PHILOSOPHERS often remark that many of our interpersonal relationships are valuable for their own sake. At the same time, even the most intimate of these relationships—friendships, family ties, and spiritual kinships—inevitably have instrumental value. Relationships provide stocks of goods that individuals can draw from to achieve a variety of ends. Friends might provide us with special advice and encouragement for a job interview, or have enough trust in us to become involved in a risky business venture. Membership within an association may generate the self-confidence that comes with having our projects recognized by others who share a similar conception of the good. The list could continue. Sociologists often refer to the goods that are distinctively produced by, and accessed through, interpersonal relationships and associations as “relational capital.”1 This term indicates two important aspects of relationship-derived goods: they are productive resources—a form of capital—precisely like economic resources. But they cannot be reduced to economic forms of capital—they are relational in kind.

In this article, I defend the novel claim that the just distribution of relational resources—resources that either distinctively exist within interpersonal relationships or are themselves constitutive of such relationships—across society should be regarded as a legitimate concern of resourcist theories of distributive justice, no different from the distribution of economic resources. 2 However, since relational resources, unlike economic resources, cannot be directly distributed or redistributed, and since these resources are generally attached to relationships which individuals should not be forced to enter against their will, I will propose that we regard opportunities for relational resources as the appropriate object of distribution (distribuenda).

If my account of opportunities for relational resources as distribuenda proves correct, it would broaden the still dominant understanding of what social arrangements should count as basic subjects of justice. Indeed, not only those institutions that control the production and distribution of economic resources and political liberties would count as basic, but also those arrangements, if they can be found, that are responsible for the production and distribution of opportunities for relational resources across society.

Expanding the range of appropriate distribuenda in the way I suggest might also strengthen the ability of liberal-egalitarian theories of distributive justice, in particular resourcist-Rawlsian theories, to give due consideration to several types of relational inequalities that afflict current societies, and which these theories have long been accused of neglecting. By incorporating relational goods within a resourcist distributive paradigm, my aim is not simply to rescue justice as fairness from long-standing criticisms, but also to embrace and further expand its critics' concerns.

One criticism comes from relational egalitarians and recognition theorists.3 In different ways, these strands of thought criticize purely distributive theories for neglecting relational inequalities. As an alternative or complement to the distributive paradigm, they propose a relational understanding of (in)equality, defined as the absence (presence) of social relations of hierarchy, domination, and misrecognition among citizens.4 In this article, I aim to bridge the gap between distributive and relational egalitarians, by providing a distributive theory of relational equality, understood as equality of opportunities for relational resources. This account incorporates within a distributive paradigm some of the non-distributive concerns underscored by relational egalitarians and theorists of recognition.

Beyond relational and recognition theories, feminists have also criticized theories of distributive justice, and especially Rawlsian theories, for neglecting relational concerns.5 In particular, they have argued that theories of justice should take dependency seriously, as a human condition, and that this requires recognizing relationships of care as both a basic need that ought to be satisfied and a social responsibility that ought to be more equally shared.

The account of relational goods as distribuenda that I aim to develop attempts both to accommodate these concerns within a resourcist paradigm and to further expand them. First, I wish to expand the range of relational goods that matter for distributive justice. While feminist and other literature overwhelmingly focus on care, I will, instead, conceive of relational resources as a more general and much wider category of goods, some of which are produced by non-intimate relationships.6

Second, whereas the feminist literature is mainly concerned with the value of care as an appropriate response to particular forms of dependency such as childhood, old age, and disability, I want to provide an account of relational goods as resources that even putatively autonomous, adult, and healthy individuals have reasons to want, whatever else they want. To be clear, my purpose is not to set aside concerns about dependency. It is rather to point out both that these concerns also apply to the putatively independent liberal subject and that dependency is not the only reason for why we should include relational resources within the metric of justice.7 For relational resources are also an important dimension of advantage—a form of capital. This provides us with a further reason to think about relationships in resourcist terms.

However, bringing opportunities for relational resources into the category of distribuenda is not an easy task. We first need a clear definition of what resources count as relational. We then need a moral justification for why including these resources into the metric of justice is a reasonable and desirable enterprise. Does it make sense to say that people are worse off or better off, from the perspective of distributive justice, depending on their access to relational resources—what kind of interpersonal relationships they have the opportunity to form and what kind of goods they are able to access through those relationships?

To address these issues is to answer what I call the metric question (Section I). I argue that although it might be easier for welfarists or perfectionists to include relational goods within their metric of justice, even those who adopt a non-welfarist and non-perfectionist, resourcist metric have good internal reasons to include (opportunities for) relational resources.

Once opportunities for relational resources have been included into the metric of justice, a further question arises, now concerning the institutional mechanism that ought to distribute them. To say that income and wealth count as distribuenda whose distribution ought to be governed by principles of justice is to say that the mechanisms for redistributing income and wealth—e.g., the taxation system—ought to be set up according to egalitarian principles of justice. Yet what are the institutional mechanisms responsible for “producing and distributing” relational opportunities? I call this question the institutional question (Section II).

I argue that relationships, including intimate relationships and voluntary associations, as well as the resources they provide, must in many crucial respects be regarded as socially distributed resources. They have identifiable institutional social bases. These institutional bases—what I will call “the relational distributive structure”—ought to be arranged according to principles of justice.

We are then left with a third and final question. What are the principles of justice that ought to govern the distribution of opportunities for relational resources and thus apply to the relational distributive structure? I refer to this as the principle question (Section III).

I argue, quite uncontroversially, that the freedom to form and participate in certain kinds of relationships should be distributed equally to all. However, I contend, more controversially, that this freedom should be understood as a positive rather than negative liberty. I add that the distribution of opportunities for relational resources ought to be regulated by a principle of fair equality of relational opportunity, properly defined. Finally, I will suggest, but leave open to further investigation, a possible way of indexing opportunities for relational resources.

The Metric Question

  1. Top of page
  2. The Metric Question
  3. The Institutional Question
  4. The Principle Question
  5. Conclusion

Defining Relational Resources

Should opportunities for relational resources be included in the metric of distributive justice? We cannot answer this question without a clear definition of what counts as a relational resource and as an appropriate metric of justice.

What is a relational resource? Let us start with what a relational resource is not. It is not a good whose production or consumption is simply facilitated by relationships. For example, we can more easily produce food by cooking together with friends. Yet, food is not a relational resource. Relational resources are rather goods that are distinctively produced through and available within relationships or that are themselves constitutive of certain relationships.8

By “relationships” I refer to ongoing and coordinated interactions among two or more persons through which some goods are produced and exchanged. An interaction is ongoing if it is repeated with a certain degree of regularity across an extended amount of time. It is coordinated if it is governed by formal or informal norms or established patterns of mutual expectation. The term relationship, as used here, refers to intimate relationships such as relationships of familial love, friendship, and kinship, as well as non-intimate relationships between colleagues, neighbors, members of voluntary associations, and participants in a common social network. It does not refer to, however, the more abstract relation of political citizenship.

