Do Obligations to Co-nationals Constrain Effective Altruism?

Abstract

Social entrepreneurs face a dilemma. When making decisions about corporate giving, should they prioritize groups with whom they share some historical, national, or emotional tie or should they maximize the overall effectiveness of their contributions? According to a thesis I call “associationist priority,” the moral reasons to favor stakeholders with whom the entrepreneur shares an associative relationship trump the reasons to promote the impartial good. An important component of the argument for associationist priority is the premise that some nonvoluntary associations, including those between corporations and members of their communities, create special moral obligations. This essay argues against associationist priority by way of denying nonvoluntary associative obligations generally. This expands the moral discretion corporate social entrepreneurs enjoy both in how they give and to whom they give.

Introduction

Philosophers representing a variety of otherwise different theoretical outlooks accept the existence of obligations to provide aid to strangers, especially when such aid can help realize urgent, morally significant needs. ¹ Defenders of effective altruism affirm that there are strong moral reasons, perhaps extending to moral obligations, to provide aid to others in ways that will yield the greatest possible positive returns. ² At the same time, many philosophers accept obligations to particular persons or groups in virtue of some shared membership or association with them. Associative obligations are those moral obligations that obtain for certain persons “just because they are members” of some group where their status as members or role within the group has “not been voluntarily entered.” ³

The effective altruist centrally holds that we should act in ways that will promote as much value as possible, though various defenders of the view may differ on what states of affairs count as valuable as well as about their relative weightings. The defender of associative obligations holds that we should act to satisfy our obligations toward those with whom we share associations. A conflict looms between these two views. Given an ordinary slate of empirical facts about inequities in wealth, resources, and influence, it will very often be true that those for whom we could make the biggest difference will also be persons with whom we share little or no networks of association. ⁴

Governments as well as private actors can face this conflict. The problem may be particularly acute for social entrepreneurs, who have both a commitment to doing good and a commitment to discharging obligations to “stakeholders” with whom they share some institutional, historical, or associational connection. ⁵ These stakeholders may include co-nationals, with whom the social entrepreneur shares a political association, as well as other agents and groups. How should a social entrepreneur respond to this conflict? One answer is to hold that there is a difference in the moral status of our reasons to help strangers and our reasons to help those with whom we share (certain kinds of) association. According to this reply, our associative duties are morally required, whereas the well-being of distant strangers does not morally bind us. If our associative duties are morally required while reasons to participate in effective altruism are not morally required, that suggests that the social entrepreneur’s obligations to co-nationals and other stakeholders will be less defeasible than their reasons to benefit nonstakeholders. This view holds that associative obligations trump reasons to give aid to strangers, even if the outcomes of the latter would be better impartially. Call this thesis associationist priority.

This essay will argue against associationist priority. I deny that nonvoluntary associations are sufficient to generate moral obligations, although I will allow that they could create moral reasons with justifying strength. This has two implications for social entrepreneurs. First, it suggests that they are not bound to privilege some special class of stakeholders above the impartial good, barring some voluntarily accepted agreement. It thereby underscores the social entrepreneur’s discretion over whom to give to. Second, it denies that the social entrepreneur has any standing debts to co-nationals that cannot be discharged through business practices alone. This highlights discretion in how to give.

The essay’s argument proceeds in four stages. First, I will describe the associationist priority in more detail, particularly as it relates to business practices. Second, I will explicate the concept of moral obligation. Third, I will hold that association alone is an insufficient ground for obligation. Fourth and finally, I will focus on the corporate social entrepreneur, showing how such actors enjoy moral discretion along two dimensions of giving.

Associative obligations in corporate social responsibility

The literature on social entrepreneurship features a suite of potential stakeholders to whom obligations might be owed. These include employees and shareholders, but also local businesses and communities, the environment, customers, suppliers, trade organizations, fellow citizens, and governments at different institutional levels. ⁶ Although corporations may have made promises to some of these parties, perhaps most of these relationships will be based on association rather than explicit agreements. My suggestion here is that many obligations ordinarily supposed by theorists of corporate social responsibility will have the form of associative obligations. For example, business ethicists affirm that there are sometimes perfect (or nondiscretionary) obligations to community members who share common citizenship, for instance. In such cases, the corporation may have a special capacity to provide a benefit, although doing so may not be optimal from an impartial point of view. ⁷

The idea of an associative obligation usually involves three elements:

1. (1) An agent is a member of some identifiable group.

2. (2) The agent did not voluntarily consent to do more for other members of the group than they would be required to do for anyone else.

3. (3) The agent’s membership in the group involves receiving various moral goods.

To make this more concrete, imagine that the city where a local supermarket chain is based is hit by a devastating hurricane. The supermarket chain has benefited from being part of the community where it has prospered financially. The supermarket has made no explicit promises to give aid in the event of a disaster. However, the associative theorist would hold that because the three conditions above are met, the supermarket may be morally required to provide relief to members of the community.

I will be interested in two central types of case. The first are cases in which an individual or company is part of some group such that they did not voluntarily choose to be part of that group. The second are cases in which an individual or company did voluntarily choose to be part of some group, but they neither accepted some particular obligation to the other member or members nor voluntarily empowered another member of the group to decide such obligations. The associative theorist believes that such parties can have associative obligations to members of a group, even though they did nothing to accept the specific obligations in question.

