, observing that some of the most salient examples of relational requirements grow out of nonnormative relations or interactions between the individuals who are bound by them, and that morality first emerged in human development as a device to facilitate cooperation in contexts of this kind.
These observations raise the natural question whether the relational model can be extended into a general framework for thinking about moral obligation, even outside of contexts of personal interaction and in-group cooperation. I argue that any account of interpersonal morality will involve some degree of abstraction and extension as we adapt moral resources to do justice to the modern idea that we are members of an extensive community of moral equals. Drawing on the relational model for this purpose has the signal advantage that it enables us to understand interpersonal morality as a source of obligations that also have direct significance for our accountability practices. Philosophers have offered some general arguments against extending the relational morality in this way, but I show that those arguments do not succeed. Morality may be understood as a set of self-standing relational obligations, ones that can link individuals who are not already connected through ties of history or causality or cultural practice.
Only once we recognize that moral obligations take this form do some of their presuppositions and consequences come into clear focus. I argue in section 4.3 that these include an anti-individualism about normativity. On many popular approaches, reasons for action and for attitudes are thought to be contingent on subjective facts about the states and capacities of the person to whom they apply. This kind of individualism, I suggest, is hard to square with the idea that there is a genuine moral nexus that links agents and claimholders through directed requirements and their corresponding claims. But it is equally hard to reconcile with the central features of interpersonal morality that the relational account seems well designed to accommodate.
The self-standing character of the moral nexus raises a further question about the agent-relative character of the duties that it defines. A natural strategy for explaining the agent-relative structure of moral requirements appeals to the nonnormative patterns of historical interaction on which they are based; but on the approach I am sketching, many directed moral obligations obtain independently of any such relational basis. In section 4.4, I suggest that the agent-relative character of moral obligations is explained, at the deepest level, by the relational character of the claims with which they are connected, and that this explanation carries over to cases in which there is no antecedent relationship between the parties. I also argue that there is a positive personal value that is realized through compliance with relational moral requirements, which is connected to, though distinct from, the value of mutual recognition.
The paradigm cases of directed or relational obligation that have so far been considered are ones that have a transactional character. The central example has been that of promissory obligation, which arises through a concrete, datable interaction between the parties whom it binds. The promisor undertakes an obligation by making a representation of some kind about what will happen in the future, one that is registered by the promisee, and that thereby gives rise to a new directed duty. Promisors come in this way to owe it to promisees to fulfill the expectations that they have attempted to raise, and these directed duties correspond to claims on the promisees’ part to such fulfillment.
The relational obligation, in this paradigm case, thus derives from a social exchange between the parties that can rightly be regarded as its basis. If we wish to extend this relational model to the whole of the moral domain, however, we will have to construe it as applying in cases where the moral nexus is not similarly grounded in a prior transaction or interaction between the parties that it links. According to the modern understanding of it that I sketched in chapter 2, morality consists in a set of directed duties that derive from the fact that we inhabit a world in common with other agents whose interests also matter, and matter equally. Its requirements correspond to claims that people have against us, just insofar as they are persons or agents. The class of people to whom moral duties are owed, then, will include vast numbers of individuals whom we have never even met, and with whom we therefore cannot have engaged in a transaction or exchange. The random person I encounter on the public sidewalk with a gouty toe is someone who has a moral claim against me, on this conception, that I not act in ways that aggravate her suffering. So do the inhabitants of a remote foreign region whose territory I am flying over as I contemplate jettisoning some excess cargo or waste. The question is whether there can be a relational nexus of this kind that is not grounded in a prior relationship between the parties, such as the transactional relationship that gives rise to a standard promissory duty.
The case of promissory obligation raises complications, however, that the discussion so far has glossed over. A brief review of some of these complications will help us to understand better why it might seem a mistake to try to generalize the relational model to the entirety of the moral domain.
According to a common and to my mind plausible way of thinking about them, promises function to generate expectations about the behavior of promisors via the awareness of the parties to them that a distinctively moral obligation has been incurred. That is, the promissory transaction generates a moral obligation, and the promisee’s assurance that promisor will perform is parasitic on the understanding that an obligation of this kind now exists, and that the promisor is morally conscientious.3 There are, to be sure, puzzles that this way of thinking about promising tends to generate. A particular difficulty confronts those who believe that the moral obligations involved in promissory exchange derive from more general principles of assurance, pertaining to cases in which people have come to rely on us to do something as a result of our having deliberately led them to expect that we will so act. For if promises give rise to assurance on the part of the promisee that the promisor will perform via the promisee’s awareness that a moral obligation has been incurred, it seems the obligation cannot derive from the fact of the promisee’s assurance. The approach appears to be circular, presupposing the existence of the moral obligation in its account of the state of mind on the part of the promisee that is supposed to give rise to the obligation in the first place.4
Putting this difficulty to the side, the moralized approach to promises assumes that promissory obligation is a special case of moral obligation. That is, there are general moral obligations that we stand under, specified by principles such as those of assurance and of fairness (pertaining to our responsibility to contribute our fair share to sustaining useful social practices from which we have ourselves benefited). These obligations apply to us when we deliberately raise expectations in other people about what we are going to do or exploit valuable social conventions to advance our ends. Making promises turns out to be a way—though not, perhaps, the only way—of incurring obligations of these more generic kinds; and the obligations that are generated are therefore distinctively moral in character.
According to a different and less moralized conception, promissory obligation is sui generis. That is, promises give rise to obligations, but it is not characteristic of those obligations that they are distinctively moral. They are, as we might put it, practice-based obligations, whose nature is connected with and defined by the specific social conventions with which promising is bound up. Proponents of this conception may agree that promises function to generate expectations on the part of promisees through their awareness that the promisor has incurred an obligation in virtue of the promissory exchange. But they will deny that morality has anything essential to do with the obligation that is involved in this process. Promising is, on this way of thinking about it, a social practice that confers on its participants a sui generis normative power, and that thereby subserves their normative interest in being able to generate new directed claims and obligations.5 There might be, and presumably there generally are, distinctively moral reasons why people ought to keep their promises, and these may even be sufficient to ground moral obligations in many cases. These moral considerations supervene, however, on a prior and independent set of practice-based obligations and claims, which represent the primary level on which promissory obligation operates.
But what story can we tell about the directional features that intuitively seem to characterize promissory obligations? How can we make sense of the relational character of such obligations? On the nonmoralized account sketched above, as I understand it, the relational features will be traced to the structure of the social practice that is said to confer normative powers on those who participate in it.6 That is, the practice just is one in which promises give rise to obligations that are understood to be owed specifically to the promisee, who has a practice-based claim against the promisor to fulfillment of the obligation that the promise generates. The practice represents, in the phrase of Michael Thompson, a self-standing order of right, which defines the manifold of persons who participate in it as potential parties to relational requirements of the now familiar kind.7
According to the moralized approach, by contrast, the relational obligations that promises create are specifically moral obligations. So if we are to make sense of them as directed obligations, owed specifically to the promisee, we will need a way of understanding how moral requirements can have this kind of character. Here, I think, it is tempting to suppose that what makes a relational moral obligation possible will be something about the antecedent relationship that the parties to the obligation stand in. One possibility would be to interpret morality as a self-standing social practice that can inscribe directed obligations and claims, in a way that parallels the practice of promising. Moral principles, such as those of assurance and fairness, might be construed as norms that some historically situated group of people already implicitly internalizes and accepts, perhaps in the style of a Humean convention. In that case, promisor and promisee will already be linked with each other, prior to entering into their promissory agreement, through their joint participation in the social practice of morality, which makes them both members of the same Thompson-style manifold of persons, and therefore potential parties to relational bonds that are constituted through this practice.
This strategy falls back on the practice-based account of directional obligation, and simply applies it to morality itself, construing the moral domain as a distinctive set of conventional arrangements. One curious result of this approach will be that there are two separate practice-based obligations that can be brought into existence through promises: a specifically promissory obligation that is prior to and independent of morality, and a separate and equally practice-based moral obligation that piggybacks on the first. A second consequence of the approach is that morality comes to have a somewhat provincial character; it will represent a set of conventional norms or understandings that implicitly regulate the interactions of a group of people in the way of a social convention, rather than universal requirements that stem from the fact that there are other persons or agents in the world who are equally real, and whose interests matter to practical thought (even if they are not participants in our social practices or conventions in particular).8
A different way to make sense of the directed moral obligation in the promising case would point to the transaction between the parties that the promise itself represents. According to this line of thought, there might well be universal moral obligations that are not based in the norms that are shared by the participants in a common social practice. All the same, morality can generate directed obligations only as a result of a specific causal interaction or exchange between the parties who are thereby linked to each other, such as the issuing of a promise that is registered as such by the promisee. The role of the transaction, on this approach, is to render intelligible the idea that there is something specific that one of the parties owes it to the other to do, morally speaking. There is a pattern of causal interaction between promisor and promisee that brings into existence the nested set of duties and claims characteristic of relational normativity, and that is necessary for a specific normative nexus of this kind to obtain.
Normatively significant patterns of causal interaction might also be present in other cases, such as those that link the parties to a relationship between friends or family members, and they can similarly give rise to directed moral obligations and claims in those cases as well. Thus, it is natural to suppose that friends and romantic partners have special moral obligations to each other, and corresponding claims against each other, in virtue of the relationships that they stand in to each other, which among other things involve extended patterns of historical interaction.9 And there are still other forms of historical relationship that involve less deep and extended patterns of interaction, such as those that arise when one person invests trust in another or makes a significant sacrifice to benefit the other in some way. In these cases, too, it is natural to suppose that there are directed moral obligations in play, including obligations of gratitude, reciprocity, and loyalty.