In economic terms, relational resources often take the form of local public goods, insofar as the consumption of these goods by one of the parties within a particular relationship does not generally reduce the consumption of the same good by the other parties within that same relationship.9 Further, people cannot generally enjoy these goods without participating in the relationships and networks that produce and reproduce them. Finally, relational resources cannot be directly bought or sold.10

To further clarify what makes a good a relational one, let us start by considering a good sociologists often refer to as a paradigmatic component of relational capital—interpersonal trust, that is, the act of trusting and being trusted.11 Is trust a relational resource? We cannot generally consume someone's trust (being trusted) without participating in the relationship that builds it. Trust can neither be bought nor acquired by contract—we cannot willingly decide to trust.12 Finally, trust can be understood as a relational resource insofar as ongoing and reiterated relations are privileged sources of interpersonal trust (as different from simple reliance).13 Let me explain.

Trust is about having certain beliefs, expectations, and attitudes with regards to other people's willingness to act in the way we entrust them to act.14 In order for reasonable trust to be possible the truster must have good reasons to believe that the trusted is (a) willing to act in, and directly moved by, the interest of the truster when need be, even if this interest may conflict or stop aligning with the trusted's self-interest. The truster must also have a confident expectation that (b) the trusted has the relevant capacity or knowledge to do what she is entrusted to do, and to do so in a way that aligns with the truster's interpretation of the task at hand. Because the truster's expectation can never be fully confident, trust also requires what Karen Jones calls (c) “a distinctive, and affectively loaded, way of seeing the one trusted” that explains “the willingness of trusters to let those trusted get dangerously near the things they care about.”15

Condition (a) provides a first explanation for why ongoing interpersonal relations, especially affective relationships, are privileged sources of trust. Even if there are a number of reasons why we may believe that a person would give significant weight in her deliberation to the fact that we are counting on her—even when this may conflict with her own self-interest, as she may just be a truly benevolent person—chances are that when we deal with strangers we expect their goodwill to be limited in substantive ways.16 By contrast, it is generally believed that certain relationships provide people with special moral reasons and direct motivations to do things for each other that they have no comparable reason to do for others. These are usually those relationships that we value non-instrumentally.17 Because of this, we have stronger reasons to trust those we share (or we reasonably believe we share) a special relationship with than we have to trust strangers or people we merely interact with.

Conditions (b) and (c) provide an additional explanation for why trust is a relational resource. Consider the case of a single parent who must entrust her four-year-old child to the care of someone else (S) for an entire week, in order to visit a very sick relative who lives abroad. It is reasonable for the parent to entrust the child, only if she has the confident expectation that S (i) has the competence to take care of the child, (ii) is willing to do so in a way that aligns with the parent's own conception of “appropriate care,”18 and (iii) will continue to do so even in the event that her own interests stop aligning with the ones of the parent. For the parent's expectation to be fully confident in these respects, she not only needs relevant information about S's expertise and conception of care. She also must have full confidence in S's goodwill toward her.19 This confidence requires an emotionally-loaded “attitude of optimism” about the trusted's goodwill which is very difficult to obtain without the support of reiterated interpersonal relationships.20 Interpersonal trust, unlike simple reliance upon the expertise of strangers or institutions, is thus a relational resource. 21 Discussing the case of interpersonal trust thus helps us clarify what it is that makes a good a relational one. However, as I will further clarify below, trust is, by no means, the only relational resource. With a provisional definition of relational resources in hand, we can now turn to define the metric of justice.

The Metric of Justice

For welfarists who believe that social institutions should be arranged so as to maximize individuals' welfare or utility, the argument that opportunities for relational resources are appropriate objects of distribution may arguably not present much difficulty. The same is true also for perfectionists who believe that institutions should be arranged so as to provide individuals with what they need to live a fully flourishing life. For both the claim that participation in interpersonal relationships generally increases individuals' welfare, and the claim that they contribute to human flourishing, seem empirically plausible and intuitively appealing.22

Hybrid metrics such as the capability approach, according to which the unit of interpersonal comparisons is constituted by effective freedoms to achieve certain valuable human functionings, are also well suited to incorporate relational aspects. For example, Martha Nussbaum claims that “unlike liberal proceduralist approaches, the capability approach is explicitly committed to a prominent place for love and care as important goals of social planning.”23 Further, she explicitly argues that the capability for affiliation—the capability to form and enjoy meaningful attachments and relationships—has an architectonical role within her list of capabilities.24 Although Nussbaum does not provide a full account of what kind of relationships and relational goods the capability for affiliation supports, her work certainly constitutes an important attempt to account for relational concerns within a non-welfarist and (arguably) non-perfectionist metric of distributive justice.

However, political liberals reject perfectionist and welfarist metrics, and many of them are also unconvinced by the capability approach, in its different variants. They fear that the conception of well-being upon which the capability metric rests is too thick, and arguably comprehensive, for qualifying as a truly “political” conception. Therefore, in order to provide political liberals with an internal reason to regard relational goods as appropriate distribuenda, it is wise to try and carve out a space for these goods within the metric of justice they themselves adopt. From now on, I shall then focus on the metric of justice developed by John Rawls himself, while acknowledging that other metrics may have their own internal resources to incorporate relational concerns.

Rawls adopts, as it is well-known, a list of social primary goods as the appropriate metric of justice. Unlike most welfarist metrics, primary goods are meant to constitute an objective metric of justice. They are objective insofar as they “are not intended as a measure of citizens' expected overall psychological wellbeing, or of their utility.”25 Unlike other metrics, their justification is (arguably) grounded in a political rather than comprehensive or perfectionist conception of citizens' interests. In other words, they are presented as “clearly not anyone's idea of the basic values of human life and must not be so understood, however essential their possession.”26

Primary goods can be regarded as material correlates (resources) of the moral powers of persons, qua political citizens. They are “things needed and required by persons seen in the light of the political conception of persons, as citizens who are fully cooperating members of society, and not merely as human beings apart from any normative conception.”27 They count as “needed” insofar as they are generally necessary, although not sufficient, for citizens to develop and exercise two distinctive moral powers and to pursue their conceptions of the good. The first moral power consists in “the capacity to form, to revise and rationally to pursue a conception of one's rational advantage or good.” The second moral power consists in a sense of justice, that is, in the capacity and willingness to understand, to apply, and to act from the principles of justice.28 Primary goods can also be seen as “all-purpose means,” that is, as resources valuable across a variety of conceptions of the good that make possible individuals' pursuit of their ends. Rawls claims that “[w]ith more of these goods men can generally be assured of greater success in carrying out their intentions and in advancing their ends whatever these ends may be.”29 The concept of primary goods includes both natural goods, such as health, as well as social goods, that is, goods whose distribution is under the direct control of social institutions. Rawls's metric only includes social primary goods—civil and political liberties, income, and wealth, opportunities to access positions of authority and responsibility, as well as the social bases of self-respect. The point of justice as fairness is to achieve a fair distribution of expectations for these goods across an entire life.

Therefore, in order to determine whether (and what) relational resources count as primary goods, we need to establish whether these goods are: (1) generally necessary for the development and exercise of (at least one of) the two moral powers; and (2) valuable across a variety of conceptions of the good, without their value being grounded in any such a conception. In order to determine whether relational resources count as social primary goods we then need to investigate (3) what role institutions play in the distribution of these resources. We can answer the question of whether a certain good should be included, as a matter of principle, in the metric of justice, separately from and before answering the further question of (4) whether and how this metric can be operationalized for institutional purposes.