It is ordinary for philosophers and business ethicists to believe that there is an associative obligation to one’s political community. ⁸ Psychologists have recently shown that such beliefs are widespread among nonphilosophers as well. Many people regard supporting pro-national policies as morally obligatory. ⁹ In addition to imposing new requirements, associative obligations might also create new permissions, such as permissions to avoid responding to the interests of outsiders. ¹⁰

Political philosophers and business ethicists alike have held that corporations can have special moral obligations to their political communities and co-nationals. In the rest of this section, I will highlight one version of the debate over corporate obligations to co-nationals. This debate centers on whether corporate social responsibility is compatible with for-profit activity. Can corporations satisfy ethical demands to citizens or governments with actions that also turn a profit? I will suggest in the sections that follow that how we think about associative obligations in general bears on how we evaluate the morality of social entrepreneurship. If there are nonvoluntary associative obligations binding corporate giving, then social entrepreneurs will likely have less moral discretion in how and to whom they give. At the limit, some theorists suggest that for-profit activity can never satisfy moral requirements to compatriots. Such a conclusion would call the very idea of social entrepreneurship into question. So figuring out whether there is a class of associative obligations in the first place, I suggest, will help resolve important moral questions about how corporate social responsibility can be pursued.

Granting arguendo that there are obligations to compensate fellow citizens for their civic labor (for example, voting, participating in politics, and the like), Jason Brennan maintains that such obligations can be satisfied through providing for-profit goods and services. ¹¹ Against this view, Brookes Brown argues that self-interested actions can never satisfy obligations to fellow citizens, because such obligations must involve assuming a burden proportional to the burdens borne by other citizens in their civic activities. ¹² Brown’s guiding moral idea is reciprocity. If someone else has done something for you that caused them to incur a burden, then, according to her view, you cannot repay them without incurring a burden as well. She proceeds by motivating this idea initially with first-personal cases and then drawing inferences about corporate giving. The methodological assumption, here as elsewhere when the method of cases is deployed, is that we can make more complex interactions analytically tractable by starting with simpler situations.

Brown considers two kinds of cases. In the first, a for-profit service provides someone with a significant benefit. (In her case, a Katy Perry album benefits a person going through a difficult personal moment. ¹³) In the second, one friend reciprocally benefits another by providing a compensatory good that the latter values: brownies but not banana bread, in accordance with the friend’s preference. ¹⁴

At issue in this dispute is whether, as Brown’s “burden-based view” supposes, one must assume a burden to satisfy an obligation to co-nationals. This is because, according to the view proposed, co-nationals have assumed burdens in performing their duties as citizens. Reciprocity requires assuming a burden proportional to the burden assumed by the initial performance. If so, then for-profit activities will never suffice, because they do not impose proportional burdens on the obligation-bearing agent. Although businesses undertake burdens in profit-seeking activities, they also aim to be compensated financially and, for this reason, such activities do not satisfy Brown’s condition. ¹⁵ Because citizens assume burdens without being repaid, only uncompensated giving can count as proportionally burdensome. Brown draws the following contrast. It would be inappropriate for Katy Perry to demand compensation for benefiting her listener. “Something would seem terribly off if Perry showed up at your door” requesting that you compensate her for the surplus benefit she provided you. ¹⁶ However, it would be appropriate to repay a friend with brownies for a burden that the friend had undertaken for you. Brown’s diagnosis is that you do not owe compensation to Perry, because she did not assume a burden, but you can owe something to a friend who did. Brown infers that Brennan’s suggestion of for-profit compensation to co-nationals is misguided because such activity likewise fails to assume a proportional burden.

I will return to this case in the section below on “Why corporate social responsibility is not moral responsibility.” For now, I will suggest that how we classify obligations depends in part on our concept of obligation. In the next section, I will characterize features of the concept of obligation. After that, I will raise questions about whether our moral reasons to aid co-nationals count as bona fide obligations at all. If they do not, then this will have implications for how to think about this debate over how members of corporations might appropriately compensate fellow citizens.

Aspects of the concept of moral obligation

This section will describe ‘moral obligation’ as delimiting a roughly unified conceptual terrain, characterized by at least four basic criteria. ¹⁷ Although other authors use the concept differently, I will suggest that nothing crucial hangs on the differences. The section will then discuss how to think about borderline cases of conceptual membership.

Consider two cases in which you have reason to meet me for a meal:

Promised Lunch: I ask you if you will have lunch with me and you agree, deciding to make time in your busy schedule. As agreed, we meet for lunch the next day at noon, and I am considerably cheered by your company.

Lonely Lunch: You notice that I am eating lunch alone and you know that I would be cheered by your company. You have other things to do, but you suspect that the benefit you provide to me by joining me would be morally more important than the alternatives, so you decide to join.

In both the promised and lonely versions, you have moral reason to meet me for lunch. In both versions, you might be appropriately praised for your actions and I might appropriately feel gratitude to you. They differ in that you are only morally obligated to join me in promised lunch. Had you decided against meeting me for the lunch that we had both agreed to, you would have done something wrong. However, if you had opted against joining me for lunch just to cheer me up, deciding that other considerations demanded your attention more, you would not have done something wrong.