On this way of thinking about things, individual persons or agents can in principle be potential parties to directed moral duties to each other, just in virtue of their common standing as persons or agents. There is no need for them to share some antecedent basis, such as participation in a common social practice, in order for them to be members of the same manifold of persons, and hence eligible to have moral claims against each other. But actual duties and claims of this kind can arise only through particular interactions between individuals, which generate patterns of causal influence that provide a concrete basis for the relational nexus that comes to bind them to each other. Duties that we owe to each other, as Joseph Raz has written, thus seem to “presuppose some kind of personal relationship.”10
Both of the views sketched above pose a challenge to the relational account of morality that I have been attempting to explicate. Either there are relational obligations of morality, but only ones that extend to participants in a common social practice or convention; the result would be that there is nothing that we owe, morally, to outsiders to our own social practices who nevertheless stand to be affected by what we do (such as the inhabitants of the region we are flying over as we contemplate jettisoning our aircraft’s waste). Or there are universal moral requirements, but ones that acquire a specifically relational character only in cases in which there is a prior interaction between the parties that grounds the moral obligation through which they are yoked together. Here, too, we will have no moral obligations that are owed specifically to individuals who could be disadvantaged by our actions, but without our having previously interacted with them in one way or another (such as the unfortunate stranger on the sidewalk who is suffering from gout). On either view, it will be, as Williams suggested, a mistake to interpret interpersonal morality as a unified domain of directed obligations.
To understand morality in this way is to extend the relational paradigm from core cases in which it familiarly operates, such as the obligations created by promises and family relationships, to cases that are perhaps less naturally interpreted in relational terms. This extension will involve the application of the relational model to situations that lack some of the features that are present in the core cases, and it is fair to ask whether the model remains tenable when it is applied in this way. Can there be directed obligations and claims between individuals who are not antecedently related to each other, via bonds of friendship or personal history or participation in common institutions or conventional practices? And can there be specific obligations that are owed to other people that do not track and supervene upon antecedent patterns of causal interaction?
These skeptical questions grow out of reflection on some of the central cases of directed moral obligations; but a similar line of thought is encouraged by reflection on the natural history of morality. Consider, for instance, Michael Tomasello’s interesting account of the evolution of our moral capacities.11 This account posits several distinct sources of moral agency that come to the fore sequentially, in response to the distinctive challenges that our ancestors encountered in different phases in the development of the species. There is, first, a prosocial tendency to sympathy and helping behavior, especially toward kin and friends, that is essentially continuous with similar tendencies in the great apes. Second, Tomasello identifies a distinct capacity for a kind of “second-personal” morality that emerged when our ancestors were forced to enter into collaborative foraging activities with other individuals. This involves a capacity to see oneself as engaged in a joint project together with other individuals, to understand oneself as subject to role requirements that are defined by the collaborative project, and to feel a sense of commitment to the individuals with whom one is engaged in a joint pursuit. Finally, Tomasello posits a move from interpersonal to objective morality, which involves a “scaling up from collaboration to culture”;12 here humans develop a sense of themselves as members of larger collectives or tribal units, and see themselves as subject to objective requirements that are partly constitutive of the groups with which they are identified.
On Tomasello’s account, then, human morality involves distinctive patterns of thought and motivation, ones that are responsive to the forms of interdependence characteristic of our species life. A similar idea figures in the developmental story that Joshua Greene has recently presented.13 Greene argues that morality can be understood, in the first instance, as a set of emotional dispositions that evolved to solve the problem of coordination within groups. Tendencies such as empathy, family attachment, gratitude, anger, and the other moral emotions implement different but complementary strategies for avoiding the familiar dilemmas of cooperation between individuals who interact with each other within a common social framework. These emotional dispositions involve swift and unreflective, “System 1”–level action tendencies that are effective under the circumstances for which they evolved, enabling us reliably to avoid mutually destructive conflict within our local communities.14 But these same tendencies promote a sense of tribal belonging and a correlative awareness of our differences from other tribal groups.
The evolutionary solution to the problem of cooperation within groups thus leads to a new problem of competition between groups, each of whose members is prone to favor other in-group individuals. Tribal membership is constituted in part through shared moral understandings, which help group members to navigate their local world, but which also put them at odds with members of other groups, organized as they are around different thick conceptions of right and wrong. The most pressing moral problems of the present age are raised by these intergroup conflicts. To resolve them, we need to develop a new “metamorality” that will be capable of universalistic application, and that may command the assent of individuals who belong to different moral tribes. Greene goes on to argue that solving this problem must fall to our reflective “System 2” processes, and that its terms are set by some version of utilitarian moral reasoning.
These two historical narratives differ on several important points. Greene, for instance, questions the authority of the emotionally tinged reactions that tend to enable cooperation within groups, which he thinks should not be viewed as reliable guides for reflection about the intergroup conflicts that loom so large today.15 Tomasello, by contrast, tends to grant the continuing legitimacy of the “second-personal” morality that enables cooperation between individuals in contexts of joint agency, suggesting that it is supplemented rather than supplanted by the forms of reasoning that enable more cosmopolitan forms of cultural life.16 Abstracting from these differences, however, both accounts emphasize the role that morality initially plays in facilitating mutually beneficial collaboration between people who stand in close, face-to-face relations with each other. From the standpoint of the natural history of the species, this appears to be the primal scene for many of the forms of reasoning and emotional response that we continue to recognize as characteristically moral.
At the same time, these developmental accounts raise an important question: what moral resources can we draw upon as we extend our horizon, as we must do, to encompass our relations to persons with whom we have not interacted in an “up close and personal” manner? For Tomasello, this extension involves the development of new forms of moral reasoning that are centered around objective norms, and that rest on our conception of ourselves as members of larger (but still parochial) groups. But this enrichment of our moral resources does not yet get us all the way to a genuinely universal framework for thinking about what to do (for instance, in cases of intergroup conflict). Greene is especially concerned about this problem. To cope with it, he contends that we need to leave behind the emotional tendencies that evolved to regulate our in-group relations, and deploy the reflective resources of utilitarian metamorality.
The challenge, in the present context, is this. Relational moral requirements, as I have explicated them, seem tailor-made for the kinds of face-to-face encounters that shaped the earliest stages in the evolution of human morality. As I argued in the last chapter, they are also closely associated with the reactive attitudes that might have evolved to help us cooperate with members of our own groups. Thus, our paradigms of directed obligation, such as those generated by promises, are cases that rest on concrete interactions between two individuals, and that give rise to resentment and reactive blame when the obligations in question are flouted. Perhaps it is reasonable to grant the legitimacy of the relational model in application to cases of this kind.17 But there is a question about whether the model can be generalized into a plausible account of modern, universalistic morality in its entirety. It cannot be taken for granted that patterns of reasoning and emotional response that are suited to contexts of close interpersonal interaction will be adequate for thinking about our responsibilities in more impersonal contexts, where we are called on to consider the implications of our conduct for people to whom we are not antecedently related in any particular way. Moral principles that function well in situations structured by close interpersonal exchange might simply not be up to the task of regulating our actions in other situations.
This line of inquiry dovetails with the questions posed earlier in this section, on the basis of reflection on the mechanics of promissory obligation. In this paradigm case, it seems that directed obligations are anchored in something that antecedently links to each other the parties whom they bind. This might be a Humean convention in which the two parties both participate (something that would connect them as members of the same tribal or cultural in-group); or it might be a personal transaction between them (the result of a face-to-face interaction on some particular occasion). But it is then natural to wonder whether this feature of the promissory case represents something that is essential to directed obligation more generally, as Raz for instance has suggested that it is. Can there be directed obligations that float free from the personal relationships and social bonds that link us antecedently to other individuals, and that might be understood to underwrite the assignment to those individuals of specific claims against us?
The questions posed at the end of the preceding section will to some extent occupy us throughout the remainder of this book. The main aim will be to articulate and defend the assumptions that need to be taken on board when the relational model is expanded into a comprehensive interpretation of the domain of modern interpersonal morality, and to consider whether it provides an illuminating framework for thinking about first-order moral issues. I shall try to identify the most important theoretical commitments of this approach, and I shall explore its implications for several challenging questions within normative ethics.
Throughout this discussion, however, it will be important to bear in mind the conclusions established in chapters 2 and 3 of this book. I argued there that the relational approach is unique in its ability to make sense of central features of interpersonal morality, understood as a distinctive and internally unified normative domain. In particular, the approach explains the idea that moral considerations have the force of obligations, ones that have their source in the equal standing of an inclusive domain of persons. It also renders intelligible the social significance of moral requirements, their standing as a basis for relations of interpersonal accountability. These are core elements in our modern conception of the moral realm, and it is a great attraction of the relational model that it promises to illuminate them. The presumptive rationale for trying to “make everything into obligations” in the relational sense is the promise it holds of enabling us to understand interpersonal morality as a unified set of considerations that reflect the equal standing of persons, and that are intelligible both as defeasible constraints on the will and as a basis for relations of accountability.
Against this background, it strikes me as natural to turn the questions posed in the preceding section around. Why should we not interpret the domain of interpersonal morality as having an implicitly relational structure? Understanding it in this way will, to be sure, involve an extension of the relational model from the contexts in which we first encounter it in practice to situations that are somewhat different in character. But any attempt to make sense of morality, as we typically understand it today, will involve some extension of this kind. A suitably modern conception of the moral must be cosmopolitan in spirit, acknowledging what I referred to in chapter 1 as the postulate of equal standing. The basic moral insight of modernity is that we are all persons, equally real, and that there is an important class of interpersonal requirements that stem directly from this important fact. But doing justice to this cosmopolitan insight necessarily requires an exercise of reflective adaptation, through which more primitive moral commitments are modified to reflect the fact that we belong to a maximally inclusive community of moral equals. We must move beyond the parochial and limited parameters within which we initially learn to operate with moral concepts in both our individual development and in the natural history of the species.
