1. 1.1 Two Views of Equality
2. I will begin with a brief and stylized contrast between two different views of equality, which I will call the distributive and relational views respectively. These views do not exhaust the possible interpretations of equality, but versions of both have played a prominent role in recent discussions. According to the first view, equality is an essentially distributive value. We can directly assess distributions as being more or less egalitarian, and justice requires that we strive to achieve fully egalitarian distributions, except insofar as other values forbid it. This is the view taken by Jerry Cohen when he says, “I take for granted that there is something that justice requires people to have equal amounts of, not no matter what, but to whatever extent is allowed by values that compete with distributive equality.”1 If one accepts this view, then the most important task is to identify the proper “currency” of “egalitarian justice.” That is, the task is to identify the thing that justice requires us to equalize (insofar as such equalization is allowed by competing values).
3.  
4. According to the second, relational view, equality is an ideal governing certain kinds of interpersonal relationships. It plays a central role in political philosophy because justice requires the establishment of a society of equals, a society whose members relate to one another on a footing of equality. For those who accept this view, one important task is to consider what kinds of institutions and practices a society must put in place if it is to count as a society of equals. The relevant institutions and practices will include those that govern the distribution of goods within the society, and so the ideal of equality, understood as an ideal that governs the relations among the members of society, will have important distributive implications. But, according to this view, equality is a more general, relational ideal, and its bearing on questions of distribution is indirect. The relevant question, in thinking about equality and distribution, is not “What is the currency of which justice requires an equal distribution?” It is, rather, “What kinds of distributions are consistent with the ideal of a society of equals?”
5.  
6. Defenders of the relational view have sometimes criticized the distributive view for offering an inadequate account of the basis for our concern with equality. The distributive view, it is said, represents equality as an excessively abstract or “arithmetic” value. It makes it seem as if the fundamental egalitarian concern is to secure conformity to a certain pattern of distribution for its own sake. It fails to recognize that the real motivation for egalitarianism, both historically and conceptually, lies in a commitment to a certain ideal of society, a conviction that the members of society should relate to one another on a footing of equality. Distributive equality matters, they claim, only because and insofar as it is necessary in order to achieve a society of equals.
7.  
8. Yet the force of the relational view is open to doubt for at least two reasons.2 First, since defenders of this view agree that it, too, supports egalitarian distributions of some kind, it may be obscure what substantive, normative difference it makes whether one accepts the relational view. If the point is simply that egalitarian distributive principles should be grounded in the ideal of a society of equals, rather than presented as self-standing or grounded in some other way, then it looks as if the relational view has no bearing on the choice of the principles themselves. It is simply addressing a different question. So, there need be no conflict between the distributive and relational views.
9.  
10. Second, it may seem that the relational view, if fully spelled out, must itself take a distributive form. For suppose that the members of society are committed to the ideal of a society of equals and are determined to structure their mutual relations in accordance with that ideal. How would they go about doing this? The answer, it may seem, is that they would take care to ensure that certain important goods, such as status, power, or opportunity, were distributed equally within the society. That is what it would mean for them to achieve a society of equals. But if that is correct, then the relational view is not really an alternative to the distributive view but is rather a version of it. It is distinguished from other versions not by placing less emphasis on distribution but by singling out goods like status and power as the ones whose distribution should be the object of egalitarian concern.
11.  
12. I believe that these two doubts can be allayed. Contrary to what the second doubt suggests, the relational view is not a version of but is rather a genuine alternative to the distributive view. And contrary to what the first doubt suggests, the relational view does have a bearing on substantive, normative questions. If we accept the relational view, this will affect the way we think about the content of distributive justice. In order to establish these claims, more must be done to develop the relational view. That is what I will attempt to do here. Before I begin, two preliminary issues need to be addressed.
13.  
14. First, consider the distinction between “prioritarianism,” which holds that benefits to those who are worse off matter more than benefits to those who are better off, and forms of egalitarianism that hold that it is bad if some people are worse off than others through no fault of their own. It is sometimes said that prioritarianism is a nonrelational view, because it is sensitive only to the absolute levels of well-being of the affected individuals, whereas egalitarianism is a relational view, because it is sensitive to essentially comparative judgments about the relations among different individuals’ levels of well-being.3 Here the term “relational” is being used to mark a difference between two different distributive views. Both prioritarianism and egalitarianism of the sort described are distributive views, and the term “relational” is being used to distinguish distributive views that are sensitive to comparative information from those that are not. By contrast, I use the term to describe a view of equality that is not distributive at all. What I call “the relational view” is not the view that distributive principles should be sensitive to comparative information. It is a view according to which equality should not be thought of as a fundamentally distributive value in the first place.
15.  
16. Second, although my sympathies lie for the most part with the relational view, this does not mean that I regard questions of distribution as unimportant or that I think economic inequality is unobjectionable. I believe that the levels of economic inequality that prevail in my country and many others are indefensible, and I am as convinced as anyone of the importance of distributive justice. The relational view does not deny that equality has a bearing on questions of distribution. Instead it holds that, in order to appreciate the bearing of equality on distribution, one must begin by understanding equality as a broader ideal that governs the relations among members of society more generally. Rather than assuming that justice requires the equal distribution of something and then asking what that something is, a relational approach asks what the broader ideal of equality implies about distributive questions. Defenders of the relational view believe that the case against distributive inequality is strengthened rather than weakened if it is linked to a broader ideal of this kind, because the ideal is more attractive than any purportedly egalitarian distributive formula considered on its own. If egalitarian social and political positions have roots in the idea of a society of equals, this gives them a critical force that they would otherwise lack. Or so defenders of the relational view believe. Whether they are right depends on whether the relational view can be successfully fleshed out. That is what I will try to do in this essay.
17.  