What explains the difference? It is not merely that in the one case you voluntarily consented to meet me. There are, after all, moral obligations that you have toward me without your consent. Although it would not be wrong for you to ignore my solitary meal, it would be wrong for you to walk over and knock my lunch off the table, no matter that you never agreed to withhold from so doing. Nor is the issue settled by the relative importance of your action. Suppose that in the case of promised lunch, you know that I would be equally content to have lunch alone as to share lunch with you. It might mean much less to me that you meet me in the promised case than in the lonely case. Still, it seems that the relative unimportance of meeting me does not cancel the obligation in the promised case, nor does the relative importance create an obligation in the lonely case. ¹⁸

I will describe five primary features separating moral obligation from other types of moral reasons. First, obligations have a kind of phenomenological urgency. We feel that we “must” comply with obligations, whereas we may regard other reasons as merely enticing or permitting action. Obligations’ must-ness gives the experience of being obligated a special character, involving urgency and a sense of priority over other deliberative concerns. ¹⁹ Obligations have a special “peremptory character” within our moral and practical experience. ²⁰ At a minimum, noncompliance with an obligation seems deliberatively inappropriate in a wide variety of cases. We would ordinarily not entertain failing in an obligation, even if we might willfully withhold from acting on moral reasons in other cases.

Second and related, obligations have a deontological character. We ought not break obligations, even when doing so would result in greater promotion of value overall. Even in cases where noncompliance with an obligation would secure greater overall compliance with like obligations by similarly situated agents, we can continue to believe that fulfilling the obligation would be right and not fulfilling the obligation would be wrong. ²¹

Third, obligations are weighty. Obligations tend to be associated with conclusive reasons for action, such that practical reasons favoring compliance with the obligation outweigh those reasons favoring noncompliance. ²² This—at least generally—remains true even as the weight of opposing reasons increases. Thus, obligation carries an idea of “presumptively decisive reasons for action.” ²³

Fourth, obligations are strict, in that they inform what an agent ought to do, all things considered. ²⁴ Sometimes, an action might be recommended from the moral point of view, but an agent might still have moral discretion about whether to perform that action. The moral point of view is not the only perspective from which agents may deliberate and, sometimes, other considerations can have a greater say in what is overall practically rational. To say that an agent is obligated, though, is strict in that it lays claim on an “overall” or “authoritative” ought rather than a defeasible one.

Fifth, obligations carry in tow certain accountability relations. ²⁵ If I am obligated to perform some action, then other members of the moral community are entitled to hold me accountable. It may be appropriate to demand that I conform to the obligation as well as to blame me or hold various reactive attitudes toward me if I fail to conform. ²⁶ In this way, obligations affect our moral relationships with other persons and mediate the appropriateness conditions for our most serious moral emotions. If you do not show up to meet me for lunch after promising you would, I can hold you accountable by asking for an explanation, resenting you if you have none, or changing my relationship with you by declining to make plans with you in the future.

These five conditions are not meant to provide an analysis of obligation. They are just meant to describe features characteristic of our moral practice. Nor are they meant to be proprietary. I do not claim that the word ‘obligation’ must be associated with these phenomena. However, I suggest that these phenomena merit philosophical reflection. If the reader wishes to reserve the word ‘obligation’ for some other aspect of morality, that will not bear on the argument that follows. In any case, the authors from whom I take these criteria largely affirm the existence of associative obligations, which should help allay any suspicion that they are frontloading a bias against my opponent.

I will make two further points of conceptual bookkeeping. The first is to underscore the difference between moral reason and moral obligation. That you have moral reason to perform some action is insufficient to establish that you are morally obligated to perform it. Even if you have sufficient or decisive moral reason to perform some action, you may not be morally obligated to perform it. In the case of lonely lunch, you may determine that you have moral reason to join me for lunch and perhaps even decisive moral reason. However, you may still not be morally required to join me. In her helpful analysis, Elizabeth Harman refers to such conditions as those which allow for “morally permissible moral mistakes.” ²⁷ That is, if you do not join me, you will be acting contrary to what you have most (moral) reason to do, but you will not be violating any moral requirement. This is a relatively weak distinction. It requires only that there are possible cases in which one has decisive moral reason to do something but is not morally required to do it.

The second point is about how the criteria described in this section are connected to each other. One possibility is that the conditions are individually necessary and jointly sufficient for the presence of an obligation. A candidate obligation must be phenomenologically urgent, deontological, weighty, strict, and display accountability relations to count as an obligation at all. An alternative, pluralist model allows that the extent to which some action is morally obligated could be partial but incomplete. ²⁸ Perhaps there are some candidate obligations that share some but not all features of the concept. Whether to valorize such candidates with the term ‘obligation’ would be a further theoretical question.

Against associative obligation

The previous section described the concept of moral obligation. This section will consider the case for associative moral obligations. Of course, it is not possible to review all the possible strategies for defending associative obligation. Instead, the present section will investigate several ways of filling in a standard argument template favoring associative obligation. This argument-type first describes some value that obtains in virtue of a nonvoluntary association, and then argues that this value is sufficient to create associative obligations. Schematically:

1. (1) Agent X is a member of association, A.

2. (2) Association A creates some value, V.

3. (3) If V obtains, X has associative obligations to members of A.

4. (4) So, X has associative obligations.

What varies, in a variety of different arguments for associative obligations, is the nature of the value created by the association. This section will proceed by considering a few leading ways of filling in these details, arguing that they are inadequate to ground moral obligations. If leading contenders fail in the same way, there is reason to doubt that some other alternative will succeed.