Relational Resources and the Capacity to Pursue Ends

Are relational resources generally necessary to develop and exercise the capacity “rationally to pursue a conception of one's rational advantage or good”? Can these goods be regarded as “all-purpose means”? Let us start by considering the case of interpersonal trust—a good that, I argued, counts as relational in a relevant sense. To provide a comprehensive assessment of the role that trust plays in supporting people's capacity to pursue ends would require a separate paper. Here I limit myself to a few considerations.

Consider first the role that trust plays in protecting whatever it is that people care about in their everyday life. For example, since I cannot possibly take my children with me on my work travels at all times, their care must sometimes be transferred to others I trust.30 Without trust what we care about would be unsafe. The sense of security provided by trust is essential to the ability to pursue our ends with stability, whatever these ends may be.

Second, as some sociologists suggest, trust constitutes the most “effective form of complexity reduction.”31 We cannot by ourselves handle the enormous complexity of our modern lives without having to sacrifice our most important projects. We must thus delegate tasks and responsibilities to others in order to pursue our ends. Yet, if we cannot trust them (or their advice) we cannot be reasonably expected to transfer important responsibilities to them.

Note, I am not arguing that if I do not trust others, I cannot transfer any responsibility to them. For example, I can reasonably transfer the task of making photocopies to a student I have never met who works in the front office of my department. I do not trust him, as I might not even know him. Note however that this is a case where the task is not in itself a fundamental project, or not in itself so vital for my overall project. The cost of failure is not very high. Further, there are not many competing interests and/or interpretations about how the task should be performed.32 It is in those cases in which the project is important, the cost of failure is high, and interests concerning or interpretations about the task at hand must align, that trust becomes generally necessary for responsibly transferring responsibility without sacrificing our most important projects.

Third, trust is generally necessary to produce stable patterns of expectations.33 Trust, for example, allows us to rely on promises. By doing so, trust reduces uncertainty in our interactions with others and helps us assess the risks attached to our life plans. Trust is thus generally necessary to secure those conditions of stability and predictability that are, in turn, necessary for us to pursue a life plan over time, whatever this may be.

Beyond trust there are other goods that count as relational, according to my definition, and that plausibly figure as all-purpose resources. Emotional support and care are exemplar relational resources. They are not only distinctively produced by and available within face-to-face, ongoing relations, but also consist of relationships between subjects. That care constitutes, in a relevant sense, an all-purpose mean, has already been successfully argued by care ethicists and other scholars.34 Psychologists also define the need for emotional support as a basic human need, on par with the need for food and sanitation, for its fulfillment is necessary to maintain a basic level of physical and mental health.35 Relational deprivation and isolation tend to lead to depressive states that make individuals' sense of worth and confidence in the pursuit of their life plans impossible. By contrast, the emotional support provided by one's own family members, friends, and associates can, by itself, notably increase one's own motivation and thus one's chances of succeeding in the achievement of one's own ends, whatever these might be.

But there are also non-affective goods that count as relational resources. The special obligations that are generated by the relationships we have reasons to value is one example. Although scholars disagree on the ground and extent of associative obligations, they tend to agree that relations and associations generate some obligations to treat with special concern those we share special relationships with.36 Now, there is an important sense in which this stock of obligations can be regarded as an all-purpose means. Beyond generating obligations to treat others with special concern, interpersonal relationships also generate obligations to reciprocate or to show gratitude for that special treatment. Reciprocity and gratitude are often discharged as favors that can be “exacted” by the recipient in different forms, at different times.37 Individuals can rely on the gratitude or reciprocity of others whenever this would facilitate the pursuit of their ends, including in occasions of crisis or emergency when they cannot rely on other resources.

Further, social influence also counts as a non-affective relational resource. “Influence” is the ability to affect the behavior and opinions of others. Associations and networks can transform an individual claim into an organized one, thereby conferring social (let alone political) influence to it. Social influence, so understood, is itself a relational good for it is about enhancing, through associative bonds, the strength of one's claims or the socially perceived worth of one's own projects, by bringing others to share and support those claims and projects. Other things being equal, the lonely and isolated person confronting a group who claims a right to the same resources for the same project, or recognition for that project, inevitably suffers from severe competitive disadvantage. Therefore, the more others associate, the more it becomes vital to belong to a group oneself in order to make one's own voice heard by others and to have the ability to pursue one's own ends, when competition arises.

Finally, some of the social bases of self-respect—“perhaps the most important primary good”38 —can be themselves understood as relational resources, insofar as self-respect results from relationships between subjects. By self-respect I mean, following Rawls, to have “a person's sense of his own value [understood as] his secure conviction that his conception of the good, his plan of life, is worth carrying out,” as well as “a confidence in one's ability, so far as it is within one's power, to fulfill one's intentions.”39 As Rawls argues, “unless our endeavors are appreciated by our associates it is impossible for us to maintain the conviction that they are worthwhile.”40 Our associates are the source of our confidence in our abilities to fulfill our intentions. Their encouragement, recognition, and support tend “to reduce the likelihood of failure and … provide support against the sense of self-doubt when mishaps occur.”41 Therefore, to the extent that they can be understood as social bases of self-respect, meaningful but non-intimate associative ties figure themselves as primary goods.

In sum, many relational resources should count as primary goods. This is not simply because they make possible citizens' successful pursuit of a variety of ends, but also because they secure those conditions of support, stability, and the self-confidence that foster the development and maintenance of the first moral power over time.

Further, relational resources also play a relevant role in helping individuals revise their life plans. This seems to be the case when these resources are accessed through participation in a diversified (e.g., cross-class) set of relationships.42 Trust-generating relationships with people different from us provide distinctive arenas for the direct, informal, and safe experience of other points of view. This experience is generally necessary to fully exercise the capacity to confront and revise one's conception of the good. For example, adolescents who grow up in “segregated friendships” and lack meaningful bonds with a diverse set of peers often lack the capacity to engage in dialogue and deliberation with people different from them, as well as to revise the role models their parents impose on them.43 Similarly, as Marilyn Friedman has shown, friendship plays a unique role, even for adults, in developing unconventional values and promoting social criticism.44

Now, it could be objected that relational resources cannot count as all-purpose means for some people do not want or value these goods (e.g., hermits). Yet, if this objection was correct the entire list of social primary goods should be rejected. For example, freedom of movement, certainly a primary good, might not be a valued or useful means if one's ultimate aim in life is to be a cloistered nun. Yet, what matters is not whether existing persons value or want a particular good, but rather whether that good is valuable from the point of view of citizens and their moral powers. I now turn to assess whether relational resources are generally necessary to develop the second moral power.

Relational Resources and the Sense of Justice

Participation in relationships and associations may profoundly influence the development of individuals' capacity for a sense of justice—the ability and willingness to understand and honor the requirements of justice. It is a widely accepted claim both within democratic theory and sociology that the morality and practice of association play a significant role in the development of civic virtues and democratic skills. 45 It is not only within families, but also within secondary associations (if properly structured) that people learn the attitudes of trust and reciprocity. Trust and reciprocity, once transformed into generalized attitudes, become necessary for acting on a sense of justice.

Rawls himself treats “the morality of association” as an essential stage of citizens' development of a sense of justice. He argues that “acquiring a morality of association … rests upon the development of the intellectual skills required to regard things from a variety of points of view and to think of these together as aspects of one system of cooperation.”46 This stage of moral development, generally necessary for the acquisition of a sense of justice, happens not only within families and educational institutions, but also through all sorts of “short-term forms of cooperation.”47

Notably, to claim that the practice of association is an important step in the development of a sense of justice is not to say either that it is sufficient or that all forms of affiliation and association directly lead to the formation of generalized trust and attitudes of reciprocity.48 It is simply to say that it is very unlikely for individuals who lack the experience of trustworthy relationships within their families, friendships, neighborhoods, and voluntary associations to develop a sense of generalized trust towards the citizenry at large, not to mention other cooperative habits.

