AMARTYA SEN’S THE IDEA OF JUSTICE: WHAT APPROACH, WHICH CAPABILITIES? Debra Satz* Much of Amartya Sen’s work has focused on inequality and poverty. Now he has written a substantial and provocative book about justice that attempts to reorient the overly idealized direction he believes political philosophy has taken.1 Animated by the avoidable suffering and destitution of the world’s most vulnerable people, The Idea of Justice argues that political philosophy should aim at clarifying and diagnosing evident injustices here and now. The book also covers a range of other important topics such as rationality, public reasoning, democratic deliberation, social choice theory, the limitations of social contract theories, the place of democracy in classical Indian thought, and the understanding of well-being, freedom, and equality. Many of these topics have also been longstanding themes of Sen’s work. In this paper, I focus on two of the book’s important topics: (1) Approach: Sen’s criticisms of what he sees as the practically useless theories of ideal justice, as opposed to his own contrasting comparative approach to social evaluation; and (2) Capability: Sen’s view of the place of the “capability” metric within that comparative approach.

* Debra Satz is the Marta Sutton Weeks Professor of Ethics in Society and Professor of Philosophy at Stanford University. She also directs the McCoy Family Center for Ethics in Society. 1. See AMARTYA SEN, THE IDEA OF JUSTICE (2009). An earlier version of this paper was presented at a symposium on Amartya Sen’s book The Idea of Justice, hosted by The Institute for Law and Philosophy at Rutgers University. Special thanks to Amartya Sen, John Oberdiek, who organized the symposium, and the other presenters and members of the audience for helpful comments. The first section of this paper draws heavily on a review I wrote of Sen’s book. See Debra Satz, Book Review: The Idea of Justice, 39 POL. THEORY 560 (2011) (reviewing AMARTYA SEN, THE IDEA OF JUSTICE (2009)).

277

278

RUTGERS LAW JOURNAL

[Vol. 43:277

I will use the phenomena of gender inequality, child labor, and unequal opportunities to show the power of Sen’s comparative approach and to point out some reasons (1) the contrast between “transcendental” justice and the comparative approach is too strong, and that, (2) for the purposes of social assessment, it is important to abstract from idiosyncratic valuations of capabilities. I. APPROACH Sen makes two powerful indictments of contemporary political philosophy. First, he argues that political philosophers spend an inordinate amount of time trying to ascertain the shape of a world with perfect justice.2 But theories of perfect justice—what Sen calls transcendental justice—are entirely redundant. We do not need a theory of what a perfectly just society would be like to know that it would be better if fewer people suffered from severe malnutrition and premature morbidity, if there were fewer child laborers, or if governments were not brutal and corrupt. Moreover, he argues that transcendental theories offer no practical guidance with respect to the choices we actually face: If we are trying to choose between a Picasso and a Dali, it is of no help to invoke a diagnosis . . . that the ideal picture in the world is a Mona Lisa. That may be interesting to hear, but it is neither here nor there in the choice 3 between a Dali and a Picasso.

Second, he claims that the quest for perfect justice leads philosophers to embrace an unobtainable standard of theoretical completeness and consistency.4 Perfect justice does not allow for conflicts about justice. Yet, conflicts about values, including the values involved in justice, are inevitable and cannot be eliminated even by the most impartial scrutiny and logical reasoning. Sen illustrates this point with a parable about three children, each of whom has a legitimate claim to a flute: one has made it, another can play it, and the third has nothing else to play with.5 In Sen’s view, there is no single answer as to the correct way to rank the claims of ownership, ability,

2. 3. 4. 5.

SEN, supra note 1, at xii. Id. at 16. Id. at 5–8. Id. at 2–15.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