Associative obligations and the identity-shaping condition

A common proposal treats one’s political community as the relevant nonvoluntary association and argues that the relevant value is the community’s formative effect on one’s identity. On this proposal, entrepreneurs have associative obligations to their political community because their identities are shaped by those communities. Consider a few cases:

1. (a) Ben & Jerry’s supports fair trade, using Fairtrade Certified ingredients in its ice cream since 2010. The company also emphasizes a tradition of hiring refugees, extending to a practice of hiring Bosnian refugees in Burlington, Vermont, in the early 1990s. ²⁹

2. (b) Verizon has made public commitments to maintain donations to a relief fund in the aftermath of Hurricane Sandy, which affected areas in New York and New Jersey near the company’s headquarters. ³⁰

3. (c) Wegmans’ corporate giving prioritizes hunger relief, emphasizing its identity as a food provider. ³¹

As these examples illustrate, it is common for corporations to prioritize giving to communities with whom they have an issue-based, historical, or geographic connection. All of these are examples of associations that in principle (and in practice) shape identities. The question posed by the associationist priority thesis is whether associations create not only reasons to give, but also full-blown moral obligations.

Associative theorists answer in the affirmative. They argue that a community’s effect on its members is not just a set of propositions (which they could, in principle, disavow) or a set of practices (which they could, in principle, abandon). Rather, the community shapes the very capacities that members have for practical reasoning. No matter how forcefully we try to leave, we are fated to use our reasoning capacities in that very decision and our capacities are partly formed by the community. In this meaningful sense, the community persists with us even in the leaving of it. ³² Against an imagined voluntarist interlocutor, Samuel Scheffler memorably puts the point:

Whether we like it or not, such relations help to define the contours of our lives, and influence the ways that we are seen both by ourselves and by others. Even those who sever or repudiate such ties—in so far as it is possible to do so—can never escape their influence or deprive them of all significance, for to have repudiated a personal tie is not the same as never having had it, and one does not nullify social bonds by rejecting them. One is, in other words, forever the person who has rejected or repudiated those bonds; one cannot make oneself into a person who lacked them from the outset. Thus, while some people travel enormous social distances in their lives, and while the possibility of so doing is something we have every reason to cherish, the idea that the significance of our personal ties and social affiliations is wholly dependent on our wills—that we are the supreme gatekeepers of our own identities—can only be regarded as a fantasy. ³³

Let us grant this point. Even in trying to leave a community, it continues to shape our identities in the form either of our rational capacities or in the form of “significant” elements of our personal histories. Our community ties are, in any of these ways, inescapable.

What should this mean for obligation? Given that we have conceded that our “personal ties” to community have ongoing “significance” for our lives (understood in one of the ways noted above), it might seem like this significance would provide us with reasons for action that meet the phenomenological urgency condition. However, recall that the motivation in question was distinctively normative. It was not just that an experience or personal history was important, but that our experience of perceived reasons was motivating. And the identity-shaping condition, for all we know, will not meet this standard. I might recognize that my community has influenced my values or social practices, but it cannot be inferred that I should adopt an appreciative response toward that influence. Compatible with the identity-shaping condition, I could have indifference between my community and any other or I could rationally prefer other communities. The same could be said, mutatis mutandis, for the community’s inescapable influence on my rational capacities. Although I might only be able to use the rational capacities I now have, it does not follow that I ought to appreciate that I have those capacities rather than some other capacities. Indeed, it does not follow that I have to value those capacities at all. I might believe that an idealized version of myself would have some entirely different set of capacities. ³⁴ In short, the identity-shaping condition only shows that certain features of one’s identity are inescapable, but that is insufficient for any necessary normativity. In a slogan, inescapability does not yield justification. ³⁵

Once this is seen, then it seems unlikely that the identity-shaping condition can meet any of the other criteria for obligation. If there is no necessary connection to normativity, then there is certainly no guarantee of deontic character. It might seem that the weightiness condition is met because of the significance of the community’s impact, but such a suspicion would again confuse “significance” with normativity. Finally, the identity-shaping condition also does not meet the accountability-relations condition. As Scheffler himself suggests, we have “every reason to cherish” the ability of persons to travel “enormous social distances in their lives.” If the identity-shaping condition empowered those in our communities to demand that we remain nonvoluntarily bound to them by networks of obligation, it certainly seems that the radius of such social movements would be constricted considerably. But more to the point, if we do not have normative reasons in virtue of the identity-shaping condition (assuming that the analysis above is right), then there is no apparent basis on which people in our community could hold us accountable to them for obligations beyond those owed to any person. In short, there are no grounds for supposing the identity-shaping condition creates any new accountability relations.

Associative obligations and the identifying condition

Perhaps a defender of the identity-shaping condition would counter that the relevant sense of identity is, itself, normatively loaded. In other words, one might propose that if something is part of my identity, then it follows trivially that it provides me with reasons to act. This avenue is worth exploring. Several philosophers have maintained that if our identities can be shaped in ways beyond our control, then our obligations also extend beyond our voluntary control. Alluding approvingly to the passage from Scheffler above, Jay Wallace writes that

it is something of a fantasy to suppose that we could have complete voluntary control over all important features of our identity and social connections; once we are clear about this, the idea that we might have complete voluntary control over all the special obligations that our relationships might bring in their train begins to seem similarly unrealistic. ³⁶

But if identity just refers to nonnormative features—however significant—then it does not seem at all unrealistic to think that they would not impose obligations on us. We lack complete voluntary control over many things—death, taxes, weather—but that seems neither here or there from the point of view of our obligations.

To put the skeptical point another way, take some subdomain of obligation that we normally believe to be subject to our voluntary control, for example, promising. Emphasizing that there are many important features of my identity over which I do not have voluntary control does nothing to undermine the belief that I do have voluntary control over when and to whom I make promises. Confidence of voluntary control in one domain is not undermined by recognizing a lack of voluntary control in some other domain, so long as I have no reason to believe the domains are connected in some way. If I do not take the relevant features of my identity over which I lack voluntary control to provide me with normative reasons, then that issue is orthogonal to my obligations.