18. 1.2 Egalitarian Personal Relationships
19. I will first consider a nonpolitical example. I do so advisedly. Equality is a central political value, and its implications in political contexts will differ in many respects from its implications in other contexts. Yet equality is not an emergent value that appears for the first time at the political level, and we should be able to see some connection between the way it functions in political contexts and the way it functions elsewhere. It would be an objection to an account of political equality if it allowed us to see no such connection. One of the advantages of the relational conception is that it represents equality as a value that applies to human relationships of many kinds, and we may learn things by looking at its nonpolitical applications that will help us to understand how it applies to the political case. So consider the assertion that a marriage or partnership should be a relationship between equals. What might this mean? Suppose we have two spouses or partners, each of whom is committed to conducting their shared relationship on an egalitarian basis. How might this affect the way they relate to each other? Presumably it will affect the attitudes they have toward one another and the ways in which they are disposed to treat one another, but what exactly will these effects be? A natural first thought is that the participants in an egalitarian relationship will have a reciprocal commitment to treating one another with respect. Each sees the other as a full-fledged agent who has the capacities associated with this agential status. Each expects the other to bear whatever responsibilities are assigned to a person in virtue of this status and, similarly, each sees the other as entitled to make whatever claims accrue to a person in virtue of this status. Moreover, neither participant is seen by either of them as possessing more authority than the other within the context of the relationship, and each sees the other as entitled to participate fully and equally in determining the future course and character of the relationship.
20.  
21. If these initial speculations are on even roughly the right track, then one thing that is already clear is that the ideal of an egalitarian relationship draws on values other than equality itself. It draws on values such as reciprocity and mutual respect, and on a conception of the rights and responsibilities of agents. This might lead one to wonder whether the term “egalitarian relationship” is a misnomer.4 Fully spelled out, perhaps the idea of such a relationship appeals entirely to values other than equality. This suggestion seems to me overstated. What is true is that the ideal of an egalitarian relationship is a complex one, and that several of its elements draw on values other than equality per se. This is an important point, and I will return to it later. At the same time, the ideal also includes some distinctively egalitarian elements, and in what follows I want to discuss one that seems to me especially important.
22.  
23. In a relationship that is conducted on a footing of equality, each person accepts that the other person’s equally important interests—understood broadly to include the person’s needs, values, and preferences—should play an equally significant role in influencing decisions made within the context of the relationship. Moreover, each person has a normally effective disposition to treat the other’s interests accordingly. If you and I have an egalitarian relationship, then I have a standing disposition to treat your strong interests as playing just as significant a role as mine in constraining our decisions and influencing what we will do. And you have a reciprocal disposition with regard to my interests. In addition, both of us normally act on these dispositions. This means that each of our equally important interests constrains our joint decisions to the same extent. We can call this the egalitarian deliberative constraint. It is a distinctively egalitarian element in the complex ideal of an egalitarian relationship.
24.  
25. Arriving at decisions that satisfy this constraint is not, in general, an easy matter. Simply identifying the relevant interests that bear on a decision can be difficult. And, of course, the interests of the participants may clash, and then there will be a question about how to forge a joint decision in the face of conflict. Different solutions may suggest themselves on different occasions. One strategy that will sometimes be available is a strategy of splitting the difference. You want badly to go to Paris for three weeks. I want just as badly to go for one week. We split the difference and decide to go for two weeks. Another strategy that will sometimes be available is a strategy of choosing the second-best. My first choice is to go to Paris and my second choice is to go to Rome. Under no circumstances do I want to go to London. Your first choice is to go to London and your second choice is to go to Rome. Under no circumstances do you want to go to Paris. So we decide to go to Rome. A third strategy is taking turns. I want to go to Paris; you want to go to Rome. We decide to go to Paris this year and to Rome next year. A fourth strategy is joint satisfaction. I want to go to Paris; you want to go to Rome. So we decide to spend half our time in Paris and half our time in Rome. A fifth strategy is one of trading off. Suppose we face two decisions that we regard as being of roughly comparable importance: where to go on our holiday, and whether to subscribe to the ballet or to the opera. I want to go to Paris; you want to go to Rome. You want to subscribe to the opera; I want to subscribe to the ballet. So we decide to go to Rome and subscribe to the ballet. A sixth strategy is separation. I want to go to Paris; you want to go to Rome. So we decide that we will not take a joint holiday. Instead, I will go to Paris, while you will go to Rome.5
26.  
27. Even in simple cases such as these, arriving at a decision may be a significant deliberative task, because multiple solutions to the deliberative problem may be available, and you and I may have different meta-preferences among the strategies embodied in those solutions. In more complex cases that implicate more important interests, satisfactory solutions of any kind may be difficult to find, and so the deliberative task may be more challenging. Moreover, different decisions may interact with one another in a variety of ways that add further deliberative complexity.
28.  
29. Six additional complications should be noted. First, it should be clear even from the simplified example just given that the egalitarian deliberative constraint is best understood diachronically rather than synchronically. The point is not that each decision taken individually must give equal weight to the comparably important interests of each party. Sometimes this will be impossible or undesirable. The point is rather that each person’s interests should play an equally significant role in determining the decisions they make over the course of the relationship.
30.  
31. Second, I have said that the relevant notion of interests is a broad one that includes needs, values, and preferences. We need not suppose that there are sharp dividing lines among these categories. But the inclusion of needs and values along with preferences and other interests reminds us that sometimes arriving at a joint decision in the face of conflict may be difficult or even impossible. If you and I have diametrically opposed values, then in decisions that implicate the opposing values none of the strategies just mentioned may be available. If I am a pacifist and you are a warrior, there may be no possibility of splitting the difference between us. Diachronic solutions like taking turns may also be unacceptable if our values are sufficiently opposed. Deciding to honor my values today and yours tomorrow will not work if honoring your values amounts to a violation of my values whenever it is done. If I am an animal rights activist and you are a hunter, then deciding that we will demonstrate against animal experimentation today and go hunting tomorrow will not work. Even separation may not always seem tenable. A joint decision that I will go to the animal rights demonstration while you will go hunting may still seem to me an intolerable compromise of my values. This gives people who want their relationships to be conducted on a footing of equality a (defeasible) reason to seek out others who share their most important values, at least for their most comprehensive personal relationships.