279

or need. The drive for a final theoretical adjudication of all such value conflicts fuel endless battles with no possibility of resolution. Sen sharply contrasts the search for a transcendental theory of justice with his own comparative approach, which has strong affinities with social choice theory.6 On the comparative perspective about justice, the point is to provide a means of ranking alternatives that might be realized, not by appeal to perfect justice, but by exploring all of the values and priorities of the people involved in the context of empirical evidence. One critical device for such exploration is public discussion. Through open debate, we bring information from the lives of others into our own deliberations about which policies to choose. To make progress in these deliberations we do not need to appeal to perfect or complete justice. Sen points out that we already engage in such comparisons when we protest against the subjugation of women— comparing the lives of men and women—or advocate against premature mortality and morbidity—comparing the lives of those who die young of treatable and avoidable illnesses with those who live far longer lives— or “reject the quiet tolerance of chronic hunger”—comparing the malnourished with the well-fed.7 We make these comparisons in the light of evidence and discussion about what to choose. Sen emphasizes that the comparative approach is general and may be open-ended.8 Application of its procedure does not necessarily yield a complete ranking of all of the social choices that are possible. A ranking is complete if and only if for any two elements X and Y in an agent’s choice set can be ordered such that X is preferred or indifferent to Y by the individual. The comparative approach can yield incompleteness because we sometimes ambivalently judge options from conflicting points of view (as in the case of the flute), don’t have enough information about options to rank them (as in the case of many of our future actions), or reject ranking some items in terms of a common metric (as in the case of incommensurable goods). Additionally, context can sometimes lead people to change their rankings or outcomes. I might rank option X over option Y but reverse that ranking in the presence of a third option Z. For example, I might prefer to read an official state newspaper Pravda as opposed to the alternative of not reading it, but reverse my preferences if the state proceeds to ban all newspapers besides Pravda. 6. Id. at 106. See generally AMARTYA SEN, CHOICE, WELFARE AND MEASUREMENT (1997) (discussing social choice theory). Sen identifies the comparative approach with Adam Smith, J.S. Mill, and Karl Marx, among others. SEN, supra note 1, at 7. 7. SEN, supra note 1, at xi–xii. 8. Id. at 105.

280

RUTGERS LAW JOURNAL

[Vol. 43:277

While the search for perfect justice aspires to an unachievable completeness, the comparative approach aims to be true to the human condition. We cannot escape from plural and conflicting values both among diverse people and even within ourselves. Sen very plausibly argues that there is no one single principle for ranking all alternatives and no standard of justice to which everyone will agree.9 Here, Sen follows those such as Max Weber who note that we live out our lives among warring gods.10 Additionally, and more formally, social choice theory has shown that, even under very weak and reasonable sets of assumptions, it is quite easy to generate conflict in the ranking of alternatives among individuals. Sen employs this finding to knock down the pretensions of theorists who assume such conflicts away or claim they can always be transcended.11 Much contemporary political philosophy is, then, according to Sen, flawed by both its practical uselessness and its unrealistic standard of completeness.12 How compelling are these two objections? I am sympathetic to Sen’s concern that too much effort in academic political philosophy is sometimes spent on theoretical distinctions that make no practical difference in the service of a degree of implausible theoretical completeness. However, I think that the distinction he draws between transcendental justice and assessing injustice here and now is overdrawn.13 Some of our comparative judgments do—and I would argue must—appeal to considerations that might be associated with a “perfect” justice. Of course, we do not need a theory of perfect justice to condemn extreme poverty, premature and avoidable mortality, or state repression of those with religious beliefs as unjust. I doubt that this point was ever in dispute. The problem we face is that many plausible cases of injustice are much more complicated than these examples. It is precisely our perplexity in the face of opposing claims with respect to these more complicated and disputed cases that lead us to clarify our ideals, abstract away from contingent features of our circumstances, and attempt to bring our competing intuitions and judgments into equilibrium. 9. Id. at x, 12. 10. See MAX WEBER, Science as a Vocation, in MAX WEBER: ESSAYS IN SOCIOLOGY 147–48 (H. Gerth & C.W. Mills eds., 1958). 11. SEN, supra note 1, at 105. 12. Id. at 15–17. 13. Sen’s distinction between transcendental justice and comparative justice is importantly different from the distinction sometimes drawn between ideal and non-ideal theories of justice. Theories that are often labeled non-ideal because they premise non-full compliance or are fact sensitive can also be transcendental if they seek the perfect non-ideal theory. See SEN, supra note 1, at 62.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