Utilitarians such as Greene and Peter Singer, for instance, emphasize the need to extend natural empathy to encompass the entire domain of sentient creatures, through the exercise of System 2 capacities for critical reflection.18 But prosocial empathy is only one among several tendencies that have emerged in our evolutionary history to promote moral behaviors and to facilitate cooperation. There is, in addition, the ability to see oneself as subject to directional requirements that specify duties that are owed to other individuals. This is an ability that plausibly emerges in situations that involve close interpersonal contact, including contexts in which we are trying to do things together with other people. But these abilities too are capable of being extended to encompass a much broader domain of individuals. We can come, through reflection, to see others as having claims against us, just insofar as they are persons who are equally real, and regardless of whether we are already connected to them through personal relationships or shared social practices of some kind. Similarly, though we initially understand ourselves to owe specific debts to people with whom we have already interacted in some way, it is perfectly coherent to think that there might also be debts we owe to others, just insofar as they are persons, equally real. Understanding interpersonal morality along these lines has the signal advantage that it enables us to make sense of it as a domain of practical requirements, ones that are at the same time suited to structure relations of interpersonal accountability; this was the argument of chapters 2 and 3 above.
As I noted earlier in this book, the postulate of equal standing is often thought to favor broadly consequentialist accounts of morality, which treat the interests of each as equally important in the determination of what it is morally right to do. Approaches such as those of Greene and Singer draw much of their appeal from the straightforward expression they give to the commitment to equality that is implicit in universal morality. But there are other interpretations of this commitment that are at least as appealing, and that can be detached from the specifically consequentialist interpretation of the requirements of the moral right.
Consider Samuel Scheffler’s compelling suggestion that the value of equality in the context of political life should be seen as a special case of a more general value, that of an egalitarian relationship.19 The hallmark of such a relationship, on Scheffler’s account, is that the parties to it acknowledge an egalitarian deliberative constraint; they take the interests of the different parties to the relationship to be equally significant to their deliberations within the context of the relationship.20 To conduct a relationship on these terms is not to be antecedently committed to realizing any particular pattern of outcomes (such as ones that equalize the welfare level of all parties to the relationship), but to participate in an ongoing interpersonal practice whereby the equal standing of the parties is reciprocally acknowledged. Scheffler believes that this egalitarian ideal can be realized in personal relationships of different kinds, including not only those between members of the same political community, but also the relationships between friends or family members or life partners. An egalitarian political practice will make different demands on members from the demands that egalitarianism makes on friends and intimates; but these differences will reflect differences in the character of the relationships that are to be conducted on a basis of equality, not differences in the ideal of equality that applies to them.21
Scheffler himself associates the egalitarian deliberative constraint with valuable relationships that have an ongoing, historical character, such as those between friends or between the members of a common political community. The assumption seems to be that egalitarian commitments get a grip on agents only in the context of existing relationships that are valuable along other dimensions.22 But while relationships of this kind might provide our initial exposure to the egalitarian deliberative constraint, the relational value that they instantiate is one that potentially has application outside the context of such relationships.
Thus, we can see ourselves as having directed obligations that are owed to individuals with whom we have not yet interacted at all, just in virtue of the fact that they stand to be affected in one way or another by the things we might do. As I have argued in chapter 2, directed obligations of this kind make sense to us as practical requirements, that is, considerations that intelligibly function as presumptive constraints on an agent’s deliberations. The constraints, moreover, appear to operationalize a commitment to equality, insofar as they acknowledge the standing of persons as equally worthy of moral consideration. The interests of others count equally, in this context, not as considerations that are to be assigned equal weight in assessing the consequences of actions that the agent might perform, but as potential bases of moral claims that are held against the agent. The generalized relational model of obligation in this way seems to exemplify Scheffler’s egalitarian deliberative constraint in its most abstract form. It enables us to understand interpersonal morality as a set of cosmopolitan obligations that potentially ground accountability relations with anyone, realizing a practice of equality distinct from, though structurally similar to, the egalitarian practices that Scheffler situates within thicker human relationships.23 This is an additional respect in which it is attractive to understand morality as an extension of the relational model of interpersonal obligation that is familiar to us from situations of face-to-face interaction with other people.
According to the resulting interpretation of interpersonal morality, specific causal interactions with other people might affect what we owe to them, but they do not represent general preconditions for the possibility of directed moral duties and claims. If I have induced other people to form expectations about my future behavior, or benefited specifically from their significant efforts to help me out of a bind, then I will have obligations to them, of fidelity or gratitude, that would not otherwise obtain. It does not follow, however, that there is nothing that I owe to people with whom I have not previously interacted in these or other ways, or to whom I am not antecedently linked through shared relationships or practices of some kind. It might be a condition for the obtaining of directed moral obligations and claims that the parties connected by them are not causally isolated from each other, so that there is at least the possibility that one of them would be affected significantly by something that the other might do. Thus, it is arguable that there is nothing that we owe to the rational denizens of distant galaxies, assuming there are such beings, at least until such time as we are able to have a discernible impact on them through exertions of our individual and collective agency; this is an issue to which I shall return in chapter 5 below. But this condition, of course, does not require that we should already have interacted with those to whom moral duties are owed. The strangers whom I encounter on my travels have basic moral claims against me, even if we are meeting for the first time, insofar as they are in a position to be harmed and benefited in specific ways by various actions that it is open to me to perform. My sense of myself as owing them consideration is a particular way of articulating the idea that they and I are moral equals.
Given the coherence and initial attraction of this way of understanding interpersonal morality, are there general theoretical considerations that would stand in the way of it? Michael Thompson has argued that there are. As noted earlier, he takes relational or “bipolar” requirements to define an order of right that potentially links individuals within the manifold of persons to which it applies. But he contends that there is an explanatory challenge we face in making sense of how individuals can belong to one and the same manifold of persons, a challenge that universalistic conceptions of relational morality may not be able to meet.24
Thompson illustrates the general problem by asking us to consider members of two different tribes, the Lombards and the Schlombards, whose systems of personal private law include directed requirements with the same content, and who encounter each other for the first time at a remote location in the Alps.25 Thompson suggests that these individuals, taking each other to belong to the same tribe, might endeavor to enter into a contract with each other, concluding their attempted agreement with the ritual mixture of song and dance that is called for under the terms of their respective private law regimes. But this exercise will not succeed in generating contractual obligations that are genuinely owed to each other, since there is no single order of right through which the parties to the putative contract are linked. As Thompson writes, “They are like ships passing in a juridical night.”26
On Thompson’s view, it seems, there are two conditions that have to be satisfied in order for individuals to belong to the same manifold of persons under a common order of right. First, it must be the case that the individuals can come nonaccidentally to have thoughts about directed duties and correlative claims that are shifted into the same normative “gear,” invoking the very order of right that they both fall under.27 If they encounter each other on a remote highway for the first time, and swerve in their automotive trajectories so as not to cause harm to each other, this must normally be because they understand themselves to have valid claims against each other, and corresponding directed duties, that are of the same normative kind. Second, Thompson suggests that a system of nonaccidentally convergent relational thoughts is possible only if there is something else that the bearers of those thoughts have in common, which gives their shared dyadic thoughts a “foothold” or represents their “common source.”28
Appealing to these conditions, Thompson goes on to argue that it will be very challenging to make sense of the kind of cosmopolitan order of relational right that I have been trying to sketch in this book.29 The common source for convergent bipolar thoughts about morality might be sought in shared Humean conventions or (along Aristotelian lines) in a common human nature that binds the individual bearers of such thoughts. But then it will be lacking in full universality, failing to apply to agents who do not participate in our natural life form or share in our conventional arrangements.30 A more abstract conception of the manifold of persons could perhaps be constructed on Kantian grounds, by appeal to the shared principles of pure practical reason that we might take to be latent in any exercise of concept-governed self-determination, construed as a kind of “intelligible cause” of such agency. But Thompson views this as an “alarming” metaphysical commitment, one that it is difficult for even a mildly naturalistic conception of ethics to maintain.31 Without it, the allegedly universal manifold of persons who are united under common moral norms will lack the kind of shared basis that it requires.
But both of Thompson’s conditions strike me as questionable.32 First, I do not see that it is built into the notion of an order of relational right that the individuals to whom it applies should be capable of converging nonaccidentally on thoughts about it. Doubts about this assumption are already raised by some examples that Thompson himself considers, such as duties that we owe to infants or to members of our species who are mentally infirm.33 It isn’t obvious to me that directed obligations in these cases are unintelligible unless we can see the inability of the claimholders to grasp thoughts about them as accidental, in some sense. More fundamentally, however, it seems at least conceivable that there might be things that we owe to individuals who are not capable of entertaining normative thoughts of any sort, not even when they come to exhibit the full spectrum of capabilities that is normal for adult members of the species to which they belong. Our dogs and cats, for instance, might have claims against us that we look after their basic animal needs, and provide them with a comfortable and secure environment in which to live in accordance with their nature.
Thompson concedes that there might be prohibitions on how we may treat individuals of this kind, things that it would be wrong for us to do to them.34 But he maintains that we cannot wrong them, insofar as they lack in principle the capacity for normative judgments about the relational order of right that allegedly links us to them. But in the present context this assumption seems to me question-begging. There is nothing incoherent about the idea of a manifold of persons that tolerates at least some internal asymmetries, insofar as it assigns claims to individuals within the manifold who are not capable of asserting those claims on their own behalf, and who also do not stand under directed duties toward other members of the manifold. But if this idea is not incoherent, then Thompson’s first condition must be rejected. It is not an a priori constraint on an order of right that the manifold of persons who stand under it must normally or nonaccidentally share thoughts about the bipolar norms that link them with each other.
Even if I am wrong about this, however, it would not be enough to motivate Thompson’s second condition. He contends that individual members of a manifold can converge in bipolar thoughts about its requirements only if there is something else that binds them to each other, and that makes possible this alignment in their normative attitudes. But this seems deeply implausible.