Further, participation in voluntary associations may facilitate the exercise of a sense of justice by making the discharging of individuals' duties of justice substantively less costly and arguably more effective. For example, associations may facilitate the discharging by individuals of their natural duty to bring about just institutions, by giving public voice and influence to otherwise isolated claims, and by having the power to transform individual claims into organized action. This means that associations can reduce the amount of effort—in terms of both energy and time—that an isolated individual would need to make in order to be heard by the relevant authorities. Associations can also make the discharging of the duty of mutual aid—itself a duty of justice—less costly. Helping others who are in need tends to be more effective and less burdensome for an individual, if she can rely on an organized aid system, for example by acting through or volunteering in charitable organizations. Now, individuals' natural duties to support just institutions and to aid others in need are by nature cost-sensitive, insofar as they become active only in circumstances where complying with them is not too costly for the individual. Since relevant costs plausibly include, among other things, time and effort, we can say that participation in associations has, under normal circumstances, the power to render obligatory actions and efforts that would otherwise remain supererogatory, by reducing relevant costs of compliance. If we assume that acquiring the capacity for a sense of justice presupposes the repetitive exercise of this virtue through, among other things, the appropriate discharging of individuals' natural duties, we must conclude that participation in certain forms of association is generally necessary to develop the capacity for a sense of justice.

To sum up, given their importance for both the development and exercise of individuals' moral powers, relational resources should be included, at least as a matter of principle, into the list of primary goods. This means that when evaluating how well-off people are, it is not sufficient to measure their economic and occupational resources. We must also consider the type of relational resources people have access to. Yet, how do we measure inequalities of relational resources, that is, how do we specify a relational metric?

Measuring Relational Resources

Within a political context, citizens must be assured that they are not getting less than their fair share of social primary goods and that others are not getting more. Inequalities of a good G need therefore to be measured according to a metric that can be publicly understood and verified to a reasonable degree of certainty without excessive intrusion within the lives of individuals.49 Relationship resources would seem to fail both tests.

In order to measure individuals' relational resources and to assess their respective claims, we would need to know how many and what kind of relations an individual enjoys, what kind of and how many relational resources she is able to derive from her friends and fellows, and so on. Beyond being an impossible enterprise, it would require a great amount of intrusion into the lives of individuals.

However, the fact that people's holdings of relational resources cannot be measured in the same way as, say, their taxable income, should not lead us to give up the central idea that relational resources are a part of the metric of justice. The fact that differences in the control of relational resources, like differences in the control of economic resources, significantly affect citizens' development of their central moral powers persists as a morally relevant fact. We need therefore to keep searching for a relational metric that can fulfill the requirements of publicity and measurability of a conception of political justice.

A possible way out of this impasse is the following. Consider the case of a relational good that is difficult, if not impossible, to measure: self-respect. Now, even if we cannot measure how much self-respect a person has, we can tell whether she has access to the social bases of self-respect, and it is these social bases that ought to count as distribuenda. The term “social bases” refers to those “features of the basic structure that may reasonably be expected to affect people's self-respect.”50 So what are the social bases of relational resources? The natural answer to this question would seem to be the “features of the basic structure that may reasonably be expected to affect people's holdings of relational resources.” However, this answer deserves a clarification. The basic structure of a society includes those basic institutions that have the role of controlling the overall production and distribution across society of the benefits and burdens of social cooperation, in the form of social primary goods. If this is the case, what counts as a legitimate element of the basic structure must depend on what counts as a primary good.51 This means that a change in the list of social primary goods may lead to a change in the list of institutions that constitute the basic structure. For example, since income and wealth figure as primary goods, private property and taxation systems figure as parts of the basic structure. Were income and wealth withdrawn from the list, so would these systems be withdrawn from the basic structure. Therefore, when searching for the social bases of relational resources, we should not start by searching for features of the basic structure “as we know it,” by reading Rawls, for Rawls's basic structure is designed with a list of primary goods in mind that excludes relational resources as a distinctive category of goods. We need rather to start by asking what are the elements of the social structure, broadly understood, that determine the production and distribution of relational resources across society. This is what I have referred to in the introduction as “the institutional question.”

The Institutional Question

  1. Top of page
  2. The Metric Question
  3. The Institutional Question
  4. The Principle Question
  5. Conclusion

In approaching the institutional question, we should distinguish between the socially-produced opportunities necessary for people to access relational resources—to form and participate in valuable relations—and the social institutions and norms that provide those opportunities. Both these elements count as relational social bases. Yet, only the former can be distribuenda, whereas the latter are distributors. Among the socially-produced opportunities prominently figure the following elements: the basic liberty to associate, the actual opportunities for socialization, the means for developing some basic level of social skills, as well as “the temporal autonomy” to cultivate relations.52 Therefore, the institutions and social norms that (1) secure the necessary freedom to voluntarily enter into relations with others, (2) determine how many opportunities they have to meet other people in a way that fosters the formation of relational-goods-producing relationships, (3) affect—by either fostering or undermining—people's sociability and relational talents, and (4) determine how much temporal autonomy people have to cultivate their relationships, should be understood as forming a “relational distributive structure.” These are the institutions that control the production and distribution of opportunities for relational resources across society. Since these opportunities count as a social (for socially produced) primary good, we have a justification for including this distributive structure into the basic social structure. Yet what are, in practice, the institutions and norms forming the “relational basic structure” in our society?

Institutions securing freedom of association. These institutions are the easiest to identify. Among these institutions prominently figures the Constitution that protects the freedom to associate.

Institutions providing opportunities for the formation and cultivation of relations. Sociological research on the social determinants of interpersonal relations can help us identify those institutions that play a critical role in providing relational opportunities. Social scientists explain how “making ties depends on how actors interact with others: how long they interact, how frequently, how intensely, and while performing what activity.”53 When looking at the institutions that govern these aspects of interactions, they tend to emphasize how civil society organizations and educational institutions play a unique role in producing interpersonal ties. As the urban sociologist Mario Small puts it, “organizations can affect not only the formation of ties but also the trust they exhibit, the obligations they carry, and the resources they exchange.”54 People's access to relationship resources fundamentally depends upon their access to civil society organizations and educational institutions, as well as upon the way in which these organizations are internally structured.

Why is civil society so special in producing and structuring interpersonal bonds? Social scientists would respond that several conditions determine the production of interpersonal bonds and that these conditions are best met by social (neither political nor market-based) institutions. The formation of close and lasting relations depends on whether agents have the opportunity to interact and to do so informally, frequently, and for a sustained amount of time.55 Membership and volunteering organizations, as well as educational organizations, provide these opportunities more than open public spaces where people come and go, and workplaces where relations tend to be more formal. The formation of affective and long-lasting bonds also depends on whether individuals interact around a common project or interest, and in conditions of cooperation rather than competition.56 Voluntary associations, by definition, are groups of people joined together by some common purpose and are often better placed for the formation of affective ties than market firms and professional associations. Certainly, these more instrumental organizations can provide access to other types of relations, such as extended social networks. Yet, the ability of individuals to initially enter the occupational structure is often conditioned by possession of relational resources and connections that they acquire through friends, as well as voluntary associations and educational institutions.57 Therefore, the overall structure of civil society appears to play a particularly prominent role in the distribution of relational opportunities. In this respect, we can argue that civil society functions as a basic “relational distributive structure” for the production and distribution of social ties and relations and, consequentially, of relationship resources across society.