281

Consider the question of gender inequality. There is a great deal of disagreement not only on how to characterize this inequality but also whether there is anything that is unjust about it. We will get very divergent results about what constitutes an “improvement” in women’s unequal social position depending on our conception of what we think the injustice of women’s inequality consists of. Is paying women for the care-giving work they do in the home an advance in terms of their position or an accommodation to gendered preferences? Should a society provide additional incentives for men to be involved in the labor of child-rearing and domestic care? We get different answers to these questions depending on our view of, among other things, what more ideal relationships between men and women should look like and what it would actually mean for men and women to have equal opportunities. Some people think that gender justice requires equal opportunities for men and women to assume different gender roles; some think that obtaining equal opportunity requires restructuring social institutions to do away with gender roles. Some people think ensuring equal opportunity for men and women means nothing more than removing formal legal barriers to entry into social positions. In evaluating specific policies, such as paying women for their household labor, people inevitably draw on their vision of what, ideally, achieving equality of opportunity between men and women would entail. Trying to figure out how to respond to such questions as those above presses us to move to a higher level of abstraction. By moving to this higher level, people who would otherwise disagree with one another might reach an agreement on options that should be rejected: for example, an agreement on why removal of formal barriers to professional entry is not sufficient for gender equality. And, importantly, an individual might, through a process of reflective equilibrium, clarify what she herself thinks. My point is that there are times when a person cannot figure out what constitutes a move to a superior, more just position without reflecting upon and working out her own conception of what “perfect” justice entails. There are times when we need a Mona Lisa to anchor our judgments about Picasso and Dali. It is useful in considering Sen’s criticisms of perfect justice to reflect on the Rawlsian approach to justice, since the difficulty of figuring out what to do in complex cases is Rawls’ point of departure. In A Theory of Justice, Rawls does not begin by trying to provide grounds for the very profound injustices that motivate Sen.14 Rather, he simply takes these injustices as given, as fixed points—e.g., slavery is wrong; extreme poverty and avoidable 14. JOHN RAWLS, A THEORY OF JUSTICE (rev. ed. 1999).

282

RUTGERS LAW JOURNAL

[Vol. 43:277

deprivation is wrong; the denial of religious freedom is wrong. Rawls does not offer a theory of justice designed to tell us that these practices are wrong or that they are unjust.15 He begins by taking the wrongness, the injustice, of these practices as his starting assumption.16 He treats the injustice of certain practices the way that Lincoln had earlier treated slavery—i.e., if slavery is not wrong, nothing is wrong. Thus, Rawls’s ideal theory was never meant to play a substantial independent role in condemning these practices, and so Sen’s charge that ideal theories of justice are “redundant” at least in this context, misses the mark. Rawls recognized, however, that there are many examples where intuitions and common sense views about whether a particular practice is unjust are confused or in conflict: for example, the just distribution of health care and the regulation of market generated inequalities There is no settled or widely shared consensus on these or other important matters.17 Accordingly, the project of Rawls’s theory is to “provide guidance where guidance is needed.”18 His theory organizes both familiar and basic ideas implicit in the public political culture of a democratic society19 and more abstract ideas in order to resolve points of contemporary controversy. For example, his theory of justice reveals to us that income and wealth inequalities have more in common with gender and race based discrimination than we might have supposed,20 leads us to see why formal equality of opportunity is a limited measure for redressing injustice,21 provides an argument for a certain minimum level of entitlement to citizens,22 and offers an alternative to other ways of looking at these issues such as utilitarianism and libertarianism.23 Rawls holds that the test for the plausibility of his theory is if, using our reflective powers of reasoning and bringing our judgments into equilibrium, we find his two principles of justice to be preferable to a set of given alternatives or, after engaging in wide reflective equilibrium, all the plausible alternatives that we can imagine.24 I actually think it is unhelpful to characterize the Rawlsian approach as transcendental if that is taken to entail that it is committed to a unique set of 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Id. at 17–18. Id. JOHN RAWLS, JUSTICE AS FAIRNESS: A RESTATEMENT 2 (2001). RAWLS, supra note 14, at 18. JOHN RAWLS, POLITICAL LIBERALISM 45–46 (1993). RAWLS, JUSTICE AS FAIRNESS, supra note 17, at 63. Id. Id. at 65. Id. at XVII. Id. at 506–13.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