Return to Thompson’s case of the Lombards and Schlombards. Individual members of these two tribes are, let us assume, capable of nonaccidentally entertaining bipolar thoughts about the private-law orders that prevail in their respective tribal societies. They can thus think about themselves as having contractual claims against other members of their tribe, as well as obligations that are owed to such individuals; this much Thompson himself clearly accepts. But if these resources are available to them, it seems they can easily, through a modest exercise of abstraction, step back from their tribal identity and think of themselves as individuals who are capable of bipolar normative thoughts of one kind or another. This in turn will open the way to their understanding themselves to be members of a larger manifold of persons whose defining feature is the capacity for bipolar normative judgment. They can then ask themselves about the directed duties and claims that link them to each other, just insofar as they are each members of this manifold. They would thus arrive at convergent judgments about what they owe to each other not as Lombards or Schlombards, but simply as individuals who are competent with bipolar concepts.35 (This is true even if their reflections lead to the conclusion that there is nothing they owe to each other in their common capacity as persons capable of bipolar thought; this negative conclusion would still be a judgment about the abstract order of moral right.) This is precisely the kind of reflective extension of our relational normative concepts that I envisaged at the start of the present section, whereby relational requirements that are originally brought to bear in face-to-face or tribal encounters come to be applied more universally to a much broader class of individual persons.
Against this, it might be suggested that the postulated abstract competence with bipolar normative concepts is something that the individual members of the larger manifold have in common, and that might therefore represent a “source” or “foothold” for the convergence in judgment that I have described. But to interpret Thompson’s second condition in this way would render it empty. As he understands the condition, it pushes us toward seeing the members of a manifold of persons as bearing substantive connections with each other that are independent from the bipolar norms that purport to link them. Examples include the historical and causal ties that bind together the individuals who participate in and benefit from a Humean convention or practice, or their joint membership in an Aristotelian natural life form, or their shared connection to pure practical reason, understood as a kind of intelligible cause. But individuals who exhibit competence with bipolar normative competence need not share any substantive connections of these kinds. They hold a property in common, but this represents a merely logical relation between them, not a thick social or natural-historical or metaphysical relationship that has independent explanatory significance (for instance, as something that might make sense of the fact that their bipolar thoughts nonaccidentally converge in their content).
A different basis for skepticism about the possibility of universalistic relational morality is suggested by some reflections of Margaret Gilbert’s on the nature of obligation.36 Gilbert notes that obligations, in a core sense, should be understood in what I have called relational terms, as owed to another party who has a special claim against the agent to compliance. In a case with this structure (involving, for instance, an obligation generated through a promissory exchange), the behavior called for is not the exclusive possession of the agent’s, but belongs in some sense to both parties whom the directed obligation links. This is connected to the important fact that the party to whom the obligation is owed has special standing to complain in case the obligation should be flouted. The challenge for an account of obligation, in general terms, is to make sense of this kind of directionality. Gilbert argues that this challenge can easily be met if we understand directed obligations to be constitutively connected with the structure of joint commitments among different parties to do something together; and she suggests that it will be very difficult to meet the challenge outside of these kinds of contexts.37 The upshot would be that directed obligations will not be possible for agents who are not already linked to each other through ties of shared agency.
A particular focus in Gilbert’s development of this position is the idea that directed obligations render the actions they prescribe the joint possession of the parties to them.38 I emphasized a similar point in section 2.3 above, noting that relational norms belong to the different parties that are connected through them, and that this helps us to see why agents lack unilateral discretion to fall short of the obligations that they define; this is among the features of these norms that suit them for the deliberative role of practical requirements. Gilbert argues that this joint ownership condition is fully intelligible in the context of shared agency. For if two parties are jointly committed to doing something together, then it is already the case that they see their individual activities as belonging to both of them, insofar as they are actions done in accordance with a commitment jointly undertaken by them. As Gilbert at one point writes, “Any one party of the joint commitment can appropriately say of the other’s conforming actions: ‘They are mine—in my capacity as co-creator of the commitment.’ ”39 She further suggests that to be the owner of an action in this sense just is to be the person to whom its performance is owed.40 So the relational structure of directed obligation goes together with the nature of joint commitment, and appears difficult to make sense of outside of this context.
Note that this account, if plausible, would potentially place surprisingly strict constraints on the applicability of directed obligations. Thus, even within the kind of close social contexts in which we encounter other people on a face-to-face basis, there are many such encounters that do not appear to rest on or involve a structure of Gilbert-style joint commitment. It seems that I owe it to the gout-afflicted stranger I encounter on the sidewalk not to step on her toe, and I would similarly be understood to owe a debt of gratitude to the stranger who made a significant personal sacrifice to help me change my flat tire in the rain. In these cases, it is far from obvious that there is any joint commitment that links the parties who are bound to each other through relational requirements and claims; and yet we intuitively feel that there is enough in the interaction between the parties to give rise to such a relational nexus.41
But Gilbert’s argument for embedding relational obligations within contexts of joint commitment does not in any case succeed. One difficulty is that the argument does not elucidate the directionality that is characteristic of relational obligation. If Olive and Roger are jointly committed to doing something that requires Olive to fly to Chicago, then it seems that each of them, in their “capacity as co-creator of the commitment,” can regard Olive’s going to Chicago as his or her own, insofar as this is part of something larger that they have undertaken together. This joint commitment sense of ownership is equally available to both of the individuals who are implicated in the commitment, including Olive (who can also regard the action as her own in a different, individualistic sense). Now Gilbert claims that being the owner of an action in the joint commitment sense is necessary and sufficient for being the person to whom the action is owed, which would have the consequence that Olive owes it to herself, as well as to Roger, to fly to Chicago. But this seems wrong. If Olive owes it to Roger to fly to Chicago, then it is Roger, and not she herself, who has a claim against her to this action, and who would have special standing to complain should she fail to carry it out.
More fundamentally, Gilbert’s argument appears to equivocate on the notion of joint ownership. The sense in which Roger owns Olive’s action in a case of directed obligation is fundamentally normative. The action belongs to him just insofar as he has a claim or entitlement against Olive to her performing it; and what Roger shares with Olive, strictly speaking, is not the action of hers that fulfills her duty, but the normative complex of obligations and claims that links them to each other. In this, the situation is precisely analogous to that obtaining with literal property relations. To own a given physical object (the green Ford that is parked a few streets down, say), is not to have the object in my physical possession at a given point in time, but to have a claim or entitlement to determine what is done with it. This normative claim is shared in common with other parties, insofar as it is constitutively connected with obligations on their part to refrain from making use of the car without my consent; it exists only as part of a Hohfeldian normative complex that includes those obligations.
Against this background, Gilbert’s appeal to joint commitment appears to be a non sequitur. Olive and Roger, in virtue of their joint commitment, might each be able to view Olive’s trip to Chicago as his or her own, in the sense of being something that they have both undertaken.42 But our quarry was the different, normative sense of ownership at issue when we say that Roger has an entitlement or claim to Olive’s action. This normative relation is not captured by appeal to the very different sense of shared ownership at stake in cases of joint commitment. For the same reasons, we could not elucidate the normative relation of property ownership by appeal to the distinct relation of physical possession; a given car can be mine, insofar as I am physically occupying and operating it on the roadway, without belonging to me in the sense of my having a property right to it. But if this is correct, then shared ownership of the kind at issue in joint commitment would appear, absent further argument, to be neither necessary nor sufficient for directed obligation.43 There would then be no principled obstacle to thinking of universal morality in relational terms, as a set of directed obligations and claims that can link individuals who are in no sense bound up in shared commitments of any kind.
The moral nexus, understood in this way, represents what we might call a self-standing domain of relational requirements and claims. This is to say that it is not grounded in prior relationships or interactions between the parties that it links. In this, it differs from some other examples of relational requirements, such as those that arguably connect friends or family members to each other. But as I noted above, interpersonal morality must differ from other normative domains on any interpretation, relational or otherwise, that aims to make sense of it as a coherent set of requirements that reflect the equal standing of all. Nor have we encountered compelling principled reasons for thinking that the relational paradigm cannot be extended in this way.
Philosophers attracted to a broadly relational approach to morality tend to gloss over its standing as a self-standing normative domain. Thus, it is sometimes said that moral duties are grounded in an abstract relationship we stand in to those who have claims against us that is analogous to the special relationships we stand in to friends and colleagues and members of a family group, a relationship we might call “fellow humanity.”44 But the analogy is specious. The familiar relationships that we recognize as sources of special obligations have a historical and social reality, consisting in patterns of causal interaction that are psychologically salient for those caught up in them. The relationship of fellow humanity, by contrast, involves joint instantiation by different individuals of a morally significant property, such as the capacity for reason or for principled self-determination. This is, as I remarked above, a merely logical relationship. To say that moral obligations are owed to those to whom we are related as fellow humans is merely to designate the maximally inclusive manifold of individuals who potentially have moral claims against us, and against whom we ourselves have claims. It does not serve to specify some antecedent relationship that might provide a “foothold” or “common source” (to use Thompson’s helpful expressions) for the normative nexus that interpersonal morality defines.45
I have been suggesting that it is open to us to interpret interpersonal morality as a self-standing domain of relational requirements. Requirements of this sort are not grounded in any independent relationship that might obtain between the individuals to whom they apply, and it is therefore possible that they might regulate our interactions with any agent who has moral standing, including individuals we have never before encountered. But the resulting relational interpretation of interpersonal morality has some significant normative implications that we now need to consider.