Institutions and norms that foster the acquisition of relational skills. Sociability, understood as the ability to enter into, cultivate, and mobilize different types of relations cannot be acquired simply through formal education. In order to develop sociability skills, individuals must actively participate in different kinds of relational practices. For example, fulfilling parent-child relationships within the family play a prominent role in developing children's sociability skills. Yet, the nuclear family is not the only school of sociability. Childcare and educational institutions, as well as civil society organizations such as sport clubs and summer camps, can also play a primary role in the development of these skills from an early age.58 For it is within these institutions that formative bonds, through which a child's personality is shaped, mainly develop.59 It is also within these institutions that the negative effects of social segregation on children's self-confidence and sociability first appear. In schools where friendship segregation is the norm, children or adolescents belonging to a minority or lacking popularity can easily loose self-confidence and develop anti-social behavior.60

Institutions and norms distributing temporal autonomy. The labor market features prominently among these institutions. On the one hand, the occupational structure provides some people with the opportunity to form and cultivate particular types of relations (usually non-intimate relations based on weak ties, e.g., professional networks) and thus to access some relational resources (e.g., trust and influence). On the other hand, this same structure also deprives individuals of the temporal autonomy necessary to cultivate other types of relations, especially relations of love, kinship, and friendship, and thus to access (and provide other with) other relational resources (e.g., care and emotional support etc.). The amount of working hours per day, the flexibility of jobs, as well as the amount of holidays per year are fundamental determinants of the temporal autonomy that people have at their disposal to enter and cultivate non-work relations.

The institutions and social norms that control the distribution of relational freedom, relational opportunities and skills, as well as temporal autonomy form what might be called the relational distributive structure. Now, what principles should regulate the distribution of opportunities for relational resources and thus the relational distributive structure?

The Principle Question

  1. Top of page
  2. The Metric Question
  3. The Institutional Question
  4. The Principle Question
  5. Conclusion

In this section I argue that different principles govern the distribution of different relational opportunities. More precisely, freedom of association should be distributed according to the liberty principle, and inequalities of relational resources should be attached to actual opportunities for association open to all under conditions of fair equality of relational opportunity. I will suggest, but leave open, the question of whether the temporal autonomy necessary for the cultivation of relationships should be indexed and distributed according to the difference principle (or an alternative principle of distribution).

Let us start with freedom of association. This complex freedom encompasses the freedom to associate and to dis-associate, as well as the freedom to exclude, within certain limits, others from one's own associations. Because this is a basic liberty, it is quite uncontroversial that it should be equally distributed to all. Within a liberal-egalitarian framework, its distribution should be governed by a principle that prohibits departures from the baseline of equality, even if necessary to increase the socio-economic advantage of the worst off.

However, more controversial is the question of how the freedom to associate should be understood. Generally this freedom is understood as a negative liberty—the state should not prevent people from associating, but has no positive duty to fulfill the effective exercise of the right to associate. I believe instead that the freedom to associate should be conceived as a positive liberty, which imposes upon government a positive duty to provide real opportunities for, and sometimes to actively encourage, association. This is required by the value of this liberty as a social basis of self-respect, and by the importance of diversified forms of association for the capacity to revise one's conception of the good and the development of a sense of justice.

To illustrate this point, it is helpful to draw a parallelism between the associational and the occupational realm. Beyond supporting freedom of occupation as a negative liberty, many would support the idea that diverse opportunities for meaningful work should be secured, according to a principle of fair equality of opportunity, on the ground that meaningful work is an important basis of self-respect.61 Yet, I have argued that the ability to participate in (one or more) meaningful forms of association is also an important basis of self-respect, as well as an important stage in the development of a sense of justice. Assuming the validity of these premises, it follows that opportunities for meaningful relations and associations should also be included into the list of distribuenda, on par with opportunities for meaningful work. Note that only those forms of association that provide the relational goods generally necessary to foster the two moral powers count as “meaningful” from the perspective of political citizens. Now, it could be argued that the Liberty Principle is already meant to guarantee each citizen a social association in which she is able to enjoy support for her plan of life. In order for self-respect to be fully secured, Rawls explains, it “normally suffices that for each person there is some association (one or more) to which he belongs and within which the activities that are rational for him are publicly affirmed by others.”62 However, it seems implausible to think that one association is sufficient for people to foster their self-respect. Self-respect is a good the value of which depends upon relative comparisons and recognition by others. A person who is accepted as a member by one minor association and excluded by the rest of civil society, while other individuals are welcome everywhere, would still count as an outsider and her self-respect would be clearly at risk, not to mention her ability to access other relational resources.

Therefore, one of the reasons that ground freedom of association as a basic liberty, i.e., its importance as a basis of self-respect, is the same reason that warrants moving beyond the language of negative liberty. A state should not limit itself to respect freedom of association as a negative liberty, but should also positively act to provide individuals with (equal or at least adequate) opportunities to participate into meaningful associations.

For example, this may provide government with a neutral justification for using tax-incentives so as to encourage engagement in voluntary associations. Government could do so by providing each citizen with a voucher or tax-credit to be spent in support of qualifying associations.63 The eligibility of associations should be, in turn, made conditional not only on the type of material goods they distribute, but also on the type of relationships they foster—how diverse and inclusive they are and how well suited to produce relationships of trust and support. To illustrate, inclusive coalition membership groups, or organizations that adopt specific policies of socialization, should be given priority for eligibility over more exclusive or sectarian groups, even if they distribute the same services to the public.

Rawls further suggests that meaningful forms of occupational association, such as worker-managed firms, should be encouraged through subsidies on the basis of political values (so compatibly with the requirement of neutrality), if they are likely to encourage the democratic political virtues needed for a constitutional regime to endure—the sense of justice certainly being a virtue of that kind.64 Now, if we agree that actual participation in (one or more) meaningful forms of civil association is also an important step in the development of a sense of justice, then it follows that individuals should not simply be free, but also encouraged, to associate. This provides political institutions with a further political justification for encouraging direct participation into specific forms of associations. Incentives for donating time (volunteering), for example, could be justified, compatibly with the neutrality principle, as long as, and to the extent that they work as effective means of encouraging individuals to (1) directly participate (2) into the life of associations that are conducive to the development of a sense of justice.

Once we grant that a liberal government is under a duty to secure actual opportunities for meaningful association, we then face the question of how these opportunities should be distributed. In order to answer this question, let us consider once again the parallelism between opportunities for meaningful work and opportunities for meaningful association. If the opportunities for relational resources count as primary goods, precisely like opportunities for prerogatives of authority and responsibilities, the principle that regulates the distribution of both types of opportunities should coincide. What is this principle?