283

principles of justice to which everyone will agree. In his later writings, Rawls is quite explicit that there might be other ways of satisfying the conditions of reciprocity besides his two principles and that many controversial issues will remain.25 At points, Sen acknowledges this but seems to think it defeats Rawls’s theory. But why exactly should the lack of unanimous agreement defeat a theory of justice? Is a theory of justice useless if we do not need it to condemn the extreme evils of our world and if it does not generate universal agreement? Theories of justice play many roles in individual and collective deliberation. They help us think through the consequences of our views, help us understand the costs and benefits of different principles, push us to make our views consistent with one another, provide bases for reconciliation with our society, and guide judgment on contested matters. One of the most important roles they play is providing us with systematic visions; roadmaps for different ways the world could be and to which we might reasonably aspire. Sen’s criticism of transcendental theory reminds me of the parable of the rabbi of Krakow. The Krakow rabbi interrupts his prayers with a wail that he has just seen the death of the rabbi of Warsaw two hundred miles away. A few days later some Jews from Krakow happen to travel to Warsaw where, to their surprise, they see the Warsaw rabbi healthy and delivering a lively talk on the Torah. After these same Jews return to Krakow rumors begin to spread among the rabbi’s congregation, and there is criticism and snickering among the skeptical. A few of the rabbi’s disciples rush to his defense claiming, although he was wrong on the details, “Nonetheless what vision!” If we think of perfect justice as comprising the attempt to organize a coherent vision focused somewhere at the outermost limits of the possible, then it can play a role in orienting people’s evaluation of their world, even if it cannot yield specific policy recommendations. In fact, it seems plausible that this is how the great political philosophers like Locke, Smith, Rousseau, Hobbes, Marx, and Mill have made their mark—and why we go back to them. It is not because of the details of their policy proposals or their close argumentation. It is, in the end, I think, because of their articulation of systematic visions of how to characterize a just society, visions that set a compass for people in their attempts to navigate complex cases. Martin Luther King used such a systematic vision when he appealed to the ideal of a society in which people would be judged only by the content of their character and not their skin color. That is not our world; we have considerable distance to go to reach this ideal in practice; and he did not need 25. RAWLS, POLITICAL LIBERALISM, supra note 19, at 152.

284

RUTGERS LAW JOURNAL

[Vol. 43:277

to appeal to this vision to criticize existing racial discrimination as unjust. Nonetheless, this ideal has played and continues to play a role in mobilizing people for racial justice and in setting a compass to guide our society’s actions, even in the face of the many practical obstacles hindering that goal.26 Now it is true that it is possible to inspire people without endorsing the existence of an “optimum point,” the perfectly just society. But I can think of three roles that the attempt to work out such an optimum plays. First, many of our judgments are in conflict with one another. There is pressure, both theoretical and practical, to resolve these conflicts. The idea of a perfectly just society acts as a regulative ideal in bringing our ideas into equilibrium. We move back and forth between visions of an ideally just society and our current beliefs and commitments to see what extent we can bring these into some kind of stable coherent system. Achieving this equilibrium will require sacrificing some beliefs, changing some commitments, and modifying some ideals. To decide what to change we need some points of pressure. Taking the implications of our views about justice to their logical conclusion is one such pressure point. Sen is correct to argue that there is unlikely to be unanimous agreement on a single theory of perfect justice.27 My point is simply that the drive for a “perfect” theory actually arises from within our everyday comparative and evaluative thinking about what we should do in the context of complicated cases. Because of this, the line between comparative assessment and ideal theorizing is less sharp than I believe that Sen takes it to be. Second, transcendental theories can play a valuable role given that some short-term improvements can conflict with long-term improvements as expressed in the catchphrase “one step forward, two steps backward.” Intentional beings employ such strategies where we embrace short-term, negative consequences as a means of achieving long-term, positive benefits. If we simply do pair-wise comparisons of states of the world, unguided by where we are trying to go, we may miss the possibility of long-term global optimization. Furthermore, since incremental improvements often set us on path dependent trajectories, we need to know what our ultimate goals are to be able to decide whether a particular improvement is warranted. There are many empirical cases where processes do not progress steadily toward some pre-determined, unique equilibrium but instead have the chance to reach one 26. Thanks to Josh Cohen for conversation about this issue. See also Sam Freeman, A New Theory of Justice, THE NEW YORK REVIEW OF BOOKS (Oct. 14, 2010) (reviewing AMARTYA SEN, THE IDEA OF JUSTICE (2009) available at http://www.nybooks.com/articles /archives/2010/oct/14/new-theory-justice/. 27. SEN, supra note 1, at 14–15.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