In chapters 2 and 3 of this book, I argued that considerations of moral rightness, interpreted in relational terms, have normative significance for two different parties. For the agent, they are considerations that are registered in reflection as presumptive constraints on the will, defining normative requirements that ordinarily have to be complied with. For the claimholder, they have a different kind of normative significance, representing considerations that make it fitting to hold the agent to the directed obligation that is owed to the claimholder, in the way that is characteristic of relations of interpersonal accountability. Relational moral requirements thus provide a kind of normative scaffold for the interactions between the two parties that are distinctively connected through them. It is in this sense that, as noted in the preceding section, they are held in common by the individuals whom they link.
This relational theory of morality, however, interacts in interesting ways with the general theory of normativity. In particular, it looks as if there might be some tension between the idea that there are moral obligations, on the one hand, and some popular approaches to practical normativity on the other. Many of the most influential accounts of reasons for action are broadly individualistic; they affirm that questions about what a given person has reason to do depend, in some way or other, on facts about the subjective attitudes of that individual in particular. Thus, Humeans of a fairly familiar kind hold that what an agent has reason to do is a function, in part, of the agent’s desires or wants, broadly construed. On this kind of approach, facts about an individual’s motivational profile are either reason-giving in themselves, or they are understood as states that condition the reason-giving force of other facts about the agent’s situation.46 Similarly, constructivists about normativity claim that normative principles of all kinds are binding for an agent in virtue of the agent’s individual attitudes, which commit the agent to treating some considerations rather than others as reasons for action and for other attitudes.47
If we understand practical reasons in these recognizably individualist terms, however, it becomes very difficult to make sense of the kind of normative nexus that I have taken moral obligations to represent. The basic idea, as I just noted, is that the obtaining of a directed moral obligation—its being the case that A owes it morally to B to act in a certain way, or that B has a moral claim against A that A so act—should have normative significance for two distinct agents. The directed obligation represents a presumptive constraint on the deliberations of A about what to do, as well as a consideration that makes it fitting or appropriate for B to adopt a distinctive practical stance toward A, holding A to account with regard to the obligation. According to the individualist approach to normativity, however, whether reasons of these kinds exist for A and B will depend on facts about their different psychological profiles, facts that go beyond the obtaining of the directed obligation itself, and that are liable to vary independently of each other. Under these philosophical assumptions, it is hard to see how a genuine moral nexus of the kind we have been discussing could really obtain.
The facts, whatever they are, that suffice to determine that A has reason to comply with a directed obligation that is owed to B should also suffice to establish that B has reason to adopt toward A an accountability-conferring stance with regard to the obligation, and vice-versa. But individualism does not appear to deliver this result. The facts that go into determining that the obligation is a presumptive normative constraint on A’s exertions of agency include facts about A’s attitudinal profile; but no set of facts of this kind could, on individualist assumptions, determine what stances B has reason to adopt toward A. A and B might come to develop attitudes that fortuitously align with each other, ensuring that, at least as long as the alignment persists, the directed obligation is a source of reasons for each of them. But this is not the kind of normative nexus that we take moral obligations to constitute. The obtaining of the directed obligation should itself provide the parties that it implicates with reasons to adopt the attitudes toward it that are appropriate to their different positions. Its normative significance for the two people who are bound by it should not depend on the fortuitous alignment of factors extraneous to it.
One response to this might be to observe that on any plausible account of normativity, there are going to be some subjective conditions on one’s reasons, if only constraints of an epistemic nature. Thus, whether B has reason to adopt toward A a responsibility-conferring attitude has to depend, in some measure, on information about A that is at least accessible to B. It would hardly be fitting for B to blame A, for instance, unless B has reason to believe that A has done something to flout duties that A owed to B, and this will be a matter of B’s subjective epistemic situation. By the same token, A can be said to have flouted an obligation owed to B only if A had reason to believe that the obligation obtains and that the action in question would violate it. But this observation does not defuse the apparent tension between individualist accounts of normativity and robustly relational accounts of the moral right.
The normative significance of a directed obligation for the two parties who are connected by it is filtered, we may concede, by facts about the individual epistemic situation of each of them. But the role of these epistemic filters is to make accessible to the two parties a normative nexus that obtains independently of whether they are in a position to grasp it.48 It is in this way only shallowly conditioned by subjective facts about the parties’ epistemic situation. Individualist accounts, by contrast, posit deeper subjective conditions on normativity. In different ways, Humeans and constructivists both want to say that considerations are made normative for a given agent by facts about that agent’s attitudinal profile, which function not to make accessible to the agent reason-giving considerations that independently obtain, but to constitute them as reasons for the agent in the first place.49 This is the kind of position that seems difficult to reconcile with the relational conception of normativity, which takes directed moral obligations and claims to provide reasons for two different persons who may have never so much as met each other before, and whose attitudinal profiles will therefore not antecedently be coordinated in any way.
Individualist approaches differ, among other things, on the scope of coverage they aspire to achieve. Constructivism, as I understand it, is at least sometimes put forward as a general account of normativity in any domain,50 but other versions of individualism are less ambitious. Humeans, for instance, typically take reasons for action to be constituted by motivating attitudes of the agents to whom they apply, but they don’t, or needn’t, affirm that reasons for other kinds of attitudes, such as beliefs and emotions, have similar subjective conditions. A more restricted position of this kind might apply, at most, only to one side of the alleged moral nexus.
Directed obligations represent presumptive constraints on the agency of the person who stands under them; but for the person to whom they are owed, their normative significance is at least partly of a different nature. They are reasons, in the first instance, not for actions on the claimholder’s part, but for emotional attitudes, including centrally those through which accountability is conferred on the agent who is subject to the obligation. Reasons of this kind are considerations that make reactive and other forms of blame fitting or warranted, in response to the flouting of a valid moral obligation on the agent’s part. But the Humean about reasons need not maintain that reasons of this kind are subjectively conditioned by the claimholder’s motivational profile. The Humean account of reasons for action therefore does not require us to think that a directed obligation can be in place only if the attitudinal profiles of both parties to the moral nexus align in a nonfortuitous way. It is enough for agents to have the attitudes that render the obligation normatively binding for them. Once this condition is satisfied, the Humean might say, there is a claim in place that the other party holds against the agent. This in turn means that the agent’s disregard of the obligation amounts to a disregard of another person’s claim, which is sufficient to render fitting reactive and other forms of blame by the claimholder.
This might appear to diminish somewhat the tension between Humean individualism concerning reasons for action and the relational account of morality. But the appearance is misleading. As noted in section 3.1 above, the social dimension of morality is complex; it is a matter not merely of reasons for accountability-conferring attitudes on the part of those affected by an immoral agent’s behavior, but of their having accountability-related reasons for action as well. The attitudes and actions characteristic of blame represent elements in an economy of social esteem and disesteem that serves to structure our social relations and to incentivize and reinforce compliance with basic moral standards. But it is open to us to step back from our participation in this system, and to reflect on whether it is reasonable for us to accede in the reactions and behaviors to which it disposes us. In doing this, we pose for ourselves a question about our reasons for action, asking (inter alia) whether we should continue to contribute normally to the economy of esteem and disesteem, or should instead attempt to withdraw from it, ignoring or disregarding our blame reactions in our interpersonal relations with other persons. Reasons for action are also at issue when it comes to nonreactive forms of blame, including such behaviors as avoidance of the wrongdoer, withdrawal from trust-based relationships with the wrongdoer, and so on.
The normative significance of relational norms for claimholders, then, is at least partly a matter of their providing the claimholder with reasons for action of this kind. These reasons may in some sense be derivative from the reasons for reactive blame that the flouting of relational obligations represents. It is the fact that such wrongdoing renders resentment and other reactive attitudes fitting that gives us pro tanto reason to accede to our participation in the economy of esteem and disesteem, and to adjust our behavior toward the wrong-doer in ways that express the reactive emotions that are appropriate under these conditions. That is why I said above that the normative significance of directed obligations for claimholders is “in the first instance” a function of reasons for attitudes rather than for actions. But reasons for action are secondarily also in play. To the extent that this is the case, a Humean individualism about such reasons will still seem at odds with the relational account of the moral that I have been trying to develop in this book.
To this, the individualist about the normative will naturally respond: so much the worse for the relational approach. Most radically, one could appeal to this kind of individualism to argue for a skeptical take on relational accounts of normativity in any domain. According to this form of skepticism, the appearance that there is ever a genuine normative nexus of the sort we have been discussing is an illusion. At bottom, there are just the reasons that people who are differently situated have for doing and feeling different things. Sometimes these individual reasons fortuitously align, in ways that might lead us to think that there is a normative connection that links two people with each other, as in cases of an ordinary promissory exchange. But strictly speaking there is no such connection; that is, the reasons of the two parties to the exchange are not constitutively tied to each other, so that it would be possible, in principle, for the promise-given reasons of one party to change, without changes either in the nature of the promise, the reasons of the other party, or the epistemic situation that the parties both find themselves in.51
A different position that retains a commitment to individualism about reasons might hold that there can be a normative nexus, but only under conditions in which there is an antecedent meeting of the minds by the parties that it links. According to this position, it is open to individuals to make it the case that they have reasons for action and attitudes that are constitutively connected to the reasons of another party. They can, for instance, undertake to do something together with the other party, forming a joint commitment that shapes the normative situation of each of them, going forward. Or they could enter into some other kind of transaction that binds them to each other, creating reasons that hang together in the right way (for instance, through participation in a common practice or convention).
Under these conditions, the individualist might concede that a normative nexus can come to exist. But the relevant conditions would precisely not be in place in all of the cases to which relational moral norms would seem to extend. As we saw in section 4.2, cosmopolitan moral requirements, when interpreted in relational terms, represent self-standing normative structures that need not be grounded in any antecedent relational interaction or connection to the parties that they link. But individualism apparently entails that there could be no self-standing normative structures of this kind, even if it does not necessarily rule out the very possibility of a distinctively relational requirement. On the present interpretation, individualism precisely requires that a relational nexus of directed obligations and corresponding claims be based in some antecedent relation between the parties to it, one that ensures the presence in them of the individual conditions that reasons for actions and attitudes presuppose. Individualism, on this interpretation, could therefore be understood as a form of revisionism about relational moral requirements.