For this reason, it might be better to understand associativist intuitions as being about identity, understood in a distinctively normative register. On this view, my identity just is a source of practical reasons, such that if I identify with X, then I have reasons to promote, preserve, honor, or in some other way appreciate X. ³⁷ To contrast with the use of the identity-shaping condition discussed in the previous subsection, I will refer to this as the identifying condition, where “identifying” is sufficient for “providing reasons for action.” With this in mind, we can reformulate the schematic argument from the beginning of the section:

1. (1) Agent X is a member of association A.

2. (2) In virtue of being part of A, X identifies with A.

3. (3) If X identifies with A, X has associative obligations to members of A.

4. (4) So, X has associative obligations.

This leaves (1) as before, but revises (2). As we have seen, this condition is stronger than the identity-shaping condition because it connects identity with normative reasons.

For the revised version of the argument, premise (3) will be crucial. (Let us grant that there are in fact many nonvoluntary associations that people do, in fact, identify with.) So it is not surprising that theorists of associative obligation present defenses of this premise. A standard defense points to valuable features of our political communities, families, or other nonvoluntary groups. These groups provide us with a sense of achievement in collective histories of which we can see ourselves as a part. They thereby provide occasions for personal pride or, in the case of collective moral mistakes, a sense of shame. In these ways, nonvoluntary associations are a locus of emotional depth, with concomitant moral significance. ³⁸ They provide important moral goods, including a sense of “home” in a given community, as well as social practices and traditions that can provide belonging or a sense of “home” in time. ³⁹

Considerations like these are likewise prominent in examples of corporate giving such as those described at the start of this section. The history that Ben & Jerry’s has with their community in Vermont shapes how the company identifies its values in an ongoing way. Verizon’s history with an area affected by Hurricane Sandy relates its corporate “home” to its altruistic practices. Wegmans’ employees’ emotional connections to food insecurity drive their corporate priorities. This is just to say that association again appears to be a plausible place to consider whether there are nonvoluntary obligations on social entrepreneurs.

In virtue of the group’s effects on our “self-understanding,” as well as our “moral sentiments, emotions, and attitudes,” we identify (perhaps inescapably) with it. ⁴⁰ Perhaps, though, our identification is not inescapable. Massimo Renzo suggests that “[i]f we do not identify with the polity in which we happen to be, we cannot be said to be its membership.” Nevertheless, “conditional on the fact that we endorse our membership,” we remain bound by associative obligations to the polity. ⁴¹ While a political state treating its members as obligated to obey does not depend on their attitudes, this view proposes that members in fact being obligated does. ⁴² As Renzo’s version nicely illustrates, associative theorists are divided on whether we can alienate associative obligations (as Renzo’s quasi-voluntarist position affirms) or whether we cannot alter the groups with which we identify, and so to which we are obligated (as the nonvoluntarist believes). In either case, identifying is thought to constitute a sufficient condition for obligation. If it is sufficient, then premise (3) is true.

This elaboration has enriched the content of premise (3) considerably, adding details about how the practice really is valuable, about the content of its value, and so on. Granting the version of this story the associative theorist most prefers, we can still ask why identification yields obligation. Consider the four criteria of obligation outlined above. First, does the identifying condition involve phenomenological urgency? Its defenders believe that it does. On the associativist view, identifying with a cause, group, or person provides not only reasons for action, but also the sense of “must-ness” that characterizes this condition. Wallace imagines cases in which a friend is stranded at a bus stop, in which his wife’s car breaks down at work, or in which his mother is ill. He writes:

The situation I describe is one in which I have to do something, whether it is visiting my mother or picking up my wife or rescuing my friend; these are all actions it strikes me that I need to perform. Second, the basis for these obligations is apparently the simple fact that I stand in a relationship of love of some kind to the persons who are prospective beneficiaries of my actions in these cases. ⁴³

Wallace is here concerned with obligations of love, but because we have abstracted from specific types of association, the point will transfer to any associative case. The psychology he describes is urgently motivated in connection with certain relationships. He must go visit his mother, rescue his friend, and so on.

This characterization also shows how the identifying condition can meet the deontic status and weightiness conditions. We do not suppose that we should maximize visiting sick parents or rescuing friends, but rather that we should visit our sick parents and save our stranded friends. This agent-relativity points to the deontic character of the associated moral reasons. ⁴⁴ Likewise, Wallace’s examples show how the actions would also have a special weightiness or tendency to override competing considerations. Even if saving a friend is very inconvenient or trades off with other goods (including moral goods), it seems likely that it would still—at least in many cases—take precedence. Similarly, a social entrepreneur might prioritize helping immigrants in their community over helping new members of other communities.

The final question is whether the identifying condition can set up accountability relations. Here the writing of associative theorists is less clear. For example, consider John Horton, writing about the family as an analogue to political society:

Moreover, it is very hard to see how any such institution could entirely eschew some sort of ethical bond between its members (including obligations of mature children towards those who have been responsible for raising them). But, even if this could somehow be envisaged, it is still harder to see why anyone would want to insist that we must reject the thought that membership of such groups can give rise to obligations. For it certainly seems that not only are non-voluntary groups integral to social life, they can (and often do) have value for us. And this is at least sufficient to support the claim that we can intelligibly and defensibly understand ourselves to be ethically bound to non-voluntary groups. ⁴⁵

On the face of it, Horton’s final inference is not obvious. The fact that some good has value for us is not generally sufficient to establish that we are ethically bound to it. Things I find valuable to me include Trader Joe’s, tennis, and the writings of Thomas Nagel—but I think it would be borderline unintelligible to say I was “ethically bound” to any of these.