285

of several equilibriums. How we get to one of these equilibria depends on clearly identifying the goal of where we are going. An ideal theory gives us a point to aim at and along which to evaluate our progress. Third, there is a danger that when we are focused only on ranking social alternatives through the comparative method we will overlook possible alternatives that are not on the table. Consider the implications of realizing equality of opportunity in K-12 education. When we explore this question, we inevitably need to grapple with the issue of how much of the given world we should take for granted: children’s temperaments and personalities (themselves influenced by family origin, genetics, and early experiences), housing policies that make integration across class and race lines difficult, the existence of elite private secondary schooling, implicit racial bias, and so on. Many of these are features of our social world that current thinking largely takes as given. I believe that one important role for “transcendental” theorizing comes down to this: it is an invitation for us to imagine that less of our current political and social context is fixed than we tend to believe. The alternative is to think that we only have to consider the options currently on offer, evaluated for their short-term or immediate effects. Let me, at the risk of being redundant, give two final examples to illustrate this last claim. My first example involves existing debates about the legitimacy of child labor.28 Some organizations call for the immediate abolition of all child labor as a violation of the rights of the child. From their perspective, drawing distinctions between different types of child labor—hazardous versus nonhazardous, part time versus full time and so on—may seem pointless because anything short of full-time formal education for children is seen as a threat to their basic interests. But for others this perspective may appear useless: it offers little guidance on how it could actually be implemented, and it is potentially harmful because it seems to preclude making marginal improvements in children’s lives. But neither criticism is correct. In the first place, such “utopian” demands have a useful policy function. Articulating the wrong of child labor as a violation of human rights helps to create, legitimize, and reinforce social understandings about what children deserve. This function is true even if no such option is currently on offer. In the second place, as I mentioned in the context of discussing Rawls’s theory, there is nothing in this ideal perspective that prevents its advocates from

28. See DEBRA SATZ, WHY SOME THINGS SHOULD NOT BE FOR SALE: THE MORAL LIMITS OF MARKETS 155–69 (2010).

286

RUTGERS LAW JOURNAL

[Vol. 43:277

criticizing some child labor practices as so massively unjust that they require immediate attention. My second example concerns the abolition of slavery. Who would have imagined in the early 18th century that slavery could in fact be completely abolished? A small group of Englishmen went out on a limb and put forward the radical notion that slavery was wrong and must be immediately ended. At first, their proposal that the ideals of equality and liberty should be extended to the millions of African slaves held in Britain’s colonies looked like a nonstarter—an appeal to a perfect justice that would never be achieved. Nevertheless, slightly more than fifty years after its humble birth in a London print shop in 1787, the British antislavery movement overturned the economic backbone of the world’s most powerful empire.29 Although I accept Sen’s point that too much work in political philosophy is practically disconnected from current struggles and aspirations, I do not think we should regard all such work as unnecessary or redundant. We still do not know where to set the limits of the possible. II. CAPABILITIES I now want to turn to a second issue that has received somewhat less attention in the many illuminating discussions of Sen’s book. Theories of justice must specify a metric that characterizes the type of goods that they hold is subject to the demands of distributive justice. Continuing themes developed in his previous work, Sen argues against focusing on income and wealth as the basis of social assessment—the sole metric of justice with respect to distribution.30 In this book, as well as in his previous work, he has repeatedly stressed the importance of capabilities as a crucial input into our normative evaluations of social states.31 To sum up briefly: “capabilities” are derivatively defined from people’s potential “functionings,” where functionings are “beings and doings” The difference between a capability and a functioning is similar to the difference between an achievement and the freedom to achieve something or between an outcome and an opportunity for an outcome. Thus, being well nourished is a functioning, while the substantive opportunity to be well nourished is a capability. 29. See ADAM HOCHSCHILD, BURY THE CHAINS: PROPHETS AND REBELS IN THE FIGHT TO FREE AN EMPIRE’S SLAVES 1–8 (2005). 30. SEN, supra note 1, at 253. 31. Id. at 254.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