As should be clear by now, I myself favor the idea that directed moral requirements and claims constitute a self-standing normative domain, one that is not grounded in any antecedent relation or connection between the parties that fall under its manifold. But I concede that this position will be difficult to sustain if one accepts an individualist account of normativity (in either a global or a more qualified formulation). If they are willing to countenance them at all, individualists will grant that there can be relational requirements and claims only in circumstances in which there is an independent connection between the parties who are thus related. The result would be a restriction in the class of situations in which genuinely directed moral requirements and claims could be in place. But that is only to be expected: insofar as the relational interpretation of normativity has metaethical presuppositions, it will be hostage to metaethical fortune.
There is obviously not the space here to undertake an independent assessment of individualism about the normative. But there are three points that may be worth emphasizing in this connection, at least very briefly. The first has to do with the modern, cosmopolitan conception of morality that I have invoked at various points in my discussion. So far, I have mainly emphasized the idea that this conception aspires to acknowledge the maximally inclusive character of the community of agents whose interests count for moral reflection. The suggestion is that we inhabit a world together with others who are equally real, and that moral obligations must be traced directly to the equal standing of the members of this inclusive notional community. This aspect of universalistic morality is straightforwardly accommodated on the relational approach, insofar as it treats moral obligations as connected to claims that others have against us, just insofar as they are members of this inclusive manifold of equals.
But it is also part of the modern conception of morality that its obligations are universal in a different sense: they not only reflect the equal standing of all persons, but they are binding on all individuals who are equipped with normal capacities for moral agency. Morality is in this way a source of requirements for everyone. But this is an aspect of the modern conception that will be difficult for individualists to make sense of, quite independently of whether moral requirements are interpreted in relational or in nonrelational terms. On individualist views, reasons of various kinds are a function of the subjective attitudes of the agent whose actions they would regulate, including their desires and dispositions or their beliefs about what there is reason to do. But it is a philosophical challenge to establish that all agents have the subjective attitudes that would secure for them compelling reason to comply with moral requirements, compatibly with individualistic assumptions. The most promising strategy for securing this result might be the familiar Kantian suggestion that the commitment to comply with the moral law is somehow built into the structure of agency, so that it can be regarded as an element in the subjective outlook of all individuals who are capable of self-determination.52 But the prospects for successfully executing this strategy are not, in my opinion, terribly bright. The upshot is that individualism about the normative seems difficult to reconcile with the modern conception of universalistic morality on any interpretation of it, relational or nonrelational.
A second point worthy of notice in this connection is that our accountability practices seem to incorporate a commitment to anti-individualism about moral reasons. I noted in the chapter 3 that these practices have a relational deep structure; they involve responses to wrongdoing that are ordinarily occasioned by a disregard for relational claims, and this is reflected in their cognitive content as well. Resentment, for instance, both is, and is understood to be, a response to actions that wrong the person subject to it. What makes this reactive attitude fitting, then, are facts about the willful violation of claims on the part of an agent. But from the standpoint of a participant in the practice of interpersonal accountability, the kind of subjective conditions that standardly figure in individualist accounts of the normative do not enter into it at all.
To be sure, resentment and other reactive forms of blame do attribute to the agent who is their target attitudes of some kind or other, specifically including attitudes of disregard for another person’s moral claims against them. But reactive blame will typically be warranted on the claimholder’s part so long as this condition is satisfied. It is not necessary, in addition, to establish the presence in the agent of the kinds of motivational and belief states that individualists take to be conditions of reasons for action.53 If A is normatively competent, as we might put it, then it would be fitting for B to blame A when A acts with contempt for or indifference to B’s claims. This is true, regardless of whether A’s desiderative profile includes an attitude of concern to respect the moral requirements that are intuitively associated with those claims, and regardless of whether A’s normative beliefs commit A to accepting the validity of those requirements. In this way, our practices of interpersonal accountability seem to reflect a pragmatic rejection of individualistic assumptions about normativity. Individualism is in tension not merely with the relational interpretation of morality, but with the participant understanding of our familiar reactions to moral infractions. This is perhaps not surprising if, as I argued in the preceding chapter, our practices of interpersonal accountability are themselves structured around relational moral norms.
The third and final point I wish to emphasize about individualism concerns obligation. As noted earlier, the normative phenomena that individualists most saliently take to have implicit subjective conditions are reasons for action rather than for other attitudes. Thus, Humeans typically contend that one can have reason to pursue a given goal only if one has a desire (broadly construed) that suitably aligns with such a pursuit (for example, a desire either to achieve the goal directly, or for something else that would be advanced by its realization). As was emphasized in chapter 2, however, the normative significance of moral considerations for the agent consists in the provision of obligations, rather than merely aspirational reasons for action. That it would be morally wrong to do X is not just a consideration that counts against doing it, but one that enters the deliberative field as a presumptive constraint on one’s agency (of the form: whatever else you do, don’t do X!). But obligations or practical requirements, in this sense, do not appear to be the sorts of normative phenomena that are naturally thought of as grounded in or conditioned by an agent’s desires. Indeed, in our developmental experience, our initial encounter with practical requirements is likely to be in contexts in which they are precisely at odds with our desires. We come to understand what it is to be under an obligation by seeing that obligations constrain us even when we do not feel like complying with them. Obligations thus do not seem to be auspicious candidates for treatment in standard Humean terms.
As I observed in section 2.1, there are individualistic approaches that provide at least potential models for understanding practical requirements. Identity-based accounts, for instance, trace obligations to features of one’s practical outlook that are constitutively bound up with one’s conception of who one is. These features might well include aspects of one’s desiderative or motivational profile, such as the care one has for another person whom one loves. But this strategy seems unpromising when it comes to making sense of the kinds of moral obligations that have been at the center of discussion in the present chapter.
Consider, for instance, the transactional obligations that are generated by promises and agreements between two parties. I suggested earlier that individualists might concede that relational obligations can be in place in such cases, insofar as their transactional character makes for a meeting of the minds between the parties to them. But further reflection raises doubts about whether relational models aptly apply even to cases of this kind. Most significantly, there is a temporal dimension to such transactional cases that seems fundamentally at odds with individualistic approaches. Let us grant that the promissory exchange brings together the attitudes of promisor and promisee at the point in time when the transaction between them is entered into, mobilizing in them some identity-based concern to take the promise seriously. The promise, once it has been made, generates an obligation that remains in force, going forward, until its terms have been discharged. But there is nothing in the promissory exchange that will guarantee the persistence of the subjective attitudes it relies on into the future. The promisor, having entered into the promissory undertaking with a sincere concern to honor it, might lose interest in it over time, coming to care much more vividly about other things. The traditional individualist about reasons for action should say that the promisor’s obligation to fulfill the terms of the promise will change with time, in accordance with these changes in the promisor’s motivational profile. But this is completely at variance with our understanding of how promissory obligation works.
In sum, standard individualist stories about the normative seem peculiarly ill-equipped to deal with central features of the modern conception of morality. They have difficulty making sense of the universal applicability of moral requirements, of the conditions that warrant reactions of interpersonal accountability, and of the status of moral considerations as obligations, and not merely as aspirational reasons for action. Individualism about normativity may be in tension with the relational approach I have been developing, but this is because it is more fundamentally in tension with basic aspects of morality that the relational approach helps to illuminate.
For my part, I do not believe that individualism about the normative is an independently plausible position to take. But that is not a view that can be defended here. It is enough, for present purposes, to observe that those who are committed to individualism will find it a struggle to make sense of many aspects of the moral domain. These include, most saliently, the very features of modern morality that the relational approach seems especially well equipped to help us understand.54
In chapter 2, I suggested that it is part of our conventional understanding of moral obligation that it involves constraints on behavior that are agent-relative in character. The moral prohibition on duplicity or deceit does not give us the objective of reducing the incidence of these forms of conduct in the population at large, but enjoins us not to act in these ways ourselves. I suggested, furthermore, that the relational notion of a directed obligation promises to illuminate this feature in our conventional understanding of morality. The idea was that the paradigm examples of directed obligation, such as the obligations that are created by promises, are intuitively understood to be in the way of agent-relative constraints, and that this feature of them is connected to the directed character of the obligations at issue. Having made a promise that I would do something, I owe it to the promisee that I should myself do the very thing (whatever it was) that I promised to do. This obligation is neither equivalent to, nor derivable from, an agent-neutral commitment to promote the general goal of promissory compliance in the local or global community.
In light of the discussion in the present chapter, however, it might be wondered whether this argument trades on an aspect of the paradigm examples that does not generalize to all of the cases that fall under the universal moral nexus. Promissory obligation, as we have seen, involves a specific transaction between the two parties that it links, and this seems to have a bearing on its agent-relative character. By making a promise to someone, I establish a relationship with the promisee that grounds a directed duty, owed specifically to the promisee, that I should fulfill my promise. The promissory transaction thus connects me with the promisee in a way that renders intelligible my agent-relative duty to satisfy the terms of the promise.
Something similar is true of other cases in which directed duties arise from causal or historical interactions between the parties that they link. If I have benefited in particular from the costly assistance of a stranger, I owe my benefactor a duty of gratitude, one that can be complied with by reciprocating in some fashion for the efforts that were undertaken on my behalf. I would not count as fulfilling this obligation if I were merely to promote the general objective of rewarding benefactors for the sacrifices they have made on behalf of people in need of assistance; rather I must do something to express gratitude toward the specific individual who has benefited me. Similarly, the directed duties that derive from the special relationships we stand in to friends and family members have an agent-relative character that is appropriate to the historical connection between two people that relationships of this kind essentially involve. Friends owe each other duties of loyalty that they would not discharge by promoting the general value of friendship in the community at large, but that require them to stand by each other in moments of anxiety and need.