Following Robert Hockett and Matthias Risse, I shall distinguish between the “formal availability” and the “consequential aspects” of a good (G).65 The formal availability of G is the opportunity to access G. The consequential aspects of G have to do with the actual value of G for those who access it. Whereas, for example, Rawls requires that the formal availability—the opportunity to access prerogatives of office—be made equally available to all according to a principle of fair equality of opportunity, he requires that the actual value yielded to the holder of offices and positions of authority and responsibility must be indexed according to the difference principle.

By analogy, we can argue that the opportunity to access meaningful relations and associations, and through those relevant relational resources, should be made available to all according to a principle of fair equality of opportunity. This leaves open the possibility that the actual value of the relational resources—their consequential aspects—that individuals are able to access through their relations ought to be indexed according to the difference principle.

Let us thus examine first how a principle of fair equality of opportunities for relational goods would work. The principle of fair equality of opportunity is generally understood as requiring equality of opportunity as between those with the same level of talent and willingness to use them, and it compares them with respect to their opportunity of achieving social and economic advantages attached to public offices and positions of responsibility. According to this principle, it is unfair if some people get more occupational achievements than others for reasons that are arbitrary or irrelevant such as their family background. Only differences in people's talents, abilities, and motivations should determine their occupational outcomes. This principle has the well-known limit of giving insufficient weight to the fact that talents and motivations are themselves shaped by arbitrary factors, such as individuals' socio-economic and relational background. I will thus try and develop a principle of fair equality of relational opportunity that overcomes this limitation.

We may start with the simple claim that it is unfair if some people get more opportunities to access relationship resources for reasons that are arbitrary. Yet, what counts as an arbitrary reason when access to relational resources is at stake? To say that arbitrary obstacles to equality of relational opportunities include all factors that are outside of the control of a person and for which that person bears no responsibility is over-inclusive. For example, the unwillingness of A to associate with B does not necessarily constitute an arbitrary obstacle to B's relational opportunities, despite certainly reducing B's relational opportunities and being outside of B's control. This is because A is, within certain constraints, morally permitted to associate with whomever she wants. A list of arbitrary obstacles to equality of relational opportunities should thus include only those factors for which a person cannot be held responsible and that do not simply result from someone else's legitimate exercise of a (constrained) right to exclude.

A list of arbitrary factors would thus reasonably include the following: (1) lack of relational talents (e.g., sociability) that is either socially produced or has genetic origins, rather than being voluntarily acquired; (2) lack of opportunities for association (lack of temporal, spatial, and material opportunities for encounter and association) that is socially produced, rather than voluntarily sought; and (3) impermissible exclusions from relations and associations. I will analyze these factors in turn.

The lack of relational talents that is socially produced includes the lack of self-confidence and social skills due to growing up within authoritarian, abusive, or segregated relations, whether within the family, neighborhoods, schools, or civil society associations. In this respect, fair equality of relational opportunity requires, as its own precondition, relational equality understood as the absence of social relations of domination, segregation, and hierarchy.

In practice, this first requires policies aimed at ending negligent parenting within the family (due, for example, to parents' lack of time free from work or opportunities for paid parental leave) as well as tackling such phenomena of bullying and friendship segregation within schools and voluntary associations. Yet, these latter policies often fail if not supported by the simultaneous desegregation of neighborhoods, through, for example, the prohibition of zoning and rent control policies. To see that these forms of urban and relational desegregation are required by the principle of fair equality of relational opportunity, is to arrive at a distributive justification of what Anderson calls “the imperative of integration.” Relational integration is a necessary condition of distributive equality, understood as including equality of the opportunities for relational resources.

Yet, lack of relational talents can also be due to disabilities such as autism or mental health problems. As Harry Brighouse and Eva Kittay have argued, a Rawlsian theory of justice should and can, if properly understood and expanded, take the needs of people with disabilities seriously.66 Among the policies that a state could adopt to foster, as far as possible, the relational talents and relational opportunities of people with disabilities or low levels of relational skills, prominently figure the following ones: (a) making psychiatric care, including communication and socialization training schemes, affordable to patients; and (b) providing psychotherapists, social workers, and care-givers with adequate pay and recognition so as to enable them to provide their recipients with adequate care and support.67 The problem with these policies, however, is that they are extremely costly and subject to a leveling-down objection. They require productive individuals to pay for benefits to generally less productive or, in some case, unproductive individuals, without the latter being able to reciprocate. However, as Brighouse has argued, there are non-arbitrary ways to limit the devotion of public resources to unproductive individuals that would pre-empt the leveling-down objection, while still allowing large public spending on the above policies.68 The first limiting principle disallows redistribution of resources from the able to the disabled that would come with the cost of impairing the two moral powers of the former. The second limiting principle would prohibit redistribution that would end up making the worst off, including people with disabilities, worse off than they would be were that money spent to foster overall economic development. The assumption here is that economic advancements would redound for more effective compensation and rectification of several disabilities and mental health issues.

However, the very idea that people should have an equal opportunity to access meaningful relations and associations faces further problems, for people have a right to freely decide with whom they want to associate. How can people have an equal opportunity to, say, enter into a given friendship or marriage? Whereas we can say that if Alpha and Beta are equally talented and motivated for a job, they should be given an equal chance to get that job, it would make little sense to say that since Alpha and Beta are equally talented and motivated would-be-husbands, they should be given an equal opportunity to marry Gamma. Gamma can decide not to marry whomever she wants, for whatever reason she wants. No reason (arguably) counts as arbitrary when such deeply personal choices are at stake.

Therefore, we should interpret the principle of fair equality of relational opportunity as requiring that individuals have an equal opportunity to access an overall adequate level of each relational resource, rather than particular relationships. The standard of adequacy should be defined against whatever is the reason for the sake of which we think opportunities for relational resources ought to be regarded as distribuenda in the first place. For those who adopt a subjective welfarist metric, people should have an equal opportunity to access whatever threshold of relational resources enhance their utility, to the point at which this utility starts decreasing. For those who see relational resources as a constitutive part of human flourishing, adequacy will be set in accordance to some standard of human flourishing. In my account, adequacy must be set according to whatever threshold of each relational resource is necessary to develop and exercise citizens' two moral powers. To the extent that different relational resources are not interchangeable—I cannot substitute trust with emotional support and vice versa—and to the extent that not all relationships provide the same type of relational goods, people should have an equal opportunity to access an adequate number of diverse types of relationships.

Further, and most importantly, the principle of fair equality of relational opportunity does not directly apply to individual choices. It rather governs the framing of the relational distributive structure, and the distribution of opportunities for relational resources. So, whereas we could agree that Gamma violates no principle of fair equality of opportunity when picking Alpha rather than Beta as her husband/wife, even if Beta suffers deprivation from relational resources, we can also agree that a social structure that produces systematic inequalities of opportunities for relational resources, where some people are precluded the opportunities to access meaningful relations and trustworthy social networks, violates this principle.69

Also, even if we assume, quite implausibly, that people are entitled to exclude others from all their associations for whatever reason, we must still acknowledge that the reasons why people exclude others are largely shaped by the social structure. Individuals tend to befriend their similar and people's understanding of who counts as “a similar” is clearly affected by social norms and institutions. Therefore, by designing the relational social structure according to a principle of fair equality of relational opportunities, the point is to simultaneously constrain the number of private actions that run against relational integration. This warrants, within the constraints of basic liberties, active state attempts to change social norms in favor of desegregated and diversified friendships, associations, and marriages.