287

Although resources like income and wealth can be part of the means for achieving capabilities, the distinction between capabilities and resources is profound and consequential. Resource based theories do not acknowledge that people differ in their ability to convert their resources into capabilities due to individual, social, or environmental factors such as physical and mental disabilities, talents, traditions, social norms, legal rules, public goods, climate, infrastructure, and so on. This can be seen if we consider, again, the case of gender inequality. Resource based theories concentrate on redressing injustice through the distribution of goods: income and wealth, rights and liberties, and so on. But some kinds of gender injustices persist by informal means, such as stigma and stereotyping. An example of a stereotype is the assumption that women are the primary caretakers of their children. In the presence of such stereotyping, women will be able to do less than men, even if they have equal resources, since their ability to achieve on the basis of their resources will be less than men’s. Welfare based metrics of advantage and disadvantage are also limited in their ability to pick out salient features of gender injustice. If men and women are equally satisfied with their lives, then a welfare approach will judge their lives equally good. But women may be satisfied with far less than men in life because they have learned to accept less, feel that they deserve less, or have simply given up hoping for more. The extent of women’s actual deprivation may be invisible if we focus only on women’s welfare and not what they are actually able to achieve. It is hard to underestimate the importance of the shift from the focus on resources and welfare to that of capabilities. Not only does this complete a shift away from a narrow focus on “things” toward a focus on what people can do with these things, but it also places freedom at the center of our concerns. I agree with Sen that capability is an especially important metric to use when making comparative judgments about social states.32 But which capabilities should we appeal to in making such comparative judgments? This is the question I would like to pursue. There are perhaps an infinite number of functionings and thus an infinite number of capabilities. Surely not all of them are equally important. Why should we care about the capability a person has to eat at the most expensive restaurant in Paris? Or care about a person’s capability to experience fifteen

32. Id. at 231.

288

RUTGERS LAW JOURNAL

[Vol. 43:277

minutes of fame.33 One way of going about answering this question is given by Martha Nussbaum.34 Nussbaum has harnessed the capability approach to defend a single determinate list of valuable capabilities.35 In particular, she argues that a determinate list of capabilities serves as a fundamental metric when ranking social states as minimally adequate or not on the scale of justice.36 She also argues that this list needs to have universal applicability, even if we will need to interpret its general concepts differently in specific contexts.37 Because of this, the capabilities on Nussbaum’s list are defined abstractly: she is interested in general categories not in specific items that will vary from place to place.38 She thus attempts to ensure that her very wide ranging and ambitious list can be the basis of a plausible public conception of justice agreeable to reasonable people with varying values. Her list includes ten capabilities: life, bodily integrity, senses and imagination, emotions, practical reason, affiliation, relations to other species, play, and control over one’s environment.39 She argues that justice requires at the very least that individuals have access to at least a threshold amount of each of these.40 Sen’s approach is quite different than Nussbaum’s. First, he rejects the idea that the capability approach is the basis of a theory of justice, a set of principles that we have reason to endorse as the basis of our obligations to each other.41 Let me elaborate on this point. Because people have diverse values, people will endorse different rankings of capability sets. It is possible that some of the elements on Nussbaum’s list will not be on a particular person’s list; certainly, people will value the items on the list differently. Sen has advocated an “intersection approach” which attempts to construct an incomplete ranking of capability sets based on points of agreement among different conceptions.42 Capability set A is better than capability set B only when the divergent rankings of capability sets in a society converge on this judgment. But if different views of what has value offer conflicting rankings 33. See Richard Arneson, Two Cheers for Capabilities, in MEASURING JUSTICE: PRIMARY GOODS AND CAPABILITY 101 (Harry Brighouse & Ingrid Robeyns eds., 2010) (discussing trivial capabilities). 34. See MARTHA NUSSBAUM, SEX AND SOCIAL JUSTICE (1999). 35. Id. at 39–42. 36. Id. at 45. 37. Id. at 47. 38. Id. at 42. 39. Id. at 41–42. 40. Id. at 43. 41. SEN, supra note 1, at 232. 42. Id. at 242–43.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

289

of A and B then those sets are incommensurable. We cannot settle what is in the intersection approach except by seeing what people actually value in different decision contexts. We should not be surprised to see differences across contexts. Second, and relatedly, Sen does not stipulate which capabilities should be taken into account, or how different capabilities should be aggregated in an overall assessment. Lists of important capabilities will be generated for specific uses in the context of public discussion and debate over policy alternatives. Sen’s capability approach then does not specify which capabilities should be taken into account. However, Sen does suggest that, in thinking about capabilities, what is important is not so much the capabilities people merely want but those that they have reason to value.43 I’ll elaborate this below. Third, as I mentioned above, Sen insists that the capability approach is misunderstood when it is viewed as a theory of justice.44 The focus on capabilities is an attempt to correct for information that is excluded by other theories of justice. In their different ways, Sen thinks that important information is excluded by both Rawlsian social contract theory and utilitarianism.45 But it is not meant to substitute for either theory of justice, neither for utilitarianism nor, especially, for that offered by Rawls. Sen’s focus with respect to justice has been much more limited, with respect to each of these two theories: as we have seen, he hopes to bring needed and excluded information to the task of our assessment of social alternatives Capabilities are not the sole “metric” of justice:46 they cannot substitute for all the various considerations that must be weighed when we seek to make improvements in a society’s justice, including other political values such as participation, accountability, and efficiency. Sen’s discussion of the capability approach points to a gulf between its more modest informational role in enriching theories of social assessment and its development as a theory of social justice To be fair, Nussbaum does not see the capabilities approach as constituting a full theory of justice but rather as marking out a minimum standard of justice or baseline beneath which no individual should be allowed to fall.47 Nevertheless, she generates a list of ten specific capabilities that she believes are universals which fill this role. 43. 44. 45. 46. 47.