32.  
33. There may sometimes be another alternative. I have said that the egalitarian deliberative constraint applies to decisions made within the context of an egalitarian relationship. However, it is not obvious when a decision counts as being made “within the context of the relationship.” If I decide to demonstrate against animal experimentation while you decide to go hunting, is it the case that our respective decisions are made “within the context of our relationship?” That may depend on the character of the relationship. In principle, one way of preserving an egalitarian relationship in the face of conflicting values may be to externalize the conflict by relegating the parties’ pursuit of their discordant values to a space that is defined as being outside the relationship. This is similar to the strategy of separation that may be used to satisfy the deliberative constraint, but it differs in that here each of the parties can disclaim even the limited endorsement of the other’s values that is involved in a joint decision to separate. It is an interesting question under what conditions externalization of this kind can be successful. At times it may seem artificial or self-deceptive. And when conflicts of fundamental values are at stake, it may be unsustainable.
34.  
35. Third, however we assess the prospects for externalizing conflict in the manner just described, it is important to emphasize that decisions made within the context of an egalitarian relationship need not always be arrived at jointly. Sometimes one of the parties to a relationship will be charged with the sole responsibility for making such a decision. This can happen, for example, if the other party is unavailable for joint deliberation, or if the parties have themselves decided on a division of deliberative labor, in which, say, decisions of some kinds are made by one of them while decisions of other kinds are made by the other of them. But these exceptions arise against the background of a presumption that each party is equally entitled to participate in decisions made within the context of the relationship. This participatory requirement follows from the more general point, noted earlier, that the parties to an egalitarian relationship view each other as equally entitled to determine the future course and character of the relationship. The participatory requirement can be modified in cases like those mentioned but only in ways that are acceptable to the parties themselves.
36.  
37. Fourth, my example of choosing a holiday destination may create the misleading impression that the egalitarian deliberative constraint requires the parties always to make decisions that will satisfy their interests (weighted for importance) to an equal degree, as all the possible decisions mentioned in the example do. But in this respect the example is unrepresentative. What the deliberative constraint requires is that the comparably important interests of each party should play a comparably significant role in influencing decisions made within the context of their relationship. This does not mean that, in general, their decisions must leave the parties equally well-off either in respect of those interests or overall. To suppose otherwise is to overlook the heterogeneity of people’s interests and the variety of ways in which their interests may constrain the deliberative process.
38.  
39. How should the deliberations of the parties be influenced by the interests of each in order to comply with the egalitarian deliberative constraint? The first requirement concerns the way in which the parties’ interests shape their deliberative priorities. The comparably important interests of each of them should be assigned comparable priority when setting their joint deliberative agenda, that is, when selecting the issues that will receive their joint deliberative attention. In addition, the parties should display comparable tenacity and imagination in seeking to address the comparably important interests of each of them.6 In these ways, they make manifest their view of one another as equals and the equal seriousness with which they treat one another’s interests.
40.  
41. Beyond this, there is the question of how the content of the parties’ decisions should be influenced by their respective interests. There is no single answer to this question. When the “interests” in question are values, then “satisfying” those interests will mean different things in different contexts. This is true even for a single individual who is not subject to the egalitarian constraint. Sometimes “satisfying” a value may simply mean not acting in ways that are inconsistent with it. In other contexts, it may mean acting in specific ways that are demanded or required by the value. And in still other contexts, it may mean acting in ways that are expressive or constitutive of the value. It follows that in joint deliberations where the parties’ values are at stake, what the egalitarian deliberative constraint requires cannot without distortion be described as achieving an equal level of interest-satisfaction. Instead, what the constraint requires is that the parties’ decisions should be equally sensitive to the diverse implications of their actions for the values of each of them.
42.  
43. Similarly, an interest that takes the form of a need or preference rather than a value may serve only to rule certain options out rather than to fully determine the content of a decision. With interests of this kind, the equal satisfaction model is again out of place. And the model fails even in cases where the parties are explicitly attempting to fulfill a need or preference of one of them. Consider a case in which they previously took action to address some need or preference of the first party. Now the other party has a comparably urgent need or preference. Here what the egalitarian constraint requires is that they should make just as great an effort to satisfy the second party’s need or preference as they did with that of the first party. It does not say their aim should be to ensure that the second party’s preference is satisfied to just the same extent that the first party’s was. For example, it does not require that if their conscientious decisions about medical treatment led to a 70% reduction in the first party’s chronic pain, then their aim should now be to produce a 70% reduction in the second party’s pain. If, by making comparably conscientious efforts to secure good treatment, they could achieve the complete elimination of the second party’s pain, the egalitarian constraint hardly forbids this.
44.  
45. The upshot is that the egalitarian deliberative constraint does not, in general, require the parties to make decisions that will leave them equally well-off either in respect of their immediately affected interests or overall. The fact that all the decisions mentioned in the holiday-destination example do leave the parties equally well-off results from special features of that example. In particular, it is an example in which (a) the parties are seeking to make a single, circumscribed decision about a joint activity, (b) the only interests bearing on the decision are the parties’ symmetrical but conflicting preferences about one aspect of that activity, (c) there is an obvious metric for determining the extent to which the preferences of each party have been satisfied, and (d) there are multiple options available that will leave the parties equally well-off in respect of their conflicting preferences. Given these simplified decision parameters, it is natural to suppose that the parties will choose one of the equalizing options. Even in this case, however, they might choose otherwise without violating the egalitarian deliberative constraint. They might, for example, flip a coin to decide on their holiday destination. The egalitarian deliberative constraint does not require them to make a decision that will leave them equally well-off. Rather, it requires them to attend with equal urgency and determination to the comparable interests of each of them. Given that they are deciding on a joint activity about which they have comparable but conflicting preferences, that there is a well-defined metric available for assessing how well candidate decisions satisfy those preferences, that there are options available that will satisfy their preferences to an equal degree, and that there are no other interests that need to be taken into account, it is natural, though not strictly necessary, that they should make a decision that will produce equal preference-satisfaction. But in many cases one or more of these conditions will fail to obtain. In such cases, there is no general reason to expect that the egalitarian constraint will require decisions that leave the parties equally well-off with respect to preference-satisfaction or anything else.