So far I have been concerned with the “accessibility” of relational resources rather than with “the consequential aspects” of these goods—the actual value these resources have for individuals. It could be argued however that if (1) relational resources are a form of “capital” on par with economic income and wealth, and if (2) the difference principle applies to the distribution of income and wealth then (3) the difference principle ought to apply also to the distribution of relational resources. This would mean that relational resources should not only be attached to opportunities for meaningful relations made accessible according to a principle of fair equality of relational opportunity; they should also be themselves distributed so as to benefit the least advantaged. However, I have argued, it is impossible to exactly measure and compare individuals' amounts of relational resources. Therefore, relational resources, unlike economic resources, seem not well-suited for indexing and, as such, are not an appropriate subject of the difference principle.

However, I would like to conclude this article by offering a sketch, by no means a definitive account, of how some of the opportunities for relational resources may provide a helpful surrogate index. Recall that these opportunities include the temporal autonomy that people have at their disposal to form and cultivate, if they wish, interpersonal relations. Now several scholars have provided detailed arguments about how temporal autonomy could be reasonably included in the list of primary goods.70 Temporal autonomy could be in principle indexed so that it could be traded against other primary goods. The question is whether temporal autonomy can be indexed, as an opportunity for relational resources. This does not seem impossible if we accept that employers could be demanded to pay smaller salaries insofar as, as a part of their employment benefits package, they offer more temporal autonomy and opportunities to participate in networks and associations (e.g., special memberships for particular associations or clubs).71 Inversely, they could be required to pay higher salaries if they offer less temporal autonomy and actual opportunities to form meaningful relationships, either within or outside the work environment.72 For example, jobs that require extensive isolation or fail to provide individuals with an adequate amount of temporal autonomy might, other things being equal, command higher salaries. These are just a few examples aimed at prompting further investigation on whether and how the actual value of relational resources could be indexed.


  1. Top of page
  2. The Metric Question
  3. The Institutional Question
  4. The Principle Question
  5. Conclusion

This paper has sought to make two original contributions. The first contribution has been to look at interpersonal relationships in distributive terms. I argued that, insofar (and only insofar) as they produce certain types of goods, relationships can be reasonably regarded as socially-produced, productive resources (a form of capital) that, like income and wealth, are necessary to pursue a variety of ends and to facilitate the development of a sense of justice. I thus argued that opportunities for relational resources ought to be included within a resourcist metric of socio-economic advantage, together with other social primary goods. By so arguing, I aimed to show that, against what is commonly thought, resourcist theories of distributive justice have a direct, inherently distributive, reason to care about how people relate to each other.

The second contribution of the paper has been to enlarge the understanding of the basic structure, by including within it an identifiable relational social structure. A natural consequence of regarding opportunities for relational resources as a social primary good—a socially distributed type of wealth—has been to include into the basic distributive structure those social institutions (e.g., the institutions of civil society) and social norms (e.g., norms of relational equality) that are often excluded from it. In this respect, this paper provides principles of justice with a new subject of application beyond and between political and economic institutions on the one hand, and the family on the other.

  1. 1

    “Relational capital” is “the valued number of resources an actor can employ and use through direct or indirect personal relations”; Hartmut Esser, “The two meanings of social capital,” Handbook of Social Capital, ed. D. Castiglione, J. W. van Deth, and G. Wolleb (Oxford: Oxford University Press, 2008), pp. 2249, at p. 32.

  2. 2

    I use the terms relational goods and relational resources interchangeably. My definition of relational resources resembles what Harry Brighouse and Adam Swift call “familial relationship goods”; see their “Legitimate parental partiality,” Philosophy and Public Affairs, 37 (2009), 4380. Yet, my account is not limited to goods provided by the family. Further, while Brighouse and Swift conceive of relationship goods as “aspects of well-being” (p. 51) and as part of an arguably comprehensive conception of human flourishing, I conceive of them as instrumental, all-purpose resources.

  3. 3

    Elizabeth Anderson has criticized theories of distributive justice for failing to take seriously the problem of social integration. See Elizabeth S. Anderson, The Imperative of Integration (Princeton, NJ: Princeton University Press, 2010), pp. 1617. Nancy Fraser has argued that the redistributive paradigm is insufficient without a theory of recognition; see her From redistribution to recognition? Dilemmas of justice in a ‘postsocialist’ age,” New Left Review, 212 (1995), 6893, at pp. 78–80.

  4. 4

    See Elizabeth S. Anderson, “What is the point of equality?Ethics, 109 (1999), 287337.

  5. 5

    See especially Eva Feder Kittay, Love's Labor: Essays on Women, Equality and Dependency (New York: Routledge, 1998) and Joan Tronto, Moral Boundaries: A Political Argument for an Ethic of Care (New York: Routledge, 1992).

  6. 6

    For a luck-egalitarian approach to inequalities of love and care see Anca Gheaus, “How much of what matters can we redistribute? Love, justice, and luck,” Hypatia, 24 (2009), 6383. See also John Backer, Kathleen Lynch and Maureen Lyons, Affective Equality (Houndsmills: Palgrave, 2009). These authors argue that an appropriate conception of equality should include the dimension of “affective equality”—equality of love, care, and solidarity. While I share the central thrust of their project, my aim is to include such relational concerns within a political theory of distributive justice, rather than a conception of equality. This effort requires developing a relational metric and a set of distributive principles so as to distinguish those relational inequalities that are unjust from those which are simply bad. Further, unlike theorists of affective equality, I conceive of relational equality as a requirement of, rather than complement to, equality of resources.

  7. 7

    The “metric of justice” refers to what counts as advantageous in terms of distributive justice. A metric is needed to assess the relative weight of people's claims to certain resources. Amartya Sen, “Equality of what?Tanner Lectures on Human Values, ed. S. McMurrin (Cambridge: Cambridge University Press, 1980).

  8. 8

    By “distinctively” I mean that the good cannot generally be produced or accessed outside of relationships of the sort I indicate below, at least not without changing the quality of the good.

  9. 9

    Benedetto Gui, “Beyond transactions: on the interpersonal dimension of economic reality,” Annals of Public and Cooperative Economics, 71 (2000), 139169; Carole Jean Uhlaner, “‘Relational goods’ and participation: incorporating sociability into a theory of rational action,” Public Choice, 62 (1989), 253285.

  10. 10

    We can buy the services of someone, say, a psychotherapist, whose duty is to produce a relational good—a relationship of trust with us. However, we can neither directly buy her trust nor are we under a contractual obligation to trust her.

  11. 11

    Esser, “Two meanings of social capital,” p. 34.

  12. 12

    Annette Baier, “Trust and antitrust,” Ethics, 96 (1986), 231260, at p. 244.

  13. 13

    This is also true of the capacity to trust (beyond trust itself) which develops through participation in certain kinds of interpersonal relationships. Ken J. Rotenberg ed., Interpersonal Trust During Childhood and Adolescence (Cambridge: Cambridge University Press, 2010).

  14. 14

    My analysis of trust is indebted to Baier, “Trust and antitrust” and Karen Jones, “Trust as an affective attitude,” Ethics, 107 (1996), 425.

  15. 15

    Jones, “Trust as an affective attitude,” p. 4 (emphasis mine).

  16. 16

    Ibid., p. 7.

  17. 17

    Samuel Scheffler, “Morality and reasonable partiality,” Partiality and Impartiality, eds. B. Feltham and J. Cottingham (Oxford: Oxford University Press, 2010), pp. 98130, at pp. 1034.