So can identifying ground obligations? On this issue Horton moves from affirming “some sort of ethical bond” to the explicit invocation of “obligations.” Such conceptual movement may be telling. Recall from above that one can have moral reason (including decisive moral reason) to perform some action without thereby being obligated to do it. Taking that point as given, it is now unclear that we can infer obligations from “some sort of ethical bond.” If we share some relationship, I might thereby be permitted to act for your well-being when it is in conflict with that of other persons or I might have sufficient practical reason to support your ends rather than the ends of others, and so on. It is natural to describe such cases as instances of “ethical bonds.” However, compatible with such bonds, I may yet have no obligations whatsoever to you. We cannot move from speaking of “some sort of ethical bond” directly to “moral obligations.”

To meet the accountability-relations condition of obligation, there must be some other member of the moral community in position to appropriately demand compliance with the obligation. Both Wallace’s and Horton’s passages eschew this question by framing the issue from the point of view of the acting agent rather than from the point of view of the putative obligation holder. This rhetorical framing emphasizes the condition of phenomenological urgency: What must I do? The literature is nearly uniform in considering obligation from this angle. Consider Simon Keller, in his subtle treatment of filial duty: “[W]hen you are uniquely placed to provide someone with an important good, you have a moral reason to do so, at least in some sense and other things being equal.” ⁴⁶ Also consider Niko Kolodny on duties of solidarity: “If I were to betray the culture, for example, I would have reason to feel not only that I had betrayed it, but also that I had betrayed them.” ⁴⁷ By calling our attention to the fit between the deliberating agent’s point of view and the phenomenology of obligation, this maneuver avoids the question of how the receiving agent’s point of view fits with the concept of obligation.

To see how this matters, it will help to clarify the argumentative role that these intuitions are meant to fill. The intuition that I must perform some action for another person is intended to support the conclusion that I am obligated to perform it. If my experience of “must-ness” is not preceded by any voluntary assent to the obligation, it seems that I am nonvoluntarily obligated. And, if we accept that accountability relations are one aspect of obligation, then it should not matter whether we consider this action from the point of view of the acting agent or the receiving agent.

However, crucially, there are cases where this symmetry does not hold. That is, there are cases in which, considered from the point of view of the acting agent, it seems that we must do something, but, considered from the point of view of the receiving agent, it seems that we are not entitled to demand the action be done. The associative theorists’ own cases can be redeployed to make this point. Recall the case of the stranded friend at the bus stop, but now suppose you are the person stranded rather than the friend at home. Would it be acceptable to call a friend and demand that they come to your rescue? Certainly, I might permissibly call and ask for help, but—at least in many cases—I will want to insist that I can get out some other way if my friend cannot make it. Indeed, I will try to assure my friend that she should not feel like she must come, and so on. ⁴⁸

The present point is that there is no contradiction between feeling like I must help, when considering a case from the point of view of the candidate assisting agent, but also feeling like I must not demand help, considering the same case from the point of view of the candidate recipient. If I am a morally sensitive person with strong moral reason to perform an action for another person and no other weighty competing reasons, it may feel like I should not even consider not helping. All of this may be true, even if no one is in position to demand that I act or resent me if I fail to act. The experience of obligation is phenomenologically urgent, but some non-obligations can have just that character. And in fact, such an asymmetry between acting and receiving agents is psychologically commonplace. Subjects in social-psychological studies express greater willingness to offer assistance than to demand it—including, interestingly—assistance in the form of picking up a friend at the train station. ⁴⁹

Perhaps the associative theorist might grant this asymmetry but maintain that it is not explained by a lack of obligation to help. For example, on the one hand, one might hesitate to ask for help, not knowing what important activities in which the other person might be engaged. On the other hand, I face no corresponding uncertainty when deciding to offer aid to another. Although this difference between the cases could sometimes be relevant, I do not believe it will eliminate the asymmetry between offering and asking for aid. One can imagine that a friend has nothing pressing to do, and you can still feel reluctant to impose on their time. Likewise, it is easy to think that I might feel strongly motivated to provide aid, even if it takes me away from other pressing concerns.

Perhaps some uncertainty about the right explanation for the asymmetry may linger. Even if the reader declines to accept the foregoing diagnosis of the bus stop or train stop cases, all that is necessary is that there be some possible cases in which one would feel that one must act, if adopting the point of view of the acting agent, but would not feel permitted to demand action from the point of view of the recipient. If there are even a few such cases, then the principle connecting the experience of must-ness to the fact of obligation is mistaken. If there even might be such cases (for all we know), then theorists are not entitled to move from the intuition of must-ness to the conclusion of obligation.

Consider Keller’s or Kolodny’s principles, except from the direction of the recipient. When someone is uniquely placed to provide you with an important good, are you in a position, other things being equal, to demand that they do so? Suppose that someone has a “personal history” of engaging with your culture. Are you entitled to require of them that they continue to engage with your culture as well as “seek to preserve it”? That these questions at least do not feel symmetrical should raise a cautionary flag. These authors provide compelling arguments to believe in moral reasons to help one’s parents or preserve one’s culture—but these moral reasons may not actually yield moral obligations. ⁵⁰ In the case of corporate giving, social entrepreneurs may have strong moral reasons to help fellow members of their communities without being obligated to prioritize their own histories, groups, or values above anyone else’s.