46.  
47. The fifth complication is this. The egalitarian deliberative constraint tells the parties something about how they should treat the comparably important interests of each of them. But how are these judgments of importance to be understood? Is the point that the parties should be guided by their own beliefs about the importance of their interests, or is there some independent standard of importance that applies? In practice, the parties have no choice but to rely on their own judgments of importance (even if they consult others in forming those judgments). Moreover, the very fact that one believes an interest to be important can sometimes make it important. But what the deliberative constraint says is that the parties should treat (what are in fact) the equally important interests of each of them as having equal significance for their decisions. This standard of importance is independent of and can diverge from their own judgments of importance, even if they have no choice but to rely on those judgments. This means they can be mistaken in thinking they have complied with the constraint.
48.  
49. Finally, the deliberative constraint is central to egalitarian relationships, but if it is kept too clearly in view or interpreted too rigidly it can encourage a kind of scorekeeping that may erode the quality of the relationship. If the participants in a relationship are constantly preoccupied with making sure that the comparably important interests of each of them are playing comparably significant roles in determining their joint decisions, that may exclude forms of intimacy and joint identification that give personal relationships much of their value. So the trick is to ensure that the egalitarian deliberative constraint is satisfied without itself becoming the focus of excessive attention.
50.  
51. It should be clear from the six complications I have discussed that conducting and sustaining a personal relationship on a footing of equality is a significant practical task. Indeed, relating to others as equals is best thought of as a complex interpersonal practice. It is a practice that makes substantial demands on the attitudes, motives, dispositions, and deliberative capacities of the participants. There is no general formula or algorithm for determining how best to engage in the practice. Instead, sustaining an egalitarian relationship requires creativity, the exercise of judgment, and ongoing mutual commitment, and even the sincere efforts of the parties are no guarantee of success, although success is a matter of degree and should not be conceived of in all-or-nothing terms.
52.  
53. 1.3 The Role of Distribution in Egalitarian Personal Relationships
54. What role do issues of distribution play in our understanding of egalitarian relationships? Such issues might be thought of as arising in two different ways. First, they might be thought of as internal questions that arise within the context of an egalitarian personal relationship. As such, they present practical challenges that must be addressed by the participants in the relationship. Given a relationship that qualifies as egalitarian by some independent standard, how are questions about the distribution of various goods to be handled within the context of that relationship? Alternatively, questions of distribution might be thought of externally, as questions about the best way to characterize egalitarian relationships in the first place. It might be supposed that in order to understand what an egalitarian relationship is, we need to ask what it is that is distributed equally between the participants in such a relationship. This alternative is in keeping with the second of the two doubts about that the relational conception that I want to investigate. It assumes that the relational conception, fully spelled out, must itself take a distributive form.
55.  
56. Let me begin by saying something about the external question. I have characterized egalitarian relationships in practical and deliberative rather than distributive terms. Equality, as I have described it, is ultimately a form of practice rather than a normative pattern of distribution. An egalitarian relationship is one in which the parties have certain attitudes, motives, and dispositions with respect to one another. Among other things, they satisfy a fundamental deliberative constraint when making decisions that fall within the scope of their relationship. And the point is not that these attitudes, motives, and dispositions must be distributed equally between the parties. Admittedly, the relationship will not have an egalitarian character if one of the parties exhibits the relevant attitudes and dispositions and the other does not. The attitudes and dispositions must hold reciprocally. But neither will the relationship have an egalitarian character if the parties possess those attitudes and dispositions to an equal but low degree. If anything, the egalitarian aim is not to equalize the relevant attitudes and dispositions but to maximize them: to ensure that both parties exhibit them to the fullest.
57.  
58. It might be suggested that the description I have given is equivalent to saying that egalitarian relationships are characterized by an equal distribution of status between the participants. However, this formulation does little to suggest the deliberative and practical dimensions of equality as I have described it and, as I have just argued, an emphasis on equal distribution seems misplaced where those dimensions are concerned. Of course, one might stipulatively define an equal distribution of status as one that obtains when a relationship satisfies the deliberative and attitudinal criteria I have outlined, but such a definition would be artificial. The bare possibility of constructing a stipulative definition does not show that there is any natural or interesting sense in which egalitarian relationships are best understood in distributive terms.
59.  
60. This leads me to think that questions of the second, external kind are misplaced. The way to understand what a relationship of equals is like is not to ask what is distributed equally in such a relationship. The distinctive feature of egalitarian relationships is not that there is an equal distribution of something. It follows that, at least as applied to personal relationships, the second doubt about the relational conception of equality is unfounded. It is not the case that the relational conception, fully spelled out, must itself take a distributive form.
61.  
62. Internal questions, however, are not misplaced. Insofar as decisions about how to use available resources arise within the context of an egalitarian relationship, the egalitarian deliberative constraint will apply to those decisions. Again, there are questions about when a resource allocation decision falls within the scope of such a relationship. In some cases, it will be clear that the resources in question belong jointly to the participants in the relationship and that decisions about how to allocate them fall within its scope. In many ordinary friendships, by contrast, it may seem clear that the participants have few if any material resources in common, and that the decisions that each of them makes about how to allocate his or her resources fall outside the scope of the relationship. I do not in general expect to have a say in how my wealthy friend decides to spend his money. But whether a resource allocation decision falls within the scope of a relationship cannot always be settled solely by reference to the prevailing legal regime of property and ownership. For one thing, we may sometimes feel that an individual participant has taken advantage of the prevailing regime to “shelter” resources, or exclude them from joint decision, in a way that is incongruous with the egalitarian character of the relationship. In addition, there may be cases in which we are in no doubt that certain resources are the legal property of one of the participants, but we nevertheless believe that his decisions about how to allocate his resources are incompatible with a relationship of equals. If my wealthy friend regularly insists on going to more expensive restaurants than I can afford and on paying the bill for both of us, then I may feel that his allocative decisions, although legally unimpeachable, are undermining the egalitarian character of our relationship. This suggests that certain allocative decisions may fall within the scope of a relationship, in the sense that matters for our purposes, even if the resources whose allocation is under consideration belong exclusively to one of the participants.