It is natural to associate directed obligations with the valuable forms of interpersonal relationship that are enabled by them. T. M. Scanlon, for instance, traces the reason-giving force of moral requirements to the value of what he calls mutual recognition.55 His idea is that a contractualist interpretation of moral requirements enables us to see how compliance with them makes possible a distinctively valuable form of relationship, whereby the parties acknowledge each other as sources of claims against each other. Though it requires qualification and development, I find this to be a deeply suggestive proposal, and I shall return to it later in this section. In the meantime, I note that Scanlon’s idea seems to have natural application to the kinds of examples we have just been discussing, where directed requirements are grounded in an antecedent relationship between the parties that are connected through them. The promisor has already entered into a relationship of some kind with the promisee through the making of the promise, defining them both as parties to a single normative nexus. Compliance with the requirements and claims that constitute this nexus enables them to conduct this existing relationship in a distinctively valuable way, on a basis of mutual recognition or regard.
Similar remarks apply to the other cases we have discussed, involving the relationships between benefactor and beneficiary and between friends or members of a close family. Directed obligations grow out of the antecedent relationships that are in place in these cases, and by fulfilling those obligations the parties are able to structure their ongoing relationship with each other in a valuable way. The agent-relative character of the requirements that are here in place seems connected to the fact that there is an existing relationship between the agents who are subject to them, one that can be managed better or worse, where compliance with the requirements is a condition for conducting them well. By flouting the terms of their agreements, promisors impair their own relationship with promisees. Doing this might, in the philosopher’s familiar scenario, bring it about that five other promissory relationships are not similarly impaired; but insofar as the promisor is not a party to those ongoing relationships, this consideration lacks the significance for the promisor’s deliberations that the impairment of his or her own relationships clearly has.
When we turn to the self-standing relational requirements that also seem to be a part of a cosmopolitan conception of morality, however, things seem otherwise. Here there is no antecedent, nonnormative relationship or interaction between the parties to the moral nexus that is the basis for its linkage of obligations and claims. It is therefore not the case that the parties are already connected to each other through an ongoing relationship that, whether they like it or not, it is up to them to conduct better or worse. In this general situation, I might well be invested in the value of relationships of mutual recognition. But why should this concern take a distinctively agent-relative form? Why should I be concerned that I myself stand in relationships of this kind with people I have not yet met or interacted with, rather than being concerned that valuable relationships of the same kind should generally be promoted in the larger social world? Why should I refrain from stepping on the gouty toe of the stranger if—as philosophers have the power to arrange—by doing so I can bring it about that five other agents respect the bodily interests of gout-stricken strangers that they will encounter sometime in the future?
In considering this question, we cannot appeal to the fact that I am already party to an ongoing relationship with the person whose gouty toe is vulnerable to being affected by the personal trajectory of my body. In the cases at issue, there is no such relationship to appeal to, and so it isn’t a matter of conducting better or worse a relationship in which I am already implicated. To render intelligible the agent-relative character of our moral concerns in such situations, we therefore need to draw on considerations of some other kind.
But an alternative basis for understanding the agent-relativity of moral requirements is provided by the relational interpretation I have been developing. In the preceding discussion, it was assumed that agents have a legitimate interest in conducting the relationships in which they are already implicated on the best possible terms. Being a participant in an ongoing relationship of this kind is a consideration that can make sense of a person’s concern for the terms on which that relationship in particular is conducted. But a distinct consideration that might ground such a concern is a person’s standing as the targeted object of a moral claim. On the relational approach, individuals can have claims specifically against us, regardless of whether we already stand in nonnormative relationships of any particular kind with them. In respecting those claims against us, we relate to their bearers in a way that acknowledges the significance of their interests for our agency, and that ensures that there will be no basis for their resenting what we have done. This is something that might reasonably matter to us even if we have not antecedently interacted with them at all.
Consider again the case in which five gout victims’ claims will be respected by others if we ourselves flout the claim of a sixth. This is a situation in which it makes all the difference against whom the various claims are held. The single victim has a claim on us in particular, and so it is naturally important to us to conduct ourselves in a way that respects that claim, going forward. That five other agents will flout claims that different strangers have against them is not in the same way important for us, since it is not a consideration that directly affects the question of how we relate to those who have claims against us. The intellectual temptation to suppose that we might appropriately honor claims by promoting their maximal satisfaction in the population at large rests on a neglect of their relational structure; they are directed at specific individuals, and it matters to our thinking about a given claim whether we are the person against whom it is held. Indeed, it seems plausible that this same feature is what ultimately accounts for the agent-relativity of directed obligations even in the cases in which there is an antecedent relationship between the parties whom they link. The relationship between the promisor and the promisee matters in thinking about what the former must do, because it gives rise to a claim on the part of the promisee against the promisor in particular. It is true that the two parties are already implicated in a historical relationship, but this is significant for our understanding of the promisor’s obligations only insofar as it bears on the question of the promisee’s claims against the promisor.
This way of thinking about morality might seem to invite the charge of moral self-indulgence. The complaint would be that conscientious moral agents should not be prissily concerned with their own moral purity, but should care in the first instance about how others fare who are apt to be affected by what they do. If there are five moral claims that can be honored as a result of an action of mine that violates the claims of a sixth, then that is a small price to pay to ensure a better outcome for the larger universe of moral claimholders.
This complaint rests on a flawed conception of moral self-indulgence, however. As Bernard Williams argued long ago, there is such a thing as moral self-indulgence, understood as a defective form of moral motivation. It involves a reflexive concern for the virtues that are expressed in the actions that one performs, rather than for the first-order reasons to which those virtues themselves are properly responsive.56 Understood in these terms, however, moral self-indulgence is a hazard to which agents are potentially subject according to any account of what morality requires of us, including nonrelational accounts such as utilitarianism. To act out of a concern for how one stands with regard to the claims of other people is not to betray an objectionable obsession with one’s own moral virtue, as opposed to a concern for the legitimate interests of other agents. Rather, it reflects a proper concern for those interests, understood as considerations that ground inherently relational duties. Persons who are unwilling to violate a claim held against them in order to bring about a reduction in the overall number of claims-violations are responding directly to the deeply relational structure of moral obligations. Of course, it is a matter of philosophical controversy whether the obligations of impartial morality are inherently relational in their structure. If they are, however, then a concern for how one stands in regard to others need not betray a preoccupation for one’s own virtue as opposed to the claims of others. Concerns about moral self-indulgence therefore do not represent an independent basis for objecting to the relational story I have been telling.
Let us now return to the idea that compliance with directed moral obligations makes it possible for us to stand in relations of mutual recognition. I noted earlier that this idea has natural application in cases in which there is an ongoing relationship between the parties who are connected by moral requirements and claims. The concern, in this context, is to conduct a relationship one is already implicated in on a basis of mutual recognition. But what about the situations we have just been discussing, in which directed moral requirements apply to people who do not already stand in any significant interpersonal relationship? In honoring the claims that others have against us in these situations, our concern is not to continue an ongoing relationship in accordance with the relational standards that connect the parties to it. It is, rather, to ensure that we relate to anyone who might be affected by our actions in accordance with such standards. If people we have never met before have claims against us, as I have argued that they can, then we will rightly strive to conduct our planning agency on a basis of respect for those claims, treating them as presumptive constraints on our activities. To do this is, among other things, to show a concern for how one relates to the bearer of the claim.
I have elsewhere suggested that moral requirements should be understood to be relationship constituting, not relationship based; they make possible valuable forms of interpersonal relationship, even if they are not grounded in some antecedent pattern of interaction between the parties related through them.57 But talk about “mutual recognition” is somewhat misleading in this context, taken as a characterization of the value that is realized through compliance with moral norms. If we interpret such talk in literal terms, it would appear to presuppose moral reciprocity. Two individuals would be party to a relationship of mutual recognition, on this literal interpretation, just in case each of them acknowledges and accords the claims that the other holds against them. The reciprocal element in mutual recognition is an important respect in which it resembles a different relational value that is sometimes mentioned in this connection, as an analogy or model for the value realized through compliance with moral requirements. This is the value of friendship, which figured in my discussion of the positive moral ideal of interpersonal recognition in section 2.4 above. But if mutual recognition involves something like the kind of reciprocity characteristic of friendship, it is also, for that very reason, an ideal that does not literally apply to many of our relationships within the manifold of moral persons.
The problem is that people have moral claims against us, on any plausible construal of relational morality, even when they do not show respect or concern for our moral claims against them. There are things that it is clearly impermissible for me to do to other people, even if they are out to kill me or to subordinate me completely to their will. (Force may plausibly be used to defend myself against such assaults, for instance, but only to a degree that is both necessary and proportionate, given the nature of the threat that is posed.) Under these circumstances, there is no prospect of realizing the literal value of mutual recognition in my interactions with those who have set themselves against me. They have, through their actions, already seen to it that this reciprocal value cannot be achieved, regardless of what I might do to them. And yet, as I just noted, they continue to have moral claims against me. It follows that the concern that leads me to honor those claims cannot be a concern for mutual recognition, construed in the literal way that would require the different parties to a relationship to honor the claims that each has against the others.