Id. at 233. Id. at 232. Id. at 279. Id. at 232. See MARTHA NUSSBAUM, WOMEN AND HUMAN DEVELOPMENT 75 (2000).

290

RUTGERS LAW JOURNAL

[Vol. 43:277

Without here endorsing Nussbaum’s thoughtful and carefully crafted list, I think that there are good reasons to hold that in the context of debates about justice we need more specification as to which capabilities matter.48 First, given the perhaps infinite number of capabilities, we need some guidance on where to focus our attention. Perhaps in extreme cases, this will be obvious, but many cases that are contested are not extreme. Second, and most critically, many capabilities that people care about are idiosyncratic. In particular, (1) Sen’s approach does not rule out a role for idiosyncratic capabilities from having a role in our discussions about justice; (2) however, justice demands a public criterion for assessing social disadvantage. Let us consider these points in turn. First, consider that Sen does not endorse a subjectivist approach when it comes to ranking social alternatives: he argues that a person’s valuations of her alternatives can be distorted; she may have adapted her preferences to unjust circumstances.49 For example, a person may place little value on being healthy if she has given up on the idea that achieving health is a realistic possibility for her. Instead of simply defining valuable capabilities subjectively, Sen argues that in social policy, a focus on capabilities should bring information about what people have “reason to value.”50 This gives some criteria for the inclusion of, and for the exclusion of, certain capabilities in making social comparisons. For example, it rules out simple counting up of the number of capabilities a person has in her set, irrespective of how she values those options. It also opens up a space between a person’s preferences and reasons. A person can have reasons not to choose what she now prefers. Some preferences are ill informed, malicious, or adaptations to injustice. A person can have reasons to choose things that she does not now prefer. But Sen’s approach does not rule out a person valuing idiosyncratic goals like achieving her fifteen minutes of fame or having the freedom to eat in the most expensive restaurant in Paris. In The Idea of Justice, Sen does not give us any reason why a metric that uses idiosyncratic capabilities should be rejected in favor of one that stresses general capabilities that are widely shared, such as the one that Nussbaum provided which I referred to above. This brings me to the second point. In ranking social alternatives in the context of making state or international policy, we need to move away from people’s individual idiosyncratic capabilities and focus on a more generic set 48. For example, not all the items on Nussbaum’s list seem equally important. 49. SEN, supra note 1, at 166–67. 50. Id. at 233.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

291

of capabilities. In this context, it is not enough to drive some wedge between capability valuation and preference. We also need a way of justifying a further restriction that rules out such personal and idiosyncratic valuations as inputs into social decision-making. It is critical for liberal societies to find justifications for public policies that abstract away from people’s personal, private conceptions of what is good. Given differing conceptions of value, it is unlikely that people will attach equal value to the freedoms such societies provide. These societies need some way to distinguish between the freedoms they provide their members and how valuable those freedoms are to individuals. They cannot do this if they tie freedom too closely to very specific preferences and values. For this reason, liberal theories of justice tend to use generic metrics of advantage and disadvantage. In Rawls’ own case, the metric involves primary goods that are meant to provide a generic, objective conception of citizen’s needs.51 Other conceptions use metrics of well being, income, or opportunities. With respect to public justification, treating idiosyncratic capabilities as inputs is misplaced, even if some individual has reason to value these capabilities. This problem is that capabilities grounded in such a manner do not provide any basis for what we owe to each other. The fact that I place a high value on a particular capability does not by itself give you any reason to provide me with the resources to obtain it. The basis for our obligations to each other arises from our (more or less) objective needs and not from the importance a particular person places on that need within their own particular view of their lives. Tim Scanlon gives an example that I think powerfully illustrates this point. 52 Assume, as most of us believe, that we have an obligation to prevent members of our society from starving; that is, that a society owes to its members a minimum sustenance. Now suppose that a starving person is not interested in sustenance, but he is desperate for money to build a monument to his god. It does not follow that society has an obligation to give this person funds for the monument. Claims of justice are precisely claims about what we owe each other; they are second person claims. They are not first person claims that we address to ourselves. I think that the upshot of this is that we should not base comparative social assessment in the context of justifying public policies on

51. RAWLS, POLITICAL LIBERALISM, supra note 19, at 187–90. 52. See T.M. Scanlon, Preference and Urgency, 72 J. PHIL. 655 (1975).