63.  
64. Let us leave these complications aside, however, and focus on cases in which the participants in a relationship of equals are considering how to allocate resources that they jointly control. These decisions are subject to the egalitarian deliberative constraint. Each participant accepts that the other’s comparably important interests should play a comparably significant role in influencing the allocation decisions that they make. This is a substantial constraint, even if we assume that it applies diachronically rather than synchronically. In the context of a face-to-face personal relationship, however, it seems unlikely that the participants will attempt to satisfy the constraint through the self-conscious application of a fixed distributive formula, such as a leximin principle or a principle of equality of welfare or resources. It would seem more than a bit peculiar if they did do this. There are several reasons why this is so. First, to rely on such a formula would seem rigid and moralistic, and would raise concerns of the kind noted earlier about excessive scorekeeping. Second, many of the allocation decisions the participants are likely to face will be decisions about how best to advance or protect their shared interests, whereas distributive formulae of the kind mentioned are used to adjudicate among conflicting interests.
65.  
66. Finally, the participants are, by hypothesis, concerned to sustain their relationship as a relationship of equals, and they are therefore concerned with the ways in which their respective interests influence their joint decisions. This means, to put it crudely, that they are concerned with the ways in which their respective interests are treated as inputs of deliberation and decision. But distributive formulae of the kind mentioned operate, in effect, on the outputs of decision. Such a formula does not directly assess the role played in deliberation by considerations about the respective interests of the parties. It looks instead at the situation of the participants once a given decision is carried out and assesses their comparative standing in respect of some dimension, such as welfare or resources, which is thought to reflect their interests. Such assessments may provide indirect evidence of the way in which considerations about the participants’ interests influenced the decision-making process. But the participants, with their normally extensive mutual knowledge and their direct access to their own deliberations, are unlikely to regard these output measures as being, in general, good proxies for the kinds of assessment of their deliberations that matter to them. Why should they look at their overall situation once a decision is carried out and make an inference on that basis about how they must have deliberated? As a way of assessing their deliberations, this would be not only indirect but also of limited reliability, since for two people to deliberate in accordance with the egalitarian constraint it is neither necessary nor sufficient that their overall situation once the decision is carried out should end up satisfying any fixed distributive formula. There is, in general, no need for the participants to rely on such indirect and unreliable inferences. They can ask themselves directly whether the comparably important interests of each of them played an equally significant role in influencing their decisions. Of course, they can be mistaken about this, and output measures may serve as correctives to self-deception and other forms of error. Distributive inequalities may be symptoms that the participants’ internal deliberations violated the egalitarian constraint even though they thought otherwise. But there is a difference between using output measures to guard against self-deception and using them systematically to satisfy the egalitarian deliberative constraint.
67.  
68. So, to repeat, the participants in a relationship of equals are unlikely, when facing decisions about the allocation of their resources, to try to satisfy the egalitarian deliberative constraint by applying a fixed distributive formula. On the other hand, the deliberative constraint will itself exert pressure in the direction of egalitarian distribution. If, in deciding how to allocate their resources, the participants treat the comparably important interests of each of them as having comparable significance, then a natural default assumption is that they will end up devoting roughly equal resources to satisfying the comparably important interests of each. And insofar as it makes sense to compare the extent to which their interests are satisfied, a natural default assumption is that their decisions will tend to produce roughly equal levels of (weighted) interest satisfaction. The fact that the egalitarian deliberative constraint exerts general pressure toward egalitarian distributions explains why distributive inequalities can serve the corrective function just noted. But the conclusion that the participants’ decisions will have distributively egalitarian upshots is a defeasible one, and the reason it holds is not because they apply any particular distributive formula in making their choices. It holds because they regard the reasons generated by the comparable interests of each of them as themselves being of comparable strength. That is the regulative principle governing their deliberations.
69.  
70. It may seem that the participants in egalitarian relationships would have a greater concern than I have acknowledged with distributive equality per se. They would regard it as intrinsically important that equal resources be allocated to meeting their respective interests, or that those interests be satisfied to an equal degree. But I do not believe that a concern for the egalitarian character of their relationship would lead them to be troubled by the bare fact of inequality in one of these dimensions. Their primary concern, insofar as they wish to conduct their relationship on an egalitarian basis, is with their attitudes toward one another and with how seriously each takes the interests of the other in contexts of deliberation and decision. If they were in other respects satisfied with the egalitarian character of the relationship, then I doubt that the bare fact of distributive inequality would, by itself, arouse their distinctively relational concern.7 By contrast, if one of the participants regularly flouted the egalitarian deliberative constraint but strict distributive equality were somehow achieved anyway (perhaps by luck or perhaps because it was imposed by an outsider), then the egalitarian character of the relationship would be compromised despite the fact that distributive equality had been achieved.
71.  
72. Let me pause to take stock. I began by identifying two doubts about the relational conception of equality. The first turned on the thought that it makes no normative difference whether or not one accepts the relational view, since in either case egalitarian distributive principles will be needed and the relational view has no bearing on the choice of such principles. The second turned on the thought that the relational view, fully spelled out, must itself take a distributive form. As applied to egalitarian personal relationships, we have seen that the second doubt is unfounded. It is not true that the relational view must take a distributive form. And the considerations we have just been rehearsing suggest that, as applied to personal relationships, the first doubt is also unfounded. On the relational view, there is strong general pressure within egalitarian personal relationships toward rough distributive equality of some kind, but there is no reason to think that such relationships are regulated by any fixed distributive formula. So, a fortiori, there is no reason to think they are regulated by the same formula that a purely distributive conception of equality might recommend.