Interpersonal morality, on the relational interpretation of it, includes principles that apply to everyone, and that define obligations that are owed potentially to each member of the extensive manifold of persons that it defines. General compliance with such principles would bring about a situation in which all individuals conduct their affairs on a basis of equality, acknowledging that the interests of members of this extensive manifold count equally, and honoring the claims that others have against them in virtue of their equal standing within the manifold. Acknowledging and honoring other persons’ claims is a way of recognizing them as bearers of moral claims. So each person’s recognition of others will be reciprocated, under conditions of general compliance with relational moral principles, through a similar attitude of recognition on the part of those who have claims against that person. This idealized situation might aptly be described as one of mutual recognition or regard.58
Each of us, as individuals, is powerless to bring about this idealized condition through our own efforts alone. It requires the collaborative agency of others, something over which we do not exercise control. What is in our power is to do our part in the scheme of interlocking ends that is represented by a situation of literal mutual recognition. The concern to act in this way, however, is not derivative from a concern for the value of the condition that would be realized if everyone else were doing his or her part, as well. If it were, then the concern would presumably dissipate as soon as it becomes clear to us that the specific claimholders we are interacting with do not respect our claims against them. Situations of this kind make it clear that the fundamental value that is realized through moral agency is not derivative from the collaborative good of mutual regard. It is an independent value that involves a distinctive way of relating to others, on a basis of what I have called interpersonal recognition. Those who realize this value in their actions acknowledge the moral claims that other individuals have against them, and honor those claims in their decisions about how to comport themselves in the world. They do this, moreover, in the understanding that their recognitional stance remains valuable even if it is not literally reciprocated.
It is sometimes said that morality enables us each to live in harmony with our fellows, and that it in this way contributes to the quality of our own lives.59 The point is suggestive, but also somewhat misleading, for reasons that dovetail with the concerns about mutual recognition discussed above. Other people have claims against us even when they do not acknowledge our claims against them, and respecting their claims under these conditions will not bring about a condition of mutual recognition. But neither will it achieve social harmony, necessarily. On the contrary, especially under the conditions of entrenched inequality that prevail in most societies, standing up for the claims of some will likely put us in opposition to others, including most notably the powerful and the privileged. A morally decent life is often one that stirs up controversy and conflict, posing awkward questions to the complacent and challenging the conventional wisdom about social arrangements. To the extent this is the case, acknowledgment of the moral claims of others will not promote harmony with our fellows, any more than it will ensure a situation of literally reciprocated recognition and regard. This is an important respect in which relational moral values are different from those involved in relationships of friendship or personal attachment.
The personal value that is realized through deliberate compliance with relational moral requirements is more subtle than these conventional formulations might suggest, though it is no less important for that. It is the value that is achieved when we have done what we could to honor the claims that other individuals hold against us, regardless of whether they have shown a similar regard for our claims against them. To satisfy this condition is to adopt a special sort of attitude toward people, relating to them (as I put the point above) in a way that seems distinctively worth caring about. This is a claim about intrinsic value and, as with other such claims, it does not admit of proof or demonstration. I cannot establish the goodness of the attitude of interpersonal recognition by deriving it from premises that are independently more compelling, for example. The best that can be done is to try to articulate the aspects of this stance that make it valuable, in the hopes that others will concur in recognizing the value once its good-making features have thus been laid bare.60
There are different strands that could be emphasized in the attempt to articulate more specifically what is valuable about interpersonal recognition. We might start with Scheffler’s idea, discussed in section 4.2 above, that there is an egalitarian form of relationship that we intuitively recognize to be valuable across a wide range of contexts. I have repeatedly referred to the postulate of equal standing that is a hallmark of modern, cosmopolitan thinking about the moral realm. We feel that it would be absurd to deny that the interests of each individual are neither more nor less important than the interests of any other. It matters to us that we should do justice to this desideratum in the conduct of our lives, and a plausible way of understanding this concern is in deliberative terms. Morality, on the relational interpretation of it, attributes equal importance to the interests of each individual as potential bases of claims that are held against others. So to make a conscientious effort to comply with its requirements is, in effect, to deliberate about conduct that might affect other people on a basis of equality with them, something we moderns rightly care about.
Granted, the ideal of relational equality plays out differently in the moral context than it does within the ongoing thick relationships that are Scheffler’s primary concern. A relationship such as friendship operationalizes the egalitarian deliberative constraint through the acknowledgment by both parties that their interests are equally important. Interpersonal recognition, however, insofar as it remains valuable even when it is not literally reciprocated, lacks this element of mutual acknowledgment. But this difference between the cases undermines neither our conviction of the importance of unreciprocated recognition nor our sense that it is a special case of a generic value that is also realized in the context of ongoing relationships between individuals. Important here is the idea, explicated above, that the principles through which we realize interpersonal recognition are the very principles that would make possible a relationship of genuinely mutual recognition, one in which the egalitarian deliberative constraint is reciprocally satisfied.
A distinct strand that might be emphasized by way of articulating the value of interpersonal recognition is that it affects concretely one’s relationships with other people, going forward, even in cases in which it is not fully reciprocated. By honoring the claims that others have against us, we ensure that those affected by our conduct will not have a legitimate complaint to bring against us on the basis of it. We will have done them no moral injury or wrong, and we will in this way ensure some degree of protection from their justified resentment. This is connected to another consideration that I mentioned earlier in connection with the value of interpersonal recognition, that of justifiability to others.61 When we conscientiously comply with relational moral requirements, we can in principle appeal to the requirements to justify our conduct to each of the persons it affects, showing in this way that we have lived up to the obligations that we owe specifically to them, and that we have honored the claims they hold against us. We could look them in the eye and give an account of ourselves specifically to them, if called on to do so. People intuitively feel that it matters whether they are able to relate to others on this kind of basis, and this reflects the widespread conviction that interpersonal recognition is a valuable way of conducting oneself.
Taking these discrete strands together, it seems to me highly plausible to suppose that our own lives go better to the extent we satisfy the condition of interpersonal recognition in which these strands are interwoven. If, as I have argued, there are directed obligations that we owe to others, just insofar as they are persons who are equally real, then people who acknowledge those considerations as presumptive constraints on their agency will be deliberating correctly. They will be, in this respect at least, good practical reasoners. But this fails to capture the significant positive contribution that acknowledgment of moral obligations has on the agent; in acknowledging the claims that others have against us, we thereby relate to them on a distinctive basis, what I have called interpersonal recognition, which has a value that goes beyond its contribution to making us generically good practical reasoners. It is a noteworthy if underappreciated feature of interpersonal moral standards, on my interpretation of them, that compliance with them has this positive effect on agents and the lives that they lead.
Now, to focus primarily on this aspect of relational morality in deliberating about what to do would perhaps betray an objectionable self-concern, something in the ballpark of moral self-indulgence. The attention of conscientious moral agents should ordinarily be directed to the claims that others have against them, as considerations that constrain their deliberations about what to do in the way we have seen to be characteristic of practical requirements. To act out of consideration for such claims just is to deliberate in a way that acknowledges the equal standing of all. But the discussion in the preceding paragraphs suggests that compliance with this egalitarian deliberative constraint will bring it about, as a kind of secondary effect, that our own lives are made better as well. They will have been conducted on a basis of interpersonal recognition, reflecting a commitment to the equal importance of others that immunizes us from legitimate complaints about the effects of our conduct on them, and that makes it possible to give a compelling account of ourselves to them.
There is a role for considerations of this kind to play in thoughtful agents’ awareness of the values realized through their agency, a role that will be salient in contexts of what we might call eudaimonistic reflection.62 This is the kind of reflection we engage in when we think about what it is for our own lives to go well, identifying things that make them meaningful or worthwhile from our own point of view. This is a context, as it seems to me, in which the intrinsic value of interpersonal recognition might have a constructive role to play. The case of friendship, though disanalogous with interpersonal recognition in some respects, may help to illustrate the point.63 To be a friend is, in the first instance, to acknowledge the special reasons and requirements that apply to us in virtue of standing in this relationship to another person. It is, for instance, to take another person to have claims against us just in virtue of being our friend, and to strive to honor and respect those claims. Attitudes of this kind track and respond to reasons that people have in virtue of standing in the relationship of friendship to others. But we also understand the relationship that is partly constituted by these kinds of recognitional attitudes to be valuable in itself, in ways that go beyond their value as generically correct responses to reasons. The relationship of friendship of which they are a part is among the things whose presence in our lives can contribute to making them meaningful and worth living, from our own point of view—for instance, as things that we take particular pride or satisfaction in, when we look back in reflection on how things have gone for us; or as personal values that give us reason to carry on in life in the first place.
It has been suggested that the value of relationships such as friendship can help us to understand the “reason-giving force” of the considerations associated with them.64 This suggestion is initially somewhat puzzling, since true friends do not generally act, in the first instance, out of a concern to have the goods of friendship in their lives, but out of a concern for their friends. But the eudaimonistic significance of valuable relationships of this kind has an indirect role to play in bolstering our conviction that the considerations associated with them are serious reasons for action. Friends have claims against each other, just in virtue of standing in this relationship to each other; we might say that it is constitutive of friendship that these relational obligations obtain and are honored. As was noted in section 2.4, however, our sense that the normative constraints at issue are significant is connected to our understanding that the form of relationship in which they are implicated is independently valuable, something that we have reason to care about when we think about what it is for our own lives to go well.
As I have emphasized above, interpersonal recognition is different from friendship in lacking the element of mutuality. It is a way of relating to people rather than a full-fledged relationship with them, one that retains its value even under conditions in which it is not reciprocated by the individual who is its object. If this claim is plausible, however, then we should acknowledge that interpersonal recognition has a role to play in eudaimonistic reflection that is similar to the role played in this context by thick relationships such as friendship. Agents who have these attitudes are not merely reasoning correctly about what they are obligated to do; they are also realizing other values that people should care about when they survey their lives retrospectively or make provision for the future. Indeed, the fact that interpersonal recognition is valuable in this way helps to reassure us that the moral requirements to which it responds are genuine constraints on our conduct, and not merely arbitrary demands associated with an alien and questionable system of values. It is a significant contribution of the relational approach that it helps us to appreciate this eudaimonistic dimension of interpersonal morality, isolating a significant respect in which our own lives go better when we comply with its requirements.