292

RUTGERS LAW JOURNAL

[Vol. 43:277

subjective valuations or idiosyncratic preferences.53 This is a reason to prefer a specified list of generic capabilities like Nussbaum’s over a non-specified list. Whether this list is defined through some combination of expert judgment and public debate54 or by a conception of what capabilities citizens need to function as equals in a scheme of cooperation,55 what is important is that we look beyond people’s subjective evaluations. Consider the issue of how we should understand equality of opportunity in the case of K-12 education.56 Which capabilities should we focus on in trying to interpret this idea? Unwieldy as the idea of equality of opportunity undoubtedly is, it cannot possibly be taken to include developing every capability that a person values—there is no apparent obligation for the State to provide free ballet lessons to anyone who asks. Any plausible conception of equality of opportunity in education will have to appeal to generic capabilities—like numeracy and literacy—otherwise it will be hostage to some person’s (e.g., parents) expensive tastes and require information that is simply not available. To respond to the expensive taste and complexity objections, I think that we will need to focus on a set of generic capabilities. In making this point, I am really suggesting that Sen should further integrate his discussion of public reason with his discussion of what it means to appeal to “capabilities that a person has reason to value.57” If I am a foodloving individual, then I have reason to value the ability to eat in the most expensive restaurant in Paris. But, my desire to realize this capability has no place in a public assessment of social alternatives which must appeal to reasons that can be acceptable even to people who dislike fine dining experiences. We need to distinguish between what people idiosyncratically care about and the nature of our obligations to one another as members of a shared society.

53. I do not think that the intersection approach, by itself, will always take care of such cases, especially when such valuations are widely shared but nonetheless controversial. But I acknowledge that the intersection approach is likely to be a filter on some idiosyncratic tastes. 54. See Martha Nussbaum, Capabilities as Fundamental Entitlements: Sen and Social Justice, 9 FEMINIST ECON. 33, 40–42, 47 (2003). 55. See Elizabeth Anderson, What is the Point of Equality, 109 ETHICS 289 (1999). 56. Leave aside, for a moment, the fact that in the case of children we should be focused on providing functionings and not capabilities. 57. SEN, supra note 1, at 233.

2012]

WHAT APPROACH, WHAT CAPABILITIES?

293

III. CONCLUSION Both the capability metric and the comparative approach have numerous advantages. The capability metric provides critical information lacking from other approaches to social justice. It shifts the focus from resources and things to the much more important question of what people can do and be on the basis of those things. It brings into focus the many ways that people can be disadvantaged—by stigma and stereotype—that are not evident within a metric focusing on things. The capability approach is also superior to views that focus solely on utility. It brings into focus the ways that a person who can achieve little may be happy with little. At the same time, I have argued that Sen’s approach faces an important challenge due to its underspecified character. It cannot distinguish between idiosyncratic capabilities valued by individuals and capabilities that give rise to obligations on the part of others. While Sen may be correct to emphasize the importance of capabilities to a theory of justice, I think Rawls was right to attempt to define advantage and disadvantage using a metric that is objectively—or at least intersubjectively—justified. The comparative approach to social justice that Sen defends has many attractive features. It places democratic deliberation at the center of social choice, has the potential to include many diverse perspectives including those of outsiders, and recognizes the persistence of reasonable disagreement and the need for any public criterion of justice to accommodate that disagreement. Most importantly, it presses us to think about justice here and now. At the same time, I have argued that the line between the here and now and the ideal is not fixed and that comparative theorizing must make use of perfect justice in circumstances of path-dependent choices or when justifying indirect strategies. I close by noting that my preferred approach draws something from Rawls and from Sen. Sometimes taking an ecumenical and catholic stance is a way of having your cake and eating it too—an opportunistic appeal to competing goods. I hope that is not the case here. I have learned powerfully from both Rawls and Sen, and I hope I will be excused if I cannot generate a complete ranking of their contributions.