73.  
74. 1.4 A Society of Equals
75. Let me now turn back to the case of a society of equals. What light, if any, can our discussion of personal relationships shed on the contrast between the relational and distributive interpretations of social and political equality? Are the two doubts about the relational conception any better founded in this case than they are in the case of personal relationships? To begin, I believe that a version of the deliberative constraint that plays a central role in egalitarian personal relationships is also central to the idea of a society of equals. In such a society, each member accepts that every other member’s equally important interests should play an equally significant role in influencing decisions made on behalf of the society as a whole. Moreover, each member has a normally effective disposition to treat the interests of others accordingly. So a society of equals is characterized by a reciprocal commitment on the part of each member to treat the equally important interests of every other member as exerting equal influence on social decisions. This gives determinate content to the otherwise vague thought that the members of such a society regard one another as equals. It means that the equally important interests of each of them constrain social decisions to the same extent. This is, I take it, a familiar ideal. And one has only to consider its application to cases of racial or ethnic or gender hierarchy to see that it has considerable critical force. To cite one topical example, it is this ideal to which advocates of gay marriage appeal when they argue that the interests of homosexuals in being able to marry are just as strong as the interests of heterosexuals and, accordingly, that both sets of interests should be accommodated in the same way in our laws and institutions.
76.  
77. Some of the problems that arise in thinking about egalitarian personal relationships have straightforward parallels in thinking about a society of equals. For example, questions about conflicting values and how to accommodate them in egalitarian decision-making present challenges in both cases. And just as there is a question about when a decision counts as being made “within the context of” a personal relationship, so too there is a question about which decisions count as “social decisions” or decisions “made on behalf of the society as a whole.” Without attempting a complete answer to this question, it seems safe to assume that decisions about a society’s constitution, its laws, and the design of its major social, political, and economic institutions all count as matters of social or collective decision in the relevant sense.
78.  
79. However, as this last observation already suggests, there are also obvious and important differences between personal relationships and the relations among the members of a political society. These differences affect the way the relational conception applies to the two cases. Let me mention some of the most significant differences.
80.  
81. First, in contrast with personal relationships, few of the relationships among the members of society are face-to-face relationships. No member of a modern society is acquainted with more than a tiny fraction of the other members. For the most part, the relations among the members of society have an anonymous character. Second, one consequence of the anonymous character of these relations is that the members of society do not in general have individualized knowledge of the needs, preferences and values of their fellow members. So in thinking about how to satisfy the deliberative constraint, they have to rely heavily on normalized assumptions about the characteristic needs and interests that members can be assumed to have.
82.  
83. Third, the anonymity of the relations among the members of society also affects the character of their collective decision-making. Although their decisions are subject to the egalitarian constraint, they are never arrived at through face-to-face deliberations in which all members participate. This again contrasts sharply with the case of personal relationships, in which face-to-face joint deliberations and decisions are common. Yet the ideal of a society of equals remains subject to the presumption that each participant in an egalitarian relationship is equally entitled to participate in decisions made within the context of that relationship. As noted earlier, this participatory requirement can be modified even in the case of personal relationships, but only in ways that are acceptable to the participants themselves. In developing the ideal of a society of equals, a crucial task will be to determine how the participatory requirement should be modified to apply to the large-scale deliberative processes that are needed in a society whose members are largely anonymous to one another.
84.  
85. Finally, the anonymity of the relationships among the members of society sets up pressure to establish clear boundaries to those relationships and clear limits to the scope of the decisions that are thought to fall within them. The members of society will be interested in preserving social space within which they can conduct their face-to-face personal relationships and pursue their conceptions of the good life without being subject to comprehensive regulatory scrutiny from the perspective of an anonymous collectivity that lacks individualized knowledge of its members’ needs, preferences, circumstances, and values. This interest is reflected in the ubiquity of such distinctions as those between the public and the private or between the political and the nonpolitical. It means that strategies of externalizing decisions—treating them as falling outside the context of a given relationship—will have a special salience in connection with the generic relations among the members of society.
86.  
87. Although the case of a society of equals differs in these respects from the case of egalitarian personal relationships, the core content of the egalitarian deliberative constraint continues to apply. In a society of equals, the comparably important interests of each member are to constrain social decisions to the same extent. This aspiration is reflected in the reciprocal attitudes and the normally effective dispositions of each member. It is a complex aspiration and not one that is easily satisfied, but it is an aspiration that is characteristic of an egalitarian society and undertaking to satisfy it is a challenge that such a society accepts.
88.  
89. 1.5 The Role of Distribution in a Society of Equals
90. As in the case of egalitarian personal relationships, the ideal of a society of equals is not well described in distributive terms. It is not the view that there is something that should be distributed equally among the members of society. Instead, it is a practical ideal concerning the kind of society the members want to construct and the way they want to relate to one another. This ideal is reflected in their attitudes and dispositions and, in particular, in their convictions about the ways in which the interests of each of them should constrain social decision. These attitudes are not themselves to be equalized but rather to be securely entrenched in the motivational outlook of each member. As in the case of personal relationships, then, “external” distributive questions are misplaced, and attempts to characterize a society of equals in purely distributive terms are bound to be procrustean. The defining feature of this type of society is not that there is an equal distribution of something among the members. So in this case as in the case of personal relationships, the second doubt about the relational conception of equality is unfounded. It is not true that the relational conception, as applied to society as a whole, must itself take a distributive form.
91.  