  18. 18

    Baier, “Trust and antitrust,” p. 253.

  19. 19

    Jones, “Trust as an affective attitude,” p. 8.

  20. 20

    Bernard Williams, “Formal structures and social reality,” Trust: Making and Breaking Cooperative Relations, ed. D. Gambetta (Oxford: Basil Blackwell, 1988), pp. 313.

  21. 21

    Baier distinguishes between trust and reliance. I may rely on the babysitter's expertise, or concern for her own professional reputation, to motivate her to take effective care of my child. But I can only trust her when I have good grounds for confiding that she is directly motivated by her good will towards me (a good will that would persist as a motivating factor even in the absence of reputational concerns). But, of course, we can interact with people operating within institutions long enough to form trust-generating relationships. On the distinction between trust and reliance see Baier, “Trust and antitrust,” p. 235.

  22. 22

    Both types of metrics run into notorious problems. See Sen, “Equality of what?

  23. 23

    Martha Nussbaum, Women and Human Development (Cambridge: Cambridge University Press, 2000), p. 245, emphasis mine.

  24. 24

    Ibid., pp. 8283.

  25. 25

    John Rawls, Political Liberalism (New York: Columbia University Press, 1993), henceforth PL, pp. 187188.

  26. 26

    Ibid., p. 188.

  27. 27

    John Rawls, Justice as Fairness: A Restatement, ed. Erin Kelly (Cambridge, MA: Harvard University Press, 2001), p. 58.

  28. 28

    Rawls, PL, p. 19.

  29. 29

    Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971), henceforth TJ, p. 79.

  30. 30

    Baier, “Trust and antitrust,” p. 231.

  31. 31

    Niklas Luhmann, Trust and Power (New York: Wiley, 1980).

  32. 32

    Thanks to Kristi Olson for prompting me to clarify this point.

  33. 33

    Lawrence Backer, “Trust as noncognitive security about motives,” Ethics, 107 (1996), 4361, at p. 45.

  34. 34

    Kittay, Love's Labor. See also Gheaus, “How much of what matters can we redistribute?

  35. 35

    Roy Baumeister and Mark Leary, “The need to belong: desire for interpersonal attachments as a fundamental human motivation,” Psychological Bulletin, 117 (1997), 497529.

  36. 36

    For a relationship-based account of special obligations, see Samuel Scheffler, “Relationships and responsibilities,” Philosophy and Public Affairs, 26 (1997), 189209. For a vulnerability-based account, see Robert E. Goodin, Protecting the Vulnerable (Chicago: University of Chicago Press, 1985).

  37. 37

    James Coleman, “Social capital in the creation of human capital,” American Journal of Sociology, 94 (1988), 95120.

  38. 38

    Rawls, TJ, p. 386.

  39. 39


  40. 40

    Ibid., p. 387. By “associates” Rawls here means members of particular associations, not citizens.

  41. 41

    Rawls, TJ, p. 441.

  42. 42

    See Emily Buss, “The adolescent's stake in the allocation of educational control between parent and state,” University of Chicago Law Review, 67 (2000), 12331289.

  43. 43

    James Moody, “Race, school integration, and friendship segregation in America,” American Journal of Sociology, 107 (2001), 679716.

  44. 44

    Marilyn Friedman, What Are Friends For? Feminist Perspectives on Personal Relationships and Moral Theory (Ithaca, NY: Cornell University Press, 1994).

  45. 45

    Robert Putnam, Bowling Alone (New York: Simon and Schuster, 2000); Theda Skocpol, “Voice and inequality: The transformation of American civic democracy,” Perspectives on Politics, 2 (2004), 320.

  46. 46

    Rawls, TJ, p. 410.

  47. 47


  48. 48

    Nancy Rosenblum, Membership and Morals (Princeton, NJ: Princeton University Press, 1998).

  49. 49

    Robert Hockett and Mathias Risse, “Primary goods revisited: the ‘political problem’ and its Rawlsian solution,” Cornell Law School Legal Studies Research, 2006, paper 55, <>, at p. 6.

  50. 50

    John Rawls, “A Kantian conception of equality” [1975], Collected Papers, ed. S. Freeman (Cambridge, MA: Harvard University Press, 2001), pp. 254266.

  51. 51

    This is not how Rawls proceeds. However, it logically follows from Rawls's understanding of what the role of basic institutions is. Further, Rawls acknowledges that the institutional content of the basic structure can be subject to (e.g., historical) variations, without this compromising the core of his theory.

  52. 52

    I borrow the concept and definition of “temporal autonomy” from Robert Goodin, James Mahmud Rice, Antti Parpo and Lina Eriksson, Discretionary Time (Cambridge: Cambridge University Press, 2008). Temporal autonomy is the time an individual has discretionary control over how to spend it. This includes the time that remains at the disposal of an individual after she has taken care of her work and life necessities.

  53. 53

    Mario Small, Unanticipated Gains (New York: Oxford University Press, 2009), p. 13.

  54. 54


  55. 55


  56. 56


  57. 57

    Mark Granovetter, “The strength of weak ties,” American Journal of Sociology, 78 (1973), 13601380.

  58. 58

    Laura Rosenbury, “Between home and school,” University of Pennsylvania Law Review, 155 (2007), 833898.

  59. 59

    Ibid., fn 41.

  60. 60

    Moody, “Race, school integration, and friendship segregation.”

  61. 61

    Rawls, PL, p. lvii. Jeffrey Moriarty, “Rawls, self-respect, and the opportunity for meaningful work,” Social Theory and Practice, 35 (2009), 441459.

  62. 62

    Rawls, TJ, p. 441.

  63. 63

    For details about how a voucher scheme could be designed, see Ryan Pevnick, “Democratizing the nonprofit sector,” Journal of Political Philosophy, 21 (2013), 260282.

  64. 64

    Rawls, Justice as Fairness, p. 178.

  65. 65

    Hockett and Risse, “Primary goods revisited.”

  66. 66

    Harry Brighouse, “Can justice as fairness accommodate the disabled?Social Theory and Practice, 27 (2001), 537560. Kittay, Love's Labor, ch. 4.

  67. 67

    Kittay, ibid.

  68. 68

    Brighouse, “Can Justice as fairness accommodate the disabled?

  69. 69

    For example, P is morally permitted not to befriend P1 just because P1 is aggressive or rude, even if P1's aggressiveness or rudeness are a matter of brute-luck—due to social circumstances or genetic features. However, institutions have a duty to secure individuals with opportunities not to develop these traits. This is, for example, a reason for schools to adopt bullying-preventing programs and for the state to fund them.

  70. 70

    Thomas Pogge, Realizing Rawls (Ithaca, NY: Cornell University Press, 1989), pp. 198199; Rawls, Justice as Fairness, p. 179.

  71. 71

    For an account of how trade-offs between different primary goods could be worked out through a system of “spontaneous indexing,” see Robert Hockett and Mathias Risse, “Primary goods revisited.”

  72. 72

    Notice that one person may have plenty of free time but no temporal autonomy to cultivate relationships. An unemployed person has plenty of time free from work, but since she must spend her day trying to find a job and may lack other resources to cultivate relationships, she lacks the relevant temporal autonomy. By contrast, a banker may work fifteen hours per day but, as part of her job, she may have relevant opportunities to join associations and form relationships.