Notice how this contrasts with paradigm cases of obligation. If your friend has promised to meet you for lunch but has not turned up and your food is getting cold, there is nothing out of sorts with calling the friend and asking, “Where are you?!” and thereby registering an implicit moral demand. In Stephen Darwall’s favored case, if someone is standing on your foot, you are entitled to ask them to get off your foot; they are obligated to respect your rights over your physical body. ⁵¹ Demands need not be impolite or gauche. Asking whether a demand can be made is simply a way of testing whether one has “second-personal authority” to hold another agent accountable for complying with moral reasons that they have. If so, then this helps to vindicate the language of obligation. Such accountability relations do not always track motivation. If anything, the intuition against issuing demands feels even stronger in cases in which the recipient of the demand might feel a phenomenologically urgent pull to provide help. If I have missed the bus and call a loyal friend, I will likely take greater precaution in canceling any implicature that my plight establishes a warranted demand.

I underscore that the associative theorist is free to disagree about these cases. Regardless of intuitions, though, the associative theorist is not actually entitled to the claim that individuals can “legitimately regard themselves as having obligations toward the group” without either (1) considering the question from both actor and recipient angles, or (2) introducing a weaker analysis of obligation. ⁵²

Associative obligations and justice

Another option the associative theorist might take is to argue that meeting three or four of my five conditions is sufficient to invoke the language of obligation, even if the accountability-relations condition is not satisfied. “Certainly, by meeting three of the conditions,” the associative theorist might venture, “associative moral reasons share some degree of membership in the category of obligation.” I doubt that this pluralist strategy can succeed. The remainder of this section will elaborate why not.

Having observed how our associative moral reasons are like and unlike obligations, the question that remains is whether we ought to apply the concept of obligation to them. There may not be any answers to such questions apart from the particular aims of a given theorist. ⁵³ Fit with folk-psychological usage is only one desideratum among a variety of philosophical goals. If using a concept has explanatory value within a given theory, then this value can work to justify the concept’s inclusion. So it may help to investigate the theoretical aims of associative theorists.

In general, associative theorists are interested in grounding special obligations to other citizens as well as obligations to the state and its institutions. ⁵⁴ Whether the bearers of putative obligations are individuals or corporations, the result is that associatively grounded duties will form part of the structure of justice. The theoretical purpose is to have associative obligations explain how citizens will have rights of justice with respect to each other and with respect to the state itself. The association figures in the explanation of why such obligations are owed to some persons (fellow citizens) and to one’s own state rather than to other persons or other states. Because of this connection to claims of justice, associative obligations are conceptually linked to the rights that citizens enjoy. If some citizens have claims of justice against others, then they are entitled not only to demand them morally, but also to use the power of the state to bring about their satisfaction. Thus, rights have a conceptually bipolar structure, connecting the claimant of the right to the addressee of the right. ⁵⁵ Moreover, the accountability relation is even stronger in the case of rights than in the case of moral obligations generally. This is why theorists of obligation appeal to the rights-like structure of obligation when explicating the accountability relation as a conceptual platitude about obligation. ⁵⁶

If associative obligations are supposed to ground claims of justice, the associative theorist’s own purposes would be ill-served by discharging accountability relations as a feature of obligation. Under that condition, the associative obligations between citizens would not entitle anyone to make moral demands, much less coercively enforceable moral demands. Thus, the notion of associative obligation could not ground the legal or political obligations of citizens, because these two uses of “obligation” would invoke different concepts. The ambiguity in “obligations” would invalidate any argument attempting to ground the political conception of obligation in the associative conception. And this would crucially undermine the associative theorist’s own program.

Even more important, though, are the specific types of legal rights and correlative duties thought to be grounded in associative obligation. In particular, these rights and duties are especially weighty and morally demanding. Consider, for example, the idea that associative obligations to co-nationals permit (or require) us to restrict the movement of outsiders into our political community. The moral issues involved are serious. Threatening force against others under circumstances of extreme inequality or absolute deprivation must require a compelling justification. ⁵⁷ Consider, instead, the proposal that associative obligations sometimes justify using lethal force against even just combatants. ⁵⁸ Seth Lazar puts it directly: “A theory of associative duties should be able to justify a duty to risk one’s life.” ⁵⁹ Consider, again, the corporate case. If a social entrepreneur withholds giving to a needier group to prioritize an associated group, they may be giving in a way that produces considerably less impartial value. These are strong conclusions. If the idea of association is going to help yield them, then it has some heavy moral lifting to do. Part of the seriousness of these obligations is that there are weighty opposing moral reasons held by individuals, so it must be possible to justify the conclusions to them. Once again, this brings us back to the necessity of accountability to other persons who have an authoritative standing.