92. Once again, however, “internal” distributive questions are not at all misplaced when thinking about a society of equals. Questions about the distribution of social resources are of obvious importance for the members of such a society. What binds the members to one another is their shared participation in a common social framework, and they are especially concerned with the way that framework structures the distribution of the resources that are necessary for them to flourish. Here the egalitarian deliberative constraint is once again relevant. In deliberating about the institutions and practices that constitute the social framework, the members accept that the comparably important interests of each of them should exert comparable influence on their decisions. As in the case of personal relationships, this constraint exerts strong pressure in favor of an egalitarian standard of distribution. For example, it is difficult to see how a pure laissez-faire market system of the kind Rawls referred to as “the system of natural liberty” could be reconciled with the egalitarian constraint.8 Such a system allows the distribution of resources to be determined to a very high degree by natural and social contingencies, such as people’s natural attributes and the social circumstances into which they were born, which themselves have no moral basis. This feature of the system is difficult to reconcile with the egalitarian constraint, for it will inevitably compromise the ability of some people to satisfy their basic interest in pursuing a conception of the good life, while allowing other people to prosper in ways that satisfy no comparably important interest.
93.  
94. But if the distributive implications of the relational conception are to this extent the same in the case of a society of equals as they are in the case of egalitarian personal relationships, there are also important differences between the two cases. And these differences suggest that the first of the two doubts about the relational conception may get more of a purchase in the case of a society of equals. The central point is that some of the reasons for doubting whether the participants in egalitarian personal relationships would rely on any fixed distributive formula do not apply to a society of equals. For example, concerns about moralism and scorekeeping seem less significant in this case. And for the members of a society of equals, who lack the kind of direct deliberative access that the participants in an egalitarian personal relationship have, an “output measure” like a distributive formula, indirect though it is, may be the best way of judging whether the egalitarian deliberative constraint has been satisfied. For them, the fact that resources have been distributed equally may be the best available indicator that the comparably important interests of all of them constrained the processes of social decision to the same extent.
95.  
96. In addition, the anonymity of the relations among the members of society militates in favor of a clear public standard to govern the distribution of resources. Without the extensive mutual knowledge that is available at the level of personal relationships, the members of society are unable to engage in the kind of sensitive individualized consideration of one another’s interests that such knowledge makes possible. Instead, they need a clear public standard governing distribution: a standard they can all accept as an appropriate basis for judging whether, on the bounded but vitally important range of issues that concern them collectively as members, their shared egalitarian aspirations have been satisfied.
97.  
98. These considerations may serve to revive the first doubt about the relational view of equality. Their tendency, it seems, is to suggest that the ideal of a society of equals supports egalitarian distributive principles of some familiar kind. But we still need to determine which principles in particular egalitarians should accept, and that is precisely the question to which the distributive view is addressed. The suspicion, then, is that there is no conflict between the distributive and relational views; they are simply addressing different questions.
99.  
100. This suspicion is likely to be reinforced when one considers that the egalitarian deliberative constraint seems to underdetermine the choice among candidate distributive principles. As earlier observed, the deliberative constraint exerts general pressure in the direction of egalitarian distribution, and it provides a basis for rejecting nonegalitarian arrangements like the laissez-faire system of natural liberty. It also provides strong grounds for opposing systems of hereditary caste and privilege, and it vindicates the familiar complaint that we do not live in a society of equals if our laws and policies are shaped to a disproportionate degree by the interests of the rich and powerful. Beyond that, it is not clear that the deliberative constraint provides a basis for selecting among the different egalitarian distributive principles that have been proposed. Although the participants in egalitarian personal relationships may not need such a principle, a society of equals does need one: or, at any rate, it needs a principled public standard to regulate distribution and provide a shared basis for the justification of decisions made on behalf of the society as a whole. The inability of the egalitarian deliberative constraint to determine such a principle seems to confirm that it provides no alternative to the distributive conception.
101.  
102. I draw a different conclusion from the fact that the deliberative constraint underdetermines the choice among candidate distributive principles. Recall that the deliberative constraint is only one dimension of the broader relational ideal, the ideal of a relationship among equals. If it is unclear whether a given principle is compatible with the deliberative constraint, then the next question is whether the principle is consistent with the broader ideal. And if two different distributive principles both seem compatible with the deliberative constraint, then the question is whether either of them coheres better than the other with the idea of living together as equals. These are practical questions in the sense that, in order to answer them, we must consider what it would actually be like to carry on human relationships on the terms specified in the proposed principles.
103.  
104. Suppose, for example, that there is disagreement about whether hedonistic act-utilitarianism is compatible with the deliberative constraint. One side maintains that it is not, since hedonistic utilitarianism would permit a person’s fundamental interests to be sacrificed in order to maximize aggregate welfare. In such a case, this side argues, the interests of the person who undergoes the sacrifice are not exerting the same influence on social decision as the comparable interests of those who are not sacrificed. The other side argues, however, that the deliberative constraint is satisfied even in this case, because the fundamental interests of the person who undergoes the sacrifice are being assigned exactly the same weight in the overall hedonistic calculus as the comparable interests of others. It is simply this person’s bad fortune that his interests are outweighed while theirs are not.
105.  
106. On a relational view, the way to make progress in resolving this disagreement is to consider which side’s position can be more readily reconciled with the broader ideal of a society of equals. We saw earlier that, in such a society, members have a reciprocal commitment to treat one another with respect. They view one another as possessing the entitlements and responsibilities associated with full-fledged agency. No member is seen as possessing any more or less authority than the other members, except by virtue of a division of responsibility that all can accept, and each member sees the other members as entitled to participate fully and equally in determining the future course and character of their shared relationship. According to a relational view, the important question is whether utilitarian aggregation is compatible with this ideal of society. This is not merely a question of logical consistency but also a question about the human implications of living together on utilitarian terms. Once the question is framed in this way, it seems clear to me, though I will not try to argue the point here, that an unrestricted principle of utilitarian aggregation is incompatible with the ideal of a society of equals.
107.  
108. 1.6 Deeper Differences between the Distributive and Relational Views
109. Still, even the broader ideal may not fully determine the choice among candidate distributive principles. But this is not an objection to the relational view nor does it show that view to be normatively inert. Instead, it points to two deeper differences between the distributive and relational views. First, as we have seen, the relational view cannot be spelled out without reference to other values. According to this view, equality is a complex ideal whose distinctively egalitarian aspects cannot be identified, nor their appeal appreciated, independently of their connections with the other values, such as reciprocity and respect, that also help to define the ideal. This marks a subtle but significant contrast with the distributive view, which takes the normative content of the concept of equality to be located simply in the idea of an equal division, and appeals to other goods or values, such as welfare or resources or status or opportunity, only to determine the things to which the idea of an equal division should be applied. To be sure, the choice among these candidate equalisanda raises issues of principle concerning the role of values such as responsibility, liberty, and desert in determining the distributive implications of equality. Nevertheless, the idea that equality requires an equal division of something is common ground among versions of the distributive view that differ on these issues and so accept different equalisanda. They all take the normative content of the concept of equality to be exhausted by the idea of a division of some “currency” into equal amounts. In this sense, the distributive view, unlike the relational view, treats equality as a normatively autonomous value.
110.  
111. The second, complementary difference between the two views concerns the relevance of equality for questions of distributive justice. Both views agree that, while equality is one of the values that helps to fix the content of justice, it is justice rather than equality that provides the ultimate normative standard for assessing distribution. But consider again the remark of Cohen’s that I quoted earlier. Cohen’s position is that there is some currency of which justice requires people to have equal amounts, at least to the extent that this is not prohibited by the values that compete with equality in fixing the requirements of justice. This implies that equality alone suffices to give us a distributive formula, albeit one whose application may at times be limited because of conflicts with other values. It follows that, in addition to being normatively autonomous, equality is also distributively self-sufficient. Not only is its normative content exhausted by the idea of an equal division—by the idea there is something “people should have equal amounts of”—but, in addition, equality is capable all on its own of generating a presumptively authoritative principle of distribution, albeit one that may have to give way if, from the standpoint of justice, other conflicting values trump equality in some cases. Once again, the relational conception takes a different view. Not only do other values enter into the definition of equality, so that equality is not normatively autonomous, but, in addition, equality so understood need not by itself yield any fully determinate principle for regulating the distribution of resources, not even a presumptive or prima facie one. Although some candidate principles will be incompatible with the ideal of a society of equals, that ideal may not fully determine the choice of a single principle. This is unsurprising, according to the relational view, for there is no reason to expect equality to be distributively self-sufficient. The regulative principles governing distribution are the principles of distributive justice, and those principles are answerable to a range of values, of which equality is just one. None of these values need determine even a prima facie principle of distribution on its own. In Rawls’ representative formulation, the principles of justice specify the fair terms of cooperation for free and equal persons. This does not mean that we first establish what principles of distribution are required by equality and then ask to what extent the “competing” values of fairness, freedom, and cooperation restrict the application of that egalitarian principle. It means that justice is the virtue that tells us how the distribution of resources should be regulated so as jointly to accommodate all of these values. This is not to deny that distributions can be assessed as more or less egalitarian in some purely arithmetic sense. It is not, for example, to deny that we can use the Gini coefficient to measure income inequality. It is rather to assert that equality as a value, considered on its own and without reference to the other values that bear on justice, need not yield a fully determinate distributive principle that enjoys even prima facie authority.
112.  
113. It is tempting to conclude from this that the bearing of equality on issues of distributive justice is weaker on the relational view than it is on the distributive view. This would be a mistake. Consider, for example, the well-known criticisms of various “luck-egalitarian” principles as having unacceptably harsh or demeaning implications in some cases.9 One reply by defenders of luck-egalitarianism is to say that these criticisms do not show that it provides the wrong account of distributive equality, only that equality may be overridden by other values in some cases. From a relational perspective, however, this reply misses the distinctively egalitarian character of the criticisms. The harshness of unadorned luck-egalitarian principles is a reason for thinking that such principles are incompatible with the ideal of a society of equals, so that they are ruled out as unjust on specifically egalitarian grounds. Rather than speaking for luck-egalitarian principles, albeit not decisively, equality speaks decisively against them. Here it is the relational view rather than the distributive view that has clearer implications for justice.
114.  
115. If what I have been saying is correct, then the first doubt about the relational conception can be allayed even as it applies to the case of a society of equals. The relational conception understands the bearing of equality on issues of distribution very differently than does the distributive conception. In assessing a candidate distributive principle, the distributive conception will lead us to ask whether that principle has correctly identified the currency of which people should have equal amounts, whereas the relational conception will lead us to ask whether the principle sets out plausible terms for regulating the relations among the members of a society of equals. Although the general tendency of the relational conception is to support strong limits on allowable economic inequalities understood in purely arithmetic terms, and although the relational conception confirms the need for a public set of principles to regulate distribution, it insists that the requisite principles are given by justice and not by equality. Equality by itself need not determine a distributive principle with even presumptive authority. It is not normatively autonomous nor need it be distributively self-sufficient. Although it is conceivable that the same distributive principles will be selected no matter which conception of equality one begins with, there is no reason to expect this and offhand it seems unlikely. So if we wish to investigate the content of distributive justice, it matters which of these conceptions of equality we accept.
116.  
117. 1.7 Conclusion
118. As I said earlier, my sympathies lie for the most part with the relational conception. However, I have provided little in the way of direct argument in its favor. My aims have been more modest. I have tried to show two things. The first is that the relational conception is an independent conception of equality, which is not reducible to a version of the distributive conception. The second is that it makes a difference which of these two conceptions we accept. If we accept the distributive conception, we will see equality as a value that is essentially concerned with distribution and that, on its own, generates a distributive formula with presumptive authority. We will think it important to identify that formula, which we may see as providing the core of “an egalitarian conception of justice.” If we accept the relational conception, by contrast, we will see equality as a broad practical ideal governing the structure of human relationships, an ideal that itself draws on a variety of other values and that has a clear bearing on questions of distribution but does not yield determinate principles of distribution in isolation from other values. We will think it important to develop this ideal across a broad front. Insofar as we are concerned with social and political philosophy in particular, we will think it important to identify the kinds of practices and institutions we would have to create, and the kinds of attitudes and dispositions we would have to possess, in order for us to live in a genuine society of equals.