In light of the specific purposes of associative theorists, the concept of associative obligation cannot discharge the accountability-relations condition. But neither, I have argued, can associative theorists satisfy this condition. The reason is that there is no sense of “identifying” strong enough to yield this conclusion. In this section I have been happy to let the content of “identifying” vary among a variety of different specifications. At this point, the rationale should finally be clearer. No matter how strong associative theorists make “identifying” out to be, it remains an essentially internal activity. Whether voluntary or decided by social circumstances controllable only in fantasy, our “identifying” with some group, person, or practice is effected within our own mental lives quite independently of the other agents with whom we identify. Even if our so identifying rises to the level of an internal commitment to a group or its practices, this may still not secure the conditions of obligation between us and them. ⁶⁰ Ruth Chang helpfully contrasts the type of “internal commitment” created through an act of identification with a “moral commitment,” which requires the “uptake” of another agent. ⁶¹

The lack of uptake helps to explain why the putative obligation holder is not intuitively ever in position to demand action from the putatively obligated party. But it is impossible to strengthen the identifying condition further. For once we add that the other party has uptake of the commitment, then it is no longer the association that generates the obligation, but rather the promise-like agreement between the parties constituted by the reciprocal commitment and uptake of commitment.

Why corporate social responsibility is not moral responsibility

At issue from the “Associative obligations in corporate social responsibility” section were two questions. First, should corporations prioritize doing good for their co-nationals or local governments above trying to maximize the impartial effectiveness of their altruistic activity? This is the question raised by the associationist priority thesis. Second, could corporations discharge obligations to co-nationals through business activity or must they accept some additional burdens?

The previous section challenged whether there are associative obligations. It may be that corporations have obligations to some stakeholders that they do not have to persons generally, but if the foregoing arguments are right, such obligations must be voluntarily accepted. If such obligations must be voluntarily accepted, then association itself is not a sufficient condition for moral obligation. And if association is insufficient for moral obligation, social entrepreneurs are not required to prioritize co-nationals or others with whom they share some identity above the impartial good. If there are no associative obligations, then the associationist priority thesis is false. Corporate social responsibility should not morally require prioritizing co-nationals or government entities. The reasons favoring effective altruism are thus not trumped by specific associative obligations, so I conclude that corporate altruism retains discretion in whom to prioritize.

What about the question of discharging obligations to co-nationals through business activity? Although I have challenged the existence of such obligations, we are also now in a position to say something internal to this debate as well. Recall the two first-personal cases from above:

Katy Perry: The singer Katy Perry records an album that provides a great benefit to a person at an important moment. However, it remains inappropriate for Perry to show up demanding additional compensation.

Reciprocal Brownies: A friend performs some act for you through which they assume a burden. It is appropriate to compensate them with brownies, though not with banana bread, if they like the former but not the latter.

One explanation for why it seems inappropriate for Perry to demand compensation is because Perry did not assume any burden in providing her music, because she had been compensated. This explanation is meant to favor the burden-based view, according to which obligations to co-nationals can only be discharged through assuming a burden. According to this explanation, there is an asymmetry between the Katy Perry case and the Reciprocal Brownies case. (I should acknowledge that I am here using my own glosses on the two cases.)

However, we are now in a position to consider a rival explanation. It is inappropriate for Perry to demand compensation while it is appropriate to pay back a friend in brownies because the cases are testing for conceptually distinct aspects of obligations. Whether Perry can demand compensation is a question of accountability relations. Whether it is appropriate to compensate the friend with brownies is a question of the weightiness of moral reasons. It may be that in cases where one has received a benefit from another person, there are weighty moral reasons to reciprocally compensate them in some way. However, compatible with such reasons, it may be inappropriate to demand compensation. As we have seen in the previous section, weighty moral reasons can sometimes be in place when accountability relations are not.

To compare, we should set up the cases symmetrically. In Reciprocal Brownies, imagine the friend showing up at your home, demanding compensation in the form of brownies for the burden they had incurred. For a wide variety of burdens incurred by friends, it may seem inappropriate to show up and issue such demands. Alternatively, imagine the fan who is deeply helped at a crucial moment by Katy Perry’s album. Imagine that rather than Perry showing up to demand compensation, the fan regards themselves as merely having weighty reasons to provide some benefit reciprocally to Perry, perhaps through loyally buying further albums, publicizing Perry’s helpful service, or writing a letter of gratitude. Any of these may seem like plausible responses.

My suggestion, then, is that none of these cases supports the burden-based view. The burden-based view tried to explain our intuitions about obligation with respect to whether an agent had assumed a burden in providing a benefit to another. But once we set up cases to investigate the same aspects of obligation, then whether a burden was incurred no longer looks important. If weighty reasons are considered, then plausibly one has reasons to compensate in both burden-assuming and non-burden-assuming cases. ⁶² If accountability reasons are considered, then neither case seems like one of genuine obligation. My tentative conclusion is that these cases do not favor the burden-based view, and so they are equally compatible with the view that corporate actors could act morally through ordinary business activity. Once again, this provides grounds for skepticism about associationist priority. Even if it is true that corporate actors have weighty reasons to compensate their co-nationals for acts of civic service, this does little to vindicate the associationist prioritiy thesis. After all, there may also be weighty moral reasons to engage impartially in effective altruism. So merely showing that there are weighty reasons to help co-nationals does little to show that corporate actors should prioritize co-nationals.

Conclusion

The central aim of this essay has been to challenge the case for associative obligations. Nonvoluntary associations may well ground significant moral reasons, but they do not ground the sort of obligations needed for the associative theorists’ aspirations. This has clear implications for how to prioritize corporate giving. It offers reason for skepticism that there are any stakeholders to whom corporate actors are nonvoluntarily obligated. Perhaps there are moral reasons to compensate or benefit these stakeholders, but admitting as much does little to establish either that benefiting stakeholders should be prioritized above other altruistic aims or that altruistic aims cannot be satisfied through business activities. If they do not have associative obligations, then corporate actors have more discretion about to whom they give and how they give.